Water Resources of Arid Areas

WATER RESOURCES OF ARID AREAS

PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON WATER RESOURCES OF ARID AND SEMI ARID REGIONS OF AFRICA (WRASRA), AUGUST 3–6TH 2004, GABORONE, BOTSWANA

Water Resources of Arid Areas Edited by

D.Stephenson Civil Engineering Department, University of Botswana, Gaborone, Botswana E.M.Shemang & T.R.Chaoka Department of Geology, University of Botswana, Gaborone, Botswana

A.A.BALKEMA PUBLISHERS LEIDEN/LONDON/NEW YORK/PHILADELPHIA/SINGAPORE

Copyright © 2004 Taylor & Francis Group plc, London, UK All rights reserved. No part of this publication or the information contained herein may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, by photocopying, recording or otherwise, without written prior permission from the publisher. Although all care is taken to ensure the integrity and quality of this publication and the information herein, no responsibility is assumed by the publishers nor the author for any damage to property or persons as a result of operation or use of this publication and/or the information contained herein. Published by: A.A.Balkema Publishers, a member of Taylor & Francis Group plc http://www.balkema.nl/ and http://www.tandf.co.uk/ This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to http://www.ebookstore.tandf.co.uk/.” ISBN 0-203-02340-4 Master e-book ISBN

ISBN 04 1535 913 9 (Print Edition)

Table of Contents Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

Preface

xii

Keynote address Sustainable water management in arid and semi-arid environments W.Kinzelbach, P.Bauer, P.Brunner & T.Siegfried

3

Theme A: Problems in obtaining hydrological and geo-hydrological data Slug tests in fractured rock formations: value, pitfalls and misinterpretations P.D.Vermeulen & G.J.van Tonder Flow simulation model performance assessment using entropy approach A.M.Ilunga & D.Stephenson Data collection experiences in water level monitoring, borehole archive and research projects in semi arid Botswana M.Magowe, T.Obakeng & P.Makobo Rainfall characteristics in semi-arid Kitui district of Kenya A.O.Opere, V.O.Awuor, S.O.Kooke & W.O.Omoto Quantification of the impact of irrigation on the aquifer under the Vaalharts Irrigation Scheme R.G.Ellington, B.H.Usher & G.J.van Tonder

21

29

36

43 60

Theme B: Groundwater recharge: natural and artificial Groundwater development—identification of artificial recharge areas in Alla, Eritrea K.S.Viswanatham, F.Tesfaslasie, M.Asmellash, A.Kumar & S.A.Drury

75

Subterraneous injection of nutrient rich groundwater to the coastal waters K.K.Balachandran & J.S.Paimpillil A new method for the estimation of episodic recharge J.Bean, G.van Tonder & I.Dennis Prioritisation of the impacts of pollutants on groundwater flow systems in South Africa I.Dennis, B.Usher & J.Pretorius Understanding problems of low recharge and low yield in boreholes: an example from Ghana A.J.E.Cobbing & J.Davies Spatial variation of groundwater recharge in semi-arid environment—Serowe, Botswana L.M.Magombedze, B.Frengstad & M.W.Lubczynski Quantification of artificial ground water recharge G.C.Mishra The architecture and application of the South African Groundwater Decision Tool I.Dennis & G.J.van Tonder The development of a groundwater management tool for the Schoonspruit dolomitic compartment B.H.Usher & S.Veltman Effects of mining and urban expansion on groundwater quality in Francistown, Botswana B.Mafa & H.Vogel In situ remediation potential for Southern African groundwater resources S.Clarke, G.Tredoux & P.Engelbrecht Coastal aquifers intrusion at semi-arid region of Turkey L.Yilmaz Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region Y.B.Mkwizu & H.H.Nkotagu

85

92 99

109

122

133 145

156

168

181

191 198

Theme C: Socio-economic aspects

KNUST experiences in capacity building in the water and sanitation sector S.N.Odai, F.O.K.Anyemedu, S.Oduro-Kwarteng & K.B.Nyarko

213

Strategic partnerships for sustainable water education and research in developing countries S.N.Odai, K.A.Andam & N.Trifunovic Assessing demand for clean and safe domestic water in eastern Zimbabwe E.Manzungu, M.Machingambi & R.Machiridza The role of supplementary irrigation for food production in a semiarid country—Palestine M.Y.Sbeih Conversion of priority water rights to proportional water permits and conflict management in the Mupfure river catchment, Zimbabwe T.Mpala Impacts of water development in arid lands of Southern Africa: socio-economic issues J.P.Msangi Institutional challenges for small towns’ water supply delivery in Ghana K.B.Nyarko Socio-economic performance of Sepeteri irrigation project in Nigeria O.O.Olubode-Awosola & E.O.Idowu

221

227

240

254

262

275

287

Theme D: Application of geophysical, GIS, and remote sensing techniques

Mapping vegetation for upscaling transpiration using highresolution optical satellite and aircraft images in Serowe, Botswana Y.A.Hussin, D.C.Chavarro, M.Lubczynski & O.Obakeng Gravity study on groundwater structure in Central Butana (Sudan) K.M.Kheiralla & A.E.Ibrahim Remote sensing and electrical resistivity studies on groundwater structure zones in Central Butana (Sudan) K.M.Kheiralla & A.E.Ibrahim Monitoring and modeling of fluxes on Kalahari—setup and strategy of the Kalahari Monitoring project. Serowe study case, Botswana M.W.Lubczynski & O.Obakeng Geoelectrical investigation for aquifer delineation in the semi-arid Chad Basin, Nigeria A.Iliya & E.M.Shemang

302

313 330

346

357

Monitoring of evapotranspiration on Kalahari, Serowe case study, Botswana O.Obakeng & M.W.Lubczynski Electro-seismic survey system S.R.Dennis, M.du Preez & G.J.van Tonder Borehole site investigations in volcanic rocks of Lolmolok area, Samburu district, Kenya J.K.Mulwa Groundwater evaluation in a complex hydrogeological environment—a GIS based approach B.Mudzingwa, J.L.Farr, R.Gumiremhete & T.Kellner Application of 2-D resistivity imaging combined with time domain electromagnetic survey to map shallower aquifers in Kunyere valley, northwest Botswana E.M.Shemang, H.Kumar & J.Ntsatsi

364

381 388

406

420

Theme E: Climate change and its impact Hydraulic studies in the design of sand dams A.S.Nzaba, H.O.Farah, T.C.Sharma & C.W.M.Sitters Designing and implementing an aircraft survey mission using highresolution digital multi-spectral camera for vegetation mapping for upscaling transpiration of Serowe, Botswana Y.A.Hussin, M.W.Lubczynski, O.Obakeng & D.C.Chavarro Relevance of groundwater interaction with surface water to the eco-hydrology of semi-arid regions J.Y.Diiwu Impacts of climate change in water resources planning and management A.Opere Turning a liability into an asset: the case for South African coalmine waters B.H.Usher & F.D.I.Hodgson Environmental hydrogeology of the dolomite aquifer in Ramotswa, Botswana M.Staudt & H.Vogel Investigation of natural enrichment processes of nitrate in soil and groundwater of semi-arid regions: case study—Botswana S.Stadler, M.von Hoyer, W.H.M.Duijnisveld, T.Himmelsbach, M.Schwiede, J.Böttcher & H.Hötzl Hydroclimatological approach to sustainable water resources management in semi arid regions of Africa U.T.Umoh

430 442

450

461

467

479

489

505

Impact of cultivation practices on multiple uses of water in the Alemaya catchment, eastern Ethiopia Y.E.Woyessa & A.T.P.Bennie Geochemical evidence and origin of salinity in the shallow basinal brine from the Makgadikgadi Pans Complexes, northeastern Botswana L.N.Molwalefhe

514

528

Theme F: Vulnerability and risk Decision support for optimal water system planning: a Wadessy case study A.A.Ilemobade & D.Stephenson The importance of constructing a correct conceptual model for an aquifer G.van Tonder, I.Dennis & D.Vermeulen Water resources development and risk assessment in mountain regions of Africa H.Scheuerlein Reliability, resilience and vulnerability for reservoir sizing and operation J.G.Ndiritu Hydrological impact of dam construction in an arid area D.Stephenson & Z.Chengeta The geochemistry of fresh water supplies in Botswana L.Molwalefhe & S.Vriend Groundwater modelling with limited data: a case study of Yobe River Basin, North East Nigeria M.Hassan, R.C.Carter & K.R.Rushton

541

551

559

572

580 589 600

Theme G: Water resources management Apple and grape vinegar application as c-source in water denitrification Ş.Aslan & A.Türkman

613

Water resources management in the National Park, central Australia E.R.Rooke Integrated water resources management and agriculture in southern Africa M.McCartney, H.Sally & A.Senzanje

623

634

Challenges for managing water resources in semi-arid areas: a case study from two rural communities in Zimbabwe F.T.Mugabe & A.Senzanje An Integrated Water Resources Management tool for Southern Africa allowing low flow estimation at ungauged sites M.J.Fry, S.S.Folwell, H.A.Houghton-Carr & Z.B.Uka Organization of water services in Malawi and strengths and weaknesses in implementing Integrated Water Resources Management (IWRM) M.Selemani Towards best water resources management practice in small town water supply system in Tanzania A.Mvungi & M.Makuya Water management in the Mauritian textile wet processing industry N.Kistamah & S.Roseunee Analysis of the microbiological situation of the quality of domestic water sources and identification of the microorganisms in them, located in the semi-arid regions of the Eastern Cape, South Africa M.Zamxaka, G.Pironcheva & N.Y.O.Muyima Dry season Kalahari sap flow measurements for tree transpiration mapping—Serowe study case, Botswana M.W.Lubczynski, A.Fregoso, W.Mapanda, C.Ziwa, M.Keeletsang, D.C.Chavarro & O.Obakeng Heavy metals and radioactivity in the groundwater of Khartoum State, Sudan A.M.Ahmed Impediments to the effective implementation of a groundwater quality Protection strategy in Botswana T.R.Chaoka, E.M.Shemang, B.F.Alemaw & O.Totolo Spatial assessment of groundwater pollution vulnerability of the Kanye wellfield in SE Botswana B.F.Alemaw, E.M.Shemang & T.R.Chaoka The effect of socio-economic activities on watershed management: the case study of Gaborone Dam catchment in Botswana G.S.Thabeng & D.B.Kemiso

Author index

643

650

661

666

678 685

693

702

707

718

725

737

Preface Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

Africa’s water resources are threatened by population growth with the resultant increase in water demand, the stresses of water use for various activities, desertification, global warming and climate change, and other interventions in the water cycle by man. These effects are more pronounced in the Arid and Semi-Arid regions of Africa in particular and the world in general. It is therefore vitally important that the water resources in Arid and Semi-Arid regions are developed and managed in a sustainable and integrated manner. Integrated management of water resources in the arid and semi-arid regions of Africa requires a spectrum of efforts from local and community stakeholders to national and transboundary river basin management. This conference aims at sharing the best practices of water use and conservation around the globe. The main objective of this conference was therefore to bring together educators, researchers, practitioners, managers, policy makers and NGO’s from Africa in particular and the world in general involved in various aspects of water resources in arid and semi arid regions. The more specific objectives of the conference were to (i) Assess the current state of the art of water resources management in arid and semi arid regions with particular emphasis on African regions. (ii) Address the future water stress due to limited water resources, population growth, increasing demand and pollution and other related risks, resulting in insufficient water supply. (iii) Promote dialog and interaction between different disciplines and professions. (iv) Forster insights into issues of global sustainable development and set concrete targets to meet the need for drinking water and water borne sanitation in arid and semi arid countries of Africa and the world in general. We received an overwhelming response to our call for papers. We received over 120 abstracts and each abstract was reviewed and more than two thirds of the abstracts were accepted. Authors were then requested to submit full text of papers. The full texts of the papers were reviewed by the conference organizing committee and 68 papers were finally accepted for conference. The papers in this book “Water Resources of Arid and Semi-Arid Regions of Africa” constitute the conference proceedings. This book is subdivided into seven sections. Section 1 deals with problems in obtaining data. Section 2 deals with groundwater recharge: natural and artificial; Section 3 deals with Socio economic aspects of water demand management; Section 4 deals with geophysical, GIS and remote sensing

techniques for groundwater exploration; Section 5 deals with climate change and its impact on water resources; Section 6 deals with vulnerability and risk assessment and Section 7 water management. This book will be of interest to researchers and practitioners in the field of surface water hydrology, groundwater hydrology, environmental engineering, agricultural engineering and earth sciences, as well as those engaged in water resources planning, development and management in arid and semi arid areas. Graduate students and those wishing to conduct research in hydrology, environmental science and engineering and water resources will find the book to be of value. Dr A.R.Tombale Permanent Secretary Ministry of Minerals, Energy and Water Resources, Botswana

Keynote address

Sustainable water management in arid and semi-arid environments W.Kinzelbach, P.Bauer, P.Brunner & T.Siegfried Institute for Hydromechanics and Water Resources Management, Swiss Federal Institute of Technology, Zurich, Switzerland Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Scarcity of water often leads to its non-sustainable use. The globally most widely spread non-sustainable practices are overpumping of aquifers, drying up of wetlands and soil salination on irrigated land. Only with much more careful management of scarce resources sustainability in the long term can be reached. Modeling is a valuable tool in the analysis of management options and scenarios. New types of data from remote sensing, airborne geophysics, and environmental tracers to name a few allow reaching a new quality of prediction. Three field studies illustrate the points.

1 INTRODUCTION Fresh water is a scarce resource on a worldwide basis. This becomes apparent when looking at the basic items of the global freshwater balance (Postel et al., 1996). Of the 110,000km3/a of precipitation on the landmass of the earth, 50,000 are returned to the atmosphere via evapotranspiration by the planet’s natural plant cover. Another 21,000 are used by man-made ecosystems (18,000km3/a by rain fed agriculture and 3,000km3/a by irrigated agriculture). This shows that agriculture and natural vegetation are already fierce competitors for the available freshwater. Of the accessible runoff of 13,000km3/a about 4,000 are appropriated by mankind. 70% of those go into irrigated agriculture. This means that a global water crisis would above all be a global crisis in food production. Compared to the 13,000km3/a available the abstracted 4,000 appear small. One should, however, not forget that these figures are averaged in time and space and therefore hide the real problem, e.g. droughts and floods. One can still use the ratio a of withdrawals and available renewable resources as an indicator. Due to the variability of the quantities involved, it is the experience that a value of a>0.4 already reflects severe scarcity. On a global scale a=0.31 is found (Alcamo et al., 2003). This indicates that scarcity on a global level is a reality today, with the arid world already experiencing very severe scarcity problems.

Water resources of arid areas

4

2 SUSTAINABILITY Sustainable water management is a practice, which avoids irreversible or quasiirreversible damage to the resource water and other natural resources linked to it, such as soil and ecosystems. Such a practice conserves the ability of the resource water to provide its services including ecological services. Water scarcity and poverty are often the causes of non-sustainable behavior as they lead to overexploitation and depletion of stocks. What are the relevant issues for global sustainability in the water sector? To identify big and possibly existential problems for whole regions we have to look for ubiquitous negative global trends. In that sense there are a number of non-sustainable practices, which are of global importance. Above all, these are – the overpumping of aquifers, – the destruction of wetlands, – the salination of soils, and – the pollution of aquifers with persistent pollutants. Globally about 800km3/a of freshwater are abstracted from aquifers. About one quarter of this abstraction is non-sustainable in the sense that it is not replaced by recharge, i.e. it is taken out of the available stock. On the Arabian Peninsula, in North Africa, China and the arid Western United States for example, abstractions for large-scale irrigation have withdrawn large quantities of fossil water, which under present climatic conditions are not replenished any more. The global area of wetlands has diminished by 50% since the year 1900. This has a dramatic impact on species diversity. It is a consequence of the competition between natural and man-made ecosystems for land and water resources. The tendency is unbroken. Of the 260 million ha of irrigated farmland in the world about 80 million are affected by soil salination. Salination is a common phenomenon in hot climates. It occurs if in a soil more salt is deposited by evapotranspiration than is removed by drainage. In irrigated agriculture, the most common mechanism leading to salination is the groundwater table rise due to seepage of irrigation water. Once a groundwater table is closer than 2m to ground level, capillary rise leads to direct evaporation from groundwater and to fast salination of the topsoil. Finally, there is the deterioration of groundwater quality by persistent pollutants. One might expect that among those chlorinated hydrocarbons are the most important. This is only true for industrialized countries while globally the most prominent pollutant is salt, especially in arid regions and coastal areas, where seawater intrusion occurs. In principle, all these violations of sustainability are reversible. But the required timescales are on the order of several generations. For all practical purposes these damages are irreversible. In the following, three examples from projects of the Institute of Hydromechanics and Water Resources Management are shown, illustrating the first three globally important sustainability problems. The common features of these examples are that – there is water scarcity (all three areas are in arid or semi-arid climate zones), – a model is developed to analyze and understand the system,

Sustainable water management in arid and semi-arid environments

5

– the model is used for the testing of management strategies and/or for optimization, and finally – the connection is made to the field of socio-economics.

3 OVERPUMPING OF THE NORTH-WEST SAHARA AQUIFER SYSTEM The North Western Sahara is underlain by one of the world’s largest aquifer systems, which covers approximately 1,000,000km2 and consists of two major aquifers, the deeper Continental Intercalaire (CI) and the shallower Complexe Terminal (CT) (Fig. 1). Their water resources are being utilized by the three countries Algeria, Tunisia and Libya mainly for irrigation purposes. The system nowadays hardly receives any recharge. At most 30m3/s are estimated as recharge

Figure 1. Overview of the NorthWestern Sahara aquifer system and its water balance. along the southern flank of the Saharan Atlas where the aquifers strike out (ERESS, 1972). Compared to the size of the system, this recharge flux is—if at all—only of importance locally. The system discharges mainly via the sink of the Chotts or salt lakes, which are the topographic lows of the endorrheic basin. Here approximately 10m3/s evaporate. A very small portion of no more than 5m3/s is thought to discharge to the sea in Libya. Until 1950 abstractions were small. Since then the population has tripled and with it the amount of water pumped for irrigation. An estimated rate of 180m3/s is abstracted today. As a consequence the large springs in the vicinity of the Chotts have run dry (Fig. 2). Artesianism has vanished over large areas and the water, which before flowed at no energy cost, now has to be pumped.

Water resources of arid areas

6

The present situation is characterized by an abstraction, which is 6 times as large as the recharge rate. This brings up the question whether a non-renewable resource should be used at all. Looking at the size of the system and its storage coefficient, an enormous amount of about 100000km3 of water is stored. About one tenth of that amount can be accessed with an economically feasible drawdown of less than 250m. With a projected future withdrawal rate of 500m3/s the total resource can still last for about 600 years. But this water comes at a price. First, energy is necessary for its pumping and distribution and investments in pipes and boreholes have to be made. Second, pumping can lead to deterioration of water quality. Sources of pollution are various. Near the Chotts for example, large drawdowns will reverse the hydraulic gradient, which under natural conditions is always directed from the oases to the Chotts. A reversed gradient mobilizes brine, which finally leads to degradation of the water quality pumped and contributes to the die-off of oases. A similar phenomenon is observed along the coast, where overpumping leads

Figure 2. Development of discharges from springs in Southern Tunisia (1887–1985) (Source: Mamou, 1990)

Sustainable water management in arid and semi-arid environments

Figure 3. Modeled head distribution in CI, 1950 (heads from 530 mamsl to 70 in steps of 35m).

Figure 4. Predicted head distribution in CI, 2050 for planned pumping (heads from 530 mamsl to –250 mamsl in steps of 55m).

7

Water resources of arid areas

8

to seawater intrusion. Salt water can also be mobilized from lower saline aquifers such as the Turonian. With a reduced pressure overburden this confined aquifer can infiltrate at a larger rate into the CT from below. A numerical model of the system has been built which demonstrates that with the required total pumping rate by 2050 large areas of the presently strongly pumped regions will face a piezometric decline with economically infeasible drawdowns of more than 250m below ground level (area with sawtooth pattern in Fig. 4). At the same time the constraints for water quality locally can no longer be fulfilled. The groundwater model was then coupled with optimization algorithms to find allocation patterns that conform to demand, drawdown and quality constraints in time while minimizing overall provision costs. The wells in an optimal scheme spread out over the area to equilibrate distribution cost with pumping cost, which depend on drawdown. They further spread to the CI from the CT. Two variants were analyzed. In the first, the existing pumping locations were used and the pumping rates at those constituted the set of decision variables. At the Chott cells, gradient constraints were introduced to prevent gradient reversal and thus preserve the productivity of the oases. On the whole the costs are exploding over time, with the running cost of water increasing by a factor of about 30 in 50 years (Siegfried, 2003). In a second variant, pumping at any location was allowed with the costs being the only criterion for choosing a specific cell. The results show that compared to the first variant much better abstraction schemes are possible with considerably lower running cost (and total costs) over the next 50 years. However interesting such a scenario is, it would require a complete renewal of infrastructure. Realistically, only a gradual transition from today’s pumping well distribution to a more favorable one in the future will be feasible. The model demonstrates that it is possible to minimize pumping cost to reasonable levels and provide water for the next 50 years. This time however must be used to develop alternatives. All optimization runs were carried out ignoring national borders in order to assess benefits from cooperative management. As the results demonstrate, cooperation between the three countries involved brings considerable advantage in the exploitation of the resource. Nevertheless, in the long run the conservation of the oasis culture requires heavy subsidies as the substitution between the production factors water and capital progresses. 4 MANAGEMENT OF THE OKAVANGO DELTA, BOTSWANA The Okavango River flows from the Benguela plateau of Southern Angola in southeastern direction through the northern tip of Namibia and then into Botswana, where it forms an inland

Sustainable water management in arid and semi-arid environments

9

Figure 5. Satellite image of the Okavango Delta (length from right to left about 550km). Delta in which it is consumed completely by evapotranspiration (Fig. 5). The Delta is one of the largest wildlife areas in Africa and is an attraction for numerous tourists. The yearly floods of the river turn a large area of the Delta into a seasonal swamp (Hutchins et al., 1976; Scholz et al., 1976; McCarthy et al., 1986; Thomas and Shaw, 1991; Ellery et al., 1993; McCarthy et al., 1993; McCarthy and Ellery, 1994; Modisi et al., 2000; Gumbricht et al., 2001). As the flood takes 3 months to propagate from the inflow at Mohembo to the distal side at Maun, it is out of phase with the local rainy season and thus increases the water availability over the year. The upstream countries are discussing plans to abstract water from the river and/or build dams for electricity production or agricultural purposes. In Botswana itself, various sectors of the economy have also proposed to make use of the Okavango water, be it for agriculture or for mining purposes. All measures proposed threaten the existence of the Delta as the unique ecosystems it is. Both abstraction of water in the upstream and acceleration of the through-flow by dredging of channels etc. will cause a decrease in the size of the seasonal swamp. In order to assess the impact of hydraulic measures on the size and distribution of the flooded area a numerical model was constructed which contains the surface water and the groundwater in two coupled layers. In an innovative approach satellite data on the time-varying size of the Delta were used to calibrate the model (McCarthy et al., 2003; Bauer et al., 2004). Further data used in this approach are a high-resolution digital terrain model obtained from the flooding patterns and the related vegetation patterns (Gumbricht et al., 2001, 2003), the inflow at Mohembo, the precipitation from METEOSAT data (Herman et al., 1997), the

Water resources of arid areas

10

evapotranspiration from multi-spectral satellite data (Bastiaanssen et al., 1998a, 1998b), and last not least local measurements which are routinely performed by the Botswana Department of Water Affairs. The model is able to reproduce satisfactorily the seasonal dynamics of the flooded areas both in total extent and in distribution over a period of 20 years for which data are available (Fig. 6). The sink of all water is essentially the evapotranspiration by the plant cover. This process also governs the distribution of salts in the Delta. Pronounced salt crusts indicate areas, which are natural disposal sites of salts. Their continued functioning is of considerable importance to the conservation of the Delta. This process will be incorporated in a future version of the model. One example of measures with potentially serious impacts on the Delta is the abstraction of water upstream of the inflow (Fig. 7). It is seen that an abstraction is amplified i.e. the relative reduction in area is considerably larger than the relative reduction in inflow. Dams have an effect on both inflow reduction and temporal inflow distribution. Model calculations showed that the change in input distribution not necessarily is detrimental to the size of the flooded area. A more stretched out flood will bring water further downstream. Morphological changes such as dredging of channels and removal of blockages by papyrus have also a pronounced effect, not so much on the total flooded area as on the distribution of flooded areas within the Delta (Bauer et al., 2004).

Figure 6. Observed and modeled flooding frequency (%).

Sustainable water management in arid and semi-arid environments

11

Figure 7. Flooded area for different abstraction scenarios in comparison with the modeled development of the last 20 years. The local abstraction for household consumption, be it directly or from the aquifer, is so small that it will at no stage be of relevance for the Delta. The tentative ranking of different interventions according to their severity is as follows: – abstractions larger than 2m3/s in the upstream, – building of large dams in the upstream, – change to a drier climate, – morphological changes (dredging, cutting of vegetation, tectonics), – local drinking water supply. The model can provide a quantitative basis for the political debate between the three riparian nations. It is clear that the conservation of the Delta must bring some revenue to the upstream in exchange for the guaranteed inflow. The key parameter for an administrable negotiated solution will be the minimum inflow at Mohembo and its seasonal variability. 5 SALINATION OF SOILS AND WATER IN YANQI BASIN, XINJIANG, CHINA The third example studies a region in China’s arid west. The Yanqi basin is formed by the lowlands of the Kaidu River and Lake Bostan (Fig. 8). The area has been used intensively for agriculture over the past 50 years. The main products are grapes, cotton and red peppers.

Water resources of arid areas

12

Figure 8. Satellite map of the Yanqi basin showing the irrigation areas along the Kaidu River, Lake Bostan and the Kongque River. As precipitation is only 70mm/a and thus negligible compared to the potential evaporation of 1800mm/a, no agriculture is possible without irrigation. The last 50 years have seen a tremendous growth of the population. This has led to a strong increase in agricultural production. The intensive irrigation with river water caused a water table rise, followed by serious soil salination. To maintain production, over-irrigation is required to push salt from the surface beneath the root zone. This practice increases the amount of water used per unit crop and contributes again to water table rise. A vicious circle is triggered, leading to higher and higher salinity in the water flowing off the irrigation area both in the subsurface and in the drains. One could argue that the applied irrigation techniques and efficiencies in the Yanqi Basin are sustainable, as a steady state has been reached (the amount of salt transported out of the Yanqi Basin is equal to the amount of salt imported by the Kaidu) and production stabilized on a level still profitable. This of course cannot be called sustainable because only the needs of the farmers in this particular irrigation system are satisfied. With the rising groundwater table and the increased non-productive evaporation of water the salinity in the lake has increased and the lake level has fallen. The amount of water available for the downstream of lake Bostan, carried by the Kongque River, has decreased thus limiting natural vegetation growth and agriculture in the so called Green Corridor. The Green Corridor is a landscape, which extends down to Lop Nor and is characterized by the riverine desert poplar forests. Today, no water reaches Lop Nor due to the high consumptive use in the upstream irrigation systems. In order to improve the situation of the system as a whole, a number of measures in the upstream have been proposed (Dong, 2001). They include

Sustainable water management in arid and semi-arid environments

13

– the reduction of irrigation area, – substitution of irrigation water from the river by groundwater thus guaranteeing that the groundwater table stays below critical levels, – changes in the crop mix and irrigation techniques (e.g. drip irrigation for grapes), – the transfer of water directly from Kaidu river to the Kongque river bypassing the lake, – the lowering of the lake level in order to reduce non-productive evaporation of the lake and others. In an integrated modeling approach all these options will be investigated. Again, some relevant data can be obtained using remote sensing techniques. In this case we constructed a digital terrain model from stereo images of radar satellites based on methods described by Zebker and Goldstein, 1986. The absolute elevations were obtained from single point DGPS measurements (Fig. 9). The ground surface elevation is of particular interest in salination problems as evaporation from groundwater is a function of the distance to groundwater table. Hence salinity observed at the

Figure 9. Digital terrain model of the Yanqi basin.

Water resources of arid areas

14

Figure 10. Correlation between measurements at ground control points (GCP) of soil conductivity and spectral match to salt pixel. ground level is a data type, which allows the regional verification of the groundwater model. The distribution of surface salinity was obtained from multispectral ASTER data and measurements in the field. Based on the spectral response of a completely salinized pixel, the closeness of any pixel to this reference is determined yielding an uncalibrated salinity map. To convert these values into salinity or its physical measure of electrical conductivity, a calibration with ground truth is required. The ground truth was obtained both by single core samples and less time-consuming geophysical measurements. A good correlation between ground truth and the uncalibrated salinity map was found (Fig. 10). Of course, this correlation only holds for the non-irrigated areas. The salinity map (Fig. 11) clearly shows the salt accumulation in the paths between fields while in the irrigated fields themselves no increased salinity is visible due to overirrigation. While a coupled groundwater-surface water model is still under development, preliminary estimates are already available on the basis of a multi-box approach with the irrigation area, the aquifer and the lake being the respective boxes. Despite the fact that the box approach is a major simplification, it demonstrates that steady states for groundwater tables as well as salt concentration exist. Depending on how water in a steady state is exported from the system, reaching a steady state salt concentration can take a very long time compared to reaching a steady state in groundwater tables (Fig. 12). The steady state salt concentration is directly determined by

Sustainable water management in arid and semi-arid environments

15

Figure 11. Salinity map obtained from a multispectral satellite image.

Figure 12. Steady state for groundwater tables (hss, m below surface) and salinity (css, g/l) in the Yanqi basin aquifer as functions of applied irrigation water (in 107m3/a). the ratio of the flux of water draining from the aquifer into the lake and the groundwater recharge. Furthermore, the box approach shows that the rate of accumulation of salt increases rapidly as soon as direct evaporation from the aquifer occurs. Pumping groundwater for irrigation purposes would not only reduce the need for overirrigation, but also directly contribute to the water availability downstream. This solution is more expensive as groundwater comes at about 10 times the price of surface water. However, if the water table can be kept low by pumping groundwater, the conservation of soil and the increased availability of surface water in the downstream might strike the balance with a higher price of water.

Water resources of arid areas

16

6 CONCLUSIONS In arid countries the problems of sustainability in the water sector are prominent. On a worldwide basis the three subjects discussed are the most widespread. They show several common features. Water management in the arid and semi-arid environment must include salt management. Modern tools such as remote sensing, geophysics and modeling hydrological science help even in regions with weak infrastructure to quantify the implications of human interaction and to give advice to decision makers on the sustainability of water management practices. Models summarize the state of affairs and are the only means to make predictions. They are bound to be crude and simulations will always be idealized. Still, they can serve as points of reference. A further common feature is that sustainable solutions require the system boundary to be taken sufficiently large, often transgressing political boundaries. While science can give some decision support, the decisions for or against sustainability are made in the political arena. REFERENCES Alcamo, J., Doll, P., Henrichs, T., Kaspar, F., Lehner, B., Rosch, T. & Siebert, S. 2003. Development and testing of the WaterGAP 2 global model of water use and availability. Hydrological Sciences Journal-Journal Des Sciences Hydrologiques, 48(3):317–337. Bastiaanssen, W.G.M., Menenti, M., Feddes, R.A. & Holtslag, A.A.M. 1998a. A remote sensing surface energy balance algorithm for land (SEBAL). 1. Formulation. Jnl. of Hydrology, 212– 213:198–212. Bastiaanssen, W.G.M., Pelgrum, H., Wang, J., Ma, Y., Moreno, J.F., Roerink, G.J. & van der Wal, T. 1998b. A remote sensing surface energy balance algorithm for land (SEBAL). 2. Validation. Jnl. of Hydrology, 212–213:213–229. Bauer, P., Gumbricht, T. & Kinzelbach, W. 2004. A large-scale coupled surface water/ground water model of the Okavango Delta, Botswana. Water Resources Research, submitted. Dong, X., Jiang, T. & Jiang, H. 2001. Study on the pattern of water resources utlilsation and environmental conservation of Yanqi Basin. In: G.Li (Ed.), Development, Planning and Management of Surface and Groundwater Resources. IAHR congress proceedings. Tsinghua University Press, Beijing, China: 333–340. Ellery, W.N., Ellery, K., Rogers, K.H., McCarthy, T.S. & Walker, B.H. 1993. Vegetation, hydrology and sedimentation processes as determinants of channel form and dynamics in northeastern Okavango Delta, Botswana. African Jnl of Ecology, 31:10–25. ERESS 1972. Etude des Ressources en Eau du Sahara Septentrional. Rapport sur les Résultats du Projet, Conclusions et Recomm endations, UNESCO, Paris. Gumbricht, T., McCarthy, T.S. & Bauer, P. 2003. Microtopography of the Okavango Delta using correlation between land cover and elevation. Earth Surface Processes and Landforms, in press. Gumbricht, T., McCarthy, T.S. & Merry, C.L. 2001. The topography of the Okavango Delta, Botswana, and its tectonic and sedimentological implications. South African Jnl. of Geology, 104:243–264. Herman, A., Kumar, V.B., Arkin, P.A. & Kousky, J.V. 1997. Objectively Determined 10 Day African Rainfall Estimates Created for Famine Early Earning Systems. International Journal of Remote Sensing, 18(10):2147–2159. Hutchins, D.G., Hutton, L.G., Hutton, S.M., Jones, C.R. & Loenhert, E.P. 1976. A summary of the geology, seismicity, geomorphology and hydrogeology of the Okavango Delta, Geological Survey Botswana, Gaborone.

Sustainable water management in arid and semi-arid environments

17

Mamou, A. 1990. Charactéristiques et evaluation des resources en eau du sud Tunisien. Dissertation, Université de Paris-Sud, Centre d’Orsay. McCarthy, J., Gumbricht, T., McCarthy, T.S., Frost, P.E., Wessels, K. & Seidel, F. 2003. Flooding Patterns of the Okavango Wetland in Botswana between 1972 and 2000. Ambio, 32(7):453–457. McCarthy, T.S. and Ellery, W.N. 1994. The effect of vegetation on soil and ground water chemistry and hydrology of islands in the seasonal swamps of the Okavango fan, Botswana. Journal of Hydrology, 154: 169–193. McCarthy, T.S., Ellery, W.N., Rogers, K.H., Cairncross, B. & Ellery, K. 1986. The roles of sedimentation and plant growth in changing flow patterns in the Okavango Delta. South African Journal of Science, 82: 579–584. McCarthy, T.S., Green, R.W. & Franey, N.J. 1993. The influence of neo-tectonics on water dispersal in the north-eastern regions of the Okavango swamps, Botswana. Journal of African Earth Sciences, 17(1): 23–32. Modisi, M.P., Atekwana, E.A., Kampunzu, A.B. & Ngwisanyi, T.H. 2000. Rift kinematics during the incipient stages of continental extension: Evidence from the nascent Okavango rift basin, northwest Botswana. Geology, 28(10):939–942. Postel, S.L., Daily, G.C. & Ehrlich, P.R. 1996. Human appropriation of renewable fresh water. Science, 271(5250):785–788. Scholz, C.H., Koczynski, T.A. & Hutchins, D.G. 1976. Evidence of incipient rifting in Southern Africa. Geophysical Journal of the Royal Astronomical Society, 44:135–144. Siegfried T. 2003. Management of internationally shared groundwater resources in semiarid and arid region s: the Northern African Aquifer System. In E.Servat et al. (eds), Hydrology of Mediterranean and Semiarid Regions, IAHS Publ. No. 278, 2003. Thomas, D.S.G. & Shaw, P.A. 1991. The Kalahari Environment. Cambridge University Press, Cambridge. Zebker, H.A. & Goldstein, R.M. 1986. Topographic Mapping from Interferometric Synthetic Aperture Radar Observations. Journal of Geophysical Research-Solid Earth and Planets, 91(B5):4993–4999.

Theme A: Problems in obtaining hydrological and geohydrological data

Slug tests in fractured rock formations: value, pitfalls and misinterpretations P.D.Vermeulen & G.J.van Tonder Institute for Groundwater Studies, University of the Free State, Bloemfontein Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Currently slug tests in South Africa are used with two objectives in mind: (i) to get a first estimate of the yield of a borehole (relationship obtained by Viviers et al., (1995) and (ii) to estimate the Kvalue (or T-value) of the aquifer in the vicinity of the borehole. The paper shows that the use of currently available slug test interpretation methods to analyse slug tests in fractured rock aquifers to estimate a T or K-value is problematic. The estimated value is dependent on the flow thickness (thickness of the part of the aquifer in which flow occurs due to the slug input). If this thickness of flow is known, the estimated K-value is more representative of that of the fracture zone. By using the total thickness of the formation for the estimation of the K-value in slug test analysis, the estimated K-value (and thus KD-value) does not represent the T-value of the formation.

1 INTRODUCTION In performing a slug test, the static water level in a borehole is suddenly lowered or raised. This is usually done by lowering a closed cylinder into a borehole. The cylinder replaces its own volume of water within the borehole, thus increasing the pressure in the borehole. As the equilibrium in the water level is changed, it will recover or stabilise to its initial level. If the rate of recovery or recession of the water level is measured, the transmissivity or hydraulic conductivity of the borehole can be determined (Kruseman and De Ridder, 1994). In South Africa slug tests are conducted for the following two reasons: ● To estimate the hydraulic conductivity (K) of the aquifer in the vicinity of the borehole and ● To get a first estimate of the yield of a borehole (Vivier et al., 1995).

Water resources of arid areas

22

Vivier et al. (1995) performed slug tests on 32 boreholes, of which the maximum yield was known and they then derived empirically the following formula (there is a 93% correlation between the actual yield and the yield estimated with the formula): Q=117155.08t−0.83 (1) where: Q=yield of the borehole in l/h and t=recession time of the slug test in seconds (90% recovery). Usually the Cooper method (Cooper et al., 1967) or the Bouwer and Rice method (1976) is used to estimate the K-value (or T-value in the case of the Cooper method). In the following section slug test results, as well as pumping and tracer test results for borehole UO5 on the well-known Campus Site of the University of the Free State, South Africa (Figure 1) will be discussed to illustrate the problems associated with the interpretation of slug tests in a borehole drilled in a fractured aquifer.

Figure 1. Map of the RSA.

Slug tests in fractured rock formations

23

Figure 2. Diagram of the geological formation at the Campus Test Site (relative thickness of the aquifers in brackets). 2 BOREHOLE UO5 ON THE CAMPUS SITE The Campus Test Site is underlain by a series of mudstones and sandstones from the Adelaide Subgroup of the Beaufort Group of formations in the Karoo Supergroup (Figure 2). There are three aquifers present on the Site. The first, a phreatic aquifer, occurs within the upper mudstone layers on the Site. This aquifer is separated from the second and main aquifer, which occurs in a sandstone layer of between 8 and 10m thick, by a layer of carbonaceous shale with a thickness of 0.5 to 4m. The third aquifer occurs in the mudstone layers (more than 100m thick) that underlie the sandstone unit.

Water resources of arid areas

24

Figure 3. Acoustic scan of borehole UO5 at a depth of 20m to 25m below the surface.

Slug tests in fractured rock formations

25

Figure 4. Borehole video image of the fracture zone in borehole UO5 showing a fracture-zone thickness of about 200mm.

Figure 5. Constant rate pumping test data of UO5. Table 1. Hydraulic parameters estimated for UO5.

Water resources of arid areas

Parameter

26

Value

T of formation* (m2/d) 19 K of fracture zone (m/d) 3600 T of fracture zone (m2/d) 576 K of matrix (m/d) 0.17 T of matrix** (m2/d) 3 *Average for fracture+matrix, obtained from Cooper-Jacob fit to late drawdown values. **For 20m thickness.

A major characteristic of the main aquifer is the presence of a horizontal fracture that coincides approximately with the centre plane of the sandstone layer, and which intersects all 11 boreholes with significant yields on the Site, of which UO5 is one. The remaining 14 boreholes all have very insignificant yields. The fracture zone thickness is approximately 10mm, but the adjacent 200mm of sandstone is also highly permeable. Figure 5 shows a graph of the data from a constant rate test conducted on UO5 at a rate of 1.25L/s. Measurements were also taken in the observation borehole UO6. These pumping test data were analysed with a numerical 3D model (Van Tonder et al., 2001), and the following parameters were estimated in Table 1. The thickness of fracture zone (referred to in Table 1) was obtained from tracer tests and the borehole video, and is 0.16m. The hydraulic parameters given in Table 1 are regarded to be accurate (Van Tonder et al., 2001). It would now be interesting to analyse the data of a slug test (Figure 6) conducted on borehole UO5 and compare the estimated values with the values given in Table 1. The 90% recovery occurred after about 9 seconds, and using Equation (1) the yield of borehole UO5 is estimated as 5.3L/s. The tested blow yield of borehole UO5 was 6L/s during drilling. The Bouwer and Rice method (1976) was applied to the data in Figure 6. The Bouwer and Rice equation reads: (2) where: rc=radius of the unscreened part of the borehole where the head is rising rw=horizontal distance from the borehole centre to the undisturbed aquifer Re=Radial distance over which the difference in head h0 is dissipated in the flow system of the aquifer

Slug tests in fractured rock formations

27

Figure 6. Data collected during a slug test conducted on UO5. Table 2. Estimated K-values with the Bouwer and Rice method (1976) for different values of the flow thickness. Thickness open to flow (m)

K (m/d)

30 20 10 1 0.16 0.001

12 17 32 231 541 3600

T (m2/d) 360 340 320 231 86 3.6

d=length of the borehole screen or open section of the borehole h0=head in the borehole at time=0 ht=head in the borehole at time t The estimated K-value of Bouwer and Rice is dependent on the thickness open to flow, d, and Table 2 shows the different K-value estimates for different flow thicknesses. Note that a flow thickness of 30m will indicate the depth from the water level to the end of the borehole and that a thickness of 0.16m is the thickness of the fracture zone in borehole UO5. 3 DISCUSSION Comparison of Table 1 and Table 2 shows the following important issues: ● An incorrect K-value is obtained from the slug test if the thickness of the aquifer (total formation) is used as the flow thickness. For a thickness of 30m, a K-value of 12m/d

Water resources of arid areas

28

(or T=360m2/d) is estimated from the slug test, which is neither the T-value of the fracture zone nor the T-value of the matrix. ● For a flow thickness of 0.16m (i.e. the thickness of the fracture zone), a K-value of 541m/d is estimated with the Bouwer and Rice (1976) slug test method. This estimated K-value is more representative of the K-value of the fracture zone. ● The average T-value of the formation, which is important for management purposes, was estimated as 19m2/d from the constant rate pump test. It is impossible to estimate the T- or K-value of the aquifer formation via a slug test.

4 CONCLUSIONS The use of the current available slug test interpretation methods to analyse a slug test in a fractured rock aquifer to estimate a T- or K-value is problematic. The estimated value is dependent on the flow thickness (thickness of the part of the aquifer in which flow occurs due to the slug input). If this thickness of flow is known, the estimated K-value is more representative of that of the fracture zone. By using the total thickness of the formation for the estimation of the K-value in slug test analysis, the estimated K-value (and thus KD-value) does not represent the T-value of the formation. REFERENCES Bouwer, H. & Rice, R.C. 1976. A slug test for determining hydraulic conductivity of unconfined aquifers with completely pr partially penetrating wells. Water Resources Research, 12:423–428. Cooper, H.H, Bredehoeft, J.D., & D Papadopulos, I.S. 1967. Response of a finite-diameter well to an instantaneous charge of water. Water Resources Research, 3:263–269. Kruseman, G.P. & de Ridder, N.A. 1994. Analysis and Evaluation of Pumping Test Data. 2nd ed. International Institute for Land Reclamation and Improvement. Publication 47. Wageningen, the Netherlands: 237–247. Vivier, J.J.P., Van Tonder, G.J. & Botha, J.F. 1995. The use of slug tests to predict borehole yields: correlation between the recession time of slug tests and borehole yields. In Conference Proceedings: Groundwater’95: Groundwater Recharge and Rural Water Supply, Midrand, South Africa. Van Tonder, G.J., Botha, J.F., Chiang, W.H., Kunstmann, H. & Xu, Y. 2001. Estimation of the sustainable yields of boreholes in fractured rock formations, Special issue of Journal of Hydrology: No 241.

Flow simulation model performance assessment using entropy approach A.M.Ilunga Civil Engineering, University of the Witwatersand, South Africa D.Stephenson Civil Engineering, University of Botswana, Gaberone, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Hydrological data (e.g. rainfall, river flow, etc) are used in water resources planning and management for planning reservoir and operation. However, it happens sometime that the appropriate site where a hydraulic structure (e.g. reservoir) should be built has no available data due for example to inaccessibility to erect a flow gauging station, etc. This is particularly a problem in arid areas. Very often hydrologists make use of simulation models to estimate the flows data series at the very site from the nearby stations and using some physical characteristics of the catchment area. In this paper, a merely methodology is proposed to evaluate the performance of simulation models in terms of entropy (e.g. reduction of the uncertainty of flows) before and after applying a model to the site. This reduction should be above a certain threshold value for the model to be retained as performing well. An example is illustrated through RAFLER model, which is used to simulate yearly flows at Braamhoek.

1 INTRODUCTION For planning, management and effective control of water resources systems, a considerable amount of data on hydrologic variables such as rainfall, streamflow, etc. are required. It sometime happens that the appropriate site where a reservoir should be built has no available data due for example to inaccessibility to erect a flow gauging station. This is particularly a problem in arid areas. Physical models, semi-distributed models, statistical models, conceptual model, embracing probabilistic, fitting curve, black box, etc are often used to simulate/estimate flows. In this paper, a merely methodology is proposed to evaluate the performance of simulation models in terms of entropy (e.g. reduction of the uncertainty of flows) before and after applying a model to the site in a similar way of Panu (1992). These authors used entropy approach for infilling hydrological data problems (e.g. reduction of uncertainty before and after infilling the

Water resources of arid areas

30

data series), while in this paper the same approach is applied to cases where no data series is available at all at the target site. In very recent paper (see, Ilunga and Stephenson, 2003b) the methodology for evaluating the model performance was roughly used but it was not explained systematically as done in this paper. It should be noted that the reduction evoked above should be above a certain threshold value for the model at hand to be retained as performing well. An example is illustrated through RAFLER model, which is used to simulate yearly flows at Braamhoek. 2 INFORMATION CONTENT OF HYDROLOGICAL VARIABLES Traditionally, the information content of a hydrological variable can be measured through variance, which shows the variability of the hydrological variable with respect to its mean. However this approach was criticized for cases where information available about the hydrological variable is little (Singh, 1998). Since 70’s hydrologists tried to find another way of measuring information by theoretic entropy (a term borrowed from communication, see Shannon & Weaver, 1949). Thus the concept has been applied in water resources (Singh & Florentino, 1992; Amorocho & Espildora, 1973) and water related fields. The entropy is considered as a measure of the amount of chaos or lack of information about a system. The entropy can be viewed as a measure of ignorance about the system described in classical sense by a probability distribution. Indirectly, it measures the information about the system. Mathematically entropy of a system {xi} is defined in its discrete form by the following expression (1) where K: is a function of the base used or the scale factor (bits for base 2, napiers for base e, decibels for base 10), i=1, 2,…, n and pi is the probability of occurrence of the event i. It can be shown that the value of H(X) reaches its maximum when all variate values xi are equally likely, that is, when the outcome has maximum uncertainty (Amorocho & Espildora, 1973). In this case the entropy becomes Hmax (X)=log n (2) The theoretic entropy definition was extended to hydrology. Hence entropy is considered as a measure of the degree of uncertainty of random variable hydrologic processes (Amorocho & Espildora, 1973). Since the reduction of the uncertainty by means of making observations is equal to the amount of information gained, the entropy criterion indirectly measures the information content of a given series of data (Harmancioglu et al., 1994). It arises that the distribution of the variable can be unknown a prior although some of its properties may be known, e.g. mean, variance, normality condition. These proprieties (information) enable to determine the distribution of the variable, which maximizes the entropy function. In this way the distribution is consistent with the available information,

Flow simulation model performance assessment using entropy approach

31

but retains maximum uncertainty within the feasible domain and thus ensures the least bias; that is the principle of maximum entropy (POME) introduced by Jaynes in 1968. This principle has been applied intensively in hydrology in the last two decades. 3 ENTROPY APPROACH AS HYDROLOGICAL MODEL PERFORMANCE EVALUATION Amorocho and Espildora (1973) suggested that the mutual information (between the observed values and the simulated ones) could be used as a criterion in the selection of hydrological models; e.g. rainfall-runoff prediction. Note that the mutual information concept is derived from entropy notion and for more details; the reader is referred for example to the above-mentioned paper. Later the directional information transfer index (DIT) appeared as a generalization of the mutual information Yang and Burn (1994) and was used for dependency evaluation between streamflow gauging station pairs. Recently, it is argued that since mutual information is used for model performance assessment, its generalization i.e. DIT can be extended to model performance evaluation (Ilunga & Stephenson, 2003a). The above considerations are valid when the estimated values have to be compared to the observed ones. In that respect statistical criteria such as root mean square error, etc can be also used to crosscheck the results (Ilunga and Stephenson, 2003a). However it becomes difficult to use these considerations when missing values are encountered in the data series. Thus Panu (1992) introduced the notion of reduction of uncertainty of the hydrological variable before and after infilling the data series. The reduction of uncertainty Re d(%) at a given site as defined by Panu (1992) can be given as follows: (3) where Hcc and Hcomp are entropy values before and after infilling the data series respectively. It should be noted that this concept was applied to cases of consecutive missing data values, e.g. hydrological data exist before and after the missing values. 4 PROPOSED METHODOLOGY Panu (1992) used expression (3) for infilling data problems, in other words some data exist before infilling process. In this paper the same expression (3) is proposed for cases where no available flow data exist at all at the site. It is more natural to say that a case where no data is available, the uncertainty is higher than a case where data exist. Thus it is assumed that the uncertainty should be maximum (e.g. if all hydrological events would occur equally likely) at a site where no data is known. Thus, in this case expression (3) can be re-written as (Ilunga & Stephenson, 2003b): (4)

Water resources of arid areas

32

Where Hmax is defined by expression (2) The following are the different steps for evaluating the performance of a flow simulation model for cases where no data exist at all at given site. 1. Having the physical parameters of the catchment area and information (e.g. rainfall) from the nearby sites, compute the simulated flows. 2. Compute the frequency (probability) distribution of the flows. 3. Compute the marginal entropy of the site using formula (1) and set the entropy before simulating flows to its possible maximum value, e.g. see formula (2). 4. Compute the reduction of entropy at the site using expression (4) and set a threshold reduction of entropy to an arbitrary value. If the computed value for the reduction is greater or equal than the threshold value, the model is considered as performing well. Otherwise, the model performs poorly. Terminate.

5 SHORT NOTE ON RAFLER RAFLER is an acronym for Rainfall Flow Erosion. A model (RAFLER) is a deterministic model based on the physics of runoff, soil infiltration and soil transport and which converts rainfall data to runoff over a length of time, e.g. years. The model uses monthly rainfall figures to reproduce monthly stream flow series and soil erosion. Some simplification is made to enable the model to be run with a minimum of data. And the rainfall period each month is estimated from the number of rain days to enable true flow rates to be calculated. This model requires a number of modules including catchment, channels and reservoirs. The general theoretical background of the model can be traced in Stephenson (2002). 6 STUDY AREA AND DATA AVAILABILITY Braamhoek is situated in the Free State, in South Africa. The catchment area is about 62km2. Neither rainfall data nor stream flow data is available at this particular site. Thus it was possible to simulate flows at Braamhoek using rainfall data from the nearby sites; viz at Van Reenen (MAP=1002mm/month); at Moorside (MAP=839mm/month) and at Baldergow

Flow simulation model performance assessment using entropy approach

33

Figure 1. Simulated yearly frequency (probability) distribution at Braamhoek. Table 1. Model performance evaluation at Braamhoek.

Before applying RAFLER Applying RAFLER

Marginal entropy (Napiers) at Braamhoek

Reduction of uncertainty (%) at Braamhoek

4.47

−

1.65

63.06

(MAP=887mm/month). The monthly rainfall data (1916–2002) were obtained from the Weather Bureau, South Africa. 6.1 Application of the methodology to Braamhoek The application of the model, i.e. RAFLER to simulate the total annual flows (from 1916–2002, e.g. 87 data points) at Braamhoek site gave the following results. Figure 1 from which the entropy calculations were possible shows the probability (frequency) distribution estimated from the model. The threshold value of the reduction of uncertainty was set to a value of 50% napiers. Table 1 shows the results of entropy calculations before and after applying RAFLER models. It is therefore concluded that the reduction in uncertainty of the yearly flows at

Water resources of arid areas

34

Braamhoek was 66.06% by this model. This value is the equivalent of information inferred about the site using RAFLER model. This model could be thought to perform well. Thus RAFLER model could be used for flow prediction at Braamhoek with regard to the total yearly flows. Nonetheless the model needs to be tested on other flow regimes for that specific site. 7 CONCLUSION The focus of this paper was to give a methodology for evaluating the performance of simulation models using entropy approach. The methodology has been tested with RAFLER model on Braamhoek site where records were simulated. Recall that this methodology was roughly used in Ilunga and Stephenson (2003a), but without presenting systematically the steps involved as been done in this paper. The computations from the entropy criterion showed that RAFLER model could be used for simulating the total yearly flows at Braamhoek when a threshold value of 50% is considered for the reduction of uncertainty before and after simulation. Investigation should also be done on other flow regimes. REFERENCES Amorocho, J. & Espildora, B. 1973. Entropy in the assessment of uncertainty in hydrologic systems and models. Water Resources Research, 9(6):1511–1522. Harmancioglu, N.B., Alpaslan & Singh, V.P. 1994. Assessment of the entropy principle as applied to, water quality monitoring network design. Stochastic and Statistical Methods in Hydrology and Environmental Engineering., 3:135–148. Ilunga, M. & Stephenson, D. 2003a. Performance of hydrological data infilling techniques using entropy approach: Expectation maximization algorithms. 11th South African National Hydrology Symposium, Port Elizabeth, South Africa: 6. Ilunga, M. & Stephenson, D. 2003b. Entropic measures for comparing flow simulation models at Bedford site. Paper submitted to the J. Hydrology, Elsevier. Panu, U.S. 1992. Application of some entropic measures in hydrologic data infilling procedures. In: Singh, V.P. & Fiorentino, M. (Eds) Entropy and energy dissipation in water resources, Kluwer Academic Publishers, The Netherlands: 175–192. Shannon, C.E. & Weaver, W. 1949. The Mathematical Theory of Communication. University of Illinois Press Urbana, Chicago, London. Singh, V.P. 1998. Entropy as a decision tool in environmental and water resources. J. Hydrology , Indian Association of Hydrologists, 21(1–4):1–12. Singh, V.P. & Florentino, M. 1992. A historical perspective of entropy applications in water resources. In: Singh, V.P. & Fiorentino, M. (Eds) Entropy and energy dissipation in water resources, Kluwer Academic Publishers, The Netherlands: 21–61. Singh, V.P. & Krstanovic, P.F. 1987. A stochastic model for sediment yield using the principle of maximum entropy. Water Resources Research, 23(5):781–793. Stephenson, D. 2002. “Modular kinematic model for runoff simulation”. In: V.P.Singh, & D.K.Frevert (Eds). Mathematical models of small watershed hydrology and applications. Water Resources Publications, LLC, pp. 183–218, Chapter 7. Yang, Y. & Burn, H. 1994. “Entropy approach to data collection network design”. J. Hydrology, 94:307–324.

Flow simulation model performance assessment using entropy approach

35

Yevjevich, V. (1972). Probability in hydrology. Water Resources Publications, Colorado, U.S.A.: 331.

Data collection experiences in water level monitoring, borehole archive and research projects in semi arid Botswana Magowe Magowe, Thothi Obakeng & Paul Makobo Department of Geological Survey, Hydrogeology Division, Lobatse, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The Department of Geological Survey (DGS) of Botswana has been carrying out water level monitoring of non-operational wellfields since 1983 and various research projects throughout the country. This involved intense collection of data on water levels and rainfall manually and later on automatic data collection instruments were introduced. This data was stored as hardcopies without quality assurance. Problems experienced with this data collection encompass logistical, equipment and human input. The logistical problems include poor accessibility due to the country’s hostile environment such as dust and extensive ponding during heavy showers. This hostile environment leads to reduction of the lifespan and the poor performance of these instruments. Poor handling of data, equipment failure, lack of the right set of equipment and local operational knowledge also poses problems. Therefore, as a consequence, valuable data is normally lost. In 2003, a quality assurance process was resumed for the 1999–2003 water level monitoring and rainfall records. Common problems that were encountered are data gaps which could be explained by the above causes. During the quality assurance of 1999–2002 data for the ten (10) monitoring network areas, data gaps or unavailable data constituted 80% of all recorded problems (DGS, 2003). This paper discusses the experience of the DGS in collecting hydrogeological data in Botswana semi arid environment.

1 INTRODUCTION The DGS Hydrogeology division has been collecting data in the areas of monitoring, borehole archive and research projects. Data collection in monitoring started in 1983 and since then boreholes from various groundwater projects have been added to the

Data collection experiences in water level monitoring, borehole archive and research projects

37

monitoring network. The data serves various purposes such as establishment of benchmarks and parameters of various systems. Monitoring is comprised of water level and rainfall measurement for establishing the natural piezometric surface. Knowledge of the natural piezometric surface is also needed as an input in water resources modelling and groundwater recharge estimation efforts. National Borehole Archive acts as storage for all borehole data whereas Research projects collects data on various hydrogeological parameters. This data collection is wrought with problems that are of logistical, equipment and human nature. 2 MONITORING Water level measurement by manual dipping and change of rainfall charts are done on monthly basis in selected boreholes. The problems recorded in the data collection sheets of ten monitoring networks are depicted in Figure 1.

Figure 1. Distribution of problems as recorded in water level records for 1999–2002 period. 3 MONITORING LOGISTICAL ISSUES However, this schedule is not followed due to several problems such as poor accessibility due to the country’s hostile environment and its vastness. Since the present monitoring network is determined by the groundwater potential projects carried out, most of the monitored areas are in remote areas far from the main office and are laden with heavy

Water resources of arid areas

38

sands and muddy soils from heavy thunderstorms. This means that a four wheel drive vehicle is a necessity. However, in some cases these vehicles are not available and this results in data gaps for some months. The other problem is late availability of transport which results in the late changing of water level/rainfall charts and water sampling. This renders the data useless as it results in the superposition of more than one line on the charts making it difficult to read. In some cases, other stakeholder dealing with water connect these monitoring boreholes to water supply out of emergency to supply people with water and this creates data gaps in the records as the borehole will not be accessible and cannot be used. In other cases, other water authorities drill production boreholes adjacent to these monitoring boreholes and the pumping creates interference and as such the borehole will be taken out of the network. Some of the boreholes are in private property and in some cases there is no access as the gates are locked. In other situations the dipper is reported to be stuck for the whole year such as in borehole 6736 in Lethakeng/Bothapatlou monitoring network. 4 MONITORING DATA COLLECTION ISSUES 4.1 Recording One of the major problems is the recording system. There are several problems associated with data recording and these include the following: – Unclear or no comments, this include things such as reporting same borehole dry and blocked/ collapsed for different months or reporting “no hole for dipper” – Water level taken from a borehole with unknown number – Incorrect entry of measured values – Measured values being different from the chart recorded value. This unclear recording resulted in data being discarded hence creating data gaps. 4.2 Quality check All the collected data between 1999 and 2002 was just filed without being checked and the lack of quality check was evident by a lot of problems encountered during the quality assurance of this 1999–2002 data. This lack of timely quality checking resulted in some data points being thrown out as it was difficult to know the exact reasons for these discrepancies/anomalies. This included problems such as recording problems. 4.3 Data storage Until 2003, all the collected hard copy data was not being digitised and it was not filed properly. The records/charts were either misfiled in different binder or thrown in drawers. This resulted in some of the records missing and these added to the issue of unavailable data.

Data collection experiences in water level monitoring, borehole archive and research projects

39

4.4 Equipment Lack of proper preparation for the field also creates data collection problems. In some cases, it is reported that dipper or its light was not working and this results in partial or no data collection. Due to lack of timely quality check it becomes difficult in ascertaining the source of the mal/non-functioning of the equipment, whether it is the sensor or the batteries. The other problem was from the mechanical water level recorders being used, these were perceived as a better replacement of manual dipping, and however, they came with their own shortcomings. In most cases, the monitoring equipment used was designed and manufactured in Europe where the environment is completely different from the semi arid conditions of Botswana. This has lead to tremendous reduction of the lifespan and the poor performance of these instruments. Some of the problems experienced include – Stuck pens rendering the data – No marking on the chart resulting in blank chart.

5 MONITORING HUMAN RESOURCES ISSUES 5.1 Knowledge In most cases, the personnel operating some of the equipment such as water level recorders, dippers, sampling pumps and rain gauges lack the technical know-how necessary to implement first line maintenance. This result in late or no acquisition of data and hence data gaps develop while the equipment is sent for maintenance or replacement. The lack of knowledge sometimes results in the equipment not being calibrated or set up properly and this indicated off scale water level and rainfall curves. 5.2 Availability The personnel used for data collection are at technician and artisan level. In most cases these personnel are shared among the various on going research projects and the normal monthly monitoring. This results in one of these activities suffering due to nonavailability of the personnel for a certain period and as such data gap will be inevitable. 6 NATIONAL BOREHOLE ARCHIVE (NBA) NBA records borehole data on daily basis. This includes registering privately drilled boreholes; entering borehole data in the database including plotting boreholes on hardcopy maps and storing rock chip samples. 6.1 NBA Logistical issues There are several logistical problems associated with NBA. One of these problems is the running out space for rock chip samples storage. The lone core shed is full and this has resulted in the new and reliable samples being piled without proper storage. The other

Water resources of arid areas

40

problem is the use of outdated topographical maps for plotting boreholes which results in difficulty in borehole location verification. The lack of physical verification of registered boreholes is also one of the problems. This is due to the fact that boreholes are drilled almost daily all over the country and due to its vastness; it is difficult to cover all the drilled boreholes. 6.2 NBA Quality check Since the NBA is the storage for all drilled boreholes in the country, the issue of data quality check is very important. However, until recently it has been neglected especially on the privately owned boreholes. As an authority responsible for registration of private boreholes, one of the main tasks is to verify the location of the boreholes as provided by the owner/driller so that the borehole could be plotted correctly in the map. However, this has not been done adequately and as such a lot of boreholes have uncertain location. 6.3 National Borehole Archive Human Resources issues 6.3.1 Availability The personnel used for data collection are at technician and artisan level. In some cases these personnel is shared between NBA and the normal monthly monitoring. This results in one of these activities being suspended for a certain period and creating backlog. 7 RESEARCH PROJECTS (RP) The Hydrogeology division has been running various projects ranging from groundwater potential survey to hardcore research projects such as Groundwater Recharge Evaluation Studies (GRES) and the Kalahari Research Project. These projects are multi-disciplinary and use different equipment and collect different data sets. In most cases these projects are carried out jointly with external partners and therefore timely bound. These projects, especially those run in-house experience a lot of problems. 7.1 Research Projects logistical issues In most cases these projects are carried out in remote areas and several problems are experienced in the field. This includes transport problems such as vehicle breakdown which takes a long time to fix due to the long process to be followed. In some cases there is a need to seek permission from other stakeholders and these requests can result in extension of the program while waiting for approval. This includes funds approval and changing the scope of the study. For example, the approval of funds and changing of project scope may take three (3) to four (4) months of valuable field activity time. This negatively affects projects that require time based data.

Data collection experiences in water level monitoring, borehole archive and research projects

41

8 RESEARCH PROJECTS HUMAN RESOURCES ISSUES 8.1 Knowledge In most cases these projects use specialised equipment that requires special operating level. However, most of the personnel have never been exposed to this equipment and this can result in loss of data or collection of unreliable data. For, example various software used to operate Skye data loggers are still unknown to a good number of hydrogeological technical personnel within Hydrogeology Division. 8.2 Availability The personnel used for data collection are at technician and artisan level. In most cases these personnel is shared among the various on going research projects and the normal monthly monitoring. This lack of technical level staff impacts negatively on the running of these projects. This results in the project using unqualified staff to fill the gap; however, that has serious implications on the quality of the collected data. In other situations, there is a need to have specialised personnel such as a welder to develop a specialised piece of equipment on site. This can result in delays especially if that person is unavailable or is occupied with other departmental work. 8.3 Equipment Some of this necessary specialised research equipment needs special care and due to harsh conditions prevailing in these remote areas, a lot of time is lost when the equipment breaks down since it must be sent overseas to be fixed. This also results in loss of data especially temporally dependent data. 9 CONCLUSION On the basis of this experience, we conclude that the following aspects are vital for a successful and reliable hydrogeological data collection effort in the semi-arid Botswana environment. – Routine analysis of the archived data should be a must rather than an option, in order to eliminate useless data before it accumulates in large amounts within records. – Routine training programmes for technicians on field equipment should be designed to enable technicians to keep abreast with the changing technology that is specific for hydrogeological applications. – Increasing manpower capacity by recruiting personnel with basic hydrogeological monitoring and database knowledge in order to facilitate data collection and reduce data losses arising from lack of knowledge. This will increase the reliability of the collected data. – The general public and other stakeholders need to be informed about the importance and relevance of hydrogeological research, borehole archiving and monitoring

Water resources of arid areas

42

activities, so that they can allow such activities in their private properties such as farms. – Manufacturers must be encouraged to design field equipment suited to the hostile semi arid and saline conditions of Botswana, so that the durability of the field equipment can be guaranteed. – Periodical inspection of water level monitoring boreholes should be a must in order to curtail issues of “dry” or “blocked” boreholes, hence maintain a continuous and an accurate water level record. – A comprehensive process map of water level monitoring program which include recording of the environmental status or changes in the vicinity of the monitoring borehole such as new pumping borehole. – Conduct a routine water sampling of observation boreholes. Currently the Hydrogeology Division is engaged in improvement of data collection and archiving through implementation of the following programs. – Acquisition of digital water level and rainfall recorders and accessories to replace mechanical ones and manual dipping. This will reduce human errors and improve data quality. – Development of proper databases and checking data immediately from the field to ensure that issue of unclear comments and data anomalies are reduced hence maintaining good quality data. – Development of process maps to improve the quality of the data being collected and being entered into the databases. This will ensure that all factors are considered before a inexplicable conclusions such as “dry” boreholes are reached. – Field programs are being carried out to review borehole location maps. This is to ensure that borehole locations are correct and indeed the plotted boreholes do exist. Regular data collection even if it is not part of a specific study, helps to build a picture of the general behavior of the system. The data collected provides valuable comparisons and context when the system is studied in more detail. However, all this will not be possible if the data collected is wreaked with a lot of problems. REFERENCES Department of Geological Survey. 2003a. Groundwater Monitoring of Non-Operational Wellfields and Other areas of Development Interest, compiled by T.Kellner, vol. 1b–1d. Department of Geological Survey (2003), Groundwater Monitoring of Non-Operational Wellfields and Other areas of Development Interest, compiled by T.Kellner, vol. 2b, 2d–2f, 2h–2k.

Rainfall characteristics in semi-arid Kitui district of Kenya A.O.Opere, V.O.Awuor, S.O.Kooke & W.O.Omoto Department of Meteorology, Nairobi, Kenya Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Variable semi-arid climate characterized by precipitation patterns unfavorably distributed in space and time and high evaporation rates reaching up to 100% of the incoming monthly precipitation is a challenge facing water resources management in arid and semi-arid lands (ASALs). Kitui district in Kenya is an example of an ASAL environment where water resources management issues are particularly important and sensitive. Sources of water are nearer to the people in the wet season, but as the seasonal rivers dry up, distance to water points get as far away as 25–30km. There is, however, great potential for rainwater harvesting. This is dependent on proper understanding of the patterns of precipitation both in space and time. This would be useful in understanding drought characteristics in order to develop strategies to capture, store and redistribute the available water. The spatial characteristics were determined through principal component analysis (PCA). Season lengths, drought severity and frequency were determined. The results indicated that, on average, the onset for the long rains (March–May) was centred on day 82.36 while cessation was on day 126.3. The longest season was 107 days during the long rains while the shortest lasted 7 days during the short rains (October–November). Severe droughts in the district were experienced in 1980, 1985, 1990 and 1995. The largest seasonal total for the long rains was 768mm with a return period of 25.9 years while the smallest total was 81.1mm with a return period of 1.0 years. For the short rains, the largest total was 1022.0mm with a return period of 15.7 years and the smallest total was 205.4mm with a return period of 1.0 years.

1 INTRODUCTION Efficient and sustainable use of available water resources is paramount for a peaceful, sustainable and equitable development of any region. Some of the challenges to integrated and sustainable water resource management include:

Water resources of arid areas

44

Changing land use, land degradation by erosion, deteriorating water quality and competing water demands by stakeholders. There exists, therefore, a strong demand for an integrated water allocation and decision support system. The backbone of such a system must be scientifically sound to be accepted and trusted by stakeholders. Efforts to improve catchment management and to impose a sustainable water resources management are of economic and political importance for any country. Variable semi-arid climate characterized by precipitation patterns unfavorably distributed in space and time and high evaporation rates reaching up to 100% of the incoming monthly precipitation is a challenge facing water resources management in arid and semi-arid lands (ASALs). Kitui district in Kenya is an example of an ASAL environment where water resources management issues are particularly important and sensitive. Sources of water are nearer to the people in the wet season, but as the seasonal rivers dry up, distance to water points get as far away as 25–30km. The main problems in water development and management in this district include: ● unreliable rainfall and inadequate supply to meet the demand, ● the available water resources are unevenly distributed and inaccessible to all, ● traditional farming methods lack water conservation principles, ● most of the water projects have since been abandoned, and ● High rates of potential evaporation on the available water resources. There is great potential for rainwater harvesting. This is dependent on proper understanding of the patterns of precipitation both in space and time. This would be useful in understanding drought characteristics in order to develop strategies to capture, store and redistribute the available water. Droughts have been the phenomena of great concern throughout the continent of Africa, because of the devastating effects they have inflicted on the economies of some of the countries in the continent. Kitui District located in Eastern Kenya is no exception and is an example of one of the most vulnerable areas to the effects of drought. Droughts are usually classified as meteorological, hydrologic or agricultural depending on the variable under investigation. Definition of droughts has also been given on the basis of theory of runs and stationary structure of time series, Yevjevich (1967). The most important variable in meteorological drought is rainfall, in hydrological drought is availability of water in rivers, lakes, reservoirs, and ground water resources; and in agricultural drought is the soil moisture content to sustain the crop growth. Drought analysis involves investigation of duration, magnitude or severity, frequency and regional spread of the event. There have been limited investigations on meteorological droughts in Kenya. However, substantial work exists on the drought characteristics for the various agro-climatic regions of South Africa; Dyer and Tyson (1977), Zucchini and Adamson (1984), Dent et al. (1987). The study investigates the duration, magnitude or severity as well as frequency aspects of drought within Kitui District.

Rainfall characteristics in semi-arid Kitui district of Kenya

45

2 DATA AND METHODS 2.1 Data: types, sources and problems 2.1.1 Historical data Historical data from thirteen rainfall stations within and around the project area were analyzed. The first part of this analysis was to determine the onset and cessation dates of rainfall in each year for the short and long rains. The duration for each season was then derived. 2.1.1.1 Filling in missing records Many rain gage records are incomplete. It is necessary to estimate the missing data in order to utilize partial records, especially in data sparse areas. The methods used include: ● Arithmetic average method ● Normal ratio method ● Correlation method ● Inverse distance method The problem of filling data at un-gauged location involves transmitting data at the nearby index gages to the un-gauged location. The missing data can be formulated as (1) ai is the weighting factor of the ith gage with record Pi and N is the number of index gages while Px is the rainfall to be estimated at x. The different methods differ in their methods of estimating ai’s where i=1, 2,…………………N 2.1.2 Data quality control One of the sources of error in rainfall measurement is the location of the gage in relation to obstructing objects such as trees and buildings. In the progress of time, trees grow and buildings come up. This means raingages must be moved. This may affect the consistency of the records from the raingage i.e. the records before and after the movement might be different. In addition, change in observational procedure might also affect the consistency.

Water resources of arid areas

46

Inconsistency in a rainfall record may be detected by graphical or statistical methods such as double mass analysis, the Vonn Neumman ratio test, cumulative deviations, likelihood ratio test and runs test. 2.2 Methods 2.2.1 Principal component analysis (PCA) The method of PCA involve the transformation of a greater number of unorthogonal (manifest) variables into smaller number of orthogonal variables, which present common causes of manifest variable changes. It can therefore reduce the dimensionality of a problem by replacing the measured variables and the inter-correlated variables by using a smaller number of uncorrelated variables. This can be useful in reducing the amount of basic data to be processed. Depending on the data, it is possible to interpret the orthogonal functions in terms of some underlying physical processes. Castell (1966) proposed a method of retaining significant factors in PCA solutions. Similar methods have been used by Ogallo (1988a,1989) and Basalirwa et al. (1995) for East Africa and Tanzania respectively. This method was used to group rainfall records from the study catchment into homogeneous zones. Mathematically, a variable Z may be transformed in terms of m common empirical orthogonal function (factors) and n unique factors as below: (2) where, Zi is variable i in the standardized form Fi represents the common orthogonal vector (factor) ui is the unique factor for variable i ai1=standardized multiple regression coefficient of the variable i on the common factor 1 (factor loading). The unique term diui=0 since principal component analysis does not consider the unique component of the variance. Details of this method are available in many referencesincluding Drosdowsky (1993), Ogallo (1989) and Basalirwa (1991) among others. 2.2.1.1 Identification of representative rainfall station Principal component analysis (PCA) solutions were used to identify the most representative station in the area of study. This formed the basis for further analyses including onset/cessation of rainfall, dry spell lengths and frequency. Two stations from the thirteen stations were chosen for detailed analyses. These were Kitui Secondary School (identified as station with highest communality from PCA results); thus is a representative station based on communality concept. Kitui Water Office was also used for comparison of results.

Rainfall characteristics in semi-arid Kitui district of Kenya

47

2.2.2 Drought indices Drought duration is a crucial component particularly if one has to plan for storage that can last certain duration of drought for given water demand. The duration portrays the season lengths hence the potential success or failure of a water supply scheme that can be put up. At each of the stations, the onset and cessation dates were determined for each year and for each season i.e. long and short rains. The actual, earliest and latest onset and cessation dates was the basis of deriving average, shortest and longest duration at each of the stations. Mhita (1990) has made similar attempt in Tanzania. Two methods used were the water balance technique and pentad method. Both methods are based on preset hydrological conditions, to determine the onset and cessation of rainfall. Details of these methods are briefly discussed in the following sections. 2.2.3 Pentad method Definition of the start of the rains that is used is based on preset hydrological conditions. The first occasion after March 1st and October 1st that the running 5 day total exceeded 25mm and there being no dry spell exceeding 7 days in the next 21 days (Successful start, threshold of 1mm). In a nutshell, the pentad method involves computing a 5-day total rainfall for each year. The cumulative values of the 5-day total are divided by the annual total for each year and expressed as a percentage, that is, ΣPi*100/Annual total. These are plotted against pentad numbers. The onset and cessation dates are then determined from the plots. 2.2.4 Water balance technique The first occasion after March 1st and October 1st for the long and short rains respectively when the water balance goes to zero (capacity 100mm, daily evaporation 6mm). A water balance approach was used with a threshold of 1mm of rainfall. Evaporation rate for the area is taken to be 6mm per day on average. The soil moisture capacity was taken to be 100mm (i.e. average soil moisture during dry days). No runoff is generated since the rainfall amounts cannot even satisfy the evaporation demand. 2.2.5 Determination of season duration The maximum duration for each season was obtained using the earliest onset and latest cessation for the period of study in each of the two stations. The longest and shortest duration for each season was also determined for the two stations. 3 RESULTS AND DISCUSSIONS The results of the PCA, onset/cessation of rainfall, dry spell lengths and frequency are presented in the following sections.

Water resources of arid areas

48

3.1 Results of the principal component analysis From the results, three Eigen vectors were retained based on the Scree method (see Fig.1).

Figure 1. Scree test of Castell. Kitui Sec. School had the highest communality and was thus picked as the most representative station for further analysis. Results from PCA can be seen in Table 1 and Figure 2. The spatial map was obtained by mapping the factor loadings at the station locations. Three homogeneous regions were delineated from these results indicating complex rainfall variability within the study area.

Table 1. Rotated loading matrix. Variable (rainfall stations) 9137012 9137010 9137003 9138000 9137020 9137028 9137045 9137058 9137073

Factor 1

2

3

0.212 0.0373 0.164 0.458 0.191 0.385 0.757 0.274 −0.027

0.288 0.119 0.154 0.122 0.786 0.293 0.179 0.531 0.141

0.325 0.656 0.643 0.549 0.171 0.396 0.331 0.251 0.770

Rainfall characteristics in semi-arid Kitui district of Kenya

9137076 9137094 9138003 9138013 Variance explained % of total Variance

49

0.796 0.196 0.056 0.063 0.792 0.082 0.848 0.129 0.106 0.546 0.147 0.256 2.908 1.876 2.290 22.37 14.429 17.612

Figure 2. Homogeneous rainfall zones of Kitui. Table 2. Examples of onset/cessation dates. Average Earliest/shortest Latest/longest Std. (Day (Day no.) (Day no.) dev. no.) (Day no.) 9137012 LR onset LR cessation LR duration SR onset SR cessation

96.9 131.4

76 115

123 11.84 141 8.19

35.45

7

56 11.37

307 342.2

280 333

335 12.89 366 8.483

Water resources of arid areas

SR duration 9137010 LR onset LR cessation LR duration SR onset SR cessation SR duration 9137003 LR onset LR cessation LR duration SR onset SR cessation SR duration 9138000 LR onset LR cessation LR duration SR onset SR cessation SR duration

50

36.18

11

69 17.31

82.35 143.1

0 124

122 25.21 163 10.04

62

18

149 30.68

306.9 349.4

287 328

319 9.462 366 12.45

46.17

14

76 19.04

93.7 138.4

64 122

129 16.35 152 6.905

46.05

11

83 20.02

307.8 342.3

289 336

324 10.26 361 8.42

37

16

71 15.02

97.92 124.5

78 99

144 14.44 172 16.38

27.63

8

47 12.15

316 353.8

296 336

341 12.51 366 11

41.13

6

60 15.66

Rainfall characteristics in semi-arid Kitui district of Kenya

Figure 3. Water balance for Kitui Water Office.

Figure 4. Pentad for Kitui Water Office 1989.

51

Water resources of arid areas

52

Figure 5. Water balance for Kitui Sec. School—1984. 3.2 Onset and cessation dates of rainfall Some results for onset/cessation dates and season duration are given Table 2. A few examples of the results from pentad method for the two stations are given in Figure 4 and Figure 6. The annual water balance plots from the water balance technique is also given in Figure 3 and Figure 5.

Figure 6. Pentad for Kitui Water Office 1988.

Rainfall characteristics in semi-arid Kitui district of Kenya

53

Table 3. Dry spell lengths for Kitui Sec. School and Kitui Water Office. Kitui Sec. School Year L S 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998

22 31 27 24 35 27 84 18 36 66 32 81 69 70 33 8 55 61 37 12 14 23 7 16

Kitui Water Office Year L S

18 1981 23 1982 39 1983 155 1984 18 1985 79 1986 23 1987 14 1988 169 1989 9 1990 17 1991 149 1992 83 1993 41 1994 18 1995 12 1996 15 1997 27 140 11 13 114 107 26

11 10 23 8 13 56 31 43 18 22 70 46 20 5 7 114 7

12 17 25 67 30 13 72 28 47 17 72 64 28 26 31 23 107

The two methods were found to be comparable. For example, Kitui Water Office: Average onset of long rains is on day 82.36 and cessation on day 126.3 using the two methods. On the other hand onset is on day 84.37 and cessation on day 137.5 using the pentad method. These results are also presented in Table 2.

Water resources of arid areas

54

Figure 7. Dry spell lengths during long rains—Kitui Sec. School.

Figure 8. Dry spell lengths during short rains—Kitui Water Office. 3.3 Season lengths/duration Some of the results of the season lengths for Kitui Water Office and Kitui Sec. School are given in Table 4 and presented in Figure 7 to Figure 10. From the results in Table 4, for instance, Kitui Sec. School shows the longest long rains duration in 1997 (107 days). The same year also shows the shortest short rains duration (7 days). The result is confirmed in Kitui Water Office for the year 1997.

Rainfall characteristics in semi-arid Kitui district of Kenya

55

3.4 Drought severity This was based on the anomalies of the seasonal totals for each year for the two stations. A normal expectation was taken to be ±0.5 s.d. For Kitui Water Office, 1985, 1990 and 1995 were severe in terms of the long rains totals. The same years are also severe for the short rains. In Kitui Sec. School, 1980, 1985, 1990, 1995 were severe during the long rains as well as during the short rains. The seasonal totals for the very dry years can be used as a basis for planning for a water storage facility. Examples of the anomalies are given in Figure 11 and Figure 12.

Figure 9. Dry spell lengths during long rains—Kitui Water Office.

Water resources of arid areas

56

Figure 10. Dry spell lengths during short rains—Kitui Sec. School. 3.5 Drought frequency Gamma distribution was used to fit the seasonal totals for the period of study. The results are shown in Figures 11–12 and Table 4, while Figure 13 indicates the gamma distribution. In Kitui Water Office (KWO), the largest seasonal total for the long rains is 768.3mm and has a return period of 25.9 years. The smallest seasonal total is 81.1mm with a return period of 1.0 years. For the short rains, the largest total is 848.6mm with a return period of 20.1 years and smallest total is 98.76mm with a return period of 1.0 years. In Kitui Sec. School (KSS), the largest seasonal total for the short rains is 1022.0mm with a return period of 15.7 years while the smallest seasonal total n is 205.4mm with a return period of

Table 4. Values expected frequency MAM(KWO) OND(KWO) Gamma dist. Gamma dist. Mean 344.1 & Mean 441.2 & k of 2.929 k of 4.133 <=81.1 0.71 81.1 to 106.60.64 106.6 to 122.30.49

<=98.760.23 98.76 to 167.10.98 167.1 to 175.80.19 175.8 to 198.60.57

OND(KSS) Gamma dist. Mean 578.1 & k of 4.799 <=205.40.73 205.4 to 244.50.57 244.5 to 261.20.30 261.2 to 319.31.30

Rainfall characteristics in semi-arid Kitui district of Kenya

122.3 to 198.6 to 178.52.15 244.11.40 178.5 to 244.1 to 183.80.23 257.40.46 183.8 to 257.4 to 192.60.38 306.31.82 192.6 to 306.3 to 197.10.20 424.90.72 324.9 to 197.1 to 20 1.70.20 325.40.02 201.7 to 325.4 to 286.63.72 342.30.66 286.6 to 333 342.3 to 1.86 405.32.40 333 to 344.50.43 405.3 to 417.70.45 344.5 to 417.7 to 391.81.61 427.20.34 391.8 to 427.2 to 477.2 425.91.00 1.66 425.9 to 477.2 to 459.70.87 507.60.91 459.7 to 507.6 to 51 465.10.13 10.10 465.1 to 511 to 546.81.59 540.70.80 546.8 to 540.7 to 594.80.68 570.40.72 594.8 to 570.4 to 658.20.67 632.61.25 658.2 to 632.6 to 768.30.72 672.10.64 >768.30.73 672.1 to 690.70.26 Maximum 690.7 to T=25.9 723.80.41 Minimum T=1.0 723.8 to 724.20.00 724.2 to 848.61.04 >848.60.95 Maximum T=20.1 Min T=1.0

319.3 to 333.30.37 333.3 to 338.30.14 338.3 to 402 1.85 402 to 537.14.21 537.1 to 571.41.01 571.4 to 583.50.34 583.5 to 592.30.24 592.3 to 626 0.90 626 to 69 1.9 1.56 691.9 to 727.30.73 727.3 to 791 1.13 791 to 848.50.82 848.5 to 909.50.70 909.5 to 979.30.61 979.3 to 10220.29 <10221.21 Maximum T=15.7 Minimum T=1.0

57

Water resources of arid areas

58

Figure 11. OND seasonal anomalies— Kitui Water Office. 1.0 years. For the long rains, the totals are 757.1mm with a return period of 29.8 years and smallest total is 73.22mm with a return period of 1.0 years. 4 CONCLUSION Three rainfall regions were delineated. The longest spell is in 1997 (107 days) during the long rains. The same year also showed the shortest spell (7 days) during the short rains Average onset for long rains is on day 82, cessation was on day 126. The years 1980, 1985, 1990, 1995 were severe during long rains as well as during short rains. Average onset for long rains is on day 82, cessation was on day 126 using the two methods. The largest seasonal totals had return periods ranging between 15–25 years while the smallest seasonal totals had return period on average of 1 year, and indication of frequent drought. This information is crucial if water management and planning is to meet a particular demand in a specified duration of water stress. REFERENCES Basalirwa, C.P.K. 1991. Raingauge network designs for Uganda, Ph.D Thesis, Univ. of Nairobi, Kenya. Basalirwa, C.P.K. et al. 1995. The climatological zones of Tanzania based on rainfall characteristics. Water Resources Engineering Research Report, University of Dar-es-Salaam. Castell, R.B. 1966. The Scree test for the number of factors. Multivar. Behav. Res., 1:245–276.

Rainfall characteristics in semi-arid Kitui district of Kenya

59

Dyer, T.G.J. and Tyson, P.D. 1977. Estimating above and below normal rainfall periods over South Africa. Journal of Applied Meteorology 16(2):145–147. Dent, M.C., Schulze, R.C., Wills, H.M. & Lynch, S.D. 1987. Spatial and temporal analysis of the recent drought in the summer rainfall region of Southern Africa, Water SA, 13(1):37–42. Drosdowsky,W.1993. An analysis of Australian seasonal rainfall anomalies 1950–1987I: J climat. 13:1–30. Mhita, M.S 1990. The onset and cessation of rains andr importance for cropping strategies in Tanzania. Ogallo, L.A. 1988a. The spatial and temporal clusters of the East African Seasonal Rainfall anomalies derived from Principal component analysis. J.Climatol. 6:1–23. Ogallo, L.A. 1989. The spatial and temporal patterns of East African Seasonal Rainfall derived from Principal component analysis. J.Climatol. 9:145–167. Yevjevich, V. 1967. An Objective Approach to Definitions and Investigations of Continental Hydrologic Droughts. Hydrol. Paper 23, Colorado State University, Fort Collins, Colorado. Zucchini, W. and Adamson, P.T. 1984. The occurrence and severity of droughts in South Africa. WRC Report No.91/1/84, Department of Civil Engineering, University of Stellenbosch, South Africa.

Quantification of the impact of irrigation on the aquifer under the Vaalharts Irrigation Scheme R.G.Ellington, B.H.Usher & G.J.van Tonder Institute for Groundwater Studies, University of the Free State, Bloemfontein, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The Vaalharts is the largest irrigation scheme in the country, with water imported from the Vaal River to supplement to rainfall in the area. Approximately 30000 hectares of land is currently being irrigated. The salinity of the irrigated water has steadily increased over time. Several previous research projects have been undertaken to determine the fate of the added salts. The conclusion in these reports is that a very large proportion of the salts added to the subsurface due to irrigation are not returned to the surface water. The underlying aquifer was postulated as sink for these salts, with limited storage capacity. Once this capacity has been exceeded, a flow reversal was postulated to occur. This process is likely to add a tremendous salt load (estimated to be approximately 100000t/year) to the Harts river system. The adverse effects of such an addition would be catastrophic to the irrigation scheme, and all downstream irrigation schemes and water users. Investigations into the hydrogeology and hydrochemistry were conducted to quantify the impact of irrigation on the groundwater resources. This included drilling, aquifer testing, groundwater monitoring and empirical and numerical modeling. Findings included a water and salt balance for the area and the understanding of the underlying lithology as fractured rock aquifer.

1 INTRODUCTION The Vaalharts Irrigation Scheme was initiated approximately 55 years ago. It is the largest irrigation scheme in South Africa at approximately 32000ha. Vaal River water is transferred via an extensive canal system from Warrenton into two subsequent canals, namely the North Canal and West Canal. This research entailed a detailed groundwater investigation including drilling of additional boreholes, aquifer parameter determination from slug, pump and tracer testing,

Quantification of the impact of irrigation on the aquifer

61

groundwater monitoring of water levels and water quality for more than a year, and the construction of water and salt balances using empirical and numerical modelling techniques. 1.1 Overview of the study area The geology within the Vaalharts valley is largely sedimentary of Karoo age, although the pre-Cambrian basement geology appears igneous. The Vaalharts valley is largely overlain by aeolian Kalahari sands. Also of Quaternary age are calcretes and alluvial gravels. Below these Quaternary sediments lie shales, tillites and mudstones. The preCambrian igneous lithologies form the lower boundary of the system. Natural drainage has been found to be poor. This is attributable to the flat topographical gradient, and typical soil profiles found in the area. The upper, generally impermeable calcretes are found at depths varying between 0m and 5m (Gombar & Erasmus, 1976). According to Streutker (1977) the water table was found to be lying at approximately 24m below ground level (mbgl) for the period between 1935 and 1940, although it seems that no comprehensive borehole drilling to determine the water levels in the irrigation area was undertaken across the entire scheme (Herold & Bailey, 1996). No extensive measurement of the water levels seems to have been undertaken during the period of 1940’s to 1970’s. To combat waterlogging, a comprehensive network of 240 subsurface drains was installed between the years 1976 and 1979 at an approximate depth of 1.8mbgl. The drains were found to successfully control the water table, and in so doing, improve the crop yields. In 1976, prior to the drains’ installation, approximately 3000ha of soils were saline or saline-sodic to a depth of 0.3mbgl. The end of 1977 had reduced this reduced to approximately 1500ha, while in 1980 there remained approximately 1000ha of saltaffected soils (Herold & Bailey, 1996). 1.2 Previous investigations Research has been conducted in the Vaalharts since the 1960’s addressing increasing water levels and salinisation within the irrigation area. Most recent was a report by Herold & Bailey (1996) discussing the long-term salt balance for the Vaalharts Irrigation Scheme. This report stated an annual loss of 100000t of salts to groundwater, and predicted that, as these salts were not being measured in the Harts River, that they would be seen in the form of a sudden salt reversal to the Harts River, thereby adding a massive strain to an already stressed river system. The basic hypothesis for this to occur was the existence of a “perched” aquifer below the irrigation scheme, which in turn is underlain by a deep aquifer. This deep aquifer would be the sink of these excess salts. Once the postulated deep aquifer’s storage was exceeded, these salts would be added to the Harts River, causing a water quality deterioration in the downstream Spitskop Dam. The irrigation schemes in the lower Orange River such as those at Douglas would be negatively impacted on. In another study Gombar & Erasmus (1976) sampled various boreholes in the North Canal area. The average TDS at the time was determined to be 1005mg/l.

Water resources of arid areas

62

2 METHODOLOGY 2.1 Hydrocensus An initial field recognisance study was undertaken. Literature indicated that 41 diamondprospecting boreholes had been drilled during the 1970’s. It was hoped that these boreholes could provide access to the aquifer. During a hydrocensus inspection to locate these boreholes, all were found to either be destroyed by farming practices, or have been blocked by stones. A second hydrocensus across the Vaalharts Irrigation Scheme to obtain an indication of the groundwater response was conducted. The boreholes located during this hydrocensus were equipped with mono-pumps, and had been encased in concrete. In total, 22 boreholes were discovered. Only pumped samples were possible from these boreholes. 2.2 Field investigation The need for drilling was made evident by the relative lack of accessibility to open boreholes, as the majority of boreholes are covered by cement blocks fitted with monopumps. The drilling method used was air percussion, with the boreholes drilled to a diameter of 0.165m, using a drill rig provided by the Department of Water Affairs and Forestry. A total of 17 boreholes drilled across the Vaalharts Irrigation Scheme, of which three were located on the riverbanks of the Harts River, and the remaining 14 were drilled on the plots. In order to determine the hydrogeologic variation over the extent of the lithology, the first seven boreholes were drilled to random depths until they reached the lava bedrock, while the remaining 10 boreholes were drilled to a depth of approximately 20m. The boreholes were predominantly cased with a 4mm steel casing, although three of the boreholes drilled on the banks of the Harts River were cased with Johnson screens. Furthermore, three of the seventeen boreholes drilled during this research were equipped with piezometers. In all boreholes, casing and screens used for the borehole construction were slotted to allow free flow of groundwater through the boreholes and accurate groundwater investigation. The boreholes were slotted from a metre below the depth of the Kalahari sands to prevent clogging of the borehole from these sands. Piezometers were installed in three boreholes across the North Canal area. The boreholes used are believed to present an accurate representation of the general geology in the Vaalharts. The piezometers were installed to test the conceptual model of Herold & Bailey (1996), where they assumed there to be two aquifers in the Vaalharts—an upper, perched aquifer relating to the calcretes, and a deeper aquifer. 2.3 Aquifer testing Three types of aquifer tests were undertaken to obtain the hydraulic parameters of the aquifer underlying the irrigation scheme.

Quantification of the impact of irrigation on the aquifer

63

2.3.1 Slug tests A slug test is a method used to measure the hydraulic conductivity or transmissivity of a borehole. This is done by measuring the rate of recovery or recession in the borehole, following a sudden addition into the borehole or extraction of water from the borehole of a known volume. The Bouwer and Rice (1976) equations were used to analyse the slug tests. 2.3.2 Pumping tests Multi-rate and constant discharge rate pump tests were undertaken on selected boreholes. Use was made of traditional analysis methods such as the Bisroy-Summers (1980) method for multi-rate tests, the Cooper-Jacob method for constant rate tests and the more recent suite of methods based on the Flow-Characteristics methodology (van Tonder et al., 2002). 2.3.3 Tracer tests Tracers are identifiable substances that, from the examination of their behaviour in a flowing medium, may be used to infer the general behaviour of the medium (Riemann, 2002). Leap and Kaplan first described the single-well Injection Withdrawal Tracer test for the estimation of groundwater flow velocities in 1988. The single-well Injection Withdrawal Tracer test is conducted by injecting a known volume of tracer solution into the test borehole, allowing the tracer to drift under the influence of the natural hydraulic gradient for a period, and then removing the tracer by pumping the test borehole to recover the tracer. Adopted methodologies as described by Riemann (2002) were used, with NaCl and NaBr as artificial tracer. 2.4 Groundwater monitoring Groundwater monitoring has been ongoing since inception of the drilling program. Six groundwater-monitoring runs have been undertaken in the period April 2003 to March 2004. Groundwater levels have been measured using electronic contact dip meters. Sampling of the boreholes has been undertaken with depth-specific sampling equipment. Hydrochemical profiling using YSI- Sonde 6000 multi-parameter probe which measures pH, specific conductance, dissolved oxygen and redox potential with depth, was done at three occasions. 2.5 Numerical and analytical methods Empirical and numerical methods were employed. 2.5.1 Analytical methods Darcy’s law was used to determine an expected flux to the river. Salt loads were calculated using the Ogata equation. The salt load estimation in the GW reserve program uses the Ogata equation (from Freeze and Cherry, 1998):

Water resources of arid areas

64

where: l=distance along the flow path, v=the average linear water velocity, t=time, C=concentration at time t, C0=initial concentration and D=the coefficient of molecular diffusion, for the solute in the porous medium. The expected salt leaching from the irrigation scheme was obtained from du Preez et al., (2000), who used the Aragues model and the Soil Water Balance to determine salt leaching from the scheme. These values were used as input in the empirical and numerical models for salt balance calculations. 2.5.2 Numerical models The numerical model used to simulate the aquifer system in the Vaalharts irrigation area was Modflow. Modflow is a modular two- or three-dimensional finite difference groundwater flow model that was developed by McDonald and Harbaugh of the United States Geological Survey for the purpose of computation of hydraulic heads in saturated porous medium with uniform water temperature and density (Harbaugh & McDonald, 1996). The Modflow mass transport program used during the Vaalharts transport modelling was MT3D. The Vaalharts numerical groundwater model is, as all groundwater models, a representation of the naturally occurring conditions. Certain assumptions therefore had to be made, while certain limitations also persisted in representing natural conditions. The following assumptions were made: ● The rivers in the area were treated as fixed heads. ● As there is no significant groundwater extraction in reality due to the water allocation from the Vaal River, no discharge was included. ● As there are large volumes of water being applied by irrigation, a higher volume for recharge was applied. ● The basic lavas in the stratigraphy were accepted as being the lower boundary within the stratigraphy due to their relatively impermeable nature. The model was assigned 320 rows and 152 columns with a cell size of 250m×250m. This equates to a model area of 3040km2. The model area’s co-ordinates are −3120000, −2000 (lower right corner) to −3040000, −40000 (upper left corner). The Vaalharts model constructed during this project made use of a two-layer model. Confined conditions were applied to these layers. The layer depths were based upon geology encountered during literature reviews of Vaalharts specific data and drilling that took place during the course of this project. The upper layer was assigned values for the sands, according to geological logs drilled during this and other projects, averaged at approximately 6m. The lower layer was assigned average values for the calcretes, clays, gravels and shales due to their relatively similar range of depths, and depths of the geological strata from borehole logs. The various pre-determined areas therefore each had

Quantification of the impact of irrigation on the aquifer

65

separate hydraulic conductivities applied to them, based upon knowledge of the geology and tested aquifer parameters. For instance, areas with a higher degree of gravels were assigned a higher hydraulic conductivity for that area. In other areas, where fractures with significant yields were encountered, an increased hydraulic conductivity was assigned. The drain package in MODFLOW was applied to the North Canal and West Canal areas during the simulations. The drains applied in the model accurately represent the processes occurring naturally in the Vaalharts system. The subsurface drains were positioned 2mbgl to simulate natural conditions. For the mass transport simulations, an initial concentration of 500mg/l TDS was assigned, while the input concentration from irrigation was obtained from the salt leaching models described by du Preez et al., (2000). 3 RESULTS 3.1 Geology From the drilling a consolidated geological model was constructed using the Rockworks (Version 3.4.1.6, Rockware Incorporated) program. The geology model shows thicker shales to the northern side of the Vaalharts Irrigation Scheme, with thinner shales in the south of the scheme. In addition, while the model indicates calcretes throughout the scheme, the calcretes are more pronounced in the southern half of the scheme. The southern half of the scheme’s geology is represented more by gravels and clays. This may be due to possible erosion and deposition as the Harts River meandered during the Vaalharts’ geological history. 3.2 Aquifer parameters The aquifer parameters obtained through pump and tracer testing show the underlying strata to be heterogeneous, with high transmissivity values where fracture zones are encountered. The tracer test results indicate a similar range of values, with high velocities occurring wherever fracture zones exist. 3.3 Water levels The water level variation over the time of measurement was not significant, with minor summer and winter levels observed. Apart from a few exceptions, the water levels in the area are very similar, varying between 1.6–2.0m below surface. This level coincides with the depth of the installed drainage systems prevalent in the area. Water levels in piezometers installed within the same borehole exhibited less than 1cm variation between piezometers installed to different depths.

Water resources of arid areas

66

3.4 Water quality The groundwater quality in the area varied spatially and is dependent on the geology and position relative to the irrigation. The majority of the electrical conductivity values are between 100mS/m and 270mS/m. The average TDS in the irrigation area was determined to be 1350mg/l. The two boreholes illustrating the highest and lowest electrical conductivities respectively, are interestingly enough both present on the same plot, and are within a 50m distance of each other. Borehole 6L16-1, illustrating the lowest conductivity of the samples obtained, is drilled within 10m of a canal, which seems to be leaking water into the groundwater system via cracks in the concrete. Borehole 6L16-2 is however located within 10m of irrigated land. Several of the groundwater samples have relatively high nitrate values, but considering this is a heavily

Table 1. Transmissivity values for tested boreholes. Borehole

Transmissivity (m2/d)

1G14-1 1K10-1 6L16-1 6L16-2 1D3-1 1D7-1 2J14_RIV-3 8H14-1 2J5-1

43.3 31.6 4.2 4.5 0.21 194.0 58.0 1.2 123.0

Table 2. Tracer test results. Borehole number 1B10-1 1D3-1 1D7-1 2J5-1 8H14-1

Darcy velocity (m/d) 2 1.5 3 1 22

Seepage velocity (point dilution) (m/d) 21 15 29 9 217

Quantification of the impact of irrigation on the aquifer

67

Figure 1. Electrical conductivity values of the samples taken from the boreholes drilled during this project. Table 3. Values used for Vaalharts Water Balance using empirical values, and model values where possible empirical values calculated for the Vaalharts Water Balance (Mm3/annum). Name North Canal West Canal Rainfall Groundwater going to river Canal Tailends Recharge Drainage Runoff Evapotranspiration Totals Difference (Inflow—Outflow)

Incoming water (Mm3/a)

Outgoing water (Mm3/a)

272.01 42.97 309.60 14.34

624.58 −0.41%

23.35 28.38 23.63 4.10 533.30 627.16

cultivated area the 50th percentile value for the groundwater of 2.2mg/l N is considerably lower than expected. As nitrate is often used as a tracer to highlight the effect of cultivation on water quality (e.g Pulido-Bosch et al., (1999)) such a low value would seem to indicate that vertical migration of salts from the cultivated lands to the groundwater is less pronounced than previously expected. This is confirmed by the low

Water resources of arid areas

68

potassium values, another key constituent in fertilizers, where 95% of the measured values fall under 15mg/l. Consideration of the major parameters using interpretive diagrams such as Piper plots, showed no dominant anion or cation, although some of the cations do tend towards the Na+K field. Comparison with surface waters and geology indicated a Mg- enrichment in the Dwyka shales, and sulphate largely from incoming Vaal River water used for irrigation. The in situ water quality, as determined through hydrochemical logging, exhibited only minor variations with depth. No significant evidence of stratification of poorer water quality was observed in any of the boreholes. 3.5 Water and salt balance Water balances were determined using the numerical model and empirical calculations for various scenarios.

Table 4. Salt balance permutations Vaalharts salt balance (tons/year). Components

Option 1 Option 2

North Canal 112884 112884 West Canal 17832 17832 Groundwater going to 15873 15873 river Canal Tailends 17977 17977 Drainage 17979 17979 Recharge1,2 84287 111758 25962 25962 Salts taken up by crops1 Fertilizer addition1 48302 48302 1900 1500 Salts in soils1 Incoming salts 179018 179018 Outgoing salts 163979 191050 Incoming less outgoing 15039 −12032 Percentage difference 8.401% −6.721% 1 Based on values obtained from du Preez et al., 2000 and the numerical model. 2 Upper and lower values of calculated salt leaching used.

For the purposes of the salt balance water qualities of various water types were obtained from DWAF, previous reports and measured in this project. Combination of the water balance, these concentration values and the output of salt leaching models reported by du Preez et al., (2000), allowed salt loads to be calculated. The following was used to for the salt balance. There are several permutations of these options but these give similar results regarding the overall salt balance. Of importance is the recharge salt addition to the groundwater system. Using the median value of approximately 98000t/year of salts added to the

Quantification of the impact of irrigation on the aquifer

69

groundwater and the assumption that the net storage in aquifer remains relatively constant over time, the expected net increase in TDS should be in the order of 14mg/l. 4 DISCUSSION AND CONCLUSIONS The main aim of the study was to ascertain the impact of many years of irrigation on the groundwater resource. More specifically, emphasis was placed on the assessment of previous hypotheses related to the aquifer system and salt migration within this system. 4.1 Previous hypotheses The regular water level measurements over a period of a year in the seventeen boreholes have shown that the subsurface drainage installed by the farmers is effectively controlling groundwater levels. All the water levels in close proximity to such drainage exhibit limited variation over time, and lie at a depth consistent with the installed drains. The installed piezometers in several boreholes have shown that the water levels in deeper and shallower systems are within a few centimetres of one another. The water quality in these piezometers is also very similar for each borehole. Hydrochemical profiling has indicated that no significant stratification of water occurs. All these factors point to the conclusion that the deeper lying aquifer, thought to be the salt sink, does not behave independently. The system is dynamic enough, and the shallow impermeable layers too localised, to have a system of cascading groundwater finding its way to a deeper system. This finding has positive consequences, in that no sudden catastrophic event of salt reversal is likely to occur. However, the indications are also that the groundwater quality is showing a steady deterioration over time. 4.2 Groundwater quality The ongoing monitoring has shown the groundwater quality to be fairly poor, but not as saline as was suspected. The average TDS of 1350mg/l compared to 1050mg/l in 1976 is cause for concern. The numerical model has shown that groundwater is expected to leave the scheme and contribute to water quality deterioration downstream in the Harts River system. The observed water quality change is approximately 13mg/l year. Based on the salt balance calculations, which takes the incoming and expected outflowing volumes into account, the increase was expected to be around 14mg/l. These two results are therefore in good agreement. 4.3 Water and salt balances The empirical water and salt balances showed good agreement with one another. These balances showed that irrigation is the most important driver on water quality and volumes in the system. The salt balance also highlighted the fact that the greatest contribution to the incoming salt load is the irrigation water sourced from the Vaal River. The salts added in this way are more than double those from fertilizer addition and management of this incoming water is therefore the key to the salt accumulation in the irrigation scheme.

Water resources of arid areas

70

Short of addressing the upstream Vaal River practices and means by which to ensure cleaner water entering the Vaalharts system, the Vaalharts itself needs to be addressed. The logical approach to ensure lower salinity water enters the groundwater in the Vaalharts system is to make use of less water. Since agriculture accounts for nearly 70% of all water withdrawn from rivers, lakes, and underground aquifers for human use, the greatest potential for conservation lies with increasing irrigation efficiencies (Clarke, 1991). What is needed in the Vaalharts is a more efficient manner of irrigation, where less water is applied per unit area, and therefore fewer salts enter the groundwater via leaching. A more efficient means of irrigating would be drip irrigation, with a field application efficiency of 95%, which is 40% to 60% more efficient than gravity systems (Postel, 1997). The installation of drip irrigation in the Vaalharts would increase the efficiency of irrigation, thereby reducing the volumes of water needed. This would simultaneously decrease the mass of salts applied to the Vaalharts Irrigation Scheme, and reduce the tonnage leaching to groundwater and, eventually, entering the Harts River. REFERENCES Annandale, J.G., Benadi, N., Jovanovic, N.Z. & Du Sautoy, N. 1998. SWB: A user friendly irrigation scheduling model. Soils and Crops towards 2000 Congress, South African Society of Crop production, Alpine Health, Kwazulu-Natal. Aragues, R.M. 1996. Conceptual irrigation return flow hydrosalinity model. In K.K.Tanji (ed.). Agricultural Salinity Assessment and Management. Am. Soc. Of Civ. Eng., New York. Clarke, R. 1991. Water: The international crisis. London, Earthscan:. 193. Birsoy, Y.K. & Summers, W.K. 1980. Determination of aquifer parameters from step tests and pumping data. Groundwater, 18:137–146. Chiang, W.-H. & Kinzelbach, W. 2000. Processing Modflow (PMWIN), Version 5.1. The Institute for Groundwater Studies, University of the Free State, Bloemfontein, South Africa, 9300. Du Preez, C.C., Strydom, M.G., Le Roux, P.A.L., Pretorius, J.P., Van Rensburg, L.D. & Bennie, A.T.P. 2000. Effect of Water Quality on Irrigation Farming along the Lower Vaal River: the Influence on Soils and Crops. WRC Report No. 740/1/00. Water Research Commission. Harbaugh, A.W. & McDonald, M.G. 1996. User’s documentation for MODFLOW-96, an update to the U.S. Geological Survey modular finite-difference ground-water flow model: U.S. Geological Survey Open-File Report 96–485, 56p. Herold, C.E. & Bailey, A.K. 1996. Long Term Salt Balance of the Vaalharts Irrigation Scheme. Water Research Commission. Gombar, O. & Erasmus, C.J.H. 1976. Vaalharts Ontwateringsprojek, Technical Report GH2897. Department of Water Affairs. Leap, D.I. & Kaplan, P.G. 1980. A single-well tracing method for estimating regional advective velocity in a confined aquifer theory and preliminary laboratory verification. Water Resources Research, 23(7): 993–998. Postel, S. 1993. Water and Agriculture. Water in Crisis. New York, Oxford University Press. pp. 55–66. Pulido-Bosch, A., Bensi, S., Molina, L., Vallejos, A., Calaforra, J.M. & Pullido-Leboeuf, P. 1999. Department of Hydrogeology, University of Almeria. Canada, Spain. Riemann, K. 2002. Aquifer parameter Estimation in Fractured Rock Aquifers using a combination of Hydraulic and Tracer Tests. PhD thesis. Institute for groundwater studies. South African Weather Service. 2002. Climate data. http://www.weathersa.co.za/

Quantification of the impact of irrigation on the aquifer

71

van Tonder, G., Bardenhagen, I., Riemann, K., van Bosch, J., Dzanga, P., Xu, Y. 2002. Manual on Pumping Test Analysis in Fractured-Rock Aquifers. WRC Report No. 1116/1/02. Water Res. Comm.

Theme B: Groundwater recharge: natural and artificial

Groundwater development—identification of artificial recharge areas in Alla, Eritrea K.S.Viswanatham, Filmon Tesfaslasie & Michael Asmellash Water Resources Department, Government of Eritrea, Asmara, Eritrea, N.E. Africa Arun Kumar Department of Earth Sciences, University of Asmara, Asmara, Eritrea, N.E. Africa S.A.Drury Department of Earth Sciences, The Open University, Milton Keynes, UK Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Remotely sensed data acquired by the EO satellite have been analyzed to demarcate potential target zones for the development of future irrigation projects in a part of Eritrea. The study area is one of the important centers of horticulture activities and is currently facing water problem. Various lithologic units, tectonic signatures, land use and geomorphology related to groundwater assessment have been identified and interpreted from the imagery. The northern part of the area, covered by alluvial fans and presence of NW-SE dykes indicates favorable potential areas for groundwater development. Based on present investigation, construction of check dam in the Ghadien River and subsurface dam near Bazit Village is recommended in order to improve groundwater potential in the area.

1 INTRODUCTION The advent of polar-orbiting satellite remote sensing has provided hydrogeologists with a sophisticated and reliable tool for rapidly assessing natural resources of an area with reasonable accuracy. The focus of this paper is on the analysis of remotely sensed data combined with ground truth to delineate geological and geomorphologic patterns and their effect on groundwater occurrence and movement. The area selected for the study lies south east of Asmara, the capital city of Eritrea, in Sub-Sahelian Africa (Fig. 1). Alla-Ghadien and its vicinity are well known for horticultural activities. Irrigation is from open dug wells and in a limited way from bore

Water resources of arid areas

76

wells. The continual exploitation of water resources has resulted in gradual decline of water levels and also a decrease in well yields (Habteab 2000). The area lies within a depression at the edge of the Red Sea escarpment, and is an intermontane basin enclosed by several outstanding ridges. Ghadien is the administrative center of five villages, Alla, Belesto, Adi-Asambo, Adi-Moya and Ghadien itself. It is located 16km northeast of Dekemhare town. The elevation of the area ranges from 1680m to 1936m above mean sea level. The horticultural activities in the area commenced during the Italian colonization. In the early times, sufficient groundwater was found at shallow depths (five meter); however at present it is difficult to get water in some areas, even at depths of 30 meters (Habteab, 2000). The total population of the Ghadien administrative center is about 2773, with a total cultivated area of 2888.24 hectares. The average annual rainfall is 463mm. Based on the information of the Ghadien administrator the number of wells is about 600, of which 200 are dry (Habteab, 2000). Some of the existing problems in the area are—acute shortage of groundwater, lowering of water table, high density of wells with close spacing, over exploitation of groundwater, mismanagement of the existing water, lack of assessment of groundwater potential, extensions of irrigated areas,

Figure 1. Location map. construction of wells and small dams without knowing the groundwater potential. A reconnaissance survey in April 2001 indicated a lowering of water levels and decreasing

Groundwater development—identification of artificial recharge areas in Alla, Eritrea

77

well yields. The main objective of the present work is to identify the existing hydrogeological problems, to test the water quality at random sites and to evaluate water use and management. With the help of image interpretation and field checks, areas for future groundwater development and suitable site for check dam and subsurface dam have been suggested. The remotely sensed data employed are from the NASA EO-1 satellite’s Advanced Land Imager (ALI) instrument (NASA, 2001). The deployment aims at testing the suitability of 9 spectral bands in the visible to short wave infrared, with 30m resolution, and a single panchromatic band with 10m resolution, for the eventual replacement of the Landsat-7 Thematic Mapper. The chosen bands include 6 similar to that of the TM instrument, with an additional 3 that augment possibilities for vegetation and mineral discrimination. Sharpened by the 10m panchromatic band, various combinations of the spectral bands offer unprecedented (for Eritrea) opportunities for resource assessment. 2 PHYSICAL SETTING Eritrea is located on the western flank of the Red Sea at about 12.5°–18° North, 36.5°– 43.5° East. Its border length in the north and west with Sudan is about 624km, that with Ethiopia to the south is about 917Km, and that in the southeast with Republic of Djibouti is about 104km. With a total area of about 124000km2, it has more than 350 islands and a coastline of more than 1200km. The population of Eritrea is approximately 3.5 million. Eritrea has five major river basins, namely Mereb-Gash, Setit, Barka-Anseba, Red Sea and Danakil depression. All the rivers (except the Setit) and their tributaries are mostly seasonal and intermittent. It has four physiographic regions namely the Central Highlands, Western Lowlands, Eastern Lowlands and Coastal Lowlands. The average temperature ranges from 3°C–28°C in the highlands and 20°C–48°C in the lowlands. Eritrea being an arid and semi-arid country is not endowed with rich water resources. It has a vulnerable environment due to recurrent and devastating droughts, being part of Sahelian Africa. Rain-fed agriculture is the main occupation of most rural people. The majority of the population depends on groundwater as the main water-supply source. Rainfall is torrential in nature, i.e. high intensity and short duration, and is monsoonal. Annual precipitation ranges from 300mm to greater than 800mm in the Central Highlands and Southwestern Lowlands, 200mm to 300mm in the Northwestern Lowlands, and 100 to 200mm in the Eastern and Southeastern Lowlands. Situated close to the Red Sea escarpment, the Alla-Ghadien area receives intermittent orographic rainfall, as well as that during the two monsoonal periods in March–April and June–September. 3 PREVIOUS WORK Drury and Berhe (1993) reported significant regional geological controls for groundwater in Eritrea and their expression on satellite images. According to them, the main potential for groundwater developments occur as fracture systems, carbonates with enhanced permeability, granites (which develop deep, coarse and porous soils and extensive joint

Water resources of arid areas

78

systems), fissile rocks in shear zones, deep basins of unconsolidated sediments in the mountains and coastal plains, igneous intrusion which act as natural barriers, and outcrops of lava and laterite. Many such features have been identified on images of Eritrea. Some of them were considered to be targets for locating bore wells in difficult terrains of northwestern Eritrea, and drilling them met with a success rate of 85%. Drury et al. (2001), focusing on the hydrogeological potential of fracture systems, indicated that NNW-SSE Precambrian shear zones, normal faults roughly parallel to those earlier structures and prominent East-South-East-West-North-West dilatational fractures offer considerable scope for groundwater development. Prospect geophysical profiling across several of these structures in both lowland and highland terrains revealed conductive features believed to relate to saturated zones in large, regionally extensive fractures. Zerai and Solomon (1993) identified five main hydrogeological units: (i) Unconsolidated sediments with variable inter granular permeability. (ii) Volcanic rocks (basalts) with fracture and fissure permeability. (iii) Fissured and karstic carbonate aquifers. (iv) Metamorphic and intrusive rocks with localized low to moderate permeability along fractured and weathered zones. (v) Aquitards and aquicludes and groundwater barriers (acid to intermediate volcanic).

4 GEOLOGY Knowledge and understanding of the geological events of an area are important for groundwater investigations. The geology of Eritrea is made up of a Precambrian basement complex comprising high-to low-grade metamorphic rocks and associated intrusives, which are overlain by predominantly Mesozoic sedimentary rocks and Tertiary to Quaternary volcanic and sedimentary rocks (JICA Report 1998 Drury and Berhe 1993). Precambrian granites are exposed in the eastern, western and northern parts of Alla-Ghadien. The southern part of the study area is a wide and flat plain between actively rising ridges of metamorphic rocks. The geological events in Eritrea are summarized as follows: (1) Precambrian: Formation of the crystalline basement complex and its associated intrusive rocks; (2) Paleozoic: Peneplanation of the basement complex and deposition of sparse sedimentary rocks; (3) Jurassic: Transgression-regression of the Mesozoic sea, which deposited lower sandstone, Adigrat sandstone and Antalo limestone during subsidence towards the Indian Ocean; (4) Upper Eocene-Miocene: Uplift forming domes, extensional fault systems, basaltic flood volcanism and opening of the Red Sea rift system; (5) Miocene Period: Formation of upper sandstone; and (6) Quaternary Period: Formation of alluvial, eluvial and colluvial unconsolidated sediments.

Groundwater development—identification of artificial recharge areas in Alla, Eritrea

79

5 HYDROGEOLOGY The reconnaissance survey done in Ghadien and Alla areas shows that the plain is covered by Quaternary alluvial and colluvial sediments and with subordinate outcrops of granitic intrusive rocks. There are two types of aquifers in the study area: alluvial sediments and weathered and fractured acidic granitoids. Moreover major and minor lineaments exist in the area, many being fracture zones or igneous dykes, which naturally serve as conduits for or barriers to groundwater flow respectively. That the NNW-SSE basaltic dikes serve as groundwater flow barriers is confirmed by high yields of wells around the dikes. Minor structures trending roughly NE-SW are found to serve as a conduit for the water flow, confirmed by the high yield of the wells in the Ghadien area, where the structures are dominantly observed. In 1994 and 2000 the Water Resources Department (WRD) did a well inventory in Alla-Ghadien area, encompassing location, static water levels and well design. Based on the well inventory data in 1994 there were 144 wells, the average well depth was 15m+ with static water level (SWL) 14.4m. In the year 2000 the average well depth was 17 meters & SWL 16.51. Therefore the lowering of water level in 4 years is 2 meters. In the Alla-Ghadien area, hand dug wells are very closely located, from approximately 20m to 100m apart. An ad-hoc assessment of the catchments of 1465km2 area based on rainfall for Alla has shown that the total groundwater recharge is 643Ha.m, while the total annual groundwater draft is 1062Ha.m leaving a negative balance of −419Ha.m (K.S.Viswanatham 2002). The above figure indicates “mining” of water. The Alla-Ghadien area could therefore be classified as an overexploited area. Areas where groundwater resource assessment shows stage of groundwater development more than 100% and both pre and post monsoon groundwater levels show a significant longterm decline are classified as over exploited areas. 6 METHODOLOGY An integrated approach that involves interpretation of remotely sensed data and groundbased ancillary investigations has been implemented for the Alla-Ghadien study. Groundwater zonation was prepared using various thematic maps at 1:50,000 scales, which include geology, geomorphology, lineament trends and land use/landcover. Field survey was done to correlate the image characteristic to ground feature to confirm the interpretation. Specific field data, such as well inventories of SWL, yield, depth, diameter, quality of water and drilling logs were collected. Images (bounded by UTM 7500000–7516000 E 1666000–1682000 N) of the Alla-Dekamhare area were from an EO-1 overpass in April 2001 and combine ALI bands 5, 4 and 3 as red, green and blue components, sharpened by the use of the ALI panchromatic band to modulate intensity. This combination is optimum for expressing vegetation cover, but does discriminate some lithologies and soil types, as well as revealing small-scale topographic features.

Water resources of arid areas

80

7 SIGNIFICANT PARAMETERS OF GROUNDWATER OCCURRENCE AS REVEALED BY IMAGE INTERPRETATION The various water-bearing and movement properties, and the controlling parameters in the study area based on image interpretation and ground check are as follows: Fluvial/alluvial sediments and alluvial fans, Tertiary laterite, Granite, Tertiary basaltic rocks, as lavas and dykes.

Geologically, the area is dominated by granitic terrain. The alluvial areas of Alla and Ghadien are eastward sloping plains surrounded by hills with mostly steep slopes. The alluvial sediments have a yellowish colour and define a roughly triangular area. Tertiary laterites are indicated by grayish white color in irregular shapes along the streams and on the plains. Areas of bedrock comprise granites, which are traversed by NNE to SSW basaltic dikes. In the imagery, the granites are bluish to dark blue in color and occupy topographical ridges. The granitic exposures form circular and semi-circular shapes, for example Bazit Hill near Bazit village. There is a possibility of marble being present as roof pendants in the granitic masses at the northern flank of the area (represented by light to dark gray tones with signs of bedding), which have to be checked and confirmed in the field. Topographically the area is surrounded by hills with undulating slopes, and almost plain on the center. These features could be observed north west of Ghadien. Dikes show as linear features mostly concentrated north of Ghadien and roughly north of the Alla plains. The major land use of the area is classified based on color, shape and texture as fallow land, barren rocky terrain, dense vegetation (mainly horticulture gardens), river sediments and sparse vegetation. The Ghadien, Sesah and Bazit Rivers drain the area. The drainage pattern is distributed and dendritic. All the rivers flow eastwards. Some flow features follow a structurally controlled direction. Low drainage density in the alluvial plains, which probably indicates high rates of infiltration, suggests good groundwater prospects in parts of them. 8 FIELD CHECKS From the fieldobservations, the area can be divided into high- and low-potential zones. In the upper Ghadien river successful hand dug wells are being pumped for 4–7 hours per day and irrigating 6–10ha of horticulture gardens, particularly citrus bushes. In addition, this high-potential area has been investigated by the geophysicist from the Ministry of Agriculture, and recommended to be a potential area for development. In the satellite image alluvial fans and fills clearly represent the upper Ghadien zone.

Groundwater development—identification of artificial recharge areas in Alla, Eritrea

81

The low-potential zones, as defined by the well inventory data, are represented in the satellite image by granitoid and granitic ridge with adjoining alluvial cover. The farmers are pumping water for 2–3 hours per day from wells there, some of which have been abandoned due to poor yields or non-availability of water. The geophysical surveys have also confirmed the absence of suitable aquifers at deeper levels in such low potential zones. A preliminary interpretation on the regional geology was attempted, based on the imagery. In order to corroborate this interpretation, a quick field trip was made. The field observations revealed additional information on the geology, which necessitated some modification to the preliminary assumptions about the geological set-up. The alluvial fans as interpreted from the imagery are observed in the field to be a thin cover of alluvium on granitic rocks. The possibility of carbonate rocks/Marble in the North West part of the area is ruled out as granites varying in color, composition and texture represent these outcrops. Only few small patches of marble, which are not mapable, have been reported, which could be the extension of carbonates that supply abundant spring water in the Maihabar area 15km to the north. Beside this, cherts and conglomerates are observed on the Northwest part (assigned as basement metasediments), diorite on the Northeast, and metabasalt on the Southeast part of the study area (Fig. 2). In the Alla area, four bore wells drilled in the range of 47 to 50 meters are reported to yield 1.4 to 2.2 liters per second. The depths of the alluvial cover at those sites are from 13 to 21 meters, followed by granites. In the Bazit area, there are 3 boreholes, each having depths of 50 meters, and yields of 1.4lps, 2.5lps and 5Lps. In the Ghadien area,

Figure 2. Geological map of AllaGhadien.

Water resources of arid areas

82

there is only borehole with a depth of 50 meters and yield of 2.5Lps. The alluvial cover is 21 meters. In Chuhot area 3 boreholes, with depths of 36 meters, 48 meters and 50 meters respectively proved to be highly successful, with yields of 4Lps, 5Lps and 5.5Lps. 9 TARGET AREAS FOR GROUNDWATER DEVELOPMENT (1) The areas around Alla (UTM 504000–508000 E and 1673000–168000) are favourable for groundwater exploitation. (2) The NNE-SSW dikes on the North flank of Alla plains, which can be, extrapolated beneath the alluvial cover show a positive indication for potential groundwater development as they form natural sub-surface barriers. (3) The source of water is mainly from the alluvial formation whose origin is by weathering of granites.

10 ARTIFICIAL RECHARGE SITES The alluvial formations in Alla plain provide suitable sites for sub-surface dams, infiltration galleries and check dams. This is due to fine sediments predominantly quartz gravels and sand derived from granitic terrain. The natural barriers namely the NNESSW dikes have to be taken into consideration while constructing the sub-surface dams and check dams for the suitability of the structures. 11 RECHARGE AND DISCHARGE AREA Areas can be delineated into recharge and discharge areas depending on whether water is added to or abstracted from the zone of saturation. In the case of the water table aquifer, usually the areas occupying higher elevations with deeper water tables constitute the recharge areas while the

Groundwater development—identification of artificial recharge areas in Alla, Eritrea

83

Figure 3. Water table contour map of Alla-Ghadien area. topographic lows with shallow water tables comprises natural discharge areas (Karanth, 1994). Therefore in the case of Alla-Ghadien area, Ghadien is a useful recharge area due to its higher elevation and deeper water table than Alla. This is shown by a water table contour map (Fig. 3) where the ground water flow is NNW to SSE of the area. To confirm this, further investigations are necessary. Implementing a program of artificial recharge and abstraction requires the construction of check dams and sub-surface dams at finalized sites. 12 CONCLUSION & RECOMMENDATIONS (1) Construction of check dams specifically in the Ghadien River because of high slopes with fractured granitic rocks and catchment area is recommended. (2) The sub-surface dam near Bazit village needs to be further probed by integrated geophysical investigations. (3) Bore wells drilled in the Chuhot area gave reasonably good yields (5ls−1), which if used through drip irrigation could cover large areas. Integrated groundwater studies are to be taken up for further bore well locations.

Water resources of arid areas

84

ACKNOWLEDGEMENTS The authors acknowledge with thanks Mr. Ghebremichael Hagos Director General for giving an opportunity for the investigations and permitting to publish this paper. The authors also thank Mr. Ghebremichael Temenewo and Mr. Michael Negash for discussions from time to time, which helped to improve the paper. Ms. Meron Teshome is acknowledged for her support in preparing the maps. REFERENCES Drury, S.A and Berhe, S.M. 1993. Remote Sensing and Water Exploration in Eritrea WRD Eritrea/EIAC/ GREADCO and Open University UK. (Unpubl) 6 pp. Drury, S.A and Berhe S.M. 1993. Accretion Tectonics in Northern Eritrea revealed by remotely sensed imagery. Geol Mag, 130(2):177–190. Drury, S.A, R.J.Peart M.E. & Andrews Deller. 2001. Hydro geological potentials of major fractures in Eritrea. Journal of African Earth Sciences, 2(2):163–177. Habteab.T. 2000. Groundwater Depletion in Alla Commercial Farm. Department of Environment, Ministry of Land Water and Environment (Unpubl). Karanth, K.R. 1994. Groundwater Assessment, Development and Management, New Delhi, Tata McgrawHill Publishing Company Limited: 720 pp. JICA Report. 1998. Study on Groundwater Development and Water Supply for Seven Towns in Southern region of Eritrea.Water Resources Department and Sanyu Consultants Inc., Japan. NASA, 2001. EO-1Science Validation Team Home Page, http://eol.gsfc.nasa.gov/science/SVTAuth.cfm Viswanatham, K.S. 2002. Water Resources Development Management of Critical Areas in Eritrea. Journal of Applied Hydrology, XV(4), Oct:21–25. Zerai Habteab. 1996. Groundwater and Geothermal Resources of Eritrea with the emphasis on their chemical quality: Journal of African Earth Sciences, 22:415–421.

Subterraneous injection of nutrient rich groundwater to the coastal waters K.K.Balachandran National Institute of Oceanography, Regional Center, Cochin Joseph Sebasgtian Paimpillil Center for Earth Research & Environment Management, Cochin, India Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Existence of a subterranean flow connecting a tropical backwater to coastal waters through submerged porous lime shell is inferred from nutrient distribution patterns in the coastal waters. The activation of ‘mud banks’ triggers high productivity in oligotrophic coastal waters of southeastern Arabian Sea. The current results represented a period when the mud banks were not activated but fertilization at certain compartments of the coastal zone by injection of nutrients by hitherto unknown processes was evident. The higher dissolved concentrations of ammonia, nitrate and silicate originating from shallow depths and extending to offshore have indicated a clear groundwater based nutrient source. The enriched particulate organic carbon and Chlorophyll a were also notable features of the nutrient injection region. It is difficult to point out a definite source for the high nutrient introduction as fresh water discharge was at the minimum during the nutrient injection duration. A band of N/P>15 funneling out during non-mud bank period gave a clear indication of an ‘external ground water source’ of nitrogenous compounds to the coastal water which deserve identification as it is traced to a region far away from any river mouth and the injection of nutrients was observed during non-monsoon months when mud banks were passive. The existence of subterraneous channels as the artifacts of porous nature of the lime shell base of the region transporting the nitrogenous compounds cannot be ruled out in the region.

1 INTRODUCTION The west coast of India is environmentally more sensitive than the east coast primarily because it is bordering one of the most sensitive ecosystems in the world, the Arabian Sea. The environmental property of the northern Arabian Sea is unique which manifests in rich biological production throughout the year through different processes and thus,

Water resources of arid areas

86

explain for the Arabian Sea ‘Paradox’ Mathupratap et al (1996). The symptoms are there to show considerable impact of deterioration of estuarine waters on the coastal ecosystem Nair et al (1991), Naqvi et al (2000), Jayakumar et al (2001). The emerging industrial establishments and human settlements along the west coast of India, thus necessitates a critical evaluation of the nature and quantum of inputs to the Arabian sea as well as their regional assimilative capacities. If there is a possible threat to the well being of the living resources of EEZ of India, then the coastal waters of southwest coast of India, and in particular, Cochin region is the prime location prone to trigger it. The booming city of Cochin has population of nearly 1.5 million Anonymous (1998) and 60% of the chemical industries of Kerala are situated in this area Cochin backwaters are the largest of its kind on the west coast of India with an area of 256Km2. The 16 major and several minor industries situated in the upstream region of the backwaters discharge nearly 0.105Mm3d−1 of effluents Anonymous (1996). The fertilizer consumption in Kuttanad region (the main agricultural field draining to Cochin backwater) alone is reported to be 20,239ty−1 Anonymous (1996). The backwater

Figure 1. Map showing the study area (A) and study region with location of stations and bathymetry (B). receives organic wastes ~260td−1 Anonymous (1998) and an annual dredge spoil from the harbor area to the tune of 107m3. Conventional understanding of coastal waters of southeastern Arabian Sea is that activation of mud banks by monsoon forcing triggers intense geochemical processes leading to high productivity. Mud banks, as they appear only during monsoon and disappear with its retrieval, are unique in their formation and functions, and have turned out to be economically important for its rich biological resources. As far as the chemical features are concerned, the general picture so far emerged out is that except during the

Subterraneous injection of nutrient rich groundwater to the coastal waters

87

monsoon periods, the southwest coastal waters remained oligotrophic and surface chlorophyll a typically ranges from 0.1 to 5.3mgm−3, while primary productivity ranges from 100 to 360mgCm−2d−1. Recent studies as the one discussed here contradict these findings and show that even after the monsoon period, fresh injection of nutrients by hitherto unknown processes fertilize the coastal waters that are either permanent or quasipermanent in nature. One of the major mudbank regions (Fig. 1 A, B) of southwest coast of India was selected for observation that indicates episodic introduction of nutrients into the coastal waters during periods when mud banks are passive. 2 RESULTS AND DISCUSSION During the typical pre-monsoon (February) months, the nitrogenous nutrients in water remained low except for the southern transects centered on Chethi and Alleppey. The phosphate concentrations did not show any spatial or vertical variation in the water column, but higher concentrations of ammonia, nitrate and silicate were observed at selected regions starting in the near shore regions and extending offshore (Fig. 2 A–D). The Nitrate-N concentrations point towards a clear source between Chethi and Pazhayangadi, where it peaked up to >8µM and decreased towards offshore. A similar trend was observed for ammonia-N with the source centered on Chethi (at about 15m depth). It may be assumed that the ammonia released were either rapidly utilized by phytoplankton or oxidized within the system itself where the waters were saturated with dissolved oxygen. Distribution of silicate-Si was similar to that of nitrate (4–10µM), higher than the corresponding values reported for the waters of Southeastern Arabian Sea. The input of these nutrients supported high primary production up to 14mg/m3 of chlorophyll a (peak column production of

Water resources of arid areas

88

Figure 2. Distribution of ammonia-N (A), nitrite-N (B), nitrate-N (C) phosphate-P (D) at the surface and bottom during October (a, b), February (c, d) & November (e, f). 1529mgCm2d−1), approximately 3 times greater than the peak values reported so far from these waters Qasim et al (1978). The peaks in chlorophyll a and ammonia showed a preference of ammonia among the nutrients for primary production. It is difficult to point

Subterraneous injection of nutrient rich groundwater to the coastal waters

89

out a definite source to these high nutrients during this period, as the fresh water discharge was at the minimum.

Figure 3. N/P peak values funneling out from mud bank region.

Water resources of arid areas

90

During post monsoon (November), homogenous mixed layer prevailed in the entire region. While the physical characteristics were more or less stable, there was considerable variability in the nutrients and in chlorophyll a concentration (Fig.2 A–D). A marked decrease in sub-surface dissolved oxygen (2.8–4.8ml/l) was the characteristic feature of this period, which was concomitant with enriched nitrite (0.5–2.0µM), phosphate (0.4–2.8µM) and silicate (0.5–14µM). The ammonia (1–7µM) and nitrate (1– 6µM) were also elevated at some regions along southern transects. The enriched particulate organic carbon (>3.5mg/l) and Chlorophyll a (14.8mg/m3) were also the notable features of this period. It is likely that chlorophyll a values were proportionate to carbon production indicating a strong positive relationship binding it with nutrient related factors rather than seasonal or diurnal fluctuation. The elevated nitrite and phosphate levels around Cochin may be due to the input from the backwaters. Higher values of nitrite, POC and chlorophyll a towards the southern offshore waters off Pallana were conspicuous and the regions with high nitrite had nitrate levels up to 6µM and the low levels of ammonia had ruled out the nitrification as a significant process responsible for nitrite accumulation. The remarkable co-existence of nitrite with nitrate strongly suggested that the nitrite production should mostly be due to assimilatory reduction. This was further substantiated by the high concentration of chlorophyll a (4–9.8mg/m3) on these transects. The N/P ratio in the coastal waters was below 15 during November (Fig. 3), possibly due to the disproportionate release of P from mudbank sediment. However, a band of N/P>15 funneling out from Alleppey region was indicative of an ‘external source’ of nitrogenous compounds into the coastal waters. A comparison of long-term (decadal) trend in the chlorophyll data of this region showed “greening” of near shore waters Devassy (1983). This suggests that phytoplankton standing crops had increased historically, possibly in response to watershed nutrient inputs. These sources of nutrients deserve identification as it was traced to a region, far away from any river mouths. The current observations in general indicated the presence of a nutrient source between Chethi and Pallana. This region has mud banks but the release of nitrogenous compounds cannot be accounted from sediments. The injection of nutrients was in nonmonsoon months when mud banks were passive and a new influence of Vembanad Lake on the coastal waters is very clear. One of the recent estimate shows that in spite of receiving 42.4×103mold−1 of inorganic phosphate and 37.6×103mold−1of inorganic nitrate from Periyar side of the estuary, the export to the coastal waters is only 28.2×103mold−1of inorganic phosphate and 24×103mold−1 of inorganic nitrate Naik (2000) and the lake acts as a sink for the nutrients, flushing out only a portion of the pollution load that it receives. Increased human population along the coastal belt has also resulted in concomitant increases in widespread use of septic tanks and nutrient inputs to coastal waters, particularly from regions occupying limestone beds. It has been found that domestic wastewater from septic tanks provide more nitrogen than that due to precipitation or use of fertilizers. The situation is exacerbated in the present study region, as more than 70% of households in these coastal belt and adjacent areas of Vembanad Lake do not have proper sanitation facilities. Significant amounts of nutrients from fertilizer applied in agricultural fields (approx. 94kg/ha) leach out into waterways, groundwater and to the coastal bays inducing coastal fertilization due to direct discharge into coastal ocean and through ground water seepage.

Subterraneous injection of nutrient rich groundwater to the coastal waters

91

3 CONCLUSIONS The nutrient fluxes into coastal region were influenced by fluxes from Cochin backwater and by the mud bank formation. The present study isolates a possible link between Vembanad Lake that supplies primary nutrients to the adjacent coastal waters and precondition it for rich primary production during non-monsoon months. The causative factors discussed are indicative of existence of a subterranean flow connecting Vembanad Lake to the adjacent coastal waters through the submerged porous lime shell beds. Continuous nutrient entry through such process is bound to upset coastal water productivity pattern. If the existence of the subterraneous channels linking Vembanad Lake to the adjacent coast is proved, it might even re-construct the historical evidence that the subterraneous flow plays a decisive role in the formation of mud banks along this region. A sub aqueous injection of nutrients into the coastal waters through this region is possible even after the rainy season. This assumption need further study to establish cause and affect mechanisms and quantify actual trends created by increased nutrient loading. REFERENCES Anonymous, 1996. Pollution potential of industries in coastal areas of India. Coastal Pollution Control Series: central Pollution Control Board Report. COPOCS/9/1995–96. Anonymous. 1998. NEERI- carrying capacity based developmental planning of Greater Kochi Region. Phase I Report. Devassy, V.P. 1983. Mahasagar, Bull Bull Nat. Inst. Oceanogr.7:101–105. Hema Naik, 2000. Budgets for Periyar estuary, Kerala. Presented at Regional Training Workshop on Biogeochemical Budgeting and Socio-Economic modeling for Coastal Scientist. APN/SASCOM/LOICZ, 18–22 September, Colombo. Jayakumar D.A., Naqvi S.W.A., Narvekar P.V. & George M.D. 2001. Methane in coastal and offshore waters of the Arabian Sea. Mar. Chem. 74:1–13. Mathupratap N.M., Prasanakumar S., Bhattathri P.M.A, Dileepkumar M., Reghukumar S., Nair K.K.C. & Ramaiah N. 1996. Mechanism of the biological response to winter cooling in the north eastern Arabian Sea. Nature, 384:549–551. Nair C.K., Balchand A.N. & Nambisan N.P.K. 1991. Heavy metal speciation in sediments of Cochin estuary determined using chemical extraction techniques. Sci.Total Environ. 102:113– 128. Naqvi S.W.A., Jayakumar D.A., Narvekar P.V., Naik H., Sarma V.V.S., D’Souza W., Joseph S. & George M.D. 2000. Increased marine production of N2O due to intensifying anoxia on the Indian continental shelf. Nature, 408:346–349. Qasim, S.Z., Wafar, M.V.M., Sumithra Vijayaraghavan, Joseph P., Royan. & Krishna Kumari, L. 1978. Ind. J. Mar. Sci.,7:84–93.

A new method for the estimation of episodic recharge J.Bean, G.van Tonder & I.Dennis Institute for Groundwater Studies, University of the Free State, Bloemfontein, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: A new stable isotope-based technique, the Modified Amount Effect (MAE) Method, was developed during this study. This technique provides insight into episodic recharge processes by estimating the proportion of preferential pathway-to-matrix-derived flow entering an aquifer, and the amount of rainfall required to initiate recharge via the respective flow paths. Significantly, the proportion of bypass flow can be determined without undertaking expensive and time-consuming unsaturated zone studies, both factors often of primary concern when undertaking recharge investigations in developing countries.

1 INTRODUCTION There can be no doubt that the South African water industry has been profoundly transformed over the last 10 years, with millions of rands invested in water infrastructure aimed at ensuring that every South African has access to fresh drinking water. In drier, more isolated, inland areas of the country, this has often meant that available groundwater resources must be exploited. As such, government and non-government organisations have invested in research associated with developing new assessment techniques so that these resources can be managed sustainably. In common with all these strategies is the need for recharge processes to be understood, and if possible, quantified. An understanding of site recharge behaviour is far more important than many geohydrologists realise, and goes beyond estimating the average proportion of rainfall entering a given aquifer. For example, from a planning viewpoint, groundwater ingress into a mine is seldom problematic to mine management, providing it is constant; problems occur when unpredicted increases occur, such as those associated with the sudden entry of recharge water into surrounding aquifers. Thus, through understanding the episodic nature of recharge in semi-arid and arid areas, and therefore the thresholds that must be exceeded before recharge occurs, geohydrologists are better able to provide predictive advice for their clients. This paper discusses a new stable isotope-based

A new method for the estimation of episodic recharge

93

technique, the Modified Amount Effect (MAE) Method. This was developed during the study, which provides insight into episodic recharge processes. 2 THE MODIFIED AMOUNT EFFECT METHOD 2.1 General This technique provides insight into episodic recharge processes by estimating the proportion of preferential pathway-to-matrix-derived flow entering an aquifer, and the amount of rainfall required to initiate recharge via the respective flow paths. Significantly, the proportion of bypass flow can be determined without undertaking expensive and time-consuming unsaturated zone studies, both factors often of primary concern when undertaking recharge investigations in developing countries. 2.2 The Methodology Recharge water is progressively enriched in δ18O due to evaporation, which occurs with movement through the unsaturated zone. The slope of the evaporated water line for matrix water in the unsaturated zone (EWL-U) is generally around 5, but sometimes as low as 2. The variable d was corrected to represent evaporation within the unsaturated zone by constructing a line through the 2H and 18O average for those samples representing background recharge (i.e. δ18O<−4.3‰), resulting in an EWL-U of δ2 H=2.5δ18O+ −21.05. Laboratory studies undertaken by Allison et al. (1984) confirm that if site recharge is constant, evaporation-induced enrichment within the unsaturated zone should also be constant, resulting in a line parallel to the local meteoric water line (LMWL), herein referred to as the Matrix Water Line (MWL-U). Thus, if exchange processes between aquifer materials and groundwater are ignored, preferential recharge areas can be inferred in cases where the LMWL and MWL-U are not parallel (i.e. the LMWL has a steeper slope and greater d-excess than the MWL-U). A similar assumption can be made if the line-of-best-fit through site groundwater data (GWL) is not parallel to the LMWL, as the isotopic characteristics of water stored in the aquifer represent a long-term average of recharge processes. An approximate estimate of the contribution of preferred pathwaysderived recharge to aquifer storage can therefore be determined by constructing lines parallel to MWL-U, GWL, and EWL-U, through the average 2H and 18O composition of groundwater derived from direct recharge i.e.:

Where PPflow=the proportion of recharge derived from preferential flow, and dGWL, dEWL2 U, and dMWL-U represent the d excess in δ H (‰) for GWL, EWL-U, and MWL-U, respectively. It should be appreciated that the calculated value PPflow is sensitive to: ● Variations in the orientation of the LMWL and EWL-U.

Water resources of arid areas

94

● The recharge threshold. At low recharge thresholds (i.e. recharge occurs rapidly in most years), particularly in more temperate areas, evaporation effects may not be represented in the stable isotopic composition of groundwater data. In these areas, transpiration, and not evaporation, probably has a greater potential to reduce the recharge flux to site aquifers. ● The source of recharge water. The suggested method assumes that recharge water is derived solely from precipitation, with no contribution from an adjacent surface water body where pre-recharge evaporation has occurred. 2.3 Adapting the MAE Method 2.3.1 Background information Groundwater samples were taken from the vicinity of Liebenberg’s Pan near Petrusburg, Free State Province, South Africa. Pans occur throughout the Western Free State, generally following the strike of the Ecca Series, which forms part of the Carboniferous to early Jurassic aged Karoo Basin sediments. The Ecca Series at the pan is comprised of mudstone, sandstone, and shale interbeds. Cretaceous-aged dolerite dykes and sills have intruded these sediments, with high yielding aquifers (>10L/s) often occurring at the structure/sediment interface. These have been locally overlain by calcretes to a maximum depth of about 15m. Petrusburg has a semi-arid climate, with an evaporation excess of 1920mm (2380– 460mm MAP) annually. Given that the water table is generally less than a metre below the pan surface, groundwater here is exposed to continuous evaporation in most years, the exception being those years where sufficient rainfall occurs to flood the entire pan for a few months of the wet season. Thus, the slope of EWL-U in these areas should be parallel to GWL because, while preferential pathway-derived water may not be as evaporated as matrix-derived during recharge events, it will eventually be evaporated to the same degree after entering the aquifer. Liebenberg’s Pan-derived brine with chloride concentrations in excess of 100000mg/L is further concentrated in evaporation ponds that have been constructed on site as part of a commercial salt-extraction enterprise operated by a local farmer. The pan itself is the lowest topographical feature in the landscape. Land use varies with soil type, topography, and access to irrigation water, with grazing and dairy farming predominant to the north and west of the pan in the steeper dolerite hills that occur there, and irrigated cropland located to the south and east on deeper soiled, gently sloping ground. 2.3.2 Calculating recharge using MAE The orientation of the EWL determined from brine samples was determined to be δ2H= 3.65 δ18O+−4.71 (refer to Figure 1). Another characteristic of Petrusburg data of interest is that evaporation-induced enrichment has not been excessive as would be expected in this type of environment, with all groundwater samples having a δ18O concentration 3.3‰ or less. This suggests that brine has mixed with isotopically depleted water from another source, the most likely being groundwater from upslope areas, a finding

A new method for the estimation of episodic recharge

95

supported by the occurrence of freshwater springs at various locations around the perimeter of the pan, and the observed decrease in brine concentrations in production bores over time. PPflow estimated using the MAE Method is between 33 and 25% assuming an EWL-U slope of 3.65.δ18O and 2.5.δ18O, respectively (refer to Figure 2). However, given the potential for brine/ fresh groundwater mixing, the lower figure would be more acceptable in this instance. Further insight into site recharge processes can be obtained when variations observed in recharge threshold estimates obtained using the cumulative rainfall departure method (CRD) and MAE techniques are considered. On the basis of 98 years of rainfall data for Petrusburg, the average

Figure 1. Stable isotope characteristics of groundwater samples taken in the vicinity of Liebenberg’s Pan, Petrusburg. “GW”, “PW”, and “Ave” denote groundwater samples taken from boreholes surrounding the pan, in the pan, and the background isotopic average, respectively.

Water resources of arid areas

96

Figure 2. Line characteristics used to determine PPflow at Liebenberg’s Pan. monthly rainfall is 35.7mm; this value also representing the long-term recharge threshold for an aquifer in equilibrium if seasonal conditions are ignored. In theory therefore, there would be no change in water levels if 35.7mm of rain fell at the site every month. Under field conditions however, this does not occur; prolonged periods of below average rainfall are evident throughout the Petrusburg dataset. Thus, in order to restore equilibrium conditions such that the average recharge threshold again decreases to 35.7mm/month, a given catchment must receive above-average rainfall. This observation is significant because it indicates that, for a given aquifer in a semi-arid and arid area, multiple recharge thresholds will be represented in site water level data. Multiple recharge thresholds that are likely to be of importance include those necessary to induce recharge via: 1. Preferential pathways after a period of below-average rainfall; 2. The matrix after a period of below-average rainfall; 3. Preferential pathways once aquifer equilibrium has been restored; 4. The matrix once aquifer equilibrium has been restored. Each of these recharge thresholds can be approximated using available site stable isotope data by applying the mass balance equation: RTave.δ18ORT−ave={RTlow.(X.δ18ORT−low)}+{RThigh.(X−1)δ18ORT−high} Where, RT=Average recharge threshold expressed as an equivalent rainfall depth (mm); RTlow=Average recharge threshold to be exceeded if recharge via preferential pathways is to occur (mm); RThigh=Average recharge threshold to be exceeded if recharge via the matrix is to occur (mm) and δ18ORT-low=Average δ18O concentration of preferential pathway-derived recharge water (‰); δ18Orw-high=Average δ18O concentration of matrixderived recharge water (‰); and, X=Preferential pathway to matrix proportioning factor. The average thresholds to be exceeded before recharge occurs via the matrix, and preferential pathways. Note that these values represent long-term averages, and not the upper and lower limits of recharge thresholds. These limit thresholds can be calculated,

A new method for the estimation of episodic recharge

97

however, by considering CRD and long-term average values together. For example, the CRD Method indicates that, for an aquifer under equilibrium conditions, the recharge threshold is approximately 35mm/month. Since, on average, the recharge threshold cannot be lower than this amount, it must represent the average lower recharge threshold. Thus, the respective average lower recharge thresholds can be calculated once the isotopic composition of rainfall for an equivalent depth of 35mm has been estimated from amount effect data. Once the lower and average long-term thresholds for both preferential pathway (RTlowand RTave-pp) and matrix-medium recharge (RTlow-uzm and RTave-uzm), the upper recharge pp thresholds RThigh-pp and RThigh-uzm can also be calculated, i.e. RThigh=2.RTave−RTlow Only 25% of recharge at Petrusburg occurs via preferential pathways. On average, recharge occurs via these pathways in more than 50% of all rainfall events (RTave−pp=56.4%). Therefore in episodic recharge environments, resource managers must ensure that allocated water can be used for the entire period between major recharge events, which where recharge via the matrix predominates, can be significant. Indeed, in many instances it may be more realistic to base groundwater allocations on the proportion of bypass flow-derived recharge entering site aquifers initially, the allocations increasing once aquifer storage, recharge threshold, and recharge event return period characteristics are better understood. 3 CONCLUSIONS Four recharge thresholds can be identified using the MAE Method; the low and high recharge thresholds that must be exceeded before recharge occurs via preferential pathways or the matrix, respectively. These represent threshold limits, the low value only of importance following successive months of wet weather, the high value representing the rainfall that must be received to restore an aquifer system to equilibrium after prolonged dry spells. Once these thresholds are known, the recharge history of a site can be modelled using available rainfall data by adapting the CRD Method. An important finding of modelling undertaken during this investigation is that in those semi-arid to arid areas where most recharge water enters, the aquifer via the matrix, the period of time that elapses between successive rainfall events that exceed the matrix recharge threshold often extends to scores of years. This has significant resource management implications for much of the region, as it indicates that the current approach of basing allocations on average recharge estimates is only justified if sufficient groundwater is available for use over the entire period between recharge events. The MAE Method was found to be sensitive to the recharge history of the site, the returned recharge estimate significantly higher when calculated immediately after recharge via the matrix had occurred. This is not to say that these estimates were incorrect (indeed they were representative of site recharge processes at the time of sampling), but that rainfall in the preceding months should be considered prior to sampling. In general however, sampling should be undertaken near the end of the dry season, which in the summer-dominant rainfall areas of Southern Africa is between

Water resources of arid areas

98

September and November (allowing for a 30 to 60 days lag time between rainfall and subsequent recharge). REFERENCES Alison, G.B., Barnes, C.J., Hughes, M.W. & Leaney, F.W.J. 1984. Effect of climate and vegetation on oxygen-18 and deuterium profiles in soils. Isotope Hydrology 1983. IAEA Symposium 270, September 1983, Vienna.

Prioritisation of the impacts of pollutants on groundwater flow systems in South Africa I.Dennis, B.Usher & J.Pretorius Institute for Groundwater Studies, University of the Free State, Bloemfontein, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Groundwater pollution can occur, as a result of various activities of man. With increased human settlement and economic development, a range of undesirable waste products are produced which can end up in the environment. If these waste products are not well handled, they can cause pollution of groundwater. The threat caused by undesirable substances on groundwater is recognized in South Africa and measures have been put in place through legislation to protect groundwater from pollution. Although groundwater pollution incidences have been reported countrywide, we do not have an indication of the extent of the problem. The results of the investigation discussed in this paper are therefore geared towards filling the gap in the understanding of groundwater pollution in South Africa’s urban environments. By doing so the principal pollutants can be identified and based on their risk prioritised. This will facilitate better management of groundwater quality through the country.

1 INTRODUCTION Groundwater pollution can occur, as a result of various activities of man. With increased human settlement and economic development, a range of undesirable waste products are produced which can end up in the environment. According to the National Water Act (Act No. 36, 1998), pollution is defined as the direct or indirect alteration of the physical, chemical or biological properties of a water resource so as to make it— 1. Less fit for any beneficial purpose for which it may reasonably be expected to be used; or 2. Harmful or potentially harmful— ● to the welfare, health or safety of human beings;

Water resources of arid areas

100

● to any aquatic or non-aquatic organisms; ● to the resource quality; or ● to property. The main aim of the investigation discussed in this can therefore be summarized as the prioritization of the type of pollutants and their associated sources which present a threat to groundwater, the environment and health in South Africa’s urban catchments. 2 METHODOLOGY 2.1 Factors taken into account The methodology followed evaluated the sources and contaminants separately. The results of the evaluation were then combined to determine a final risk based prioritization. 2.1.1 Sources Sources, in this context, refer to the origin of the substances (inorganic species, organic compounds or microbial agents) that are causing, or may potentially cause, the pollution. The term

Table 1. Systems for classification of groundwater contamination sources. Classification Examples system based on Way of release Loading history Location Degree of localization Origin Likelihood of occurrence

Discharge sources, transport sources Spill or continuous Above ground surface, below surface Point (or line) and non-point sources Industrial sources, mining sources For example petrol service stations found more often than chemical manufacturing plants

is used very broadly over a range of scales and may describe physical entities (e.g. a pond, a tank, a pipeline); human activities (e.g. mining, irrigation, wastewater treatment); the site at which potential pollutants are stored, used or disposed (e.g. wastewater treatment works, cemeteries, fuel filling stations) or even large scale phenomena (e.g. atmospheric deposition).

Prioritisation of the impacts of pollutants on groundwater flow systems in South Africa

101

Source of pollutant plays a large role in whether the pollutant will reach the groundwater table and if it does the rate at which the pollution will enter the groundwater system. There are also several existing methods for classifying the sources of groundwater pollution. A simplified classification based on that of Nonner (2002) was used to classify South African sources (see Table 1). 2.1.2 Pollutants Pollution refers to levels of hazardous substances in the environment over and above what would ordinarily be found in the absence of local activities. Groundwater pollution therefore refers to the occurrence of substances (inorganic species, organic compounds or microbial agents) in concentrations above those that would naturally be found in an aquifer. The substances themselves, both chemical and microbial, are called pollutants. There are various ways in which to group or classify groundwater pollutants. Each of these has major classes which can then be broken down into smaller categories. The choice of system and level of detail of the classification is dictated by the purpose of the classification for the sake of this investigation pollutants were classified according to: ● Fate in the environment – Degradable pollutants, which can be rendered harmless by natural processes and need therefore cause no permanent harm if adequately dispersed or treated; and – Persistent pollutants, which eventually accumulate in the environment and may be concentrated in food chains. – Pollutants may also be divided by their behaviour in water into: (a) Soluble pollutants, which includes most inorganic species and some organics. (b) Insoluble substances, which are small enough to be carried through the aquifer matrix, including microbial pollutants and colloidal inorganic pollutants. (c) Non-aqueous phase liquids (NAPLs), which are organic compounds that do not dissolve readily in water and remain as a separate liquid phase. These are further subdivided into Light Non Aqueous Phase Liquids (LNAPLs) and Dense Non Aqueous Phase Liquids (DNAPLs). ● Human health impacts – Non-harmful substances, which have no observed effects on human health. – Toxic substances, which cause various effects on the body from short-term exposure or long term accumulation, ranging in severity depending on the dose e.g. nausea, rashes, kidney failure or neurotoxic effects. – Carcinogenic substances, which are known to cause cancer. – Pathogenic substances, which are known to cause diseases in humans. ● Other aspects that are taken into account: – Duration of pollution—if the pollution results from a single (once-off) spill, the impact will probably be smaller than that resulting from continuous pollution. – The vulnerability of the aquifer represents the intrinsic characteristics that determine the sensitivity of an aquifer to the adverse effects resulting from the imposed

Water resources of arid areas

102

pollutant (Lynch et al., 1994). Factors taken into account include depth to groundwater, recharge, aquifer media, soil media, topography and impact of the vadose zone. 2.2 The risk-based methodology Rating occurs when contaminant sources are given a quantitative or qualitative measure of the potential hazard they pose to groundwater. Prioritisation methods focus on aspects such as contaminant loading, mobility, persistence and hazardousness while risk assessment develops these further into potential human health impacts. A risk analysis estimates the probability and consequences of a contaminant event and usually considers both the properties of the contamination source and the hydrogeological environment. Conventional set theory (Boolean) states that an element is either a member of a set or not. Fuzzy logic is an extension of conventional set theory enabling an element to belong to a set to a degree. The degree of membership is a function that defines the membership of an element to a set according to the value of the element. Membership is expressed as a value between 0 and 1. Zero implies 0% membership and 1 implies 100% membership. Linear membership functions are seldom used in practice in contradiction to sinusoidal functions which are very popular. In most cases risk analysis will involve more than one input to be considered in the analysis. Fuzzy logic makes it possible to generate a set of decision rules according to the number of inputs and these rules must then be evaluated by an expert in the field of study. The number of rules generated is given by the following equation is: n=2inputs where n represents the number of rules generated. The rules consist of all possible binary combinations of the respective inputs with a weight assigned to each rule representing the risk. The risk is then calculated using the following formula:

where n=number of rules, DOM=degree of membership and Wn=weight of rule n. 2.3 A tiered approach Based on the amount of data available a tiered approach is followed when considering risk assessments. The first tier (LEVEL 0) is a rapid assessment of sources in which minimal data are required and it produces low confidence results. This assessment should be completed within a few minutes and is based on a rating system. LEVEL 1 is the second tier which is a rapid assessment of contaminants on a local scale. It is intended to

Prioritisation of the impacts of pollutants on groundwater flow systems in South Africa

103

give the assessor a guideline of the risks. The assessment should take a couple of hours to complete. The next tier (LEVEL 2) is an intermediate assessment. The first step in the intermediate assessment is to collect all relevant data. Data requirements include aquifer and contaminant parameters, as well as health information. General information will be obtained from databases, but it is sometimes necessary to have site-specific data. The confidence attached to this assessment should be medium to high. Both the second and third tiers include risk assessments based on a fuzzy logic methodology. Figure 1 is a schematic representation of the tiers and the function performed on each level of assessment.

Figure 1. Tiered approach to South African prioritization methodology. In order to protect boreholes wellhead protection areas (WHPAs) need to be delineated. A WHPA can be defined as the surface and subsurface area surrounding a

Water resources of arid areas

104

borehole or wellfield, through which contaminants are reasonably likely to move and reach such a borehole or well field. In many cases it is difficult to protect the whole area, therefore various zones are established within the area. 3 CASE STUDY: CAPE FLATS WASTEWATER TREATMENT WORKS 3.1 The wastewater treatment works within the City of Cape Town prioritization of sources and contaminants on a regional scale The City of Cape Town (CCT) is located in the Western Cape Province on the southeastern corner of South Africa. A major portion of the CCT consists of the area known as the Cape Flats, which has an elevation of between 20 and 45m above sea level. CCT has a mean annual rainfall of 515mm/annum and an average temperature of 16.7°C. It is a winter rainfall area. The current population of the CCT is estimated at 3.2 million with the highest population density occurring on

Table 2. Source prioritization for CCT (incomplete list). Source prioritisation (from highest to lowest risk) On-site sanitation Petrol service stations (underground storage tanks) Cemeteries Stormwater/sewer systems Agriculture (general and crop cultivation) Feedlot/poultry farms Wastewater treatment

Table 3. Contaminant prioritization for CCT. Contaminant prioritisation (from highest to lowest risk) Nitrate Chloride Phosphate Potassium Ammonia & sulphates

Prioritisation of the impacts of pollutants on groundwater flow systems in South Africa

105

Table 4. Information used in site-specific risk assessment. Parameter

Assigned value

Vulnerability –Recharge 65mm/yr –Soil media Sa-LmSa –Aquifer media Intergranular –Vadose zone Beach sand –Groundwater depth 8m –Topography 1% Duration Continuous Contaminant* Nitrate Level of management Low *Once the contaminant is entered the software automatically pulls in the health risk information and physio-chemical behaviour from a database.

the Cape Flats and there are approximately 90000 consumers on informal sites. There are 21 wastewater treatment plants within the CCT. According to TIER 0 the wastewater treatment works are rated as the 7th highest pollutant source within the CCT. Due to the length of the complete list only the 7 highest potential polluters have been documented in Table 2. Typical contaminants found at wastewater treatment works include ammonium, nitrate, potassium, phosphate, chloride, sulphate and faecal pathogens. Micro-organisms were not included in the investigations and will therefore not be included in the prioritization list. The prioritization of the above-mentioned chemicals is listed in Table 3. 3.2 The Cape Flats wastewater treatment works risk assessment The wastewater treatment works has unlined sewage sludge drying ponds. The wastewater treatment works are situated on an unconfined primary sand aquifer. The information used to determine a site-specific risk is listed in Table 4. For the sake of demonstration only the risks for nitrates will be determined in this paper. The results of the assessment are summarized in Table 5.

Table 5. Results of risk assessment. Assessment

Risk (%)*

Source 58 Vulnerability 52 Health 99 Physio-chemical 75 Total 68 * Higher the risk higher the negative impacts.

Water resources of arid areas

106

Table 6. Data used to calculate protection zones. Parameter

Assigned value

Abstraction rate Transmissivity Effective porosity Hydraulic gradient Saturated thickness

10l/s 100m2/d 0.1 0.01 20m

Table 7. Calculated protection zone. Definition

Radius (m)

Zone 1: Highly protected area around the borehole. Its purpose is to protect the borehole from the direct introduction of pollutants into the borehole and its immediate area from spills, surface runoff, or leakage from storage facilities or containers. Potential pollutant sources in Zone 1 should be strictly monitored. 25 Zone 2: Is established to protect a borehole from contact with pathogenic micro-organisms which can emanate from a source located close to the borehole, as well as to provide emergency response time to begin active cleanup and/or implementation of contingency plans should a chemical contaminant be introduced into the aquifer near the borehole. 470 Zone 3: Is designed to protect the borehole from chemical contaminants that may migrate to the borehole; it typically includes a major portion of the 750 recharge area or the capture zone.

The results of the risk assessment for nitrates indicate there is a 68% chance that there are going to be negative impacts on the environment (including human health) as a result of groundwater becoming polluted with nitrates as a result of the wastewater treatment works. 3.3 Protection of boreholes The distance between a pollution source and a protected borehole can be calculated to ensure the borehole is not polluted. The zone of protection can then be delineated around the borehole. These wellhead protection zones can also be used to plan new boreholes. If all pollution sources are known then the ‘safe’ distance from a source can be calculated.

Prioritisation of the impacts of pollutants on groundwater flow systems in South Africa

107

Wellhead protection zones were calculated for boreholes in the Cape Flats. Table 6 contains the information needed for the calculations and Table 7 defines the protection zones and gives the radius of protection zones. 4 CONCLUSIONS AND RECOMMENDATIONS Pollution of South Africa’s urban aquifers presents a threat to the sustainability of this water resource. Man’s activities, use of chemicals and generation of wastes tend to concentrate potential sources of pollution in the urban areas. The threat caused by undesirable substances is recognized in this country, but the understanding of the extent of the problem in South Africa’s urban catchments is poor. This paper therefore briefly outlined a risk-based methodology to prioritise and determine the impacts of pollutant sources and pollutants. The methodology takes the following into account: ● Characteristics of pollutant sources ● Characteristics of pollutants ● Human health impacts of pollutants ● Vulnerability of South African aquifers ● Duration of pollution The methodology was then applied to determine: ● National list of priority chemicals and sources ● Regional list of priority chemicals and sources for the large South African urban areas ● Local risk assessments to determine the risks of certain pollutants ● Delineation of protection zones The results are intended to help groundwater practitioners and water authorities in assessing the likely transport, fate an impact of pollutants in the subsurface in an urban environment. It is recommended that the following aspects receive more attention in future research investigations: ● Based on the paucity of groundwater-related microbial data encountered in this project, the inclusion of these aspects in urban groundwater management must be regarded as a priority. ● Petroleum products, industrial thinners and mineral oils and other non-aqueous phase liquids represent a category of potential pollutants that have been largely overlooked by regulatory agencies and legislature, despite their harmful effects at small concentrations. ● A general lack of data on groundwater pollution from pesticides is evident. This is due to: (i) surface waters are the main source of water supply in the country; (ii) cost and difficulty to measure organic contaminants; (iii) private companies are often sensitive to make public data related to pollution problems. Therefore there is a need to investigate pesticides in groundwater.

Water resources of arid areas

108

REFERENCES Lynch, S.D., Reynders, A.G. & Schulze, R.E. 1994. Preparing input data for a national-scale groundwater vulnerability map of Southern Africa. Water SA, 20(3):239–246. National Environmental Management Act. Act 107 of 1998, Pretoria, South Africa. National Water Act. Act 36 of 1998, Pretoria, South Africa. Nonner, J.C. 2002. Chapter 3: Sources of groundwater contamination. In: A. Zaporozec (ed.) Groundwater contamination inventory: A Methodological Guide.UNESCO, IHP-VI, Series on Groundwater No. 2. 23–38.

Understanding problems of low recharge and low yield in boreholes: an example from Ghana A.J.E.Cobbing & J.Davies British Geological Survey, Wallingford, Oxon, UK Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The Afram Plains region of Ghana experiences acute seasonal water shortages during the four to five month long dry season. The long-term development of the limited groundwater resources of the region has proved to be difficult as the hydrogeology is poorly understood. Failure of boreholes is common, and there is little or no monitoring of groundwater levels. A two-year study led by the British Geological Survey, including the monitoring of borehole-drilling programmes, has led to a better understanding of the groundwater resources in the area and has provided guidelines for data collection.

1 INTRODUCTION The Afram Plains area is located in the Eastern Region of Ghana, in the Volta River basin between latitude 6°30′ and 7°30′N and longitude 1°00′W and 0°15′E (Figure 1). The area is about 4285km2 in extent, and lies between lake water level at 76m and 300m above mean sea level. The topography is subdued, with the main feature being a low northeast to southwest trending ridge 200–300m high. The Afram Plains supports savannah vegetation that is being progressively cleared for agricultural use. Coarse tussock-grass with a few stunted trees covers the low-lying lakeside plain and dense bush with large trees covers the better-drained ridge area. Since initial settlement in 1930, the rural population of the Afram Plains has increased rapidly following the construction of the Akosombo Dam in the 1960s. Between 1970 and 1984 census data show a 250% increase in the farming population, attracted by fertile soils and improving infrastructure. There are now more than 140 villages on the Afram Plains. Access to the area is poor, with the principal route by ferry across Lake Volta. The main town is Donkorkrom, which has a hospital, secondary school and post office.

Water resources of arid areas

110

Figure 1. Map of Ghana showing the Afram Plains study area. 2 THE WATER SUPPLY PROBLEM IN THE AFRAM PLAINS Before construction of the Akosombo Dam, village water supplies were obtained from the perennial Afram and Volta Rivers, seasonal flows and pools along ephemeral tributary streams and shallow water-filled dugouts. Rainfall on the Afram Plains is seasonal, with an average of about 1200mm/year falling almost entirely between April and October. Surface drainage is mainly ephemeral, storm water draining by sheet flow as short-lived floods. The seasonal rainfall and limited surface water storage result in acute water shortages during the November to March dry season. Reliance on unprotected pools and dugouts for water supply results in water-washed and diarrhoeal disease, and much time and effort in water collection. Guinea worm infections occasionally occur in the Afram Plains. 2.1 Regional geology The Afram Plains are located at the southern end of the large (>100,000km2) Voltaian Sedimentary Basin formed during the Precambrian to early Palaeozoic Pan-African Orogeny of 730–550Ma. (Black and Liegeois, 1993, and Shackleton, 1976). The Voltaian Basin is interpreted as a foreland basin; with sediments of marine and terrestrial origin filling a flexural depression at the margin of the West African Craton (Ako and Wellman, 1985). Kesse (1988) and Anani (1999) describe the Voltaian basin sediments as fairly flat bedded sandstones, shales, pebble beds, mudstones, limestones and siltstones deposited unconformably upon older Precambrian rocks. The molasse type sediment pile, that is estimated to be more than 4km thick, resulted from erosion of mountain chain fold belts that occurred along the present Ghana-Togo border to the east. The Voltaian Formation

Understanding problems of low recharge and low yield in boreholes

111

Obosum Beds that underlie much of the southern Afram Plains have yet to be studied in detail. Present geological knowledge has been derived from rapid reconnaissance surveys, several deep exploration boreholes and a number of shallow groundwater boreholes. 2.2 Previous groundwater development Development of the groundwater resources of the Voltaian sediments of the area began in 1963–65 when the Geological Survey of Ghana and the Volta River Authority (VRA) drilled a series of test and production boreholes in response to populations displaced by the rising lake waters. During the late 1960s and early 1970s the Catholic Church funded the construction of 28 hand-dug wells to supply small villages. These were mainly located in valley sites to replace unprotected shallow dugout sources. Additional boreholes were installed by the VRA at Donkorkrom and Kaklakoklope in 1983/84. UNICEF provided a borehole for the secondary school at Donkorkrom in 1983. During 1984, Prakla Seismos drilled 47 village boreholes for the German NGO Misereor. Of these, 19 boreholes were dry and 17 had yields greater than 301min−1. Although Lake Volta forms the eastern boundary of the area, the underlying low permeability rocks are the main source of water especially in the more remote western area. During the 1990s, more than 300 boreholes were drilled to meet the water supply needs of the expanding population.

Table 1. Summary of borehole drilling on the Afram Plains, 1963 to 2001. Organisation

Period No. Wet Dry of Bhs Bhs Bhs

Volta River Authority Prakla Seismos for Misereor World Vision International World Vision International WaterAid/Afram Plains Dev. Org. DANIDA exploration boreholes Totals

1963– 1965 1984

10

6

4

47

28

19

1990– 1995 1999– 2000 1996– 2001 2001

152

92

60

66

?

?

101

67

34

5

5

0

381

198 117

Many of the boreholes drilled were dry whilst other nominally successful boreholes showed a progressive decline in yield to fail after two to three years of use, especially in the west of the area. Due to their short period of land tenure, communities have yet to develop coping strategies to manage the limited water available during dry periods.

Water resources of arid areas

112

Populations attracted to the area by the groundwater supply have no effective alternative water source if borehole yields fail after several years of use. 3 RECENT WORK ON THE AFRAM PLAINS The British Geological Survey (BGS) first worked on the hydrogeology of the Afram Plains in 1985–86, in a project examining shallow wells and boreholes in the Donkorkrom area (Buckley, 1986). Beginning with a visit to the area in February 2000, the BGS collaborated with the Afram Plains Development Organisation (APDO), WaterAid, DANIDA, Legon University (Accra) and other partners in a two-year project specifically aimed at investigating the hydrogeological problems of the area. The BGS work was funded by the British Department for International Development (DfID). The project was timed to coincide with the drilling of 36 village water supply boreholes on the Afram Plains, funded by WaterAid working with the APDO. In addition, DANIDA funded the drilling of a further 5 deep (>100m) exploration boreholes, the first four of which were sited and geologically logged by BGS hydrogeologists. Studies carried out by the BGS in collaboration with local partners included: ● A reconnaissance geological and hydrogeological survey, and the creation of a GIS base map of the area. ● The geophysical survey of four of the five deep exploration borehole sites using frequency domain electromagnetic induction (EM34). ● The geological logging of rock chip samples produced during drilling, and the recording of penetration rates and drill stem yields. ● The test pumping of boreholes, including the demonstration of bailer tests and low yield “whale” pumps. ● The sampling of groundwaters for hydrochemical analysis of major and minor ions, and isotopes. ● The geophysical logging of the deep exploration boreholes was carried out by DANIDA.

4 SURFACE GEOPHYSICS ON THE AFRAM PLAINS Electrical resistivity and EM geophysical exploration surveys have been undertaken in the Afram Plains during other development projects. As in other hydrogeologically “difficult” areas in Africa, these methods have fallen out of favour in the Afram Plains, being seen as relatively expensive and time consuming for little benefit. This is due to a combination of: ● the mode of groundwater occurrence in the area, such as deep fractures, often thick weathered zones, that cannot be defined using geophysical surveys, ● the lack of experienced personnel capable of interpreting geophysical results for sedimentary environments.

Understanding problems of low recharge and low yield in boreholes

113

BGS undertook 11km of EM34 surveys at the first four exploration borehole sites, using 10m, 20m and 40m inter-coil separations. Readings were made in both vertical and horizontal orientations. The survey results were correlated with the geological logs from the exploration boreholes. Geophysicists from the University of Ghana, Legon, undertook max-min EM and electrical resistivity geophysical traverses along the main road in the eastern Afram Plains (Banoeng-Yakubo and Armah, 2001). The result of these studies demonstrate that geophysical surveys can be used in the eastern Afram Plains to differentiate between near surface shale, siltstone, sandstone and conglomerate bands, as well as delineate possible fault zones. In the west of the area, re-cemented sandstones up to 60m thick form a low permeability homogeneous layer below an ancient weathered surface. Thin water bearing fracture or weathered zones beneath this layer cannot be detected using EM34 or VES equipment. 5 DATA GATHERED DURING BOREHOLE DRILLING Useful geological and hydrogeological data that can be gathered during the drilling of a borehole includes: ● Geological data ● Penetration rate data ● Flow data ● Hydrogeological data Rock chip samples produced during drilling were collected at 1m intervals. Weathered zones (colour changes) and fracture zones (calcite and quartz mineralisation) enabled identification of water bearing zones. The chip samples were placed in a marked half pipe and photographed to produce pseudo-core logs. This procedure allowed zones of water inflow to be correlated with changes in lithology, and deductions regarding the nature of groundwater occurrence to be made. The rate of drill penetration and flow rate, determined at water strike zones and at the end of each drilling rod can be correlated with changes in lithology and weathered zones. Photo logs can show the nature of the weathered zones. The results obtained from exploration borehole showed that the rock types present are generally tight and fine-grained, with water being produced from horizontal weathered zones and along lithological boundaries rather than near-vertical fractures. In the western half of the area, the presence of a thick duricrust weathered zone, stopping recharge to underlying aquifer systems, was recognised. 6 BOREHOLE GEOPHYSICS A suite of geophysical logs was obtained from six boreholes in the study area, i.e. the five deep exploration boreholes together with a water supply borehole located at the APDO office in Tease. The calliper logs show the fracture zones, which can be correlated with the drillers report and the chip sample logs. The fracture zones are also indicated by lower resistivity measurements.

Water resources of arid areas

114

Pumped fluid logging of the boreholes clearly shows that fluid inflows occur at discrete fractured or weathered horizons, and that most of the water obtained from the boreholes is derived from these features. The discontinuous nature of the fracture systems that supply water is illustrated by drilling at the APDO office in Tease: in 2000, a 70m deep borehole (“hole no. 28”) was drilled in an attempt to provide a water supply for the office. This borehole proved to be dry and was backfilled. In 2001 a further two boreholes were drilled within 20m of this hole, one to 54m and a deep exploration borehole to 152.8m. Both of these boreholes yielded water. 7 TEST PUMPING OF BOREHOLES Pumping test data in fractured aquifers is more difficult to interpret compared with intergranular systems. There is often a distinct change between early and late time drawdown rates, due to the effect of fracture dewatering. This can allow erroneous interpretations to be made, particularly if pumping tests are carried out over only short periods of time. Pumping test interpretation requires specific training, and pumping tests have sometimes been done on the Afram Plains merely as required by the contract, without the pumping test information being used to inform the borehole completion. BGS developed a simple bail test, which allows field personnel on the Afram Plains to decide in a general way whether or not to equip a borehole, without going through the lengthier and more complex process of a pumping test (Davies and Cobbing, 2002). There are cases however where the bail test is inconclusive and the borehole requires a pumping test. Bail tests are

Understanding problems of low recharge and low yield in boreholes

115

Figure 2. Geophysical logs of deep exploration boreholes at Gazeri Camp (left) and Samanhyia, near Tease. Fractured and weathered zones can be seen on the calliper and induction resistivity logs, and the pumped flowmeter logs show that most flow into the boreholes is derived from these horizons. recommended as a rapid and simple field procedure to be used by staff not trained in pumping test interpretation to decide whether or not to equip a borehole with a pump. Simple pumping tests give indications of the productiveness of the systems but the results obtained are from “fractured” aquifer systems with high secondary permeability zones are difficult to reconcile. Such systems can initially give high yields but when they are dewatered during extended periods of over-pumping these systems can suddenly fail. 8 HYDROCHEMISTRY RESULTS Water samples for hydrochemical analysis were obtained from 29 boreholes and wells during the 2001 visit. Samples were taken from sources after several minutes of pumping where possible. Measurements of pH, specific electrical conductance (SEC), temperature and bicarbonate were taken at each site. Filtered acidified and non-acidified samples were obtained from each source for laboratory analysis. A GPS was used to locate the areal co-

Water resources of arid areas

116

ordinates of each sample site. Stable isotope analysis (δ2H and δ18O) was carried out on twelve samples by mass spectrometry. The results of these determinations plot close to the world meteoric water line. There is some evidence for the possible mixing of lakederived waters with aquifer waters in some areas. The major and minor ion analyses show that most determinants are within World Health Organisation (WHO) Guide Values, with the exceptions of boron and sodium that are a problem in the unfractured shale and sandstone area. Nitrate and ammonium levels in a few boreholes were evidence for anthropogenic pollution, which can occur because water is able to move relatively rapidly through fractures. The fluoride concentration in one sample exceeded WHO Guide Values. 9 HYDROGEOLOGY OF THE AFRAM PLAINS A five-fold hydrogeological division of the rocks of the Afram Plains can be produced, based on the conclusions of Bannerman (1990) and Acheampong (1996), and taking the current study into

Figure 3. Five hydrogeological divisions on the Afram Plains. account (Figure 3). The hydrogeology of each of these units is summarised in Table 2. 1. Massive conglomerate and sandstone. 2. Fractured shale and grey sandstone. 3. Quartzitic sandstone and conglomerate.

Understanding problems of low recharge and low yield in boreholes

117

4. Feldspathic sandstone, arkose, siltstone and mudstone. 5. Unfractured shale and sandstone.

10 DISCUSSION In regions of seasonal or low rainfall with ephemeral drainage patterns rural settlements may be totally dependent upon groundwater supply during the dry part of the year. Such is the present and future shortage of land in many areas that communities once settled in water poor areas are difficult to move. Therefore understanding of groundwater resources is a vital factor for long-term development plans of such marginal areas. Groundwater development in the Afram Plains has followed a pattern that is typical of areas underlain by low permeability rocks in sub-Saharan Africa. Reconnaissance level geological and hydrogeological surveys were first undertaken with limited drilling more than thirty years ago. Some borehole drilling by the VRA was undertaken at the time of population resettlement following the building of the Akosombo Dam and consequent flooding in the 1960s, but these boreholes have fallen into disuse following lack of maintenance. NGO-led water supply programmes, undertaken by World Vision International, a Catholic Church Group and WaterAid, funded the drilling of some 370 boreholes on the Afram Plains during 1984–2001. During these programmes the economic design and construction of boreholes, and borehole drilling “success rates” were emphasised. A borehole was judged a success if “wet” at the completion of drilling. The hand pump equipped boreholes were expected to supply 250 people with at least 20 litres per capita of water per day. In the Afram Plains the acceptable yield minimum is about 121min−1, due to the low borehole yields obtained. The high borehole “failure rate” (40%) has led to further study of the distribution of fracture and near surface weathered zones, these being perceived as the best groundwater bearing targets. Although many boreholes have been drilled, the geology of the area, groundwater occurrence, and the nature of the water resource remain poorly understood. This problem is exacerbated by the failure of apparently successful boreholes after 3–4 years of use.

Table 2. Summary of the hydrogeology of the five hydrogeological units. Description Ground of rock/ water hydrogeology targets unit Obosum Massive Beds— conglomerate Upper and sandstone Voltaian System

Ground Ground Field Technology Comments water water techniques potential quality

Weathered ** zones and fracture zones. Success rate ~66% wet 38%≥30l/min

Good. Presence of NO3N and NH4 indicates pollution in heavily used

Weathered Boreholes conglomerate 60–100m gravel often visible at surface: EM34—used to locate fractures and sandstones/ conglomerate

Good recharge, best sites located in valleys. Boreholes should be drilled to below present day

Water resources of arid areas

118

boreholes near surface. in village VES— centre. indicates depth of weathering

Quartzitic Weathered ** sandstone and zones and conglomerate fracture zones. Success rate ~67% wet 40%≥30l/min

Good.

Quartzitic Boreholes sands often 100–150m visible at surface. EM34—used to locate fractures and sandstones/ conglomerate near surface. VES— indicates depth of weathering

Feldspathic sandstone, arkose, siltstone and mudstone

Good.

Weathered Boreholes purple brown 100–150m sandstone platform surface beneath thin ferrecrete. Difficult to identify fractures with EM34, sandstones have been recemented to 60 m. VES—may indicate

Weathered */** zones and fracture zones. Success rate ~66% wet 39%≥30l/min

lake level. May be able to induce flow from the lake along fracture zones, Problems with pollution in villages. Moderate recharge, best sites located in valleys. Boreholes should be drilled to below present day lake level. May be able to induce flow from the lake along fracture zones. Problems with pollution in villages. Very poor recharge potential due to recemented layer down to ~60m. Deep holes may intercept weathered zones, Remoteness precludes direct recharge from lake

Understanding problems of low recharge and low yield in boreholes

119

depth of weathering

along fractures, Fractures poorly defined.

Description Groundwater Groundwater Groundwater Field of rock/ targets potential quality techniques hydrogeology unit

Technology Comments

Unfractured Weathered * shale and grey zones and sandstone fracture zones. Success rate ~50% wet 14%≥30l/min

Poor to saline. Low lying Boreholes— low altitude 50–100m lake side areas. EM34— moderate to high conductivities, used to locate fracture zones VES— indicates depth of weathering

Fractured shale Weathered *? and sandstone zones and fracture zones. Success rate Unknown due to lack of data

Poor to saline?

Poor to moderate recharge to tight formation except where conglomeratic bands area present. Boreholes should be drilled to below present day lake level. Boreholes— Unknown 50–100m

Low lying low altitude lakeside areas. EM34— moderate to high conductivities, used to locate fracture zones VES— indicates depth of weathering

KEY: Groundwater potential: *Low; **Moderate; ***High. Note: Groundwater Potential is an overall function of groundwater storage, groundwater yield and groundwater residence time (length of time groundwater remains in the unit, i.e. rate of groundwater throughflow). It indicates both the available yields and the length of time these are available for: i.e. high, moderate or low yields, available only during the wet season and immediately afterwards, or yearround. See below for more detail. EM34 conductivity response: High>50mmhos/m; Moderate 20– 50mmhos/m; Low <20mmhos/m. Yield: High >1l/s; Moderate ~0.5l/s; Low <0.2l/s. Note: Where groundwater residence times are long, groundwater availability is likely to be less vulnerable to variations in seasonal rainfall—e.g. one year of drought. Where few data are available locally, the interpretations given here are preliminary, and should be updated as new data are provided by continuing groundwater development work.

Water resources of arid areas

120

11 CURRENT CONCEPTUAL MODEL OF THE AQUIFER: A SUMMARY The main features of the aquifer model for the Afram Plains as a whole are as follows: ● Groundwater is thought to occur in discrete fracture systems or zones of weathering. ● The geological units have different hydrogeological characteristics but all are relatively low yielding. ● In the west of the Afram Plains in particular, the aquifer units may not be adequately recharged during successive wet seasons, leading to the progressive mining of groundwaters that leads to the failure of boreholes with time. Old water is often present in the fracture systems. ● In areas where recharge of surface water occurs, rapid movement through near surface weathered zones and fracture systems can lead to rapid transport of contaminants into boreholes below sanitary seal zones, as indicated by high ammonium and nitrate levels discerned in the central village borehole water sources on the Afram Plains. ● Water bearing weathered zones may be too deep and discrete to be determined using geophysical survey methods. ● Drilling deep boreholes to below the present day lake level may allow interception of fracture and weathered systems that can potentially be recharged by lake water. This process of recharge from the lake remains to be proven. ● The collection of accurate geological and hydrogeological data is vital for better understanding of the aquifer systems present. The use of currently available data is hindered by a lack of accurate site locations. The interaction of geological factors such as lithology, diagenesis, recent weathering, ancient weathering, tectonism with ancient and modern water level changes needs to be understood.

12 CONCLUSIONS The water supply problems on the Afram Plains cannot be solved by borehole drilling and groundwater development alone. The failure of boreholes after two or three years of use is particularly serious since in that time communities come to rely on the groundwater resource. Conjunctive use with rooftop rainwater catchment systems and small dams may need to be considered as well as artificial recharge to aquifers. There is a need to understand recharge mechanisms before borehole drilling commences. This project has demonstrated the types of data that can be easily collected at little additional cost during borehole drilling, and the uses to which such data can be put to the benefit of subsequent water supply programmes. A regional summary of groundwater occurrence in this “difficult” hydrogeological area has been built up, and presented in a format that can be used in subsequent groundwater development. The general shift from centralised groundwater development towards demand-driven, private organisation or NGO led work in Africa has had some benefits in terms of sustainability, community involvement and ownership issues, and the targeting of resources at the poorest communities. However, the negative effect has been the non-collection, storage and sharing of basic groundwater data, which leads to a lack of understanding in those areas where the groundwater resources are limited or difficult to access. At present, data collection is frequently seen

Understanding problems of low recharge and low yield in boreholes

121

as an unaffordable “optional extra”, adding mainly difficulty and expense to a project Such data as are collected often become difficult to access, since no effective central repository for data is currently in operation. The project aimed to overcome this by depositing the data collected in easily retrievable Word and Excel based packages with the Afram Plains Development Organisation staff. REFERENCES Acheampong, S.Y.1996. Geochemical evolution of the shallow groundwater system in the Southern Voltaian Sedimentary Basin of Ghan. A dissertation submitted in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Hydrology/Hydrogeology, Univ. of Nevada, Reno. Acheampong, S.Y. & Hess, J.W. 1998. “Hydrogeologic and hydrochemical framework of the shallow groundwater system in the southern Voltaian Sedimentary Basin”. Hydrogeology Journal, 6:527–537. Ako, J.A. & Wellman, P. 1985. The margin of the West African craton: the Voltaian Basin. Journal of the Geological Society of London, 142:625–632. Anani, C. 1999. Sandstone petrology and provenance of the Neoproterozoic Voltaian Group in the southeastern Voltaian Basin, Ghana. Sedimentary Geology, 128:83–98. Bannerman, R.R. 1990. Afram Plains borehole drilling programme, hydrogeological survey for WaterAid Ghana. Banoneng-Yakubo, B. & Armah, T. 2001. Hydrogeological and geophysical test investigations in the Afram Plains, Ghana. Department of Geology, Univ. of Ghana, Legon, for DANIDACWSA Project, Eastern Region, Ghana. Black, R. & Liegeois, J.-P. 1993. “Cratons, mobile belts, alkaline rocks and continental lithospheric mantle: the Pan-African testimony”. Journal of the Geological Society, London, 150:89–98. Buckley, D.K. 1986. Report on advisory visit to WaterAid projects in Ghana. British Geological Survey Technical Report. Davies, J. & Cobbing J. 2002. An assessment of the hydrogeology of the Afram Plains, Eastern Region, Ghana. British Geological Survey. Technical Report CR/02/137N. Grant, N.K. 1967. Complete Late Precambrian to Early Palaeozoic orogenic cycle in Ghana, Togo and Dahomey. Nature, 215:609–610. Kesse, A.O. 1988. The Mineral and Rock Resources of Ghana. A A Balkhema. Shackleton, R.M. 1976. Pan-African structures. Philosophic Transactions of the Royal Society, London. 280: 491–497. World Vision 1995. The Conrad N Hilton Foundation Funded World Vision Ghana Rural Water Project, Hydrogeological Report, Second Phase.

Spatial variation of groundwater recharge in a semi-arid environment—Serowe, Botswana L.M.Magombedze & B.Frengstad Geological Survey of Norway, Trondheim, Norway M.W.Lubczynski International Institute of Geoinformation Science and Earth Observation, Enschede, Netherlands Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The estimation of groundwater recharge by conventional direct methods of subtracting evapotranspiration (ET) from rainfall is practically limited in semi-arid regions by the difficulty in determining ET with sufficient accuracy for calculation of very low recharge fluxes in the order of a few mm/yr. In the Serowe study area two alternative methods are used to estimate spatial distribution of groundwater recharge. Based on 127 point measurements of chloride concentration in groundwater, the chloride mass balance method gave recharge rates ranging from 2mm/yr to ~30mm/yr with a harmonic mean of 12mm/yr. In order to present recharge spatially the data were interpolated by kriging of the logarithms of the net recharge values. GIS map modelling technique was used to integrate the influence of various recharge attributes in a single semiquantitative recharge potential map. The relative influence of factors such as soil type, vegetation, lineament density and slope among others was weighted and subsequently validated by site specific recharge rates. This method can give a valuable first estimate of the recharge rate based on surface properties identified from maps and remote sensing in areas where detailed hydrogeological information is limited.

1 INTRODUCTION Serowe area, like the whole of Botswana, is characterised by a low rainfall pattern and lack of surface water resources. People in this area depend mainly on groundwater. Recently, the population of Serowe has increased tremendously, making it the largest village in the country, thus widening the gap between demand and availability of water. In this regards, it is important for sustainability and management purposes to determine the renewable groundwater resources and how net recharge varies in space and time. This

Spatial variation of groundwater recharge

123

study focuses on estimation of net recharge rates using the chloride mass balance method and the assesment of spatial variability of both net recharge and potential recharge. Serowe study area is situated in the Central District of Botswana, at the eastern fringe of the Kalahari Basin and is about 275km northeast of the capital, Gaborone (Fig. 1). It is characterized by semi-arid climate with cool dry winters (May to September) and hot moist summers (October to April). It receives an average annual rainfall of 447mm/year. The most prominent feature in the area is the 90–150m high escarpment, which extends NNE-SSW splitting the area into two hydrologically contrasting areas: eastern and western. The western part slopes gently to the west and is covered by thick Kalahari sands. The eastern part slopes steeply to the east and Kalahari sand cover is thin or absent where the Ntane Sandstone or the basalt outcrops. The study area boundaries were assigned after the boundaries of the numerical model set up by Wellfield Consulting Services (WCS 1998, Lubczynski 2000). The northern boundary is a regionally stretching impermeable graben. The eastern and south-eastern boundary is delineated along the

Figure 1. Serowe study area. eastern wedging limit of the Ntane sandstone. The southern and south-western boundary is an impermeable dolerite dyke. The western boundary is an artificial one with no physical meaning and it stretches approximately from UTM coordinates (400000, 7517249) to (420150, 7544376). The study area is located entirely in the Karoo strata, where the most important aquifer is the ~100m thick Ntane Sandstone layer (Lubczynski 2000). The Ntane aquifer is underlain by the impermeable Mosolotsane Formation, mainly siltstone and mudstone and overlain by locally fractured Stormberg Basalt with thickness ranging from 0m to 143m. Groundwater recharge is expected where the basalt is absent so the Ntane aquifer lies directly under the Kalahari and where basalt is thin and fractured. The Kalahari sand cover, which overlies the Stormberg Basalt, is thick in the western part of the study area 50–75m and thin <10min the eastern part. The Kalahari sands have a high recharge potential where they do not contain calcretes and silcretes. The structurally dominant feature in the study area is the WNW-ESE trending fault system that is intruded by dolerite dykes of 10−40m width (SGC 1988, Lubczynski 2000). These features have divided the area into a series of horsts and grabens (Fig. 1), which enforces structurally controlled groundwater flow outward to the west and to the east of the groundwater divide located on the western side of the escarpment (Fig. 1).

Water resources of arid areas

124

2 GROUNDWATER RECHARGE ASSESSMENT The net groundwater recharge is defined by Lubczynski (2000) as the amount of water that reaches the groundwater minus groundwater evapotranspiration representing discharge of groundwater by tree transpiration and evaporation from groundwater in the form of upward convective water flux. Groundwater recharge is the most important factor in evaluating the renewability of groundwater resources of regional aquifer systems in arid and semi-arid environments and it is unfortunately the most difficult to quantify (Allison 1988). Several methods of estimating recharge have been developed. These can be divided into physically based, chemical and isotopic methods (Simmers 1997, Lerner 1990). Recently, also numerical models have been used to estimate groundwater recharge. While making site-specific recharge measurements, one of the most important problems to overcome is the spatial data presentation (Allison 1988, Lerner 1990). In this study, two techniques i.e. kriging interpolation (Ahmed et al. 1995, Gieske 1999) and spatial extrapolation with intergrated GIS recharge modelling technique were applied to assess spatial variability of recharge. 2.1 Chloride mass balance measurements The chloride mass balance method, developed by Eriksson & Khunakasem (1969), allows calculation of the average recharge rate in groundwater, RT in mm/yr provided the Cl ion behaves

Figure 2. Interpolated recharge map with net recharge estimates from the chloride mass balance method. conservatively and mass is conserved (Beekman et al. 1996): (1)

Spatial variation of groundwater recharge

125

where RT=total recharge rate (mm/yr), TCl=average annual total chloride deposition at the surface (mgm−2/yr); P=rainfall (mm/y); ClP=chloride content in rainfall (mg/l); Clgw=chloride content in groundwater (mg/l); D=dry deposition of chloride measured during dry season (mgm−2/yr). Due to lack of total dry deposition data during this investigation, estimation of recharge was based on the 1986/93 dry deposition of chloride (D) for Serowe of 442±124mgm−2/yr determined from rain gauge measurements (Selaolo 1998). Also Gieske (1992) recommended a similar D of 400–500mgm−2/yr for Serowe area. Based on 127 measurements of chloride in groundwater collected in September 2001, site-specific net recharge rates ranging from 2mm/yr to 74mm/yr were calculated according to equation 1. The results are shown in Figure 2. Two high values were regarded as outliers and have been omitted from further calculations. The method gave a harmonic mean recharge rate of ~12mm/yr, which is within range of estimates from previous studies. For example, SGS (1988) calculated net recharge rate of 11.7mm/yr from the chloride mass balance method. 2.2 Recharge interpolation Groundwater recharge is a function of space and time and is often highly variable. With regard to spatial variability, usually geostatistical analysis is carried out using only quantitative point data and qualitative geological information is often neglected (Ahmed 1991 in Ahmed et al. 1995). The key to geostatistics is the semi-variogram, which is a graphical presentation of spatial correlation in a given data set (Cohen & Spies 1990). The method assumes normally distributed spatial data. In this study the recharge data set was tested for normality using the Anderson-Darling normal probability test applying Minitab software. The test showed that the data is not normally distributed. The same data was successfully normalised by taking the logarithm of the net recharge values. These values were then further analysed geostatistically. The calculated semi-variogram and the spherical semi-variogram model are shown in Figure 4. A value of R2=0.71, indicating a good fit between the experimental data and the semivariogram model was found. Based on the parameters obtained from the semi-variogram model, the interpolation of the logarithms of the net recharge values was done using the ordinary kriging method. Finally, the antilogarithm of the interpolated values was taken resulting in the net recharge

Water resources of arid areas

126

Figure 3. Groundwater recharge potential map with point net recharge estimates.

Figure 4. Semi-variogram and semivariogram model for recharge. map (Fig. 2). The map of the net recharge distribution shows a general decrease in net recharge from the south-east and east towards the north and the west. Bulbs in the map

Spatial variation of groundwater recharge

127

indicate places of locally high and low net recharge. Low net recharge in the north could be explained by confinement by very thick basalt layer hampering net recharge. High evapotranspiration resulting in high chloride concentration in groundwater is the likely cause of locally low net recharge in the east. High net recharge in the northeast could be due to the absence of a confining basalt layer. Generally, recharge is lower in the western part than in the eastern part. This is due to the thick sand and basalt cover, deeper groundwater table and also lower rainfall in the western part of the study area. 2.3 Recharge attributes for GIS modelling Spatial variation of recharge is influenced by the spatial variability of factors specific to the study area like rainfall, vegetation type and density, soil type and texture, geology, landuse, topography,

Table 1. Kalahari thickness scores. Kalahari thickness (m)

Score

0–10 10–20 20–30 30–40 >40

5 4 3 2 1

Table 2. Depth to water table scores. Depth (m)

Score

<50 50–60 60–70 70–80

5 4 3 2

landform and depth to water table (Lerner et al. 1990). However, these factors influence recharge with different weights. Therefore, GIS modelling, which involves combining maps of different recharge attributes was used to come up with a recharge potential map, qualitatively displaying the spatial distribution of groundwater recharge potential. The modified index overlay method described by Bonham-Carter (1997) was applied in this study. Each attribute map was subdivided into six classes, with most recharge suitability assigned score 5 and 0 representing the least recharge suitability. Each attribute map was then assigned a weight according to its significance in controlling recharge. The average score of each pixel is defined as: (2)

Water resources of arid areas

128

where S=the weighed score for each pixel, Wi=the weight of the ith map and Sij=the score of the jth class of the ith map. The recharge potential map of the study area was derived considering the following factors influencing recharge potential: – Basalt cover: Basalt lacks primary porosity and impedes vertical flow of water except where it is fractured. Therefore high recharge suitability is expected where basalt is absent (score 5) and least recharge where basalt is present (score 0). – Kalahari thickness and depth to groundwater table: The thinner the Kalahari layer the faster water passes the unsaturated zone and reaches the aquifer and the less chances of water encountering duricrusts. The shallower the water table the faster water reaches the saturated zone. Assigned scores are shown in Tables 1 and 2. A correlation matrix showed that Kalahari thickness and depth to groundwater table are positively correlated. Therefore, the two scored maps were crossed and combined in a new map which was used for assigning recharge suitability scores. – Soils: Porous and coarse textured soils have high infiltration capacity and low field capacity and therefore enhance recharge processes. Soils with high silt and clay content do not release water to the lower zone fast enough to allow recharge processes to take place. Scores were assigned according to soil type and average infiltration rates as described by De Wit & Nachtergaele (1990), see Table 3. – Vegetation density: Areas with high vegetation density have also large density of root network and are therefore less suitable for recharge. High vegetation density may also facilitate infiltration by reducing runoff, but this is less important in case of Serowe due to the generally high infiltration capacity of the Kalahari sands. A NDVI map constructed from the Landsat TM 7 image of 24 April 1998 was classified into three recharge suitability classes (Table 4). – Slope: Slope of the landscape influences recharge rate. Generally, the steeper the slope the more runoff and the less the amount of water infiltrating the soil. A slope percentage map was assigned scores according to the FAO slope classification (Allen et al. 1998) see Table 5. – Lineament density: The influence of lineaments on recharge is greatest when the fractures and faults are deep, continuous over some distance and are not filled with secondary material. Not

Table 3. Scores for soils. Soil type Average infiltration rate (cm/hr)

Score

Arenosols 25–33 Regosols 22–30 Luvisols 0.05–0.8 Lixisols 0.05–0.8 Vertisols <0.05

5 4 3 3 1

Table 4. Scores for vegetation density. NDVI

Vegetation density class

Score

0–130

Low

5

Spatial variation of groundwater recharge

130–140 Medium 140–250 High

129

3 1

Table 5. Scores for slope. Slope (%)

Description

Score

0–2 2–8 8–16 16–30 >30

Flat Gently sloping Undulating Rolling Hilly

5 4 2 1 0

Table 6. Lineament density scores. Density

Class

Score

<20 20–40 >40

Low Moderate High

1 3 5

Table 7. Scores for vegetation cover. Vegetation cover

Score

Bare Open savanna Open savanna shrub Open savanna tree Pans Escarpment dunes Escarpment woodlands Outcrop area Hardveld shrub Hardveld woodland Agricultural area Riverine woodland Grassland Hardveld savanna

3 5 4 3 2 3 2 4 3 2 3 2 5 4

Table 8. Weight assigned to recharge attributes. Attribute

Model Model Model Model 1 2 3 4

Soil type Vegetation density Basalt cover Lineament density Slope

4 6

5 6

5 6

4 6

4 3

4 3

4 3

7 3

2

2

2

2

Water resources of arid areas

Kalahari 7 thickness/ depth to water table Vegetation cover 7

130

7

8

7

6

7

6

every lineament, but the majority of them can provide paths for infiltrating water. The density of lineaments was assessed from a combination of a lineament map obtained from WCS (1998) and lineaments derived from satellite TM5 data (Table 6). – Vegetation cover: Evapotranspiration and interception vary with vegetation cover. More evapotranspiration is expected from woodlands than from shrubs and grass. Therefore high recharge is associated with grasslands and low recharge with woodlands. In this assessment the vegetation map prepared by Ecosurve (1998) was used (Table 7). 2.4 GIS map modelling for mapping of recharge potential The recharge attributes considered for GIS map modelling do not bear equal importance to recharge potential mapping. Each map was therefore assigned a weight according to its perceived importance (Table 8). Most importance was attached to Kalahari thickness, depth to groundwater table, and vegetation cover, followed by soil type and vegetation density since they are

Table 9. Classification of recharge zonation maps. Weighted score

Recharge class

0–2.8 2.8–3.0 3.0–3.5 3.5–4.2 4.2–5

Very low Low Moderate High Very high

fundamental determinates of amount of water available for recharge followed by basalt cover. Lineament density and slope were considered the least important. There is not much variation in slope in the study area and the faults and fractures on the Kalahari are masked by the thick Kalahari sand cover. In the eastern part of the study area higher weights were applied to the lineament occurence. GIS map modelling refers to adjustments of attribute scores and attribute weights in order to provide the most realistic spatial distribution of recharge potential. Four different models based on weight sets given in Table 8 were used in this process. Four recharge potential map models were then produced following equation 2. In order to come up with recharge zonation models, each of the four recharge potential maps was subdivided into five recharge potential classes namely, very low, low, moderate, high and very high according to Table 9. Conceptually, all four recharge potential models are realistic when visually assessed. The four recharge potential models were then overlaid with recharge rates from the chloride mass balance method for comparison and validation of the GIS recharge potential models. In principle GIS recharge modelling should be manipulated by

Spatial variation of groundwater recharge

131

changing scores and weights to fit the point distribution of recharge (Lubczynski & Gurwin 2004). However, due to time constraints, this could not be done resulting in deterministic recharge potential rather than recharge potential modelling. Though verification data is missing for the far eastern and northern part of the study area, Model 3 is in best agreement with most of the point chloride mass balance results and was therefore considered the most realistic. Figure 3 shows the spatial variation of recharge potential based on Model 3 overlain with point net recharge values. In this map there are places in the central part of the study area with inconsistencies, where the moderate recharge values of the chloride mass balance fall in the very low recharge category of the recharge potential map. This could be explained as inaccuracy of the lineament assessment in terms of the size i.e. depth, width and openness which could not be incorporated in the GIS map modelling. Also, it is difficult to assess which lineaments are important for recharge and which ones are not. Other inconsistencies such as those in the northeast, north and center where low chloride mass balance recharge rates fall in the moderate to high recharge potential zones are most likely attributed to the subjective way of scoring and weighting in map modelling. It could also be attributed to the underestimation of recharge due to the presence of groundwater evapotranspiration in that area. 3 CONCLUSIONS – The chloride mass balance technique gives an insight into the spatial variation of net recharge. Net groundwater recharge in the Serowe study area is spatially variable and it generally ranges from 2mm/yr to ~30mm/yr with a harmonic mean of 12mm/yr. – The recharge potential map obtained by GIS recharge map modelling provides a semiquantitative distribution of recharge. This technique is a very useful tool in reconnaissance studies and in numerical model calibration, particularly when quantitative data is limited, because it is able to integrate spatial data from various sources. The zones do not necessarily show that recharge occurs or how much recharge occurs but give an indication of where recharge is most likely to occur. Such an assessment can therefore also be used as a planning tool taking into account local hydrogeological knowledge and constraints in the development, management and use of groundwater resources. – The process of GIS modelling as presented is largely based on expert knowledge of the modeller as well as knowledge of the area. More objective solutions can be obtained by optimization of the scores and weights to best fit the recharge attribute zones with the point measurements (Lubczynski & Gurwin 2004). – The pattern of net recharge interpolated by kriging is more or less similar to the extrapolated recharge pattern obtained by intergated GIS modelling of recharge potential. The differences can be attributed to inaccuracies in both methods and also to the fact that chloride mass balance gives net recharge while GIS map modelling shows potential recharge.

Water resources of arid areas

132

REFERENCES Ahmed, S., Sankaran, S. & Gupta, C.P. 1995. Variographic analysis of some hydrogeological parameters: Use of Geological soft data. Journal of Environmental Hydrology 3(2). http://www.hydroweb.com/ Allen, R.G., Pereira, L.S., Raes, D. & Smith, M. 1998. Crop evapotranspiration, guidelines for computing crop water requirements, Food and Agriculture Organisation of the United Nations, FAO Irrigation and Drainage paper 56, Rome, Italy. Allison, G.B. 1988. A review of the physical, chemical and isotopic techniques available for estimating groundwater recharge. In I.Simmers (ed.) Estimation of Natural groundwater recharge. NATO ASI series, C222, Reidel, Dordrecht: 49–72. Beekman, H.E., Selaolo, E.T. & Nijsten, G.J. 1996. Groundwater Recharge at the Fringe of the Botswana Kalahari-The Letlhkeng-Botlhapatlou Area. Botswana Journal of Earth Sciences 3. Cohen, W.B., Spies, T.A. & Bradshaw, G.A. 1990. Semi variograms of Digital Imagery for analysis of conifer canopy structure. New York: Elsevier Inc. Ecosurve. 1998. Vegetation mapping and ground truthing for Radar Imagery (Vegetation report). Serowe. Ecosurve project. Eriksson, E. & Khunakasem, V. 1969. Chloride concentration in groundwater, recharge rate and rate of deposition of chloride in the Israel coastal plain, Journal of Hydrology 7:178–197. Hendrickx, J.M.H. & Walker, G.R. 1997. Recharge from precipitation. In I.Simmers (ed.) Recharge of Phreatic Aquifers in (Semi-) Arid Areas. Rotterdam: Balkema. Gieske, A. 1992. Dynamics of groundwater recharge: A case study in the semi-arid eastern Botswana. PhD thesis, Vrije Unversiteit Amsterdam, The Netherlands. Gieske, A. 1999. Geostatistics for hydrologists, Principles and applications, Lecture notes, Adapted from de Marsily, 1986, ITC, Enschede, The Netherlands. Lerner, D.N., Issar, A.S. & Simmers, I. 1990. Groundwater Recharge: A guide to understanding and estimating natural recharge. International contributions to Hydrogeology 8. Lubczynski, M.W. 2000. Groundwater evapotranspiration—Underestimated component of groundwater balance in a semi-arid environment—Serowe case, Botswana. In Oliver Sililo et al. (eds), Groundwater: Past achievements and future challenges: 199–204. Rotterdam: Balkema. Lubczynski, M.W., Gurwin, J. 2004. Integration of various data sources for transient groundwater modelling—Sardon study case, Spain. Journal of Hydrology—in revision. Selaolo, E.T. 1998. Tracer studies and groundwater recharge assessment in the eastern fringe of the Botswana Kalahari, The Letlhakeng-Botlhapatlou Area. PhD thesis, Vrije Universiteit Amsterdam, The Netherlands. Swedish Geological Survey (SGS) 1988. Serowe Groundwater Resources Evaluation Project, Final Report, Ministry of Mineral Resources and Water Affairs, Department of Geological Survey, Lobatse, Botswana. Wellfield Consulting Services (WCS) 1998. Serowe wellfield 2 extension project (TB10/3/10/95– 96), Main report, DWA, Gaborone, Botswana. Wellfield Consulting Services (WCS) 2000. Serowe wellfield extension project, Groundwater Modelling report, DWA, Gaborone, Botswana.

Quantification of artificial ground water recharge G.C.Mishra Water Resources Development Training Centre, Indian Institute of Technology, Roorkee, India Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Sedimentary groundwater basins are mostly comprised of alternate layers of sand and clay. Occurrence of a clay layer at the top surface prevents direct recharge from rainfall. Natural and man-made surface drains in such a region are likely to carry away most part of the rainfall as direct runoff. These drains, while conveying the runoff, can be considered as a source of water for artificial recharge. Vertical shafts (recharge well) or a pit may be constructed in the bed of the surface water body through the intervening clay layer to facilitate recharge to the underlying confined aquifer. If the piezometric surface of the aquifer stands below the water level in the drain, the recharge would take place under the action of gravity. The recharge rate is governed by (i) the difference in the hydraulic heads at the water body and in the confined aquifer under the shaft, (ii) diameter and length of the shaft, (iii) transmissivity and storage coefficient of the aquifer being recharged, and (iv) hydraulic conductivity of the coarse material the shaft may be filled with.

1 INTRODUCTION Sedimentary groundwater basins are mostly comprised of alternate layers of sand and clay. Occurrence of a clay layer at the top surface prevents direct recharge from rainfall. Natural and man-made surface drains in such a region are likely to carry away most part of the rainfall as direct runoff. These drains, while conveying the runoff, can be considered as a source of water for artificial recharge. Vertical shafts (recharge well) or a pit may be constructed in the bed of the surface water body through the intervening clay layer to facilitate recharge to the underlying confined aquifer (Sandford, 1938). If the piezometric head in the aquifer stands below the water level in the drain, the recharge would take place under the action of gravity. The recharge rate is governed by (i) the difference in the hydraulic heads at the water body and in the confined aquifer under the shaft, (ii) diameter and length of the shaft, (iii) transmissivity and storage coefficient of

Water resources of arid areas

134

the aquifer being recharged, and (iv) hydraulic conductivity of the coarse material the shaft may be filled with. Appropriate methods of artificial recharge for different geohydrological conditions have been described in detail (Todd, 1985; Oaksford, 1985). In the present study, an analytical method is described to quantify time variant recharge from a surface water body to a confined aquifer through a vertical shaft or a recharge well under the action of gravity. In the paper, analytical solutions have been obtained applying unit response function coefficients known as discrete kernels and convolution technique to quantify time variant recharge from a surface water body to a confined aquifer through a vertical shaft and a recharge well. The following cases have been dealt: (i) a vertical shaft marginally penetrating into an aquifer and filled with coarse sand; (ii) a vertical shaft marginally penetrating into an aquifer; (iii) a recharge well fully penetrating an aquifer. A shaft with radius ranging from 1 to 2m filled with coarse sand can recharge at a significant rate between 250 to 700m3/day. If the shaft is not filled with sand, the rate of recharge at large time is twice that of when the shaft is filled with coarse sand. Recharge through fully penetrating well is more than 10 times that of the recharge through a vertical shaft filled with coarse sand. 2 STATEMENT OF THE PROBLEM A sedimentary groundwater basin consists of a confined aquifer overlain by an aquiclude and underlain by an impervious stratum. The aquifer is homogeneous, isotropic, and of infinite areal extent. The water level in the surface water body is at a height h1 above the bottom impervious base chosen as the datum. The thickness of the upper clay layer beneath the surface water body is L. Prior to onset of recharge, the piezometric surface in the aquifer stands at a height h2 above the datum. The height h2 is lower than h1. The aquifer can be recharged by constructing a vertical shaft in the bed of the surface water body through the intervening clay layer. The shaft may be filled with a filter material such as coarse sand to restrict contamination of groundwater. The aquifer can be recharged through a fully or partially penetrating recharge well. Quantification of recharge rate is sought for the following structures: (i) a vertical shaft marginally penetrating into the upper aquifer and filled with a coarse material; (ii) a vertical shaft marginally penetrating into the confined aquifer; (iii) a recharge well fully penetrating into the aquifer. 3 ASSUMPTIONS The assumptions made to quantify the recharge rate are:(i) the time span is discretised by time steps of uniform size ∆t (day); within each time step the recharge rate is uniform; the varying recharge is a train of pulses, (ii) an unsteady state is a succession of steady states, (iii) within a time step, Bernoulli’s equation is applicable.

Quantification of artificial ground water recharge

135

4 ANALYSIS 4.1 Case 1: A vertical shaft penetrating marginally into the aquifer and filled with a coarse material A vertical shaft penetrating marginally into an aquifer can be treated as a recharge well of zero penetration. Hantush (1961) has derived an analytical expression for evolution of piezometric surfaces in response to continuous uniform pumping from a well with zero penetration. The corresponding Hantush’s well function can be used to compute the evolution of rise in piezometric surface due to a unit pulse recharge. The response of a linear system to a unit pulse perturbation has been designated as discrete kernel coefficient (Morel-Seytoux, 1975). Morel-Seytoux and Daly (1975) have demonstrated the use of kernel coefficients in solving complex ground water flow problems. The rise in piezometric surface is expressed in terms of varying recharge and kernel coefficients derived from Hantush’s well function using a convolution technique. Let the time span be discretised by time steps of uniform size ∆t and let the time varying recharge through the shaft be treated as a train of pulses. Let R(γ) be the recharge during γth time step. Let δp(m, ∆t) be the rise in piezometric surface at the well face at time m∆t due to unit recharge (unit pulse input) that occurs during the first time step only. The expression for δp (m, ∆t), the kernel coefficient, is given in Appendix-1. The rise in piezometric surface, s(rw, m∆t), at the recharge well face at time m∆t due to variable recharge, R(γ), γ=1, 2, …, m, is given by: (1)

The hydraulic head at time m∆t at the bottom of the shaft is summation of the initial height h2 and the rise in piezometric surface, s(rw, m∆t). Applying Darcy’s law, the recharge during mth time step is given by: (2)

in which, kf=hydraulic conductivity of the coarse material the shaft is filled with. The term within the bracket is the hydraulic head difference dissipated in length L of the shaft. Solving for the recharge during the mth time step from (2)

Water resources of arid areas

136

(3)

R(m), m=1,2,…n, can be found in succession starting from m=1 to n.

Figure 1. A vertical shaft penetrating marginally into the aquifer and filled with a coarse material.

Quantification of artificial ground water recharge

137

4.2 Case 2: A vertical shaft penetrating marginally into the aquifer, with no filling material An analytical expression for recharge is derived applying Bernoulli’s equation (vide Streeter and Wylie, 1981). Accounting for the entry, exit, and friction losses and applying Bernoulli’s equation between points 1 and 2 for mth time step: (4)

(5) in which, ce=coefficient of entry loss, f=friction loss factor, L=length of the shaft, γw=unit weight of water, g=acceleration due to gravity (m/sec2), v=velocity of water in the shaft (m/sec) during mth time step, R(m) recharge volume (m3) during mth time step, ∆t=time step size (day). The first term in the right hand side of equation (4) accounts for entry loss, the second term accounts for friction loss in the shaft and the third term is the expansion loss at the exit of the shaft. s(rw, m∆t) is the rise in piezometric surface at the recharge well consequent to the recharge taken place. Incorporating (1) and (5) in (4) the following quadratic equation in R(m) is obtained: (6)

Considering the positive root of the equation (7)

in which,

For m=1, c=h2−h1. R(m), m=1, 2,…n, can be found in succession starting from m=1. The recharge rate during mth time step is equal to R(m)/∆t (m3/day).

Water resources of arid areas

138

4.3 Case 3: A recharge well fully penetrating into the aquifer The procedure for finding recharge through a fully penetrating well is same as that for the partially penetrating well described above except that the rise in piezometric surface at the well face is to be computed using discrete kernel coefficients pertaining to the fully penetrating well. The unit pulse response function coefficients δ1(m, ∆t) are obtained from the unit step response function derived by Hantush (1964) for a fully penetrating well of finite radius (Appendix II). 5 RESULTS The kernel coefficients are generated assigning values to the aquifer parameters. The thickness of the intervening clay layer is taken as 10m; the hydraulic conductivity of the packed porous medium is 10 times the hydraulic conductivity of the aquifer medium and is assumed to be 380 m/day. The friction factor f=0.02; entry loss coefficient ce=0.05; bl/rw. The corresponding non-dimensional recharge rate, R(m)/∆t T(h1−h2), with dimensionless time factor, , are presented in Figure 2. Since, the recharge rate is governed by the difference in the hydraulic heads at the entry and exit points of the shaft, and the head difference decreases with time, the recharge,

Figure 2. Variation of dimensionless recharge rate with time factor for

Quantification of artificial ground water recharge

139

Table 1. Average recharge rate for various radii; ce=0.05; f=0.02; kf=380m/day. Radius of the shaft (m) 0.1 0.2 0.4 0.5 1 2

Rate of recharge (case 1) (m3/day) 5.4213 19.8253 67.4339 97.8430 283.9961 700.755

Rate of recharge (case 2) (m3/day) 59.0824 116.8527 229.4277 284.1505 541.7252 991.7594

Rate of recharge (case 3) (m3/day) 1979.721 2078.622

therefore, decreases with time. When the shaft is filled with the coarse material, the rate of recharge at large time is half of the recharge rate that would occur without filling material in the shaft. Recharge through fully penetrating well is more than 10 times that of the recharge through a vertical shaft with a filler material. Numerical results are presented for the following aquifer parameters for various radii of the recharging structure: transmissivity, T=655.5m2/day; storativity, ; thickness of clay layer, L=10m; initial hydraulic head difference, h1−h2=5m. The average recharge rates during 120 days for different well radii are presented in Table 1. A vertical shaft with 2m radius, 10m length filled with a filter material having hydraulic conductivity of 380m/day, can recharge at an average rate of 700m3/day under an initial hydraulic head difference of 5m. 6 CONCLUSIONS Analytical methods are presented to estimate unsteady recharge, that can occur under the action of gravity, through (i) a vertical shaft filled with coarse sand, (ii) a well penetrating marginally into an aquifer, and (iii) a fully penetrating well. Application of unit response function coefficient is illustrated while quantifying the recharge rate. A vertical shaft with radius ranging from 1 to 2m filled with coarse sand can recharge at a significant rate between 250 to 700m3/day. 7 APPENDICES 7.1 Appendix I: Discrete kernel, δP (m, ∆t) Let the unit step response function for piezometric rise at the well face of a marginally penetrating recharge well and a confined aquifer system be designated as U(rw, t). According to Hantush(1961)

Water resources of arid areas

140

(1)

in which, T=transmissivity (m2/day), ф=storativity, and b1=thickness of the upper aquifer(m); rw=radius of the well or shaft(m),

and

Let the time domain be discretised by time steps of uniform size ∆t. The unit pulse response function of the system, δp(m, ∆t), is given by: (2)

W(u) and Wn(u, nπrw/b) are improper integrals as the upper limit of integration is infinite. W(u) is Theis’ Well function and can be computed using the polynomial and rational approximation (Abromwitz and Stegun, 1970) Wn(u,n πrw/b) is evaluated using Gaussian quadrature after converting the improper integral into proper integral and changing the limit. The procedure is as follows.

Quantification of artificial ground water recharge

141

As x→ −1, the value of the integrand in the second integration is found as follows:

These integration can be performed numerically using Gauss quadrature. 7.2 Appendix II: Discrete kernel, δ1 (m, ∆t) Hantush(1964) has derived the well function for computation of drawdown in an artesian aquifer due to pumping from a fully penetrating well of finite radius starting from the basic solution given by Carslaw and Jaeger (1959) for an analogous heat conduction problem. Let the unit step response function for piezometric rise at the well face of a fully penetrating recharge well and a confined aquifer system be designated as U1(rw, t). According to Hantush(1964) it is given by: (1)

in which,

Water resources of arid areas

142

functions of first kind of zero and first order respectively; Y0(x) Y1(x)=Bessel functions of second kind of zero and first order respectively; T=transmissivity (m2/day), and ø=storativity of the upper aquifer; rw=radius of the well or shaft(m). The integral in (1) is an improper integral as the upper limit of integration is infinite. The improper integral is reduced to a proper integral as described below.

Expanding the exponential term, and applying L’ Hospital’s rule, it can be shown that as v tends to −1, the integrand tends to 0. The integral I1 is a proper integral and can be evaluated numerically using Gauss quadrature.

Limit of the integrand at the lower is found as described below.

Quantification of artificial ground water recharge

Similarly,

Thus I2 can be evaluated using Gauss quadrature.

143

Water resources of arid areas

144

REFERENCES Abramowitz, M. & Stegun, I.A. 1970. Handbook of Mathematical Functions. Dover Publications, Inc, New York, 231pp. Carslaw, H.S. & Jaeger, J.C. 1959. Conduction of Heat in Solids. New York, Oxford Univ. Press: 338pp. Hantush, M.S. 1961. Drawdown around a partially penetrating well. J. Hydr. Div., ASCE, 87(HY4):83–98. Hantush, M.S. 1964. Hydraulics of wells. Advances in Hydroscience, Ed. Ven Te Chow, Vol. 1, 340pp. Morel-Seytoux, H.J. 1975. Optimal legal conjunctive operation of surface and ground water. Proc. Second World Congress. Intl. Water Resour. Assoc., New Delhi, Vol. IV:119–129. Model-Seytoux, H.J. & Daly, C.J. 1975. A discrete kernel generator for stream-aquifer studies. WaterResour. Res., 11 (2):253–260. Oaksford, E.T. 1985. Artificial Recharge: Methods, Hydraulics, and Monitoring. Artificial Recharge of Groundwater. Ed. Takashi A. Butterworth Publisher: 69–127. Sandford, H.J. 1938. Diffusing pits for recharging water into underground formation: chemical well cleaning methods. American Water Works Association Journal, 30(11):1755–1766. Todd, D.K. 1985. Groundwater Hydrology. New York, John Wiley & Sons: 458–493.

The architecture and application of the South African Groundwater Decision Tool I.Dennis & G.J.van Tonder Institute for Groundwater Studies, University of the Free State, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Groundwater forms part of an integrated water resource and needs to be managed accordingly. Currently, however, there are limited tools available for groundwater professionals as well as water resource regulators to make informed decisions concerning groundwater use and management as part of Integrated Water Resource Management (IWRM). The South African Groundwater Decision Tool (SAGDT) was developed to assist the Department of Water Affairs and Forestry (DWAF) and Catchment Management Agencies (CMA) in decision-making with regard to aquifer protection and management in South Africa. This paper discusses the legal requirements and policies that the SAGDT complies with together with the application architecture to highlight the methodology used in the design. Finally a simple case study is given to demonstrate the application of the tool.

1 INTRODUCTION Groundwater forms part of an integrated water resource and needs to be managed accordingly. Currently, however, there are limited tools available for groundwater professionals as well as water resource regulators to make informed decisions concerning groundwater use and management as part of Integrated Water Resource Management (IWRM). Traditional water resource planners and engineers find it difficult to conceptualise groundwater hydraulics and to come to terms with the estimated impact of groundwater utilisation on surface water sources. Groundwater professionals, on the other hand, need to know at what level they have to do their investigations to satisfy the requirements of the regulator.

Water resources of arid areas

146

The purpose of the SAGDT is to assist the Department of Water Affairs and Forestry (DWAF) and Catchment Management Agencies (CMA) in decision making with regard to aquifer protection and management. The following applies to the SAGDT: ● Consists of a standard system of consistent methods/rules to guide planning and decision making about water resources. ● Allows transparency, accountability and long-term goal-setting to be incorporated into water resource management. ● Calculates the level of confidence of results obtained. This paper discusses the legal requirements and policies that the SAGDT complies with together with the application architecture to highlight the methodology used in the design. Finally a simple case study is given to demonstrate the application of the tool. 2 LEGAL REQUIREMENTS AND POLICIES The SAGDT is aligned with existing legal requirements, policies and DWAF activities which are discussed in the sections below. 2.1 Legal requirements There are 3 Acts that are taken into account, namely; Constitution of the Republic of South Africa (Act No 108 of 1996), National Water Act (Act No 36 of 1998) and Water Services Act (Act No 108 of 1997). The Water Acts are aligned with the Constitution, which states: Everybody has the right to an environment not harmful to their health and well-being, to have an environment protected for the benefit of present and future generations, and to have access to sufficient food and water. The underlying principles of the Water Acts are therefore the sustainability and protection of water resources balanced by the use thereof. The SAGDT is based on the same underlying principles namely sustainability and protection of groundwater resources (quantity and quality). Warning systems are included when the sustainability of the system is at risk. For more information concerning the warning systems, refer to Section 3.3. 2.2 Resource Directed Measures (DWAF, 1999, 2003 & 2004) To implement the National Water Act, the DWAF has initiated resource directed measures (RDM). The steps in the RDM process includes delineating the area under investigation, classifying the resource, quantifying the reserve, setting resource quality objectives, and implementing monitoring. The first four of the five can be done within the SAGDT. In addition there are different levels of investigation, namely: desktop, rapid, intermediate and comprehensive. The SAGDT provides information, methodologies and guidance to perform the various levels of assessment. A DWAF initiative Framework program for Education and Training in Water (FETWATER) focuses on training and capacity building in the water sector in South Africa. A groundwater RDM network has been developed under this initiative. One of the

The architecture and application of the South African groundwater decision tool

147

objectives of the groundwater network is to develop training material and workshop the groundwater RDM process. The SAGDT has been included in this training and the relevant components will be presented at the workshops. 2.3 National Resource Water Strategy (DWAF(a), 2003) A National Water Resource Strategy (NWRS) is currently being developed as a framework for managing water resources in this country. The NWRS aims to provide a framework balancing the sustainability and protection of water resources and the use thereof. This is the focus of the SAGDT. In addition the 19 Water Management Areas (WMAs) defined, the NWRS will be included in the SAGDT. According to the NWRS a water allocation plan must be developed for each of these WMAs, the SAGDT can assist the water planner in developing the groundwater component of these water allocation plans. 3 FUZZY LOGIC IN RISK ANALYSIS Conventional set theory (Boolean) states that an element is either a member of a set or not. Consider the following real-life problem: A person is said to be young when they are under the age of 25 and a person is said to be old when they are over 40. In which group would we place a person of the age of 30? Fuzzy logic is an extension of conventional set theory enabling an element to belong to a set to a degree. The degree of membership is a function that defines the membership of an element to a set according to the value of the element see the Figure 1. Membership is expressed as a value between 0 and 1. Zero implies 0% membership and 1 implies 100% membership. The solid line describes the membership function for the set of people older than 40 and the dotted line describes the membership function for the set of people younger than 25. Note that in most cases the membership functions of the two sets will be inverses. To answer the question of where will a person of the age of 30 fit in can be as follows: That person belongs 75% to the set of young people and 25% to the set of old people.

Water resources of arid areas

148

Figure 1. Age membership function. Table 1. Fuzzy logic rule set for 3 inputs. Rule Weight Input 1 no. 1 2 3 4 5 6 7 8

0.0 ? ? ? ? ? ? 1.0

Input 2

Input 3

Favourable Favourable Favourable Favourable Favourable Unfavourable Favourable Unfavourable Favourable Favourable Unfavourable Unfavourable Unfavourable Favourable Favourable Unfavourable Favourable Unfavourable Unfavourable Unfavourable Favourable Unfavourable Unfavourable Unfavourable

Selection of the membership function is done by an expert on the field of study. Linear membership functions are seldom used in practice in contradiction to sinusoidal functions, which are very popular. In most cases risk analysis will involve more than one input to be considered in the analysis. Fuzzy logic makes it possible to generate a set of decision rules according to the number of inputs, and these rules must then be evaluated by an expert in the field of study. The number of rules generated is given by Equation 1. n=2inputs (1) where n represents the number of rules generated. The rules consist of all possible binary combinations of the respective inputs with a weight assigned to each rule representing the risk. Table 1 shows the decision rules generated for 3 inputs. Instead of true and false the terms favourable and unfavourable are used to make the rules easier to read.

The architecture and application of the South African groundwater decision tool

149

Rule number 1 is read as: If input 1 is favourable and input 2 is favourable and input 3 is favourable then the risk is 0%. All of the other rules are read in the same fashion and an expert must evaluate each rule individually to assign the appropriate risk. For each input a membership function must be defined with a favourable and unfavourable limit defining the two sets. One function will represent the favourable set and the other the unfavourable set. Thus, for each input, a favourable and an unfavourable value can be read from the membership functions. For each input the table of decision rules is then populated with the respected favourable and unfavourable degree of membership and the risk is calculated using Equation 2: (2)

where n=number of rules DOM=degree of membership Wn=Weight of rule n Note that the minimum function must return the minimum value of all inputs for each rule. 4 SYSTEM ARCHITECTURE The sub-systems comprising the SAGDT are discussed in the sections that follow. Refer to Figure 2 as reference for the sections to follow. 4.1 3rd Party Software The 3rd party software included in the SAGDT is to provide a one-stop application to the user. The SAGDT makes provision for most user requirements without having to search for additional resources. This software will not interface automatically with the SAGDT environment, but should be seen as additional utilities provided to the user.

Water resources of arid areas

150

Figure 2. High level architecture of the South African Groundwater Decision Tool. 4.2 GIS Tool The GIS database will consist of a core set of shape files that ship with the application. Users are able to extend this database by importing custom generated shape files. New shape files added to the database will not be used in the risk assessment unless the user maps the shape file over an existing core shape file (certain criteria must be met for this operation). The GIS Tool comes with a query builder that interfaces with the whole GIS database, and users are encouraged to extend their personal database to take advantage of the query builder functionality. Using the GIS Tool provides some of the functionality of a full-blown GIS application, with a fraction of the complexity of commercial GIS systems. 4.3 Risk tool The risk tool is a fuzzy logic-based multi-criteria risk assessment tool. Two spatial environments exists in the SAGDT, i.e. the external and internal worlds as shown in Figure 2. The external world represents the whole of South Africa and the user selects a single coordinate as starting point of the scenario. The internal world then uses the specified coordinate as reference to query the GIS database. The user builds a scenario in the internal world through the use of objects available in the object repository. A simplified version of the object model is shown in Figure 3.

The architecture and application of the South African groundwater decision tool

151

In some instances more than one object performing the same function exists, and they differ only in the detail their attributes require as well as their respective confidence levels. In building a scenario, an object tree will result due to the parent child relationships of the objects used. The base object will always be a GIS object representing the area of the internal world. This object tree is then passed to the fuzzy logic engine to determine the risk assessment result, and further analysed to determine the confidence levels associated with it. A warning system exists so that the user will be notified when to do a more detailed scenario. The warning system uses the confidence level, risk assessment result and policy as inputs to

Figure 3. Simplified SAGDT object model for illustration purposes. determine if a warning should be issued. The SAGDT supports the following risk assessment categories: Sustainability, Health, and Ecological. After the completion of a risk analysis the SAGDT produces a risk profile report containing the following information: ● Description of area (area object name) ● Picture of area (internal world snapshot) ● Summary of object properties and calculated values ● Risk assessment per applicable category together with the confidence level ● Warning system response

Water resources of arid areas

152

4.4 Scenario wizard Tutorials are available in the form of a scenario wizard. The wizard gives users step by step instructions to build predefined scenarios. The wizard also guides the inexperienced user in not only learning the software, but also the methodology used in the design of the software. 4.5 Application help file The application help file is a Windows® based help file providing the user with help on the Graphical User Interface (GUI), operation and functionality of the application. 4.6 Groundwater Dictionary The Groundwater Dictionary contains terms and definitions related to groundwater illustrated by graphics, photos and animations where possible, to ensure that concepts are correctly understood. 4.7 Object help A scenario is built using available objects in the repository or library. An extensive help database is available that specifically describe each object and its functionality. This help file contains the mathematical description of each object and all associated properties, as well as interfacing with the other objects. By making this available to the user, the methodology applied in each object can be understood and an object is not just a “black box” to the user, but allows him to make educated decisions when populating the object attributes. 5 CASE STUDY 5.1 Geology The region consists of mudstones, shales and sandstones from the Adelaide Subgroup of the Beaufort Group within the greater Karoo Super Group. Post-Karoo dolerite sill intrusions are present, which have to a large extend been eroded, exposing the underlying sedimentary rocks. The surface dolerite on the Campus are highly fractured with little ground cover and it is assumed that recharge over these areas is probably high. 5.2 Study area The Campus test site is located on the grounds of the University of the Free State, South Africa, and covers an area of approximately 180×192m2. The aquifer is intersected by thirty percussion and seven core-boreholes as shown in Figure 4. Core samples indicate parallel horizontal fractures, the most significant of which is at a depth of 21m In more

The architecture and application of the South African groundwater decision tool

153

weathered sections of the aquifer, diagonal fractures intersect the bedding plane fractures. The sandstone containing the most horizontal fractures also forms the main watercarrying formation.

Figure 4. Borehole positions on the Campus test site. 5.3 Scenario The scenario that will be evaluated for this case study is the determination of the sustainable yield of UO5 when pumped continuously for 2 years. The assumptions used in the case study are as follows: ● Only UO5 will be pumped at variable rates over the assessment period. ● The assessment period is 2 years. ● Three levels of evaluation will be done, that is each successive analysis uses objects with higher confidence levels than the previous set of objects. This implies that more detailed data are needed for objects used with higher confidence levels. 5.4 Results Figure 5 shows the results obtained from the SAGDT for the specified scenario. From the graphs it is clear that the higher the confidence level of the scenario, the more accurately the risk of failing can be determined. There exists a good correlation between the 99%

Water resources of arid areas

154

risk of failure for each of the confidence levels evaluated, but the higher confidence scenarios give opportunity for better management. As a recommendation one could propose a 0.4L/s abstraction rate from the 89% confidence scenario, which indicates a 20% risk of failure. From extensive field investigations it has been proved that UO5 can be pumped for 6 months at 0.33L/s without failing, which correlates well with the proposed recommendation. It is important to note that the tool will produce an overestimate for the risk when the confidence is low. This is why a warning system was implemented to make sure the user is aware of the fact that the risk is too high according to policy and that a more detailed analysis is required to confirm the high risk situation. This will prevent users from making management decisions based on high-risk results with low confidence.

Figure 5. Risk of UO5 failing for various scenarios. 6 CONCLUSIONS The SAGDT has proven to be a powerful groundwater management tool. The tool provides a common framework for all groundwater practitioners in South Africa in which they can perform groundwater risk assessments that relate to policy. By employing fuzzy logic to do the risk analysis the user has the knowledge of an expert captured in the application to assists in the decision-making process. The SAGDT also acts as a groundwater educational environment, due to the extensive groundwater dictionary and object help files available.

The architecture and application of the South African groundwater decision tool

155

REFERENCES Constitution of the Republic of South Africa. Act No 108 of 1996, Pretoria, South Africa. DWAF 1999. Water Resources Protection Policy Implementation—Resource Directed Measures for the Protection of Water Resources Version 1.0 Volumes 2–6; Department of Water Affairs and Forestry, Pretoria. DWAF 2003. Resource Directed Measures—Module 1—Introductory module; draft edition August 2003, Department of Water Affairs and Forestry, Pretoria. DWAF 2003a. National Water Resource Strategy; current draft edition, Department of Water Affairs and Forestry, Pretoria. DWAF 2004. Groundwater resource directed measures training manual. Sponsored by Fetwater, Department of Water Affairs, Pretoria. National Water Act. Act 36 of 1998, Pretoria, South Africa. Water Services Act. Act No 108 of 1997, Pretoria, South Africa.

The development of a groundwater management tool for the Schoonspruit dolomitic compartment B.H.Usher Institute for Groundwater Studies, University of the Free State, Bloemfontein, South Africa S.Veltman Department of Water Affairs and Forestry, Geohydrology Division, Free State Region, Bloemfontein, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The Schoonspruit dolomitic compartment is situated in the relatively arid Northwest province of South Africa. The compartment is a critical management area since it feeds the Schoonspruit “eye”, a perennial spring feeding into a stream along which several communities and irrigation farmers abstract water. Due to large-scale groundwater abstractions and declining water levels, the area was declared as a groundwater protectorate in 1995. Water management has been devolved to local level, and in this instance a user-friendly practical management tool was needed for the Water User Association. The compartment was investigated to identify the compartment boundaries, groundwater flow directions, recharge relationships and the eye flow response over time. Use was made of long-term monitoring data, isotopic and hydrochemical data to identify recharge zones. Two zones were identified as groundwater management units in the compartment and groundwater balances for the two zones were defined. The cumulative rainfall departures method and a method of moving averages were among techniques employed to define calibrated recharge relationships for the eye. Different threshold rainfall values and recharge factors have been determined for two zones. These equations were incorporated into a simplified electronic management tool. Time-dependent inputs into this tool include rainfall, hydrochemical and water level data, while factors such as the basic human needs, reserve requirements and allocated volumes are built in. This is translated into allocable volumes for irrigation into the compartment and predicted spring flow volumes. With this tool, groundwater management is facilitated and the sustainable use of the water resources in this area can be more accurately considered.

The development of a groundwater management tool

157

1 INTRODUCTION 1.1 Overview of the study area The Schoonspruit dolomitic compartment is a dolomitic aquifer situated to the North and Northwest of the town Ventersdorp in the Northwest province. The compartment has been named after the Schoonspruit Eye, which is dependent on the compartment for flow. The Schoonspruit Eye, in turn, is the sole reason why the Schoonspruit has a constant flow and provides a municipality and two surface water irrigation boards with surface water all year. The Schoonspruit dolomitic compartment is situated north and northwest of Ventersdorp in the Northwest province. With the proclamation of The National Water Act, Act 36 of 1998, a new responsibility towards groundwater and groundwater management developed. Regional Offices were given the responsibility of managing these resources as acting Catchment Management Agencies. Groundwater is a resource that needs management and decisions based on sound scientific principles, regarding allocable volumes from the groundwater resources must be made. With the above principles in mind, the aim of the study was the development of a technical methodology and a first-order technical groundwater management tool to manage groundwater in the Schoonspruit compartment according to geohydrological principles, within a Groundwater User Association. This management tool had to be able to, on a year-to-year basis, determine the volumes available in the aquifer for allocations. 2 THE SCHOONSPRUIT COMPARTMENT 2.1 Overview The setting can be described in more detail as the compartment is categorised as Transvaal Highlands with elevation changes of more than 100m over a 40-km distance. The topography slopes downward from the northeast to the southwest. The Pretoria Formation in the north forms the water divides, in the north of the compartment, between the Vaal and Limpopo rivers. The Schoonspruit compartment falls within the surface water drainage area C24, drained by the Schoonspruit, and circular depressions can be found in the area that show elements of karstic evolution. Most of the rainfall occurs from November to February. The average rainfall for the area is 606mm and the average evaporation is in the vicinity of 1900mm. 2.2 Geology The geology of the area can best be described by differentiating between the main geological systems. In general the geology is known as dolomites of the Malmani Subgroup that plunge regionally northward and are overlain by the Pretoria Group. Outcrops of the Witwatersrand Supergroup appear along the southern boundary of the dolomites (Fleisher, 1981). The Malmani Subgroup is described as dolomite, banded iron formation, chert and shale. This series consists mostly of layered strata of calcium

Water resources of arid areas

158

magnesium carbonates (CaMgCO3), some layers massive and some with chert bands. Secondary limestone also occurs in the dolomites and is widely mined for the manufacturing of cement. Dolomites in this area are generally easily weathered and form undulating landscapes. (Kok, 1972) The Subgroup is further described as representing the dolomitic sequence and is largely concealed by overburden throughout the study area and therefore difficult to trace (Polivka, 1987). The majority of outcrops and aquifers in the area are associated with the Malmani Subgroup (Kotze, 1994). 2.3 Geohydrology The dolomitic aquifer of the Schoonspruit dolomitic compartment consists of four different formations (Polivka, 1987). Of these the chert-rich formations, Monte Christo and Eccles, are better aquifers compared to the chert-poor formations, Oaktree and Lyttleton, and boreholes drilled on fault intersections also gave high yields (>10l/s) (Kotze, 1994). The strata dip northward and are overlain by the Pretoria Group. Average borehole yields differ for the different formations and range from 11l/s in the Eccles to 3l/s in the Lyttletone formation (Polivka, 1987). Agriculture has the most important influence on the compartment’s water quality. As such, factors such as nitrate pollution are of particular importance. 2.4 Springs and water users The area is solely dependent on groundwater either directly, as abstraction from boreholes, or indirectly, as abstraction out of the Schoonspruit. The Ventersdorp Local Municipality indicates an average daily consumption of 16.3m3, including the four communities within the Ventersdorp Municipal District and on the dolomitic compartment: Ga-Mogopa, Ga-Motlatla, Goedgevonden and Tsêtsê. The largest water use within the compartment is abstraction for agricultural irrigation

Figure 1. Monthly abstraction values in the Schoonspruit dolomitic compartment.

The development of a groundwater management tool

159

Figure 2. Annual and monthly flow volumes of the Schoonspruit Eye. purposes. In the Schoonspruit river below the eye, there are four water users, claiming water rights from the Schoonspruit Eye. They are Ventersdorp municipality, the Klerksdorp Irrigation Board (IB), the Schoonspruit Governmental Water Scheme and the vlei down stream from the weir. (Darcy Consultants, 2002). Figure 1 shows the monthly abstraction volumes as used in the water balance methods. Note that these abstraction volumes include abstraction from all boreholes, spring flow volumes from the system, leakages and water consumed by evapotranspiration. There are several springs of which the Schoonspruit eye is by far the most productive, although the flow of the eye has decreased over time. This decrease is most likely associated with increased groundwater abstraction, as well as decrease in rainfall. Figure 2 shows the annual and monthly flows of the Schoonspruit Eye up to September 2002. Annual flows were calculated for hydrological years, starting October and ending in September of the following year, e.g. 1-Sep-81 would be flows for October 1980 to September 1981. It is clear from the flows that an average flow volume would not be an indication of the true situation in the compartment and it is necessary to do an analysis of the effect of rainfall and the lag-time effect of recharge through the compartment on the flow of the eye.

Water resources of arid areas

160

3 GEOHYDROLOGICAL DESCRIPTION The most important aspect for the geohydrological description was the determination of an appropriate water balance methodology for the compartment. Details regarding the verification of the compartment boundaries, current water quality and aquifer parameter determination are contained in Veltman (2003). 3.1 Recharge and water balance methods The most important input into the water balance of the area is incoming water. While some water does migrate into the compartment across the boundaries of the compartment, the vast majority of the water in this dolomitic terrain is derived from recharge from rainfall. Several recharge methods, based on methods described by van Tonder et al., 1999, were used. Recharge in the area is high due to soils that are transmissive and areas of karstification, which allow rapid recharge. Several recharge methods were used, a level of certainty assigned to each method and a weighted recharge estimate obtained. By using the weighted values, a value that is reliable for recharge can be obtained. Recharge was estimated as 6.0% of annual rainfall, amounting to the average volume of 70.68Mm3/a. 3.2 Springflow simulation Simulations of the groundwater monitoring borehole levels were done using the CRD and MA methods and the Schoonspruit Eye flow were done using MA springflow software currently being developed by Bredenkamp, 2003. 3.2.1 Groundwater simulations Groundwater levels and rainfall data are a critical input into most simulations and evaluations, of which the Saturated Volume Fluctuation (SVF), Cumulative Rainfall Departure (CRD) and Moving Average (MA) methods give the best simulations for determining natural water levels and various aquifer parameters. Where spring flows are linearly related to the CRD and MA of a rainfall series, and thus also linearly related to the groundwater system from which it is recharged, both the natural flows and the effect of abstraction can be simulated (Bredenkamp, 2000). The simulations of the groundwater monitoring borehole levels were done using the CRD and MA methods and the Schoonspruit Eye flow using MA springflow equation. 3.2.1.1 Moving average method This method mimics the groundwater level of a specific month to the average rainfall over a number of preceding months (Bredenkamp, et al., 1995) and is described by Equation 1 (Bredenkamp, 2000):

The development of a groundwater management tool

161

hi=(b/s)/n.Rfj+F (1) Where: hi=the groundwater level for month I, b=coefficient of rainfall representing recharge, s=aquifer storativity, n=number of months, Rf=average rainfall for preceding months J, F=inferred depth of aquifer below surface. 3.2.1.2 CRD method The groundwater balance can be explained by the concept that equilibrium in an aquifer is established over time between recharge (average gains) and drainage (average losses) and is expressed with Equation 2 (Bredenkamp, et al., 1995): Rfave=ROave+REave+EVTave (2)

Table 1. Aquifer parameters determined with the MA & CRD methods. Parameter

Zone A

Zone B

Aquifer thickness (m) Recharge (%) Threshold monthly rainfall (mm) High rainfall recharge factor (%) S

8 8 26

13 7.63 24.33

30

30

0.027

0.023

where: Rfave=average rainfall, ROave=average runoff, REave=average recharge, and EVTave=average evapotranspiration. The CRD method corresponds to the concept that equilibrium is established in an aquifer over time, therefore matching the groundwater level fluctuations to the cumulative rainfall departure from the average rainfall, can mimic the hydrological balance of an aquifer (Bredenkamp, et al., 1995). Defining the CRD relationship for different time intervals yields Equation 3 CRDi=CRDi−1+Rfi−k.Rfave (3) Where: CRDi=CRD at month I, CRDi−1=CRD at the month preceding month I, Rfi=rainfall at month I, Rfave=average rainfall and k=coefficient representing abstraction. 3.2.1.3 Springflow simulations When doing the CRD_MA simulations one has to incorporate inflows and outflows (abstractions) and adjust storativity and recharge values to attain the best possible curve to fit the actual groundwater level measurements that was taken. The simulation therefore

Water resources of arid areas

162

incorporates the groundwater balance to a degree, although levels and not volumes are of concern. The storativity and recharge values, attributed to the specific borehole’s reaction to averaged rainfall over an area, are obtained. In the dolomitic aquifer these values are not as related to fracture flow, since the dolomites characteristics cause variations in groundwater levels to be smoothed over an area. The values are therefore indicative of the aquifer characteristics. Further information gained from these simulations is the effective depth of the aquifer and the threshold values of rainfall before recharge will take place. Table 1 summarise the information as acquired with these simulations. The simulations provide valuable information for use in regional modelling of the aquifer, and for determining the groundwater balance. However, for proper management of the relationship between the system’s response and the flow of the Schoonspruit Eye needs refinement. 3.2.1.4 Spring flow When relating the CRD relationship to spring flow, Equation 4 can be used to simulate this relationship (Bredenkamp, 2000): QI (4) spring=J/S.ρ.CRDI+CFLOW Where: QI spring=spring flow at month I; J=hydraulic coefficient+flow cross section width constant; S=aquifer storativity; ρ=coefficient of rainfall representing recharge; CRDI=CRD at month I and CFLOW=long-term average spring flow around which the flow fluctuates. Various factors can be introduced to simulate different situations, e.g. aerial extent, abstraction or different lag time effects. When the moving average of rainfall is used, lag time effects of rainfall events can be included and its integration over the aquifer (Bredenkamp, et al., 1995). When simulating the spring flow all known parameters are incorporated and the unknown parameters are calibrated to attain the best fit. Spring flow parameters have been incorporated together with the different moving averages of rainfall. The Schoonspruit Eye simulation is shown in Figure 3 and simulated with the following equation: Schoonspruit Flow (Mm3/m)=ReN+ReF−AbsGW

The development of a groundwater management tool

163

Figure 3. Schoonspruit Eye simulated flow with a 96 month moving average. where: ReN=recharge under normal rainfall events, ReF=recharge under flood rainfall events and AbsGW=groundwater abstractions from the drainage area (Mm3/m) And ReN=ReN%/100*Rf24MMA/Rf120MMA*(Rf96MMA−ThN)*A/1000 ReF=ReF%/100*(RfFLOOD)*A/1000 RfFLOOD=IF((Rf120MMA−ThF)>0, Rf120MMA−ThF) The calibrated parameters for the system have been determined as ReN (7%), ReF%(44), ThN(26mm) which the recharge threshold and ThF(43mm) which is the flood recharge threshold. The Rf values all refer to the month lag included. The equation therefore amounts to Equation 5: Schoonspruit Flow (Mm3/m)=(0.07*Rf24MMA/Rf120MMA*(Rf96MMA−26)*0.842) (5) +(0.44*(IF((Rf120MMA−43)>0, Rf120MMA−43))*0.842)−AbsGW This equation is used in the groundwater management tool of the Schoonspruit dolomitic compartment. The biggest advantage of this method is that abstractions can now be incorporated into the simulation and predictions can be made with long-term predicted rainfall. The effective recharge for the Schoonspruit Eye was determined as 13% for 2002.

Water resources of arid areas

164

4 MANAGEMENT TOOL 4.1 A first order groundwater management tool The basic principle of a first order tool is to include the essential mechanisms in an understandable format, which will be used by the most basic groundwater manager. Inputs into the tool must be simple and outputs easily usable, while a layman should not change the driving equations. When new information becomes available, the tool should be easily modified by professionals, to include refined parameters or simulations. Developing a groundwater management tool is dependent upon the geotechnical controls essential to the management of the dolomitic compartment and which are only beneficial. Therefore the tool cannot be developed before the geohydrological evaluation has been complete and all essential controls have been defined and determined. Essential outputs from the tool include groundwater balances for various zones in the compartment, annual recharge volumes and therefore allocable volumes in the compartment, spring flow simulations for predictions from rainfall, including allocable volumes for both groundwater and surface water users and a classification of the groundwater quality based on the standards for the use of the water on the compartment. Beneficial outputs from the tool include a warning system if the Resource Quality Objectives are not met and the management class of the aquifer incorporated into the allocable volumes. 4.2 Users The users of such a tool range from the groundwater user’s association to the regulators and also groundwater consultants operating in the area. The tool needs to be versatile and contain all the necessary geohydrological equations, yet at the same time to be userfriendly. Equations for inclusion in this tool included the Schoonspruit Eye simulation equation and incorporation of domestic and ecological requirements. 4.3 Input The tool was constructed in such a way that only the latest rainfall and water quality need to be included as time-variant data. Aspects such as the reserve requirements and rainfall/recharge equations are built in. 4.4 User-friendly tool The management tool was programmed with macros which guide the user, in MS Excel. The Title page of the Tool is an information page only and a navigational button move to the next sheet, the Menu sheet. The “Enter Data” button navigates to the Data sheet, the “Compartment Map” button to the Map sheet and the “Assign/check Volumes” button to the Volume sheet. This is a navigational sheet only and the pathways are inserted here is up to the developer. The Data sheet of the Tool, allows input of the groundwater

The development of a groundwater management tool

165

quantities and qualities. Data input includes compulsory data inputs and optional data inputs. The optional data are helpful if available, but the simulations are not dependent on these cells to run. Navigational buttons to other sheets and data entry points are also included. The Drinking Water Quality Classes for the different parameters are included as fixed parameters. On the Prediction sheet of the Tool all the calculations for the simulation of the Schoonspruit Eye flow, therefore allocable volumes are done. The simulated flow (Mm3/m) is then determined using the spring flow equation as only rainfall values, and not equations, are now incorporated. Input data to this sheet is obtained from the Data sheet. Allocable volumes are determined with the simple equation of subtracting surface water demand from the simulated flow, as this has already taken into account current groundwater use. Figure 4 shows the spring flow and allocable volume graph. 5 DISCUSSION The aim of the groundwater management tool was to provide a first order technical tool, which is a practical and workable tool, for use by the WUA in determining allocable volumes. The following conclusions are made with regard to the groundwater management tool: ● Input and output parameters as outlined in this paper were used and proved to be sufficient for defining quantity and quality concerns in the Schoonspruit dolomitic compartment. ● Allocable volumes can be determined for the two zones using predicative rainfall data. ● The Schoonspruit Eye can be simulated using predicative rainfall data with the following equation:

Schoonspruit Flow (Mm /m)=(0.07*Rf24MMA/Rf120MMA*(Rf96MMA−26)*0.842) +(0.44*(IF((Rf120MMA−43)>0, Rf120MMA−43))*0.842)−AbsGW 3

Water resources of arid areas

166

Figure 4. Spring flow and allocable volume graph of the SGM tool. ● The drinking water quality classes were introduced, as a useful parameter, as part of an early warning system where drinking water quality is of concern. ● The tool is sufficient to continue with groundwater management in the dolomitic compartment. ● Verification of the lawful users is of utmost importance for groundwater management to be successful. ● The tool is a practical and useable tool for all groundwater managers and planners. The following recommendations are made with regard to the groundwater management tool: ● Groundwater management should commence at once and the tool tested against annual data. ● Verification of lawful water uses should continue and be completed as soon as possible. ● The tool should be tested and applied to other dolomitic areas.

REFERENCES Bredenkamp, D.B., Botha, L.J., Van Tonder, G.J. & Van Rensburg, H.J. 1995. Manual on quantitative estimation of groundwater recharge and aquifer storativity. Report no. TT 73/95. Water Research Commission, Pretoria. Bredenkamp, D.B. & Swartz, A. 1987. Reconstruction of the flow of springs by means of annual recharge estimates. Technical report no. GH 3525. Department of Water Affairs, Directorate Hydrology, Pretoria. DARCY Groundwater Scientists and Consultants. 2002. A catchment management plan for the Schoonspruit and Koekemoer Spruit catchments: A groundwater situation analysis. Department of Water Affairs & Forestry, Bloemfontein.

The development of a groundwater management tool

167

Fleisher, J.N.E. 1981. The geohydrology of the dolomite aquifers of the Malmani Subgroup in the SouthWestern Transvaal, Republic of South Africa. Technical report no. GH 3169. Department Water Affairs & Forestry, Directorate Hydrology, Pretoria. Kok, T.S. 1972. Wes-Transvaal en Noord-Kaap waterbeplanningstreek—geologie, fonteine en myne in opvanggebied. Technical report no. GH 1758. Department of Mines, Geological Survey, Pretoria. Kotze, J.C. 1994. Summary of the Geology, Geohydrology, and Boundaries of the proposed SGWCA, District Ventersdorp, Drainage Area C24. Technical report no. 3833. Department of Water Affairs & Forestry, Directorate Hydrology, Pretoria. National Water Act, Act No. 36 of 1998. Polivka, J. 1987. Geohydrological investigation of the Schoonspruit compartment in the dolomitic area of Ventersdorp. Technical report no. GH 3524. Department of Water Affairs, Directorate Hydrology, Pretoria. Schoeman & Vennote. 1996. Ventersdorp Oog Ondergrondse Staatswaterbeheergebied. Report no. B0307/2. Department of Water Affairs & Forestry, Sub directorate Water Allocation, Pretoria. Selaolo, E.T. 1998. Tracer Studies and Groundwater Recharge Assessment in the Eastern Fringe of the Botswana Kalahari. Ph.D. thesis, Free University of Amsterdam. GRES Project Publication. Van Tonder, G. & Xu, Y. 2001. A guide for the estimation of groundwater recharge in South Africa. The Institute of Groundwater Studies, Bloemfontein. Vegter, J.R. 2001. Groundwater development in South Africa and an introduction to the Hydrogeology of groundwater regions. Report no. TT 134/00. The Water Research Commission, Pretoria. Veltman, S. 2003. A Methodology for Groundwater Management in Dolomitic Terrains with the Schoonspruit Compartment as Pilot Area. Unpublished M.Sc thesis. University of the Free State, Bloemfontein, South Africa.

Effects of mining and urban expansion on groundwater quality in Francistown, Botswana Benjamin Mafa Department of Water Affairs, Gaborone, Botswana Horst Vogel Department of Geological Survey, Lobatse, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: This study was carried out as part of a technical cooperation project between the Department of Geological Survey in Lobatse and the German Federal Institute for Geosciences and Natural Resources. The aim was to determine if groundwater pollution had taken place in Francistown (NE Botswana) due to urban expansion and/or historic gold mining activities, and to delineate affected areas as well as potential groundwater hazards on thematic maps which were designed in a digital and easily readable form for future development planning by urban planners. The results revealed that groundwater in Francistown had indeed become polluted through pit latrines, gold mine tailings dumps, and waste disposal sites (landfills). Different pollutants were associated with specific pollution zones. Groundwater from boreholes located within these zones was not suitable for human consumption because it exceeded certain World Health Organization and Botswana Bureau of Standards recommendations for drinking water. The study revealed that groundwater pollution due to nitrates constitutes a real health hazard and environmental and health hazards emanating from abandoned mines jeopardized human safety and environmental protection and was obvious from observed chemical “cocktail” conditions of tailings dumps and trace element concentrations in some boreholes.

1 INTRODUCTION Francistown is the oldest established town in Botswana. Born during the late 19th century as a gold mining town at the confluence of the ephemeral Tati and Ntshe sand rivers,

Effects of mining and urban expansion on groundwater quality in Francistown, Botswana

169

Francistown is the commercial hub in the NE of Botswana. The city’s rapid economic development, in particular since the 1970s, has caused its population to triple over the last three decades to approximately 100,000 inhabitants. Today Francistown is the second largest city in Botswana. In the not too distant past, water demands were entirely met by groundwater locally available from shallow alluvial and fractured volcanic rock aquifers. However, in the 1970s it was found that groundwater produced from the city’s public wells contained elevated concentrations of nitrate. In addition, the available limited groundwater resources could no longer meet the steadily rising demand for water. For these reasons public water supply was shifted in 1982 to surface water from the Shashe dam, which is located at a distance of approximately 30km to the SW of Francistown. The Shashe dam was built during the 1970s to supply the copper-nickel mine in Selebi-Phikwe. 2 GEOLOGY The prevailing dendritic drainage pattern consists of a system of irregularly branching tributaries and forms junctions at various acute angles. This is a manifestation of the complex folded and contorted metamorphosed rocks where lithological variations (in terms of weathering and erosion) are insufficient to modify this pattern. A significant portion of the Francistown study area consists of rocks of the basement complex including meta-volcanics of the so-called Tati schist group. The basement complex is divided into various granitic formations and two non-granitic lithostratigraphic units (Gibb & Partners, 1987). These are subdivided into three formations, the first of which is correlated with the Lady Mary volcanic formation. This formation consists of a homogeneous succession of dark coloured, fine-grained amphibolitic schists. The Lady Mary formation is overlain by the Penhalonga formation, which includes both metasediments and meta-volcanics. The latter are predominantly meta-andesite (greenstone) lavas, tuffs and agglomerates with amphibolite and meta-tuff beds (Key, 1976). The Selkirk formation at the top of the schist relic is laterally more restricted than the other two formations and consists of mainly felsic meta-volcanic extrusives with minor intercalations of meta-sedimentary schists. Gold mineralization in the Francistown area is mainly from quartz reefs and fissure veins of the Tati schist relic. Indeed the Tati schist relic has also been recognized for its base metal potential. Copper and nickel deposits have been identified and are now mined at the Selkirk and Phoenix mines near Matsiloje, 40km further to the SE of Francistown. Copper-zinc anomalies have also been reported near the contact between the Penhalonga and Lady Mary formations as well as in several ironstones in these formations. 3 HYDROGEOLOGY Very little detailed groundwater monitoring of the Francistown aquifers was undertaken since the first abstractions in the early 1950s and since the recommendations made by

Water resources of arid areas

170

consultants in 1974 (Colquhoun et al., 1974) and in 1979 respectively (Gibb & Partners, 1987). Groundwater consultants identified the major aquifer in Francistown as the Penhalonga mixed formation about 1.5km wide extending for at least 7km downstream of the Tati and Ntshe river confluence (Colquhoun et al., 1974). The most productive aquifers were recognized as relatively shallow discontinuous zones of fracturing. These fracture zones have a high transmissivity and draw from storage in the overlying weathered rock and alluvium. They may be up to 4m thick and are usually semi-confined by alluvial sediments and clayey weathered rock. Confining layers composed of sandy horizons contain water and contribute leakage into the underlying aquifer thereby acting as perched aquifers. Weathering appears to be confined to certain horizons within the Penhalonga mixed formation where it appears to be restricted to the easily weathered acid meta-volcanics. Indeed the river Tati is an excellent outward expression of this feature since it also follows the geological strike of this formation within these acid metavolcanics. The river tends to change its course where it traverses more competent members of the Penhalonga mixed formation. Groundwater also occurs in the sandy channels of the rivers Tati and Ntshe and this perennial baseflow component may also be regarded as an aquifer. Upstream of the confluence, the river Tati is 35 to 40m wide with the average thickness of the sand bed being 1.7m. However, sand pockets of up to 3m deep exist and increase the saturated storage of this aquifer. Downstream of this confluence, larger volumes of water can be stored since the river becomes wider with widths ranging from 20 to 100m and deeper sand beds of more than 2m in parts. 4 METHODS AND MATERIALS The study commenced with a census of all existing wells so as to establish their distribution, usage, and availability for sampling. A Garmin 40 hand-held GPS (http://www.garmin.com/) was used for coordinate acquisition. Similarly, all industries and other sites that may have a negative impact on groundwater quality were mapped. Boreholes that were found to be accessible in terms of water level measurement were used together with the topographical elevation to infer groundwater flow directions. An electrical dipper was used for water level measurements and a Trimble high-precision GPS for ground elevation measurements as well as more accurate Cartesian coordinates. The sampling of accessible boreholes involved the use of a Grundfos MP1 submersible pump (http://www.grundfos.com/) equipped with riser pipes of up to 90m. All discharge water generated while pumping was released at least 30m away from the borehole down gradient of the prevailing land slope. The method of sampling was such that electrical conductivity and groundwater reaction were measured continuously until both parameters had stabilized. Once they had stabilized a groundwater sample was taken from the particular borehole. The sample bottles were all made of plastic. Upon sampling, water reaction (pH), electrical conductivity (EC), and dissolved oxygen (DO) were measured using hand-held meters (http://www.wtw.com/). Bicarbonate determined through titration.

and carbon dioxide (CO2) were

Effects of mining and urban expansion on groundwater quality in Francistown, Botswana

171

The data obtained from the various fieldwork exercises and the hydro-chemical laboratory analyses were used to produce several environmental geology maps. For this to materialize, all data were transferred to the ArcView GIS software (Version 3.2) environment (http://www.esri.com/) where the various data layers were put together to produce the thematic maps. Data obtained from the chemical analyses were also used to deduce redox conditions, to delineate redox zones, and to determine the predominant redox processes. 5 RESULTS Out of the total of 202 boreholes that were identified during the well census, only 48 could be sampled for groundwater. All the others were inaccessible because of collapse, vandalism, or else, they had fallen dry. The vast majority of the accessible boreholes were concentrated along the two rivers Ntshe and Tati. However, groundwater yields were generally low. Several of the few known borehole yields were below 2m3/h, hence their proximity to the rivers. Only very few such as the monitoring boreholes at the abandoned and the new waste disposal site were beyond the rivers. The chemical analyses revealed that there was not much variation in groundwater reaction (pH). Most of the samples had neutral pH levels around 7, which is normal for groundwater. No groundwater sample showed acid conditions. Magnesium (Mg2+), calcium (Ca2+), and bicarbonate were the most important ions. Hence, Mg-Ca-HCO3 type of water was dominant. In some places, NaMg-Ca-HCO3 type of waters were prevalent that also featured elevated concentrations of , chlorine (Cl¯), and sulfate The concentration of total dissolved nitrate solids (TDS) was less than 1000mg/L in these particular boreholes. Over most of the built-up city area the groundwater was strongly influenced by anthropogenic activities. This was evident from TDS levels greater than 1000mg/L, and and constituted the dominant anions. In order to identify and delineate Cl¯, distinct groundwater pollution zones, all chemical groundwater parameters were used as environmental indicators and mapped individually. The concentration of oxygen allowed to identify zones with different aeration status, namely zones with aerobic (oxic) and those with anaerobic (probably reduced) groundwater conditions. This was the starting point towards defining likely pollution zones and also towards predicting redox states. In order to allow for a sound investigation it was necessary to examine the main indicator species for redox state, namely sulphate

ferrous iron (Fe2+),

nitrite and ammonium . manganous manganese (Mn2+), nitrate As was to be expected, a comparison between these species revealed that areas with high and . Equally, levels of Fe2+ and Mn2+ had at the same time low levels of areas rich in sulphates and nitrates coincided with zones high in dissolved oxygen (O2), indicating oxidizing (aerobic) conditions, while zones high in ferrous iron and

Water resources of arid areas

172

manganous manganese overlapped with zones very low in dissolved oxygen, thus indicating reduced (anaerobic) environments. The change from one zone to another was gradual. Important information in order to identify buffering systems was the presence of carbon dioxide (CO2) and bicarbonate . The distribution of these two species did not show a significant relationship to the redox state of the water. It rather was related to the calcium (Ca) and magnesium (Mg) distribution. The next parameter under consideration was the distribution of chloride (Cl¯). Due to its significant mobility, Cl¯ was meant to point to possible pollution sources. Yet, only few zones of high concentration could be identified. The next step was to seek out possible pollutants, that is heavy metals and other trace elements. All heavy metals that were detected in the study area showed distributions quite different from each other and were possibly related to mine waste sites. Zinc (Zn2+) however was not connected to mine dumps only; very strong concentrations of reduced Zn were much wider spread. The spatial distribution of the different pollutants revealed that they formed zones. Thus, once the areas with oxic (aerobic) and anoxic (anaerobic) groundwater environments and the spatial distribution of pollutants were identified, the study area could be divided into different pollution zones (Fig. 1):

Effects of mining and urban expansion on groundwater quality in Francistown, Botswana

173

Figure 1. Groundwater pollution zones in Francistown in the year 2000. Zone I—Oxidizing conditions, lots of dissolved oxygen, lots of nitrate, sulphate present, very low concentrations of ferrous iron and manganous manganese; mostly organic pollution. Zone Ia—A local disturbance within Zone I showing presence of trace elements. A small mine dump was next to this site. Zone II—High concentrations of ferrous iron and manganous manganese, enhanced zinc, no sulphate, no nitrate, ammonia present indicating very reduced conditions possibly due to organic pollution but also danger from precipitation of sulphides of different metals.

Water resources of arid areas

174

Zone III—Typical mining waste point. Elevated heavy metal and trace element concentrations, very high arsenic concentration, but also high nitrate and sulphate levels. Ferrous iron and manganous manganese strongly decreased. Zone IV—Reduced zone with high Zn2+ and Cr concentrations; Fe2+ and Mn2+ enhanced. Some trace elements present. Zone IVa—Disturbance within Zone IV with high oxygen concentration and some elevated heavy metals and trace elements; Cl¯ also highly concentrated. Zone V—Oxidised, very high sulphate and chloride concentrations; no nitrate; trace elements in significant concentrations; some heavy metals present; controlled landfill. Zone VI—Reduced, but no Fe2+ and Mn2+; Cl¯ highly concentrated; Mg2+ and Ca2+ ; some trace elements present; old and abandoned enhanced as well as CO2 and landfill. Zone VII—Oxidising conditions; no organic pollution but very diverse trace elements present (some of them concentrated); no potential source of pollution could be identified from the groundwater hazards map. 6 DISCUSSION The results of this study revealed that groundwater in Francistown had become polluted through three major sources, namely pit latrines (Zones I and II), mine tailings dumps (Zones Ia, III, IV, IVa, and VII), and waste disposal sites (Zones V and VI). 6.1 Pit latrines—Zones I and II Chemical analyses showed that nitrate concentrations (Fig. 2) well above the Botswana drinking water standard of 45mg/L (BOS, 2000) were frequent within the city area, that is zones I and II. The areas where the boreholes revealed elevated nitrate levels matched well with the areas where pit latrines where to be found. Pit latrines were located all along the river Tati throughout the built-up area. They constitute a constant source of organic pollution in the form of human excrements. This problem is made worse by the fact that pit latrines are also being used to discharge household wastewater. It is likely that a considerable amount of pollution may have been transferred into zone I from the reduced zones IV and VII upstream to the north. Zone I was characterised by a high concentration of dissolved oxygen. Nitrogen originating upstream as well as from zone I itself was probably oxidized to nitrate, which showed an extremely high concentration in the centre of the zone. Downstream the concentration of dissolved oxygen decreased. At the same time the nitrite concentration increased, probably due to denitrification. The high nitrate concentrations gradually decreased towards zone II. In contrast, the concentrations of Fe2+ and Mn2+ were very low in the centre of zone I and gradually increased towards the reducing zone II. This was typical of a redox state controlled by bacterial activity. In zone I there was a lot of organic matter input, which may have been used by bacteria as a source of carbon for the oxidation of Fe2+ (cf. Christensen et al., 1995). In the southern part of Francistown, the river Tati flows in south-easterly direction and then bends back towards the west in the middle of zone II (cf. Fig. 1). Between the two

Effects of mining and urban expansion on groundwater quality in Francistown, Botswana

175

bends ephemeral river flows are slowed down and the observed depth to the groundwater was shallower here than elsewhere in Francistown. In the crest of the second river bend there were big alluvial

Figure 2. Nitrate concentrations (mg/L) in Francistown in July/August 2000. deposits, which probably resulted in the accumulation of organic pollution and reduced groundwater conditions. At this point was hardly measurable but showed strong concentration. Fe2+ and Mn2+ were also strongly concentrated along with Zn2+. This was indicative of strong anaerobic bacterial reduction processes. On the edge of zone II towards zone III there was an old sewage pond. This location could be picked up in the form of a prolonged reduced zone characterized by lower and very high Mn2+ levels. Surprisingly though, the concentration of Fe2+ was low. From this it appeared that Mn reduction was somehow favoured over Fe reduction, which may have been controlled by the redox state of the pond (Mn needs less energy for oxidation than Fe). 6.2 Mine waste dumps—Zones Ia, III, IV, IVa, and VII Several groundwater zones were indicative of pollution due to historic gold mining activities. The strongest evidence came from the wider surroundings of the Lady Mary mine, which is located in the SE corner of the study area (zone III). Two boreholes located close to this abandoned mine site (strongly) violated international and Botswana drinking water arsenic standards, which allow for a maximum of 10 ppb (µg L−1). Yet, the groundwater in the two boreholes featured levels of 26 and 244µg (ppb) As L−1

Water resources of arid areas

176

respectively. Arsenic is very problematic in the environment because of its relative mobility over a wide range of redox conditions (Smedley & Kinniburgh, 2001). Zinc (Zn), copper (Cu), cadmium (Cd) and nickel (Ni) were also present in very high concentrations. In addition, other compounds such as cobalt, titanium, scandium, antimony, mercury, tellurium, rubidium, and thallium also showed elevated concentrations in zone III. Since zone III is situated in the most downstream spot of the study area, it is likely that all sorts of organic groundwater pollutants and products of anaerobic processes originating from zones I and II were also transferred into this area. Because the concentration of dissolved oxygen was rather high, the redox processes obviously went towards oxidation. The nitrite and nitrate levels were also rather high, which was indicative of active oxidation of ammonia that must have originated from reduced zone II. Sulphate was also very high and possibly originated from the oxidation of FeS or MnS. At the same time the concentrations of Fe2+ and Mn2+ were strongly decreased, which was probably the result of oxidation and the formation of insoluble Fe3+ or Mn4+ compounds. All this suggested very strong bacterial processes. In addition, the Cl distribution in this zone was also indicative of a site characterized by pollution input. All in all, the broad range of organic and highly toxic inorganic pollutants in zone III calls for urgent attention. Another menacing mine site is Monarch (Vogel & Kasper, 2002), which is located north of the confluence of the rivers Tati and Ntshe. Surprisingly though, zone IV did not indicate elevated levels of heavy metals or other trace elements. The only irregularities compared to the surroundings were very high O2 and elevated cobalt and silver concentrations but low nitrate and very low Fe2+ and Mn2+ levels. Possibly this was due to a combination of factors such as limited rainfall in this semi-arid environment, the fine grain-size distribution of the tailings material, and the huge size of the Monarch tailings, which may not easily provide for acid mine drainage (leaching). Rather most of the pollutants may remain in the oxidised crystal form. In contrast, a couple of areas (Ia, IV, IVa, VII), which at first had not appeared conspicuous, revealed strange irregularities in their groundwater composition. Zone IV was very reduced with a medium concentration but high Fe2+ levels. Surprisingly, it also showed a high zinc and a very high chromium concentration. Thallium, rubidium, tellurium, and cadmium were also present in increased concentrations. The distribution of Cl¯ indicated a strong pollution input upstream from this zone. The data obtained from this zone suggested that somewhere there must have been a very strong but unrecognized source of pollution, or else, the natural geological environment may have caused the formation of reduced groundwater conditions and the release of metal ions into the water. The latter was however unlikely given the granitic nature of the resident rock. The situation was similar in zone IVa. A low dissolved O2 level and therefore a low concentration, increased Fe2+ and Mn2+ but also increased arsenic, copper, selenium, beryllium, tin, caesium, yttrium and tungsten concentrations. Data from this site also revealed strong inorganic pollution even though no obvious inorganic waste source was detected. Zone VII was located north of zone IVa. Again, data showed enhanced concentrations of heavy metals but not of the elements identified in zone IVa. Because no pollution

Effects of mining and urban expansion on groundwater quality in Francistown, Botswana

177

source could be detected in these two zones it is suggested that remnants of old mine deposits may still exist in these two areas. A small mine dump within zone I (cf. Lehmann, 2001) caused raised concentrations of zircon, tantalum, hafnium, cerium, niobium, bismuth scandium and titanium and it was therefore separated out as mine waste zone Ia. Increased concentrations of Cl¯ and clearly pointed to anthropogenic pollution. The oxygen and nitrate concentrations at this site were strongly reduced but nitrite was increased. This indicated a change in bacterial populations from nitrifying to denitrifying bacteria. Given the obvious similarities in groundwater pollution between the above sites, they were put in the same pollution risk group. They may be even more hazardous than zone III since they are situated upstream from the built-up areas. Clearly, more investigations need to be carried out and immediate attention must be given. 6.3 Waste disposal sites—Zones V and VI Waste disposal sites pose an environmental hazard if they give rise to the formation of leachate plumes. The two most important factors governing the biogeochemical processes within a leachate plume are (1) the redox state, and (2) the content of the leachate. Determining the redox state of polluted groundwater is not easy. It is based on the identification of redox-sensitive species. The primary redox-sensitive species in groundwater are the dissolved ions of Fe2+, Mn2+, , , , , HS¯, the dissolved gasses CH4, N2O and O2, and also some organic substances (Christensen et al., 2001). Most of these processes are driven by bacteria and therefore slow. Bacterial populations are differentiated according to the presence (aerobs) or absence (anaerobs) of oxygen. Hence, a crucial step for this part of the study was to determine the presence of dissolved oxygen (O2) in the groundwater samples. It was obvious that the two waste disposal sites in Francistown were quite different in terms of aeration. The old landfill site (zone VI) was very poor in dissolved oxygen (O2). It is assumed that the long-lasting deposition of waste had formed the reduced environment and that anaerobic processes had probably taken place. In contrast, the new landfill site (zone V) had not yet developed a reduced zone of influence. There the concentration of O2 was quite high. In both zones, Fe2+ and Mn2+ were only present in very low concentrations. Similarly, was also only present in a very low concentration, and and were probably absent. Since no significant increase in Fe2+ and Mn2+ levels and no decrease in could be observed, and given the fact that there was only very little groundwater in both areas (in fact, during pumping one of the sampled boreholes dried up), it is assumed that the geochemistry and the redox states were not governed biologically. Bacteria need water in order to thrive. The very enhanced concentration of sulphate in zone V was probably the result of the presence of oxygenated water and the deposition of ash and building material at this site. Spreading out in a radial manner, sulphate looked like a serious problem. Very

Water resources of arid areas

178

similar pictures emanated from the spatial concentrations of rubidium, thallium, silver, uranium, molybdenum, lanthanum, zircon, titanium, sodium, bromide, and boron. may have caused The observed slightly enhanced concentrations of CO2 and the dissolution of Ca2+ and Mg2+ out of the carbonates. Probably as a result of this, the concentrations of these two cations were slightly raised (cf. Christensen et al., 2001). This could have influenced the buffering system of the sediments. Both waste disposal sites also featured high Cl¯ concentrations, though the concentration was much wider at the new (zone V) as compared to the old landfill site (zone VI). This supported the assumption that there was no new input of pollution at the old landfill site. So far, the new landfill is only used to deposit inorganic waste. Once it will be used for other kinds of waste, different processes may set in. Considering the semi-arid environment in Francistown it may be assumed that pollution at both landfill sites is localized, will not move readily from place to place, and is probably confined to the soil only. From this it would follow that the two landfills had no significant adverse effect on groundwater quality in the study area. On the other hand, natural remediation in the form of transporting pollutants to other places or through bacterial degradation is also not likely to take place. Thus pollution would probably stay as a hazard for a long time. 7 CONCLUSIONS The study highlighted that groundwater quality in Francistown had deteriorated drastically due to the influence of urban expansion and historical mining. The three dominant sources of pollution were identified as pit latrines, mine waste dumps, and waste disposal sites (landfills). However, pollution from these sources was spatially confined to those zones within which pit latrines, mine waste dumps, and landfills were located. Groundwater from boreholes located within these zones was not suitable for human consumption because it exceeded certain World Health Organization (WHO, 1998) and Botswana Bureau of Standards (BOS, 2000) recommendations for drinking water. Amongst the three pollution sources, pit latrines were found to have had the worst impact on groundwater quality. The chemical analyses of groundwater samples from a total of 48 public and private wells sampled within and around Francistown showed that nitrate concentrations were frequently well above the maximum allowable level of nitrate in drinking water. Groundwater sampled from boreholes situated in remote areas outside the city featured considerably less nitrate. In most cases the nitrate levels in remote areas outside the city were below 40mg/L, which supported the assumption that the cause of nitrate contamination was anthropogenic. Finally, the addition of nitrate through faecal waste had in turn triggered complex redox processes that had raised the ferrous iron (Fe2+) and sulphate concentrations of the groundwater. Mine dumps and/or tailings also contributed to the deterioration of groundwater quality through the addition of heavy metals, and by raising the sulphate concentration in certain zones. However, since the vast majority of the sampled boreholes were located along the rivers Tati and Ntshe and thus far away from the tailings, the real groundwater

Effects of mining and urban expansion on groundwater quality in Francistown, Botswana

179

hazards emanating from the tailings may have gone unnoticed. Clearly, further investigations are required. Amongst the three major pollutants, landfills had the least impact on groundwater quality. They are sited away from the main aquifer and within rock formations that yield little groundwater. Because of the limited rainfall in the study area, pollutants within these zones are likely to stay contained within the area. Only occasionally will they be flushed out during the rainy season and become diluted. 8 RECOMMENDATIONS Groundwater from a substantial number of boreholes was found to be not suitable for human consumption. It is therefore necessary to determine which boreholes are used for humans so as to discontinue their use. As a rule, the Francistown city council ought to adopt a development strategy that places more emphasis on an environmental approach to planning taking into account the existing water resources. For example, all new infrastructures should be placed as far away as possible from the rivers because the aquifers in the area are dependent on rainfall and river recharge. Activities such as the recent aligning of the sewage pipelines along the riverbanks must in future be avoided by all means. Such activities not only destroy a natural flood barrier but they may in fact lead to serious water pollution. Similarly, any new development must not include pit latrines. Since a sewage reticulation system has been put in place throughout the city, it is necessary to educate the residents on the need to connect to the sewerage and put an end to the use of pit latrines. So far, connection to the sewerage system is on a voluntary basis and pit latrines (and septic tanks) are currently still the main means of wastewater discharge in the newly connected areas. The study also confirmed that environmental and health hazards emanating from abandoned mine tailings must be dealt with in a way that guarantees human safety and environmental protection. The reported chemical “cocktail” conditions of tailings and the observed trace element concentrations in some boreholes make this obvious. The waste disposal (landfill) sites appeared to have been well sited in areas of low groundwater yields. But continuous monitoring is necessary in order to determine the dynamics of possible plume development so as to act upon possible groundwater pollution. Further investigations are also necessary to determine the source of heavy metals and other pollutants at the new landfill site. REFERENCES Colquhoun, B., O’Donnel, H. & Partners 1974. Redevelopment of the Francistown groundwater studies report. Phases I, II and III. Australian Groundwater Consultants. BOS 2000. Water quality—Drinking water—Specification. BOS 32, Botswana Bureau of Standards, Gaborone, Botswana. Christensen, T.H., Kjelsden, P., Bjerk, P.L., Jensen, D.L., Christensen, J.B., Baun, A. Albrechtsen, H.J. & Heron, G. 2001. Biogeochemistry of landfill leachate plumes. Applied Geochemistry: 659–718.

Water resources of arid areas

180

Gibb, A. Sir & Partners 1987. Francistown Water Development. Pre-Investment Study. Appendices B1 and B2. Water Resources. Water Utilities Corporation, Botswana. Key, R. 1976. The geology of the area around Francistown and Phikwe, Northeast and Central Districts, Botswana. District Memoir 3, 121p. plus maps, Dept. Geological Survey (DGS), Lobatse, Botswana. Lehmann, A. 2001. Conceptual map of the urban soils of Francistown. Draft map and explanations with special reference to town planning and environmental quality. Report by the Environmental Geology Division, Dept. of Geological Survey (DGS), 48p, Lobatse, Botswana. Smedley, P.K. & Kinniburgh, D.G. (2001). Source and behaviour of arsenic in natural waters. In: United Nations Synthesis Report on Arsenic in Drinking Water. Vogel, H. & Kasper, B. 2002. Mine soils on abandoned gold mine tailings in Francistown. Report by the Environmental Geology Division, Dept. of Geological Survey (DGS), 43p., Lobatse, Botswana. WHO (1998). Guidelines for drinking water quality. World Health Organization, 2nd ed., Volumes 1 and 2, Geneva, Switzerland.

In situ remediation potential for Southern African groundwater resources Sumaya Clarke, Gideon Tredoux & Pannie Engelbrecht Water Programme, Environmentek, CSIR, Stellenbosh Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: In situ groundwater remediation is practised in Europe, the United States, New Zealand and Canada. Widely accepted treatment methods include, permeable reactive barriers, redox manipulation, the Vyredox® method and biological denitrification. The permeable reactive barrier is widely used for contaminant removal. It consists of a constructed trench filled with a contaminant specific reagent, such as sawdust or wood chips for the promotion of biological denitrification. Other techniques include in situ redox manipulation which requires dithionite injection and in situ biological denitrification, which uses substrates such as ethanol or methanol. The Nitredox® and Vyredox® methods use a sophisticated arrangement of injection and aeration boreholes to manipulate oxidation and reduction to control nitrate, iron and manganese concentrations in the aquifer. Concern about nitrate as a chemical constituent of groundwater is increasing, especially in the arid and semi-arid regions of Southern Africa. Livestock losses, and “blue baby syndrome” in humans, result from high nitrate concentrations in drinking water. Hence, there is a need to remediate groundwater with nitrate concentrations above the required standard. Low cost, robust and simple treatment technologies are needed for rural water supply in Southern Africa. This paper gives an overview of the performance of full scale and pilot scale treatment plants. An estimate was made of the costs of applying selected in situ treatment options for a South African town. The order of difference in cost between in situ and ex situ treatment plants is calculated. The geological and hydrogeological parameters required for successful operation of most in situ treatment systems are described. Advantages and disadvantages of in situ treatment are also mentioned.

Water resources of arid areas

182

1 INTRODUCTION Research into groundwater remediation methods has been intensified and various techniques have been tested and applied. “Pump and treat” technologies have been widely used in the USA. The success of this method has been questioned, considering its excessive costs, (Simon et al., 2001). As an alternative, in situ technologies are being developed and applied for removing contaminants in the aquifer. Literature references to more than 100 successfully operating sites confirm that permeable reactive barriers can remove a variety of contaminants including halogenated organic compounds, metals, nitrates, acid mine drainage, phosphorous, chromium and gasoline/petrol derivatives using various reactive materials in the barrier. The currently operating sites range from household scale permeable reactive barriers to industrial sites, to mining and wastewater treatment plants and municipal well fields (Robertson and Cherry, 2003). Slowly degradable carbon sources are placed in barriers perpendicular to the flow and such treatment occurs with a high success rate. The Nitredox® plant in Vienna, Austria to treat nitrate, iron and manganese; has been operated successfully for more than a decade. Various low cost, robust treatment techniques like permeable reactive barriers and biological denitrification have proven to be successful in Canada, New Zealand, Austria, France, and the USA. Cost implication of implementing any in situ technology is important and need to be taken

Figure 1. Map showing the distribution of nitrate (as NO3) in Southern Africa.

In situ remediation potential for Southern African groundwater resources

183

into consideration. A cost estimation, performed for a town in the Northern Cape Province, South Africa demonstrates the cost difference between in situ and ex situ treatment methods. Internationally a nitrate concentration of 10mg/L as N (45mg/L as NO3) is accepted as guidelines for health risk. The maximum allowable level is set as 20mg/L as N (90mg/L as NO3) in South Africa. Nitrate concentrations in groundwater are alarmingly high in some parts of Southern Africa as shown in Figure 1. The northern provinces of South Africa all have many groundwater sources with nitrate concentrations ranging from 251– 500mg/L. In some areas concentrations up to 1000mg/L occur, while the Southern Kalahari has concentrations of up to 2000mg/L, particularly in the more saline areas, (Marais, 1999, Tredoux et al., 2000). In view of the prevalence of nitrate in groundwater, this paper focuses on in situ denitrification as a viable treatment option for town and rural applications. It is crucial that groundwater pollution be taken seriously, and that remediation and protection of the groundwater resources available be considered as a priority in countries affected by pollution. Surface water resources are limited, more particularly in arid and semi arid regions, and with groundwater being unfit for use by inhabitants of these regions, a serious threat is posed to the survival and growth of communities affected. 2 AVAILABLE DENITRIFICATION TECHNOLOGIES Many methods are successfully being employed to denitrify groundwater. Permeable Reactive Barriers and Biological Denitrification methods are the most widely used of the many methods identified, hence these methods will be discussed in further detail. 3 PERMEABLE REACTIVE BARRIERS Permeable reactive barriers (PRB) are constructed across the flow path of the migrating plume of contaminated groundwater. These systems are typically designed as a continuous trench, filled

Water resources of arid areas

184

Figure 2. Permeable reactive barrier in situ denitrification system, shown here on a rural/local scale. with permeable, reactive material. Alternatively, a funnel and gate configuration is used, which includes impermeable sections, directing the groundwater flow through the permeable treatment “gates” (Robertson and Cherry, 1995, 2000, Blowes et al., 2000, Schipper & Vojvodic-Vukovic, 2000). These treatment systems may be applied for the removal of various anions, cations, organic compounds and inorganic compounds. Configurations and system design is generally site and contaminant specific, e.g. for mitigation of nitrate at on site sanitation (see Fig 2). Requirements for the denitrification barriers (“walls”) include the following: ● The site should have a shallow water table; ● Aquifer parameters should be well understood; ● The aquifer thickness and composition should allow for constructing the wall i.e. not more than 10m deep. ● Boreholes should be placed on either side of the wall to sample groundwater to monitor chemical and microbiological changes. ● Analysis of groundwater and soil should be done prior to installation of the PRB to estimate the amount of carbon substrate required.

4 BIOLOGICAL DENITRIFICATION Biologically enhanced denitrification requires injection of readily available carbon substrates such as ethanol, methanol, sucrose and glucose to serve as a source of energy for promoting microbiological activity. Various configurations of the method may be

In situ remediation potential for Southern African groundwater resources

185

used to suit site-specific requirements. The most successful configurations are those used in Vienna and Nebraska. A list of operational systems is presented in Table 1. The daisy configuration is shown in Figure 3. Most of the biological denitrification systems use variations of this basic configuration. The area in the sketch labelled (IV) represents nitrate polluted water. The “daisy” represents the area (in plan view) progressively affected by the denitrification due to substrate (carbon source) injection. The reaction takes place in zone I, followed by filtration of any by-products in zone II while the nitrate free water (or water with a lowered nitrate concentration) is found in zone III. The Nitredox® system consists of one pumping borehole located at the centre of two concentric circles of injection boreholes. It involves injection of an organic substrate (outer ring) to enhance denitrification, but includes an additional phase of aerated water injection for the oxidation and removal of iron (inner ring) once the nitrogen is removed. The groundwater recovered from the

Table 1. Some pilot and field operational denitrification sites and their experiences. Method and Period NO3location Nmg/L

Aquifer

Carbon substrate

Injection/ barrier

Nitrate removed

PRB, Canada 5 yrs+ 5–57 Primary Sawdust/ Emplaced 58–91% (1) woodchips barrier PRB, New 5 yrs+ 5–15 Unconfined, Sawdust Emplaced 95+% Zealand (2) sandy barrier Electrokinetics/ Test Controlled Primary/ None: Abiotic Emplacement 84–87% Fe-wall, period amounts secondary of wall and USA(3) electrodes NitrEI system, Many Up to Primary/ None: Electrodes Reduced Canada (4) Currently 1000 unsaturated Electrochemical down to operating zone electrodes 0.1mgN/L sites Daisy wheel, 40 Sand and Ethanol C and P 35%-cISBD, gravel injection injection; Nebraska (5) 90–100% pinjection Nitredox, 15 years 14 Primary Ethanol P injection 75% ISBD, Vienna aquifer (6) ISBD, line of 226–565 Chalk Ethanol 80% injection boreholes, France (7) C-Continuous injection, P-pulse injection. (1) Robertson and Cherry (1995, 2000), Blowes et al., (1999), (2) Schipper & Vojvodic-Vukovic, 2001, (3) Chew and Zhang, 1998 and Loo, 2000, (4), (5) Khan & Spalding (1998), (6) Braester and Martinell (1988), Jechlinger et al., (1991), (7) Chevron et al., (1998).

Water resources of arid areas

186

Figure 3. Biological denitrification treatment system with “daisy” configuration (after Mercado, 1988). Table 2. Cost estimation of implementation of in situ denitrification compared to ex situ treatment (amounts in S A Rand). Method

PRB ISBD Ex situ Method

Capital investment Operation and maintenance per m3 Projected running expenses over 5 yrs Projected total cost over 5 yrs

61332 100289 350000 0.1 0.3 2 71144 21343 2845740 132476 313720 3195740

central production borehole is partly free of nitrate but completely free of iron, manganese or other by-products (Braester and Martinell, 1988). This method has been applied to coastal aquifers and primary aquifers. Where biological denitrification is implemented, it is important to know and monitor the permeability and porosity. Clogging may result when the carbon substrate injection exceeds the amount required for denitrification. The method has been applied mainly to primary aquifers where flow dynamics are well understood.

In situ remediation potential for Southern African groundwater resources

187

5 OPERATIONAL SITES In Southern Africa, denitrification per se is not applied. Treatment methods use expensive ex situ pump and treat systems such as desalinisation by ion exchange. These do not specifically treat nitrate and does not obtain optimum results with respect to nitrate concentrations. There is a number of test and full-scale in situ denitrification plants all over the world. Table 1 lists some of these sites, their experiences and shows the variety of configurations, carbon sources and aquifer types to which in situ denitrification has been applied. These methods are mainly applied in sand; gravel and other primary aquifer type settings, although it has been used in chalk aquifers as well. Implementation in secondary aquifer settings is said to be possible and has been modelled for sites in the UK (Cartmell et al., 1999). 6 OPERATIONAL EXPERIENCES The PRB systems in Canada have been operational for more than 5 years and are used on various scales including household, municipal, and huge water treatment plants. In New Zealand, the reactive barrier had to be replaced after 5 years of operation as reactive material had clogged parts of the aquifer. In the USA, the electrokinetic methods worked better when combined with iron walls. The biological denitrification used in Nebraska used both continuous and pulse injection regimes. The continuous carbon source injection was more efficient in denitrification but led to complete biofouling after 10 days. The system used inner oxidation ring to remove possible nitrite, iron and manganese. In Vienna, where the Nitredox® method is currently operational, clogging was experienced. Pulse injection and reduction of the amount of ethanol (carbon source) prevented clogging of the system. In France, natural in situ denitrification was carbon limited. Remediation by carbon source addition was selected to accelerate denitrification. Denitrification was achieved in long time operation (450 days). Rates of denitrification were improved when trace metals were supplied in conjunction with the carbon substrate. 7 ECONOMIC FEASIBILITY OF TREATMENT TECHNOLOGIES The costs of implementing denitrification were estimated for Marydale in the Northern Cape Province of South Africa. The results are shown in Table 2. In situ application was compared with a conventional ex situ method. In Marydale, a well field containing 10 boreholes is used as the town water supply. Half of these boreholes, produce groundwater of nitrate concentration above the maximum allowable 20mg/L, (Hofmann, 1997). Microbiological sampling showed that coliform counts of 15/100ml in some boreholes were three times as much as the SABS specification (Hofmann, 1997). No faecal

Water resources of arid areas

188

coliforms were detected. This gives an indication that no human or animal waste reaches the boreholes. Exploration boreholes drilled in the area revealed that a primary and a secondary aquifer are present in the area. The secondary fracture system is not well understood, but it is believed that the bed rock is not very permeable. The main water bearing unit is the alluvial cover of more or less 12m thick. The aquifer material consists of sedimentary layers containing primarily sand and silt. The Projected water demand of the town for 2005 is 142287m3/a (Shand and VSA, 1997). Capital expenses for permeable reactive barriers (PRB) include excavation costs, wall emplacement costs and dewatering prior to wall emplacement. Woodchips or sawdust was considered as suitable permeable reactive barrier material as these are cheap and slowly degradable carbon sources. The barrier size is based on the size of the well field and the depth to bedrock. The largest contribution to capital costs for in situ biological denitrification (ISBD) methods include borehole construction costs, purchasing of injections pumps among other costs. Erecting infrastructure is a major expense for conventional treatment plants. Running costs were based on the projected annual water demand and estimated chemical costs. The PRB method requires limited maintenance. In the calculation, operation and running costs are included; however, they may not occur frequently for methods like PRB. Operational and maintenance costs are relevant especially when clogging or partial clogging of wells occurs. Pump and treat methods and other ex situ methods (in this case ion-exchange) generally cost an order of magnitude more than in situ methods. It is clear from this information that rural communities for which funding is not always in surplus may capitalize on this advantage as well as the ease of use of some of these methods. Proper management and monitoring of sites are essential to detect potential clogging cases early and to put remedial measures in place. 8 DISCUSSION Field scale plants have proven in situ technologies to be successful. The nitrate removal rate at most currently operating sites are high. It is evident from Table 2 that the permeable reactive barrier method is the most cost effective method. Capital costs are relatively low and it requires little or no additional treatment of groundwater after passing through the system. Installation and running costs of ex situ treatment exceed that of in situ methods. Operational sites in the US, Canada and New Zealand showed that barrier material replacement was only required after 5 years, while ex situ methods have set running expenses per cubic meter of water. The largest full-scale in situ denitrification plant uses the Nitredox® principle. This plant is located at Bisamberg, Vienna (Austria) and has been operating successfully for more than a decade (Jechlinger et al., 1991). It uses ethanol as the carbon substrate and the process is regulated to ensure that the raw water nitrate, which exceeds 65mg/L, is reduced to approximately 35mg/L in the product water. There are advantages as well as some disadvantages of in situ treatment technologies. Some advantages of implementing such a treatment system include minimal exposure to

In situ remediation potential for Southern African groundwater resources

189

dangerous chemicals, job creation in rural villages, little or no treatment required at the surface, possible treatment of other contaminants due to redox changes, costs savings in comparison to conventional ex situ treatment plants, low maintenance costs, simple to use technologies, and no need for electricity (PRB- method). Disadvantages include possibilities of clogging of boreholes. This occurs when the carbon dosage is in excess of the required amount. Sulphate reduction may occur when carbon substrate dosage is too high and result in acetate production as a by-product of microbial activity (Israel, 2004, unpublished data). Loss of hydraulic permeability of the aquifer is possible if carbon addition is not effectively managed. Preferential flow of groundwater can occur, where a great contrast develops between the treatment zone and the rest of the aquifer and the path of least resistance is taken by the groundwater. Hence monitoring of the above mentioned parameters is very important. Cautionary measures include proper estimation of the required amount of carbon substrate, and monitoring the effective porosity and permeability of the aquifer before, after and during treatment. Management of implementation and monitoring are essential for success. It is important to note that no microbes are added to initiate the process, as this would affect the ecosystems that are already established at any specific site. There are many strains of bacteria that occur naturally under the various environmental conditions, which are capable of denitrification. Although some scientists may prefer to add appropriate bacteria to initiate the process but the addition of a carbon source is sufficient to activate resident bacteria. 9 CONCLUSION In situ groundwater treatment methods are widely used and accepted in the US, Canada, Europe and New Zealand. Literature shows that various in situ methods for a range of heavy metals, organic compounds and other constituents have been successfully implemented at field scale in these countries. In situ denitrification methods are also viable treatment methods which are successfully implemented and currently operating. The cost analysis performed in South Africa, showed that there is an order of magnitude difference between the costs of ex situ and in situ treatment plants. Optimal conditions for most in situ treatment methods include the following: ● Primary aquifer systems, or well understood secondary aquifers (with respect to flow characteristics and porosity/permeability). ● Aquifer material can include sand, gravel, and chalk material. ● A known concentration of nitrate-nitrogen is important for estimation of the appropriate quantity of carbon substrate. ● A maximum aquifer thickness of 20m for injection type methods (e.g. ISBD) and 10m for emplacement methods (e.g. PRB). ● Monitoring of aquifer parameters (permeability, hydraulic conductivity etc.), chemical changes (pH, Eh, etc.) and microbiological changes with time. It is important to note that no foreign microbes are added, as this would affect the ecosystems that are already established at any specific site. Although some scientists may

Water resources of arid areas

190

prefer to add appropriate bacteria, addition of a carbon source is sufficient to activate resident bacteria. REFERENCES Blowes, D.W., Ptacek, C.J., Benner, S.G., McRae, C.W.T., Bennett, T.A. & Puls, R.W. 2000. Treatment of inorganic contaminants using permeable reactive barriers. Jnl. of Contaminant Hydrology, 45:123–137. Braester, C. & Martinell, R. 1988. The Vyredox and Nitredox method in situ treatment of groundwater, Wat. Sci. Tech., 20(3):149–163. Cartmell, E., Clark, L., Oakes, D., Smith, S. & Tomkins, J. 1999. Feasibility of In situ Bioremedition of Nitrate in Aquifer systems, R & D Technical Report P277, WRC report no. EA 4683. Chevron, F, Lecomte, P., Darmendrail, D. & Charbonnier, P. 1998. Rehabilitation de qualitè physicochimique d’un aquifere contaminepar des nitrates d’origine industrielle- un example en region Nord-Pas de Calais. L’Eau, L’Industrie, Les Nuisances, 208(31–35) (In French). Chew, C.F. & Zhang, T.C. 1998. In situ remediation of nitrate contaminated ground water by electrokinetics/ iron wall process. Water Science and Technology, 38(7):135–142. EPA, 1995. In situ remediation technology status report: Treatment walls. Report No. EPA/540/K94/004. Office of Solid Waste and Emergency Response, US Environmental Protection Agency. Israel, S., 2004, Subsurface Manipulation of the Nitrogen Cycle: In-Situ denitrification and its potential for remediation of contaminated soil and ground water resources: Case Study: Marydale, Northern Cape, MSc research, unpublished data, University of Stellenbosch. Jechlinger, G., Schöller, F., Seidelberger, F., & Zibuschka, F. 1991. Denitrification In Situ. In: Proc. of I.W.S.A workshop: Inorganic nitrogen compounds and water supply. Hamburg, 27–29 Nov:113–122. Khan, I.A., & Spalding, R.F. 1998. Denitrification using a daisy well system. Presentation to National Sanitation Foundation International Symposium, Safe Drinking Water in Small Systems: Technology, Operations, and Econimics. Washington D.C., May 10–13. Kruithof, J.C., Van Paasen, J.A.M., Hijnen, W.A.M., Dierx, H.A.L. & Van Bennekom, C.A. 1985. Experiences with nitrate removal in the eastern Netherlands. Proc. Nitrates Dans les Eaux, Paris 22–24 October. Loo, W.W. 2000. Electrokinetic treatment of hazardous wastes. Standard Encyclopedia of Environmental Science and Technology, New York, McGraw Hill,: 14.69–14.84. Mercado, A., Libhaber, M. & Soares, M.I.M. 1988. In situ biological groundwater. denitrification: Concepts and preliminary field test. Wat. Sci. Tech., 20(3):197–209. Ninham Shand and VSA Consulting, 1997, Geohidrologiese Ondersoek van die Groundwaterbronne by Marydale, Noord-Kaap Provinsie, VSA Conculting pty. Ltd. Robertson, W.D. & Cherry, J.A. 1995. In situ denitrification of septic system nitrate using reactive porous media barriers: Field trials. Ground Water, 33(1):99–111. Robertson, W.D., Ford, G. & Lombardo, P. 2003. Wood-Based Filter for Nitrogen Removal in Septic Systems, (Submitted to: Journal of Environmental Quality), (unpublished). Schipper, L.A. & Vojvodic-Vukovic, M. 2000. Nitrate removal from groundwater and denitrification rates in a porous treatment wall amended with sawdust. Ecol Engineering, 14:269–278. Schipper, L.A. & Vojvodic-Vukovic, M. 2001. Five years of nitrate removal, denitrification and carbon dynamics in a denitrification wall. Wat. Res. Research, 35(14):3473–3477. Tredoux, G, Talma, A.S. & Engelbrecht, J.F.P. 2000. The increasing nitrate hazard in groundwater in the rural areas. Paper presented at WISA 2000, Sun City, RSA, May 2000.

Coastal aquifers intrusion at semi-arid region of Turkey L.Yilmaz Technical University of Istanbul, Civil Engineering, Hydraulic Division, Maslak, Istanbul, Turkey Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Coastal aquifers are important sources of water for domestic, agricultural and industrial uses. Under natural conditions the hydraulic gradient is towards the sea so that there is a natural outflow of fresh groundwater. Frequently the hydraulic gradient is small. Therefore very little extraneous activity is required to disturb the natural system and cause the fresh water to become. This situation poses a difficult management problem which is best addressed by means of mathematical models. This research describes the use of such models together with the difficulties likely to be encountered.

1 INTRODUCTION 1.1 Relationship between the level of the water table and the depth to the saline wedge A relation between the level of the water table and the depth to the saline wedge in an unconfined aquifer under steady conditions of flow (Badon Ghijben, 1889; Herzberg, 1901, Davis, 1978) points out that Joseph DuCommun (1828) made similar observations. Prior to the work of these pioneers it was thought that saline water occurred at a depth close to sea level. The saline water close to the sea shore is defined by Badon GhijbenHerzberg equation, which was derived by a simple application of hydrostatics. The weight of a column of fresh water of height hf+z is equal to the weight of a column of saline water of height z. If rof and ros are the densities of fresh and saline water respectively, it is given in equilibrium conditions rosgz=rofg(hf+z) (1) or (2) If the relative densities of fresh and saline water are taken as 1.0 and 1.025 respectively, then

Water resources of arid areas

192

z=40 hf (3) This expression is a good approximation in steady state conditions when the zone of dispersion is only a small fraction of the saturated thickness of the aquifer. Since fresh water is flowing along the interface some mixing will occur due principally to microscopic and macroscopic dispersion. When the saline and fresh water mix in the zone of dispersion then the diluted saline water becomes less dense and will rise along a seaward path. The resulting mechanism is similar to thermal convection, the only difference being that the gradients are caused by changes in density due to changes in salinity instead of temperature (Cooper, 1964). This flow will advect some saline water towards the sea. Therefore, in order to preserve the saline mass balance, a small flow of saline water must occur in the landward direction. This flow creates a head loss, thus a reduction in pressure at the interface and a reduction in the level of the interface. This application gives the position and movement of a saline front in a coastal aquifer. 1.2 Groundwater flow equation Using the Darcy’s law and the law of conservation of mass to a control volume, (Pinder and Bredehoeft, 1968; Konikow and Bredehoeft, 1978) gives (4) where Tij=transmissivity tensor (L2T−1) h=hydraulic head (L) above a reference point S=storage coefficient (−) W=source or sink volume flow term (LT−1), positive for outflow (=W (xi, t), i=1, 2. This is usually the recharge, pumping and evapotranspiration). xi, Xj=Cartesian coordinates (L) t=time (T) The advection-dispersion equation is given using the notation of Konikow and Bredehoeft (1978) as; ∂(Cb)/∂t=∂/∂xi(bDij∂C/∂xj)−∂/∂xi(bCVi)−C′W/ε (5) where Dij=coefficient of hydrodynamic dispersion (L2T−1) Vi=seepage velocity in the direction xi (LT−1) C=concentration of the pollutant (ML−3) C′=concentration of the pollutant in the source or sink fluid (ML−3) b=saturated thickness of the aquifer (L) E=effective porosity of the porous medium (−) This equation gives the change in chemical concentration due to kinematic dispersion and diffusion, the effect of advective transport and the removal of pollutant due to fluid sources and sinks.

Coastal aquifers intrusion at semi-arid region of Turkey

193

2 MAIN OBJECTIVE 2.1 Solutions of the groundwater equations Therefore the interface is not sharp and a mixing zone exists, the thickness of which depends upon the hydrodynamics of the aquifer. If this transition zone is only a small fraction of the saturated thickness of the aquifer then the assumption of a sharp interface is reasonable and a good mathematical description of the shape of the saline wedge can be obtained. The thickness of the fresh water wedge decreases in the seaward direction and the slope of the water table steepens towards the coast. Therefore the shape of the interface is concave upwards. If the more realistic view is taken that the fresh and saline water are miscible, then the interface cannot be sharp and the mathematical description of the problem becomes more complicated. The assumption of a sharp interface cannot be considered reasonable if the flow situation varies with time since the hydrostatic pressure distribution will vary and the assumed interface will move either landwards or seawards. This results in the sharp interface being replaced by a zone of dispersion in which the salinity of the water varies from fresh to very saline. Clearly the simplifying assumption of a sharp interface makes for a mathematically simpler but less accurate model. 2.2 Sharp interface models The relationship between groundwater levels and the depth to the saline wedge is given by Badon Ghijben and Herzberg who, working independently, developed a relationship between the level of the water table and the depth to the saline wedge in an unconfined aquifer under steady conditions of flow (Badon Ghijben, 1889; Herzberg, 1901). Davis (1978) points out that Joseph DuCommun (1828) made similar observations. Prior to the work of these pioneers it was thought that saline water occurred at a depth close to sea level. In contemporary practice the result is always referred to as the Badon GhijbenHerzberg equation. This equation is derived by a simple application of hydrostatics. Since the interface is stationary then the weight of a fresh water above the interface is exactly balanced by the pressure of the saline water below the interface. By consideration of the Figure it can be seen that the weight of a column of fresh water of height hf+z is equal to the weight of a column of saline water of height z. If rof and ros are the densities of fresh and saline water respectively, then for equilibrium rosgz=rosg(hf+z) (5) If the relative densities of fresh and saline water are taken as 1.0 and 1.025 respectively, then z=40 hf (6) This simple expression gives a remarkably good first approximation to the depth below sea level of the interface under steady state conditions when the zone of dispersion is only a small fraction of the saturated thickness of the aquifer.

Water resources of arid areas

194

This concept can be further developed to determine the extent of the penetration of the saline wedge inland. Many analyses can readily be developed, for example the determination of the shape of the interface when the seepage surface is submerged beneath the sea (Glover, 1964) and the shape of the saline upcone beneath a pumping well in a coastal aquifer (Schmorak and Mercado, 1969; Sahni, 1972). 2.3 A sharp interface with some mixing When a sharp interface is assumed, then this interface is a flowline in the same way as the water table is a flowline. Hence it is a boundary condition for the problem. Since fresh water is flowing along the interface some mixing will occur due principally to microscopic and macroscopic dispersion. When the saline and fresh water mix in the zone of dispersion then the diluted saline water becomes less dense and will rise along a seaward path. The resulting mechanism is similar to thermal convection, the only difference being that the gradients are caused by changes in density due to changes in salinity instead of temperature (Cooper, 1964). This flow will advect some saline water towards the sea. Therefore, in order to preserve the saline mass balance, a small flow of saline water must occur in the landward direction. This flow creates a head loss, thus a reduction in pressure at the interface and a reduction in the level of the interface as predicted from the Badon Ghijben-Herzberg equation. The mechanism is shown in Figure. It is possible to extend these concepts to determine the solutions to various moving interface problems. However, except for some very restrictive cases, analytical solutions do not exist. Hence numerical methods are required to solve the resulting equations, which usually means that it is more convenient to use commercially available groundwater quality models. Approximate solutions for moving interface problems, including numerical ones, are discussed by Bear (1979). 2.3.1 Equations of groundwater flow and advection—dispersion The above section dealt with some very simple first approximations for determining the position and movement of a saline front in a coastal aquifer. Whilst these are useful in the early stages of a study they do not permit a full solution to the majority of aquifer problems. For example they cannot deal with spatial variations of geology or aquifer parameters nor with multi-layered aquifers. In order to achieve this two equations are required, one to describe the groundwater flow and one to describe the movement of the salt. These will be considered in turn. 2.4 Solutions of the groundwater equations The finite difference method is used for solving this type of differential equations. The first step is to give the area of the model in mathematical terms, which is called the solution domain. This solution domain is covered by a rectangular grid which can be either regular or irregular. The differential equation is replaced by a set of difference equations, one for each grid point. This results in n×m simultaneous equations which have to be solved, where n is the number of rows and m is the number of columns of the

Coastal aquifers intrusion at semi-arid region of Turkey

195

grid. The finite difference method is, perhaps, the most frequently used technique for solving the flow equation. However, it is not often used to solve the advection-dispersion equation because of a phenomenon known as numerical dispersion. The numerical solution usually appears to advance the solute at a rate which is greater than is physically possible. Finite difference schemes can be developed to minimize dispersion. However they are liable to cause either overshooting or undershooting which appear in the solution as oscillations. Van Genuchten (1976) analyzed and gave as a result that the finite element schemes will usually yield more accurate solutions than finite difference ones. There are some rules which can be helpful in minimizing the effects of dispersion. These use a form of the Peclet number, Pe, and the Courant number, C. The grid should be designed such that Pe(=dx/De)<4, where dx is some characteristic grid size and De is some characteristic dispersivity. The finite element method was first developed in the solid mechanics (1950). Then it was used to solve the groundwater flow equation. The finite element method is an integral (as opposed to differential) approach in which the regular grid of the standard finite difference method is replaced by an irregular polygonal mesh which allows the modeller to describe natural shapes more precisely. In groundwater the polygonal shape is, almost always triangular. The finite element mesh can be adapted to describe the irregular shape of the boundary and obtained an accurate description of rapidly varying phenomena. In this approach the piezometric surface is approximated by a series of small triangular surfaces which can be flat or curved. If the chosen basic functions are linear then the surfaces will be flat and the variation of head within each element will be linear. The point of intersection of the triangles is called a node and each triangle is called an element. The equation is solved by a weighted residual technique of which the most popular one is the Galerkin method. In this method the weighting functions are made equal to the basic functions and the integration is then performed over each element and summed to yield the contribution from all the elements that make up the solution domain. The finite element method is a powerful and mathematically elegant technique but it is difficult to program. The method of characteristics was developed to solve hyperbolic partial differential equations (Courant and Friedrichs, 1948) and was first used for flow through porous media by Gardner et al. (1964). They proposed the method because they argued that when flow velocities become large the dispersion equation is, essentially, hyperbolic. The method has been extensively applied for solving the advection-dispersion equation and is now the basis of one of the standard solute transport models (Konikow and Bredehoeft, 1978). The solutions are x=x(t), y=y(t) and C=C(t), where x and y are the coordinates in a Cartesian system, C is the pollutant concentration and t is time. These are called the characteristic curves of, in this case, the advection-dispersion equation. Once these solutions are available then a solution of the advection-dispersion equation can be obtained by following the characteristic curves. Gardner et al. (1964) state that “Each point corresponds to one characteristic curve and values of x, y and C are obtained as functions of t for each characteristic”. Essentially this is the Lagrangian approach of classical hydrodynamics. It is particularly useful for making cross-sectional models of saline intrusion. There are many other methods for solving the groundwater flow and advectiondispersion equations, as integrated finite differences (Tyson and Weber, 1963; Goodwill,

Water resources of arid areas

196

1980), boundary element methods (Liggett and Liu, 1983) and analytic elements (Strack, 1989). 3 RESULTS The model requires substantial amounts of field data, the collection of which is both time consuming and expensive. The hydrological and geological data are used for the area to be modeled. These data are: ● Surface and subsurface geology. ● Piezometric levels for all the aquifers contained in the system. ● Aquifer characteristics and likely boundaries, soils, land use and vegetations. Since aquifers are subject to recharge and pumping, data on the quantities and timing of these will be required. Such data will include precipitation, evapotranspiration and pumping. If irrigation is undertaken, then also rates of application and return flows and river flows are required, including flows to and from the rivers to the aquifers if they are not in direct hydraulic contact. Since the concern here is with saline intrusion then data on salinity, both areally and vertically, will be required. If any of these data do not exist or are too scanty, then a field programme will be required to collect them. In order to collect and plot all these data an accurate topological map is essential, the scale of which will depend on the size of the aquifer and the scale of the problem being studied. This map should show the surface contours, surface water bodies, streams and man-made watercourses such as irrigation canals and drainage ditches. REFERENCES Bear, J. 1972. Hydraulics of Groundwater. New York, McGraw-Hill Book Co., 567 p. Bras, R.L. & Rodriguez-Iturbe, I. 1976. Evaluation of mean square error involved in approximating the areal average of a rainfall event by a discrete summation, Water Resources Research, 12(2), 181–184, a. Hubbert, M.K. 1940. The Theory of Ground-Water Motion, The Journal of Geology, 48(8), Part-I, Nov.–Dec. Pinder, G.F. 1982. Finite Element Simulation in Surface and Subsurface Hydrology, Gallagher, Vol. 4. Pinder, G.F. & Abriola, L.M. 1982. Calculation of Velocity in three space dimensions from hydraulic head measurements, Groundwater, 20, 205–213. Pinder, G.F. & Gray, W. 1982. Finite Elements in Water Resources, edited by P.Holz, V.Meissner & C.A.Brebbia (eds), Berlin, Springer Verlag: 4. Rushton, K.R. & Redshaw, S.C. 1979. Seepage and Groundwater Flow, Wiley, Winchester, UK, 339 pp. Sarma, S.V.K. & Silva, T.C. 1987. Hydraulic response to pumping in free aquifers, ABAS, 11, 26– 32. Sarma, K.V.S. & Antonio, A.P. 1997. Decontamination of pollutants from aquifers using the concept of induced flow from adjacent rivers. Intl Conf. on Large Scale Water Resources Projects, Oct. 20–23. Kathmandu, Nepal, EI 17–24.

Coastal aquifers intrusion at semi-arid region of Turkey

197

Sumer, B. 1980. The Determination of Water Quality at the Sapanca Lake, TUBITAK Project No. QA6–4, Sakarya. U.S. Environmental Protection Agency, 1988. Model Assessment for Delineating Wellhead Protection Areas, Office of Groundwater Protection, Washington DC, 210 pp. Yilmaz, L., Agiralioglu, N. & Saltabas, L. 1999. The determination of the water-use capacity of the Sapanca Lake in Turkey. Proc. Intl. Conf. on Water, Environment, Ecology, Socio-economics and Health Engineering (WEESHE), Oct. 18–21, Seoul National University, Seoul, Korea, Water Resources Pubs., LLC, 162–166.

Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region Y.B.Mkwizu Lawyers’ Environmental Action Team (LEAT), Dar es Salaam, Tanzania H.H.Nkotagu Department of Geology, University of Dar es Salaam, Dar es Salaam, Tanzania Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: This paper focuses on the need to manager groundwater exploitation by comparing its abstraction rates with natural recharge rates and a case study was done in the Kizinga catchment. Despite the increase in groundwater exploitation, there is no comprehensive legal mechanism to ensure proper management of groundwater resources and no specific provision for groundwater abstraction under The Water Utilization Act. Instead groundwater exploitation management has been provided for in the same way as surface water. The Act has not demanded operators of wells or boreholes to submit data and records to water authorities. It is therefore possible that substantial commercial drilling of groundwater has been conducted without adequate monitoring and controls. It is suggested the proper management of groundwater utilization be established to balance recharge and discharge. This will ensure that groundwater abstraction is done sustainably and thus avoid negative consequences of groundwater depletion.

1 INTRODUCTION The high demand for freshwater in the Dar es Salaam City, suggests clearly that, surface water can no longer meet the projected total demand. The second best alternative remains on groundwater. Even with the presence of surface water, groundwater can still be preferred on the basis of easier protection from pollution, better dependability during drought periods, and on the supplying costs. It is necessary however that groundwater resource is used with proper management focusing on both its quality and quantity.

Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region

199

Evaluation of groundwater recharge rates and areas is an important step towards thorough understanding of its quality and quantity. 1.1 Location and climate of the study area The study area is located to the south of Dar es Salaam region bearing geographical coordinates 39°02E and 39˚18E and 6°50S and 7°00S. It is built on a low lying coastal plain with an altitude varying between 20m to 240m above sea level in Kurasini area and in Pugu and Kisarawe hills respectively with a total surface area of 191km2 (Service Plan 1997). The study area is drained by Kizinga river having its upper reaches in Pugu and Kisarawe hills. The river flows in a NE direction towards Indian Ocean (Fig. 1). Climatically the daily temperature ranges from 18.1°C to 32.2°C with a mean value of 25.75°C, while actual evaporation has a mean monthly value of 160mm ranging from 128mm to 181mm. The mean annual precipitation is 1124mm.

Figure 1. Location of the study area. 1.2 Objectives The purpose of the study was to evaluate groundwater recharge rates in the area. Specifically, the research aimed at: Determining the mean annual groundwater recharge rates in the study area.

To determine the source of groundwater recharge in the study area.

Water resources of arid areas

200

1.3 Groundwater evaluation in the study area Very few studies on the behavior of groundwater resource have been documented in the study area. The study on Coast/Dar es Salaam Water Master Plan by the Ministry of Energy and Minerals (MEM) (1979) for example, found that the coastal sedimentary deposits of Coast and Dar es Salaam regions did not include aquifers that provide major groundwater supplies. The project revealed further that within the area; almost seventy per cent of the area was found to be underlain by material that could yield water of insufficient quantity or inadequate quality. On another study, Matondo (1978), found that the Kizinga basin is a potential source of groundwater. He also noted that groundwater from drilling is very close to the ground surface. He further found that, only 8% of annual rainfall appears as total runoff while the major part of the rainfall volume was stored within the aquifer. Both studies didn’t make evaluation on groundwater recharge rates. 2 METHODOLOGY Data that incorporate various hydrogeological units and taking into account all flow components such as discharge, infiltration, subsurface inflow to and outflow from the basin’s aquifer and abstractions through pumped wells were collected. The main data/and data sources were: – Boreholes and wells and their hydrologic information drilled in the study area up to 1999 from the Ministry of Water, borehole drilling unit-Ubungo – Monthly rainfall data, maximum and minimum temperature and evaporation for Dar es Salaam from Tanzania Meteorological Agency, Dar es Salaam office – Runoff measurements from rivers Kizinga from 1967 to 1980, From Ministry of Water, RBM-Ubungo

Table 1. Chloride concentration of rainwater in the study area in (mg/l). Location Ukonga Kiwalani Yombo Temeke Mbagala Pugu Gongolamboto Total Mean

Mar Apr May Mean 3.1 2.8 3.0 1.9 3.2 2.4 3.7 2.6 3.6 2.8 3.2 1.8 3.0 2.8 22.8 17.1 3.26 2.44

2.6 3.9 2.8 3.4 3.0 2.4 3.2 21.0 3.04

2.83 2.93 2.80 3.23 3.13 2.47 3.00 20.39 2.91

Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region

201

Figure 2. The groundwater flow net in the study area. Abbreviation Site Pg Gt Uk Airport Kw Tr Ym Ch Tm Mb Mt Kr

Mean hydraulic head (m)

Pugu Gongolamboto Ukonga Kiwalani Tazara Yombo Chang’ombe Temeke Mbagala Mtoni Kurasini

224 63.5 43.2 37.4 22.2 20.1 14.1 12.6 10.9 11.2 9.0 6.3

The fieldwork involved rainwater sampling in different sites within the study area for the rainfall period of March, April and May in the year 2000. Another fieldwork activity was to collect water from boreholes in some selected locations in the study area. Chloride determination from water samples was undertaken following standard method as reported by (APHA 1985). The results obtained from rainwater samples are given in Table 1. 3 RESULTS 3.1 Groundwater resource accumulation and flow Hydraulic heads from water level measurements at various points and elevation at various borehole points indicate the presence of a close correspondence to topographic heights.

Water resources of arid areas

202

The comparison between hydraulic head and topographical height shows that the direction of groundwater flow is roughly to the drainage pattern which follows the gradient of the land surface. Groundwater movement in the study area is therefore approximately to the north east which is generally the direction of Kizinga river (Figs 1 & 2). The above observations on the groundwater movement in the study area suggest that groundwater starts flowing from Pugu and Kisarawe hills towards low lying plains of Yombo; Changombe; Temeke; Mbagala; Mtoni and Kurasini. This shows that precipitation within Pugu and Kisarawe hills is the major source of groundwater recharge in the study area. 3.2 Aquifers properties in the study area Layers of sand are important for the hydrogeology of the study area as they are mostly accompanied with good water bearing capacity and they allow significant quantities of water to be drawn from them (Fig. 3(d)). Generally the texture of sands are medium to coarse with gravels and pebbles existing in clay matrix. Clay layers tend to hold water which can not be withdrawn easily and therefore they don’t have good water bearing capacity (Figs 3(a & c)). Significant amount of groundwater has been found basically to occur in two types of aquifers namely sands and gravels and limestone. Clay and clay bound sands are poor aquifers and their importance is mainly on the formation of confining layers for most of sand and gravels aquifers (Fig. 3(b)). 4 ESTIMATION OF GROUNDWATER RECHARGE RATES Evaluation of groundwater recharge has significant implications for not only the study of groundwater quantity, but also water quality. Infiltrating water can carry contaminants from

Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region

203

Figure 3(a & b). Lithology of aquifers with their scales in (m) in some sites of the study area. ground surface to the aquifer. Understanding the rate and mechanism out of which such infiltration takes place can therefore lay a foundation in setting down strategies for the prevention of groundwater contamination. Three methods were used to estimate groundwater recharge. These are: the water balance method, hydraulic method and chloride profile method.

Water resources of arid areas

204

Figure 3(c & d). Lithology of aquifers with their scales in (m) in some sites of the study area.

Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region

205

Figure 4. A comparison between annual groundwater recharge, rainfall and actual evaporation. 4.1 Water balance method The method was used under the following assumptions: – Precipitation is the only inflow into the basin i.e., there is no leakage or underground channel to the basin – River discharge and evapotranspiration is the ultimate water outflow – Basin storage is steady and hence the in storage is considered to be zero for long term period. The water balance method requires the use of a combination of actual evaporation, surface runoff and precipitation data in order to estimate annual groundwater recharge. Groundwater recharge in this method is calculated as a remainder when losses, identified in the form of runoff and evaporation have been deducted from precipitation. This can be presented in the following equation: P=E+R+∆S Where: P—Precipitation (mm), E—Actual evaporation (mm), R—Runoff over the catchment (mm) and ∆S—Change in Storage (mm). The value of groundwater recharge obtained using water balance method gave the average value of 81.3mm/year. The comparison between annual groundwater recharge, annual rainfall and annual actual evaporation shows that, the values of recharge in most cases increase with increase in rainfall and decrease with increase in actual evaporation (Fig. 4) and that, annual groundwater recharge and annual rainfall are highly correlated.

Water resources of arid areas

206

The annual actual evaporation in the study area always exceeds annual rainfall. However a plot of mean monthly actual evaporation and rainfall for the one year period (Fig. 5), shows that there are few months when rainfall exceeds actual evaporation. These months are mainly, March; April; and November and it is expected that during these months, groundwater recharge takes place. 4.2 Hydraulic method This method was used to estimate the amount of annual groundwater in flow (Q) in to the Kizinga river catchment area. An average hydraulic gradient (I) of 8.063×10−3, hydraulic conductivity (k)

Figure 5. Mean monthly variation of rainfall with actual evaporation for the year 1968 in the study area. of 3.14×10−5m/s along a cross section at the center of the basin having the maximum aquifer thickness of 28m and a surface width of 27km, were used. Darcy’s law, Q=KIA were used Where: Q is the quantity of water (m3/s), K is the hydraulic conductivity (m/s), I is the hydraulic gradient and, A is the area (m2). The value obtained was 6036076.153×109mm/year and when extrapolated to the entire basin’area of 191km2, this amount of annual groundwater in flow in mm per year was calculated such that, Recharge rate=(6036076.153×109mm/year)/191×1012mm=31.6mm/year.

Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region

207

4.3 Chloride profile method The technique regards chloride as an inert element, and compared with other inorganic ions, it is not added or removed by water rock interaction. The element is considered as an inert in the hydrological cycle having its source from the atmosphere. It has the advantage over tracers involving water molecule in that atmospheric inputs are conserved during recharge processes allowing a mass balance approach to be used (Nkotagu 1996). On using the chloride profile method, it is assumed that the amount of water and chloride added at the surface should equal the amount of water and chloride percolated down. This however is not always true and therefore the following assumptions were considered when using this method. (i) Recharge is only that derived from precipitation (ii) Recharge is largely by piston flow mechanism (iii) Chloride in soil water is from precipitation and dust only (iv) The precipitation amount used in the recharge and soil water age equations is reasonable for the time represented by samples (v) The total chloride input value used in the recharge and soil water age equations is reasonable for the time represented by the samples (vi) Dispersive mixing of water and chloride is small (vii) The chloride uptake by plants is negligible Knowing that there is a possibility of groundwater being affected by marine intrusion or marine connate source as the study area is situated in the coast, the use of the method involved boreholes with water whose chloride concentration values fall within acceptable ranges of freshwater and

Table 2. Summary of the groundwater yield in the study area. Total No. of boreholes Maximum yield Minimum yield Total annual yield

130 60.923m3/h 0.220m3/h 9.4×106m3/year

out of seawater. This was achieved by including in the determination of groundwater recharge, only boreholes with chloride content whose ratio is less than 1. Groundwater recharge using this method was estimated as follows. The mean chloride content of precipitation was found to be 2.9mg/l and the mean chloride concentration of groundwater was found to be 71.8mg/l. The ratio of chloride content of precipitation to groundwater was determined to be 0.04. A long term mean annual precipitation of 1124mm was used in the calculation. Estimation resulted in groundwater recharge rate of 45.4mm/year.

Water resources of arid areas

208

5 DISCUSSION OF THE RESULTS The three methods gave a mean value of 52.8mm/year(Table 2), equivalent to 10.1×106m3/year. This indicates that, 4.7% of the long term mean annual precipitation, which is 1124mm, ends up as annual groundwater recharge. It can then be concluded that at present, annual groundwater production rates which is 9.4×106m3/year is approximately equal the annual natural groundwater recharge rates from the study area. The value obtained using water balance method is much larger as compared to the values of the other two methods. The big value in this method is likely to have been contributed by errors in the process of data taking. The system that was used in measuring runoff which based on water level recording two times a day is likely to have introduced some errors. Such errors may be missing of flood peaks especially that occurring at night and therefore registering less storm runoff than real values happening in nature. According to Matondo 1978, the personnel engaged in data collection by that time in most basins including Kizinga, were unskilled or only semi skilled and therefore less accurate and efficient in collecting data. There was no even distribution of rain gauge stations in the basin. Mostly only Dar es Salaam airport had a recording gauge (Matondo 1978). The existing raingauge therefore, did not facilitate studying the rainfall distribution in respect to time and space. This restricts further the accuracy of investigations. It can therefore be seen that the rainfall and runoff records might have errors, which then could possibly contribute to more errors in the computation of effective rainfall and average precipitation, and therefore the final results. In additional to shortcomings mentioned above, the data used in this method were recorded about 20 years ago while the other two methods used the data recorded within last two years. It is possible therefore that, the soil condition 20 years back supported more groundwater recharge. This is supported by the fact that the land in the study area has been disturbed through construction and cultivation. 6 CONCLUSION AND RECOMMENDATION The major contribution in recharging the study area has been found to be the faults on the slopes of Pugu and Kisarawe hills, which are quite permeable and that the direct infiltration of rainwater is the main source of groundwater recharge in the study area. The average value of annual groundwater recharge rates has been found to be 52.8mm/year after combining all the three methods used. This value is approximately 10.1×106m3/year and it is 4.7% of the long term mean annual precipitation of 1124mm. A projection of maximum groundwater abstraction showed an increase of more than 100% after every two years. However the natural groundwater recharge is expected to remain constant, and therefore the maximum production rate will be more than natural recharge rate within a period of few years to come. This relation when projected over a long period may result in negative consequences such as depletion of groundwater supply, reduced

Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region

209

stream flow, deterioration of water quality and more importantly land subsidence may be expected in the near future. On the other hand there is no comprehensive legal mechanism to ensure proper management of groundwater resources. The Water Utilization Act has not provided a holistic approach to the management of water resources especially with regard to the management of groundwater resources. There is no specific provision for groundwater abstraction under the Act and instead groundwater exploitation has been provided for its management in the same way as surface water. There is no separate provision under the Act to regulate groundwater-drilling operations. However, abstraction of groundwater of more than 22,700 litres per day requires a water right issued under the Act. The Act has not demanded for operators of wells or boreholes to submit data and records to the water authorities. It is therefore possible that, commercial drilling of groundwater of substantial scales has been conducted without adequate monitoring and controls. Clear understanding on the aquifer parameters and recharge rates for other parts of Dar es Salaam and the country at large is fundamental before embarking on further exploitation of the resources. Studies on the subject however are very limited to academic purposes and have not been able to comprehensively provide a clear understanding on the aquifers that provide the resource. In general, the studies that have been conducted so far indicate a negative trend in the status of groundwater in different parts of the country. The absence of adequate data and legislation has been impinging on effective management of groundwater resources. The Kazimzumbwi forest which is in the Pugu and Kisarawe hills is suffering a massive deforestation from illegal harvesting of forest products. The capacity of the faults to serve as groundwater recharge will consequently be affected. It is possible therefore that, in the near future the rate of natural recharge will decrease while that of abstraction will keep on increasing. It is recommended that, the faults on the slopes of Pugu and Kisarawe hills, which serve in recharging the area be conserved and protected from disturbance to ensure continuation of safe and enough groundwater supply. Possible alternatives for freshwater supply should be used. Measures that reduce water wastage must be introduced and encouraged. More studies on groundwater recharge and aquifer performance need to be conducted in the study area and others around Dar es Salaam using other methods so as to have a complete understanding on the aquifer system in Dar es Salaam city. ACKNOWLEDGEMENTS We would like to thank all those people who contributed to the preparation and completion of this work. Our special thanks go to the Department of Geology of the University of Dar es Salaam for the facilities they have provided to us during our research. Many thanks go to MHO program through the Faculty of Science University of Dar es Salaam for the financial support in undertaking the research. We wish to acknowledge the assistance of the Staff of the institutions in that time; Ministry of Water, Ubungo Maji; DAWASA headquarters and Tanzania Meteorological Agency Dar es Salaam office for supporting us with necessary data. We are grateful to the Management

Water resources of arid areas

210

of Lawyers’ Environmental Action Team (LEAT), for their support on access to the organizational facilities and information during the preparation of this paper. REFERENCES APHA, 1985. Standard Method for the Examination of Water and Wastewater. 16th Edition. American public Health Association, Washington, D.C. Matondo, J.I. 1978. A kinematic conceptual model for Kizinga basin for Estimation of Hydrological variables. M.Sc. Thesis, University of Dar es Salaam. Ministry of Energy and Minerals (MEM) 1979. Coast/Dar es Salaam regions water master plan. Dar es Salaam. Nkotagu H.H. 1996. Hydrological and Isotopic characterization of a fractured basement groundwater flow system in Semiarid Area of Dodoma, Tanzania. Znge: Berlin, Techn. Univ. Diss. Serviceplan, 1997. Report on the evaluation of groundwater sources of Dar es Salaam. Supporting Report B. Dar es Salaam.

Theme C: Socio-economic aspects

KNUST experiences in capacity building in the water and sanitation sector S.N.Odai, F.O.K.Anyemedu, S.Oduro-Kwarteng & K.B.Nyarko Department of Civil Engineering, KNUST, Kumasi, Ghana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Capacity building for the water and sanitation sector in Ghana has been in existence in the Civil Engineering Department (DCE) of Kwame Nkrumah University of Science and Technology (KNUST) since the university was established. The activities of the department in the area of capacity building for the water and sanitation sector have become very popular after the establishment of the water supply and environmental sanitation programme (WSESP) in 1996. In fact, this recent development has brought the department to the forefront of capacity building and applied research in the water and sanitation sector of Ghana. The project, aimed at capacity building for sustainable development and growth in the water supply and sanitation sector in Ghana and the West African sub-region has so far produced several professionals from Ghana, from three West African states and one from an East African State. Capacity building through short courses and tailormade programmes has helped in training several institutions and communities in Ghana leading to sustainable development. The project has impacted positively on both institutional and human capacities in the country. The paper looks at the experiences of KNUST in providing highlevel and low-level capacity building for Ghana.

1 INTRODUCTION The human resources requirements of the water and sanitation sector are largely similar to those of other professions. Factors militating against capacity building in the sector have been mainly due to lack of funds and lack of understanding of the urgency of the need to improve the sector. In many developing countries of Africa, improvement in the water supply and environmental sanitation sector has lately been recognised to have a direct positive impact on public health (Ghana Government, 2003). The development of the sector is seen as crucial to the successful control and eradication of communicable diseases in general. As a result, governments, policy makers, non-governmental agencies,

Water resources of arid areas

214

and external support agencies have begun directing attention to issues related to these sector in an effort to improve the overall health of citizens by facilitating easy accessibility to water and sanitation. Until the last decade, the progress and growth in the water supply and environmental sanitation sector has been little and very slow. One of the major factors hampering the desired rate of progress and growth in the sector has been the lack of adequate personnel and professionals with the requisite skills, expertise and experience to lead and manage the sector. In fact, during the 1991 UNDP symposium at IHE-Delft, on A Strategy for Water Sector Capacity, it was acknowledged that capacity building in the water supply and environmental sanitation sector was essential for the development, growth and sustenance of the sector at the local, national and even sub-regional levels (KNUST, 2002a). In the wake of this symposium, it was felt that the training and re-training of professionals for and within the water supply and environmental sanitation industry must become part of the central focus of academic institutions that have long traditions of providing quality leadership training as well as professional expertise. The UNESCOIHE Institute for Water Education in Delft and the Department of Civil Engineering of the Kwame Nkrumah University of Science and Technology (KNUST) in Kumasi quickly took the initiative to develop a programme for supporting the restructuring and strengthening of the water supply and sanitation sector in Ghana and the sub-region. The programme was preceded with a needs assessment of the situation in several neighbouring countries, to assess the percentage coverage of sanitation and water supply in urban and rural communities, and the statistics were stunning. Following this needs assessment, the DCE of KNUST in collaboration with UNESCO-IHE in 1996 initiated a programme for developing human resource capacity in the water supply and environmental sanitation for a wide range of beneficiaries, spanning from sector professionals to low-level operators and sometime even uneducated water board members. The high-quality training programme in the water supply and environmental sanitation sector was conceived under the project name “Water and Environmental Sector Capacity Building and Sustainable Development in Ghana and the Region” designed by the KNUST-Kumasi and UNESCO-IHE. The project aims at providing the sector within the West African sub-region with the necessary skills, knowledge and expertise to meet the demands, challenges and opportunities anticipated with the projected growth in the region. This paper presents the experiences of KNUST in capacity building in Ghana and the sub-region. 2 HISTORY OF CAPACITY BUILDING IN THE WATER SECTOR AT KNUST The Kwame Nkrumah University of Science and Technology (KNUST) was established in 1951 to train Scientists and Technologists for both Ghana and other African countries. Academic programmes are run at the undergraduate and post-graduate levels by different faculties, schools and institutes. Training in water-related disciplines is offered in several departments in the university.

KNUST experiences in capacity building in the water and sanitation sector

215

Presently, the DCE, the Chemistry Department, and the Geological Engineering Department collaborate in delivery of education and training in water quantity and quality. The programmes of the Department of Civil Engineering cover both surface and groundwater resources and water quality analyses, while the Chemistry Department is involved mainly in water quality analyses, and the Geological Engineering Department involved in groundwater resource development. In addition, the Agricultural Engineering Department looks at irrigation and water for food, the Biological sciences Department works on environmental science, while the Physics Department masters in groundwater and limnology. Since 1996, the DCE and UNESCO-IHE have been developing human resource capacity in water supply and environmental sanitation, targeted at all levels of sector professionals (sometimes including persons with low-level education). Since the establishment of WSESP which has the primary focus of postgraduate training for the sector, several, short courses such as Public Private Participation, Water Treatment, Wastewater Treatment, Solid Waste Management, Urban Water Transportation and Distribution, etc., have been offered annually. 3 INSTITUTIONAL CAPABILITY OF KNUST The institutional capacity of KNUST to act as a capacity building centre in the water and sanitation sector is discussed under the following sub-headings. 3.1 Human resources The departmental academic strength comprises of 22 highly qualified lecturers, mostly PhD graduates from prominent universities in addition to experienced technicians who support the execution of educational programmes and consulting services; eleven out of the 22 staff members in the department specialise in water and sanitation issues. Specifically there are 5-PhD, and 6-Msc holders in the two sections. Three out of the six MSc’s are currently pursuing sandwich PhD programmes with UNESCO-IHE; one in Wastewater Treatment, one in Utility Management, and the third in Water Treatment. 3.2 Facilities The facilities at KNUST for training purposes include refurbished classrooms, a computer laboratory that gives access to each participant, a refurbished laboratory that allows for water and wastewater quality analyses for training and research purposes. Students and lecturers in the DCE, together with other departments in the School of Engineering, have access to a well-equipped library, apart from the main library of the University. In addition, a collection of specialised books is at the disposal of the staff and participants in WSES project. A 30-room hostel has been built purposely for the MSc participants. These rooms are also available for use by short course participants when these are organised during the vacation periods. The university now has Internet connectivity for research, and a website has been created for the project.

Water resources of arid areas

216

Lecture notes, mostly from UNESCO-IHE, in addition to modern ICT equipment facilitate our objective of capacity building. 3.3 Networking and outreaching capabilities At the national level, the Department has strong links with sector organisations in Ghana, such as the Ghana Water Company Limited, Community Water and Sanitation Agency, Ministry of Local Government and Rural Development, The Environmental Protection Agency and the Water Resources Commission. This link is maintained through quarterly meetings held together with these agencies. The main objective of this networking is to enable the university identify the problems of industry at first hand and provide the necessary advice. The department also depends on its international links to facilitate information acquisition. Presently, the university has active links with UNESCO-IHE in Delft, the University of Newcastle in UK, the University of Bristol in UK, and the National University of Rwanda. In fact, the department’s links with industry and the sector professionals have led to the active participation of these professionals in part-time lectures and the running of some short courses. The strategy is such that use is made of these professionals for training and lecturing where there is a gap. The departmental strategy is to maintain this alliance. 4 EXPERIENCES OF CAPACITY BUILDING UNDER THE WSES PROJECT The Water Supply and Environmental Sanitation Project is aimed at capacity building for sustainable development and growth in the water supply and sanitation sector in Ghana and the West African sub-region. It seeks to strengthen the sector through training of high-level personnel for institutions and organisations that have a stake in the water supply and environmental sanitation industry and professionals with active careers in the sector, in addition to training of low-level personnel. In Ghana, the programme targets the following sector organisations and their professionals: ● Ghana Water Company Limited (GWCL), ● Community Water and Sanitation Agency (CWSA), ● Environmental Protection Agency (EPA), ● Ministry of Works and Housing MWH), ● Ministry of Local Governments and Rural Development (MLGRD), ● Ministry of Health (MoH), ● Ministry of Environment and Science (MES), ● Water Research Institute (WRI), ● Consulting firms and contractors, ● Water-related industries, ● Small Towns Water Operators and Managers.

KNUST experiences in capacity building in the water and sanitation sector

217

4.1 Postgraduate studies The training is a formal type that aims at providing advanced level training at MSc degree level for people with interest in the water supply and sanitation sector and related fields. This group of trainees takes up top positions in agencies such as Community Water and Sanitation Agency and Ghana Water Company Limited. Besides young BSc holders, the programme is also designed to cater for professionals with industrial experience who want to improve their analytical skills and enhance their career skills and professional expertise. The table below gives the statistics of students who have graduated from the official MSc programme since its commencement in 1997. To our satisfaction, most of the graduates are working in responsible positions to implement the techniques they have acquired from their study. The major challenges facing potential participants of the programme include: – Employers not willing to release employees for two years – Employers not willing to grant study leave with pay – Duration of two years seems too long for some participants – Discontinuation of scholarship for foreign students. To help overcome some of the challenges above we consider using distance and electronic learning and modularised programmes. Scholarships for foreign students may be difficult to obtain but through our partnerships schemes, these may be possible (Odai et al, 2004). 4.2 Short/refresher courses Besides the MSc programme, the project develops short and refresher courses to enrich and refresh the knowledge of professionals already working or engaged in projects in the sector. Several topics have been treated since the establishment of the project. The topics are selected based on needs assessment in the sector through interviews and informal discussions. The target group includes staff from the organizations mentioned above in addition to consulting firms and contractors. Topics of the short courses ran include: – Operation and maintenance of water distribution systems (with computer applications) – Solid waste management – Water treatment technologies – Wastewater handling, management and disposal – Water quality analysis

Table 1. Statistics of graduates. Year Total Ghanaians 1999 2000 2001 2002

7 6 15 15

Foreigners 7 6 13 12

0 0 2 3

Water resources of arid areas

2003 Total

10 53

218

9 47

1 6

Table 2. Statistics of short courses (since 1999). Year Number of short Total number of courses participants 1999 2 2000 6 2001 3

17 50 27

– Groundwater flow pollution modelling – Public private participation in the water sector. It has been realized that most organizations that benefit from our training have more and interconnected needs, hence the most likely way forward is to develop tailor-made programmes for them, which will meet their needs holistically, rather than individual courses. The other reason being, marketing of the courses are becoming more demanding. 4.3 Tailor-made training The tailor-made courses are unique because they are usually developed through interactive and feedback discussions, in order to meet the requirements of the client institution. Target groups include technicians and professionals from the district assemblies, civil servants from the ministries, NGOs and rural water systems operators. Professionals from these agencies are given these specialised trainings to improve their performance. The WSESP is presently running two tailor-made programmes for sector agencies. They are municipal engineering course and small towns’ water supply operators’ course. Municipal engineering course: This course is organized with the primary aim of upgrading the technical and managerial capacity of the District Assemblies within the Ministry of Local Government and Rural Development (KNUST, 2002b). The personnel upon completion of the course are required to be able to do the following in addition to other assignments. – Plan their needs in infrastructure and resources, – Operate and maintain infrastructure, – Estimates materials and manpower needs for the construction of infrastructure, etc. The course is organised for duration of four weeks and the courses offered are – Planning and Management – Refuse management – Excreta disposal – Drainage – Infrastructure – Water supply – Roads and Highway – Electricity Service

KNUST experiences in capacity building in the water and sanitation sector

219

– Telecommunication – Structural Building works

Table 3. Statistics of municipal engineering trainings. Year

Total no. of participants

February 2003 August 2003 January 2004 Total so far

10 23 15 48

Table 4. Small towns’ water systems trainings. Year

Total

August 2003 February 2004 Total so far

145 38 183

Small towns’ water supply operators’ course: This course is organised with the primary aim of upgrading the technical and managerial capacity of water board members and the system operators to enable them achieve sustainability. Needs assessment showed that there are more than 2000 people in the 210 districts in Ghana needing the first round training. The objectives of the training are (KNUST, 2003) – To update the knowledge, technical and management skill, and attitude of operating staff of small town water systems for effective and efficient operation and maintenance, and management of the water systems. – To train water and sanitation board members to effectively oversee and manage the water system and deal with consumer complaints and requests satisfactorily. The global objective is to attain sustainability, which will subsequently lead to improved health conditions and productivity. The attendance statistics is shown in the table below, and it is impressing to note that the number of participants has increased since we started the tailor-made courses. The course is organised for duration of one week and the courses offered are – Water supply system operation and maintenance – Borehole pumping system operation and maintenance – Managing system information – Water and health – Roles of the water board – Budgeting and tariff setting.

Water resources of arid areas

220

5 CONCLUSION In Ghana, statistics show that there is low capacity in the water and sanitation sector. Our experience in capacity building shows that organizations are aware of their need but sometimes they need help to articulate such needs. The capacity building programme of sector professionals and low-level personnel will be strengthened, and sustainability enhanced. It is anticipated that the 210 districts in Ghana will eventually benefit from our programmes. The department is gradually shifting to tailor-made programmes since the short courses do not proof to be financially sustainable and looking for clients who will patronise the course places extra demands on us. The recent thinking of tailor-made programmes is catching on since there is usually money available for particular organizations to build capacity. We therefore develop programmes, which we discuss with donors and such organizations; upon approval of the courses we then prepare teaching materials and the cost estimates. This approach seems to becoming popular with us because of the high response of participants and the monetary value. Thus we can ensure sustainability in this process of capacity building in the water and sanitation sector in Ghana. REFERENCES Ghana Government, 2003. Ghana’s poverty reduction strategy. National Development Planning Commission, Accra, Ghana. 112–113. KNUST, 2002a. Brochure for MSc programme in Water Supply and Environmental Sanitation. KNUST, Kumasi, Ghana. KNUST, 2003. Evaluation report on operation and maintenance. Department of Civil Engineering, KNUST, Kumasi, Ghana. KNUST, 2002b. Proposal for municipal engineering and infrastructure management course. Department of Civil Engineering, KNUST, Kumasi, Ghana. Odai, S.N., Andam, K.A., & Trifunovic, N. 2004. Strategic partnerships for sustainable water education and research in developing countries. Proc. Int. Conf. on Water Resources of Arid and Semi Arid Regions of Africa, Gaborone Botswana, 3–6 August 2004.

Strategic partnerships for sustainable water education and research in developing countries S.N.Odai Department of Civil Engineering, KNUST, Kumasi, Ghana K.A.Andam Vice-Chancellor, KNUST, Kumasi, Ghana N.Trifunovic UNESCO-IHE, Delft, The Netherlands Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The past decade has seen tremendous investment in water and sanitation for both the urban and the rural communities in developing countries. For effective and sustainable implementation of these projects, local capacity is required. The training of such professionals is becoming increasingly expensive for governments of developing countries, in addition to the fact that there are few institutions offering such programmes. The Kwame Nkrumah University of Science and Technology (KNUST) realising the modern trend in water education and research has entered into strategic partnership with some twenty other institutions around the world to jointly train professionals in the sector; and other external support agencies for financial support. This strategic partnership for water education and research is seen as one of the major emerging options for water education and research around the world. This paper looks at the benefits and challenges of such global partnerships for developing countries.

1 INTRODUCTION The development of the water and sanitation sector is seen as a crucial process to the successful control and eradication of communicable diseases in general. Increasingly, more governments are realising that improvement in the water supply and environmental sanitation sector has a direct positive impact on public health. There are stories about communities where charging for the provision of potable water caused the inhabitants to go back to other sources, e.g. nearby streams, which are not potable. This practice resulted in some persons suffering from diarrhoea, while others suffered from guinea

Water resources of arid areas

222

worm infestations. These are some of the challenges encountered by governments of developing countries. As a result of these attitudes, the government, external support agencies and policy makers have begun directing attention to issues related to the sector in an effort to improve the overall health of the people (Ghana Government, 2003). To achieve this noble objective, water education and research is being enhanced at the tertiary institutions. Water education and research is becoming increasingly diverse and subsequently, more expensive for most developing countries around the world. Thus providing the needed education and research in the sector requires more input which is not accessible locally in one country, hence the need to enter into strategic partnership with other institutions to complement each other and to share experiences. In addition, it is necessary to collaborate with external support agencies working in the sector for financial assistance. In general, the financial base for service provision is quite weak for most institutions. For example, the Department of Civil Engineering (DCE) of the Kwame Nkrumah University of Science and Technology (KNUST) depends on subventions from central government in the provision of its services. However, this is not adequate for postgraduate education in water and sanitation, hence the need to collaborate with agencies who can support us financially. The department organises short courses for sector professionals and offers consultancy services to make additional income. However, as mentioned above these efforts must be augmented for sustainable delivery of the services of the department. Current trends encourage the need to enter into strategic partnerships with institutions and agencies to enhance the sustainable delivery of our services. Generally, in a large number of developing countries there are still only a few people and institutions that have sufficient knowledge to solve the complex technological as well as institutional sector problems of concern. As indicated by different groups of professionals on many occasions, the major causes are (UNESCO-IHE, 2002): ● Lack of sufficient capacity provided by qualified professional staff; ● Lack of highly developed knowledge centres; ● Lack of sufficient institutional and governance qualities and capabilities to ensure an integrated approach to sector problems; ● Lack of ‘communities of practice’ through which professionals and institutions can effectively exchange information, knowledge, experience and good practices.

2 STRATEGIC PARTNERSHIPS To overcome some of these challenges, the DCE has entered into partnerships with several institutions and external support agencies for water education and research. This is in recognition of the modern trend of education, which takes advantage of virtual classrooms and uses expertise around the world as guest lecturers.

Strategic partnerships for sustainable water education and research in developing countries

223

2.1 Partnership with WANet The West Africa Network for Capacity Building in IWRM (WANet) was established in June 2002 for the purpose of training Policy/Decision Makers, IWRM Professionals, and Technicians. Three institutions, namely, the KNUST of Ghana, the National Water Research Institute of Nigeria, and EIER/ETSHER of Burkina Faso are involved. The various training activities in IWRM proposed for these identified target groups are presented below. 2.1.1 Policy/decision makers This group is made up of professionals, executives and, bureaucrats who are involved in policy-making and implementation of IWRM programmes. Specifically, the group includes government policy makers (political leaders, ministers of water, agriculture, environment etc, Governors, Mayors), and senior managers of IWRM-related institutions. 2.1.2 IWRM professionals This group is also quite broad, made up of Water Experts/Consultants, Future Managers in the Water/IWRM domain, and Trainers/Educationalists. IWRM consultants include professionals with diverse specialisations in water management and its use. These include Planners, Engineers, Economists, Social Scientists, Agricultural Workers and Ecologists (Donkor and Nyarko, 2002). The Future Managers sub-group will concentrate on students at the postgraduate level specialising in IWRM-related subjects. Trainers/Educationalists include academic staff of universities in the field of IWRM and trainers and professionals at training institutes. 2.1.3 Technicians This group is made up of those charged with the operation and maintenance of facilities used in the direct provision of IWRM services such as water treatment plants, maintenance of water supply system, etc. Cap-net helped establish this partnership and they continue to support WANet financially for its activities in getting the institutions involved to build the capacity and the materials they need to kick-start their trainings. WANet is however faced with the challenge of offering training in English and French. 2.2 Partnership with UK institutions through British Council British Council (BC) offers opportunities for linkage between universities in UK and other countries. The British Council wholly provides the budget for these links. The activities under such links include staff exchange between the two universities, equipment purchase, research and sometimes students’ exchange. Currently, the DCE has two of such links with two universities in the UK, under different themes. There is one with the University of Newcastle upon Tyne under the

Water resources of arid areas

224

theme “shelter and sanitation for the homeless” (1999–2005), and the other with the University of Bristol under the theme “sustainable water delivery for the poor” (2003– 2006). These links are established based on the mutual consent and interest of the two institutions involved since they will have to make their expertise and facilities available for use by the other institution. Under the link arrangements, staff from UK institutions may serve as lecturers in Ghana, with the cost borne by BC. On the other hand, staff from Ghana benefit from sitting in some courses, having discussions with experts, having access to their library materials and electronic journals. The staff exchange grants staff from both institutions the mutual benefit of learning of new areas of research. Ghanaian students on exchange to a UK institution benefit from having access to facilities for research, while students from UK benefit from best practices in Ghana. Over 50 Ghanaian students and 40 British students have benefited from this partnership exchanges. This students exchange programme has been at the BSc level, and recently extended to PhD students. MSc students are yet to be included in the programme. 2.3 Partnership with DANIDA DANIDA is a bilateral agency, very active in the water and sanitation sector in Ghana. The agency supports Ghana in physical and institutional developments in the water and sanitation sector. They have already invested millions of euros in the sector. They have been working with the Water Resources Commission, Water Research Institute, Hydrological Service Department, and the Meteorological Service Department for data collection for water resources management. In December 2002, the DCE was invited to attend a stakeholders’ workshop at which it presented its strategic plan for education and training in the sector. The department’s strategic plan included networking with sector agencies for education and research, undertaking problem-oriented research to help develop the country. The result of that workshop led to the signing of agreement between the DCE and DANIDA. DANIDA has since been supporting the DCE with scholarships to train MSc students in water resources engineering and management, and providing funds for external examiners from any part of the world. This partnership is making financial resources available for the department to enhance research and education. The department also now has a better relationship with the above-mentioned agencies for easier access to data. These agencies can also now fall on the department for consultation in research or capacity building. 2.4 Partnership for Water Education and Research (PoWER) The Partnership for Water Education and Research (POWER) was founded in November 2002 by 17 institutions around the world (from countries in Africa, Asia, Europe, and South America). UNESCO-IHE has been in the center of all these activities. The main objectives of the partnership are: ● To develop a sustainable and mutually beneficial global partnership in water education and research between UNESCO-IE and regional collaboration centres that promotes life-long learning through generation and sharing of knowledge in integrated and sustainable water and environmental systems relevant to the developing world.

Strategic partnerships for sustainable water education and research in developing countries

225

● To combine the strengths of all partners and enhance the capacity of each partner in order to produce joint products, such as, deliver capable professionals in the sector, find innovative solutions for sector challenges, and build institutional capacity for better efficiency. In the process of combining strengths and levelling the capabilities of the individual partners, joint products in the field of education, training, communities of practice, staff exchange, and collaborative research will be developed in a multidisciplinary manner. These shall be demand-responsive and duly accredited. 3 BENEFITS OF SUCH PARTNERSHIPS The list of the benefits of such partnerships is endless. Some of them are ● Combining expertise from various institutions to do research and publish papers ● Combining strengths to prepare lecture materials ● For very expensive experiments, if one institute has the equipment the other institutions can have access to them for their work ● There is leadership in research, since one institute may have all strength in a particular area; and the other institution may depend on such an institution for direction ● Mutual benefit of learning of new areas of research ● Encourages distance and electronic learning ● Financial assistance usually available for research and training ● Knowledge from the north is made available to the south ● Best practices developed in the south are made available to the north.

4 CHALLENGES OF SUCH PARTNERSHIPS The challenges that come with such partnerships are numerous but not destructive. Some of them are mentioned below. ● Each institution must strive to attain excellence and international recognition/accreditation ● Strive to stay modern by improving ICT equipment and providing or having access to video-and tele-conferencing facilities ● Make your strengths available for other institution to benefit from ● Source funding for the partnership ● Difference in languages of present and potential partners. The present partnership is expected to grow into an international collaboration within which partners will complement each other’s effort as done in the aviation industry, e.g., the alliance between KLM, Northwest and Kenya Airways.

Water resources of arid areas

226

5 CONCLUSIONS Strategic partnerships for sustainable water education and research in developing countries has come to stay, because modern technologies are available for easy communication and sharing of knowledge. It is also getting more expensive to maintain all the facilities and the internationally renowned experts in every institution, therefore partners will share their facilities and expertise to help reduce cost of education in water. Partner institutions are also continually challenged to stay abreast with modern trends in water education. REFERENCES Donkor, E. & Nyarko, K.B. 2002. Establishing nodal resource center in West Africa for capacity building in integrated water resources management. Department of Civil Engineering, KNUST, Kumasi, Ghana. Ghana Government 2003. Ghana poverty reduction strategy. National Development Planning Commission, Accra, Ghana: 112–113. KNUST 2002. Brochure for MSc programme in Water Supply and Environmental Sanitation. KNUST, Kumasi, Ghana. Odai, S.N., Anyemedu, F.O.K., Oduro-Kwarteng, S. & Nyarko, K.B. 2004. KNUST experiences in capacity building in the water and sanitation sector. Proc. Int. Conf. on Water Resources of Arid and Semi Arid Regions of Africa, Gaborone Botswana, 3–6 August 2004. UNESCO-IHE 2002. PoWER and knowledge for sustainable development. http://www.ihe.nl/power/%20knowledge.htm

Assessing demand for clean and safe domestic water in eastern Zimbabwe E.Manzungu, M.Machingambi & R.Machiridza Department of Soil Science and Agricultural Engineering, University of Zimbabwe, Harare, Zimbabwe Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: This paper assesses rural people’s demand for clean and safe domestic water in two districts in eastern Zimbabwe. It explores the role played by physical, socio-economic and cultural factors in influencing rural people’s willingness to pay for domestic water. A semistructured questionnaire was administered in January 2002 to representatives of some 239 randomly selected households in Chitakatira and Nyanyadzi wards. Willingness to pay for clean and safe domestic water was found to be influenced by the availability of other water sources, rainfall received in the catchment area, perceived safety of a water source, age and occupation of respondents. Season and gender did not significantly affect respondents’ willingness to pay. Demographic characteristics influenced willingness to pay for clean and safe water, which should be taken into account when implementing cost recovery policies in the domestic water sector. For completeness, effective demand, best illustrated by ability to pay, should be determined.

1 INTRODUCTION In 1996 it was estimated that 2.5 million Zimbabweans had no access to safe water (Chenje and Johnson, 1996). The situation was worse in the rural areas where only 64% of the population had access to safe water compared to 99% in the urban areas. A recent survey confirmed that water in the rural areas was largely unsuitable for human consumption due to bacterial contamination (Moyo and Mtetwa, 2000). The situation has deteriorated in the last 5 years because of severe economic problems, worsened by the withdrawal of donor support. Since independence in 1980 the donor community has heavily financed Zimbabwe’s Rural Water and Sanitation Programmes. Reduction in funding has resulted in poor maintenance of water supply facilities forcing rural communities to revert back to unsafe water sources (NAC, 1997).

Water resources of arid areas

228

Local communities are increasingly being called upon to contribute in cash and kind to the operation and maintenance of the domestic water sources, a development linked to the Economic Structural Adjustment Programme that the Government of Zimbabwe started in 1991. This World Bank/International Monetary Fund-supported programme advocated for cost recovery in many areas including social services. Provision of domestic water was not spared. For example, the enactment of both the Water Act and the Zimbabwe National Water Authority Act in 1998 incorporated policies like cost recovery and economic water pricing. This signified a policy shift towards the concept of treating water as an economic good (Manzungu, 2001). However the concept, borrowed from the international community, as captured in the World Water Vision (Cosgrove and Rijsberman, 2000), is characterised by a number of inconsistencies (Savenije and van der Zaag, 2002). In a country like Zimbabwe, where 75% of the rural population is regarded as poor, (GOZ, 1995) there are legitimate grounds to ask whether such a policy best serves the population. A recent survey found that, at the local level, there were mixed signals regarding people’s willingness to pay for water (Machingambi and Manzungu, 2003). Respondents wanted the cost of water point establishment and repairs shared between the community (69%), the government (11%) and the donors (5%). Sixty-three percent of the respondents wanted the government to take the responsibility of establishing water points. Close to half (43.9%) indicated that they had individually contributed towards the establishment of the water points they were currently using. There was also a willingness to participate in the maintenance of most domestic water sources except in the piped water scheme apparently because of the high costs involved. The question is: Does this willingness to participate in operation and maintenance of domestic water facilities translate to a demand for clean and safe water by rural communities in Zimbabwe? This study sought to determine whether there was a demand for clean and safe water among the rural people by assessing their willingness to pay for domestic water in the Lower Odzi subcatchment in Chimanimani and Mutare districts in eastern Zimbabwe. In many respects this area typifies most rural areas in the country. In the study the demand for safe and clean water was assessed using the contingent value method (Pearce, Markandya and Barbier, 1989). The method is based on eliciting, from respondents, valuations/bids, which to some extent reflect the strength/ depth of feeling i.e. degree of concern about access to clean and safe water on the basis of a hypothetical market. The hypothetical market is taken to include, not just the good itself, but also the institutional context in which it would be provided, and the way in which it would be financed. The respondent is asked to indicate whether or not they would be willing to pay (WTP) a “starting-point bid/price (SPP)”. An iterative procedure then follows: the SPP is increased to determine whether or not the respondent would still be willing to pay the increment in the price. The last accepted bid, then, is the “maximum willingness to pay (MWTP)”. Besides WTP, respondents were asked about their ability to pay for clean and safe water.

Assessing demand for clean and safe domestic water in eastern Zimbabwe

229

2 MATERIALS AND METHODS 2.1 The study area Chitakatira ward in Mutare district, and Nyanyadzi ward in Chimanimani district in eastern Zimbabwe, were selected as the study areas because of the different rainfall amounts received by the two areas. Water availability was hypothesised to influence demand for water. Nyanyadzi is located in natural region V that receives less than 650mm per annum. Chitakatira falls in natural region III where annual rainfall amounts of 680–800mm are received (Vincent and Thomas, 1960). The amount and distribution of rainfall received in region V is less reliable than that received in region II. The ward was used as a sampling unit since it is the government’s planning and administrative unit. Six villages, each made up of 100 homes, normally constitute a ward. A ward is therefore made up of approximately 600 homes (Makumbe, 1996). Nyanyadzi ward had a total of 832 households (CSO, 1992). The Nyanyadzi Rural Service Centre, which falls within the ward, is supplied with piped water as well as some of the surrounding villages. An irrigation scheme near the service centre also constituted a domestic water source for some respondents. According to the 1992 census there were 1224 households in Chitakatira (CSO, 1992). It also has a rural service centre, and villages, supplied with piped water. The water is drawn from Zimunya dam. It is treated by the national water utility, Zimbabwe National Water Authority (ZINWA). 2.2 Data collection and analysis Respondents were drawn from villages with and without piped water, and from the nearby rural service centres. A semi-structured questionnaire was administered in January 2002 to a total of 239 randomly selected households from the two districts. The questionnaire was administered to one respondent per household who was either the head of the household or a representative. A total of 118 people (82 males and 36 females) and 121 people (70 males and 51 females), were interviewed in Nyanyadzi and Chitakatira respectively. In Nyanyadzi a total of 83 respondents with access to piped water were interviewed compared to 100 in Chitakatira. The questionnaire sought to obtain answers to the role played by physical, socio-economic and cultural factors in influencing respondents’ willingness to pay for water from boreholes, deep and shallow wells, rivers, dams, canals and water taps. Informal interviews were also conducted with ZINWA, Department of Agricultural Technical and Extension Services (AGRITEX) now split into the Department of Agricultural Research and Extension (AREX) and the Department of Agricultural Engineering (DAE) and local government officials. Data was analysed using the Statistical Package for Social Sciences (SPSS) for windows version 10. Descriptive statistics were run for different variables so as to obtain the frequencies, means and cross tabulations. Further analysis was done to determine the effect of different variables on the SPP and MWTP using analysis of variance (ANOVA). The Levene test was used to assess whether the used ANOVA model had a good fit. To validate the ANOVA results, non-parametric tests were done. The Mann-Whitney test

Water resources of arid areas

230

was used to determine which of the factors; season, age, region, gender, occupation and access to piped water affected respondents’ WTP regarding the establishment and repairs of different water points. The Moses test was used to determine whether the observed variation between variables was due to the influence of some of the aforementioned factors. The Kruskal-Wallis test was used to assess if the means of the variables in each category were the same at 5% significance level. 3 RESULTS In some cases, there were no significant differences in the observations between the two wards that were studied. For this reason some data is presented in a collated form. Where there were differences the data is presented separately for the two wards. 3.1 Profile of respondents The profile of the respondents was hypothesised to have an effect on the demand for clean and safe water. Table 1 presents the age distribution of the respondents. At least 70% of the respondents from both wards were communal smallholder farmers while 5% of the respondents from each of the wards were government officials. The proportion of student respondents in Chitakatira and Nyanyadzi was 2% and 5% respectively. Nyanyadzi also had traditional healers among its respondents. 3.2 Conceptions of water Table 2 shows the various uses to which water was put as well as whether it was regarded as primary or commercial water use. Water was classified as primary when it was regarded as a basic need i.e. when used for daily subsistence requirements for most households. Commercial water use, for which people were supposed to pay, according to the respondents, was any use that was not regarded as primary, especially if the use generated a financial income. Most of the water uses were classified as primary.

Table 1. Age distribution of respondents. Age (years) 15–30 31–45 46–60 61–75 >75

Number of respondents Chitakatira Nyanyadzi (n=121) (n=181) 29 46 34 10 1

22 42 32 17 3

Respondents did not make a distinction between clean and safe water although the general perception was that clean and safe water was free of bacteria. Table 3 gives perceived characteristics of clean and safe water.

Assessing demand for clean and safe domestic water in eastern Zimbabwe

231

The perception of whether water was clean or safe was influenced by the source of the water and the season in which that water was used as shown in Table 4. In the wet season, across all water sources, water was perceived to be unsafe. 3.3 Willingness and ability to pay Eighty-two percent of the respondents felt that primary water use should be accessed free of charge, and 79.5% of the respondents felt that people had to pay for commercial use of water. Some 2.1% of the respondents felt that water had to be paid for to enable the maintenance of

Table 2. Patterns and classification of water uses. Water use Drinking Cooking Bathing Laundry Irrigating gardens Livestock watering Brick making Irrigating plots

Classification (%) % citing Primary Commercial water use 99.2 99.2 99.2 99.2 80.8

98.3 54.4 98.7 98 70.3

0.8 0.4 0.4 0.8 10.9

31

16.7

14.3

23.4 25.9

9.2 7.9

14.6 18

Table 3. Characteristics of clean and safe water. Characteristic % attributing % attributing it it safe clean Clear Bacteria free Chlorinated Protected Tasty Not rusty Not dirty Piped Treated

43.1 34.3 3.8 9.6 1.3 0.4 0.4 0.4 0.8

38.1 36.4 3.8 9.6 1.3 0.4 0 0 2.9

Water resources of arid areas

232

Table 4. Perception of water quality of various sources across seasons. Water quality in Water quality in summer (%) winter (%) Water Safe Not Clean Not Safe Not Clean Not source safe clean safe clean Borehole Deep well Shallow well River Dam Canals and water taps

66.1 1.3 6.7 2.1

67 6.7

0.4 53.1 1.7 2.1 3.3 1.3

53.6 3.3

1.3 1.3

0.4 2.1

0.4

2.1 0.4 1.3

0.4

1.3

4.2 31.8 3.3 5 25.5 1.7

5.4 30.5 3.3 12.6 2.5 5.9 2.9 3.3 25.3 1.7 16.7 1.7

4.6 11.3 2.1 4.2 16.7 1.7

infrastructure. About 80% of the respondents said it was the responsibility of the government to ensure that the communities had enough water. Domestic water provision was said to involve some cost by 28.9% of the respondents. Of these, 4.8% put the costs as ranging between Z$50 and Z$400 (Z$4000=US$1). This was however misleading since the figures coincided with the monthly water bills for respondents, especially those staying at rural service centres. The ability to pay different sums of money for water for the month is as shown in Figure 1. Ability to pay, it should be noted, is a function of affordability. It, however served, as a good indicator of respondents’ demand since it showed the price respondents would want their water supplied at. It was observed that generally the number of respondents decreased with the increase in the amount to be paid. However Z$10 was the most common WTP figure. Ability to pay was linked to the SPP and MWTP, which was affected by a number of factors such as the source of the water, the season and the treatment water was subjected to. 3.4 Factors affecting SPP and MWTP A regression model to find the effect of rainfall received (β1), season (β2), gender (β3), age (β4), occupation (β5) and access to piped water (β6) on the SPP and MWTP was run at 5% significance level. The general model was defined as: SPP/MWTP=a+β1+β2+β3+β4+β5+β6+ (1) The following hypotheses were then formulated to test the assumption of normality on equal variance using the Levene’s test. H0: δ12=δ22=δ32=δ42=δ52=δ62 (2)

Assessing demand for clean and safe domestic water in eastern Zimbabwe

(homogeneous variance) H 1: δ12≠δ22≠δ32≠δ42≠δ52≠δ62 (heterogeneous variance)

233

(3)

Accept H0 if p>0.05 Further analysis was undertaken using non-parametric tests, namely the MannWhitney, the Moses and the Kruskal-Wallis.

Figure 1. Respondents’ ability to pay for domestic water per month. Table 5. Comparison of respondents’ MWTP in different seasons. MWTP range (Z$/bucket—25 litre container) 0 1–20 21–40 41–100 >100 Source S W D S W D S W D S W D S W D of water Shallow 8.8 7.9 7.1 33 32.9 33.9 0.4 0.8 1.3 0.4 0.8 1.3 0 0 0 well

Water resources of arid areas

234

Deep 6.7 5.9 5.4 12.6 22.6 23.4 0.4 2.1 0.8 0 0 0.8 0.4 0 0.4 well Borehole 4.6 5 4.2 34.8 31.8 33.1 1 7 4.3 1.3 2 1 1.7 2.5 0.8 1.3 1 7 River 9.6 8.8 7.9 17.5 18 19.3 0.4 0.4 0.4 0.8 1.3 1.7 0 0 0 Dam 7.9 7.9 6.3 18.7 17.9 18.9 0 0.4 0.4 0.8 1.3 2.1 0.4 0.4 0.4 Piped 2.5 2.5 2.5 48.9 47.6 45 2.5 2.9 4.2 8.2 8.7 9.5 2.1 2.5 2.8 scheme Canals 0.4 0.4 0.4 2.1 2.1 2.1 0 0 0 0.4 0.4 0.4 0 0 0 & water taps Key: S—Summer; W—Winter; D—Drought; MWTP— Maximum willingness to pay.

Rainfall: There was a significant effect of rainfall received in a particular area on the demand for clean and safe water. Chitakatira generally had lower SPP and MWTP values compared to Nyanyadzi. Water was therefore perceived to be more valuable in the generally drier Nyanyadzi than in the wetter Chitakatira. The ANOVA test on the means showed that except for the SPP and MWTP to repair broken down pipes, all other variables were significantly affected by the amount of rainfall received. The MannWhitney test confirmed that the amount of rainfall had an effect on all variables except the SPP for repairing a broken down pipe. The Moses test showed that all variables, except the SPP and MWTP for establishing a new borehole, had major differences within them because of the rainfall factor. The amount of rainfall received in a particular location can therefore be said to influence respondents’ WTP. Source of water: Water from boreholes, piped water schemes, canals and water taps, shallow and deep wells had a modal MWTP figure of Z$10 whilst that of dam and river water was Z$0. Piped water was the most popular water source followed by borehole, shallow well, deep well, river, dam, canals and water taps. Season: The measures of association between season and the different variables were also very small (ranging between 0.015–0.098) confirming minimal association between season and the SPP and MWTP values. However, drought had the highest mean followed by winter and summer. Table 5 shows the distribution of respondents’ MWTP across the seasons. Water treatment: Treatment of water marginally changed the proportion of respondents willing to pay more than Z$40 across all the water sources, especially for piped, river and dam water. When treatment was factored in, the modal MWTP figures remained the same for the respective water points although the MWTP figures rose to Z$999. The proportion of respondents willing to pay Z$40 or less actually increased for water sources such as the piped scheme and river, whilst it decreased in the case of water from boreholes, shallow and deep wells. The modal MWTP figures for different water sources did not change. The distribution of respondents’ MWTP with and without treatment is shown in Tables 6 and 7. Gender: The ANOVA, Mann-Whitney and Moses tests revealed that gender did not significantly affect the SPP and MWTP values of the respondents. Except for the SPP and MWTP of establishing a new borehole and repairing it, all other variables showed that males had higher mean WTP figures than their female counterparts.

Assessing demand for clean and safe domestic water in eastern Zimbabwe

235

Age: The means for the different variables were found to have small variances with the 30–45 and 60–75 year age groups having the highest SPP and MWTP figures whilst the dependent (15–30 and 75+ year) age groups had the lowest values. ANOVA tests showed that the SPP and MWTP for a new deep well were affected by age. The KruskalWallis test showed that the SPP for a new water source, as well as the SPP and MWTP for the repair of a borehole, were different for

Table 6. Comparison of respondents’ MWTP for treated water. Treated water—MWTP range (Z$/bucket—25 litre container) Source of 0 1–20 21–40 41–100 >100 water Shallow well Deep well Borehole River Dam Piped scheme Canals & water taps

6.7 32.9

2.5

1.3

0.4

5.4 4.6 7.1 5.9 2.9

2.1 3.7 0.8 0.8 5.1

1.7 2.5 2.1 2.1 9.2

0.4 1.3 0.4 0.4 2.4

0

0.4

0

22.1 32.5 20.6 19.8 44.7

0.4 2.5

Table 7. Comparison of respondents’ MWTP for untreated water. Untreated water—MWTP range (Z$/bucket—25 litre container) Source of 0 1–20 21–40 41–100 >100 water Shallow 9.2 32.5 well Deep well 6.7 23.4 Borehole 5 34.2 River 8.8 18.4 Dam 7.1 19.2 Piped 5.4 45.5 scheme Canals & 0.4 2.1 water taps

0.8

0.4

0

1.7 1.7 0.8 0 3.7

2.1 0.8 0.4 0.8 7.8

0.4 1.7 0 0.4 1.2

0

0.4

0

respondents with the different ages. It can therefore be concluded that age affected these variables particularly the 15–30 and 75+ year age groups. Occupation: Traditional leaders’ MWTP was Z$1 for borehole water, Z$5 for piped water and Z$0 for all the other water points as well as for all repairs. Communal farmers,

Water resources of arid areas

236

local government officials and students had higher SPP/MWTP figures for most variables in decreasing intensity than traditional leaders, Agritex and ZINWA officials. ANOVA tests revealed that the SPP and MWTP for establishing a new deep well, SPP and MWTP for a new borehole, SPP and MWTP for repairing a borehole, and SPP and MWTP for repairing a broken down piped scheme, had significant variances due to occupation of the respondent. There was minimal negative association between occupation and the SPP and MWTP for a new water source, SPP and MWTP for a new borehole, SPP and MWTP for a new deep well, and the SPP and MWTP for borehole repairs. The SPP and MWTP for establishing a new deep well, SPP and MWTP for a new borehole, SPP and MWTP for repairing a borehole and SPP and MWTP for repairing a broken down piped scheme were shown to be significantly affected by occupation according to the Kruskal-Wallis test. Occupation had a weak association with these variables. This could be explained by the fact that more respondents did not have a stable income hence their responses masked those of respondents with stable sources of income. Access to piped water: The Levene test for equality of variances on the impact of access to piped water on respondents’ SPP/MWTP showed that there was homogenous variance, which implied that access to piped water did not affect differences in the SPP/MWTP. However, the Mann-Whitney test showed that access to piped water affected WTP for all water sources although respondents with piped water had lower SPP/MWTP figures compared to those without piped water except for borehole water. In order to establish whether demand for a better water service delivery existed in the communities, respondents were further asked whether they were willing to contribute towards the establishment of a new water source that would save women time compared to an old source, new borehole and deep well. Investment in a new water source for women’s needs: Respondents’ SPP and MWTP were not much different for a new water source that would save women time compared to an old source across the seasons. However, during drought the SPP/MWTP figures were higher than in winter and summer. The Levene test showed that variances in the SPP and MWTP values observed were due to the effect of the amount of rainfall received. Parameter estimates showed that Chitakatira had lower SPP and MWTP values than Nyanyadzi. Agritex/ZINWA officials, traditional leaders and local government officials had a decreasing effect that is lower SPP and MWTP values compared to communal farmers and students. Gender did not affect the respondents’ SPP and MWTP values although males had higher SPP and MWTP figures than females. The Mann-Whitney test showed that access to piped water affected the SPP/MWTP. Respondents with piped water had higher SPP/MWTP than their counterparts without. For the establishment of a new borehole, Chitakatira had lower WTP figures compared to Nyanyadzi, which decreased the SPP and MWTP values. Traditional leaders had lowest WTP figures whilst the Agritex/ZINWA officials had the highest SPP and MWTP figures. Drought had higher SPP and MWTP values followed by winter then summer. In this case females were found to have higher SPP and MWTP values than males. Access to piped water was shown not to affect respondents’ SPP/MWTP for establishing a new borehole. Occupation, age and region had significant effects on the SPP and MWTP for the establishment of a new deep well. Traditional leaders still had a decreasing effect on the SPP and MWTP whilst students had the highest SPP and MWTP values. The 15–30 year

Assessing demand for clean and safe domestic water in eastern Zimbabwe

237

age group had the lowest SPP and MWTP values. Chitakatira had lower WTP values than Nyanyadzi, which lowered the SPP and MWTP values. Gender did not significantly affect the SPP/MWTP values although males had higher SPP and MWTP values than their female counterparts. Season did not have a significant effect on the SPP and MWTP values for the establishment of a new deep well. 4 DISCUSSION The evidence gathered in this study showed interesting perceptions held by rural people in relation to WTP for clean and safe domestic water. At a general level it can be said that there is no substance in the assertion that poor people do not want to pay for water. In Zimbabwe poor people have already begun to meet operational and maintenance costs in domestic water sources (Machingambi and Manzungu, 2003), and in publicly owned irrigation schemes, contrary to claims that the government maintained these schemes (Manzungu, 1999). Worldwide it has been documented that poor people tend to pay the highest amounts for domestic water (Cosgrove and Rijsberman, 2000). Where payments are not forthcoming the problem may be a lack of money rather than willingness to pay (Machingambi and Manzungu, 2003). Poor community mobilisation methodologies may also be another reason (Global Water Partnership, 2000). The study has also provided insights into specific issues concerning the supply of domestic water in rural areas, which may be of interest to policymakers and practitioners. It was clear that there was a high awareness of the potential danger caused by consumption of water containing bacteria. Respondents characterised clean and safe water as being free of bacteria. There was also a realisation of the likely causes of the contamination. This was shown by the fact that water was perceived to be generally unsafe in the wet season (hence lower WTP figures than for the dry season). Piped water had the highest WTP figures as it was rated the safest. River water was rated the most unsafe; it had the largest proportion of respondents not willing to pay anything for it in summer. Treating water had the effect of increasing MWTP figures. In some cases physical scarcity of water also affected SPP and MWTP. This explains why Nyanyadzi, the drier of the two regions, had respondents who were willing to pay higher amounts of money for their water than their counterparts in Chitakatira. The impact of physical scarcity of water on shaping the management of water resources is increasingly being acknowledged internationally. It is not absolute scarcity of water that is a problem but an economic scarcity regarding the availability of finances for the development and management of water resources (IWMI, 2000). This explains the paradox of a country like Zambia with more water resources than South Africa, but has a greater percentage of the population suffering from water scarcity more than the latter. Some commentators have also argued that water scarcity can lead to better adaptive capacities, which may mean the adoption of more intensive water uses (Turton and Ohlsson, 1999). This underlines the importance of analysing the role of social, cultural and economic factors in influencing willingness to pay for water. Socio-economic and cultural factors affected respondents’ WTP for water. The effect of the respondents’ economic circumstances on WTP for water was illustrated by the fact that economically dependent individuals’ (15–30 and the 75+ year age groups) were not

Water resources of arid areas

238

interested in contributing towards their water use. Their WTP was not affected by whether the water was safe or unsafe since they showed no interest in paying for water. Generally traditional leaders were not willing to pay for water because they considered themselves as owners of the water, underlining the role of cultural factors in influencing the demand for water. Perceptions about who owns water also affected WTP (Machingambi and Manzungu, 2003). However safety was a fundamental factor in influencing respondents WTP as even the traditional leaders who were unwilling to pay for water from any other water source wanted to pay for the “safe” piped water. Females were found to have lower WTP than males probably due to the fact that they normally do not handle finances in the home. They therefore tended to be more conservative regarding money issues than the males. However females had low WTP figures in relation to investing in new water sources to reduce labour upon women. The influence of physical, socio-economic and cultural factors on the WTP provided a basis for respondents to portray their degree of concern about access to reliable, safe and clean water, the ideal institutional context in which water could be provided and the way in which it would be financed. Addressing such issues constitutes a more holistic intervention in water issues affecting respondents, rather than merely focusing on cost recovery. 5 CONCLUSION Demographic characteristics of respondents are important in influencing WTP for clean and safe domestic water. Cost recovery policies should therefore be related to demographic characteristics of the intended beneficiaries. While demand for reliable, accessible, clean and safe water was shown to exist in rural areas, success of cost recovery policies depends on the ability to pay. ACKNOWLEDGEMENTS The authors wish to thank the Water Research Fund for Southern Africa (WARFSA) for providing the grant that made the study possible. Mr Chimedza of the University of Zimbabwe’s Department of Statistics is greatly acknowledged for the assistance with data entry and analysis. REFERENCES CSO, 1992. Central Statistics Office, Census 1992, Provincial Profile: Manicaland. Government Printers. Chenje, M. & Johnson, P. (eds.) 1996. Water in Southern Africa, SADC/IUCN/SARDC. Harare, Print Holdings. Cosgrove, W.J. & Rijsberman, F.R. 2000. World Water Vision: Making Water Everybody’s Business. London. Earthscan Publications Ltd. Global Water Partnership, 2000. Towards Water Security: A framework for Action. GWP, Stockholm.

Assessing demand for clean and safe domestic water in eastern Zimbabwe

239

Government of Zimbabwe, 1995. Poverty Assessment Study, Harare, Government Printers. International Water Management Institute, 2000. World Water Supply and Demand: 1995 to 2025. DTP Unit, IWMI—January 2000, Colombo. Machingambi, M. & Manzungu, E. 2003. An evaluation of rural communities’ water use patterns and preparedness to manage domestic water sources in Zimbabwe. Physics and Chemistry of the Earth: Water Demand Management for Sustainable Use of Water Resources, 28(20–27):1039– 1046. Pergamon Press. Makumbe, J.Mw, 1996. Participatory development: the case of Zimbabwe. Harare, University of Zimbabwe Publications. Manzungu, E. 1999. Strategies for smallholder irrigation management in Zimbabwe. PhD Thesis, Wageningen University, The Netherlands. Manzungu, E. 2001. A lost opportunity: The case of the water reform debate in the fourth parliament of Zimbabwe. Zambezia, XXVIII (i):97–119. Moyo, N.A.G. & Mtetwa, S. 2000. Water Quality Management Strategy for Zimbabwe. A paper prepared for the Ministry of Environment and Tourism, Harare. National Action Committee, 1997. Sustainability strategy for the National Rural Water Supply and Sanitation Programme, Government of Zimbabwe, Harare. Pearce, D., Markandya, A. & Barbier, E.B. 1989. Blueprint for a green economy. London, Earthscan Publications. Savenije, H. & van der Zaag, P. 2002. Water as an economic good and demand management: paradigms with pitfalls. Water International, 27(1):98–104. Turton, A.R. & Ohlsson, L. 1999. Water scarcity and social adaptive capacity: Towards an understanding of the social dynamics of managing water scarcity in developing countries. Paper presented at the Workshop on Water and Social Stability at the 9th Stockholm Water Symposium on Urban Stability through Integrated Water-Related Management, hosted Stockholm Water Institute (SIWI), Stockholm, Sweden, 9–12 August. www.soas.ac.uk/Geography/WaterIssues/OccasionalPapers/home.html Vincent, V. & Thomas, R.G. 1960. An agricultural survey of Southern Rhodesia: Part I—agroecological survey. Salisbury, Government Printer.

The role of supplementary irrigation for food production in a semi-arid country—Palestine Mohammed Yousef Sbeih Irrigation Project Coordinator, American Near East Refugee Aid (ANERA), Ramallah, West Bank, Palestine Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Palestine consists of the West Bank and the Gaza Strip. The proclaimed state of Palestine has a land area of 6657km2. Water is considered an essential factor of life and needs to be developed in arid countries. Reuse of treated wastewater for irrigation as supplementary irrigation will increase the irrigated area in Palestine and replace fresh water.

1 INTRODUCTION Palestine consists of the West Bank and the Gaza Strip. The proclaimed state of Palestine has a land area of 6657km2. Water is always considered as an essential factor of life and development in arid and semi-arid countries. In Palestine the total per capita water consumption is 139m3. The total available water for Irrigation is 239 M.C.M. which is responsible for irrigating only 330000 dunums out of 2314.000 dunums cultivated that can be irrigated if water is available i.e. 5% of the total cultivated land. The average rainfall is 450mm and unfortunately there isn’t any water harvesting structures i.e. dams, most of this rainwater flowing towards the Dead Sea or the Mediterranean Sea as waste. So harvesting this water in individual farmer land and using this water for supplementary irrigation to irrigate olive trees, almonds, grapes and cereals will be of a great impact on the Palestinian land for feed production. It should be noted that there are few farmers who practice supplementary irrigation for production of vegetables that are planted in summer as individual initiative. The quantity and quality of production that they have is extremely tangible. Since most of the land in Palestine is planted by olive, grape, and cereals, supplementary irrigation should be introduced and practiced where the production of wheat via irrigation by treated wastewater was three times that under rain fed planting project implemented in a pilot project.

The role of supplementary irrigation for food production in a semi-arid country-Palestine

241

Reuse of treated wastewater for irrigation as supplementary irrigation will increase the irrigated area in Palestine and will replace the fresh water that can be used for domestic purposes. 2 THE NEED FOR SUPPLEMENTARY IRRIGATION IN PALESTINE As it was mentioned before, Palestine is a semi-arid country, where the average rainfall is 450mm. The availability of water is questionable. Furthermore, the availability of water for agriculture is reducing in a tangible way due to the following: 1. The normal increase in growth rate, the population of the country is increasing, so the demand for domestic water is also increasing. This will affect the availability of water for agriculture. 2. Since rainwater is the only source of water, the quantity of rainwater (rainfall) has been decreasing in the recent years. 3. There is a huge conflict on water issues at this stage between the Palestinians and the Israelis since Israel occupied Palestine. It should be mentioned here that during early negotiations in the peace process, four main issues have been delayed since 1992; they are Jerusalem, refugees, water and borders. Still after 8 years of negotiations, there hasn’t been any significant movement on these issues. So the quantity of water that can be available for the Palestinians will probably not be increased. 4. The quality of ground water wells especially in Gaza and Jericho becomes saline and shortly it cannot be safely available for agriculture. From the above, it seems that extra availability of water for additional irrigated area or even to sustain the irrigated area is not an easy task. Total cultivated area in the West Bank is 2100.00 dunums, but the irrigated area is 110000 dunums. From the small experience (pilot project) for this field as well as other country experience i.e. Syria. It has been proven that the production of crops under supplementary irrigation is 3 times higher than under rain fed crop, in addition to the increase in the quality of the product. So if supplementary irrigation has been practiced we can easily increase the production of rained crops to three times or twice. This will play a major role in providing food for the people and even exports can take place and the net income of the country will be increased. 3 BACKGROUND It is foreseen that the world’s food production has to be doubled in the next 25 years, and thus, the agriculture continues to be an important sector in the 21st century. Meanwhile, the agriculture sector remains the largest user of the water resources, and it is evident that there is a decline of agricultural water due to increasing demands from cities, industries, and hydropower utilities in the developing countries such as Asia. Much of the water has to come from irrigation water savings.

Water resources of arid areas

242

Population and economic growth in many developing countries of Asia have created serious problems, such as the shortage of food, the scarcity of water, and the deterioration of the environment. Some of the irrigation and drainage projects have been seriously criticized due to their high-cost and low-efficiency for the construction and maintenance. The concept of maximum yield is now changing to optimum yield for creating an efficient irrigation schedule. The water saving is the most sustainable conservation, because it reduces the new construction needs to meet the increased water demand. The major issues of agricultural water are how to increase withdrawals about 15–20% by water saving, how to increase storages 10–15% by new irrigation facilities, and how to conserve the water quality of irrigation. 4 SUPPLEMENTAL IRRIGATION 4.1 Definition ICARDA defines supplemental irrigation (SI) as; the addition of essentially rain fed crops of small amounts of water during times when rainfall fails to provide sufficient moisture for normal plant growth, in order to improve and stabilize yields. Accordingly, the concept of SI in areas having limited water resources is built on three bases: First: water is applied to rain fed crops, that would normally produce some yield without irrigation; Second: since precipitation is the principal source of moisture for rain fed crops, SI is only applied when precipitation fails to provide essential moisture for improved and stabilized production and; Third: the amount and timing of SI are not meant to provide moisture stress-free conditions rather to provide minimum water during the critical stages of crop growth to ensure optimal instead of maximum yield. The management of supplemental irrigation is seen as a reverse case of full or conventional irrigation (FI). In the latter the principal source of moisture is the fully controlled irrigation water, while the highly variable limited precipitation is only supplementary. Unlike FI the management of SI is dependent on the precipitation as a basic source of water for crops grown. Water resources for supplemental irrigation are mainly surface, but shallow ground water aquifers are being increasingly used lately. Non-conventional water resources are of a potential for the future, but an important one emerging is water harvesting (Dwas 2001). 4.2 Improving production with SI Research results from ICARDA and other institutions in the dry areas as well as harvest from farmers showed substantial increases in crop yields in response to the application of relatively small amounts of supplemental irrigation. This increase covers cases with low as well as high rainfall. Average increases in wheat grain yield under low, medium and high annual rainfall in Tel Hadya reached about 400%, 150% and 30% using amounts of

The role of supplementary irrigation for food production in a semi-arid country-Palestine

243

SI of about 180, 125 and 75mm respectively. Generally, optimal SI amounts range from 75mm to 250mm in areas with annual rainfall between 500 to 250mm, respectively. Determining the optimal amount under various conditions will be discussed later (Oweis 2001). When rainfall is low, more water is needed but the response is greater, but increases in yield are remarkable even when rainfall is as high as 500mm. The response was found to be higher when rain distribution over the season is poor. However, in all rain fed areas of the region it was found that some time in the spring there is usually a period of stress, which threatens, yield levels. This soil moisture stress usually starts in March, April of May, if total annual rainfall received is low, average or high respectively (Oweis 2001). In Syria average wheat yields under rain fed conditions are only 1.25t/ha and this is one of the highest in the region. With SI the average grain yield was up to 3t/ha. In 1996 over 40% of rain fed areas were under SI and over half of the 4 mil tons national production was attributed to this practice. Supplemental irrigation does not only increase yield but also stabilizes farmer’s production. The coefficient of variation in rain fed production in Syria was reduced from 100% to 10$ when SI was practices. This is of special socio-economic importance since it affects farmer’s income (Oweis 2001). 4.2.1 Introduction Historical Palestine is located between the Mediterranean Sea and the Jordan River, as well as to the Red Sea from the south. The present proposed Palestinian state consists of West Bank and Gaza Strip. The other part of Palestine is occupied by Israel in 1948. This study focuses on the West Bank and the Gaza Strip. The proclaimed state of Palestine has a land area of 6657 square kilometres (Kateeb 1993). Population senses has been taken place recently by the Bureau of Statistics early 1998. It is reported that the population of the West Bank is 1571571 and Gaza Strip is 963026 where the total population of the Palestinian people is 2534598 people. Ground water is the main water source in the country. It is recharged by rainfall. Rainfall varies from 100mm in the south east to 800mm in the north. The average rainfall is 550mm (Sbeih 1995). Where the average rainfall in Jordan Valley is from 100mm to 270mm/year (Zaru 1992), and in Gaza is 200–400mm/year (Abu Safieh 1991). Not all the rainwater is available to the Palestinian due to Israeli Military orders. Water is abstracted from the ground water through 340 wells in the West Bank and 1781 wells in Gaza. In addition to that springs contribute a lot, where half of the irrigation water in the West Bank is due to springs. The quality of the available water varies from almost rain water to brackish water. In the Jordan Valley where it is the lowest point in the elevation in the world where temperature is very high in this area especially in summer. As example, the chloride content is reaching 68mg/l and the SAR reaches 11.7 where the TDS reaches 5000PPM. Still the utilization of this saline water is not as efficient and environmentally safe as it should be where further utilization of this water could play a major role in developing the area where still the irrigated area consists of not more than 6% of the cultivated area in the West Bank.

Water resources of arid areas

244

It should be mentioned that not only saline water does already exist and utilized unproperly, but it also seems to be that the additional water that can be allocated for irrigation is also saline water which is going to be from: 1. The Eastern aquifer to be used in Jordan Valley. 2. From the treated waste water from different cities and villages in the West Bank. 4.2.2 Water sources in occupied Palestine 4.2.2.1 West Bank Two main water sources are available for Palestinian in the occupied Palestine (West Bank and Gaza Strip) for agricultural, domestic and industrial use. These are rainfalls and ground water sources—Palestinians consume water mainly through ground water wells and springs (where rainfall is considered the main recharge). The total annual water springs discharge varies according to the rainfall. The total annual flow of the 113 fresh water springs in the West Bank ranges between 24 M.C.M. (as in the year 1978/79) to 119.9 M.C.M. (as in the year 199/92) and with an average of 52.9 M.C.M. as calculated from the annual flow in the past 24 years. Around 86% of the total annual flow of these 113 springs is within the eastern drainage (in/or toward the Jordan Valley), while the other 14% is within the western and south-west (Nusseibeh 1995) where the total estimated annual water discharge from ground wells is 60 M.C.M. (Awartani 1992). So that the total annual water available to Palestinian is 113 M.C.M. In addition to that there is another 2.5 M.C.M. is collected directly from the rainfall in cisterns in Palestinian houses. So that the total available water is 116 M.C.M./year, for more information see Table 1. 4.2.2.2 Gaza Strip Water situation in Gaza Strip is very critical. The Gaza Strip lies on top of two water strata. The upper is fresh water, the lower carries saline water. The annual consumption of water is at present in the vicinity of 100 M.C.M. These aquifers get replenishment of some 60% leaving a deficit of 40 M.C.M. of water (Shawwa 1991). Even the Gaza water is lower in quality than West Bank, but due to the complication of the situation there and due to the geographic location where my work is more in the West Bank. This paper will address West Bank issues more clearly.

Table 1. Basic land and water indicators for Israel and the occupied Palestinian and other Arab territories. West Bank

Gaza Strip

Israel

Total area 5573000 360000 20000000 (dunums) Population (1988) 900000 600000 4300000

The role of supplementary irrigation for food production in a semi-arid country-Palestine

245

Area of land 2100000 214000 4250000 cultivated (dunums) Area of land 110000 120000 850000 irrigated (dunums) Percentage of total 5 56 44 irrigated land (%) Percentage of total 38 59 21 land cultivated (%) 95 80 1320 Annual water consumption for irrigation (million m3) 27 21 325 Annual water consumption for households (million m3) 3 2 125 Annual water consumption for industry (million m3) Total annual water 125 103 1770 consumption (million m3) Total per capita 139 172 411 water consumption (m3) 35 75 Per capita water 30 consumption per household (m3) Per capita water 3.3 3.3 29 consumption for industry (m3) Per capita water 106 133 307 consumption for irrigation (m3) 1 dunum=1000m2. Source: Israeli land and water policies and practices in the occupied Palestinian and Arab territories, unpublished study in Arabic (Economic and Social Commission for Western Asia, Baghdad, 1990), p. 8.

4.2.3 Irrigated areas in the occupied Palestine In Palestine, being a semi arid country, we are confronted by a demographic growth, and agricultural development as well livestock and industrial development. Thus in essential growing water requirement makes the rational management of water resources supremely important in order for development to be lasting and for environment to be served.

Water resources of arid areas

246

On a global basis at least 60% of all water abstracted at present is used for agricultural production. In Palestine 70% of all water consumed is due to agriculture. Here in Palestine, agriculture is considered to be one of the main national income. Agricultural production contributes 47.61% of the total national income in 1970. The potential for irrigation to raise both agricultural productivity and the living standards of the rural poor has long been recognized. Irrigated agriculture occupies approximately 17% of the world’s total available land but the production from this land comprises about 34% of the world total. In Palestine, irrigation is considered to be the spinal chord of plant production for the following reasons: 1. Palestine is considered as a semi arid region where some of the crops cannot be grown without irrigation (example, citrus). 2. In the Jordan Valley, which constitutes the main agricultural production for the country, irrigation is a must due to low rainfall and high temperature. 3. With irrigation the same plot of land can be planted up to three times per year while it cannot be planted more than two times with dry farming. 4. Different varieties and crops can be planted in any region due to the availability of water i.e. more flexibility of planting several crops at different regions in different times of the year. 5. Job creation: Since the labour requirement per irrigated dunum is more than double that of job required per dry farming per one season. This has now become more vital due to continuous of closures of the West Bank and Gaza Strip where the number of labourers that are working in the Palestinian part that occupied in 1948 is sharply reduced. 6. Agricultural production is much higher for irrigated farming than for dry farming per dunum per season. As example average tomato production per dunum is as follows: – Dry farming: 2–3 ton per dunum per season. – Irrigated (open land) 6–8 ton per dunum per season – Irrigated (greenhouse) 12–16 ton per dunum per season 7. Net income per dunum of dry farming does not exceed $150 while from irrigated area the net income can exceed $1500 per dunum. 8. Especially in Palestine, where the horizontal expansion in agriculture by increasing the total cultivated area due to the Israeli occupation, and shortage of water. The vertical expansion could be the main parameter to play with. Irrigation will be the main element in this formula. So that providing extra water for irrigation to irrigate as much as possible of the cultivated area is a must. This implies that Palestinian should use any drop of water. Regardless the quality of that water practically and efficiently: Table 2 shows the irrigated area in each district in Palestine where the total irrigated area in 1993–94 was 217000 dunum (PSBS 1996).

The role of supplementary irrigation for food production in a semi-arid country-Palestine

247

Table 2. Distribution of area that could be irrigated in the West Bank (Source: Awartani, 1991). Location

Dunum

Plains in Jenin and Tulkarem Highland Eastern slopes Jordan Valley Total

99600 27740 64.6 93.5 535.1

4.2.4 Available area that is ready for irrigation Where in Gaza Strip the irrigated area could be doubled or tripled in terms of topographical situation but due to the limitation of the water both quality and quantity it is very difficult to increase the irrigated area while in the West Bank the area that could be irrigated in terms of topographical conditions estimated to be 535 thousand dunums (Awartani 1991) as in Table 2. Where in the study conducted by PWA in 1992 in order to develop a plan for the western Ghore the following locations could be the most suitable area to be ready for irrigation. 4.2.4.1 Northern Ghore The areas suitable for irrigated agriculture in this region include: 18000 dunums in Ein Al Beida, Bardalla villages 5300 dunums in the Ghore 3500 dunums in the Ghore But the Ghore and Zhor areas are mostly closed by the Israeli Military orders. 4.2.4.2 El-Bique Valley This is a large flat area to the west of the hills of northern Ghore. This area includes about 18500 dunums of fertile smooth deep soil. The Palestinian farmers as rainfed excluding 5500 dunums where the two settlements their (Baquat and Roi) are occupying cultivate all this area. 4.2.4.3 Upper El Fara’ Valley area (Semi-Ghore) In this area, there are 13100 dunums that are suitable for irrigation and can be easily irrigated as follows: Sahl Tubas Sahel Tayassear Sahel Tammun Sahel El Fara’ El Nassarieh (additional)

3600 dunums 900 dunums 1900 dunums 5000 dunums 1700 dunums

Water resources of arid areas

248

Where there are another 7000 dunums, which are already irrigated. 4.2.4.4 The middle and south Ghore This region extends from approximately grid north 180 (northern of Marj Najeh) in the north to the Dead Sea in the south and from the Jordan River in the east to the feet of the west-bank mountains. The total area that could be ready for irrigation in this area is 145500 dunums. In summary, the total area that can be used in irrigated agriculture in the western Ghore will be: Northern Ghore Biquia Valley Semi-Ghore Southern Ghore Total

26800 dunums 18500 dunums 201000 dunums 145500 dunums 210900 dunums

Where about 44000 (PCBC 1991) dunums of this area is currently irrigated. So the total additional area that could be irrigated in the West Bank is (210900−44 000+(535100−935000)= 608500 dunums. It should be mentioned that the Jordan Valley produces more than 59% of the vegetables produced in the West Bank. It also produces 100% of the bananas produced in Palestine. 5 PALESTINIAN EXPERIENCE OF SUPPLEMENTARY IRRIGATION Still the term supplementary irrigation is not even used formally and officially in Palestine. Until this time there is not any plan of implementing any project of supplementary irrigation. This is mainly due to the lack of qualified staff at the Ministry of Agriculture as well as to the lack of great interest to agriculture from M.O.A. due to the following reasons: 1. The lack of responsibility of the Palestinian Authority on most of the agricultural land due to the occupation. 2. The lack of finance and funding to development projects. Nevertheless, there are individuals who attempt to use supplementary irrigation, an example of that are few farmers in Sinjel town in the Ramallah area. 6 DESCRIPTION OF AGRICULTURAL AREA IN SINJEL This village is located just between Ramallah and Nablus cities, situated 20km to the north of Ramallah. The total agricultural area in the village excess 4000 dunums, out of these areas. About 1000 dunums are plain and flat.

The role of supplementary irrigation for food production in a semi-arid country-Palestine

249

This 1000 dunums is planted with vegetables in summer and cereals in winter. All of this area is rain fed, there are no source of water for irrigation since this area is located close to the village (houses), it is easy for the farmers to bring water by mobile tanks. Usually the farmers in summer, bring some water and store them in a container (barrel) of 200 liter capacity each, since the ownership of land is between 3–5dunums, the number of barrels used are 6–8. In summer farmers used to plant vegetables, at the time of planting the seedlings, farmers used to irrigate the seedling by a bucket. Farmers used to mix the fertilizer water and irrigation at the time of planting the seedlings. Later on, after 20 days the second irrigation with fertilizer is applied. The third one and the last one are provided with fertilizer before flowering. The total amount of water applied per each plant is not more than 1 liter, for a dunum of 1000 plants, 1000 liter is applied 1 cubic meter of water applied for the whole season per one dunum. While for the irrigated area the minimum irrigation water requirement is 70m3/dunum per the season. In this village, Sinjel, and through my investigation, in the year 2000 I found 3 farmers who are using this approach technology, when I asked one of them what is the result that you will expect, he broadly replied: 1. The quality of agricultural product that I used to obtain for the last two years where I used to use supplementary irrigation is much better than the product of my neighbor in the same plot of land in the village, so the price per 1kg. That I got is much higher also. 2. The total production is much higher than that of my neighbor, i.e. I got 4 tons each per dunum, my neighbor got 2 tons of squash per dunum. 3. The period of production that I have is much bigger than that of my neighbor has, this means that total income that I gained is much higher. I used the produce vegetables for 2 months, while my neighbor only one month, i.e. the harvesting period is much higher when supplementary irrigation used. I informed this farmer that I am working on an irrigation project coordinator for an NGO that provides funds for farmers. Since this farmer believes that he was happy from his production since he has only 3dunums and all of his family working in this plot of land, he did not ask what service that since that we offered, this totally indicated that he is happy, and he did not need any further assistance. At that time there was visiting irrigation professor from Canada. This professor told me that we should use him as a model to encourage people using appropriate technology. Another example of using supplementary irrigation is found in Hebron where a farmer from Al Tamimi family, who has a grape field and luckily a pipe water pass through his field and used to get some water from this pipe and provide some water for his grape. In winter since the rainfall in Hebron is not exceeding 300mm, as well as in July.

Water resources of arid areas

250

Table 3. Results of El Bireh wastewater treatment pilot plant using treated wastewater. Treatment

Production of wheat (anber variety), all the plants (kg/dunum)

Irrigation with treated wastewater with fertilizer Irrigation with treated wastewater without fertlizer Without irrigation, with fertilizer Without irrigation, without fertilizer

2520

20036

1600 572

It is well known in Hebron, that the quality of grape of that man is the best in Hebron, since Hebron is of the biggest producing city (country) in Palestine. Since the municipality constructed a pilot treatment plant, it thought of planting crops using the treated effluent. This was funded by American Near East Refugee Aid (ANERA). Three crops were selected by the Agriculture Department to be planted for the first time in Palestine using treated wastewater: ● Artichokes on 150m2—planted on October 31, 1993. ● Onion frozen production on 500m2—planted on November 6, 1993. ● Wheat on 1000m2—planted on November 22, 1993. ● Drip irrigation as well as sprayers were used. Several treatments were made as follows: ● Irrigation with wastewater used, fertilization was used. ● Same as above, but without application of fertilization. ● Irrigation not used but fertilization was used. ● No irrigation and no fertilization (dry land farming). All the agricultural practices were used (pesticides, ploughing, seed control, etc. Table 3 shows the production of each kind of treatment. The impact of using treated wastewater appears clear. Notes: 1. Time of planting was October 1993; all the crops received rainfall during the growing period. 2. Time of harvesting was June 2, 1994. 3. Production with irrigation with treated wastewater with fertilization was five times without irrigation and fertilization. 4. Production increased the soil when irrigated with treated wastewater where fertilization was applied on both cases (irrigated and non-irrigated).

The role of supplementary irrigation for food production in a semi-arid country-Palestine

251

7 METHODOLOGY OF PRACTISING SUPPLEMENTARY IRRIGATION IN PALESTINE Since the ownership of land is very small in size i.e. from 5–10 dunums, supplementary irrigation can be easily implemented for vegetables, trees and to cereals to some extent constructing of small ponds of 40–50m3 capacity, i.e. this pond can be located on a 14–18 meter square area. This pond can be located on the lowest point in elevation of the individual land. This land serves two farmers if agreed upon where it can be sited on the border of each farmer land. Distributing of water to the plant can be done manually by lifting the water and distributing it to the plants by a bucket. Another way of distributing this water that this water can be lifted manually from the pond and poured into a barrel that can be located on the dip of the pond with 1/2meter raised over the surface so water can be distributed to the plant by gravity through pipe line. The farmer can distribute the water pipe from the plant to another. These methods can be implemented

Table 4. Results of Al Beireh Pilot Wastewater Treatment, 1994.

Crop Wheat 870 type

Kind of treatment

Production (kg/dunum) Seed Hay

Irrigation with 687.5 fertilizer Irrigation 656.70 without fertilizer Rainfed with Rainfed with 537.5 fertilizer fertilizer Rainfed without 500 fertilizer Wheat anber Irrigation with 864 type fertilizer Irrigation 824 without fertilizer Fertilizer Rainfed with 600 fertilizer Rainfed without 236 fertilizer

1375 1373

1187.5 1531.25 1656 1212

1000 336

easily with zero operation cost. Since only the farmer himself can conduct this job easily, another method of distributing water is by using a small pumped electricity is available since the head required is very small. In the case of cereals water can be distributed easily by establishing ponds, so water can be discharged into the farm then water can flow by gravity. In order to reduce the

Water resources of arid areas

252

cost of pumping farms can cooperate between themselves when each farmer can construct his pond on the highest point in elevation on his land. His pond can receive water from his neighbour’s field and so on. 8 THE ECONOMY OF SUPPLEMENTARY To construct a pond of 50m3 the following is needed with estimated costs: 1. 2. 3. 4.

Excavation of 50m3 Construction works Plastering Parallel, pipes, buckets Total estimatated cost:

=$3900 =$2000 =$500 =$120 =$3100

9 REVENUES Assume a plot of land of 5 dunums planted with vegetables. The production of vegetables of rainfed per dunum is 3 tons/dunum, the production of dunum with supplementary irrigation is 4.1 ton. The price per ton is $200 for rainfed crops.

The income per supplementary irrigation is 4×250=$1000.

The price per ton for supplementary irrigation is $250.

The net income due to supplementary irrigation will be 1000?=400 per dunum.

S the income per rainfed dunum=3×200=$600.

5 dunums×400=2000 per session per 5 dunums.

10 CONCLUSION AND RECOMMENDATIONS 1. In Palestine the total cultivated area is 2314000 dunums, while the irrigated area is 230000 dunums, so any efforts for increasing the productivity of the cultivated area should be considered due to the large area, while the production of the irrigated area is on its maximum. 2. Providing of extra water or even to sustain the existing water for both irrigation and domestic purposes is questionable due to the increase demand for domestic purposes first and due to the Palestinian-Israeli water conflict. 3. Practicing supplementary irrigation is not costly and did not need that much complicated technology. 4. The irrigated area only represents 6% of the cultivated area, where the land that can be easily irrigated is estimated to be 608600 dunums. In the West Bank only, which is 6 times the land that is already irrigated but water is needed. 5. The salinity of the ground water is deteriorated by time due to over pumping, sea intrusion and the low rainfall especially in the Jordan Valley and in the Gaza Strip, so providing fresh water for irrigation is questionable.

The role of supplementary irrigation for food production in a semi-arid country-Palestine

253

6. The additional water that will be available for the Palestinians will be either from (a) Eastern aquifer, (b) Jordan River, or (c) Treated wastewater. Where all of this water is saline water, where there are another source such as the mountain aquifers, but this seems to be difficult to be secured soon. 7. The early possible of expansion in irrigation will be in Jordan Valley where the existing water wells and the future water that might be available is saline. 8. Since the treated water is in the full control of the Palestinians, more attention and care should be paid in order to better and safe utilize of this water for developing the agricultural sector in Palestine, and this water can be used for supplementary irrigation. 9. The productivity of one cubic meter of water with supplementary irrigation is much higher than that of irrigated land since the water prepared by irrigated dunum is 7 times more than the required for supplementary irrigation. 10. The existing irrigated area is already exhausted since this land used to be planted two or three times a year where the other land used to be cultivated once a year even it kept fallow on some years. 11. Palestinian Agricultural Ministry and Palestinian Water Authority should recognize the situation and consider supplementary irrigation as a major element for food supply.

REFERENCES Abdul Jabar, Q. 1996. Chemical analysis of Jericho wells. PhD, Jerusalem. Abu Arafeh, A. 1894. Jordan Valley, Jerusalem, published by Arab Studies Society. Al Khateeb, N. Palestinian water supplies and demands. A proposal for the development of a required water master plan, IPC, Jerusalem. Awartani, H. 1991. Irrigated Agriculture in the Occupied Palestinian Territories, Al Najah National Univseristy, Nablus. Awartani, H. 1992. Groundwater wells in the Occupied Palestinian Territories, PhD, Jerusalem. ICID, 2001. General Report to the First Asian Regional Conference of International Commission on Irrigation and drainage. Agricultural, Water and Environment, Seoul, Korea. Nusseibeh, M. & Nasser Eddin, T. 1995. Palestinian Fresh Water Springs, Jerusalem, Palestine. Palestinian Water Authority, 1992, Water Development Plan, Jerusalem, Palestine. Sbeih, M. 1995. Recycling of treated wastewater in Palestine: Urgency, obstacles, and experience to date, Elsevier. Oweis, T. 2001. Supplemental Irrigation for field crops, water saving and increasing water productivity: challenges and options. University of Jordan, Amman, Jordan.

Conversion of priority water rights to proportional water permits and conflict management in the Mupfure river catchment, Zimbabwe Tamsanqa Mpala Scientific & Industrial Research & Development Center (SIRDC), Harare, Zimbabwe Water Resources of Arid Areas–Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The Zimbabwe Water Act of 1976 was repealed and replaced by the Water Act of 1998. The Water Act of 1998 (Chapter 20:24) came into force on the 1st January 2000 and with it came the abolishment of the prior date system of water allocation. In the promulgation of the new Water Act, existing water rights used under the priority date system were to be converted to proportional water permits. The objective was to promote equitable access to water resources for all Zimbabweans and to encourage the sustainable utilization of the resource. The challenge for the newly established catchment councils was therefore to determine all existing water rights in their respective catchments and conduct the conversion exercise based on water generated in the catchment and water use by water right holders. This study analyzed the existing water rights and water generation for the Mswenzi River catchment and conducted the conversion to water permits with the Sanyati Catchment Manager.

1 INTRODUCTION Any resource, such as water, when used by more than one user in a single catchment or river basin, tends to attract conflicts about how it is shared and distributed. This paper focuses to a large extent on a highly committed small river catchment; namely, the Mswenzi River Catchment that forms part of a greater basin, which eventually drains off into the Zambezi River and into the Indian Ocean. The Mswenzi is a tributary of the Shuru Shuru River in the northwestern region of Zimbabwe. The Shuru Shuru River in turn is a tributary of the Mupfure River that drains northwesterly and into the Zambezi River. The study catchment has a total surface area of approximately

Conversion of priority water rights to proportional water permits and conflict management

255

158km2 with the Mswenzi River stretching a distance of 22km. The study catchment generates on average an estimated 70mm of surface water per year, or 11.0*106m3/year. Hence, this paper looks into the operationalization of water rights in the Mswenzi river catchment and gives an analytical recommendation of converting the old water rights used under the Zimbabwe Water Act (1976) to new water permits as recommended under the new Zimbabwe Water Act of 1998. A water right, for the purpose of this study, has been defined as a right to use beneficially a certain volume of water expressed in absolute volumetric units per time unit, whilst a water permit is a permit or allowance for the use of water, which specifies and restricts the use of water allocated. At the same time the paper attempts to bring out the conflicts involved in the Mswenzi catchment between upstream and downstream commercial water users. The paper discusses the legislative policies by looking at the old and new Water Acts and highlights the important principles that will govern conflict management and help spells out recommendations for water authorities and catchment management institutions. 2 THE WATER ACTS In Zimbabwe, the Water Act of 1976, which was repealed by the new Water Act of 1998, vested all public water in the President and private water was water belonging to the owner of land on which it was found. The right to use water was dependent on the type of use. For primary use no right was required. Access of water for non-primary use was based on the prior appropriation doctrine, where an appropriated right was based on the application of the appropriated public water to some beneficial use. The granting of any water right was the exclusive function of the Administrative Court sitting as the Water Court. The right would only be granted if public water was available and if it could be ascertained that the water would be put to beneficial use. The right granted was dependent on the date on which the application for the right was made. This date determined the applicant’s priority in the use of water applied for. The new Zimbabwe Water Act of 1998, which replaces the old Water Act (1976) sought to bring about equal and fair distribution of the available water resources in the national interest for the development of the rural, urban, industrial, mining and agricultural sectors. The major principles of the new Water Act were that all water would now be owned by the State and any use of it other than for purposes of primary use should be approved by the State. All stakeholders should be involved in decision-making processes and contribute to sustainable management of water resources. Water resources would be managed at catchment and sub-catchment levels, and the environment would also be considered a legitimate user of water. One of the important changes of the Water Act that is brought out in this study is that the priority date system of first come first served of water allocation was abolished to enable the principle of equitable access to water and sharing of water at all times. A fractional allocation system is now the recommended allocation system for non-primary water use.

Water resources of arid areas

256

3 THE CONVERSION EXERCISE The Mswenzi River Catchment has an area of approximately 160km2, and on average generates 11.0*106m3/a of blue water. Water generated in the catchment is based on the annual unit runoff and the catchment area, where the water generated (m3/a), is a product of the two. Mean unit runoff for the catchment is 70mm/year and the calculated catchment area is 158.15km2. This gives an amount of 11.0*106m3/a as water generated in the catchment. The Mswenzi River catchment has a total of 17 existing water rights owned by mainly commercial farmers, who have built reservoirs with a total capacity of 7.3*106m3. Among the 17 water rights, the total commitment level of the catchment is 65%. This is water used, (m3/a) as a fraction of water generated (m3/a). The major dams in the catchment are Balwearie and Tawstock dams, which have a combined capacity of 4.9*106m3, and serve a total of 8 water rights to various farmers in the catchment. The other 9 water rights are served from smaller dams along the Mswenzi River. Table 1 shows existing water rights and current users in the catchment as well as the priority date for each property. The procedures for the conversion of water rights to water permits were that, (i) the priority date attached to each water right be removed, (ii) volumes of flow and storage rights as allocated in the old system remain the same and shall be used as permits until such a time as the water authorities see fit to amend or revise the permit, and that (iii) water rights be converted according to the applicants ability to beneficially use the water. In the conversion exercise a simple formula has been recommended for the purposes of the study catchment, which follows: Permit (m3/a), P=[S1+S2+S3…]+F1+F2… Where, S=storage (m3/a) abstraction from stored water. F=flow (m3/a) abstraction from river flow. The storage permit is the volume of water permitted to be stored and used while a flow permit is the volume of water, which can be abstracted and used directly from the river. The validity of the permits, is for a limited period of 20 years, thereafter they are reviewed.

Table 1. Review of existing water rights and current users in the Mswenzi catchment. Water Property right number 7573 8881

Kasama

River Priority Abstraction Abstraction Dam right from right from flow storage (103m3/a) (103m3/a)

Mswenzi 19/10/66 tributary Dodington Mswenzi 06/03/70 tributary

0

45 Weir

0

136 Farm dam

Conversion of priority water rights to proportional water permits and conflict management

10626

12007

Rem. of Mswenzi 31/10/73 Luton Balwearie Mswenzi 31/05/73 Rem. of Mswenzi 05/10/83 Luton Balwearie Mswenzi 01/06/81 Balwearie Mswenzi 14/11/73 Strathspey Mswenzi 28/05/49 Strathspey Mswenzi 23/09/70 Strathspey Mswenzi 24/10/73 Handley Mswenzi 16/01/80 cross Cornucopia Mswenzi 24/10/73 Handley Mswenzi 24/10/73 cross Merchiston Mswenzi 28/01/81

14322

Merchiston Mswenzi 20/10/88

0

14321

Merchiston Mswenzi 29/10/90

0

12014

Merchiston Mswenzi 20/02/81

0

10156 10626 10156 10156 2276 2276 2276 12398 10364 9101

Total Source: Field notes, 2002.

0

910 Balwearie

0 0

1040 Balwearie 520 Balwearie

0 450 0 0 0 0

130 Balwearie Balwearie 400 Tawstock 182 Tawstock 518 Tawstock 1530 Suri Suri canal 523 Tawstock 714 Tawstock

0 0 0

450

257

611 Farm dam 20 Farm dam 5 Farm dam 20 Farm dam 7304

Balwearie Farm, which is served by Balwearie Dam, utilizes an estimated 1.59*106m3/a of water for its various crops under irrigation, hence the permit using the above formula will be: P=[1040*103m3/a+130*103m3/a] +450*103m3/a =1620*103m3/a for a validity of 20 years. Conversion for other properties in the catchment include: Kasama, P=[45*103m3/a] for a validity of 20 years. Dodington, P=[136*103m3/a] for a validity of 20 years. Rem. Of Luton, P=[910*103m3/a+520*103m3/a] =1430*103m3/a for a validity of 20 years. Strathspey, P=[(400+182+518)*103m3/a] =1100*103m3/a for a validity of 20

Water resources of arid areas

258

years. Handley Cross, P=[2244*103m3/a] for a validity of 20 years. Cornucopia, P=[523*103m3/a] for a validity of 20 years. Merchiston, P=[(611+20+5+20)*103m3] =656*103m3 for a validity of 20 years. The new permits as recommended for the study catchment after the conversion exercise is shown in the Table 2. 4 CONFLICT MANAGEMENT IN THE UPSTREAM, DOWNSTREAM CASE The case singles out the dispute between upstream and downstream commercial farmers holding water rights in the Balwearie and Tawstock Dams where the latter is downstream. Downstream

Table 2. Water permits for Mswenzi river. Water Property permit number 7573 8881 10626 10156 2276 12398 10364 12007 Total

Kasama

River Abstraction permit (103m3/a)

Mswenzi tributary Dodington Mswenzi tributary Rem. of Mswenzi Luton Balwearie Mswenzi Strathspey Mswenzi Handley Mswenzi cross Cornucopia Mswenzi Merchiston Mswenzi

Dam

45 Weir

Validity 20 years

136 Farm dam

20 years

1430 Balwearie

20 years

1620 Balwearie 20 years 1100 Tawstock 20 years 2244 Suri 20 years Suri/Tawstock 523 Tawstock 20 years 656 Farm dam 7754

farmers who held earlier water right priorities in Tawstock Dam were outraged that they were not receiving sufficient water from upstream Balwearie Dam and as a result jeopardized their operations. As earlier applicants, Tawstock farmers were entitled to water first, which meant that Balwearie farmers had to open the outlet gates at Balwearie Dam and release water for Tawstock farmers before the former could store and use any water. In 1982 An investigation came about as a result of a submission for a decision made to the Administrative Court by Tawstock farmers who possessed water rights no, 2276,

Conversion of priority water rights to proportional water permits and conflict management

259

10364 and 9101 of Tawstock Dam and who are referred to as the applicants. The applicants were concerned that the holders of water rights no, 10156, 10626 and 10659 (Balwearie Farmers) of Balwearie Dam who are referred to as the respondents, were unable to pass sufficient water from their storage works to satisfy the downstream priorities. The applicants maintained that the reason for this was that the outlet pipe of Balwearie Dam was of insufficient internal diameter for this purpose and therefore the operation of their prior rights was jeopardized. In this case the two dams have a similar capacity and are separated by a mere 2km where each dam has three participants in the utilization of the stored water. Two parts of water right no. 2276 have the earliest priority, after which water right no. 10156 has its turn. Then the remaining part of water right no. 2276 followed by the other two participants is satisfied. The other two participants in water right no. 10156 then follow each with separate priorities. This rather complex situation involves 6 separate water rights and 9 priorities. An agreement was reached at the Administrative Court between the applicants and the respondents. It was agreed that: ● A siphon of 12 inches diameter be constructed and installed together with an outlet pipe of not less than 12 inches diameter in Balwearie Dam. Both devices were to be used to pass water that flows into Balwearie Dam and down to Tawstock Dam to meet the entitlements of holders of water rights no, 2276, 9101 and 10364 together with the primary requirement of 85 liters per second. ● Not less than 425l/s will be released from Balwearie Dam, and the construction and installation of the siphon and the gauging weir immediately upstream of the headwaters of Balwearie Dam shall be carried out by Balwearie farmers so as to be in full operation.

5 SUMMARY The case has described the way in which the operation of priorities of this complex situation worked under the old system and how the issue of satisfying earlier priorities was resolved. In arbitration (a conflict management tool), the Tawstock farmers submitted their argument before the Administrative Court who acted as the judge and a solution was reached whereby both the Tawstock and Balwearie farmers signed an agreement. 6 RECOMMENDATIONS It is recommended that measures be taken to establish actual water use rather than assume values of water rights so that true commitment levels are achieved. It is also recommended that conflict management play a more important role in water resource management so as to empower local water authorities to handle such situations involving disputes over water allocation. It is hoped these recommendations will lay the platform for increased participation, negotiation and dialogue for better basin management.

Water resources of arid areas

260

7 CONCLUSION The main objective of this study was to establish how existing water rights in a small catchment were operated under the old system based on the 1976 Water Act and to describe the conversion process with the requirements of the permit system. The results from the study show that most storage rights were operated with staff gauges installed in the basin that enabled the stored volume and abstraction to be determined for any reservoir level. The study also showed that senior water rights consumed water impetuously without much consideration for downstream users and therefore new users found it difficult to receive a full entitlement of water allocated. The new system now allows new users the opportunity to be given an abstraction permit for their beneficial use therefore disregarding priority. The study revealed through a questionnaire that there was little cooperation and communication over data between upstream and downstream users that often resulted in disputes over water allocation. Of the seven farmers interviewed, six of them said they were not aware of the water reform. It is important therefore that water authorities and catchment agencies seriously consider the issue of enhancing dialogue and cooperation between different users and assist in the issuing of water permits to improve the management of water at catchment and basin level. The results of the study showed that the Mswenzi is a highly committed catchment with a total use of flow and storage water rights totaling 6900*103m3/a for all the water right holders in the catchment. The Mswenzi generates on average per year, 11070*103m3/a of water, bringing the water commitment of the catchment to 62.3%. The study showed that in the conversion process from the old water rights into new water permits, the permits would have to use the same volumes as previously granted for their water rights and discard the priority date. Therein catchment councils have the obligation of amending or revising the water permit according to beneficial use of the permits and accommodation of new entries among other criteria. Perceptions of the commercial farmers in the catchment have shown that the majority, almost 80%, of the big stakeholders in the catchment are unaware of the principles of the water reform and how their new water permits will be operated. Conflicts over water allocation have emerged under the old system due to misperceptions and lack of adequate data and it is anticipated that the new permit system will mitigate the grounds for conflicts in the future. REFERENCES Ashton, P. 1999. Southern African water conflicts: Are they inevitable or preventable? In: Water for Peace in the Middle East and Southern Africa, Geneva, Green Cross International, 2000. Biswas, A.K. 1993. Management of International Waters: Problems and Perspectives, Water Resources Development, 9. Heun, J. 1998. Water resources planning and analysis, Lecture Note, Dept. of Civil Engineering, University of Zimbabwe, Harare. Huffaker et al. 2000. The Role of Prior Appropriation in Allocating Water Resources into the 21st Century, Water Resource Development, 16.

Conversion of priority water rights to proportional water permits and conflict management

261

Jaspers et al. 1999. An external review of the Mupfure Catchment Integrated Water Management Project, Prepared for the Royal Netherlands Embassy, Harare. Lang, H. 1997. Options for a New Water Rights System. Draft Unpublished Paper, October. Manzungu, E. 1999a. Conflict management in the Umvumvumvu Catchment, In: Water for Agriculture in Zimbabwe: Policy and Management Options for the Smallholder Sector, University of Zimbabwe, Harare. Manzungu, E. 1999b. Strategies of smallholder irrigation management in Zimbabwe, PhD Thesis, Wageningen University, Netherlands. Manzungu, E, Senzanje, A & Van der Zaag, P. 1999. Water for Agriculture in Zimbabwe: Policy and Management Options for the Smallholder Sector. University of Zimbabwe Publications, Harare. Mazvimavi, D. 1998. Water Resources Management in the Water Catchment Board Pilot Areas, Phase I: Data Collection, CASS Technical Paper, NRM Series; CPN 95/98, University of Zimbabwe, Harare. Natsa, T.F. 1999. From priority date to fractional allocation: Towards equitable distribution of surface water resources in Zimbabwe. MSc Thesis, University of Zimbabwe, Harare. Resolve Inc. et al. 2000. Participation, Negotiation and Conflict Management in Large Dams Projects, Final version, Cape Town, Republic of South Africa. Van der Zaag, P & Nyagwambo, L. 1998. Water Allocation Criteria for the Mupfure Catchment, Final Document, December 1998, Harare. Van der Zaag, P. 2001. Water Law Lecture Notes, Department of Civil Engineering, University of Zimbabwe, Harare. Wallensteen, P & Swain, A. 1997. Comprehensive assessment of the freshwater resources of the world. International fresh water resources: Conflict or cooperation, Stockholm, Sweden. Wolf, A.T. 2000. Indigenous Approaches to Water Conflict Negotiations and Implications for International Waters, Published in: International Negotiation: A Journal of Theory and Practice, December 2000. Zimbabwe 1996. Zimbabwe Water Act 1976, (Chapter 20:22), Government Printers, Harare. Zimbabwe 1999. Zimbabwe Water Act 1998, (Chapter 20:24), Government Printers, Harare.

Impacts of water development in arid lands of Southern Africa: socio-economic issues J.P.Msangi University of Namibia, Windhoek, Namibia Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Aridity characterizes an expansive area of Southern Africa. By manipulating their environment and available resources, the inhabitants of this area have devised mechanisms that enable them eke a living. Although unfavorable climatic and environmental conditions contribute to precarious living conditions for the inhabitants of the arid lands all over the world, those in Southern Africa are particularly vulnerable due to low technology and high dependence on natural resources exploitation. Many water development projects have been undertaken causing varying environmental impacts that have reduced the performance of the economy and undermined sustainability of projects meant to off set the difficult situations prevalent in the arid lands in Southern Africa. Impacts of such undertakings must be identified prior and after project implementation and mitigative measures taken into consideration during the planning stage. The multisectoral nature of water resource projects should be taken into consideration during the planning and implementation phases. Partnerships and indigenous knowledge are vital in ensuring success.

1 INTRODUCTION Southern Africa has an expansive area that is characterized by aridity, a condition of perpetual moisture scarcity. The inhabitants of such areas have devised mechanisms which enable them eke a living through manipulations of the environment and available resources, water being key to all the activities. The activities of these areas are predominantly changing due to adjustments that must be made in response to the prevailing climatic conditions. Although recurrent droughts, climatic variability and uncertainty heavily contribute to precarious living conditions for the inhabitants of the arid lands all over the world, those in Southern Africa are particularly vulnerable due to low technology and high dependence on natural resource exploitation. Due to the prevailing climatic conditions and high sense of uncertainty, many water development projects have been undertaken or are proposed to provide more dependable water

Impacts of water development in arid lands of Southern Africa

263

supplies to meet both animal and human requirements in Southern Africa. In some areas, irrigation water has been provided to overcome inadequate and/or unreliable rainfall while in others domestic water supply schemes have been constructed including dam building and borehole construction. The main justifying reason for such undertakings is always that water in its natural state is seldom in a position to satisfy the requirements which include public water supply for domestic use and/or livestock production; regulated flow for hydro-electric power production; adequate supplies for industrial processing and irrigated agriculture. Water development projects have been recorded to cause adverse impacts to the environment the world over, Southern Africa included. Such impacts are known to reduce the performance of the economy and undermine the sustainability of projects implemented to off set the difficult situations found in the arid lands. Such impacts need to be identified and/or predicted during the project planning stage so that appropriate mitigative measures can be taken into consideration before and after the project is implemented. Environmental impact issues in Southern Africa include high population concentrations (both human and animal) attracted by the putting up of a reliable water source; soil erosion; agricultural and chemical pollution from irrigated fields as well as over exploitation of groundwater aquifers which may lead to collapse and eventual destruction of the aquifer. Others include denying down stream populations and habitats fresh water supply through damming or excessive abstractions to meet upstream water demands. Through environmental impact studies, the multisectoral nature of water resource development can be taken into consideration during the planning and implementation phases of water resource development projects. Strong partnerships and indigenous knowledge considerations are necessary to make sure all aspects of the resource are included in such studies. 2 INTRODUCTION 2.1 Environmental impact An impact is described as a strong impression or effect on something or even somebody so that a lasting impression is observable and/or measurable (Oxford advanced Learners’ Dictionary). An impact or repercussion of something can also be anticipated or expected or predicted before it takes place. On the other hand, an EIA has been described as a formal study that is used in achieving successful development of major projects through incorporation of environmental considerations in project planning and management (Westman, 1985). An EIA has been identified as both a planning and management tool for sustainable utilization of natural resources. It seeks to ensure that development options are environmentally sound and that any environmental consequences are recognized early and taken into account in project planning, design and implementation. EIA has its origins in USA where during the 1970s initial developments focused on impacts to the biophysical environment and subsequently moved on to encompass and integrate social, health, economic, improved public participation, risk and uncertainty. During mid and late 1980s emphasis included cumulative effects, the integration of project level environmental impact assessment with policy, planning, legislation, monitoring and auditing. During recent times, EIA has been described as “a planning and

Water resources of arid areas

264

assessment process that involves forecasting the environmental consequences of a proposed development process” (Mubvami, 2000). It involves “identifying, predicting, evaluating and mitigating the biophysical, social and other relevant effects of proposed projects and physical activities prior to major decisions and commitments are made” (Mutter, Topfer & Wichterich, 2002). EIA has evolved into a flexible planning tool that allows governments, donor agencies and project developers to evaluate the environmental implications of project proposals during the planning stage. Since mid 1980s, many investors and funding agencies including the World Bank and other multilateral banks require that their borrowers carry out EIA for proposed projects and programmes. During this time, many lending institutions and international environmental agencies like UNEP issued guidelines to assist such assessments and ensure that projects are designed and implemented in an environmentally sound manner. EIA is now widely accepted in both developed and developing countries as an important tool for project planning. The role, fully acknowledged at the 1992 Earth Summit, has led to several countries putting in place legislation that requires that an EIA be conducted before projects are implemented. The purpose of EIA focuses on providing a systematic, holistic and multidisciplinary view of the impacts of a proposed project or undertaking such as the impacts of constructing a dam across a river valley. These impacts include those affecting the natural environment (both living and non-living) and the people who inhabit and use the specified natural environment. In short, EIA examines the environmental and socioeconomic consequences of a proposed undertaking such as a river development project. It emphasizes prevention or minimization of adverse impacts of the project on the environment and the people. It also looks at the effects of the environmental factors on the proposed project as well as the impacts of the people’s activities on the proposed undertaking. Further, an EIA ensures that the ability of the biosphere to absorb effects of proposed activities is not diminished. It is undertaken in order to identify, analyze and assess potential environmental effects of a proposed project and where possible mitigate against negative effects. An EIA exercise can have varied consequences on a proposed undertaking. It can be used to modify and improve the design of a proposal, it can ensure efficient use of resources, it can enhance the social aspects of a proposal and it can be used to identify measures for monitoring and managing impacts and to provide justification for a proposed activity. The effectiveness of the EIA process will have a direct bearing on how many of these results will be achieved. This is more so in highly fragile and vulnerable ecosystems found over much of the arid lands of the world including those in Africa. Since an EIA is conducted before an undertaking, its ultimate goal should be to ensure that current development meets the needs of the present generations without comprising the ability of future generations to meet their own needs. Thus EIA would contribute towards the attainment of sustainable development. 3 PURPOSE OF WATER DEVELOPMENT Most always, the purpose of water development is to provide adequate supplies of water to meet various water demands. This is necessitated by the fact that in most instances,

Impacts of water development in arid lands of Southern Africa

265

water in its natural state is not in a position to satisfy the numerous demands placed upon it, which may include: – Public water supply for domestic use; – Hydro-electric power production; – Irrigated agriculture development; – Water for livestock production; – Industrial water for processing or for cooling machines; – Water for sewage treatment; – Fishing; – Navigation; – Recreation. Water development projects are known to cause both positive and adverse environmental impacts to the environment that need to be identified and/or predicted prior to project implementation. Examples range widely depending on the nature and scale of the project, its location and the type and level of technology required for its sustainability. A good example is where a project involving groundwater recharge using wastewater must take into consideration the danger of pollution from the wastewater from industrial and residential sources. The cost of treatment of the water before using it to recharge ground water must be considered included in the assessment. Similarly, the impacts of other land use management activities taking place in the project area are known to affect the water projects. Activities of unprotected catchment areas including unplanned deforestation and/or overgrazing would produce sediments which would reduce the capacity and adversely affect the life span and operation of a down stream reservoir greatly undermining the project performance and disrupting its sustainability such as happened in Kisongo dam in Arusha and Imagi dam in Dodoma, Tanzania (Msangi, 1987; Kitheka, 1993; Christiansson, 1981). On the other hand a well-managed cultivation system (including terracing and aforestation) of a catchment area will prolong the life span of the reservoir making it possible to meet multiple demands from its waters. All the occurring impacts need to be identified and/or predicted during the project planning stage so that appropriate mitigative measures can be considered before a project is implemented. Needed also is post planning project monitoring to assess the impacts resulting from the undertaking. 4 ASSESSMENT OF ENVIRONMENTAL IMPACT OF WATER DEVELOPMENT PROJECTS In the process of considering whether to carry out a major water project or not, the first main concern is the extent of environmental change that will result from the construction and particularly after the project is implemented. This change will be both to the physical environment as well as to the people and other living organisms inhabiting that environment. The changes to the physical environment will influence the socialeconomic environment just like the people’s response to the physical change will affect the physical environment. Thus both types of change must be considered when impacts are predicted and later on monitored. Most of the changes will be permanent and in some

Water resources of arid areas

266

cases cumulative. Alternatives form part of the pre-implementation phase when impacts are weighed and compared to provide careful and sufficient consideration of all possible impacts, both negative and positive ones. The assessment process for water development projects mainly addresses environmental impacts resulting from the project itself and land use management practices occurring in the river catchment area. These impacts could include high surface runoff; soil erosion and increased sediment flow and raised concentrations that are a result of induced land use changes in a river catchment where a project is proposed. Moreover, lack of knowledge or lack of apprehension of the consequences of overgrazing the catchment areas contributes to the shortening of the life span of a project (Msangi, 1996). Changed ecological conditions, such as the creation of an ideal habitat for disease carrying flies, snails and mosquitoes, have also made areas around a reservoir created behind a dam undesirable and unhealthy (Kaduma, 1977; 1972). Other impacts may include changes to groundwater levels, changes to river flow and flow peaks, flooding or drying up of a river (Christelis & Struckmeier, 2001). Others could be introduction of agricultural chemicals and fertilizer residues by surface runoff from farmlands where irrigated agriculture is part of the project or where the project will induce such undertakings. An indirect impact on an undertaking emanates from the attitudes of the population towards the proposed project as well as attitudes held on water resource use and management. In some parts of the dry lands in Southern Africa, attitudes centering on cattle numbers as wealth are most likely to override environmental conservation so that once water becomes available, the number and intensity of grazing increases without due regard to carrying capacity of the range (Msangi, 1992; 1996; Darkoh, 1989; Ellis & Swift, 1998). Such attitudes and other related social practices should form part of the assessment during and after implementation of a water development project. The actual impacts of a water resource project depend on the purpose, scale and location of the project. For a small water supply project for example, the positive impacts will include the expected socio-economic benefits such as drinking water and water for other domestic uses. Raised health and sanitation standards and the general well being of the people will also be included. An indirect impact will include elevated economic standards, as people enjoying good health will be able to work to produce more and thus generate some economic returns. On the other hand negative impacts will include over grazing of the land around watering points and beyond, along animal tracks and adjacent land etc (Darkoh, 1989; Msangi, 1991, 1996; Stone, 1991). For a large water supply project with pipelines and house connections and staggered animal watering points, the adverse impacts on the physical environment would be minimized or controlled completely. Thus alternatives ought to be considered carefully before implementation. A water development project involving the construction of large structures such as dams and canals will produce a varying range of impacts on both the physical and socioeconomic environments. Possible adverse impacts could include the displacement of people and animal populations as happened when Lake Kariba and other large dams in Southern Africa (Kaduma, 1997) and others including loss of flora and fauna; changes in groundwater conditions, triggering seismic activity due to the presence of a large body of water; deterioration of the health status of the environment through creation of ideal habitats for disease vectors; lowered water quality from rotting inundated vegetation and altered river flow characteristics (Kaduma, 1997). Conversely, the advantages of a large

Impacts of water development in arid lands of Southern Africa

267

scale water development would include creation of new habitats such as wetlands; the production of much needed electricity for irrigation water to support agricultural production; provision of hydro electric power for homes and for industrial establishments; job creation from undertakings utilizing the water and generated power; support improved economic conditions and the well being of the people; regulated river flow and improved utilization of a river including the establishment of a fishery. The list of indirect impacts is long and varied. The creation of sub-surface dams that are more environmentally feasible than surface dams are faced with various limitations including water recovery. High investments required during construction and maintenance and operation of pumps (be it petrol/diesel, solar energy or windmills) are often not economically justified given the low land productivity inherent in most parts of the dry lands of Southern Africa. Hand pumps are only feasible if recovery is from shallow wells (Msangi, 1996). 5 DRY LANDS IN SOUTHERN AFRICA Southern Africa has an expansive area that is characterized by aridity, aridity being a condition of perpetual moisture scarcity. The inhabitants of such areas have devised mechanisms which enable them eke a living through manipulations of the environment and available resources. The activities of these areas are predominantly changing due to adjustments that must be made in response to the prevailing climatic conditions. An exceptionally wet year may see the cultivation and harvesting of quick maturing crops that dry years will not. More often than not the survival techniques include livestock rearing, mostly keeping of small stock such as goats and sheep. Cattle are kept for milk, export beef and as a source of wealth in the areas that enjoy relatively humid conditions as opposed to those that are very dry (Msangi, 1996). Due to excessively high temperatures, the little moisture that may be received in the form of rain gets evaporated very quickly soon after a downpour (Msangi, 1996). Climatic variability and uncertainty has led to precarious living conditions for the inhabitants of the arid lands all over the world, those in Southern Africa high in the list. Due to the prevailing climatic conditions and high sense of uncertainty, many water development projects have been undertaken or are proposed to provide more dependable water supplies for both animal and human requirements as well as industrial water needs. In others, irrigation water has been provided to overcome inadequate and/or unreliable rainfall (Chenje & Johnson, 1996). The main justifying reason for such undertakings is always that water in its natural state is seldom in a position to satisfy the requirements which include public water supply for domestic use; regulated flow for hydro-electric power production; adequate supplies for irrigated agriculture development and for livestock production. Water development projects have been recorded to cause negative impacts to the receiving environment (Kaduma, 1977). While these have been investigated and documented worldwide, in Southern Africa, the need still exists to identify and/or predict them before the proposed projects are implemented and follow up monitoring after they are in operation (Chenje & Johnson, 1996; Wood, Stedman & Mang, 2000). Such projects need to be monitored as they may be affected by environmental factors caused by other land use management activities taking place in the

Water resources of arid areas

268

project areas. Such activities are known to reduce the performance of projects and do undermine the sustainability of projects implemented to off set the difficult situations found in the arid lands (Biswas, 1978; Msangi, 1996). All these impacts need to be identified and/or predicted during the project planning stage so that appropriate mitigative measures can be considered before the project is implemented. The inhabitants of the dry lands in Southern Africa have a rich heritage of managing and living with their environment including water. They have been irrigating their lands for centuries. The communities inhabiting the dry lands have lived and adapted to the environmental conditions arising from many years of experience and folklore handed down generations. This harmonious existence with nature was interrupted and interfered with during the last a hundred years or so through the introduction of western cultures and new ways of viewing the environment. The introduction of improved health and nutritional facilities as well as monetary economy together with the institutional requirements that go with it, has disrupted and partially changed the lifestyles of these communities (Msangi, 1996, 1992; Stone, 1991; Ellis & Swift, 1988). The population of both people and animals has increased rapidly as food aid and western medicine have increased survival chances and increased fertility rates. Before this interruption, land, water and vegetation successfully supported the life styles and economic activities of the dry lands inhabitants. The forces of nature had adequately checked imbalances between man and nature so that simple social and economic patterns had developed and had been harmoniously maintained. The installation of schools, hospitals, central governments and all their branches imposed new requirements on the communities and therefore the environmental resources. Water being the central and most scarce resource in these lands has been subjected to various manipulations and new development approaches geared towards meeting both the communities’ traditional and new institutional demands. Due to increasing populations of both people and animals, water demand far exceeds supply, thus the need to practice wise use, management and conservation of water resources in the dry lands of Southern Africa. This requires that social attitudes be reoriented so that communities appreciate the implications of limited supply as opposed to the ever increasing demands on the scarce water resources, limited groundwater recharge rates and the need to conserve the resources such as controlling pollution and recycling, all new concepts to most of the dry land communities in Southern Africa. Many cases of efforts to conserve, develop and manage the water resources in the dry lands have been made and are documented in numerous plans and consultant reports (Msangi, 1992). Few successful cases have been recorded and many failures have been experienced. The reasons for failure are mainly due to the inappropriateness of the technology adopted to the existing environmental conditions or, most often, to the wants and wishes of the local communities. Many times the wrong sector of the community has been targeted for training. Women and their children who are the ones mainly responsible for collecting and managing water for domestic use and sometimes tending small stock, tend to be side lined for the men who are users rather than managers of the resource. Women should be at the center of any training aimed at improving existing management technologies or introducing new ones. Sustainability rests on clear understanding of the people’s social organizations and gender roles and economic patterns in a given community.

Impacts of water development in arid lands of Southern Africa

269

Economic and other development activities intended to be introduced into the dry lands should be focused more on the needs and wants of the people bearing in mind environmental constraints. For example instead of introducing irrigated agriculture, dry land farming based on indigenous crop varieties should be employed instead of sprinkler irrigation to grow exotic crops with high water demand. Flood irrigation and other high water requiring methods of crop growing have rendered useless large tracts of land through salinization. The high evaporation rates inherent in the dry lands of Southern Africa do not favor these methods. 6 LAND USE MANAGEMENT SYSTEMS AND THEIR IMPACTS ON WATER DEVELOPMENT PROJECTS Water development projects are affected by the land use in many different ways. Unregulated land use system such as indiscriminate clearing of tree cover from a catchment area can lead to reduced water yield and cut short the lifespan of a project. Cultivation and/or heavy grazing of such a catchment area can lead to soil erosion and subsequent sedimentation and siltation of a reservoir, intakes and irrigation channels. Such a system can also lead to increased surface runoff, flooding in the lower reaches and lowered groundwater in the upper reaches thus jeopardizing a water development project. Other impacts include reduced water quality from suspended sediment and agricultural chemicals and residues from farmlands. This will lead to increased costs for water purification or adverse health conditions for those depending on the water source either for domestic use or industrial processing. Habitats for fish and other aquatic animals will be damaged and the economic standing of those dependent on them will be adversely affected. Furthermore, maintenance costs for structures will increase dramatically if sediment has to be cleared regularly, unless ofcourse this was foreseen and budgeted for right from the beginning. Closely connected to decreased infiltration and reduced water yield due to compaction is the loss of water sources such as wells and springs. Reduced infiltration leads to reduction in levels of ground water table that may cause ground subsidence (ground surface collapse and curving in) due to over pumping; or if close to coastal areas lead to intrusion of coastal salt water and soil salinization that may reduce crop production through increased accumulation of harmful salts in soil particularly where irrigated agriculture depend on wells or boreholes (Christelis & Struckmeier, 2001). Other types of land use such as urban land use may cause water pollution due to inadequate water and waste management from dwellings and industries. Pathogens as well as organic and chemical pollution can lower the water quality necessitating expensive water treatment to meet set water quality standards. Alternatively, high concentrations of discharged organic compounds may create excessive demand on oxygen resources of a body of water during the conversion process to the extent that the oxygen concentration in the water is reduced and eventually depleted resulting in death of living organisms including fish. High concentration of organic matter may also raise the fertility of the water body to the extent that eutrophication occurs leading to life decimation in the water body (Wood, Stedman & Mang, 2000).

Water resources of arid areas

270

7 CONCLUSIONS AND RECOMMENDATIONS It becomes apparent that water resource development projects produce serious and definable impacts on a community both socially and economically while other activities in the vicinity of the water projects affect and influence the performance of the projects. Feasibility studies for all impacts must be carried out prior to water projects implementation. Successful water development and resource conservation should always strive to incorporate environmental considerations during project planning and project implementation stages. Similarly, integrated catchment management should encompass the various resource components and associated management practices to achieve stable systems. Environmental legislation should make EIA mandatory in all water development projects in order to ensure sustainability and high quality water supply for industrial, agricultural and domestic usage. People centered planning should be adopted where social, economic and environmental consequences of an undertaking are given deserving emphasis. Therefore social-economic as well as environmental impacts should be considered alongside the often-emphasized physical and technical impacts. 8 APPENDIX 1 (AFTER WESTMAN, W.E. 1985) Questions Useful in Planning the Pre-Impact Phases of an Impact Assessment: 8.1 Phase I: Defining study goals What information is needed and how precise must it be for: – The proponent to minimize environmental impact. – The government agency to reach a decision on approving the project. – Concerned groups to know how they will be affected What resources are needed for the study? What resources are available? – Needed expertise; available expertise – Needed time for baseline and experimental studies – Remaining time before the project is supposed to begin – Required funds to conduct the proposed study; available funds. 8.2 Phase II: Identifying potential impact What are the boundaries of potential impacts? – Area affected – Organisms or ecological functions affected – Duration of the project

Impacts of water development in arid lands of Southern Africa

271

– Interval before effects occur – Duration of effects with and without mitigation. What is the range of potential impacts? – Major direct actions – Major ecological components (air, land, biota, structure) affected – Major ecological processes affected – Secondary or higher-order interactions – Indirect effects triggered at a future time or different place – Other actions (past, present, reasonably foreseeable future) that may add to the present action, causing cumulative effects. Which potential impacts are most significant? Which effects will: – Violate existing laws, plans or policies. – Cause major disruption to ecosystem processes, processes, affecting species significantly. – Cause major adverse effects on species numbers. – Cause health risks, economic losses or significant social disruption to people. 8.3 Phase III: Measuring baseline conditions and predicting significant impacts Baseline conditions: What are the significant features of the ecosystem presently? – What is the current pattern of fluctuation in popular sizes for important species (measured over sufficient time to characterize the range of variations)? – Which species are playing a dominant or critical role in maintaining ecosystem processes? What is their abundance, distribution and function of behaviour? – What is the condition (Quality, quantity, dynamics) of physical resources of the ecosystem? – What are the major pathways of interaction between ecological components? – What sources of stress from natural or human-induced sources already exist (fire, air pollution, grazing etc)? With what intensity and periodicity do these stresses occur? Predictions: What will be the major effects of the proposed action? What is known from each of the following? – Case studies: Extrapolation of effects from similar instances of disruption to the same or similar ecosystems elsewhere. – Modelling: Predictions from conceptual or quantitative models of ecosystem interaction. – Bioassay and Microcosm Studies: The effects of simulated disturbances on ecosystem components under controlled conditions. – Field Perturbation Studies: Response of a portion of the proposed project area to experimental disturbance.

Water resources of arid areas

272

– Theoretical Considerations: Predictions of effect from the current ecological theory. Estimation of likelihood: – What is the probability of occurrence of the predicted events? – How precisely can the magnitude and likelihood of impacts be estimated? Summarizing and analyzing findings: – How can findings be summarized in table, graphs or indexes so that key findings emerge? – What is the ecological interpretation of the findings? 8.4 Phase IV: Evaluating significance of findings How are the effects distributed among the affected groups? – What is the nature and magnitude of impact on each affected group? – What weight shall be given to the concerns of each group? – What weight does each group give to the significance of predicted effects? How well are goals achieved by the proposal? – Proponent’s goals? – Goals of affected groups? What is the overall social significance of the predicted ecological effects? – How can effects be expressed in terms that allow meaningful comparison with other social goods, services and values? – If monetary values are placed on normally unpriced goods and services, what features are inadequately evaluated by this procedure? 8.5 Phase V: Considering alternatives to the proposed action – What would be the effect of not proceeding with the project? – What would be the effect of achieving ultimate project goals by an entirely different means? (e.g. maintaining electrical service to a growing population by conserving energy rather than building a new power plant) – What alternative designs could achieve project goals? What steps could be taken to mitigate adverse environmental effects of the proposed project? – Could parts of the proposal be reduced or eliminated? – Could expected damage be repaired or rehabilitated? – Could ongoing management procedures be instituted to reduce damage?

Impacts of water development in arid lands of Southern Africa

273

– Could affected components be replaced or owners compensated? – Could project design be modified to reduce effects? – Could effects be monitored, and provision made for future mitigation of project effects when the exact nature and extent of effects are better known?

REFERENCES Chenje, M. & Johnson P. (eds.) 1996. Water in Southern Africa. SADC, IUCN SARDC, 238 Biswas, A.K. 1978. Environmental Implications of Water Development for Developing Countries, Water Supply and Management Journal, 2, 283–297 Christelis, G. & Struckmeier, W. (eds) 2001. Groundwater in Namibia: An explanation to the Hydrological Map, 128 Christiansson, C. 1981. Soil Erosion and Sedimentation in Semi-arid Tanzania: Studies on Environmental Change and Ecological Imbalance. Uppsala: Scandinavian Institute of African Studies and Department of Physical Geography, University of Stockholm. Darkoh, M.B.K. 1989. Combating Desertification in the Southern African Region: An Updated Regional Assessment. Nairobi: UNEP Ellis J.E. & Swift, D.M. 1988. Stability of African Pastoral Ecosystems: Alternative Paradigms and Implications for Development. Journal of Range Management, 41(6), 450–211. Kaduma, J.D. 1972. Some Development and Economic Aspects of the Mindu Dam Project, Morogoro: A Background Analysis for Decision Making, M.A thesis, University of Dar es Salaam, 140 Kaduma, J.D. 1977. Man-made Lakes: Their Social, Economic and Ecological Impacts—The Case in Tanzania, PhD thesis, University of Dar es Salaam, 400 Kitheka, J.U. 1993. Soil Erosion and Its Impacts on Surface Water Reservoirs: A Case study of Nguu Tatu Catchment, NE Mombasa District, Kenya, Proc. 4th Land and Water Management Workshop, Nairobi, Kenya, 309–329 Mubvami, T. 2000. Environmental Impact Assessment as a Policy Tool for Environmental Management. IUCN-ROSA A Handbook on Approaches to the Environmental Policy Analysis in Southern Africa. IUCN—The World Conservation Union. Mutter, T., Topfer, J. & Wichterich, C. 2002. A Comprehensive Study of the Heinrich Boell Foundation’s projects abroad in Political Ecology and Sustainability. 1st Ed. Heinrich Boell Foundation. Msangi, J.P. 1987. Conservation of Water Resources in the Semi-arid Areas of Tanzania. Journal of Eastern Africa Research and Development. Vol 17, 63–73. Msangi, J.P. 1991. Sustainability in Exploitation, Development and Management of Hydrological Resources of Turkana District. Journal of Eastern African Research and Development, 21, 21– 39. Msangi, J.P. 1992. Social-Cultural and Demographic Factors in Desertification Control in Kenya’s Arid and Semi-arid Lands. Proc. of the Workshop on Desertification Monitoring, Assessment and Control. Nairobi: National Environment Secretariat, 21–32. Msangi, J.P. 1996. Social-Cultural Factors Affecting Non-Adoption of New Water Harvesting Technology Among the Dryland Communities in East Africa. In Yue-man Yeung (ed) 1996 Global Change and the Commonwealth. Hong Kong Institute of Asia-Pacific Studies, Chinese University, Hong Kong, 233–253. Stone, J.C. (ed) 1991. Pastoral Economies in Africa and Longterm Responses to Drought. Aberdeen: Aberdeen University African Studies Group. Wood, A. Stedman-Edwards P. & Mang, J. 2000. The Root Causes of Biodiversity Loss. Earthscan, 399.

Water resources of arid areas

274

Westman, W.E. 1985. Ecology, Impact Assessment, and Environmental Planning. John Wiley & Sons.

Institutional challenges for small towns’ water supply delivery in Ghana Kwabena Biritwum Nyarko Kwame Nkrumah University of Science and Technology, Kumasi, Ghana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The small towns’ water supply sector emerged in Ghana after 1994 when water supply delivery was separated into urban water supply (served by a public utility) and Community Water Supply (CWS) decentralised under community ownership and management. A small towns’ water system, which falls under CWS, is defined as a piped system serving a community with inhabitants between 2,000 and 50,000 that is willing to own and manage its water supply system. A national programme for community water delivery has been in place since 1994. Ensuring the sustainability of the small towns’ systems are fundamental concerns, which makes the study of institutional issues that affect the sector timely. The paper describes how the institutional arrangement to support the delivery of small towns’ water services in Ghana has evolved, and also discusses the experiences, lessons and the challenges. The paper also makes recommendations to improve service delivery.

1 INTRODUCTION About 32% of the Ghanaian population do not have access to safe water (WDI, 2002) making access to safe drinking water a challenge. The situation is even worse in the rural an small communities where the majority of the population lives. In 1994, water supply delivery in Ghana has been separated into urban water supply (served by a public utility, GWCL) and Community Water Supply (CWS) under community ownership and management (Nyarko, 2000 & CWSA, 2003a) to improve the supply of water to the people of Ghana in a sustainable manner. Community water supply consists of rural and small towns’ water supply. The small towns’ water supply is a piped system, serving communities with inhabitants between 2,000 and 50,000 who are willing to own and manage the water system. As at the end of 2001 there were 254 small towns’ water supply systems under community ownership and management (CWSA, 2003a). These systems are decentralised and do not enjoy any cross subsidies and other benefits of economies of scale as the urban water supply.

Water resources of arid areas

276

Consequently, interest in the delivery of water services in small towns has grown rapidly in recent years due to the peculiar characteristics of the small towns and the number of inhabitants it serves. To ensure the sustainability of the small towns’ systems are fundamental concerns, which makes the study of institutional issues that affect the sector timely. This paper examines the institutional framework of the small towns’ water supply component of community water supply with the aim of enhancing the sustainability of the service delivery.

2 BACKGROUND TO THE SMALL TOWNS’ WATER SECTOR 2.1 Historical development of community water supply sub-sector The Ghana Water and Sewerage Corporation (GWSC), the predecessor of Ghana Water Company Limited (GWCL), was established in 1965 by Act 310 for the provision, distribution and conservation of both the urban and rural water supply in Ghana, for public, domestic and industrial purposes. In 1986, the first attempt to enhance service delivery in the rural areas (defined as communities with less than 5000 inhabitants) led to the establishment of the rural water department within the GWSC (Asamoah, 1998). At that time, the approach of providing water services to customers was a supply driven one. With more promising revenue from the urban areas (with higher income levels) as well as technically challenging “engineering” of providing urban water services, GWSC focused more on the urban areas. In 1991, the sector ministry for water, Ministry of Works and Housing (MWH) organised a workshop to discuss the provision and sustainability of rural water supply and sanitation. The outcome of the workshop was the National Community Water and Sanitation Programme (NCWSP), which aims at accomplishing the following objectives: ● To provide basic water and sanitation services to communities that will contribute towards the capital cost and pay the normal operations, maintenance and repair cost of their facilities ● To ensure sustainability of these facilities through community ownership and management, community decision making in their design, active involvement of women at all stages in the project, private sector provision of goods and services, and public sector promotion and support ● To maximise health benefits by integrating water sanitation and hygiene education In line with the NCWSP, the urban and rural water supply systems were separated in 1994. The rural water division of GWSC was transformed into a semi-autonomous department known as the Community Water and Sanitation Department (CWSD), with the responsibility of implementing the NCWSP. CWSD was further transformed by Act 564 of 1998 into an agency, the Community Water and Sanitation Agency (CWSA) (GOG, 1998) with the responsibility of facilitating community water and sanitation services under community ownership and management.

Institutional challenges for small towns’ water supply delivery in Ghana

277

2.2 NCWSP and GOG decentralisation reforms The implementation strategy of the National Community Water and Sanitation Programme (NCWSP) is consistent with the Ghana decentralisation policy, which transfers authority, responsibility and capacity from the Central Government, Ministries and Departments to the District Assemblies. The decentralisation policy is backed by the Local Government Act, 1993, Act 462, which aims at devolving central administrative authority and divesting implementation responsibility to district levels (GOG, 1993). It re-assigns functions making Central Government Ministries/Departments undertake policy planning, monitoring, evaluation and promotion; and makes regions, (through the Regional Co-ordinating Councils and their respective Regional Planning Co-ordinating Units), play the role of co-ordination, whilst, the District Assemblies become responsible for implementing development programmes (CWSA, 2000). 2.3 The small towns water supply sub-sector A small town is defined in the CWSA Act as “a community that is not rural but is a small urban community that has decided to manage its own water and sanitation systems”(GOG, 1998). A small town water system is also defined as a piped system serving communities of between 2,000 and 50,000 inhabitants who are prepared to manage their water supply systems in an efficient and sustainable manner (CWSA, 2003b). The Act further defines a rural community to be those with less than 5000 inhabitants. The MWH’s Comprehensive Development Framework 1999 for the water sector also defined a small town based on a population range of between 5–15,000 (MWH, 1999). The implementation of each Small Towns’ Project follows the following cycle (CWSA 2003b): ● Project Promotion—for the prospective Community to be familiar with the project cycle and procurement procedures. ● Community Selection and Approval—by the District Assembly in collaboration with the CWSA. ● Community Mobilisation—An extension team is engaged to provide relevant community mobilisation and extension services in each beneficiary community. ● Hygiene Education and Sanitation ● Participatory Planning—to ensure that beneficiary communities are adequately informed and are responsible for decisions made on the system ● Design-Water supply systems shall be adequately designed to provide reliable and good quality water in sufficient quantity over the design period. ● Construction, Operation and Maintenance of the facility ● Post Project—The CWSA shall provide relevant post project support (up to one year) to beneficiary communities to promote achievement of system sustainability. A typical small towns’ water system consists of the following (Jonah. E, 2003): ● a source (usually a mechanised borehole),

Water resources of arid areas

278

● pump house (a submersible pump powered by a 3-phase voltage transformer), ● source of power (AC power from the national grid, local diesel Power generator or Solar panels (only few cases in the northern region) ● Pipelines (transmission and distribution pipes made of uPVC and HDPE), ● An elevated reservoir, standpipes and appurtenances.

3 POLICY, LEGAL AND INSTITUTIONAL FRAMEWORK 3.1 Institutional arrangements The institutional framework is shown in Fig 1. CWSA is under the oversight of the Ministry of Works and Housing (MWH), the sector ministry responsible for water. The District Assemblies (DA) is the highest political and administrative authority in the district, with responsibility for development and management of basic infrastructure, municipal works and services (GOG,

Figure 1. Institutional arrangement for small towns’ water supply delivery. 1993). The Regional Co-ordinating Councils (RCC) and their respective Regional Planning Co-ordinating Units, play the role of co-ordination, whilst the District Assemblies are responsible for implementing development programmes.

Institutional challenges for small towns’ water supply delivery in Ghana

279

The Environmental Protection Agency (EPA), under the Ministry of Environment Science and Technology (MEST) is charged with environmental regulation. The Water Resources Commission (WRC) has the obligation to allocate and grant water rights. The Ministry of Finance (MF) is responsible for negotiation and approval of credit facilities (loans) in the water supply and sanitation industry. The CWSA regional officers constitute the Regional Water and Sanitation Team (RWST), which is composed of a Hydrogeologist, Water and Sanitation Engineer, an Extension Services person and a Financial Specialist. The District Water and Sanitation Team (DWST) is the focal point in the District Assembly (DA) for water service delivery. It’s a three-member team with members seconded from the Public Works Department, the Department of Community Development and the Department of Environmental Health. Their role includes the identification of interested communities and providing support to the Water and Sanitation Development Board (WSDB). The External Support Agencies (ESAs) provide technical and funding support to the subsector. The WSDB is responsible for the management of the small towns’ water supply system. It is composed of elected community (small towns) members. The WSDB is also responsible for appointing the operational staff, promoting and disseminating information within the community, ensuring that all community members participate in decision making, setting tariff and ensuring proper financial management. 3.2 Policy framework Based on the January 2001 draft policy for small towns’ water supply (CWSA, 2001) the key policy statements were: ● Beneficiary communities would pay a part of the capital cost and take up all operations and maintenance costs. The community contribution depends on the levels of service selected by the community. It is 5% of the capital cost for Basic Water Supply Services, which is the supply of 20l/c/d (standpipes) for 80% of population and 60l/c/d (house connection) for 20% population. For higher levels of service the community contribution shall be 50% of the capital cost. ● District Assembly shall contribute 5% of the capital cost. ● Water produced shall meet WHO International Drinking Water Quality guidelines. ● Delivery of water should be in a cost effective manner (not exceeding the cedi equivalent of $1.0/m3) CWSA has reviewed the 2001 draft policy in attempt to improve service delivery. The main changes and additions based on the 2003 draft policy for small towns’ water supply (CWSA, 2003b) are: ● Community contribution for capital expenditure would be 2.5% for Basic Water Supply Services and 50% of the Additional Cost for Levels of Service Higher than Basic Water Supply Services; ● The membership of WSDBs shall exclude Traditional Authorities and DAs. Where necessary, they may participate in WSDB meetings as observers. ● Tariffs shall be set by the WSDBs in accordance with CWSA approved tariff setting guidelines. DAs shall review and approve all tariffs. Any reduction in expected tariff

Water resources of arid areas

280

revenue as a result of action by the DA, e.g., reduced tariff, etc., shall require that the DA pay the difference in revenue into the WSDB account. The CWSA in collaboration with the Regional Co-ordinating Council (RCC) shall ensure compliance. ● The implementation of small towns’ water supply and sanitation projects shall be in accordance with the regulations of the Environmental Protection Agency (EPA) and Water Resource Commission (WRC). The CWSA Guidelines for operations and maintenance (CWSA, 2003c) also stipulates that: ● communities through tariffs shall undertake all major repairs and replacements. But for total rehabilitation, cost sharing arrangement and procurement procedures shall be the same as for new systems. ● Water produced shall meet Ghana Standards Board Drinking Water Quality standards. 3.3 Legal authority of the WSDB The concept of community management is achieved by having community representatives, the Water and Sanitation Development Board (WSDB) in charge of the water supply management in the community. As part of the decentralization policy of the Government and in accordance with section 15 of Act 462 the District Assembly has the power to delegate its functions other than its legislative functions to an individual or group (GOG, 1993). The Water and Sanitation Development Board (WSDB) takes its legal authority from the District Assembly through a byelaw. CWSA has developed generic bye-laws for WSDBs to adapt for their local circumstances.

4 INSTITUTIONAL ANALYSIS OF THE SMALL TOWNS’ SUBSECTOR The research approach utilised both quantitative and qualitative methods to gain insight into the institutional issues of the sector. A literature review of small towns’ water sector project documents was first conducted to get a thorough understanding of the sector. Literature on institutional issues such as what the institutional framework should offer was also reviewed. Based on the literature review, the research instruments (interview guides and questionnaires) were developed for the various stakeholders, in the sector to identify the main institutional issues and challenges. Specific institutional analysis tools used were a combination of the Activity Responsibility Matrix (ARM) and the Strength Weakness Opportunity Treats (SWOT) as well as National Macro-environment Analysis. The data was collected from field visits conducted in 20 small towns’ water supply systems, five district assemblies and three regional CWSA offices. In addition discussions with 20 WSDB Chairmen and 22 Technical Managers who attended a short course at Kwame Nkrumah University of Science and Technology, Kumasi, Ghana in August 2003 were used to validate the study.

Institutional challenges for small towns’ water supply delivery in Ghana

281

4.1 SWOT Analysis of the STWSS sub-sector See Table 1. 4.2 Activity responsibility matrix See Figure 2.

5 DISCUSSION OF KEY ISSUES 5.1 Institutional arrangement and framework The District Assemblies (DAs) are the managers of the NCWSP at the district level. The DAs full time field team for water and sanitation is the DWST. As mentioned already the DWSTs have been seconded from different departments and are ultimately answerable to their respective regional directors, who can effect transferred without consulting the district assemblies. There are cases of such transfers to different district on a totally different assignment. It was revealed that the time taken for the vacancy to be filled normally spans 3–12 months, after which re-training would have to be organised. It was also realised that the monitoring and supervision of the WSDB activities by the DAs is weak. For example, there are no proper monitoring mechanisms to ensue that records (operational,

Table 1. SWOT analysis. Strengths • •

•

•

•

Weakness

• The DWST is not well anchored in High sense of ownership among the DA structure since the communities Almost universal acceptance of the need individual members have been seconded to the DA making DWST to pay for water services even at rates staff subjected to indiscriminate higher than what prevails in the urban transfers by their mother water sector organisations. Users’ perception of the water service delivery is high, since in most cases the • The technical/administrative water situation was poor before the capacity at the local level (DAs, boards took over. WSDB) is weak. • Data collection and record-keeping In a number of situations, the responsiveness of WSDBs/Communities have been poor, regular water quality monitoring has not yet in servicing breakdowns is high, started; compared to the previous situation of centrally-managed systems under • Even though the various boards GWCL. have been taught the guidelines for Accountability to the users in the tariff-setting, in a majority of the community via public fora. systems visited, the tariff was not based on a rational analysis of the cost components

Water resources of arid areas

Opportunities • Government of Ghana support

•

Good will and support from External Support Agencies

•

Government decentralisation policies

282

• Regulation and monitoring from DAs is poor and in most cases no reports are sent by the WSDBs. Even where these are sent no analysis or follow-up is done by the DAs; • Lack of appropriate institutions at the local level to manage the water resource • Revenues accruing from water sales are sometimes mis-appropriated for other purposes; • Membership of some boards is dwindling due to a lack of interest, presumably because of the poor remuneration. This leaves a few who take decisions that may not always be in the interest of the whole. Threats • Political interferences in community management (WSDB affairs). • Relatively high levels of tariff in small towns’ compared with the urban water supply • ESAs/Donor fatigue • Inadequate attention to ensure water resource management • Inadequate attention to ensure financial sustainability

technical and financial) are well kept and that reports are submitted to DAs and CWSA. In addition, when reports are prepared and sent to the DAs it hardly get comments from the DAs. Majority of small towns does not perform routine water quality tests as stipulated by CWSA. This has been attributed to low capacity at the DA level especially the DWST to perform their function and inadequate resources at the DWSTs disposal for their duties. The DAs is also expected to play the role of the Water Resources Commission (WRC) at the district level in the areas of water abstraction rights and permitting. This aspect is not yet operational at the DAs level. The Ministry of Works and Housing (MWH) is responsible for policy making in the water sector. The provision of infrastructure is the responsibility of the DAs, which is under the Ministry of Local Government and Rural Development (MLGRD). CWSA as a facilitating agency cannot force the DAs or the community (WSDB) to execute its water related activities (eg. ensuring the submission of periodic reports, water quality monitoring and using appropriate water tariffs). This is attributed to the following: lack of effective accountability mechanisms between MLGRD/DA and CWSA; the location of

Institutional challenges for small towns’ water supply delivery in Ghana

283

CWSA and DAs in different ministries and the fact that the DAs is an authority on its own.

Figure 2. Scored activity responsibility matrix. To improve on the situation the following option are considered important. First, establish a works department under the DAs structure with responsibility for infrastructure (including) delivery so as to deal with CWSA. This has been accepted in principle but its implementation is yet to start. Secondly, establish clear and explicit accountability mechanism with benchmarks between the RWST and the DAs within the region would help. The introduction of yard stick competition with incentives for the DAs in a particular region or even nationwide would be useful to provide a check on the DAs to enhance performance. It is important for the CWSA regional team to have good collaboration with the RCC the appropriate institution to supervise the DAs activities to enhance CWSA monitoring role of the NCWSP. The roles and responsibilities of the WSDB demand certain skills, such as technical, financial, managerial etc. The selection criteria initially specified gender and interest groups representation, without mention of the skills required. The new draft policy dated July 2003 adds that the membership of WSDBs shall exclude Traditional Authorities and DAs, but where necessary, they may participate in WSDB meetings as observers (CWSA, 2003b). This is laudable since there were interference from traditional authorities. For example in one small town the WSDB chairman was the chief and his nephew was also the treasurer. In a number of communities political agitation and social tensions have affected the membership, tenure and therefore the effectiveness of WSDBs. Some of the examples are:

Water resources of arid areas

284

● In a small town (Bimbilla), after a meeting with all stakeholders to increase the tariffs “a youth movement” managed to convince the District Chief executive (DCE) to reverse the decision. ● In a small town (Bekwai), the District Security Council dissolved the WSDB in response to a proposed demonstration threat by some community members. However, the WSDB were re-instated after about six months. ● In a small town (Juaso) the chief requested for money for farming and was granted by the WSDB. The DA got to know of it and demanded the money back and the WSDB was dissolved. As at now (over 6 months) the new WSDB do not have access to their Bank accounts because signatories have not changed. At a recent training course organised at KNUST for WSDB treasures, majority of the participants confirmed having illiterates on the WSDB and indicated that it affects performance. They attribute that to the community sensitisation during project preparation, which made them understand that the WSDB is the community representative and that any one elected by the community could do it. As a result some of the WSDBs members do not understand the issues, and this reflects in the system’s performance. In cases where members have the required skills, performance has been exceptional (Arthur, 2002). The field visits also revealed that, WSDB perceives themselves as owners of the water supply, which seems to explain why periodic operational reports are not sent regularly to the DAs and CWSA. The DAs also do not have incentive mechanisms in place to enhance the WSDB performance. Most or some of the WSDB do not have approved byelaws and hence do not have legal recognition. 5.2 Cost recovery Initially CWSA policy for the small towns made it clear that water tariffs would have to cover all the operations and maintenance cost, but was not explicit on the recovery of capital expenditure (CWSA, 2001). The new policy indicates that, after the initial community contribution, water tariffs should cover operations and maintenance, major repairs, replacements, and extension to new areas (CWSA, 2003b). However, the operations and maintenance guidelines also states that for total rehabilitation of existing system components cost sharing arrangement and procurement procedures for new projects would be followed (CWSA, 2003c). The lack of definitions to differentiate between major repairs and the total rehabilitation makes the policy unclear. Furthermore, a blanket policy, which does not consider special cases such as a small town with a rather small population but requiring a complex technology, could worsen the plight of some communities with respect to the achievement of public health benefit. 5.3 Interface between “CWSA” and “GWCL/urban” small towns’ water supply In addition to small towns’ water supply being facilitated by CWSA and there are small communities in the urban areas that receive service from the urban public utility (Ghana Water Company Limited, GWCL) as part of urban water supply. The potential transfer from “CWSA small town” to GWCL small town” and vice versa, raises the following

Institutional challenges for small towns’ water supply delivery in Ghana

285

questions (Sarpong Manu, 2001): Whose prerogative is it to make this decision—MWH, DAs or the WSDB?, What will be the criteria for any such transfer? At the moment a small town (Ejisu) with a population of about 15,000 but under GWCL supply service areas wants to come under CWSA and benefit from the small towns facility. They claim that for the past 15 years water supply from the urban water supply utility has been basically non existent.

6 CONCLUSION AND RECOMMENDATIONS At the national level the institutional linkages and the accountability mechanisms between MLGRD/DAs, CWSA, MWH and the Water Resources Commission (WRC) are weak. This results in ineffective monitoring and management of the small towns’ water system as well as the water resources. At the district level the DWSTs members seconded from the other department does not make the DWSTs permanent in the DAs structure affecting delivery of water services. The policy on cost recovery is not clear especially on the recovery of capital expenditure such as rehabilitation and major replacement. Based on the conclusions, the following recommendations are made: ● At the national level there is the need for a closer collaboration between CWSA, MWH, WRCand MLGRD through the inter-ministerial coordination and at the regional level DAs, RCC and CWSA (regional office). In addition there is the need to include accountability mechanisms in the Memorandum of Understanding (MOU) between the DAs and CWSA. ● The Government of Ghana should speed up the process of establishing the Works Department within the DA to strengthen the DWST position in the DA. ● CWSA should clarify the policy on cost recovery. Where there is room for subsidies, the source of the subsidy and criteria for eligibility should be made explicit.

REFERENCES Arthur, E. 2002. Manpower survey at the district assembly and community levels for small towns’ water supply.. BSc Thesis, Kwame Nkrumah Univ. of Science and Technology, Kumasi, Ghana. Asamoah, K. 1998. Ghana: The Community Water and Sanitation Project. Paper presented at the Community Water Supply and Sanitation Conference at the World Bank, Washington, DC. CWSA, 2000. The Project Operational Manual (POM) of CWSP-2, CWSA: 10–11. CWSA, 2001. Small Towns Water supply and Sanitation Policy, CWSA. CWSA, 2003a. Investment Opportunities in the Community Water and Sanitation sub-sector. A presentation to an Americo-German Investors in Ghana, CWSA, pp 2, 8. CWSA, 2003b. Small Towns Water and Sanitation Policy. Community Water and Sanitation Agency, Ministry of Works and Housing, Government of Ghana. CWSA, 2003c. Small Towns Water and Sanitation Policy. Operation and Maintenance Guidelines. Community Water and Sanitation Agency, Ministry of Works and Housing, Government of Ghana. GOG, 1993. Act 462, Local Government Act, Ministry of Local Government and Rural development. Government Printer, Assembly press, Accra.

Water resources of arid areas

286

GOG, 1998. Act 564, Community Water and Sanitation Agency Act, 1998. Government Printer, Assembly press, Accra. Jonah, E. 2003. Performance Assessment of Small Towns Water Supply System: The role of management models an institutional structure, MSc Thesis, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana. MWH, 1999. WATER. Comprehensive Development Framework. Ministry of Works and Housing, Ghana. Nyarko, K.B. 2000. Ghana Water and sanitation sector: Drivers for water performance. PhD Proposal, IHE Delft, The Netherlands. Sarpong Manu, K. 2001. PPIAF/CWSA PSP in Small Towns Water Study. CWSA, Sept 2001. WDI, 2002. World Development Indicators, http://www.worldbank.org/data/dataquery.html

Socio-economic performance of Sepeteri irrigation project in Nigeria O.O.Olubode-Awosola & E.O.Idowu Department of Agriculture Economics, Obafemi Awolowo University, IleIfe, Osun State, Nigeria Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: This study tried to assess the economic, social and financial viability of irrigation the project and examine the efficiency of resource use among the project farmers. Data from all the Sepeteri Project’s farmers during the 2001/2002 seasons were used. Factor share approach was used to examine the resource use efficiency among the farmers. Records of the project’s activities from 1995/96 to 2001/2002 seasons were summarised to some socio-economic performance indices. All the farmers perceived irrigation fee cheap however, 77% attributed their low demand to lack of credits. The irrigation service was acceptable to the farmers with ease of collection. About 67% of farmers do not accept responsibility of making the project a success. Farmers were not efficient in resource use. It was concluded that while the irrigation fee is far below its economic value, it is high enough for the farmers and this prompts them to prefer rain-fed to irrigated cropping. The project was not financially viable due partly to insufficient funding and low level of demand from farmers.

1 INTRODUCTION Up to 1960s, Nigeria was almost self-sufficient in staple food crops from the relatively abundant rainfall. However, from the 1970s, the long drought; the resulting recurrent desert encroachment and the substantial rate of increase in population brought set back Nigeria’s agriculture. The emergence of these three phenomena necessitated public investment in formal irrigation. Irrigation involves development of water resources, conveyance and distribution of water supply at the field coupled with necessary water management exercises (Chukwuma 1993). River Basin and Rural Development Authorities (RBRDAs)’ Irrigation projects were established between 1973 and 1979 to cover every part of the country. They were to carry out a number of functions particularly the development of irrigation infrastructure in their respective areas of operation. Consequently, thousands of tons of crops such as tomatoes, groundnut, wheat, cotton,

Water resources of arid areas

288

millet, maize, etc, were grown by small-holder and commercial farmers. However, the overall performance of the existing irrigation facilities had been on a decline owing to a combination of technical, socio-economic and institutional factors (Nwa 1993). The projects are saddled with inadequacy of untimely funding (Akinkoye 2001) hence, not self-sustained. Consequently, the 1995–97 Corporate Plan mandated RBRDA to generate funds internally to cushion the dwindling funds from the budgetary allocations in order to meet substantial portions of their recurrent costs. Ordinarily, a guided increase in agricultural commercialization leads to increase in purchase of farm inputs, scale (farm size) and specialization in farm enterprises and changes in the role and nature of farm labour inputs. However, this move, if not guided, has the tendency to expose the RBRDAs to been more concerned with the activities that return highest internally generated revenue and possibly less concerned with meeting the irrigation needs of the intended beneficiaries is vital. To this end, while most research efforts on improving the performance of public irrigation projects have focused on the structure, technology and environmental issues, this research focused on social and economic performance of the system as well as the resulting effect on its sustainability and achievements of its statutory functions in the face of commercialization and eventual privatization. The broad objective of this study is the assessment of the socio-economic performance of Sepeteri Irrigation Project. The specific objectives are to: i. Examine the socio-economic characteristics of the farmers; ii. Examine the factors that affect demand for and supply of irrigation services in the project area; iii. Assess the economic, social and financial viability of the irrigated cropping and; iv. Determine the efficiency of resource use among the irrigation project farmers.

2 CONCEPTUAL FRAMEWORK Farmers are primarily concerned with the profitability of their enterprises at individual farm levels. Hence, economic performance of irrigation project farmers is based on a production function and viewed in terms of the efficiency with which farmers combined irrigated plot with other resources such as labour, fertilizers, etc., in the context of institutional framework and management practices of irrigation projects. Farmers are expected to meet economic optimum criteria by adjusting inputs and outputs to relative prices. According to the concept of marginal productivity, a rational producer allocates each variable input according to its market price. This concept is supported by the theory of equilibrium in factor markets under profit maximization, which requires that a factor input be paid its value of marginal physical product (VMP). If a factor is paid higher than its VMP, it is over-utilized and if otherwise it is under-utilized (Henderson & Quandt 1980). However, public irrigation projects in developing nations like Nigeria usually do not meet the conditions for competitive market analysis as explained above because its outputs are natural resources (i.e. land and water), which are developed for national economic efficiency and development. According to Schreiner et al. (1989), such a project is characterized by concepts like natural monopoly, derived demand, etc.

Socio-economic performance of Sepeteri irrigation project in Nigeria

289

Therefore, it is quite possible for project to record negative returns to the agency management because of high cost of capital, yet farmers are expected to make profit consistently. Therefore, rational irrigation policy should ensure that the only sound reason for fixing irrigation fee is the net additional benefits it offers. These benefits accrue to the region and the society as a whole. The major impact of these benefits is to be found in land use, employment, cropping pattern, farm inputs, etc. So, priorities are always given to these benefits above financial returns accruing to the government from irrigation fees. 3 LITERATURE REVIEW 3.1 The state and roles of irrigation projects in Nigeria’s agriculture The practice of irrigation in Nigeria dates back to 700AD. Formal irrigation scheme started in 1926 in Kware, Sokoto State. Subsequently, an irrigation policy for the Northern Nigeria was promulgated in 1963 to construct village-level irrigation schemes. Studies were then conducted to examine the water resources and irrigation development potential in Nigeria. The productivities of Sokoto-Rima and Chad Basin Development Authorities established in 1973 were huge and impressive (Adegbola & Akinbode, 1986). The then Federal Military Government established nine (9) more RBRDAs by 1976 to promote irrigated agriculture in order to enhance food self-sufficiency programmes (OORBRDA 1998). Sepeteri Irrigation Project is a farmer-based irrigation project under the Ogun-Oshun River Basin and Rural Development Authority (O-ORBDA), a parastatal of the Federal Ministry of Water Resources and Rural Development. Consequent to the commercialization programme and the addition of rural water supply function, the River Basin Development Authorities (RBDA) became River Basin and Rural Development Authorities (RBRDAs) since January, 1995. Through the RBRDAs, a number of hectares of lands were available under irrigation. However, the sustainability and efficiency of these formal irrigation projects have started to decline. To arrest this situation and to further explore irrigation potentials, corporate farmers, State’s Agricultural Development Projects (ADPs), private organizations and local governments started investments in small-, medium-scale irrigation projects. Despite these efforts, the various irrigation systems developed so far have not regained the initial performance especially production of import-substitute and export crops. Research findings showed that the productivity of the existing irrigation schemes is on the decline as a result of a combination of technical, social, economic, institutional and political factors (Kolawole 1988). 3.2 Empirical irrigation project performance evaluation The conscientious definition of irrigation performance was reiterated by Bos (1997) as a measure of the degree at which irrigation agent responds to the irrigation needs of its farmers and the efficiency with which the farmers use the resources. Omezzine & Zaibet (1998) used both allocating and irrigation efficiencies as indices of modern irrigation performance in Batinah Region of Oman to examine the economic and technical

Water resources of arid areas

290

efficiencies respectively. Both revealed inefficient water use. The report is that size of irrigated farms and unit cost of water are factors to be considered in the studies of water use and management. Mandal et al. (1995) in an attempt to examine resource use efficiency with respect to farm size of Irrigated HYV Boro Rice Cultivation in Mymensingh District of Bangladesh used factor share approach by estimating a CobDouglas production function. The study revealed that no farm size group allocated resources efficiently. 3.3 Water pricing and irrigation project performance Moore et al. (1994) in a study of four regions of the Western United States discovered that farmers respond to increase in water price by shifting to crops that require less volume of water, hence reduction of acreage of crops requiring high volume of water. This implies that levying too high a charge results in under utilization of facilities such as has occurred on the Sarda Canal in India (NCAER 1959). However, Krishna (1963) found that increase in general water rates would no doubt increase the technical efficiency with which water is applied. It is observed that water abstracted and lifted to field level from wells by human or animal power or by pumps at high private cost, is utilized with much greater efficiency than cheap government canal supplies. Kwanashie et al. (2000) investigated the extent to which poor pricing, poor planning, lack of good management and poor project monitoring and evaluation have affected water resource use in Nigeria. They studied Bakalori Irrigation Project in Nigeria and concluded that these factors undermined water resource management in Nigeria. They then recommended market-based strategies for allocating water between competing users for efficient and cost effectiveness. 4 METHODOLOGY Sepeteri Project is one of the O-ORBRDA’s Irrigation projects supplying irrigation services up to 2001/2002 cropping periods. O-ORBRDA has seventeen (17) farmer-based irrigation projects. However, only two (2) Sepeteri and Itoikin still supplied irrigation services as at 2001/2002 cropping periods. The Sepeteri Project is located in the SakiEast Local Government Area of Oyo State. This is a typical agrarian community. The project was then purposively selected for the study. The project was planned to irrigate 2000ha with sprinkler system. It has 2 dams—Sepeteri I and Sepeteri II of 2.1mcm and 1.3mcm storage capacities respectively for the production of dry season vegetables and Okro. Primary and secondary data were used. One set of structured questionnaire was used to purposively collect primary data from all the forty-four (44) project farmers. They all cultivated a total of 22.35ha. Cross-sectional data of the farming activities during the March 2001 to October 2001 rain-fed cropping period and the November 2001 to March 2002 irrigated cropping period were collected. Secondary data were also obtained from the project records for the periods of 1995/1996 to 2001/2002 cropping seasons. Descriptive and inferential statistics were used to summarize the distribution of data on respondents’ socioeconomic characteristics and factors affecting the supply of

Socio-economic performance of Sepeteri irrigation project in Nigeria

291

irrigation service. Index numbers were used to summarize series of figures over years. The indexes show how much one-year figures differ from another. Usually, a fairly typical year’s figure is taken as a base year figure and others are compared to the base year’s figure. The commonest application of index numbers is in comparison of a series of annual figures. In this study, 1995/96 irrigation season was taken as a base year such that subsequent figures were compared with 1995/96 figures. This year was chosen as a base year to examine how fare have the projects been performing in irrigation services since commercialization move in 1995. This year was assumed to be a logical base year for evaluating performance of a previously public irrigation project. To observe the average annual percentage change over the years, the average annual percentage change was computed as follows: (1) where Indexi=base year Index; indexl=last year index and n=number of years over which the trend is studied (Harper, 1991). To examine the performance of the irrigation projects, performance indices according to Bos (1997) were used. The indices used included Fee Collection Performance, Relative Water Cost, Users Stake in Irrigation System, Financial Self-Sufficiency and Relative Cropping Profit Indices were used to assess the operational and strategic performances of the project management agency. The indexes are specified below: (2) where Irrigation fees collected=total revenue collected on irrigation service during an irrigation season

and Irrigation Fees Due=total revenue collectible on irrigation

service during an irrigation season . The Fee Collection Index reveals the level of acceptance of irrigation delivery as a public service to the farmers i.e. the ease of enforcement of irrigation fee or how affordable the fee is among the intended beneficiaries. (3) where Number of Active project farmers=number of farmers in attendance and Total number of project members=number of project farmers informed and expected to be in attendance. Users Stake in Irrigation Project Index reveals the social capacity of intended beneficiaries and organization in managing and sustaining the project i.e. the level of acceptability of responsibility in making the project a success. The “activeness” of members were quantified using acquired data on the attendance of farmers during the last five consecutive regular meetings called by the management for an agreed upon task such as water distribution, conflict resolution, plot maintenance, etc. (4)

Water resources of arid areas

292

where Actual Income=Total internally generated revenue from irrigation related services and Total MO+M Expenditure=total expenditure on irrigation related services. Financial Self-sufficiency Index reveals the financial viability of the project. (5) where Irrigation Cost per ha=cost of irrigation service per ha (

) and Total

production cost per ha=average total cost of irrigated cropping ( ). Relative Water Cost Index reveals the tendency of farmers abandoning or continuing with irrigated cropping. It is computed on the average. However, The Relative Water Cost Index is perceived to be inadequate to measure the tendency of the farmers abandoning or continuing with irrigated cropping since for some farmers in the developing nations, ends justify the means i.e. (Alimi, pers. comm.) they consider profit far more than the cost. It will then be modified to incorporate the ends, profits from irrigated and rain-fed croppings as specified below: (6) To determine the efficiency of resource use by the respondents, Ordinary Least Squares (OLS) technique was used to estimate parameters of explanatory variables in the postulated Cobb-Douglass production function. The marginal values of inputs used were computed indicating the proportion by which value of crop output changed with one per cent change in the quantity of each input when the quantities of other inputs were kept constant. The production elasticities of the inputs were added together to obtain the returns to scale indicating the proportion by which value of crop output changed with one percent change in the quantities of all the inputs. It is assumed that the value of output depends on level(s) of input(s) such as land, labour, capital and management used and that the production function is a one-equation model (Ogunrowora et al. 1979; Omotesho et al. 1993; Ayanwale 1995). Thus the production function for project farmers was specified as follows: Y=f(X1, X2, X3, X4, u) (7) where Y=Total value of crop output

; X1=size of irrigated farm plot (ha);

X2=expenses on fertilizer and other agro-chemicals ; X3=number of farm household members that assisted in farming activities (man-day); X4=amount spent on hired labour and u=error term. Crop output (Y) was measured in monetary term because two crops—Vegetable (Amaranth sp.) and Okro (Abelmoscus esculentus) were grown together without measuring for sale in standard unit like kilogram. The log-log stochastic production function was fitted for the respondents’ values of crop outputs as follows: ln Y=ln α0+α1 ln X1+α2 ln X2+α3 ln (8)

Socio-economic performance of Sepeteri irrigation project in Nigeria

293

X3+ln α4X4 where ln α0 is the regression constant and ln αi is regression coefficient of Xi. The condition of optimum use of inputs as postulated by the theory of equilibrium in factor market under profit maximization is given by the equation: VMPi=MPPi×Py (9) where VMPi=value of marginal physical product from using additional unit of input Xi; MPPi=marginal physical product from using additional unit of input Xi and Py=market price of the output. So the Allocating/Pricing Efficiency Index is given as (10) where MICi=marginal input cost of input i. (i=1, 24). a priori, the expected signs of the explanatory variables are positive. 5 RESULTS AND DISCUSSION 5.1 Respondents’ socio-economic characteristics Table 1 reveals that about 57% of the farmers are indigenes of the project village. The rest 43% are of distant origin. Almost all the respondents, (95%) resides within the project village. This

Table 1. Summary of socio-economic characteristics of respondents. Characteristic Village of origin: Project/Neighbouring villages *Distant villages Village of residence Project/neighbouring villages *Distant villages Age: 21–30 31–40 41–50 Above 50 Sex: Male Female Level of formal education:

Frequency

% Frequency

25 19

56.82 43.18

42 02

95.45 04.55

04 19 13 08

09.10 43.20 29.50 18.20

40 04

90.90 09.10

Water resources of arid areas

294

No formal education 07 15.91 Primary education 16 36.36 Secondary education 18 40.91 Vocational studies 03 06.82 Years of farming experience: 5–10 10 22.73 11–20 18 40.91 above 20 16 36.36 Year of project participation: 3–6 16 36.40 6–10 23 52.20 above 10 05 11.40 Nature of farming occupation: Full-time 05 11.36 Part-time 39 88.64 Factors affecting size of irrigated plot demand: Credit availability 34 77.27 Cost of irrigation plot in high 05 11.36 Other non-farm engagement 04 09.09 Others 01 02.27 *These are villages farther than 50Km away to Sepeteri, the project village.

revealed that the indigenes within the Project area participated well in the project. About 73% of respondents were within age bracket of 31 and 50 years while about 18% are above 50 years. Also, women scarcely participated in the projects. About 91% of farmers were male. The reason may be that the community is a typical agrarian community where men are predominantly engaged in farming and women engage in other economic activities or assist the male household heads in farming operations. About 16% had no formal education while none had tertiary education, while 77% had between primary school and secondary school education. All the Project farmers had above 5 years of farming experience. In fact, about 37% had above 20 years of farming experience. This result justifies locating the project in the area to help agricultural development and also support the tenet that land should be allocated to farmers with proven commitment to farming as a career. In the same vein, the project farmers had long years of participation. Above half had been with the project for over 6 years. However, few (about 11%) of farmers were full-time farmers. Majority, 88.6% were involved in other economic activities. About 77% acknowledged credit availability as

Socio-economic performance of Sepeteri irrigation project in Nigeria

295

*Table 2. Farm specific characteristics of the project and factors affecting supply of irrigation services. Characteristic

Description

Irrigated plot cost (irrigation fee)

–

3300 per ha and collected after marketing of produce

Upland cost

–

Operation problems

– – – – –

800 per ha and collected after marketing of produce Insufficient resource provision Deterioration of physical structures Inadequate finance Lack of tractor and equipment Occasional invasion by Fulani cattle nomads Harsh harmattan between January and February Monkey pest invasion by February Fuel shortage for water pumping Break down of vehicle to transport produce to market Lack of ready markets for produce Electricity failure for pumping irrigation water 4 months (usually between January and April) Implementation committee comprising of project agents and farmers

Management constraints

Other risks and problems peculiar to the project

– – – – – –

Irrigation period

–

Who is responsible for water – allocation * Response from the project staff.

limiting factors of irrigated plot size while 11% perceived irrigation service fee as high while others expressed engagement in other activities as a constraint. 5.2 Factors affecting delivery and supply of irrigation services Table 2 shows some of the farm characteristics and factors affecting supply of and demand for irrigation services in the project. There is crop restriction to vegetable and Okro. The irrigated and upland costs were 3300 and 800 per ha per cropping season respectively and were constant over years. There was no coordinated Water Users Association but the project manager and farmers met occasionally as matter arose especially to allocate land. The project often witnessed inadequate finance, breakdown of tractor and occasional invasion by Fulani cattle nomads. The physical structures are deteriorated. The irrigation period was usually between 4 months of January and April each year. Some risks and peculiar problems that usually discouraged project participation included harsh harmattan between January and February each year, monkey pest invasion by February each year and lack of ready markets for the produce.

Water resources of arid areas

296

5.3 Socio-economic and financial performance of the project Table 3 reveals that the average irrigation fee collection performance index is 96%. This implies ease of enforcement of irrigation service charge i.e. despite the payment is allowed until after marketing of produce. The User’s stake performance index is 67%. This may be interpreted to mean that about 67% of the farmers were actively involved in the last five (5) obligations required of them for running the project. However, the performance declined at an average rate of 8%. Average Financial self-sufficiency index is 29%. This implies recovering about 29% of the expenditure on the irrigation services rendered to farmers. This implies decline in costs recovery at the annual rate of 25%. Cost of irrigated plot as a percentage of total cost is 20%. This is not much different from 18% reported by Mandal et al. (1995) among irrigation farmers in Mymensingh area of Bangladesh. However it should be noted that the percentage of irrigated plot cost to total production cost is high enough to make farmers abandon irrigated for rain-fed cropping because most mentioned lack of credit facility as a limiting factor to demand for irrigated plot. Besides, the ratio of profit from irrigated cropping to profit from rain-fed cropping is 1.08. This implies that there is no statistically significant difference between profits to irrigated and rain-fed croppings.

Table 3. Socio-economic performance indices of the project. Index Fee collection Users’ stake Financial selfsufficiency Relative irrigated plot cost (as a % of total cost) Relative cropping profit

Average index (%)

Average annual % change in index

96 67 29

4.3 −8 −25

20

1.08

Table 4. Estimates showing efficiency of resource use from the 2001/2002 irrigated cropping. Variable

Average Regression VMP coefficient

MIC

Dependent variable: Y 41,569.91 Explanatory variables: Intercept – 10.67 – – – 0.51 0.718*(4.55) 58,686.93 3300 17.78 X1 1900 0.069 (0.89) 3.16 1900 0.00167 X2

Elasticity of production

– 0.72 −0.069

Socio-economic performance of Sepeteri irrigation project in Nigeria

X3

1.0

X4

500

−0.078 −648.49 500 −1.30 (0.44) −0.094 −1.90 500 0.0038 (0.51)

297

−0.078 −0.094

Other statistics: N=44 R2=0.39 F=5.79* Return to scale=0.62 *Significant at 5%.

5.4 Efficiency of resource use among the respondents From Table 4, 39% of variability in the value of output was explained by the set of explanatory variables captured in the model. The 5.79 F-statistic is statistically significant at 5% level indicating that joint effect of these explanatory variables is significant. The coefficient of irrigated plot size is positive and statistically significant at 5% level. The coefficient of amount spent on chemical has positive sign but not significant. This implies negligible increase in output value results from additional unit increase in the amount spent on chemicals. The coefficient of family labour is negative and not significant. This implies additional use of family labour brings about decrease in output value. This is contrary to expectation and may result from cultivating too small a plot. Coefficient for hired labour is negative and insignificant. This implies that additional use of hired labour results in decrease in output value. The regression constant is 10.67. This is positive and implies that on the average farmers are technically efficient in realizing as much as 11 times in value of input used. This agrees with result reported by Ogunfowora et al. (1979), Omotesho et al. (1993) and Ayanwale (1995). However, the farmers were not efficient in resource allocation. The values of marginal physical products are far different from corresponding marginal input costs (MIC). VMP of irrigated plot is much higher than MIC of irrigated plot. This implies there is scope to increase irrigated plot size to generate higher income. The small and or negative ratios of VMP to MIC of other inputs imply they were over utilized in combination with irrigated plot. Also, the return to scale is 0.62. This indicates decreasing return to scale that the farmers operating under irrational zone of production. 6 CONCLUSION In conclusion, there is higher level of participation from the neighbouring villages to the project. They were mostly resident farmers. Also, they were mostly male with low level of formal education but they had long years of farming experience and project participation. Majority was aged between 31 and 50 years and above. Most of them were part-time farmers and attributed their small level of irrigation participation to lack of

Water resources of arid areas

298

credit facility. Inadequate funding and deteriorating structures hindered the level of irrigation supply of the project. The performance indices revealed that the irrigation service is acceptable to the intended beneficiaries. Higher cost per hectare of irrigated cropping connotes tendency among the farmers to abandon irrigated cropping for rain-fed cropping. However, the farmers do not accept responsibility of making the project a success. The project is not financially self-sufficient partly for insufficient funding for operations and partly low level of demand from farmers. The farmers were inefficient in resource use. The results call for reform in management of irrigation system such that privatization efforts should not tie down peoples’ land unused; Specifically farmers should be encouraged into coordinated and recognized WUA that incorporate credit lending and efficient marking. Further research should look at ways to rationalize family labour. In the same vein, on hired labour other means of bargaining should explored. REFERENCES Adegbola, A.A. & Akinbode, I.A. 1986. A review of old and current agricultural development schemes in Nigeria: Lessons for future programme designs. In Agricultural Development in Nigeria. Ife Journal of Agriculture special publication, 8:1–34. Akinkoye, O. 2001. An overview of organization and management of public sector irrigation schemes. Paper presented at the National Workshop on Participatory Irrigation Management organized by National Agricultural Extension and Research Liaison Services (NAERLS), Ahmadu Bellow University, Zaria in collaboration with the Department of Irrigation and Drainage, Federal Ministry of Water Resources, Abuja. 26–30 March 2001:12pp. Alimi, T. (Personal communication). Ayanwale, A.B. 1995. Resource use efficiency in cassava processing in Oyo North Area of Oyo state, Nigeria. Ife Journal of Agriculture 16, 17:123–135. Bos, M.G. 1997. Performance Indicators for irrigation and drainage. Irrigation Drainage Systems 11(2): 119–137. Chukwuma, G.O. 1993. Some Considerations in Developing Irrigation Research Priorities for Nigeira. Proc. National seminar on Irrigation Research Priorities for Nigeria held at the University of Ilorin, Nig. 20–23 April 1993:65–71. Henderson, J.M. & Quandt, R.E. 1980. Microeconomic Theory: A mathematical Approach 3rd ed., McGraw-Hill Kogansha Ltd. Japan, 420pp. Herpar, W.M. 1991. Statistics, London, Pitman Publishing: 501pp. Kolawole, A. 1988. RBRDAs and vulnerability to hunger in Nigeria, the case of the South Chad Irrigation Project. Food Policy 13(4):389–396. Krishna, R. 1963. Farm Supply Responses in India-Pakistan: A case study in the Punjab Region: Economic Journal, Sept, 1963. Kwanashie, M.A., Togun, A., Ajobo, O. & Ingawa, S.B. 2000. Nigeria Water Resources Management Strategies—Economic and Financing. Technical Report, 16pp. Mandal, K.C., Sabur, S.A. & Molla, A.R. 1995. Resource use efficiency of irrigated HYV boro rice cultivation by difference farm size groups and its impact on employment and distribution of income in DTWII project area of Mymensingh Bangladesh J. Agric. Econs, 8(1):71–87. Moore, M.R., Gollehon, N.R. & Carey, M.B. 1994. Multi crop production decisions in western irrigated agriculture: the role of water price, American, Journal of Agricultural Economics, 76(4):859–874. NCAER (National Council of Applied Economics Research) New Delhi 1959. Criteria for fixation of water rents and selection of irrigation projects, London, Asian Publishing House.

Socio-economic performance of Sepeteri irrigation project in Nigeria

299

Nwa, E.U. 1993. Irrigation Research Priorities for Nigeria. Proc. National Seminar held at the University of Ilorin 20–23 April, 1993, Ilorin, Nigeria, Nwa, EU, Pradhan, P. (eds) IIMI, 104pp. Ogunfowora, O., Esang, S.M. & Olayide, E.O. 1979. Resource productivity in traditional agriculture: a case study of four agricultural divisions in Kwara State of Nigeria. Journal of Rural Economics and Development 9(2):119–131. Omezzine, A. & Zaibet, L. 1998. Management of modern irrigation systems in Oman: allocative vs. irrigation efficiency. Agricultural Water management, 37(2):99–107. Omotesho, O.A., Olufe, J. & Oladeji, S.O. 1993. Resource productivity in food crop production in some selected villages of Oyi Local Government Area, Kwara State, Nigeria. Ife Journal of Agriculture 14(15): 90–97. O-ORBRDA 1998. Federal Republic of Nigeria, Ogun-Oshun River Basin and Rural Development Authority 1997 Annual Report, January, 1999. 49pp. Schreiner, D.F., Badger, D.D., Welsh, M.P. & Suprato, A. 1989. Policy Applications in Natural Resource Projects. In Agricultural Policy Analysis Tools for Economic Development (ed) L.Tweeten, London, Westview Press, 279–321.

Theme D: Application of geophysical, GIS, and remote sensing techniques

Mapping vegetation for upscaling transpiration using high-resolution optical satellite and aircraft images in Serowe, Botswana Y.A.Hussin1, D.C.Chavarro1, M.Lubczynski1 & O.Obakeng1,2 1

International Institute for Geoinformation Science and Earth Observation (ITC), Enschede, The Netherlands 2 Geological Survey of Botswana, Lobatse, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: High spatial resolution images of multi-spectral digital TETRACAM camera were used to map vegetation for upscaling transpiration from tree-bush-shrub of Serowe, Botswana. The camera was mounted on a small aircraft to collect data in 30, 60 and 100cm spatial resolution. The results were compared with one-meter improved multispectral IKONOS satellite images. The high resolution airborne images show high potential for mapping tree-bushy-shrubby vegetation of the study area for up-scaling transpiration. The spectral characteristics of the high spatial resolution images are similar to IKONOS satellite images, while the spatial characteristics of the high spatial resolution images are much better than the one-meter MS IKONOS satellite images.

1 INTRODUCTION As part of the water cycle in the nature or what is well know as the hydrological cycle, surface water is heated by solar radiation and thus evaporated to the atmosphere. However, water in plant is emitted through leaves by a process called transpiration. Vegetation cover is a major component of the hydrological cycle. It has influence on the hydrology of both ground and surface water and on soils. The illogical use or abuse of natural vegetation can have a major effect and consequently changes the hydrological cycle and produce adverse effects. These effects can be very dramatic in arid and semiarid regions. In these regions the climate is very sever with extremely hot temperatures up to 50°C and high evaporation and low annual rainfall. In Botswana, just like any other semi-arid country, water is scarce and surface water is available only for short periods after the rainy season. The main water supply is groundwater. Vegetation in Botswana, as

Mapping vegetation for upscaling transpiration

303

in any other semiarid ecosystems, is often characterized by tree-bush-shrub-grass savannas and woodlands. The increasing demand for fodder and fuel wood has led to drastic declinations on the vegetation cover. The transpiration from vegetation is contributing to the development of cloud cover and consequently to the precipitation. Thus, the amount of ground water is related to the vegetation cover. Research work in Botswana have proven the assumption that water loss comes from plants relies under the experience of rooting plant systems depth for (68m) Le Maitre et al. 2000. The transpiration measurements where done during dry season were soil evaporation was almost negligible. Accurate ground transpiration estimates become more important being the connecting link between groundwater balance and transpiration model from plant. Therefore, a relationship can be established between the ground water and the biophysical parameters of vegetation cover. Accurate mapping of vegetation cover would lead to assessment of the biophysical parameter of vegetation and consequently to transpiration. The objective of this paper is mapping vegetation cover for up-scaling transpiration using high-resolution optical satellite (IKONOS) and aircraft images (TETRACAM). The study area of this research is located in Serowe, Kalahari, Botswana. 2 VEGETATION MAPPING WITH REMOTE SENSING Forests cover large areas of the global land surface. For many developing countries, it represents an important income source for their economies. Due to over exploitation, forests are currently under constant risk. The protection of forests from disasters (e.g. fire, disease, erosion, deforestation, over grazing) over extensive area is difficult without having any information such as condition, area, species, age classes and volume. With these types of information, it is possible to make a proper management of the forest by identifying and selecting the appropriate area for different management purposes, such as, harvesting, protection, etc. Having all these information collected, there is a need to store the referred information properly, for better and comprehensive use. For this reason, forest maps play an important role in organizing gathered information for further strategies and policies determination in order to make the best use of forest. Different approaches for mapping, like land survey, aerial photographs and satellite imagery can be used depending on the level of detail required and the extension of the area under study. For wide areas, satellite imagery has been shown effective for forest classification and consequently mapping. It is recognized that different satellite or airborne imagery can give different results in terms of information extraction. These different results relate to differences in spatial and spectral resolution. Vegetation mapping involves the evaluation of the existing data and information, collecting field data or ground truth, analysing the data and finally developing the vegetation map and validate it (USGS, 1994). Riquene (2002) have studied the vegetation condition of the current research area using Landsat-TM images and ASTER Optical scanner sensors. The study concluded that ASTER images resulted in better vegetation map than TM because of its higher spatial resolution of the 15 meter than TM of 30 meter. Further more the results showed that ASTER data gave more

Water resources of arid areas

304

reliable vegetation maps than Landsat TM data. Mapanda (2003), following the steps of Riquene research, who used high spatial resolution (4 meters) multi-spectral satellite images of IKONOS in comparison with ASTER images of 15 meter spatial resolution, have concluded that IKONOS images gave much better results in mapping the vegetation cover of Serowe, Botswana. This research is going one step further in using higher spatial resolution of multispectral images. These images were acquired using Tetracam multi-spectral digital camera. The camera was mounted on a small aircraft and has collected images in 30cm, 60cm and 100cm spatial resolution. This paper is presenting the first preliminary results of the use of the Tetracam airborne multi-spectral digital images for mapping tree-bushshrub-grass-savannas and woodlands vegetation cover of part of Serowe, Botswana. These images will hopefully be used for up-scaling transpiration of the vegetation in this area in conjunction with IKONOS MS data. 3 STUDY AREA The study area is located in the Central District, about 275km NE of Gaborone the capital of Botswana. Topography is gentle, which varies from 1060 meters above sea level to approximately 1240. It is characterized to be lower in the east and southeast of the region, and the highest location in the vicinity of the escarpment edge. From these ones the average slope is 5% and it gradually decrease to less than 1% towards the east and southeast. Soils units, which can be found in that region, are related to arenosols, regosols, lixisols, luvisols and vertisols. Arenosols are the most common soil units in the study area. It has low moisture retention capacity than the other soil units. Climate is a semi-arid with a mean annual rainfall of 447mm. Rainfall occurs mainly in the summer fallowed by a dry winter season. Summer season stretches from October to April and the winter begins in May to September (Tyson, 1986) (Obakeng, 2000). Main vegetation type is thought that belong to the Northern Kalahari Tree and Bush Savanna. Trees are mostly of Acacia species, which are characterized by the marked tendency to occur in cluster, and are normally accompanied by a variety of grass species such as Ariatida and Eragrotis. Vegetation communities are determined by location on either sandveld or hardveld areas. Dense vegetation is found within and along river courses. This suggests that the vegetation density is governed by the availability of water, which may be partly controlled by topography and geomorphology (Obakeng, 2000). 4 DATA AND ANALYSIS The airborne multi-spectral data was collected using TETRACAM multi-spectral digital camera, which collects its data in three spectral bands namely red, green and near infrared. The data is collected in a rectangular frame of 1280×1024 pixels. The size of the pixels (e.g. ground resolution) would depend then on the altitude of the aircraft above the ground. The camera would saved the image in DCA format (Digital Camera Format), which is a compressed file that can be un-compressed and transferred to Bitmap format

Mapping vegetation for upscaling transpiration

305

that can be imported to any image processing software. The Airborne data was collected in three different spatial resolutions 30, 60 and 100cm. An area of 10×10km was selected as a study site. It is located in the Hardveld part of the Serowe terrain, on which two multi-spectral IKONOS satellite scenes of November 2001 and February 2002 were collected. These images are collected with 4 spectral bands (blue, green, red, and NIR). The spatial resolution of these data sets is 4 meters. A MS image, so called panchromatic sharpen, was available too. The Pan-Sharpen image is a MS image fused with the Panchromatic image of IKONOS, which has 1 meter spatial resolution. Thus the spatial resolution of the new MS image will be improved to 1 meter. For this study area, two aerial surveys were implemented to collect the multi-spectral digital camera data. The first aerial survey was done in November 2003 and the second one was done in February 2004. These surveys were designed and implemented using Aerial-Photography types of survey. The survey divides the area into flight lines. Within each flight line, images were collected with a front overlap of 20% and a side overlap between flight lines of 20% too. The following data where collected: 1. 30cm spatial resolution: 39 flight lines with a total of 910 images 2. 60cm spatial resolution: 21 flight lines with a total of 333 images 3. One meter spatial resolution: 14 flight lines with a total of 187 images. A qualitative approach was used in the analysis of airborne multi-spectral digital images, which mainly involves visual interpretation, spectral signature measurements, spatial features measurements and comparisons of different spectral and spatial data resolution. The same approach was used with the MS IKONOS satellite data to be compared to the airborne data. 5 RESULTS AND DISCUSSIONS The results presented in this paper are the first preliminary findings of some exploratory analysis of airborne MS digital images. Supervised classification and accuracy assessment was not done because an organized fieldwork to collect ground truth was not done yet in the study area. The authors are planning for one in early May of this year (2004). A spectral signature analysis of the 30cm, 60cm and one-meter spatial resolution images of the MS digital Tetracam airborne camera using the digital interpretation of the false color composite, unsupervised classification, and Normalized Difference Vegetation Index (NDVI) (Figures 1–6) showed that the sample used has 3 different spectral classes which referred to 3 different species of the bushy vegetation in the area. It also showed two high contrasted spectral classes, which refer to a soil and a grass classes the selected sample shown in the mentioned figures. A comparison of the above findings with the same signature analysis to the MS IKONOS image of improved one-meter spatial resolution showed similar results of 3 spectral classes of the bushy vegetation and two other high contrasted classes representing the soil and grass classes.

Water resources of arid areas

306

Figure 1. Color composite image of 30cm spatial resolution of MS airborne MS camera.

Figure 2. NDVI map the 30cm spatial resolution of MS airborne image.

Mapping vegetation for upscaling transpiration

307

Figure 3. Unsupervised classification map of the 30cm spatial resolution of MS airborne image.

Figure 4. Color composite image of 60cm spatial resolution of MS airborne MS camera.

Water resources of arid areas

308

Figure 5. NDVI map the 60cm spatial resolution of MS airborne image.

Figure 6. Unsupervised classification map of the 60cm spatial resolution of MS airborne image.

Mapping vegetation for upscaling transpiration

309

Figure 7. Color composite image of one-meter MS IKONOS satellite image.

Figure 8. NDVI map of one-meter MS IKONOS satellite image.

Water resources of arid areas

310

Figure 9. Unsupervised classification map of one-meter MS IKONOS satellite image. Figures 7–9 show the false color composite, unsupervised classification, and Normalized Difference Vegetation Index (NDVI) of the MS IKONOS satellite image. A general comparison of IKONOS images and the airborne MS image showed that both types of images are close in their spectral characteristics, especially the improved one-meter resolution of IKONOS as compared to the 30 and 60cm resolution images of the airborne data. A spatial analysis of several selected objects (e.g. bushes, trees, soil and grass) on the images showed that when spatial resolution increases the accuracy of the information extracted increase. For example a canopy diameter of a single tree measured on the ground is 5.5 meter. The same tree canopy diameter measured on the one-meter, 60cm and 30cm spatial resolution of the airborne MS images are 7.1, 6.13 and 5.38 meter respectively. This is a clear evidence that the higher the spatial resolution of the image the better the accuracy of the analysis or interpretation output. Moreover, the higher the spatial resolution the higher the amount of details extracted from the images as a results of the interpretation or any image analysis technique (e.g. classification or image transformation such as NDVI). For example, some of the tree species in this study area are likely to grow in clusters structure. These clusters may include 3–5 trees in one cluster. On average the crown diameter of these trees say 5 meters. Consequently a total area of the canopies from such cluster may reach up to 125 square meters. The lower the spatial resolution (e.g. one-meter or lower), the less details that represent such a cluster on the image and vis versa. However, in the case of a small bush or shrub, which, has a canopy of approximately one-meter diameter may not show clearly on the image because of the open crown structure. In such a case the spectral reflectance of the soil will dominate such the

Mapping vegetation for upscaling transpiration

311

reflectance from the canopy of the bush. While when using higher resolution (e.g. 30 and 60cm) the representation or the appearance of a small bush will be possible. Therefore, as the spatial resolution increase the spatial information extracted about an object on the images (e.g. a bush or a tree) will consequently increase. The information will include the surrounding area (e.g. soil or grass) of the tree or the bush targeted. This means that using higher spatial resolution we can define the size and shape of any tree or bush much accurate than using lower resolution. The size of the crown of a bush or tree is effecting the estimation of the transpiration of that bush or tree. Consequently this will effect the process of up-scaling transpiration from the area in general. As far as the spatial resolution is concern, a general comparison of IKONOS images and the airborne MS image showed that the information extracted from the 30cm and 60cm spatial resolution images of the airborne MS images is much better than the onemeter resolution of IKONOS image. 6 CONCLUSIONS The following conclusion remarks can be drawn: – The high spatial resolution 30 and 60cm multi-spectral digital Tetracam images have high potential for mapping tree-bushy-shrubby vegetation of semi-arid area (e.g. Serowe, Botswana) for up-scaling transpiration. – The spectral characteristics of the high spatial resolution images are similar to IKONOS satellite images. – The spatial characteristics of the high spatial resolution images are much better than the one-meter MS IKONOS satellite images.

ACKNOWLEDGMENT This research work was partly supported by the internal research fund of GWFLUX Project at ITC. However, Botswana Geological Survey (BGS) has offered the main financial support of the aerial survey missions, fieldwork logistics and transportation. The authors appreciate and acknowledge the support of Botswana Geological Survey. REFERENCES Le Maitre, D.C., Scott, D.F. & Colvin, C. 2000. Information on interactions between Groundwater and Vegetation relevant to South African Conditions: A review. Groundwater: Past Achievements and Future Challenges, Silili et al. (eds). Balkema, Rotterdam, 959–962. Mapanda, W. 2003. Scaling-up and Mapping Transpiration Using Remote Sensing and GIS: A Tool for Water and Forest Management. Unpublished MSc, ITC—International Institute for Geoinformation Science and Earth Observation, Enschede. Obakeng, O.T. 2000. Groundwater recharge and vulnerability: A case study at the margins of the south-east Central Kalahari Sub-basin, Serowe region, Botswana. Unpublished MSc, ITC— International Institute for Geoinformation Science and Earth Observation, Enschede.

Water resources of arid areas

312

Riquene, A.H. 2002. Vegetation mapping in Arid Zones: A multi-sensor analysis, the relationship between Vegetation Distribution and Environmental Factors: A case study in Serowe, Botswana. Unpublished MSc, ITC—International Institute for Geoinformation Science and Earth Observation, Enschede. Tyson, P.D. 1986. Climatic Change & Variability in Southern Africa. Cape Town, South Africa: Oxford University Press. USGS, 1994, 19 July 2001. Field Methods for Vegetation Mapping. USGS-NPS. Available: http://%20biology.usgs.gov/npsveg/fieldmethods/ [2001, 24 August 2001].

Gravity study on groundwater structure in Central Butana (Sudan) K.M.Kheiralla TU Bergakademis, Freiberg, Germany A.E.Ibrahim El Neelain, University, Sudan Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: A number of isolated outcrops of Pre-Cambrian Basement Complex rocks scatter over the central plains of Sudan. In the Butana region, several hills occur prominent landmarks within the predominantly flat clay plains. The solid geology of the Butana plain is rather concealed under a veneer of variable thickness of superficial clays covering over 70% of its total area. The present study was suggested in an effort to delineated and define the mentioned anomaly in more detail and give a reasonable geological and hydrogeological exploration for its existence. Integrated gravity methods and geological feature were applied to achieve the above aims. A total of 200 gravity points were conducted in the study area and they are compile to 275 gravity points acquired, by Sun Oil Company (approximately 2475Km2 were covered in the study area). Gravity data analysis was performed by “GEOSOFT” packages. The result show that the gravity lows is largely attributed to the occurrence of low-density rocks (granite intruded) into the high-density rocks (green schist) of the Butana region. On the other hand the gravity high zones unambiguously coincide with the areas of known shallow Basement Complex. This gravity low is largely attributed to the occurrence of lowdensity rocks (granitic intrusion) into the high-density rocks (green schist) of the Butana region. Thus for it is not unusual gravity measurements in such areas introduced by granitic into country rock of higher densities, to reflect apparently anomalies of typical sedimentary basins. However such ambiguity can be resolved by computation of the second derivative.

1 INTRODUCTION Geophysical prospecting conducted by Sun Oil Company revealed strong negative gravity anomalies in Butana region. It is suggested that sedimentary basins might have caused these anomalies (Ibrahim, 1993; Ibrahim et al., 1996). In the study area the

Water resources of arid areas

314

anomaly (Around Jebel Mundara) lies partially in the Basement rocks. Therefore this research aims to verify the existence of these anomalies and clearly define their extension. In addition, the study aims at determining the type and the dimensions of the rocks causing these anomalies. The study area lies between the Blue Nile River and River Atbara and occupies the Central Butana area, It is bounded by latitudes 14°60′N and 15°80′N and longitudes 34°00′E and 35°20′E. The study area covers about 10,000km2 (Fig. 1). The study area is generally flat, with a gentle slope to the North. The general altitude of the plain is about 500m above mean sea level (m.s.l). The flat monotony of the plain is occasionally broken by some protruding low to moderately high hills or hill chains, which hardly exceed 200m above the ground surface. The climate of the area is arid to semi arid zone of Sudan, characterized by a short duration of a rainy session in summer (July–September), and along dry season for the rest of the year. The average annual rainfall is about 200mm. Average annual temperature over the Butana area is around 40°C in summer (March–October) and 25°C in winter (November–February).

Figure 1. Location map of Central Butana. Table 1. Geological column for butana region (after Iskander et al., 1993). Formation

Age

Superficial deposits Tertiary volcanic

Quaternary/Recent Tertiary

Gravity study on groundwater structure in Central Butana (Sudan)

EL Burg Basal conglomerates The Basement complex

315

Jurassic to cretaceous Precambrian

2 GEOLOGY OF THE STUDY AREA A number of isolated out crops of Precambrian Basement complex rocks scatter over the central plains of Sudan. In the Butana region, several hills occur and make prominent landmarks within the predominantly flat clay plains. The solid geology of the Butana plain is rather concealed under a veneer of variable thickness of superficial clays covering over 70% of the Butana area. Several metasedimentary sub-parallel belts extend for 10–25km in the form of low to moderately elevated ridges surrounded by Butana clay cover (Iskander et al., 1993). The generalized geological column for Butana region can be summarized as shown in Table 1. The majority of the Basement rocks are concealed under the cover of the Butana clay plain. As mentioned, the structural domain in the area is characterized by northeast trending lineaments (Fig. 2). Exposed structures in the metasediments display complex shearing/faulting and tight folding with dipping axial planes where the axes generally have NE-SW trends and with SW mergence Iskander et al. (1993). 3 GRAVITY METHODS 3.1 Introduction The gravimeter used in this survey is a Lacoste & Romberg gravimeter, model D108, which has calibration constant of 1.0863mGal/div and 200mGal measuring range. A total of 200 gravity readings were measured in the study area following a loop survey, with a spacing of 2km between stations (Fig. 3). An area of approximately 2412km2 has been covered. In addition, about 275 gravity points (approximately 2475km2) acquired, by Sun Oil Company were compiled, in the study.

Water resources of arid areas

316

Figure 2. Geological map of Central Butana area modified after (Ahmed & Ayed, 1996).

Figure 3. Location of measured gravity points.

Gravity study on groundwater structure in Central Butana (Sudan)

317

Elevation of the gravity station was determined by Global Positioning System (Garmin II GPS Model 1999), with ±10m accuracy. The Global Positioning System, (GPS) device was used to determine the altitudes of the gravity stations and also for navigation. Control was provided by the available altitude benchmarks in the area and by elevation contour maps of scale 1:100,000, determined by the Survey Department in Khartoum. In addition the altitudes of the previous data (Sun Oil data), were determined by micro barometric altimeter. In gravity measurements effects are produced by sources not directly related to the geological objectives or interest of the study, therefore certain reductions or corrections are necessary to remove these effects. Such corrections include drift, tidal, latitude and elevation correction. This compensates for the earth tides, generated by the complex gravitational interaction on the Earth by the Sun and Moon. These effects are often calculated from table published by Geophysical Journal worldwide. In this work it was done by “GEOSOFT” computer program. 3.2 Gravity data processing 3.2.1 Production of Bouguer anomaly map Presentation of the corrected gravity values is commonly made in the form of contour maps, particularly where the survey has covered a grid of more-or-less evenly spaced stations. Exceptionally, where well-isolated profiles have been surveyed, to obtain crosssectional information on a structure, then the results may be presented in the form of profiles. A sequence of profiles may be shown in stacked form, in proper relative location, on a plan map. In either case, the horizontal scale of the presentation should be inversely related to the distance between the gravimeter stations. Also, the contour interval (mGals) may be inversely related to the scale of the presentation. Software programs are available, by means of which either contour map or profile presentation may be conveniently and quickly made. In this work, gridding was performed by “RANGRID”, program of the GEOSOFT package (GEOSOFT manual, 1989). “RANGRID” produces a minimum curvature grid from data randomly distributed or along non-parallel traverses. The method utilizes different available interpolation options, (e.g. Akima, Cubic etc) to calculate the value of object function at the grid points (original data) that falls within a circle with a given radius centered at the grid points. RANGRID roughly smoothes gaps in the acquired data. In the present study, although data have been acquired along lines, but their irregular points spacing tend to make the gravity data look randomly scattered thus fore they have been subjected to interpolation or gridding process by “RANGRID” which seemed to be a suitable technique (GEOSOFT manual, 1989). The resolution of the produced Bouguer gravity map depends on the choice of the grid cell size, as demonstrated by comparison between the maps shown in Figures (4, 5, and 6), which have been produced by 0.0025, 0.04, and 0.4 cell sizes respectively. To judge on the optimum cell size that resolves the Bouguer gravity map, variations of GB.A has been plotted against corresponding variation of the cell size. It is clearly that no practical displacement (change) has occurred beyond 0.4 cell size. Thus this cell size

Water resources of arid areas

318

(0.4) is consider as the optimum grid cell size suitable for production of the Bouguer gravity map. Contouring of the observed gravity data was performed by “CONTOUR” program of the GEOSOFT package, whose basic function is to thread contour lines through constant levels, defined in a gridded GEOSOFT data file (GEOSOFT manual, 1989).

Figure 4. Bouguer anomaly map, cell size 0.0025. 3.2.2 Production of the residual map The construction of a residual anomaly map due to local structures is therefore a process by which one removes the regional gravity effects. This task could be performed by numerous methods mentioned in the geophysical literature (Seigel, 1995), however in a broad sense they might be classified into graphical or analytical methods.

Gravity study on groundwater structure in Central Butana (Sudan)

Figure 5. Bouguer anomaly map, cell size 0.04.

319

Water resources of arid areas

320

Figure 6. Bouguer anomaly map, cell size 0.4. The gravity measurements at surface determine the sum of all effect from grass root down to the earth crust. Therefore gravity interpretation frequently begins with some procedure, which separate anomalies of interest from superficial disturbance or from deep regional effects. Various methods may be elaborated to perform the separation of the anomalies in order to emphasize the important and interesting features and to suppress the others. These methods include the following. 3.2.2.1 Graphical methods The regional is far away from measuring points it is represented by a long wavelength anomaly due to deep structure. The regional is sometimes shown as a straight line as a result of smoothing a X-Y plot or contour map anomalies. The regional (long wavelength) anomalies mainly due to the effect of deeper structure (lower crust, mantle and core), while the residual (short wavelength) anomalies represent the shallow structure (near surface and crust structure) and may be upper part of the mantle. The graphical methods involve estimation of the regional field from profile plots or contour maps. The advantage of these methods is that, control could be provided by the available geological information (e.g. Basement depth), obtained from boreholes in the survey area.

Gravity study on groundwater structure in Central Butana (Sudan)

321

3.2.2.2 Filtering techniques One of the most important problems in the interpretation of potential field data is to characterize it into different geological structures. The regional one of these methods is the analytical or the filtering techniques were employed which are characterized by: Various methods may be employed to separate anomalies. These include: regional, residual, derivative, low pass filter and high passfilter. However, the selection of the appropriate method depends on the nature of the Bouguer anomalies and the empirical judgment of the interpreter, which is of course vital. On the other hand several of the analytical methods are commonly used for determining the regional and then the residual fields. Griffin (1949), Agocs (1951), Fajklewics (1959), and Abd el Rahman et al. (1983), used in their respective techniques, linear combinations of the average fields on a number of concentric circles of different radii to represent the residual at the common center. The residual field is given by: R= G−Z=G−(ax+by+c) Where, G is the observed gravity, R is the residual field, Z is the regional field, a, b and c are constants. The condition for the above equation is that ∑R2=min These are called residual, but do not posses any relationship with local anomalous mass (Paul, 1967). The only physical significance of this residual is their proportionality to the second vertical derivative value; hence their zero contours coincide with the zero contours of the second derivatives (Nettleton, 1976). Lately dependable method have been introduced by Paul (1967), for computing the second vertical derivatives, consequently the residual determined by the previous method loose much of their significance. 3.2.2.3 Second vertical derivative The second vertical derivative technique was used as a two dimensional filter for interpretation of potential field data (Dobrin, 1976). The second derivatives that have been applied in the present study are the second vertical derivative of the vertical component of gravity. If we used the symbol “g” to denote gravity and choose axes so that Z is vertical downward, then the second vertical derivative is the quantity d2g/dz2. The importance of the second derivative for potential field interpretation arises from the fact that the double differentiation with respect to depth tends to emphasize the smaller, shallower geological anomalies at the expense of larger, regional features (Elkins, 1951).

Water resources of arid areas

322

Figure 7. Second vertical derivative of gravity anomaly map. On the practical side, the second vertical derivative has its disadvantages as seen on map (Fig. 7), that shows a number of anomalies of no actual existence, but they tend to be an interpretation of contours rather than of observed gravity field. The main objective of applying the derivative in this research study is for the delineation of shallow faults. It is interesting to prove how the regional is completely eliminated by the second or higher vertical derivative. This may be shown as follows:

The condition is that,

∑nCn=0

Hence,

Where, Z0 is the average regional at the canter of a particular grid, the residuals around a circle of radius, “n” on substituting,

is the average of

Gravity study on groundwater structure in Central Butana (Sudan)

323

Therefore, for the second vertical derivative only the equation is given as

This equation clearly shows that, the regional field constitutes nothing to second vertical derivative value, and thus this derivative represents entirely the residual value only. A profile of second vertical derivative of gravity in a direction outward from the center of a negative anomaly usually shows an outer maximum value minimum value ratio r,

and an inner

Bott (1965). The source of the anomaly may be determined by the

In case of a sedimentary basin r>1 while for a granite batholiths r<1. 3.3 Gravity data interpretation There are two basic approaches to gravity interpretation. One is to determine a plausible mass distribution directly from the gravity data (Qualitative interpretation). The other is to assume various models conforming to all known constraints and to match gravity effects predicted for each model with the gravity field that has actually been observed. The model that gives the best fit is then considered to be the most probable (Quantitative interpretation) one even though it cannot provide a definitive sub-surface picture. Running “GRAVRED” of the “GEOSOFT” package, and both graphical and analytical methods, were used for the processing did processing and interpretation of the gravity data. 3.3.1 Qualitative interpretation Interpretation of gravity, especially in qualitative sense is constrained by a number of inherent limitations and fundamental ambiguities (Dobrin, 1976). To reduce (offset) these ambiguities usually other geophysical methods and/or geological and borehole information should be incorporated. Physically, the Bouguer gravity map represents anomalies from the entire vertical and lateral density variation with the earth and may be used to qualitatively deduce geological structure. As shown in the Bouguer gravity map (Fig. 6) a prominent gravity high occurs in the eastern and the western parts of the surveyed area where basement is shallow or crops out. Small variations in gravity values in this region probably reflect density variations within the shallow basement caused by variations in weathering, especially in the NW and SE portion of the area where schist rocks occur. The northern part of Bouguer map shows a rounded-shaped strong anomaly, trending NW and with a minimum value of approximately −60mGal which is referred to as Wad Burwa.

Water resources of arid areas

324

Figure (6) shows that the gravity lows over central part the south of Jebel-Mundara Bouguer anomaly map has an amplitude of about −50mGal, a rounded-shaped strong anomaly and of general contours strike which indicates that the study area is narrowly oriented in a NW-SE structural domain. This anomaly is connected to another gravity low over J.Qeili, which extends NW beyond the border of the study area. The abrupt change in the trend direction of the anomalies from northwest (J.Mundara and Qeili), to northeast (Wad Burwa), may be due to the existence of a structural path that might have facilitated the emplacement of low-density bodies whose effects are expressed by the low gravity anomalies. Figure (8) shows a fault or lithological contact as exhibited by the second vertical derivative profile. 3.3.2 Quantitative interpretation The quantitative interpretation determines the shape of the mass excess or deficiency, which cases the gravity anomaly measured on the earth surface. The interpretation of the (residual) gravity anomaly in relation to the sub-surface causal features can be approached into two ways: (1) Linear inverse problem. (2) Non-linear inverse problem. The linear inverse problem arises when the shape of the body is specified and the problem is then to determine the distribution of density as a function of 2-D or even 3-D form of the anomaly on or above the datum plane. In practice the linear inverse problem receives less attention and at most qualitative gravity interpretation is concerned with non-linear inverse problem. The non-linear method calls for approximation of the geological bodies, which are considered to be the gravity source, by assuming simple geometric model from which the theoretical gravity effect can be compared with the observed (residual) gravity data and the shape of the body can be changed (modified) to minimize the difference between the observed and the computed gravity effects, often by interactive and/or iterative computer inversion methods (Kearey et al., 1988). 3.3.2.1 Density measurements In qualitative interpretation of gravity anomalies, it is necessary to determine the density of the subsurface rocks before one can postulate the shape or structure of the source body. For this reason

Gravity study on groundwater structure in Central Butana (Sudan)

325

Figure 8. Fault or lithological contact along profile A-A′ as exhibits second derivative. Table 2. Rock densities of the basement complex. No. of Density Density Rock sample (gm/cc) (gm/cc) type range mean Locality Source Gneiss 2 Green 40 Schist Granite 5 Granite 7 Syenite 16 Gabbro 4

2.85– 2.85 2.65– 2.98 2.59– 2.68 2.59– 2.65 2.57– 2.68 2.92– 2.95

2.84

Gadaref Author

2.81

J.Qeili

2.64

J.Qeili

2.62

Butana

2.63

J.Qeili

2.93

Ahmed (1968) Ahmed (1968) Author

Ahmed (1968) Es Sada Author

some attention has been drawn to the densities and density contrast between the representative rocks in the study area. In fact the density contrast between the rocks are the primary cause of the measured gravity effect. Densities of main rocks composing Butana (Igneous complex) had been measured by Ahmed (1968) and range from 2.57 to 2.68gm/cc. A value of 2.65gm/cc is considered as

Water resources of arid areas

326

the average for the whole complex. Densities of the country rocks vary from 2.65 to 2.98gm/cc, and a value of 2.81gm/cc is considered to be the average. Generally densities of igneous rocks, increase with decrease of silica content or in other words, it follows the acidity line regardless of the rock being plutonic or volcanic (Table 2). On the other hand densities of the metamorphic rock increase with the degree of metamorphism. 3.3.2.2 Modeling of the observed anomalies Modeling of the anomalies in this study was performed by “Grav2dc” program, written by Cooper (1991). It uses the Talwani et al. (1959) type a logarithm, to calculate the gravitational anomaly over one or more 2 D/2.5 D bodies. The construction of models due to local structures is therefore a process by which one removes the regional gravity effects. It eliminates the regional completely and thus enhances the residual anomaly. Thus this models entirely the residual value only (Figs 8, and 9). 3.3.2.3 Modeling along profile A-A′ The profile passes across the central part of the area, generally trend in a NW direction and extends to 50km. No Basement outcrops along this profile, however it passes through Butana clay

Figure 9. (a, b) Two dimensional model along profile A-A′. cover. The profile displays a rapid decrease in gravity anomaly in NW direction. These suggest emplacement of low-density body (mass deficiency) into the green schist. To

Gravity study on groundwater structure in Central Butana (Sudan)

327

account for this low (−ve) gravity anomaly a model representing a granitic intrusion into the green schist with a density contrast of −0.16gm/cc has been simulated as shown in Figure (9a). The model (Fig. 9a) shows intrusive granitic mass of a thickness of about 2km into the green schist, and bounded by several step faults, especially at SE side. 3.3.2.4 Modeling along profile B-B′ The profile starts at about 5km southeast of Jebel-Mundara. The profile extends for about 60km and generally oriented in a NE-SW direction (Fig. 9b). The surface geology along this profile consists entirely of the granite at Jebel Mundara. Modeling of the observed gravity anomaly has reveled emplacement of a granitic body, which extend to 2.5km depth and bounded by near-surface step faults. The simulated (modeled) granitic body correlates with the shape of the Jebel Mundara low anomalies, which trend in the same direction of the fracture system in the old Metamorphic rocks i.e. NNE-SSW. 4 CONCLUSION The interpreted gravity data in this research were measured to confirm the existence of low gravity anomalies in the area that is referred to as Wad Burwa anomaly (Ibrahim, 1993) or otherwise. Filtering of the gravity data comprises the second vertical derivatives of the gravity anomaly. The techniques have proven to be effective in revealing local features more clearly than their respective potential fields. The vertical derivative enhanced and resolved the regional-residual anomalies more clearly. on the other hand have delineated density boundaries (lithological boundary). Modeling of the anomalous field was performed by gravity inversion program (Cooper, 1991), which simulate two-dimensional geological model of irregular geometry, mostly representing the mode of occurrence of the granitic into the country rocks of the Butana region. The surface outcrop of granite controls the shape of the uppermost part of the model, while the −ve density contrasts between the granitic rock and the other green schist account for the −ve Gravity lows in the study area, which are generally known to be shallow Basement Complex terrain. This gravity low is largely attributed to the occurrence of low-density rocks (granitic intrusion) into the high-density rocks (green schist) of the Butana region. Thus for it is not unusual gravity measurements in such areas introduced by granitic into country rock of higher densities, to reflect apparently anomalies of typical sedimentary basins. However such ambiguity can be resolved by computation of the second derivative (Bott, 1965). By verifiable of their occurrence and contact with host rock, granitic bodies slope outwards, thus display (−ve) second derivative gradient as shown in section 3.7.2.1 of this thesis. On the other hand the gravity high zones un-ambiguity coincide with the areas of known shallow Basement Complex. As has been revealed by the quantitative interpretation, the granitic bodies are bounded by near-surface (shallow) step faults. The appearance of these faults on the ground surface is completely masked by the Butana clay

Water resources of arid areas

328

soil and probably they can be exposed by intercepting deep cut water courses (Wadi or Khor). In such conditions the faults can channel water, from surface runoff, to facilitate occurrence of groundwater storage in the basement (granitic rocks) in the Butana area. Filed work for ground water trotting and checking of interpreted data is crucial to up grade the quality of decision. Concurrently geophysical and geological investigation could be carried out in quest to acquire more precise data in the study area (Wadi and flood delta, etc.). The study area of the Butana region has a lack of water supply for both population and livestock uses, because Basement Complex, which is outcropping, or of shallow depth, dominates it. However, further more detailed gravity work in addition to the seismic survey may lead to reveal saturated depression or fracture zones. The presence of water supply in these zones can be expected. REFERENCES Abd el Rahman, E.M., Yehia, A.Y. & Amin, Y.A. 1983. Methods of determination of the proper regional gravity from Bouguer anomaly profile. E.G S. Proc. of 2nd Ann. M. Agocs, W.B. 1951. Least squares residual anomaly determination. Geophysics, 16:686–696. Ahmed, F. 1968. The geology of the Jebel Qeili, Butana and Jebel Sileitaat-Es-Sufr igneous complex, Nile valley, Central Sudan. Unpublished M.Sc. thesis, Univ. Khartoum. Ahmed, F. & Ayed, M.A. 1996. Applied geophysical and satellite imagery techniques, for ground water studies in Central Butana area; ADS report, 25pp, 10–15. Bannister, A. & Raymond, S. 1989. Surveying Catalog, Singapore, Longman Scientific and Technical. Bott, M.H.P. 1965. A geophysical study of the granite problem. Quart. Journ. Geol Soc. London, 112(445): 45–62. Cooper Ltd. 1991. Program “Grav 2dc”, written by G.R.J. Dep. Geophysics, Witwatersrand, South Africa. Dobrin, M.B. 1976. Introduction to Geophysical Prospecting. Mc Graw-Hill, New York. El kins, T.A. 1951. The second derivative method of gravity interpretation. Geophysics, 16:29–50. Fajklewicz, Z. 1959. The use of Cracovian Computation in estimating regional gravity. Geophysics, 15:&QJ;465–478. GEOSOFT reference manual, 1989. Software for earth sciences. GEOSOFT INC, Toronto, Canada. Griffin, W.R. 1949. Residual gravity in theory and practice. Geophysics, 14:39–56. Ibrahim, A.E. 1993. Interpretation of gravity and magnetic data from the Central Africa rift system, Sudan. Unpublished. Ph.D. Thesis Univ. Leeds, 209pp. Ibrahim, A.E., Ebinger, C.J. & Fairhead, J.D. 1996. Lithospheric extension NW of the Central Africa Shear Zone (CASZ) in Sudan from potential field studies. Tectonophysics, 255:70–97. Iskander, W., Ahmed, A.A., Mokhtar, A. & Fadle, A.S. 1993. Appraisal of mineral and water resources of central Butana, Eastern region-Sudan ADS report 85pp. Kearey, P. & Brooks, M. 1988. An Introduction to Geophysical Exploration. Dep. Geol. Univ. Bristol., 296 pp, ch-6, 138–169. Nettelton, L.L. 1976. Gravity and magnetic in Oil exploration. Mc Graw Hill, New York, 464p, 138–169. Paul, M.K. 1967. A method of computing residual anomalies from Bouguer gravity map by applying relaxation technique. Geophysics, 32:708–719. Seigal, H.O. 1995. High precision gravity guides. Canada, Ontario, L4K 1B5:120pp. Sun Oil Company, 1984. Nile blocks gravity survey. Final report, Unpublished.

Gravity study on groundwater structure in Central Butana (Sudan)

329

Talwani, M.J., Worzel, L. & Landisman, M. 1959. Rapid gravity computations for Two-dimension Bodies with application to the Mendocio submarine fracture zone. J. Geophys. Res., 64:49–59.

Remote sensing and electrical resistivity studies on groundwater structure zones in Central Butana (Sudan) K.M.Kheiralla TU Bergakademie Freiberg, German, Gustav-Zeuner-Str, Freiberg, Germany A.E.Ibrahim El Neelain University, Sudan Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Remote sensing techniques combined with the resistivity electrical methods adopted for locating the potential groundwater zones in central Butana. Land sat image is helpful in the location of major deepseated fracture zones. The lineament patterns derived from TM image show strong NE-SW orientations and drainage pattern present in the area. The longer NNE to NE trending feature may be important from a regional hydro-geological point of view, where as the NW trending features are significant in that they intersection the major fault. Groundwater occurrence is mainly due to the secondary porosity, such as weathering, joints, fissures and fracture. Good quantity of groundwater potentials has been identified in the high density of drainage and lineament zone in Butana region, lineaments intersections are important with well yield than are individual. Moderate to good yield of groundwater are tapping from weathered zones and good yield are tapping from fracture zones.

1 INTRODUCTION Groundwater has become an important source of water and has played an importance role in developing industry, agriculture, livestock and domestic purpose. The groundwater condition in a crystalline rock terrain is multivariate because of the heterogeneity of the aquifer, due to the varying composition, compaction, and degree of weathering and density of fracturing. As a result, exploration of groundwater in a crystalline rock terrain has proved to be a complex phenomenon. However, the presence of a vast crystalline rock terrain cannot be neglected as an unfavorable zone. The application of Remote sensing techniques and surface geo-electrical methods is highly helpful for groundwater exploration lies in delineating potential zones of groundwater from a large area. Generally the occurrence of groundwater in crystalline

Remote sensing and electrical resistivity studies on groundwater structure zones

331

rock terrain is associated with the geological structural features like lineaments, fractures/fissure, fault zones. The study area lies between the Blue Nile River and River Atbara, It is bounded by latitudes 14° 30′N and 16°00′N and longitudes 33°30′E and 35°30′E. The study area covers about 20,000km2 (Fig. 1). The study area is generally flat, with a gentle slope to the North. The general altitude of the plain is about 500m above mean sea level (m.s.l). The flat monotony of the plain is occasionally broken by some protruding low to moderately high hills or hill chains, which hardly exceed 200 meters above the ground surface (Kheiralla, 2001). According to the 1993 census in Sudan the total population of the study area is roughly estimated at 30,000 persons, Livestock rising is the major activity of 70% of the Butana population. The ecological conditions as well as the long experience of the inhabitants turn pastoralism as the most worthwhile occupation. The animal population within the area is roughly estimated as some 35,000 heads, composed of about 30,000 sheep and goats, 5,000 cattle and camel (Abd el Ati, 1993). Human and animal populations in the Butana area receive their water

Figure 1. Location map of Central Butana. supply from surface and groundwater. However Groundwater is the only permanent source of water supply in the Butana area. Different than elsewhere in Sudan, the weathered and/or fractured Basement Complex are the main sources of groundwater in

Water resources of arid areas

332

Central Butana region, as they can store and yield reasonable quantities of water from their joints and fractures. The alluvial deposits are none water bearing formation, especially close to the blind deltas of the main Wadis (valley) (Ahmed & Ayed, 1996). However, groundwater occurs in sand alluvial fans are expected to be found under thick cover of the Butana clayey. The surface runoff soon evaporates and/or infiltrates. The crystalline rock terrines, underlying the Butana area is assumed to be groundwater devoid. Under specific geological and hydrogeological condition, the crystalline rock terrines can store and yield appreciable quantities of groundwater variable quality. The weathered and/or fractured crystalline rocks underlying the drainage system from local aquifer zones, in the Butana area the hydro-geological significance of the sandy alluvial deposits is that they act as a membrane through which surface flow can infiltrate to recharge the underlying fractured zones. The study aims to investigate the extent of the influence of the drainage and fractured by the use of Remote sensing and applying geo-electrical methods to delineate the general hydro-geological aspects of the sediments overlying the crystalline rocks for exploration of groundwater. In order to understand the significance of the fracture pattern, geological, hydro-geological, drainage system and lineaments map have been prepared with the help of Land sat TM imagery. An isoresistvity map is prepared by conventional survey of equal apparent resistivity (AB/2=60) and then comparing a lineament map and drainage system map to identify the extent of correlation. 2 GEOLOGICAL SETTING AND HYDROGEOLOGY A number of isolated out crops of Pre-Cambrian Basement complex rocks scatter over the central plains of Sudan (Fig. 2). In the Butana region, several hills occur and make prominent landmarks

Remote sensing and electrical resistivity studies on groundwater structure zones

333

Figure 2. Geological map of Central Butana Area. Table 1. Geological column for Butana region. (After Iskander et al., 1993) Formation

Age

Superficial deposits Tertiary volcanic El Burg Basal conglomerates The Basement complex

Quaternary/Recent Tertiary Jurassic to cretaceous Pre-Cambrian

within the predominantly flat clay plains. The solid geology of the Butana plain is rather concealed under a veneer of variable thickness of superficial clays covering over 70% of the Butana area. Several metasedimentary sub-parallel belts extend for 10–25km in the form of low to moderately elevated ridges surrounded by Butana clay cover (Iskander et al., 1993). The generalized geological column for Butana region can be summarized as shown in Table 1. The majority of the Basement rocks are concealed under the cover of the Butana clay plain. As mentioned, the structural domain in the area is characterized by northeast

Water resources of arid areas

334

trending lineaments. Exposed structures in the metasediments display complex shearing/faulting and tight folding with dipping axial planes where the axes generally have NE-SW trends and with SW mergence. Tectonic activity and the associating metamorphism have resulted in variable folding, faulting and shearing giving rise to complex contact relations between the different basement units and within the metasedimentary assemblages.

Figure 3. Map showing lineaments and rose diagram of Central Butana Area. The structural domain in the area is characterized by northeast trending lineaments (Fig. 3). Exposed structures in the metasediments display complex shearing/faulting and tight folding with dipping axial planes where the axes generally have NE-SW trends and with SW vergence. Iskander et al. (1993) Have interpreted that Riera, Es Subagh and Wad Gidair occupy the limbs of a synform whose vergence is to words SSW with a general N trending axis, while a major anticline occupies the area between Es Subagh in the NW limb and Hosheib-Suruj Jebel El Tawill in the SE limb with a NNE-trending axis and a

Remote sensing and electrical resistivity studies on groundwater structure zones

335

NE vergence. In this search three phases of deformation recognized has reported in the Butana region (Fig. 2) these are: ● F1—Regional foliation, lineation, and shear zones. ● F2—Planar and linear structures, upright folds and faults bricca. ● E-W faulting and fracture cleavage. The tectonic events that terminated by continental collision at the end of the Pan-African developed or reactivated a conjugate set of strike-slip faults and shear zones in both the Nubian and Arabian shield. The fault/shear zones have two main trends NW-SE (Najd trend) and the ENE-WSW (Central African Lineament). In the central Butana regions, major faults and shear

Table 2. Sites proposed for drilling of boreholes. Well Locality S.W.L Expected Aquifer No. (m) yield type (103m3/yr)

Apparent resistivity at depth of 60m (Ωm)

1

Es 45 90 FBC 125 Subagh 2 Es 39 80 FBC 90 Subagh (El Buqaa) 3 Es Sada 25 50 ALL/WBC 60 4 Abu 27 40 WBC 55 Gerad 5 El Fuel 36 70 FBC 115 6 El Bresi 33 65 ALL/WBC 70 7 El Edeid 32 50 FBC 85 El Tawill 8 El Edeid 24 60 FBC 100 El Hamur FBC—Fractured Basement Complex; ALL—Alluvium; WBC—Weathered Basement Complex.

zones display apparently complicated sets of fractures generally follow the regional foliation/ schistosity trends in both the metasediments and the underlying Basement rocks. The structural domain in the north and southeast of Es Subagh area is NNE to NE (Fig. 3). Subordinate N-W and E-W trending faults affect mainly the northwestern part of El Butana (Iskander et al., 1993). The prominent NE trending sets of faults are mainly strike-slip with a dextral sense of movement. Those discontinuities extend up to 70km as attested by the linear-controlled drainage system. Some faults sheared the metasedimentary rock assemblage to significant proportions creating 2–1.5m wide zone

Water resources of arid areas

336

of shearing or brecciation (e.g. Jebel El Rabbda), thrusting the evident in Jebel El Tawill with NE trends and E or SE dip direction. This orientation lineament direction of ophiolite transport. Very similar trends have been suggested for the ophiolite and allochthonous sheets transport in the Engassana Hills (Vail and Duggue, 1986). This deformation created an important fracture, which have provided Es Subagh area in Central Butana with fresh water. To the south, J.El Tawill ultramafic belt has been affected by two boundary thrusts (Fig. 2) 2km apart, resulting into variable degrees of shearing on the eastern and western flanks. Similar sub-basin has been created and provides potable water for the inhabitants of the area. The area has complex hydro-geological conditions owing to complexity in the origin of the rock units encountered. The compact pre-Cambrian suites of rocks are poor aquifers. Groundwater occurs in confined conditions in these rocks due to the development of secondary porosity such as fracturing. Fractured crystalline rocks are less permeable at greater depth because stress variations that cause fractures are larger and, over geological time, occur more frequency near the ground surface. Fractures tend to close at depth because of vertical and lateral stresses imposed by overburden loads and horizontal stresses of tectonic origin. Apparently these basic conditions, which control ground water occurrence in crystalline rocks, apply to a large extent to the Butana area. Groundwater generally occurs in the upper weathered/or fractured zone, which may extend down to 70m depth as indicated by the lithological logs of Es Subagh two boreholes. The sheared rocks, which form the bulk of the aquifer, are composed of acid gneisses, quartzite, marbles and granites. The brecciate marbles in the Butana normally form the good aquifers. The formation easily dissolvable by moving waters, and thus forms wide fissures and cavities, which facilitated groundwater storage and transmission (Iskander et al., 1993). The best aquifer zones comprise the marble cavities along Khor Abu Gimbil, Es Sufeiya, fractured Basement rocks of Adeid El Tawill, Es Subagh, Reira, and El Hagar, and the alluvial deposits along Wadi Abu Grad and Wadi Abu Matariq. Depth to ground water level in the study area varies from 20 to 25m at Reira, between 15 to 50m at Es Subagh, from 20 to 40m below the ground surface at Es Sufeiya, 10 to 20m at El Tawill, and much deeper ground water level (60m below the ground surface) at El Hagar (Table 2). Groundwater in the fracture Basement Complex occurs under free water table conditions and at depths ranging from few meters to over 60 meters below the ground surface. The water table attains its highest level during and shortly after rainy season and drops to its lowest level immediately before the next rainy season. The amplitude of the seasonal or annual fluctuation depends to a great extent on the balance between the recharge and discharge of water to and from the aquifer. Generally productivities of the wells directly correlate the intensity of rainfall in the study area. Noticeably in dry years, the water levels drops to the bottom of the wells or completely dry out. Apparently Groundwater movement in crystalline rocks is non-committal to flow direction, because fracture trends tend to variation that cause movement are largely expected variable to flow direction. Basement outcrops from the main catchments area and ground water movement coincides with the drainage systems controlled structurally. Groundwater moves away from the surface water divide and generally in the eastern and NE direction, ground water moves to joint Atbara River. Eight boreholes have been drilled along watercourses or fractured aquifers at variable depths (Table 2). These include four at Es Subagh, Qeili, Abu Gimbil, Husheib and Umm

Remote sensing and electrical resistivity studies on groundwater structure zones

337

Sarha villages. Only five of the eight boreholes are successful with the total out put hardly exceeding 10m3 per hour, per well. Presently only a few are in use. Two hundreds and fifteen hand-dug wells were excavated into different aquifers (e.g. Fractured, weathered Basement, alluvial or fan deposits) to depths ranging between 20–55m (Kheiralla, 2001). Their static water level varies between 15 and 40m, though wells drilled into the alluvium aquifers are relatively shallower not exceeding about 5m deep. The static water level in these hand-dug wells fluctuates between 3–5m per year, which may be considered as a good sign of recharge. The total out put of the hand-dug wells is estimated to be about 2869*102m3/year. Chemical analysis of samples collected from some hand-dug wells indicates fair to good water quality suitable for human as well as animal consumption. Based on the above estimates, the grand total yield of the existing water sources in the Central Butana is around 33*104m3/year. This amount does not exceed 2/5th(40%) of the actual demand. These indicate an actuate water supply shortage. 3 LINEAMENTS One of the objectives of this work is to delineate the lineaments in more details using the geoelectrical method and Remote sensing technique. Waters (1990) suggests that there are two stages involved in hydro-geological investigations based on remotely sensed images: first, the identification of photo lineaments representing crustal fracturing; and second, the interpretation of these features with respect to their significance in terms of potential groundwater flow. Thus, lineaments visible on the land sat TM images my be expressed by: 1) geomorphologic features such as valleys, straight drainage channel segments, linear scarp faces, or pronounced breaks in the crystalline rock mass, 2) tonal differences at the boundaries of contrasting lithological units. The digitalization of lineaments was carried out through visual analysis at the screen of land sat TM and linear structure features such as faults and fractures were studied in the field. Faults can be distinguished from the fracture by the observations of the slickenside. The Rose diagram (Fig. 3). Constructed from the lineaments map shows the structural domain is NNE to NE trend. Main trend coincide with the Central Africa lineaments with an average direction between 5° and 75°, but most of the long and high frequency lineaments are clustered around 90°, while in the NW direction the NE trending sets of faults are mainly strike-slip with dextral sense of movement. Tensional faults, that is those parallel to the direction of the tectonic stress or orthogonal to the direction of crustal extension, my be believed open and some what wider than compressive/shear faults, which are orthogonal or inclined with respect to the direction of tectonic stress and consequently tend to be tighter. Thus, it should be much easier to recognize tensional faults in a land sat than shear faults and this should be reflected in the lineaments frequency histogram. These preferred orientation of deepseated fractures are responsible for the groundwater potential zones in the study area.

Water resources of arid areas

338

Figure 4. Lineament density map of Central Butana Area. The lineament pattern was subjected to further analytical treatment and a lineament density map has been prepared to identify the fracture concentration (Fig. 3). This was generated by gridding the whole area into 1km2 cell and counting the length of the lineament in each cell and counting these values (Fig. 4). An integrated survey involving location of lineaments by resistively survey for location of fracture openings has indicated that in some areas development. Well yield of groundwater potentials has been observed in the high density of the lineament areas, and was thus indicated by high apparent resistivity value (50–100Ωm) as well as by more alluvial followed by weathered thickness encountered along high density lineament zones. Additional analyses of well yield and lineaments show that lineament intersection, and not the lineaments directions are important. Point of intersection of these lineaments with well yield than are individual lineament. These intersections coincide closely with the main drainage system. More intersection deep-seated fractures are present in high-density lineaments area, which act as groundwater channels, and some of those intersection deep-

Remote sensing and electrical resistivity studies on groundwater structure zones

339

seated fractures are responsible for the formation of groundwater potential zones in the Butana region, where the density of lineaments is found to be between 0.5 and 2 (Fig. 4). These zones may also have the continuity of the lineaments extending from high to low altitudes, which may be buried under transported deposits (20m). This is in conformity with the well yields of the wells. In this region, lineaments are the most significant predictors of groundwater occurrence and general geological structures are less significant. 4 DRAINAGE A drainage map was prepared with help of land sat TM data (Fig. 5). The drainage system, which develops in an area, is strictly dependent on the slope, the nature and attitude of bedrock and on the regional and local fracture pattern. Drainage is studied according to its pattern type and its texture (Way, 1973). Whilst the first parameter is associated to the nature and structure of the substratum, the second is related to rock/soil permeability. Actually, the less a rock is permeable, the less the infiltration of rainfall, which conversely tends to be concentrated in surface runoff. This gives origin to a welldeveloped and fine drainage system. The low hills of the Butana are mostly composed of the Precambrian Basement Complex rocks (Ahmed, 1968), these hills and hill chains are arranged to form a disrupted low regional ridge, which acts as a flat watershed dividing the Butana drainage system to the Blue Nile River in the

Water resources of arid areas

340

Figure 5. Map showing drainage system of Central Butana Area. west and to River Atbara in the northeast (Fig. 1). Because of the amount of the rainfall and the flatness of the terrain, the Wadis (valleys) in the Butana area flow only after heavy thunderstorms. However, none of this flow survives to reach its final destination but usually ends in flood deltas. This means that the drainage within the Butana area is completely internal. The flood deltas at the end of the major Wadis normally offer sizeable areas for rain-fed agriculture in the study area. The major Wadis appear in welldefined channel at their headwaters but when reaching the flatlands down stream, their flows meander in several diffused courses and finally end in deltas (Fig. 5). Figure 6 showing drainage density map has been prepared to identify the drainage concentration. This was generated by gridding the whole area into 1km2 cell and counting the length of the lineament in each cell and counting these values. The superimposition of the drainage density map on the lineament density map show the relationship between them. It also reveals the complete matching between the drainage and lineament densities with well yield distributions.

Remote sensing and electrical resistivity studies on groundwater structure zones

341

The drainage system was classified as first order and second order based on their role in groundwater storage. The first order drainage pattern represent fractures or faults controlling a large part of the study area, affecting a deeper portion of the bedrock and thus can be play an

Figure 6. Drainage density map of Central Butana Area.

Water resources of arid areas

342

Figure 7. Apparent resistivity contour map of Central Butana Area. important role in groundwater storage and transmission. Second order drainage control the patterns and morphology of the rock type, are not important role in groundwater. 5 CORRELATION OF ELECTRICAL RESISTIVITY OF LINEAMENT AND DRAINAGE PATTERN The resistivity values of rocks vary depending upon the presence of secondary porosity such as weathered, fractured and joints. Groundwater prospecting is often combined with geo-electrical measurements. Vertical Electrical Sounding (VES) are executed to detect variation resistivity transition with depth. A total of fifty five (55) Vertical Electrical Sounding measurements utilizing Schlumberger array used in the present study. The objectives of the resistivity survey in the study area are to determine the lithology, weathered, fractured pattern, depth to the basement rock and resistivity variation. Vertical Electrical Soundings were taken at two interest areas, these area are exempted from

Remote sensing and electrical resistivity studies on groundwater structure zones

343

agricultural and livestock activity. Hence, no groundwater exploration is possible in these areas. In the qualitative interpretation, the contour map of the apparent resistivity distribution for the separation AB/2=60m is prepared to delineate high and low zones (Fig. 7). Few resistivity soundings have been taken and correlated with lineaments density zones. Resistivity sounding falling under high-density lineament zones proved favourable results when compared to sounding that fall under other zones. Table 3 shows the thickness of the different formation based on the

Table 3. Resistivity values of rock unite in the Central Butana Area. Rock unit Butana clays Weathered basement Fractured basement

Range in Ωm Min Max 2 10 50

20 100 >500

Figure 8. Geo-electrical section of the study area, showing three hydrogeology units. resistivity values. Using gravity model, geo-electrical section of the study area in (Fig. 8) showing three hydro-geological units (Kheiralla, 2001), weathered rocks and weathered rocks underlain by fractured rocks underlie alluvial layer. 6 RESULTS AND DISCUSSION The results of a comparative investigation of drainage and lineament mapping from TM imagery using vertical electrical soundings data are described. Initial results show that the land sat image is most useful for mapping detailed fracture pattern while the combination of vertical electrical sounding technique is helpful in the location of major deep-seated fracture zones. The longer NNE to NE trending feature may be important from a regional hydrogeological point of view, where as the NW trending features are significant in that they intersection the major fault.

Water resources of arid areas

344

Groundwater occurrence is mainly due to the secondary porosity, such as weathering, joints, fissures and fracture/lineaments. The iso-apparent resistivity contour map (Fig. 7) depicts the horizontal variations in the sub surface lithology of the study area. Figure 7 it is found that the high resistivity zones of more than 50Ωm occur from north-eastern part and from southern part of two interested area. Most of the well located in this zone yield a good quantity of water. Good quantity (more than 90,000m3/yr) groundwater potentials have been identified in the high density of drainage/or lineament zone in Butana region, lineaments intersection are important with well yield than are individual. Assuming that wide variations are not present within a few kilometers, groundwater potential zones have been delineated based on surface lithology, drainage, lineaments/fracture pattern from land sat TM imagery and from electrical resistivity studies (Fig. 9). 7 CONCLUSIONS ● The study of land sat TM images identified a lineament trending NE-SW direction and drainage pattern present in the area. ● Moderate to good yield (40–65*103m/yr) are tapping from weathered zones, good yield (70–90*103m3/yr) are tapping from fracture zones.

Figure 9. Different groundwater potential zones.

Remote sensing and electrical resistivity studies on groundwater structure zones

345

● The area as covered with high alluvial and more fractured zones are providing copious amounts of groundwater. ● Range of resistivities and Expected Yield of different zones are presented in table (2). ● The comparatively high density obtained by lineaments concentration/and or drainage system indicated the presence of groundwater potential zones.

REFERENCES Abd el Ati, H.A. 1993. A base line survey Report on Central Butana. ADS project area-UNDPKhartoum. Ahmed, F. 1968. The geology of the Jebel Qeili, Butana and Jebel Sileitaat-Es-Sufr igneous complex, Nile valley, Central Sudan. Unpublished M.Sc. thesis, Univ. Khartoum. Ahmed, F. & Ayed, M.A. 1996. Applied geophysical and satellite imagery techniques, for ground water studies in Central Butana area; ADS report, 25pp, 10–15. Iskander, W. Ahmed, A A. Mokhtar, A. & Fadle, A.S. 1993. Appraisal of mineral and water resources of central Butana, Eastern region-Sudan ADS report 85pp. Kheiralla, K.M. 2001. Geophysical study on groundwater structure at two localities in Central Butana, Central Sudan. Unpublished M.Sc. thesis, Univ. El Neelain. Waters, P. 1990. Methodology of lineament analysis for hydro-geological investigation. In Satellite Remote Sensing for Hydrology and Water Management. E.C.Barret, Power, C.H. & Micallef, A. eds., New York, Gordon & Breach: 1–23. Way, D.S. 1973. Terrain analysis, a guide to site selection using aerial photographic interpretation, Stroudsburg, Dowden, Hutchinson, Ross Inc. Vail, J.R. & Duggue, J.P. 1986. Bibliography of geological sciences for the Republic of the Sudan. 1837–1985, Center Int. Formation Echanges Geol. Paris, Spec. Publ.

Monitoring and modeling of fluxes on Kalahari—setup and strategy of the Kalahari Monitoring project Serowe study case, Botswana M.W.Lubczynski1 & O.Obakeng1,2 1

The International Institute for Geoinformation Science and Earth Observation (ITC), Enschede, The Netherlands 2 Geological Survey, Lobatse, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The ongoing discussion about the presence and the rates of recharge in Botswana, which constrains groundwater sustainability in the country scale, has led to the initiation of a new recharge project in Botswana called Kalahari Monitoring Project. In contrast to previous attempts this project focuses on temporal flux monitoring by using automated data acquisition systems (ADAS). The framework of that project is discussed on the base of the example of the Serowe study area, located at the eastern fringe of Kalahari, where an extensive monitoring network was installed to provide data for spatio-temporal flux assessment. This network allows measurements of saturated, unsaturated and surface zone fluxes. It consists of groundwater table fluctuation monitoring in 21 piezometers, soil moisture and soil suction pressure monitoring in 7 identical profiles comprising measurements at 0.5, 2, 4, 6, 8m b.g.s., one deep suction pressure profile down to 76m (sensors at 15 different depth levels), transpiration monitoring using 51 sap velocity thermal dissipation probes installed at 9 ADAS locations and monitoring of climatic variables for potential and actual evapotranspiration in 10 towers scattered over the study area. This data is either interpreted directly (rainfall, transpiration) or used in 1-D models to calibrate surface and subsurface fluxes such as evapotranspiration, groundwater evapotranspiration and recharge. For spatial data assessment the remote sensing (RS) method is proposed. The evapotranspiration is obtained with RS solution of energy balance, transpiration by RS upscaling of the sap flow measurements and recharge by RS and GIS modeling. The final integration of spatial and temporal data for spatio-temporal flux assessment is carried out by transient groundwater model calibration with spatio-temporally variable recharge and groundwater evapotranspiration. The aspect of partitioning of tree

Monitoring and modeling of fluxes on Kalahari

347

transpiration fluxes into saturated and unsaturated zone is tackled by isotopic depth dependent tracing of groundwater and tree response analysis. The preliminary results of this study indicate already that the net recharge in incidental hydrological seasons can be substantially higher than the average recharge defined by isotopic and chemical methods. In other years however, the net recharge is usually negative due to the typical excess of groundwater evapotranspiration over the recharge, which is mainly due to the substantial role of transpiration in the overall groundwater balance.

1 INTRODUCTION A long lasting debate is continued in Southern Africa, particularly in Botswana, regarding the presence, rates and the nature of groundwater recharge on Kalahari. Based on the completed GRES II project, the Kalahari recharge was defined on average in the order of 5mm/y or less, using hydrochemical and isotopic methods (de Vries et al. 2000). The GRES recharge rates obtained mainly with chemical and isotopic methods provide the long-term average recharge. In groundwater modeling such recharge can only be applied as steady-state flux input, which is justified if the temporal variability of fluxes is low. Otherwise, as it often happens in arid and semiarid countries, models have better setup if fluxes are provided in spatio-temporal manner (Lubczynski 2000, Lubczynski & Gurwin 2004). In Botswana for example, in the wet season of 1999/2000 when many places in Southern Africa experienced incidentally high rainfall, the monitored groundwater table rise indicated recharge values several times higher than 5mm/y stated by GRES II project. In the other years however, when rainfall and recharge (R) were low, groundwater table declined to the stages lower than at the beginning of the hydrologic year. This happened not only due to the lateral groundwater outflow but also due to the groundwater evapotranspiration (Lubczynski 2000). Groundwater evapotranspiration (Eg) consists of two types of fluxes: groundwater transpiration (Tg) representing root groundwater uptake and groundwater evaporation (Cg) representing convective and diffusive water flux originated from groundwater table which evaporates while reaching a shallow zone of a few m b.g.s. Similar flux components, such as unsaturated zone transpiration (Tu) and unsaturated zone evaporation (Cu) are also defined with reference to unsaturated zone. The similarity between Tg and Tu as well as between Cg and Cu makes difficult partitioning of the flux contributions of saturated and unsaturated zone. Such difficulty occurs for example when transpiration is assessed by tree sap flow measurements, which represent the combined effect of transpiration originated from groundwater and from unsaturated zone. The assessment and partitioning of Cg is even more difficult because so far there are no methods of measuring of this component and moreover Cg and Cu can also be confused with the surface evaporation when assessment is made from the land surface. Not only temporal assessment of groundwater fluxes is considered as a problem but also the assessment of spatial variability of fluxes. Very often point data characterizing local behavior of saturated-unsaturated fluxes is available, like from specific chloride

Water resources of arid areas

348

mass balance measurements or from 1-D recharge modeling, but the method of spatial representation of such data is not well defined. It is an ongoing dispute on what are the best and the most efficient schemas to present point flux information spatially, using interpolation, extrapolation by GIS modeling, stochastic modeling or discrete groundwater modeling? Thus in 2001a complex Kalahari Monitoring research project was established in Botswana focusing on monitoring and modeling of surface and subsurface fluxes. The main objective of this project is spatio-temporal assessment of subsurface fluxes for better management of groundwater resources in Botswana. As study areas three hydrologically different semi-arid locations were selected, Maun area, Localane-Ncojane area and, Serowe area. Maun area represents relatively wet conditions of the Delta Okavango with very shallow groundwater table of only few meters below the surface. Localane-Ncojane area represents the western, driest part of Botswana Kalahari with very deep groundwater table on average 100–150m b.g.s. The Serowe area of ~2500km2 on which this study focuses (Figure 1), is currently the most instrumented and the most intensively investigated research area of the Kalahari Monitoring project. 2 WHY SEROWE AREA AS STUDY AREA? The Serowe study area was selected as target area of the Kalahari Monitoring project following the previous research in the same area delivered by SGAB (1988) and WCS (2000). The latter one included also the numerical groundwater model, after which the present study area boundaries were assumed (Figure 1). The study area consists of two contrasting parts, Kalahari sandveld and hardveld, which have different natural and hydrological conditions. The western sandveld part is elevated as compared to the hardveld along the prominent escarpment feature. This elevation is due to the 60–100m eolian Kalahari sand cover on the western, sandveld part, overlying solid rocks such as Stormberg basalts and Ntane sandstones which in the eastern part outcrop or are covered by thin, 0–5m Kalahari sand cover (Figure 1). The sandveld part slopes gently to the west, is fairly flat and featureless without prominent drainage lines. In contrast, the hardveld part slopes steeper and there is a drainage system of the intermittent streams, discharging water mainly after the heavy showers. The majority of the villages in the study area are concentrated along the eastern edge of

Monitoring and modeling of fluxes on Kalahari

349

Figure 1. Serowe study area and its monitoring network. the escarpment where some ten years ago, springs were supplying water for habitants. The escarpment line as well as the drainage lines in the eastern part of the study area where groundwater table is relatively shallow (<20m), are marked by green riverine woodland vegetation, denser and taller than elsewhere. The hardveld is dominated by acacia savannah type of vegetation, which can vary from dense shrub land to true tree savannah. The Kalahari sandveld is represented by open savannah vegetation type characterized by continuous grass layer and discontinuous sparse tree and bush layer. It is noticeable that in the eastern area, vegetation remains green even at the peak of the dry season in contrast to the western part, which is generally dry except for sparse evergreentrees. The sandveld area is quite flat and dominated by free draining coarse to loamy fine Kalahari sands with high permeability and relatively low water holding capacity so the surface runoff is negligible there. In the hardveld part the surface runoff is more pronounced due to the diverse relief, solid rocks at shallow depth and less permeable soils enriched in clay materials originated from weathering of basalt and dolerite outcropping rocks. 3 MONITORING WITH ADAS The automated data acquisition system (ADAS), is a combination of sensors or just only one sensor installed in the field and operated by a multiple or single channel logger managing the performance of the sensors. ADAS are very useful in hydrology because they provide high temporal data resolution so they are well applicable in setting and calibration of transient models. The simplest example of ADAS is a combination of one sensor with one logger such as e.g. discussed below automated groundwater table recorder (AGTR). More sophisticated version is a multi-sensor ADAS operated by one, multi-channel logger. Such systems composed of various combinations of electronic sensors are usually mounted as towers on the masts (Lubczynski 2000) and can focus on monitoring of above-surface, surface, unsaturated zone and saturated zone temporal variability. The appropriate selection of the sensors and the programming of the loggers

Water resources of arid areas

350

depend on the objectives of the monitoring and both are critical for the success of the hydrological investigation program determining also the cost-effectiveness of that program. In the study area there are eleven multi-sensor ADAS towers named GS00 to GS10. Two of them are 18m high, GS00 on hardveld and GS10 on sandveld, one on sandveld, GS09 is 10m high and eight on sandveld GS01–GS08 are 2m high (Figure 1). The concentration of most of the ADAS towers on the sandveld illustrates current research focus of the project. The ten towers GS00–GS08 and GS10 are installed permanently whereas the GS09 is a mobile, retractable mast tower, which if not in mobile campaign (Obakeng & Lubczynski 2004 in the same issue) then it is temporally located as indicated in Figure 1. The multi-sensor ADAS towers provide input for assessment of rainfall, potential evapotranspiration, actual evapotranspiration, transpiration, unsaturated zone moisture and suction pressure (Lubczynski 2000). 3.1 Rainfall monitoring There are no perennial rivers in the study area so the recharge originates mainly from rainfall. The rainfall is monitored by ten tipping bucket rain gauges of Wallingford type characterized by nominal resolution of 0.2mm per tip. All the rain gauges are installed at the ADAS tower’s locations (Figure 1) at the height of 1.2m above the ground and all of them acquire data with 0.5-hour resolution. Additionally, in the Serowe village, there is one more rain gauge belonging to meteorological department, where rainfall has been recorded daily since 1922. That record indicates high, temporal variation ranging from 200mm/y (1991/92) to more than 1100mm/y (1997/98 and 1999/2000). Considering seasonal rainfall variation, in the study area, typical dry cold season starts in May and lasts to September. The rainfall in that period is negligible. The rainy hot season starts in October and usually lasts till April with rainfall peak in January. In the wet season rains occur in the form of isolated, very high intensity, localized and short duration storms (sometimes even of more than 100mm/d), which constitute the principal source of groundwater recharge. The high intensity and localized storms contribute not only to large temporal but also to the large spatial rainfall variability in the study area. 3.2 Monitoring of climatic components for potential and actual evapotranspiration The objective of monitoring of actual and potential evapotranspiration in the Kalahari Monitoring project is to provide in a cost efficient way a support for modeling of subsurface fluxes. This research therefore, is not oriented towards the most accurate and expensive evapotranspiration solutions such as e.g. eddy covariance method but instead, to develop and verify on Kalahari the methodology, which with the given cost of micrometeorological instrumentation and reasonable accuracy provides the maximal spatial output coverage. The first phase and the first results of that research are reported by Obakeng & Lubczynski (2004) in the same issue.

Monitoring and modeling of fluxes on Kalahari

351

3.3 Monitoring of unsaturated zone moisture and suction pressure The main objective in monitoring of unsaturated zone moisture and suction pressure at various depth levels (profiles) is to provide a temporal data support for flux simulation models (at this time mainly on Kalahari sandveld) and to answer to which extent recharge passes the Kalahari sand. In the study area there are seven moisture-suction pressure profiles in seven ADAS locations GS01–GS07 (Figure 1). Each profile consists of 4 dielectric, soil moisture sensors and 4 gypsum block suction pressure sensors installed in pairs at 2m, 4m, 6m and 8m below ground surface (b.g.s.). In each of the seven sites the shallow moisture is additionally recorded at 0.5m b.g.s. In order to investigate recharge at large depths on Kalahari sandveld, an additional deep suction pressure profile has recently been installed at the GS10 location characterized by absence of basalt and therefore unconfined aquifer conditions (Figure 1). That deep monitoring profile consists of 15 gypsum block suction pressure sensors distributed in logarithmically increasing with depth intervals starting at 0.25m and ending at 76m b.g.s., just above the groundwater table level. 3.4 Monitoring of transpiration by sap flow measurement The presence of green vegetation in dry season on Kalahari as well as the recent information about the deep tree rooting systems on Kalahari reaching up to 60–70m b.g.s. (Le Maitre et al. 2000) in an environment where recharge typically is very low, in order of few millimeters per year, points at the importance of tree transpiration on Kalahari. The tree transpiration in the study area is accessed by sap flow measurements. The sap flow (Qs) is defined as a product of sap velocity (ν) and sap wood (xylem) area (Ax). The ν is monitored in the study area by thermal dissipation probes (TDP) following Granier’s method (Granier 1987). In total there are 51, sap velocity monitoring points in the study area, six in each of the eight ADAS locations GS00-GS07 and three in the GS08. They cover most of the variety of the tree species in the study area. The Ax of the monitored trees is considered as time invariant at least in the time frame of the Kalahari Monitoring project and was estimated from the biometric characteristics established for each species separately in the transpiration monitoring campaigns (Lubczynski et al. 2004—in the same conference issue). 3.5 Monitoring of groundwater Monitoring of groundwater table provides direct response of the aquifer to recharge or discharge of groundwater including the most important hydrogeological information on groundwater flux regime. Groundwater monitoring as a standard is nowadays provided by automated groundwater table recorders (AGTR). All the AGTRs used in the study area are based on the principle of recording hydrostatic pressure above the sensor suspended in the groundwater of the well. Groundwater monitoring network in the study area consists of twenty-one well measurement points. There are three differential (automatically compensating for

Water resources of arid areas

352

barometric pressure) “Troll” AGTRs installed by the Department of Water Affairs of Botswana (DWA), three absolute (compensated by the external barometric pressure measurement) “Tirta” AGTRs’ and one “Diver” AGTR installed by ITC, and thirteen absolute “Diver” AGTRs’ installed by the Geological Survey of Botswana (GS). All the AGTRs are programmed to acquire data at one-hour resolution. Additionally, there is one more GS manually dipped groundwater table monitoring point with monthly data acquisition. All the groundwater table-monitoring points are also monthly tested with regard to the basic ionic, hydrochemical components of groundwater. 4 SPATIAL FLUX ASSESSMENT Groundwater fluxes such as recharge and groundwater evapotranspiration vary not only temporally but also spatially. The spatial distribution of groundwater recharge was first evaluated in the study area by groundwater modeling (Lubczynski 2000, WCS 2000). Later two series of chloride data for recharge assessment were collected from the wells, all over the study area, first by Obakeng (2000) and next by Magombedze (2002) and assessed spatially by interpolation and also by extrapolation applying integrated GIS modeling. The summary of those approaches is presented by Magombedze et al. (2004) in the same conference issue. An assessment of groundwater evapotranspiration (Eg) is a very difficult issue. A first attempt to determine Eg spatially in the study area was made by Lubczynski (2000). For that purpose he applied groundwater modeling in which Eg was considered as state variable with spatial distribution derived from RS solution of energy balance (Timmermans & Meijerink 2000). Certainly this was not the ideal procedure since Eg fluxes were small and likely comparable with the eventual error of the calibrated model. The recent attempts in defining Eg, lead through the determination of its tree transpiration (T) component applying sap flow measurements. The methodology of sap flow measurements on Kalahari and plot level upscaling is discussed in Fregoso (2002), in Mapanda (2003) and is finally summarized in Lubczynski et al. (2004) in the same conference issue. The RS upscaling of sap flow measurements for the 10×10km experimental area covered by multispectral IKONOS image (Figure 1), was attempted by Fregoso (2002), by Mapanda (2003) and by Keeletsang (2004). A similar attempt for the same study area but using multi-band TETRACAM digital camera built on the aircraft is described in Hussin et al. (2004a) whereas the multi-band aerial-photography aircraft mission itself is described in Hussin et al. (2004b), both in the same conference issue. Due to the difficulties in classification of tree species, closely related to the large biodiversity, on Kalahari the RS upscaling protocol is still being improved. The transpiration mapping by RS upscaling of sap flow measurements unfortunately does not provide the estimation of the demanded in groundwater management Eg but provides T. Equalizing the two is only possible if two critical assumptions are fulfilled, first, that considering large depth of groundwater table in the study area, the Cg is negligible or definable and the second, that the Tg component of T, can be separated from unsaturated zone root water uptake (Tu). The first assumption will be tested by setting up 1-D saturated-unsaturated models (see below) for each unsaturated moisture and suction pressure monitoring profile available in the study area. The action with regard to the

Monitoring and modeling of fluxes on Kalahari

353

second assumption, based on the species-specific partitioning of transpiration, is currently assessed in the study area by labelling of groundwater with Li+, H+2 and O+18 tracers following the methodology proposed by Haase et al. (1996). 5 SPATIO-TEMPORAL INTEGRATION OF GROUNDWATER FLUXES Groundwater fluxes such as recharge and groundwater evapotranspiration are highly spatially and temporally variable on Kalahari. If the depth-wise, spatio-temporal data is available, then such data can be assessed with regard to temporal and depth dependent flux regime by using either complex coupled flow models such as MIKE-SHE (DHI 1993) applicable rather to areas in scale of hectares or as proposed in this study by using semi-coupled modeling procedures combining information from different models. The 1-D numerical models such as EARTH lump parameter model (Van der Lee & Gehrels) or more complex HYDRUS (Simunek et al. 1998) based on the Richard’s equation are efficient because they are relatively simple. The disadvantage of all 1-D models however is that they do not account for lateral fluxes which implies the additional non-uniqueness of such models. For example groundwater table rise can be originated either from direct rainfall recharge or from lateral inflow recharged elsewhere and also groundwater table decline can be either resulted by groundwater evapotranspiration or by lateral groundwater outflow. Such non-uniqueness in assessment of groundwater fluxes affects less distributed parameter watershed models such as e.g. SWAT (Arnold et al. 1993), that generate as output a groundwater recharge further applicable as spatiotemporally variable net recharge in groundwater model such as e.g. MODFLOW (McDonald & Harbaugh 1996). Sophocleous & Perkins (2000) have successfully linked SWAT with MODFLOW. If spatio-temporal knowledge of the R and Eg is of concern, the recharge and groundwater evapotranspiration have to be reassigned and calibrated in groundwater model. This can be done following guidelines of 1-D saturated-unsaturated models. The ideal situation in that respect is when 1-D monitoring profiles or at least groundwater table monitoring points are available for each zone of spatial flux variability of R and Eg. The 1-D models allow for better understanding of groundwater regime and for reasonable simulation of temporal flux variability. Once calibrated, the 1-D flux variability can further be implemented and adjusted in MODFLOW. In such modeling procedure the most efficient way of flux adjustment is by using the automated calibration techniques such as PEST (optimization technique). This technique provides the option of parameter and flux optimization within the predefined variability ranges and with automated assessment of uncertainty. In the Serowe study area there is already a numerical groundwater MODFLOW model available with spatially variable but time invariant R and Eg fluxes. This model was calibrated in transient mode with regard to the regionally expanding and measured in boreholes drawdowns, developed in response to the increased in last years’ well abstraction. The extensive monitoring network installed within Kalahari Monitoring project, generate large amount of high temporal resolution data, and therefore provides the opportunity to upgrade the Serowe model calibration to the stage characterized by spatio-temporally variable fluxes. Such models are more reliable with regard to the

Water resources of arid areas

354

applied parameters and as explained by Lubczynski and Gurwin (2004) for Sardon granite catchment in Spain, they can provide not only the prediction scenarios but also accurate information on where, when and at which rates fluxes such as recharge and groundwater evapotranspiration occurred in the analyzed area. The availability of historical record of rainfall in Serowe village starting in 1922, allows to run backward model scenario, which will finally allow to provide the demanded in Botswana long-term temporal characteristic of recharge on Kalahari. 6 CONCLUSIONS The acquisition of temporal data with ADAS provides unique opportunity for direct temporal measurement of various flux processes such as rainfall and transpiration. Other processes such as recharge and evapotranspiration cannot be measured directly but have to be modeled. ADAS provides full data acquisition support for such models. The integration of temporal data from ADAS with the spatial data extrapolated with GIS and RS techniques in numerical models provides the opportunity of model calibration with spatio-temporally variable fluxes. In semi-arid and arid climates only models calibrated with spatio-temporally variable fluxes can provide a reliable system parameterization, reliable spatio-temporal flux regimes and reliable flux rates. This means, that such models provide the optimal tool for groundwater management. ACKNOWLEDGEMENTS We acknowledge Geological Survey of Botswana for financial support and extensive help in sap flow field campaigns. In particular we would like to thank Mr Phofutsile for his support to the project and Mr Ramatsoko and his field crew for the extensive professional and logistical help in the field. REFERENCES Arnold, J.G., Williams, J.R., Srinivasan, R., King, K.W. & Griggs, R.H. 1994. SWAT (Soil and Water Assessment Tool) user’s manual. USDA, Agricultural Research Service, Grassland, Soil and Water Research Laboratory, Temple, TX. DHI—Danish Hydraulic Institute. 1993. MIKE SHE water movement-user’s guide and technical manual, ed.1.0 DHI, Denmark, pp. 81. De Vries, J.J., Selaolo, E.T. & Beekman, H.E. 2000. Groundwater recharge in the Kalahari, with reference to paleo-hydrologic conditions. Journal of Hydrology 238, 110–123. Doherty, J. 2000. PEST—Model-Independent parameter estimation. User’s manual. Watermark Computing, Australia. Fregoso, A. 2002. Dry-season transpiration of savanna vegetation. Assessment of tree transpiration and its spatial distribution in Serowe, Botswana. MSc thesis, Library of ITC— International Institute for Geoinformation Science and Earth Observation, Enschede, The Netherlands.

Monitoring and modeling of fluxes on Kalahari

355

Fregoso, A., Chavarro, A. & Lubczynski, M.W. 2004. Sap flow measurements of tree transpiration on Kalahari, Serowe study case, Botswana. Proc. WRASRA conf. Gaborone 3–7 August, 2004, Rotterdam, Balkema. Granier, A. 1987. Evaluation of transpiration in Douglas-fir stand by means of sap flow measurements. Tree Physiology 3:309–320. Haase P., Pugnaire, F.I., Fernandez, E.M., Puigdefabregas, J., Clark, S.C. & Incoll, L.D. 1996. An investigation of rooting depth of the semiarid shrub Retama sphaerocarpa (L.) Boiss. By labeling of groundwater with a chemical tracer. Journal of Hydrology 177:23–31. Hussin, Y.A., Chavarro, D. Lubczynski, M.W. & Obakeng O. 2004a. Mapping vegetation for upscaling evapo-transpiration using high-resolution optical satellite and aircraft images in Serowe, Botswana. Proc. WRASRA conf. Gaborone 3–7 August 2004, Rotterdam, Balkema. Hussin, Y.A., Lubczynski, M.W. & Obakeng, O. 2004b. Designing and implementing an aircraft survey mission using high-resolution digital multi-spectral camera for vegetation mapping for up-scaling evapotranspiration of Serowe, Botswana. Proc. WRASRA conf. Gaborone 3–7 August 2004, Rotterdam, Balkema. Keeletsang, M. 2004. Assessment of dry season transpiration using IKONOS images, Serowe case study, Botswana. MSc thesis, Library of ITC—International Institute for Geoinformation Science and Earth Observation, Enschede, The Netherlands. Le Maitre, D.C., Scott, D.F. & Colvin, C. 2000. Information on interactions between groundwater and vegetation relevant to South African conditions: A review. In: Past Achievements and Future Challenges. Balkema, ISBN 9058091597, Rotterdam, 959–961. Lubczynski, M.W., 2000. Ground water evapotranspiration—underestimated component of groundwater balance in a semi-arid environment—Serowe case Botswana. In: Past Achievements and Future Challenges. Balkema, ISBN 9058091597, Rotterdam, 199–204. Lubczynski, M.W. & Gurwin, J. 2004. Integration of various data sources for transient groundwater modeling—Sardon study case, Spain. Journal of Hydrology—in revision. Lubczynski, M.W., Fregoso, A., Mapanda, W., Ziwa, C, Keeletsang, M., Chavarro, D.C. & Obakeng O. 2004. Dry season Kalahari sap flow measurements for tree transpiration mapping— Serowe study case, Botswana. Proc. WRASRA conf, Gaborone, 3–7 August 2004, Rotterdam, Balkema. Magombedze, L.M. 2002. Spatial and temporal variability of groundwater fluxes in a semi-arid environment—Serowe, Botswana. MSc thesis, Library of ITC—International Institute for Geoinformation Science and Earth Observation, Enschede, The Netherlands. Magombedze, L.M., Frengstad, B. & Lubczynski, M.W. 2004. Spatial variation of groundwater recharge in a semi-arid environment—Serowe, Botswana. Proc. WRASRA conf, Gaborone 3–7 August 2004, Rotterdam, Rotterdam. Mapanda, W. 2003. Scaling-up tree transpiration of eastern Kalahari sandveldof Botswana using remote sensing and geographical information system. McDonald, M.D. & Harbaugh A.W. 1996. A modular three-dimensional finite difference groundwater flow model. Washington, D.C., U.S. Geological Survey. Obakeng, O.T., 2000. Groundwater recharge and vulnerability: A case study at the margins of the south-east Central Kalahari Sub-basin, Serowe region, Botswana. MSc thesis, Library of ITC— International Institute for Geoinformation Science and Earth Observation, Enschede, The Netherlands. Obakeng, O.T. & Lubczynski, M.W. 2004. Monitoring of evapotranspiration on Kalahari, Serowe case study, Botswana. Proc. WRASRA conf. Gaborone 3–7 August 2004, Rotterdam, Balkema. SGAB, 1988. Serowe groundwater resources evaluation. Swedish Geological Co report for Department Geological Survey, Botswana. Simunek, J., Sejna M. & van Genuchten M.T. 1998. The HYDRUS-1D software package for simulating the one-dimensional movement of water, heat and multiple solutes in variably saturated media, version 2.0, IGWMC—TPS—70, 2002pp., Colorado School of Mines, Golden Colorado.

Water resources of arid areas

356

Sophocleous, M. & Perkins, S.P. 2000. Methodology and application of combined watershed and groundwater models in Kansas. Journal of Hydrology 236:185–201. Timmermans, W. & Meijerink, A., 2000. Remotely sensed actual evapotranspiration; implications for groundwater management in Botswana. In: JAG: International Journal of Applied Earth Observation and Geoinformation, 1(1999)3/4, 222–233. Van der Lee, J. & Gehrels, J., 1990. Modelling aquifer recharge—Introduction to the Lumped Parameter Model EARTH. Free University of Amsterdam, The Netherlands. WCS—Wellfield Consulting Services, 2000. Serowe wellfield extension project. DWA, Department of Water Affairs, Report No. TB10/3/10/95–96.

Geoelectrical investigation for aquifer delineation in the semi-arid Chad Basin, Nigeria A.Iliya1 & E.M.Shemang2 1

Rural Water Supply and Sanitation Agency (RUWASA) Damaturu, Nigeria 2 Department of Geology, University of Botswana, Gaborone, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Aquifer characterization of the SW, Chad Basin on the basis of geo-electrical investigation was carried out. Forty two vertical electrical sounding (VES) were carried out using the Schlumberger array technique. The results of the VES interpretation revealed that three distinct geoelectric layers (surface unit; shallow conductive unit; and deep resistant unit) can be identified. The surface unit whose resistivity range between 7Ωm to 64Ωm is 10–25m thick and appear to be discontinuous. The shallow conductive unit whose thickness is of the order of 16–22m has resistivity range of 317Ωm to 499Ωm and is thought to correlate with the Paleocene Kerri-Kerri Formation. The deep resistant unit whose thickness range between 98m–322m shows resistivity range of 899Ωm to 1927Ωm and appear to be present throughout the study area. The last two units are thought to water bearing. Based on the interpreted results, aquifer Transverse resistance (T) and Longitudinal conductance (S) were also computed and on these bases the study area was subdivided into three zones.

1 INTRODUCTION The rapid increase in urbanization as well as industrial and agricultural expansion has focused attention upon the diminishing volume of available groundwater in most major urban centres within the sahelian zones of West Africa. The study area, which falls within the Yobe portion of the Chad Basin (Fig. 1) comprises major towns as Damaturu, Potiskum, Nguru, Gashua, and Geidam, etc. whose combined population is about two million. As a young state, rapid expansion in industrialization, socioeconomic growth, etc. is expected and hence the need to address the problem of groundwater resources evaluation and management. This groundwater development stage, however, can be viewed as a sequential process consisting of exploration, evaluation and exploitation. The

Water resources of arid areas

358

exploration stage in which surface and subsurface geological and geophysical techniques are utilized to search for suitable aquifers involve the use of electrical resistivity (VES) survey and borehole data. The evaluation stage, however, comprises the measurement of hydrogeological parameters, calculation of aquifer yields (Transmissivity, Storativity,) as well as hydrogeochemical analysis of water samples collected from boreholes in the study area. The exploitation or management stage includes the consideration of optimal development strategies and assessment of the interactions between groundwater exploitation and regional hydrological system. The present study is therefore, aimed at delineating the aquifer system (s) through the use of surface electrical resistivity techniques thereby pin pointing possible productive zones, correlate aquifer hydraulic properties with those obtained from VES and hydro geochemical data. It is also aimed at forecasting the future water requirements/utilization of the study area. To achieve these objectives therefore, forty two (42) vertical electrical soundings (VES) using the Schlumberger configuration with a maximum total current electrode separation of 1000m was carried out in the study area. Some of these have already been confirmed through drilling of boreholes.

Figure 1. Map of Nigeria showing the location of the study area.

Geoelectrical investigation for aquifer delineation in the semi-arid Chad Basin, Nigeria

359

2 REVIEW OF GEOLOGY The study area located between Longitude 11°N and 13°N and Latitude 11°E and 13°E (Fig.1) is composed of the Chad Formation outcrops. In general, it consists of successions of sands, clays, sandy clays and silts with interbedded lenses and layers of sands and gravels of various levels. The deposits are generally of lacustrine origin or were formed during periods when rivers had very low discharges because of climatic and geomorphological conditions. The beds dip gently towards the centre of the basin not only because of their original attitude but also mild regional tectonic movements which have affected the basin in recent times. The Chad Formation may reach a thickness of 600–700m in the central part of the basin (Offodile, 1992) but thins out rapidly towards the edges. Such a very abrupt reduction in thickness of sediments near the margins of the Plio-quarternary lake basin could well be the result of step faulting in the basement rocks of the char depression. This is illustrated in the lithological data from boreholes (Fig. 2). The products of such activity are sometimes found at the base of the Chad Formation, as in the case of granitic rocks encountered in boreholes in Goniri and environs. These may well be associated with the faults bounding the Chad basin and may to some extent be contemporaneous with the deposition of the Chad Formation. Data collected so far indicate that the lithostratigraphic sequence in the study area consists of sandy clay alteration of the Chad Formation that very probably lies directly on the Basement Complex rocks. 3 HYDROGEOLOGY The Chad Basin is described as the largest area of inland drainage in Africa and occupies parts of Nigeria, Central African Republic and Cameroon. The Nigerian sector of the basin slopes gently towards the Lake Chad which is the main geographical feature.

Water resources of arid areas

360

Figure 2. Typical lithologic profile of study area. As far as groundwater is concerned, the most important formations in the basin are the Chad and Kerri-Kerri formations which are characteristically Pleistocene (Pliocene) and

Geoelectrical investigation for aquifer delineation in the semi-arid Chad Basin, Nigeria

361

Paleocene respectively. Surface water in streams appears seasonally usually from August to November. For the rest of the year, the streams are dry and the only source of water is groundwater. This is contained in the three aquifer systems designated as Upper, Middle and Lower Aquifers (Kogbe et al., 1992) especially in Maiduguri. While the Lower is considerably deep (over 500m) are tapped by few boreholes, the Middle and Upper aquifers are on average depth of some 250m and 40m respectively are obviously over exploited and on many cases have dried-up. This had already focused attention on the possibility of a perchy aquifer for the Upper aquifer in the Chad basin (Kogbe et al., 1992). It is worth while mentioning, however, that the above multi-aquifer systems do not extend throughout the Chad Basin. In Damaturu area, which lies on the edge of the Chad Formation lake basin, the hydrogeological situation may be summarized as follows: a) Total thickness of the Chad Formation is about 130–170m b) Marked discontinuity of water-bearing levels. c) Vertical and lateral changes in their hydraulic properties of water-bearing levels. d) Presence of perchy aquifers where impervious layer levels occur in the upper part of the formation.

Figure 3. Typical Interpretation of vertical electrical sounding curve Azbak VES1. e) Presence of two artesian aquifers, consisting of fine to medium grained sands, at a Depth varying from 30 to 70m and from 90 to 120m. f) The static water levels (SWL) of the aquifers range between 30 to 60m. g) Discharge of most boreholes range between 4–15 liters/sec.

Water resources of arid areas

362

4 DATA ACQUISITION AND INTERPRETATION Forty two (42) vertical electrical soundings (VES) using the Schlumberger array configuration with a minimum and maximum current electrode separation of 320 and 1000m respectively. The equipment used was the ABEM SAS 300B Terrameter. Sounding was carried out with aim of selecting sites for water supplies to villages and points were therefore located in and around villages. The VES data was first interpreted using the conventional curve matching techniques and later using the IPI2WIN software. Figure 3 shows an interpretation of VES 1 sounding carried out in the area of study. 5 RESULTS AND CONCLUSION The result of preliminary assessment of groundwater resources of SW Chad Basin on the basis of surface geophysical and hydrogeological investigation suggests that the surface unit whose resistivity range between 7Ωm to 64Ωm is 10–25m thick and appear to be discontinuous. The shallow conductive unit whose thickness is of the order of 16–22m has resistivity range of 317Ωm to 499Ωm and is thought to correlate with the Paleocene Kerri-Kerri Formation. The deep resistant whose thickness range between 98–322m shows resistivity range of 899Ωm to 1927Ωm and appear to be present throughout the study area. The last two units are thought to water bearing. Based on the interpreted results, aquifer. Transverse resistance (T) and Longitudinal conductance (S) were also computed. The results of these led to the subdivision of the area into three zones.

Figure 4. Isoresistivity map of the third layer in the area of study.

Geoelectrical investigation for aquifer delineation in the semi-arid Chad Basin, Nigeria

363

An Isoresistivity map of the third layer was constructed and the results show that the area can be separated two zones, the Western and Eastern zones. The boundary between these two zones probably suggest the contact between two lithological units in the area (the Kerri-Kerri formation and the Chad formation) REFERENCES Bunu, Z.M. and Iliya, A.G. (1992) Understanding the Rainfall Pattern of a Semi-Arid Region: A case study of Maiduguri. Paper presented at the Fifth National Conference of the Nigerian Association of Hydrogeologists, Shiroro Hotel, Minna, Nigeria. Carter, J.D. Barber, W. and Tait, E.A. (1963) Geology of Adamawa, Bauchi and Bornu Provinces in Northeastern Niogeria. Bull. Geol Surv Nigeria 30, 1–108. Cratchley, C.R. (1960) Geophysical Survey of the Southwest Part of the Chad Basin, C.C.T.A. Publication No. 13. Kogbe, C.A. Schoeneich, K. and Ebah, E.I. (1992) Hydrogeological Framework of Maiduguri Metropolis in the Chad Basin, NE, Nigeria. Paper presented at the fifth Conference of the Nigerian Association of Hydrogeologists, Shiroro Hotel, Minna, Nigeria. Matheis, G. (1965) Short Review of the Geology of the Chad Basin in Nigeria. Journal of Mining and Geology, 289–294. Offodile, M.E. (1992) An Approach to Groundwater Study and Development in Nigeria, Mecon Services Ltd. 300pp.

Monitoring of evapotranspiration on Kalahari, Serowe case study, Botswana O.Obakeng1,2 & M.W.Lubczynski2 1

Geological Survey of Botswana The International Institute for Geoinformation Science and Earth Observation (ITC), Enschede, Netherlands

2

Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The estimation of evapotranspiration under natural conditions at different spatial and temporal scales is crucial for water management. Knowledge of evapotranspiration is also needed in water transport models. In this study, a discussion of the driving variables is first given followed by estimation of actual evapotranspiration rates using the energy balance equation in which the sensible heat flux density is derived from temperature profiles. These estimations are performed over two typical, but different Botswana environments; Kalahari (sandveld) and hardveld areas at two high tower locations equipped with micrometeorological sensors. The actual evapotranspiration rates for the Kalahari sandveld are 0.01–2.09mm/day, and for the hardveld area 0.01– 3.74mm/day. Finally, an attempt is made to correlate the wind speed data of permanent stations with the wind speed obtained from the 10m high mobile tower that is moved between the locations of other towers not equipped with anemometers, primarily for calculation of potential evapotranspiration at these sites. The potential evapotranspiration calculated as a result of that experiment was largely variable and ranged from 0.01–8.11mm/day.

1 INTRODUCTION As 70% of the precipitation depth may evaporate annually in semi-arid climates, careful consideration should be given to the determination of actual evapotranspiration, as well as potential evapotranspiration. Estimates of potential evapotranspiration in semi arid climates are an order of magnitude greater than the rainfall depth. In the Serowe study area the annual potential evapotranspiration amounts to 1350–1450mm (Choudhury, 1997) and the mean annual rainfall is 447mm. Consequently, the actual

Monitoring of evapotranspiration on Kalahari, Serowe case study, Botswana

365

evapotranspiration rates are much smaller than the potential rates because of the limited amount of water stored in the topsoil. Evapotranspiration (ET) plays an important role in a groundwater balance in semi arid climates, as demonstrated by Lubczynski (2000). In a general sense the groundwater balance equation can be written in the following form Qin+R−Qout−Eg−A±∆S=0 (1) where Qin=groundwater inflow, R=groundwater recharge, Qout=groundwater outflow, Eg=groundwater evapotranspiration and A=well abstraction (the injection well would have opposite sign and then would be considered as input) as output. The ±∆S=change of groundwater storage. Evapotranspiration has either direct or indirect impact upon groundwater resources. The direct impact relates to groundwater evapotranspiration attributed mainly to direct water extraction by deep root systems of savanna vegetation. The indirect impact relates to the loss of water in the unsaturated zone, which reduces groundwater recharge due to reduction of the unsaturated hydraulic conductivity in the upper soil layers. Within the framework of an on-going research project titled Kalahari Research Programme, a network of monitoring stations were established primarily to monitor components of the groundwater water balance, which include evapotranspiration. This paper discusses preliminary evapotranspiration rates found for the Kalahari (mainly) and hardveld areas of the Serowe study area. 2 GENERAL DESCRIPTION OF THE STUDY AREA The Serowe area has in general, a gentle topography, which varies from ≈1020m.a.s.l. to ≈1240m.a.s.l (Fig. 1). A major geomorphic feature within Serowe area is the escarpment, which forms part of a geologically recent axis of uplift known as the Zimbabwe-Kalahari axis (Smith, 1984). It represents the eastern limit of the Kalahari sandveld. All rivers are ephemeral, and flow occurs only during exceptionally high rainfall events of the annual wet season. Otherwise, they are dry for most of the year, with groundwater levels often situated at shallow depths (4–6m) beneath riverbeds. The surface topography is lower in the E and SE of the region, higher in the western Kalahari plateau and the highest on the NW side of the escarpment edge, which is a prominent topographic feature in this area. SE from the escarpment, the average slope is 5%, and it gradually decreases to less than 1% towards the E and SE. Rock outcrops are found mainly at the escarpment and along river valleys below the escarpment. Elsewhere Kalahari sands and superficial deposits overlie rocks. The soil types found in the study area are arenosols, regosols, lixisols, luvisols and vertisols. Arenosols are by far the most common soil unit covering most of the Serowe area to the west, north, south and the extreme east. The climate of Serowe study area, like in other parts of Botswana, is characterized as semi-arid, with a mean annual rainfall of 447mm (SGS, 1988). Rainfall is seasonal, with the highest intensity in summer followed by a dry winter period. The summer stretches

Water resources of arid areas

366

from October to April whereas the winter begins in May and ends in August (Bhalotra, 1987). The main type of vegetation in the study area is thought to belong to the Northern Kalahari Tree and Bush Savanna (Vossen, 1989; Nash, 1992), despite the existence of significant spatial

Figure 1. Topographic map of Serowe area and aerial distribution of automatic data acquisition stations (ADAS). variations in species and community members. Within the Serowe study area, Ecosurv Botswana (1998) and Hernandez (2002) identified 4–10 vegetation communities. The western part of the study area (sandveld) is quite homogeneous with regard to species composition. Species such as Terminalia sericea, Ochna pulchra and Boscia albitrunca are strongly represented there. In the eastern part of the study area (hardveld) vegetation is generally taller particularly along and in the vicinity of river courses and depressions. Species such as Acacia karoo, Acacia tortilis and Acacia mellifera are strongly represented in the hardveld. At the escarpment edge the vegetation is taller, denser and more diverse than in the rest of the study area. Species like Combretum apiculatum, Croton gratissimus and Ricinodrendrum rutananii are strongly represented there. Terminalia sericea, Dicrostachys cinerea, Grewia retinervis and Combretum apiculatum are found everywhere across the study area. 3 THE EVAPOTRANSPIRATION-MONITORING NETWORK The evapotranspiration-monitoring network consists of ten automated data acquisition station (ADAS) towers of various sensor configurations and mast heights. This network consists of:

Monitoring of evapotranspiration on Kalahari, Serowe case study, Botswana

367

GS00—the most versatile, 18m high, galaxy ADAS tower, installed on the hardveld in September 2001 (Fig. 1). It is equipped with one net radiometer CNR1 installed at the top of the tower construction, 3 anemometers and 3 relative humidity & temperature (RH/T) sensors all installed at 2m, 13m, and 18m heights. In addition there are also: soil heat flux plates buried at a depth of 2cm, two soil temperature sensors buried at depths of 2cm and 15cm, a tipping bucket rain gauge raised to a height of 1.2m above the ground and 6 sap flow sensors of Granier type measuring velocity of water transport in tree stems. The data acquisition and the data storage with 0.5h resolution is managed by the Delta-T logger. GS01 to GS07—there are seven of the same, 2m high ADAS towers in the study area (Fig. 1). Each of these towers is equipped with one RH/T sensors attached to a mast at a height of 2m above the ground surface, a tipping bucket rain gauge raised to a height of 1.2m above the ground, 6 sap flow sensors, one soil moisture and one soil suction pressure profile with sensors at 0.25m, 2m, 4m, 6m, 8m. The Skye DataHog2 logger logs the data at half hour intervals in all seven towers. GS08—this tower was installed as a backup of the tower GS00 in case of its failure (Fig. 1). It consists of anemometer, RH/T sensors and radiometer CM3 for measuring incoming short-wave radiation all mounted at the height of 2m above the ground surface. Other instruments include two soil temperature sensors buried at depths of 2cm and 15cm below the ground surface and a tipping bucket rain gauge raised to a height of 1.2m above the ground and 6 sap flow sensors. The Skye DataHog2 logger logs the data at half an hour interval. GS09—this is a mobile, retractable 10m high tower, equipped with two anemometers, two RH/T sensors, installed at 2m and 10m heights, a pair of soil temperature sensors buried at depths of 2cm and 15cm in the soil and 6 sap flow sensors. During field campaigns GS09 is moved between stations GS01-GS08 every ten-days, otherwise it is fixed at its semi-permanent location (Fig. 1). The Skye DataHog2 logger logs the data at half an hour interval. 4 THEORETICAL BACKGROUND OF ET DETERMINATION Many methods exist for estimating actual evapotranspiration (e.g. Bastiaanssen, 1995) and potential evapotranspiration (e.g. Hargreaves and Samani, 1985) using micrometeorological measurements. In this study actual evapotranspiration (AET) was computed from the energy balance equation in which soil heat flux (G) and the net radiation (Rn) were considered as known (measured) and the sensible heat flux (H) was calculated using the temperature profile method (Holtslag and Ulden, 1983). The potential evapotranspiration was calculated with the FAO Penman-Monteith formula (Allen et al., 1998). 4.1 Calculation of H by temperature profile (T-profile) method H is related to friction velocity (u*) and the temperature scale (θ*) by H=−ρCpu*θ* (2)

Water resources of arid areas

368

where Cp (kJ kg−1 °C−1) is the specific heat capacity of air taken as 1013 kJ kg−1 °C−1, ρ (kg m−3) is air density. The effect of the modification of forced convection by temperature gradients on momentum and heat (and water vapor) transfer can be corrected by dimensionless parameters. One of the widely used stability parameter is known as the Monin-Obukhov correction factor. The Monin-Obukhov length L (m) is given by (3) where k is von Karman’s constant (0.41), g (ms−2) is acceleration due to gravity and T (°C) is mean air temperature. When L is greater than 0 stable atmospheric conditions exist and when L is less than 0 unstable atmospheric conditions prevail otherwise the conditions are neutral. A simplified method for determination of momentum flux and sensible heat flux (H) which requires wind speed (uz) (ms−1) at level z(m), a surface roughness length (z0) and a temperature difference ∆θ (K) between two heights z1(m) and z2(m) in the atmospheric surface layer as input is provided by Holtslag and Ulden (1983). In this method the integrated flux-profile relations of Dyer and Hicks, 1970 are used to calculate u* and θ* from the aforementioned parameters according to (4)

(5)

The integrated stability correction function for heat transfer (ψh) and momentum transfer (ψm) for unstable conditions (L<0) can be estimated from Equations 6 and 7 respectively (6)

(7) where x is given by (8) For stable atmospheric conditions (L>0) ψm and ψh are given by

Monitoring of evapotranspiration on Kalahari, Serowe case study, Botswana

369

(9) The sensible heat flux (H) can be calculated from Equations (2)–(9), starting with a first guess for L (Monin-Obukhov stability parameter). With L=−5, then u* and θ* are calculated from Equations (4)–(5). Using Equation 3, L is calculated by using the estimated values of u* and θ*. The new value of L is substituted into Equations (4)–(5), primarily to get improved values for u* and θ*. This usually takes about 5 iterations, until the value of L do not change significantly (<5%). Then H is calculated with Equation 2. This scheme is referred hereafter as temperature profile method (T-profile). The surface roughness length (z0) was estimated from vegetation height, leaf area index and other data according to Raupach (1994). This method however, will not work when no temperature differences are observed between two measurement heights, a situation that was occasionally encountered in the present research work. Once H is known then the actual evapotranspiration can be calculated from the energy balance equation by applying as input, also soil heat flux and the net radiation, both directly measured in the study area. 4.2 Penman-Monteith evapotranspiration model One of the most frequently used evapotranspiration model is the Penman-Monteith model. It combines the energy balance with mass transfer method. According to this model actual evapotranspiration is calculated as (10)

where λE (MJ m−2day−1) is latent heat flux (evapotranspiration), Rn (MJ m−2 day−1) is net radiation, G (MJ m−2 day−1) is soil heat flux, γ (kPa °C−1) is the psychrometric constant, Cp (kJ kg−1 °C−1) is the specific heat capacity of air taken as 1013 kJ kg−1 °C−1, ∆(kPa °C−1) is the rate of change of the saturation vapor pressure with temperature, es (kPa) is the saturation vapor pressure, ρa is mean air density at constant pressure, ea (kPa) is the actual vapour pressure, rs (s m−1) and ra (s m−1) are surface and aerodynamic resistances respectively. Not only actual evapotranspiration but also FAO potential evapotranspiration (Allen et al., 1998) derived from the Penman-Monteith formula (Equation 10) is used as a standard in hydrology. PET represents water demand (stress) of the hydrological system being also the upper limit of evapotranspiration (E). The FAO formula is expressed

Water resources of arid areas

370

(11)

where PET (mm/day) is potential evapotranspiration, u2 (m s−1) and Ta (°C) are wind speed and mean daily air temperature at 2m respectively and other notations are as described earlier. Basic assumptions in the formulation of Equation 11 is that the surface resistance (rs)=70 (s m−1) and aerodynamic resistance (ra)=208/u2 (s m−1). The PET Penman-Monteith formula (unlike other potential evapotranspiration methods takes into account most parameters that affect evapotranspiration. Most of the parameters necessary to calculate PET in the study area according to Equation 11 were either available or could be defined by regression analysis. This allowed assessment of PET at GS01–GS08 ADAS locations. Similar assessment of E as per Equation 10 is by far more difficult because of ra and rs parameters. ra determines the transfer of heat, momentum and water vapour from an evaporating surface into the air above the vegetation canopy and is inversely proportional to wind speed and changes with height covering the ground (Maidment, 1993). The ra is expressed as (12)

where ra (s m−1) is aerodynamic resistance, d (m) is the zero plane displacement height, uz (m s−1) is wind speed at a measurement height z (m), zoh (m) is the surface roughness length for heat transfer and water vapor, which is approximated as 10% of zom, where zom (m) is defined as the roughness length for momentum transfer. d and zom can be estimated from other parameters following Raupach (1994). Several attempts are made in the literature to evaluate rs by means of empirical rules (e.g. O’Toole and Real, 1986). One such an attempt is the so-called Jarvis type models (Jarvis, 1976: Stewart, 1988: De Rooy & Holtslag, 1999), in which stomatal (canopy) resistance is expressed as a minimum rs multiplied by a series of independent stress functions combined in a multiplicative way, through which each function is representing the influence of each factor. The main weakness of Jarvis type models is the assumption that environmental factors operate independently (Monteith, 1995). Another way in which rs can be estimated is through the inversion of the Penman-Monteith equation (Equation 13), in which the actual evapotranspiration is considered as known input parameter (Gash & Stewart, 1975), obtained by other methods (e.g. the temperature profile and Bowen ratio approaches). (13) where the notations are as described earlier.

Monitoring of evapotranspiration on Kalahari, Serowe case study, Botswana

371

The surface resistance can also be estimated by substitution of stand transpiration (Ts) derived from sap flow measurements in place of λE in the inverted Penman-Monteith equation (Equation 13). This procedure is however practically valid only for dry season estimates of Ts when the assumption is E=Ts can be made. 5 ASSESSMENT OF BASIC METEOROLOGICAL VARIABLES As mentioned, in the study area there are a number of ADAS towers monitoring various hydrological variables. The most important with regard to evapotranspiration are: radiation, temperature, relative humidity and wind speed. 5.1 Radiation components All net radiation components such as short-wave incoming and outgoing radiations, longwave incoming and outgoing radiations are monitored only in GS00. Additionally, shortwave incoming radiation is monitored in GS08. Figure 2 illustrates a typical example of the diurnal course of the radiation components measured at GS00 site for the clear-sky day of 01/04/02. The presented net radiation was post-processed from the other radiation components. It can be observed that at noon, both the incoming short-wave and net radiation reached their maximum whilst the outgoing long-wave radiation and the outgoing short-wave radiation reached their lowest values at about the same time. The incoming long-wave radiation was more or less stable throughout the day.

Figure 2. Diurnal courses of radiation components at GS00 monitoring site.

Water resources of arid areas

372

5.2 Relative humidity and temperature Relative humidity and air temperature are measured in all the ADAS towers in the study area (Fig. 1). In order to demonstrate typical diurnal courses of air temperature and relative humidity during the end of summer and wintertime two daily records of 01/02/02 and 19/06/02 at 2m height at GS00 site were selected and presented respectively as Figures 3 and 4. In both daily records the relative humidity has a parabolic shape characterized by large values in the nights and decline starting ≈07:00 and a minimum at ≈15:00 and rises to a maximum at 24:00, whilst the air temperature depicts an opposite trend, being also characterized by rise (at ≈07:00) to a maximum (also at ≈15:00), followed by a decrease again to a minimum at 2400 hours. The main differences between the two days refer to longer time with the low relative humidity in the day and lower temperatures in June than in April. 5.3 Wind speed The wind speed monitoring is available in the study area only in GS00 at 2, 13 and 18m height and in GS08 at 2m height. In order to provide wind speed characteristics in the other monitoring sites such as GS01–07, not equipped with wind speed meters, at each of this site periodic measurements with the mobile 10m tower (originally located at GS09) equipped with 2 wind speed meters, one at 2m and the second at 10m height were made. These measurements were carried out between

Figure 3. Diurnal course of the relative humidity and air temperature on the

Monitoring of evapotranspiration on Kalahari, Serowe case study, Botswana

01/04/02 at a height of 2m at GS00 site.

Figure 4. Diurnal course of relative humidity and air temperature on the 19/06/02 at a height of 2m at GS00 site.

Figure 5. A bar diagram depicting correlation coefficients between permanent measurements at GS00 and

373

Water resources of arid areas

374

GS08 sites and mobile mast measurements at GS01–GS07 locations. September 2002 and September 2003 in four series with 10 days intervals, so each location was assessed 4 times, every time in different hydrological conditions. The main purpose of that experiment was to correlate the wind speed at the sites permanently monitored with wind speed recorded at the mobile tower moved between the locations GS01–07 for PET, rs and ra assessment. Figure 5 depicts the variation of correlation of wind speed at monitoring sites that are not permanently equipped with anemometers (sites GS01–07) with those, which are continuously logging wind speed (GS00 and GS08). The following observations are summarized from Figure 5. Comparatively better correlations were obtained for wind speed measurements at one specific site for 2m and 10m sensor heights than between different locations. In this regard, GS02 site has the highest correlation coefficient (0.97) and GS01 lowest correlation coefficient of 0.86. The correlations of the wind speed at the mobile tower locations (GS01–07) with wind speed at GS08 were substantially better than with wind speed at GS00. This perhaps was a result of the shielding effect of the adjacency of GS00 to the escarpment, which did not influence GS08, which is located like other mobile tower locations uphill of the escarpment on the sandveld plateau. The half-hourly wind speed regression models presented in Table 1, were established between the permanent record at GS08 (2m height) and 2m height wind speed measurements at the GS01–07mobile tower locations. These models were developed to create the missing wind speed records at those monitoring sites not equipped with the wind speed monitoring devices. The half-hourly estimates were finally averaged to daily values for the use in PET calculation according to equation 11. 6 RESULTS AND DISCUSSION The temporal variability of actual evapotranspiration at GS00 and GS09 is shown in Figure 6. Apart from a few discontinuities occurring in Figure 6 because of data loss, the majority of discontinuities in the time series analysis of actual evapotranspiration are where the temperature difference between two measurement heights was zero. In such situation the T-profile method runs into problem, which is a more prevalent case in dry periods. The actual evapotranspiration rates found by solving the energy balance equation (using sensible heat flux density derived by T-profile method) for Kalahari sandveld area represented by GS09 range from 0.01–0.63mm/day in the dry season and from 0.01– 2.09mm/day in the wet

Monitoring of evapotranspiration on Kalahari, Serowe case study, Botswana

375

Table 1. Results of regression between permanent wind speed measurement at the height of 2m at GS08 site and mobile wind speed measurements at 2m height at GS01–07 sites (on Kalahari). Monitoring site n GS01 GS02 GS03 GS04 GS05 GS06 GS07

R2 Regression model

1222 0.72 Y=0.89x+0.39 865 0.82 Y=1.19x+0.17 736 0.76 Y=1.11x+0.46 1267 0.73 Y=0.93x+ 0.26 1258 0.60 Y=0.91x+0.68 843 0.76 Y=1.08x 0.45 847 0.79 Y=0.69x+1.30

Figure 6. Comparison of actual evapotranspiration determined from the energy balance equation (using sensible heat flux density derived from T-profile method as input) at GS00 and GS09 sites. season. The actual evapotranspiration rates for hardveld area represented by GS00 range from 0.01–2.46mm/day in the dry season and 0.14–3.74mm/day in the wet season. The temporal variations in daily actual evapotranspiration are evident in Figure 6. The seasonal trends in actual evapotranspiration of both sites are characterized by higher actual evapotranspiration rates in summer and lower winter periods. The higher evapotranspiration rates in summers are mainly related to increased availability of water for evapotranspiration, higher ambient temperatures and higher solar radiation. The

Water resources of arid areas

376

comparative analysis of evapotranspiration records in GS00 and GS09 shows also that in most cases (with a few exceptions) the daily actual evapotranspiration rates were higher at GS00 (hardveld area) than at GS09 (Kalahari sandveld area). This most likely must have resulted from the larger groundwater evapotranspiration i.e. groundwater root extraction and upward convection-diffusion of groundwater (lubczynski, 2000) at the hardveld area where groundwater table was much shallower (often <10m) than in the Kalahari sandveld area where groundwater was generally deep in order of 70m b. g. s. Under thick sandveld unsaturated zone, covered by extensive savanna vegetation, the chances of groundwater evapotranspiration, if present are lower and if so arise solely from deep tree root extraction such as e.g. of Boscia albitrunca.

Table 2. Minimum and maximum potential evapotranspiration rates for GS01-GS08 situated on Kalahari. 2002–03 dry season 2001–03 wet season daily potential daily potential evapotranspiration evapotranspiration (mm/day) (mm/day) Monitoring Minimum Maximum Minimum Maximum tower rate rate rate rate GS01 GS02 GS03 GS04 GS05 GS06 GS07 GS08

0.08 0.01 0.35 0.24 0.55 0.32 0.05 0.03

6.20 3.85 6.38 6.92 6.19 6.16 5.33 4.62

0.76 0.38 1.04 0.90 1.21 1.14 0.78 0.70

7.43 7.59 7.90 7.72 8.11 8.01 7.40 6.85

Monitoring of evapotranspiration on Kalahari, Serowe case study, Botswana

377

Figure 7. Temporal variability of aerodynamic resistance at GS00 and GS09 sites. The potential evapotranspiration calculated with FAO Penman-Monteith model was estimated for GS01–08 ADAS locations (Fig. 1) in the study area using a combination of measured and regressed data input. The assessment indicated that PET is largely variable in the study area. Table 2 illustrates the spatial variability of PET. This variability is largely due to the substantial variability in the input parameters used for PET calculation. As mentioned, in the study area the relative humidity and air temperature are monitored in all ADAS locations. The short-wave incoming radiation PET input is available only in GS00 and GS08 location. Because it is spatially invariable and therefore does not seem to require more data coverage. The most critical wind speed input (available only at GS00 and GS08) for GS01–07 was obtained through the regression analysis using wind speed data from ‘mobile tower’ campaign on Kalahar sandveld. The correlation coefficients of the regression models turned out to be surprisingly high as for the usually weakly correlated wind speed measurement. This was likely due to the homogeneous wind characteristics on Kalahari sandveld resulting from short and quite sparse vegetation having aerodynamically uniform wind characteristics. The use of Penman-Monteith formula largely depends on surface resistance and it depends on the aerodynamic resistance. The aerodynamic resistance (ra) have been derived directly in this study according to Equation 12 and presented for GS00 and GS09 locations in Figure 7. Surface

Water resources of arid areas

378

Figure 8. Temporal variability of surface resistance on the Kalahari and hardveld situated monitoring sites of GS09 and GS00 site respectively. resistance (rs) depends on a number of factors such as sunlight, leaf water potential, vapor pressure deficit and soil water content. Forward use of Penman-Monteith equation for calculation of actual evapotranspiration rates is limited by deficiency in knowledge of rs of natural vegetation. In order to investigate the nature of that parameter, the available actual evapotranspiration obtained by solving the energy balance equation (with temperature profile derived sensible heat flux density) was used as a missing unknown in the inverted Penman-Monteith formula. The temporal variability of that parameter (Fig. 8—note the log scale) is substantially larger than of ra. This variability put in question the forward applicability of the Penman-Monteith formula for actual evapotranspiration calculations applying common rs simplifications by ‘lump, most likely’ estimates of this value. The larger variability of rs and ra in GS00 than in GS09 is attributed to larger diversity of vegetation, particularly with respect to tree size and shape and to larger variability of surface relief in GS00 than in GS09, in GS00 affected e.g. by the escarpment. For the same reason also ra in GS00 is generally higher than in GS09. With regard to rs such tendency is opposite due to an unlimited supply of moisture from the groundwater table, as a consequence of the shallow unsaturated zone at GS00 site, which facilitates direct root extraction and an upward convection-diffusion of groundwater.

Monitoring of evapotranspiration on Kalahari, Serowe case study, Botswana

379

7 CONCLUSIONS The actual evapotranspiration rates obtained by solving the energy balance equation for GS09 site situated on Kalahari range from 0.01–2.09mm/day, whilst actual evapotranspiration rates for the hardveld area (GS00) range from 0.01–3.74mm/day. Actual evapotranspiration rates were generally high at GS00 than on the Kalahari situated GS09 site. PET is largely spatially variable in the study area; it ranges from 0.01– 8.11mm/day. This variability depends largely on the widely unavailable wind speed characteristics. However, the correlations between half-hourly wind speed measurements on Kalahari sandveld showed satisfactory correlation range of 0.63–0.87. This allowed for reasonably accurate extrapolation for calculation of Penman-Monteith PET in seven (GS01–07) ADAS locations. The applicability of such extrapolation with regard to actual evapotranspiration is still being tested mainly because of the difficulties in estimates of rs. The inversion of Penman-Monteith formula with actual evapotranspiration considered as dependent variable allowed the calculation of rs for GS00 and GS09 locations. The comparison of rs and ra derived directly from Equation 12 for the Kalahari and hardveld areas indicate that aerodynamic resistances were generally higher on the hardveld than on the sandveld site whereas the surface resistances were indicating an opposite trend. ACKNOWLEDGEMENTS This study was conducted as part of a project entitled Kalahari Research Programme that was cooperatively funded by Botswana Geological Survey and ITC, in The Netherlands. We appreciate the field assistance of Mr. Ramotsoko and his technicians in collecting data for this study. REFERENCES Allen, R.G., Pereira, L.S., Raes, D. & Smith, M. 1998. Crop evapotranspiration, guidelines for computing crop water requirements, Food and Agriculture Organization of the United Nations. FAO Irrigation and Drainage paper 56: Italy. Bastiaanssen, W.G.M. 1995. Regionalisation of surface flux densities and moisture indicators in composite terrain, a remote sensing approach under clear skies conditions in Mediterranean climates. PhD thesis, Wageningen Agricultural University: The Netherlands. Bhalotra, Y.P.R. 1987. Climate of Botswana, Part 2: Elements of Climate. Department of Meteorological Services, Ministry of Works, Transport and Communications: Botswana. Choudhury, B.J. 1997. Global pattern of potential evapotranspiration calculated from the PenmanMonteith equation, using satellite and assimilated data. Remote sensing of the Environment 61:64–81. De Rooy, W.C. & Holtslag, A.A.M. 1999. Estimation of surface radiation and energy flux densities from single level weather data. Journal of Applied Meteorology 38:526–540. Ecosurv Botswana 1998. Vegetation mapping and ground truthing for radar imagery. Ecosurv, 30 September 1998, Gaborone: Botswana.

Water resources of arid areas

380

Gash, J.H.C. & Stewart, J.B. 1975. The average surface resistance of a pine forest derived from Bowen ratio measurements. Boundary-Layer Meteorology 8:453–464. Hargreaves, G.H. & Samani, Z.A. 1985. Reference crop evapotranspiration from temperature. Applied Engineering in Agriculture 1(2):96–99. Hernandez, A.R. 2002. Mapping of woody vegetation in arid zones-a multi-sensor analysis, a case study in the Serowe area, Botswana. MSc. Thesis. ITC—International Institute for Geoinformation Science and Earth Observation: The Netherlands. Holtslag, A.A.M. & Van Ulden, A.P. 1983. A simple scheme for daytime estimates of the surface fluxes from routine weather data. Journal of Climate and Applied meteorology 22(4):517–529. Jarvis, P.G. 1976. The interpretation of the variations in leaf water potential and stomatal conductance found in canopies in the field. Philosophical Transactions of the Royal Society of London. Ser. B 273:593–610. Lubczynski, M.W. 2000. Groundwater evapotranspiration-Underestimated component of the groundwater balance in a semi arid environment-Serowe case, Botswana. In Sililo et al. (eds), Groundwater, past and achievements and future challenges: 959–963. Rotterdam: Balkema. Maidment, D.R.1993. Handbook of Hydrology, McGraw-Hill: Texas. Monteith, J.L. 1995. Accommodation between transpiring vegetation and the convective boundary layer. Journal of Hydrology 166:251–263. Nash, D. 1992. The development and environmental significance of the dry valley systems in the Kalahari. central Southern Africa. PhD thesis. University of Sheffield: United Kingdom. O’Toole, J.C. & Real, J.G. 1986. Estimation of aerodynamic and crop resistances from canopy temperature. Agronomy Journal 78:305–310. Raupach, M.R. 1994. Simplified expressions for vegetation roughness length and zero plane displacement as functions of canopy height and area index. Boundary-Layer Meteorology 71:211–216. Smith, R.A. 1984. The Lithostratigraphy of the Karoo Supergroup in Botswana. Bulletin 26, Department of Geological Survey: Botswana. Stewart, J.B. 1988. Modelling surface conductance of pine forest. Agricultural and Forest Meteorology 43:19–35. Swedish Geological Survey (SGS) 1988. Serowe Groundwater Resources Evaluation Project, Final report, Ministry of Mineral Resources and Water Affairs. Department of Geological Survey: Botswana. Vossen, P. 1989. An agrometeorological contribution to quantitative and qualitative rainy season quality monitoring in Botswana. PhD thesis. State University of Ghent: Belgium.

Electro-seismic survey system S.R.Dennis, M.du Preez & G.J.van Tonder Institute for Groundwater Studies, University of the Free State, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The electro-seismic survey can be used with great success in geophysical surveys where the aim is to site groundwater. The presence of water in the porous rock media is directly responsible for the generation of electro-seismic signals. An analysis technique developed through a quantitative approach produces satisfactory results from survey data sets. These include possible water strike positions as well as layer interface information. Two case studies are presented in this paper to illustrate the capabilities of the current system used for electro-seismic surveys.

1 INTRODUCTION Geophysical techniques have been employed for many years to locate groundwater in South Africa. Magnetic airborne surveys are very useful in the structural mapping of an area, particularly to determine the dominant direction of tectonic movement, and the location of features such as faults, dykes and fracture zones. The majority of geophysical surveys (magnetic, electromagnetic, electric and gravitational) can yield valuable information on the global geometry of an aquifer. The results are also often ambiguous especially in Karoo aquifers with its numerous layers of mudstone and siltstone. In many instances it is necessary to use a combination of these techniques to overcome these ambiguities. High resolution radio and seismic tomography are two methods that have shown some promise for groundwater investigations in the Karoo formations. The majority of the geophysical methods focus on obtaining information concerning the rock matrix and subsurface structures. The electro-seismic effect on the other hand is a direct result of relative movement of an electrolytic fluid with respect to the rock matrix. Thus the presence of water in the porous rock media is directly responsible for the generation of electro-seismic signals. This has huge cost saving implications due to the fact that the presence of water could be determined before any drilling has taken place. This paper discusses the theory surrounding the electro-seismic effect and finally two case studies are presented to illustrate the interpretation of the electro-seismic survey results.

Water resources of arid areas

382

2 ELECTRO-SEISMIC EFFECT 2.1 Background The electro-seismic effect describes the conversion from seismic to electromagnetic (EM) energy. Several mechanisms are likely to generate couplings between seismic and EM energy in the subsurface (Garambois & Dietrich, 2002). The main effects of interest to geophysicists are electrokinetic and piezoelectric phenomena and variations in electrical resistivity. The macroscopic governing equations were derived from first principals by Pride (1994) which coupled Biot’s theory and Maxwell equations via flux/force transport equations. In this theory the coupling mechanism is explained by electrokinetic effects taking place at pore level. 2.2 Wave behaviour A seismic wave propagating in a medium can induce an electrical field or cause radiation of an electromagnetic wave. There are two electro-seismic effects that are considered in this paper (Oleg et al., 1997). The first effect is caused when a seismic wave crosses an interface between two media. When the spherical P-wave crosses the interface, it creates a dipole charge separation due to the imbalance of the streaming currents induced by the seismic wave on opposite sides of the interface. The electrical dipole radiates an EM wave which can be detected by remote antennas as shown in Figure 1. The second effect is caused when a seismic head wave travels along an interface between two media. It creates a charge separation across the interface, which induces an electrical field. This electric field moves along the interface with the head wave and can be detected by antennas when the head wave passes underneath as shown in Figure 2.

Figure 1. Seismic wave crossing an interface generating an electromagnetic wave.

Electro-seismic survey system

383

Figure 2. Head wave traveling along an interface generating an electric field. 3 SYSTEM DESCRIPTION The Electro-Seismic Survey (ESS) system consists of a probe, a trigger and base units. The probe unit is a low noise amplifier connected to a 16-bit A/D converter. A horizontal dipole antenna is connected to the input of the amplifier. The seismic source used is a hammer and plate. The hammer is connected to the trigger unit. An inertia switch is fitted to the hammer and acts as the physical trigger. The base unit is connected to a laptop via the serial port. Custom software has been developed for data acquisition and processing. The probe and trigger units interface to the base unit via a wireless link to reduce external noise coupling into the system. The recorded data set represents 200ms which translates roughly to 200m at the slowest seismic velocity for subsurface media. Practical results indicates an average maximum penetration depth of 150m when using the hammer and plate seismic source. 4 METHODOLOGY Two methods are used to analyse recorded data. These methods are discussed in the sections below. 4.1 Numerical model The first approach in analysing the recorded signals was to setup a numerical model that simulated the electro-seismic effect. A ray path model, using the generalized matrix method to solve the governing equations, was setup for this purpose.

Water resources of arid areas

384

Problems experienced were the instability of the model for certain inputs and the fact that there was little correlation between the simulated and recorded data. 4.2 Quantitative approach The second approach taken was a quantitative one where surveys were done before drilling took place without making any recommendations with respect to the ESS survey results. After drilling all borehole logs and water strikes were recorded. By doing this the authors could gather enough data to develop an analysis technique that provides accurate results. 5 CASE STUDIES In this section two case studies are presented. The first is to illustrate that the principle of electro-seismic surveys does work and the second case study illustrate the current capability of the system used. 5.1 Shallow water pipe A shallow concrete water pipe was chosen to illustrate the electro-seismic effect. The water pipe is 1m below the surface with a diameter of 20cm. A profile was done across the pipe in a direction perpendicular to the pipe with a station spacing of 0.5m. The result of the survey is shown in Figure 3. From the results it is clear that the position of the pipe is correctly determined when the velocity of the seismic wave is estimated as 5000m/s. Typical seismic velocities lie in a range of 1500 to 6800m/s depending on the medium. Contours indicate activity beneath the pipe, but this could be a result of reflections because of the close proximity of the surface or due to the free movement of water in the pipe as opposed to water moving in a rock matrix.

Electro-seismic survey system

385

Figure 3. Results of pipe survey. 5.2 Borehole in layered media For this case study a profile was done across the area of interest and the results of the possible water strike positions together with the layer information are shown in Figure 4. Firstly, consider the graph on the left hand side. After the survey was completed this contour plot is produced to determine possible water strike positions. The depth of the water strikes is dependant on the seismic velocity used to do the depth scaling. For this particular survey an average seismic velocity was used that was calculated from a borehole about 200m away where the borehole log was available. The difference between the predicted and actual water strike was 3.5m. If no information is available prior to drilling a guesstimate is made regarding the average seismic velocity. The contour plot will still indicate the line with the best response although the estimated water strike could be inaccurate. At a first glance it would seem that the main water strike is at 20m below the surface but this is not necessarily the correct assumption for the following reasons: ● The seismic wave attenuates with depth as it crosses interfaces. ● The type of geological interface has an influence on the response generated.

Water resources of arid areas

386

Currently the analysis technique used is not capable of indicating the main water strike as related to possible yield estimation. Now consider the layer plot on the right hand side of Figure 4. Before drilling starts the layer interfaces can be visualized from the data set obtained but the layer types can not be identified until after drilling has taken place and a geological log is available. The main water strike was found at roughly 47m below surface on a doleritesandstone interface and a blow yield of 4L/s was measured. Very good correlation exists between the borehole log and the layer plot. Only a few layer types are shown on the layer plot for illustration purposes.

Figure 4. Possible water strike positions and layer information of the layered media survey. 6 SYSTEM LIMITATIONS The system is very susceptible to power line noise which makes the data analysis difficult due to the fact that the 50 to 60Hz noise and the associated harmonics fall within the bandwidth of the electro-seismic response. Various filtering schemes (notch filters and sinusoidal subtraction) have been tested with limited success because valuable data gets lost during the filtering process. Data stacking has provided the best results to date of all schemes tested. For very noisy data 50 point stacking gives good resolution otherwise 10 point stacking is used in the surveys.

Electro-seismic survey system

387

The analysis technique employed does not account for the attenuation of the seismic signal with depth and this could lead to the misinterpretation of the data because the strongest signal response does not necessarily indicate the main water strike. The current analysis technique does not lend itself to yield estimation and further research on this topic needs to be done to determine if it is feasible or not. Accurate depth of features is dependant on the seismic velocity used to do depth scaling. In most instances this information is not available before hand although good results have been obtained in areas where the average seismic velocity has been determined from existing borehole logs. 7 CONCLUSIONS From the results it is clear that the electro-seismic survey does provide accurate results if the correct seismic velocity is used for depth scaling. The electro-seismic survey provides valuable information regarding possible water strike positions and physical layering of the media but no successful yield estimation could be done to date. Results indicate a localized response from the electro-seismic effect which is in contradiction to articles written regarding this subject. Further research is needed to fully understand the electroseismic effect and the application thereof. REFERENCES Garambois, S. & Dietrich, M. 2002. Full waveform numerical simulations of seismoelectromagnetic wave conversions in fluid-saturated stratified porous media. Journal of Geophysical Research, 107(B7). Oleg, V., Mikhailov, O.V., Haartsen, M.W. & Toksoz, M.N. (1997). Electroseismic investigation of the shallow subsurface: Field measurements and numerical modeling. Geophysics, 62(1):97– 105. Pride, S.R. 1994. Governing equations for the coupled electromagnetics and acoustics of porous media. Phys. Rev. B, 50(15):678–696.

Borehole site investigations in volcanic rocks of Lolmolok area, Samburu district, Kenya J.K.Mulwa University of Nairobi, Department of Geology, Nairobi, Kenya Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: A systematic approach has been applied in the selection of suitable sites for borehole drilling in a quest to provide adequate water supply to a rural pastoral community in Lolmolok area. The study area lies in Samburu district in Kenya and is bound by latitudes 0°56′21″N and 0°57′58″N and longitudes 36°34′42″E and 36°36′35″E. The geology of this area is comprised of tertiary volcanics. Basalts, which have weathered into residual black cotton soil, are underlain by phonolitic lavas and tuffs. The systematic approach for the exploration of groundwater was followed to enable selection of an optimum drill site(s) within a quadrant with three-kilometer radius identified by the pastoral community. The approach consisted of the following multi-steps: (i) Hydrogeological reconnaissance of the whole area, mapping different groundwater potential areas on the basis of aerial photo interpretation; (ii) Geophysical field surveys involving very low frequency electromagnetic (VLF-EM) and Vertical Electrical Sounding (VES). (iii) Processing and interpretation of the data acquired in the field, which led to selection of suitable drill sites, indication of potential yield and depth of aquifers. This paper describes the success of combined geophysical survey techniques in siting boreholes whose yield ranges between 5m3/hr and 10m3/hr.

1 INTRODUCTION 1.1 Background information Lolmolok area is situated on Loroki plateau in Samburu district (Figure 1). In this area there existed a borehole (C-2847) drilled in 1958. It was reported to have an approximate yield of 2.1m3/hr and a total depth of 180m, with an aquifer depth of 110m bgl. This borehole provided a strategic water source for the pastoral community living in the area

Borehole site investigations in volcanic rocks of Lolmolok area, Samburu district, Kenya

389

especially during times of drought. The borehole broke down forcing the pastoralists to move with their livestock to Lonkewan area approximately 10km south during the drought periods where an operational borehole exists. Due to the influx and concentration of large numbers of livestock in Lonkewan area, land degradation and pasture depletion was inevitable. The soils in Lonkewan area were exposed to erosion by wind and surface run-off. In 1998, drought preparedness intervention recovery programme (D.P.I.R.P), a non-governmental organization in charge of poverty alleviation for the Samburu pastoral community, commissioned a contractor to rehabilitate the borehole in Lolmolok area.

Figure 1. Location and geology of Loroki plateau.

Water resources of arid areas

390

This was not successful since metallic objects in the borehole rendered it impossible to rehabilitate. A replacement borehole was drilled about 1.2km north of the old one down to a depth of 140m. However, no water was struck up to this depth. 1.2 Scope of the project In January 1999, D.P.I.R.P sought the services of a competent groundwater consultant to carry out borehole site investigations in Lolmolok area of Samburu district. The principal objectives of the investigations were as follows: ● To locate three potential borehole sites which could produce a yield of at least 5m3/hr each within a quadrant with three kilometer radius selected by the community; ● To locate the most suitable site for drilling a borehole within the area set aside by the community; ● To assess the general prospects for groundwater development in the southern part of Lolmolok situated on the Loroki plateau. This paper describes the hydrogeological assessment in Lolmolok area. The area is comprised of a flat to gently undulating volcanic sheet approximately 5.5km wide of the Loroki plateau. The western edge of the plateau drops steeply forming one of the scarps of the Sukuta valley, whereas the eastern edge is punctuated by deeply incised valleys in which shallow groundwater is abstracted during dry seasons. In these valleys, groundwater discharges from shallow phonolitic “bedrock” after the rains. The surface of the plateau is marked by sporadic hummocky phonolite outcrops, seperated by boulder-strewn plains. The clay soils are derived from the weathered basalts. Shallow pans are associated with poorly drained black cotton soil which cracks when dry and swells when wet. It is of medium to high fertility. In well drained areas, brown to reddish brown soil is found. On hill slopes and along streams, shallow loamy and gravelly soil is found. Most of the Loroki plateau is covered by Savannah grassland, while the seasonal stream valleys support bushland and low trees. 2 GEOLOGY The investigated area is covered by black cotton soil developed from the weathering of underlying basalt and/or phonolites. The basalt in the area is underlain by Losiolo and Rumuruti phonolitic groups, Figure 1 (Hackman 1988). Five flows of the Losiolo phonolite sequence are recognized on the Loroki plateau. The phonolites are typically black in colour and fine grained. On weathering, the phonolites attain a purplish grey colour. However, the uppermost flow is tough, brownish grey and fine grained. Parts of the flow show alignment of alkali-feldspar phenocrysts and occasionally biotite flakes. It is pitted where small nepheline phenocrysts have weathered out. Clastic bases and scoriaceous tops to the flow may be up to 5m thick (Hackman 1988). The individual thicknesses of these flows have not been established. However, on the eastern shoulder of the rift valley, the thickness is estimated to be 600m. Metamorphic basement system rocks are likely to be encountered beyond this depth.

Borehole site investigations in volcanic rocks of Lolmolok area, Samburu district, Kenya

391

The structures within the area deduced from aerial photo interpretation are faults and fractures which occur both in the basement system rocks and in the volcanics. The general trend of the major lineaments is roughly north-south. These lineaments are easily discerned on both aerial photographs and Landsat images at higher altitudes, but are less discernible in the lower regions where they are obscured by topsoils developed from the weathering of basalts and phonolites. From the aerial photography interpretation, it can be concluded that the river channels are structurally controlled as is manifested by the sudden angular changes in the channel courses (Figure 2). 3 GROUNDWATER POTENTIAL AND OCCURRENCE The investigated area is marked by a medium potential for groundwater abstraction since it is exclusively volcanic, with a rainfall of about 600mm per annum, which is higher when compared to that of adjacent areas. This supports the vegetation on the Loroki plateau. The plateau is marked by three rainy seasons which comprise the monsoon controlled long rains (March–May) and the short rains (October–November). In addition, humid air streams with their origin in the west cause a third rainy period in July–August, known as the continental rains. The continental rains are of great importance as they provide an extended growing period for crops and pasture development (Flury 1987). The most distinct dry periods occur from December to February and during the month of September. Assuming a conservative effective rainfall of 1% and a catchment area of 1480km2 over the Loroki plateau, the available recharge is estimated to be 8.9×106m3/year (Water

Water resources of arid areas

392

Figure 2. Surface drainage system in Loroki plateau. resources assessment project (WRAP) 1991). Assuming a safe yield of 20% of the available recharge, the total abstractable groundwater amounts to 1.8×106m3/year. Considering an average borehole yield of 4m3/hr at 12 hours per day pumping regime, each borehole will produce about 18,000m3/year. Consequently, recharge of 1% of

Borehole site investigations in volcanic rocks of Lolmolok area, Samburu district, Kenya

393

rainfall over the Loroki plateau can theoretically sustain approximately 100 boreholes (WRAP 1991). The Loroki plateau is drained by two main intermittent streams; Amaya in the further west of the plateau and Enkare Narok (Figure 2) in the east. A network of seasonal streams (laggas) collect surface water on the western part of the plateau and drain into Amaya river. The laggas on the eastern part of the plateau drain into Enkare Narok which in turn drains into Ewaso Nyiro river. The main NNE-SSW watershed along the central part of the Loroki plateau divides streams flowing westwards into the Rift valley from the major eastward flowing river systems. Ewaso Nyiro, Ewaso Narok and Ol Keju Losera are the largest perennial rivers on the eastern side of the water divide (Figure 2). Groundwater within the area is located within old land surfaces between lava flows, fractures, fault zones and contact zones of various rocks. The fresh and massive phonolites do not contain significant amounts of water. To the east of the plateau, lavas wedge out against the basement rocks whereas the rift valley shoulder forms the western boundary. Since the geology, physiography and drainage patterns are uniform over a large area, a regional aquifer system is expected. There are extensive aquifers of old land surfaces which are interconnected by fractures and faults. Although the boreholes drilled in the area are few and their distribution insufficient to ascertain the regionality of the aquifer system, evidence from similar areas in Laikipia and Meru shows the existence of a regional aquifer system in the volcanic sheets (Ground water survey 1998, unpublished). 3.1 Old land surfaces (OLS) Aquifers located in weathered horizons are only a few metres thick. Hence, significant number of boreholes may be necessary to obtain a reasonable yield. The top regolith developed within the fine dark lavas is generally too clayey to support significant volumes of water for abstraction. Different sequences of lava flow underlying the Lolmolok area indicates the presence of a number of old land surfaces. To the north and northeast of the investigated area, shallow groundwater occurs within the upper lava layers at general depths ranging between 25m and 55m below ground level (Table 1). The best prospects for groundwater occur where local recharge is supplemented with water entering through an alluvial system, or in places where a direct connection with the regional phonolite aquifer exists.

Table 1. Aquifer characteristics in Loroki plateau. Borehole Year Elevation Total WSL SWL Yield number drilled (m msl)* depth (m (m (m3/hr) (m bgl)** bgl)*** bgl) C-444

1946

2130

C-479 C-1505 C-2434 C-2847

1946 1951 1955 1955

2130 2130 1981 2045

122 41, 91, 122 90.2 55, 69 105.4 25 183 169 180 110

28

0.4

11 22 146 71

0.4 0.7 2.5 2.0

Water resources of arid areas

C-2972

1959

1981

C-3609

1969

2134

394

259

45.7, 174 70 37.7, 41.6 123 38 100 26, 58

49

1.3

35.4

6.1

C-3833 1972 2134 12 C-7921 1989 2040 1.2 * meters above mean sea level. ** water struck level in meters below ground level. *** static water level in meters below ground level.

2.6 12

3.2 Faults and fracture zones Aquifers within the weathered layers and old land surfaces have low yields ranging between 1–2m3/hr. Higher yields of more than 5m3/hr are obtained from boreholes intersecting “open” faults and fracture zones. The water potential of structurally altered rocks is two fold: ● Along faulted and fractured rocks, weathering penetrates much deeper, thus creating sub-vertical zones filled with relatively coarse, weathered material. These zones generally have much higher transmissivity than their surroundings. ● Recharge occurs over large areas as faults may extend well beyond the surface catchment, thus intercepting adjacent aquifers or surface sources (Mulwa 2001). The fractures and faults in Lolmolok area may have been formed during cooling of lava and periods of regional faulting, respectively (Hackman 1988). Observations from shallow hand-dug wells shows that fractures form conduits for rainfall percolating down to the saturated zone. The most productive boreholes on Loroki plateau are those which intersect fault and fracture zones. Linear structures observed on aerial photographs are dominant on the western fringe of the plateau where it borders the Amaya embayment. These structures trend in northsouth and northeast-southwest directions and are associated with the extensive faulting along the main rift system. 4 SITE INVESTIGATION The investigated area is shown in Figure 3. The main objective of borehole site investigation is to determine the optimum drilling site and reasonable depth of drilling which should yield sufficient

Borehole site investigations in volcanic rocks of Lolmolok area, Samburu district, Kenya

395

Figure 3. Locations of VLF-EM and VES survey traverses in Lolmolok area. water of good quality. For the purpose of this project, a borehole yield of about 4m3/hr or greater was required. The investigations were carried out in order to identify appropriate sites on faulted and/or fractured and weathered zones in the volcanic rock formations. The groundwater exploration programme consisted of a systematic multi-step appraoch initially applied for borehole site investigations in Gachoka, Embu District in Kenya (Anyumba et al., 1993) as follows: ● A hydrogeological reconnaissance of the whole area, including mapping different groundwater potential areas on the basis of aerial photo interpretation. ● Detailed structural mapping based on aerial photo interpretation for selection of suitable lineaments. ● Geophysical field surveys involving very low frequency electromagnetic (VLF-EM) and vertical electrical sounding (VES) methods.

Water resources of arid areas

396

● Processing and interpretation of the data acquired in the field which led to selection of suitable drill sites, indication of potential yield and depth of aquifers and general prospects for groundwater development in Lolmolok area. The VLF-EM survey was executed in order to evaluate the pattern and form of the underlying geology. Qualitative evaluation of the data enabled the identification of structural anomalies (fault zones, fractures, lithological contacts) and thicknesses of the weathered zones. The identified structural anomalies were targeted for detailed investigations using the VES method. The VES method was used mainly to identify thickness of the weathered formation and indirectly identify potential water bearing zones (Beeson and Jones, 1988). VLF-EM profiling was carried out with an ABEM WADI VLF instrument while ABEM SAS 3000B terrameter was used for VES by application of Schlumberger array. A total of six VLF-EM profiles and nine vertical electrical soundings were carried out. Six VES measurements were executed on high conductivity anomalies identified from the VLF-EM profiles. Two soundings were executed next to the abadoned borehole (C-2847) and at the dry replacement borehole about 1.2km north of C-2847. The other sounding (VES 9) was carried out within the study area at a random location with no apparent conductivity anomalies or borehole for comparison purposes. Figure 3 shows the locations of VLF-EM and VES traverses in the study area. 5 RESULTS AND INTERPRETATION The geophysical site investigations in Lolmolok area have been backed by resistivity data, vertical electrical sounding curves and very low frequency electromagnetic profiles. The VES data has been processed using Schlumb software which is based on the inverse filter coefficients of Ghosh (1971). The results of the modeled VES curves are presented in Figures 4, 5, 6, 7, 8 and 9 below. The VLF-EM profiles carried out along the VES sites 3, 4, 5 and 6 are shown in Figures 10, 11 and 12. VLF-EM profiles along traverses 4, 5 and 6 did not indicate the presence of any subsurface conductive zones and are not presented in this paper. 6 RECOMMENDATIONS On the basis of the control sounding carried out at the abadoned borehole (C-2847), four vertical electrical sounding sites were recommended for drilling within the area initially set aside by the local community. These include VES sites 6, 5, 4 and 3 (Figure 3) in that order of priority. Groundwater is expected to occur within the weathered formations, faults and fracture zones at depths ranging between 30m and 170m below ground level. From the results and recommendations of this survey, two boreholes herein referred to as C-2000A and C-2000B were drilled at VES sites 6 and 4 respectively. Borehole C2000A was drilled

Borehole site investigations in volcanic rocks of Lolmolok area, Samburu district, Kenya

Figure 4. Hydrogeological interpretation of VES 3.

397

Water resources of arid areas

398

Figure 5. Hydrogeological interpretation of VES 4.

Borehole site investigations in volcanic rocks of Lolmolok area, Samburu district, Kenya

Figure 6. Hydrogeological interpretation of VES 5.

399

Water resources of arid areas

400

Figure 7. Hydrogeological interpretation of VES 6.

Borehole site investigations in volcanic rocks of Lolmolok area, Samburu district, Kenya

Figure 8. Hydrogeological interpretation of VES 7 next to borehole C-2847.

401

Water resources of arid areas

402

Figure 9. Hydrogeological interpretation of VES 9.

Borehole site investigations in volcanic rocks of Lolmolok area, Samburu district, Kenya

403

Figure 10. VLF-EM profile along traverse 1.

Figure 11. VLF-EM profile along traverse 2. up to a depth of 200m bgl. Water was struck at 100m bgl and rested at 80m bgl. This borehole was pump tested and yielded 10m3/hr. Borehole C-2000B was drilled up to a depth of 200m bgl. Water was struck at 105m bgl and rested at 90m bgl. The tested discharge yield was found to be 5.4m3/hr.

Water resources of arid areas

404

Figure 12. VLF-EM profile along traverse 3. 7 CONCLUSION A geological and hydrogeological assessment of Lolmolok area has been carried out using geophysical investigations. Aerial photo interpretation aided in the identification of lineaments and other lateral anomalies on the surface. Geophysical exploration techniques aided in the location of lateral variations and evaluation of thicknesses and resistivities of rock formations. The results of VES 9 show the effectiveness of combined geophysical survey techniques in borehole site investigations. In Loroki plateau, aquifers associated with phonolites have the potential of yielding up to 5m3/hr. Fractured and contact zone aquifers have the potential of yielding 10m3/hr. Recharge estimated to be 1% of rainfall over the Loroki plateau can sustain up to 100 boreholes (Q=4m3/hr) and the safe drilling depth is 200m below ground level. Water occurs at various depths between 30m and 170m below ground level. REFERENCES Anyumba, J., Van Dongen, P. & Nzomo, J. 1993. Borehole site investigations in fractured hard rock aquifers in Gachoka division, Embu district, Kenya. Proc. 5th conference on the Geology of Kenya, Geol. Soc. Kenya: 116–121. Beeson, S. & Jones, C.R.C. 1988. The combined EMT/VES geophysical method for siting boreholes. Groundwater, 26(1):54–63. Flury, M. 1987. Rainfed agriculture in the central division, Laikipia district, Kenya. Univ. of Berne, Switzerland.

Borehole site investigations in volcanic rocks of Lolmolok area, Samburu district, Kenya

405

Ghosh, D.P. 1971. Inverse filter coefficients for the computation of apparent resistivity standard curves for a horizontally stratified earth. Geophys. Prospect., 19:769–775. Hackman, B.D. 1988. Geology of the Baringo-Laikipia area. Ministry of Environment and Natural Resources, Mines and Geology dept., 97:1–25. Mulwa, J.K. 2001. Geological and structural set up of Kiserian-Matasia area and its influence on groundwater distribution and flow. M.Sc. Thesis, Univ. of Nairobi, Kenya. Water Resources Assessment Project (WRAP), (1991). Water resources assessment study in Samburu district; District water development study 1993–2013, part 1, water demand and water resources. Water Resources Assessment Division, Nairobi, Kenya; TNO-DGV Institute of Applied Geoscience, Delft, The Netherlands.

Groundwater evaluation in a complex hydrogeological environment—a GIS based approach B.Mudzingwa, J.L Farr, R.Gumiremhete & T.Kellner Wellfield Consulting Services, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Lack of predictable variation in hydrogeological characteristics within tectonically complex and multi-phase deformed preKalahari environments make identification of good potential aquifers elusive. Aquifers are discontinuous, water chemistry highly variable within short distances, hydraulic continuity is absent and hydraulic gradients are subdued making groundwater flushing impossible, leading to poor groundwater quality. Nonetheless pockets of fresh groundwater resources can be found. The complexity of the hydrogeological regime, compounded by data sparsity and clear quantitative understanding of aquifer parameters, precludes assembling a numerical groundwater model to quantitatively evaluate groundwater potential. Alternative methods to assess ‘groundwater potential’ of the area can however be employed using GIS based applications through combining various data sets and applying semi-quantitative spatial weighting factors. This approach was applied in the complex fractured Proterozoic environment of Werda-Sekoma-Mabutsane area in southern Botswana. With further evaluation, this approach may also be applied in other areas for groundwater exploration, evaluation, planning and management.

1 INTRODUCTION The project location is shown in Figure 1. The project’s aim was to establish groundwater potential of all aquifers in the study area and to address the key issue of providing adequate quantities of suitable quality water to support both livestock and the associated local communities. The area is underlain by sedimentary strata of the Karoo Supergroup to the north and Proterozoic metasediments (quartzites, shales) of the Transvaal and Waterberg Supergroups in the south and central parts of the area. Intrusive rocks of the Molopo Farms Complex (mafic and ultramafic rocks) underlie the southeastern portion of the area (Fig. 2). Other intrusives found within the project area are dolerites, granites and

Groundwater evaluation in a complex hydrogeological environment

407

mafic rocks of gabbroic composition. The Karoo consists of siltstones, carbonaceous mudstones, sandstones, coals and mudstones, with sandstone units of the Ecca Group constituting the main aquifer. The geological history of the project area is long and complex and no extensive areas of strata that are readily identifiable as ‘aquifers’ are apparent. In addition, all these lithological units are largely obscured by the ubiquitous unconsolidated Cretaceous to Recent Kalahari Group sediments varying in thickness from 0m to 250m. 2 EXPLORATION PHASE ACTIVITIES AND RESULTS Satellite imagery (Landsat ETM+) and aerial photographs were interpreted and then integrated with regional geophysical data to develop a tectonic model and locate potential borehole targets. Geophysical work involved interpretation of regional geophysical data sets (gravity, aeromagnetics and seismics) and subsequent detailed ground geophysical surveys over selected zones.

Water resources of arid areas

408

Figure 1. Project location map. Test pumping data revealed that hydraulic characteristics of the fractured quartzite were highly spatially variable with presence of barrier boundaries and closed aquifer conditions. From hydrochemistry data interpretation, the Molopo River to the south presents the only known Kalahari aquifer within the project area with potable water (TDS<1000mg/l) and indicates recently recharged water with an essentially Ca-Mg-HCO3 water type. Anywhere else away from the valley within the Kalahari Group, chemical composition of groundwater is dominated by Na and Cl. The presence of both Ca-Mg-Cl-HCO3 and Na-Cl/Na-Cl-SO4 water types within the Karoo tend to indicate lithological variation, different flow systems and structural compartmentalisation.

Groundwater evaluation in a complex hydrogeological environment

409

Within the Proterozoic quartzite aquifer, hydrochemistry and recharge mechanisms are largely controlled by the distribution and interconnectivity of the fracture systems, structures and Kalahari thickness. Water type is dominantly Na-Cl/Na-Cl-SO4 down the hydraulic gradient. Temporal variations in rainfall, recharge rates and groundwater abstraction regimes are likely to induce temporal changes in water type and quality as seen in Borehole BH 1102 in Figure 3. All water types have been identified in the project area. Group I—Ca-Mg-HCO3, Group II—Ca-Mg-Na-HCO3-Cl, Group III—Na-Cl-HCO3, Group IV—Na-Cl and Group V—Na-Cl-SO4/ Na-SO4-Cl Group I, II and III are common in areas with thin Kalahari cover, Group II and III are regarded as a mixture of Groups I and IV. Group V water type is dominant in Karoo to the north and within the Kalahari occurring along the Molopo River. Ground water level and rainfall monitoring data was not sufficient to accurately determine recharge rates as the monitoring period was limited (1 year). Both newly acquired and historical isotope data from boreholes sampled within the project area were used for isotope analysis for 18O, 13C, 14C, 2H (deuterium), 3H (tritium) and N. The relationship between 18O and deuterium in the study area follows the pattern that has been seen in other parts of Botswana. The two isotopes fall along an evaporation line with slope 5.3

Water resources of arid areas

410

Figure 2. Geology map of the project area.

Groundwater evaluation in a complex hydrogeological environment

411

Figure 3. Hydrochemistry time series data from BH 1102.

Figure 4. 18Ovs. 2H plot for Werda rainfall and groundwater. (Figure 4). The Werda slope indicates fractionation and enrichment of 18O isotopes compared to the GMWL suggesting evaporation prior to recharge or predominance of recharge from light rainfall. Tritium values range from 0 to 1.6 TU and results generally

Water resources of arid areas

412

show pre-1950 recharged water. 14C ranges from a young 101pmc to values of 2 and 4pmc, the latter certainly representing fossil water. Where both 14C and tritium have been analysed, the high values for both isotopes are thought to represent recharge. The Molopo valley exhibit different isotopic signatures from other samples within the project suggesting different mechanisms of recharge. Mean Residence Time (MRT) were semi-quantitatively estimated depending on a combination of 14C, 13C and tritium contents based on the following criteria (Talma & Tredoux, 2003): MRT<200 years: 14C>70 pmc and MRT between 1000 and 15000 years: 14C tritium >1 TU. between 20 and 50pmc and tritium <1 TU, MRT<1000 years: 14C between 50 Fossil water>15000 years: 14C<5pmc. and 70 pmc and tritium <1 TU,

Quantitative recharge estimates employed hydrogeological, hydrochemical and isotope data while qualitatively a GIS based approach was used. Stable isotopes of 18O and 2H gave an average value of 8mm/year employing the deuterium shift method while application of the Chloride Mass Balance Method was not possible. Groundwater and rainfall monitoring data was too short to give conclusive results on recharge. A GIS based approach involved integrating different attributes that influence or are directly related to groundwater recharge such as soil characteristics, water quality, geology, structure (lineament density), vegetation type and density. Results of the GIS based approach showed strong correlation when compared to spatial distribution of 14C pmc values. 3 CONCLUSIONS DRAWN FROM THE EXPLORATION PHASE Conclusions relevant to this paper have been highlighted below; ● The area has a long and complex tectonic history with many periods of brittle deformation leading to development of at least five main fracture directions. ● Groundwater occurrence and movement in the quartzite aquifer is controlled by fissures, fractures, and faults induced by tectonic disturbance and possibly enhanced by weathering. ● Fracture connectivity has significant influence on piezometry, aquifer hydraulic properties, hydrochemical evolution, groundwater quality distribution and recharge. ● Kalahari cover thickness has a significant influence on groundwater recharge and quality. ● Localised flow systems with varying water quality over short distances and low recharge rates are prevalent due to variably connected fracture systems. ● The complex nature of the hydrogeological regime makes quantification of groundwater resources difficult.

Groundwater evaluation in a complex hydrogeological environment

413

4 GIS APPLICATIONS 4.1 General overview A GIS model is simply ‘a process of combining a set of input maps with a function to produce an output map.’ (Bonham-Carter, 1994) The purpose of GIS applications was to combine spatial data from diverse sources in order to describe and analyse interactions and to come up with maps that show spatial variables that satisfy certain criteria. In this exercise the index overlay method was chosen for the GIS application, which can be expressed as; (1)

where is the weighted score of an area (polygon or pixel), Sij is the score of the j-th class of the i-th map and Wi is the weight of the i-th input map. 4.2 Possible future groundwater management tools The complexity of the hydrogeological regime compounded by a paucity of data and clear quantitative understanding of regional aquifer parameters precluded the assembly of a numerical groundwater model on which quantitative evaluation of groundwater potential can be based. It was thus proposed that the ‘groundwater potential’ of the area be assessed in a qualitative manner by combining various data sets in a GIS environment and applying semi-quantitative weighting factors to such data sets. Three maps have been produced in a GIS application, namely, Groundwater Availability, Groundwater Development and Livestock Water Prospectivity maps. 4.2.1 Groundwater Availability map Groundwater Availability was defined as the ability of an aquifer at a specific location to supply groundwater in desired quantities that is suitable for exploitation by the end user. The same area can thus have different Groundwater Availability depending upon the ultimate intended use of the resource; a groundwater availability map for cattle will thus be different from one developed for human potable water needs. Groundwater Availability examines the hydrogeological attributes that directly or indirectly affect the availability of ‘suitable’ groundwater. Such attributes include borehole yield, likelihood of recharge and groundwater quality (TDS) within the limits of the intended use, as well as aquifer geology and other hydrogeological attributes that are indicative of availability of suitable groundwater (e.g. groundwater chemical and isotopic signatures). Borehole yields vary widely within the fractured Proterozoic Quartzite Aquifer environment and are quite pronounced at a local scale, although on a large scale it is

Water resources of arid areas

414

possible to delineate a general trend in their spatial distribution. A borehole yield map is thus useful in showing areas with a high probability of Groundwater Availability. High groundwater recharge potential areas have higher Groundwater Availability as a result of regular groundwater replenishment. The groundwater quality map was classified with much emphasis given to the distribution of TDS of a certain limit with respect to availability of groundwater suitable for livestock with a cut-off value of 10,000mg/l, with values above this cut-off figure designating zones of non-available groundwater.

Figure 5. Groundwater availability map. Groundwater Availability in terms of geology was delineated based on the three major aquifer groups, viz, the Kalahari Aquifer, the Karoo Aquifer and the Proterozoic Quartzite Aquifer.

Groundwater evaluation in a complex hydrogeological environment

415

The Karoo Aquifer had the highest groundwater potential, 12% of the data indicating yields greater than 20m3/hr, with the Kalahari and Proterozoic Quartzite Aquifers having values of 9 and 1% respectively. A Groundwater Availability map was thus produced using the algorithm shown in equation 2 below. (2)

where Ywght, Gwqwght, Geowght and Rechwght represent the yield, groundwater quality, geology and recharge weighted maps and Wy, WGwq, Wgeo and WRech represent the weights of each of the maps respectively. After statistical evaluation of the map histogram, the map was then sliced into 3 classes of high, medium and low groundwater availability as shown in Figure 5. Highest potential is found within the Karoo Aquifer to the north, the Kalahari Aquifer to the south and along the ENE-WSW ridge which runs through the centre of the project area. These zones are associated with high borehole yields, high recharge and relatively moderate TDS. The lowest groundwater availability is found in areas where borehole yields are very low to dry, TDS is high (generally >10,000mg/l) and recharge potential is low. This area coincides with thick Kalahari zones; illustrating that Kalahari thickness has a significant influence on groundwater availability. 4.2.2 Groundwater development map Groundwater development in the present context looks at the feasibility of developing and exploitation of a potential groundwater source. The aspect dealt specifically with drilling constraints in the context of cost implications given prevailing hydrogeological conditions. Shallow depths to water level zones were given the highest scores.

Water resources of arid areas

416

Figure 6. Groundwater development map.

Groundwater evaluation in a complex hydrogeological environment

417

Figure 7. Livestock water prospectivity map. Deep-water strikes will require deeper drilling thereby increasing drilling costs, as does thick overburden since as the overburden has to be penetrated first in order to reach the aquifer below. Also, if the overburden is loose, a larger diameter hole, grouting and casing also increase costs. In competent formations, casing may not be required, thus drastically reducing construction costs since boreholes do not have to be cased to total depth to prevent collapsing. Maps were generated for the different attributes noted above with scores assigned to different spatial units within the attribute maps. The Groundwater Development Map was than calculated as follows: (3)

Water resources of arid areas

418

Dwlwght, Dwswght, Obtwght and Afcwght represent the weighted depth to water level, depth to water strike, overburden thickness and aquifer formation competence maps respectively. WDwl, WDws, WObt and WAfc represent weights of each of the four maps. The Groundwater Development Map is shown in Figure 6. Highly favourable to favourable conditions are found along the ENE/WSW ridge associated with thin Kalahari cover. The Karoo Aquifer to the north and the Kalahari Aquifer to the south are characterised by unfavourable groundwater development conditions due to poor aquifer formation competence, deep water strikes and deep water levels. Unfavourable areas within the Proterozoic Quartzite Aquifer coincide with thickest Kalahari cover. 4.2.3 Livestock water prospectivity map Livestock Water Prospectivity (LWP) zones were delineated by combining the Groundwater Availability and Groundwater Development Maps as below LWP=GWP+GWD (4) Most favourable livestock water prospectivity zones occur in areas characterised by good groundwater quality, high potential recharge and thin Kalahari cover making groundwater development most favourable. 5 CONCLUSIONS The use of GIS in geosciences is proving to be a versatile tool in planning and management of groundwater resources. In complex environments such as the fractured Proterozoic aquifer of the project area the use of GIS data integration has been successfully applied to facilitate hydrogeological evaluation and reveal trends that are not easily discernable when analyzing individual maps, and it is apparent that GIS techniques can provide a very useful semi-quantitative methodology that can be utilized in regions where numerical models are inapplicable or not possible due to geological complexity or paucity of data. Within the framework of the project a methodology to evaluate the groundwater potential of the project area with respect to livestock water supply has been developed that could easily be applied in other regions. Furthermore, it is apparent that this GIS approach could be further researched and refined to include other controlling agricultural and sociological parameters such as livestock carrying capacity, grazing conditions, land allocation and other factors in order to develop a series of regional agricultural planning and management tools. REFERENCES Bonham-Carter, G.F. 1996. Geographic information systems for geoscientists, Modelling with GIS, Computer Methods, Geosciences, Vol. 13, Canada, Pergamon Press: 398pp.

Groundwater evaluation in a complex hydrogeological environment

419

Carney, J.N. & Aldiss, D.T. & Lock, N.P. 1994. The Geology of Botswana, Bulletin Geological Survey Botswana. pp37, 113. Department of Geological Survey, 1999. The National Geological Map of the Republic of Botswana, Scale 1:1,500000, First Digital Draft Copy. Department of Geological Survey, 2003. Werda-Mabutsane-Sekoma TGLP Groundwater Survey. Final Report volume 1 by Wellfield Consulting Services.

Application of 2-D resistivity imaging combined with time domain electromagnetic survey to map shallower aquifers in Kunyere valley, northwest Botswana E.M.Shemang1, H.Kumar2 & J.Ntsatsi3 1

Department of Geology, University of Botswana, Botswana 2 Water Resources Consultants Ltd, Gaborone, Botswana 3 Department of Water Affairs, Gaborone, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The Kunyere River, located in northwest Botswana, forms one of the major distributory of the Okavango Delta. A geophysical survey using time domain electromagnetic (TDEM) and 2-D resistivity imaging were carried out in the Kunyere valley with the main objectives of mapping different lithologies and identifying the potential fresh water aquifers in unconsolidated sediments. The TDEM survey results were used to map the depth and lateral extend of the fresh water aquifers and also define the relative variation in water quality on the basis of resistivity. The follow-up 2-D resistivity imaging was used to map the 2dimensional geometry of the aquifers in the area, and define different litho-units to assist in borehole design. This paper presents the results of 2-D resistivity imaging combined with TDEM survey results. The TDEM survey results clearly demarcated the fresh water aquifer boundaries, whilst resistivity imaging differentiated various litho-units at shallower depths. This suggests that 2-D resistivity imaging can be a used as a complementary method to TEM for better understanding and development of the shallower aquifers in the Kunyere Valley and in similar hydrogeological environment elsewhere in the Delta.

1 INTRODUCTION The area of study (Kunyere valley) is located in the North western part of Botswana (see Figure 1). Geologically, the area is underlain by sediments of the Kalahari beds which consists of unconsolidated sands, whose grain sizes range from very fine to medium grained

Application of 2-D resistivity imaging

421

(occasionally coarse grain); clays, silts, calcretes, mudstones, siltstones and sandstones (DWA, 1997). The Kunyere valley system comprised of three tributary valleys (Marophe, Xudum and Matsibe) and the Kunyere Fault. Hydrogeologically, the valley system comprised of multi-layered fresh water aquifer systems with intervening semi-confining units. Shallow unconfined and/or semi-confined units comprise the uppermost freshwater aquifer system which are underlain by an upper semi-confining unit. The lower semi-confined systems contain fresh water and the fresh/ brackish groundwater interface appears to be within the fine-medium sands that comprise these lower semi-confined aquifer systems. Figure 2 is a geologic cross section for the Kunyere Valley oriented southwest to northeast. A two-dimensional (2-D) Electrical Resistivity Imaging (ERI) survey was conducted with the objective of assessing if ERI cross-sections would provide better definition/resolution of aquifer geometry than TEM interpreted sections that were derived from data obtained at discrete locations. Two profiles were surveyed near boreholes BH8255/57 and BH9712/13 in the Kunyere Exploration Area (Figure 1). The first line (Line-A) runs through two exploration boreholes, BH8255 and BH8257, and stretches for a distance of 1km. The

Figure 1. Location map of the area of study.

Water resources of arid areas

422

Figure 2. Geologic cross section of Kunyere valley (After DWA, 1998). second line (Line-B) runs through monitoring borehole BH9712 and exploration borehole BH9713 (DWA, 2003), and stretches for a distance of 800m. In the ERI method, data is acquired using closely spaced (5m spacing) continuous vertical electrical soundings (CVES) along a profile line. Integrating the results of all these soundings provides a vertical slice, or image, of resistivity variation with depth. This information can be utilized to infer the 2-D geometry of different aquifer units as well as water quality variations. The main objectives of 2-D ERI investigation were to: ● Map the two-dimensional geometry of the aquifers, mainly lateral and depth extents. ● Determine resistivity variations with depth, and by inference, borehole yields variations. ● Optimized the location and design of production boreholes.

2 DATA ACQUISITION TEM soundings in the area were conducted using a Geonics™ time-domain EM unit comprised of a PROTEM-D receiver and a TEM-47 transmitter. Data was acquired at base frequencies of 25Hz, 62.5Hz, and 237.5Hz using a square transmitter loop, with a high frequency receiver coil (area 31.4m2) located at the centre of the loop. The 2-D ERI data was acquired using an ABEM™ Lund Imaging system using a 400m long multi-core cable with 5m electrode spacing in the Wenner Array. Data was acquired using continuous, roll-along, vertical electrical sounding method described by Dahlin and Bernstone (1997). A microcomputer, together with an electronic switching unit, was used to automatically select the relevant four electrodes for each measurement. To extend the profile line the roll-along method was employed. This involves moving the cable past one end of the profile line, by several electrode spacings, after completing a sequence of measurements. To plot the 2-D ERI data, a resistivity pseudo-section contouring method is used. In this method, the horizontal location of the plotting point is placed at the mid-point of the

Application of 2-D resistivity imaging

423

set of electrodes used to take that measurement. The vertical location of the plotting point is placed at a distance that is proportional to the separation between electrodes. 3 DATA PROCESSING AND INTERPRETATION One-dimensional inversions of the TDEM sounding were carried out using Interpex™ Temix-XL software. The inversion results were then utilized to draw geoelectric crosssections along each profile line. An interpretation of each section, in terms of different hydrogeological units was carried out by grouping the various resistivity ranges. A broad classification of different hydrogeological units, based on the TEM resistivity, is given in Table 1. Inversion of the 2-D CVES field data was carried out using Geotomo™,s “RES2DINV” software programme. The inversion routine used by this program is based on the smoothness-constrained least-squares method (De Groot-Hedlin and Constable 1990; Sasaki 1992 and Dahlin, 1996). The output of the inversion routine provide a pseudo-section of the apparent resistivity and more significantly a true 2-D resistivity cross-section from which all array geometry factor has been removed along the profile line. Resistivity variations then reflect true subsurface variations in lithology and water quality.

Table 1. Resistivity ranges of different hydrogeological units. Formation

Resitivity (Ωm)

Dry surficial sand (top 10m) Saturated Sand (Fresh water, TDS <1500mg/L) Saturated Sand (Brackish to saline water, TDS >2000mg/L) Clayey or Silty Sand within the freshwater zone

20–5000 10–70 <10 <3

4 RESULTS AND DISCUSSION Inversion of the TEM soundings provided estimates of the depth to the base of fresh water aquifers. The results show that the depth to the base of the fresh groundwater aquifers varies considerably in the area. In some localized areas, the base of the aquifer occurs up to 100mbgl. Two-dimensional resistivity imaging sections along Line-A and Line-B with borehole locations are shown in Figures 3 and 4 respectively. These figures also show the location of boreholes crossed by these profile lines. With the exception of monitoring borehole BH9712, all the boreholes are exploration boreholes that have been pump tested. The tested yield and formation water EC of each borehole is also shown on these figures. The numbers on the left side of

Water resources of arid areas

424

Figure 3. 2D resistivity imaging section along profile line-A at BH8255/8257.

Figure 4. 2D resistivity imaging section along profile line-B at BH9712/9713. boreholes BH8255 and BH9712 represent the formation resistivity obtained from the TEM sounding data. Section Line-A (Figure 3) shows several low and high resistivity zones within the top 35m followed by the geoelectric basement. Low resistivities (generally less than 10m) are interpreted as clay units while high resistivity zones are interpreted as sand units. Below

Application of 2-D resistivity imaging

425

35m, the ERI sections are unable to resolve different lithological units or the base of fresh water aquifer, which might be due to insufficient formation resistivity contrasts. The TEM sounding data also does not resolve different lithological units, however, it does resolve the base of the fresh water aquifer to be at 68m. Another feature of interest in this ERI section are the two low resistivity zones that occur between stations 0–200m and between stations 280–400m. These zones are separated by vertical boundaries and represent areas with dominantly clay and saline water. Boreholes BH8255/57 are located near the boundary of a clayey formation/saline water zone. This example shows advantage of 2-D ERI over TEM soundings in delineating lateral sub-vertical lithological boundaries cost-effectively. Such features can not be picked up by conducting TEM soundings at discrete locations and are pertinent for mapping the aquifer geometry. The results show that while widely spaced TEM soundings are adequate for broad delineation of the aquifer there are significant resistivity versus depth variations between these stations which require a much closer station spacing to ensure their adequate mapping. Section Line-B (Figure 4) shows a variably thick (7–35m) resistive layer (>400Ωm) overlying low resistivity (30–193Ωm) formations. The top layer is interpreted as dominantly sandy formation that overlies low resistive clayey sand. 5 SUMMARY From the above discussion it may be summarized that: ● The 2-D ERI method was helpful in delineating the lateral boundaries of saline water aquifers/ clayey formations. ● ERI is a cost-effective acquisition of high-resolution resistivity mapping for shallower aquifers. This method, when used in conjunction with the TEM, could be a useful tool for delineating the geometry of aquifers. ● The 2-D ERI method was effective in resolving different lithological units to a maximum depth of 35mbgl. However, it did not resolve base of fresh water aquifer or different lithological units below 35m. The depth limitation of ERI could be due to poor formation resistivity contrast in the area. Longer array and powerful current transmitter may be helpful in improving depth of investigation.

REFERENCES Dahlin, T., and Bernstone, C., 1997, A roll-along technique for 3D resistivity data acquisition with multielectrode arrays; Proc. Symposium on the Application of Geophysics to Engineering and Environmental Problems, Reno, Nevada, 927–935. Dahlin, T., 1996, 2D resistivity surveying for environmental and engineering applications: First break, 14, 275–284. deGroot-Hedlin, C., and Constable, S., 1990, Occam’s inversion to generate smooth, twodimensional models from magnetotelluric data: Geophysics, 55, 1613–1624.

Water resources of arid areas

426

DWA, 1997, Maun Groundwater Development Project. Phase 1 Exploration and Resources Assessment. Final Report. pp 134. DWA, 2003, Maun Groundwater Development Project. Phase 2 Resources Assessment and well field development. Main report. pp 80. Loke, M.H., and Barker, R.D., 1996, Rapid least-squares inversion of apparent resistivity pseudosections by quasi-Newton method: Geophysical Prospecting, 44, 131–152. Sasaki, Y 1992, Resolution of resisivity tomography inferred from numerical simulation: Geophysical Prospecting, 40, 453–464.

Theme E: Climate change and its impact

Hydraulic studies in the design of sand dams A.S.Nzaba, H.O.Farah & T.C.Sharma Moi University, Eldoret, Kenya C.W.M.Sitters Faculty of Civil Engineering and Geosciences, Delft, The Netherlands Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: A sand dam is a weir wall that holds a reservoir full of sediment within which water is stored. The weir is usually built in stages to trap sediments as water flows during different storms. The amount of water that can be drained depends on the characteristics of the sediments composing the sand dam. Problems associated with reduced yields of water from sand dams have been blamed on the use of high step increments during the construction of the different stages of the weirs. The effect of step height increments during construction of weirs on the water yield of sand dams has been investigated on a laboratory scale model. Other ways of trapping sediment which ensure sediment with high water storage and yield properties were also investigated. This was done by use of experimental units with different step height increments, flow rates, slopes and weir types. The main findings were that the slope, weir type and flow rates affect the storage properties of the trapped sediment, while step height increment during the different stages of construction has little effect on the properties of the sediment.

1 INTRODUCTION A sand dam or subsurface dam is a weir wall placed across a watercourse behind which successive floods deposit layers of sediment. The sediment absorbs substantial amount of floodwater which can later be drained for domestic or livestock use. In Kenya the use of sand dams has been proposed as a viable option for water supply in arid and semi arid area with seasonal water course and plenty of sediment flowing within the channels. Sand dams have been utilized in Namibia, Indian, Germany and the semi arid regions of Arizona and California of U.S.A (Barrow, 1987). The advantage of sand dams is that evaporation is restricted and contamination of water from the surface is minimal. A sand dam is built in suitable channels where conditions to create a water tight subsurface dam exist. Such conditions exist in rocky or stony catchments. The amount of

Hydraulic studies in the design of sand dams

431

water stored or drained depends on the characteristics of the trapped sediments. The coarser the sediments, the higher the storage capacity and yield of the sand dam. In areas where significant quantities of fine sediments may be transported during floods, certain precautions may be taken during construction of the weir wall to ensure that only the coarse fraction of the sediments is retained in the dam. The weir is raised in stages to prevent fine sediments from settling in the dam basin. The height of each is carefully controlled and each new stage is constructed only after the previous stage has effectively filled up with sediments. Some of the problems that hamper the effective utilization of sand dams is low capacity of water storage due to low porosity of sediments and low quantity of extractable water due to low specific yield of the sediments. The first objective of this study was to investigate the relationship between the porosity and specific yield of the sediments and step height increment used during construction of sand dams. The second objective is to investigate ways of trapping sediments with high water storage and high specific yield. 2 MATERIALS AND METHODS Experiments were conducted to simulate actual conditions in an existing prototype called Nakusuta sand dam, situated in semi arid lands of West Pokot district in northern western Kenya. A geometrically distorted existing laboratory flume model was used for the experiments. The issue of similarity in hydraulics between a prototype and laboratory model are very complicated and complete similarity is difficult to achieve. However, if some rules are followed, it is possible to attain similarity within acceptable limits. Similarity relations can be derived from the dimensionless numbers obtained from the conservation momentum of a small channel stretch conveying a gradually varied unsteady flow. The similarity relations for attaining correct data from research on rivers and open channels can also be obtained from energy equations (Novolk and Cabelka, 1981). In this study the scale ratio relationships used are obtained from derivation of energy equations. The variables that are used for the description of the prototype and model relations are as follows: ML, MB, MH—scales of length, width and height MR, MD, MQ—scales of hydraulic radius, depth and water discharge Mλ, MK, (Md)—scales of friction coefficients and decisive roughness size (decisive diameter of bed load mixture) Mγ′s, Mq′s, Ms—scales of specific weight of bed load, specific discharge of bed load and slope where M refers to prototype to model ratio. The similarity equations are: MQ=MBMD3/2 (1) (2) (3)

Water resources of arid areas

432

(4) (5)

where Mc is the Strickler coefficient scale. The following assumptions are made: (a) The influence of the roughness of channel walls are not taken into account. (b) Relative intensity of turbulence is the same on the model and the prototype. (c) Shapes of the grain distribution curve for sediments in prototype and model are identical. (d) The shapes of the sediment grains in the model and in the prototype are identical. From the above equations it is possible to determine all basic scales for the model studies and from these scales other quantities such as geometric (area and volume), kinematic (time, velocity and acceleration) and dynamic (pressure, energy and power) can be determined. 2.1 Description of the laboratory model The model is designed for bench mounting (Armfold, 1973). The sediment is placed in the channel and water flows along with it over the weir. The sediment carried by water is allowed to settle at the bottom of the collecting/settling tank. The tilting bed and sediment channel is mounted on the manual jackings systems which provides a positive step adjustment, via a hand wheel with graduations of 1% from 0 to 10%. The channel has an aluminium bed and clear perspex sides with a numerated graticule grid provided. The dimension and other properties of the flume are: ● available length for dam=1.2m ● available depth of flume for dam=0.1m ● width of flume=0.075m ● preset discharge rates=0.25, 0.4, 0.6l/s ● Range of slope of flume=0 to 10%. 2.2 Prototype description The following information for Nakusuta dam is available (Kibyii, 2001): ● Original slope and slopes after every stage=1.7, 0.97, 0.89, 0.77% ● Mean channel width at 300m from weir=12.27m ● Depth at dam wall=2.5m. 2.3 Model scales

Hydraulic studies in the design of sand dams

433

The ratios of width, specific weight of bed load, decisive diameter of bed load mixture and the Strickler coefficient obtained from corresponding prototype and model properties are as follows: MB=163.6, Mγ′s =1.0, MD=1.0 and Mc=1.9. The following set of linear equations are obtained from the above ratios: −2.21=1.51og MD−Log MQ =Log MD−log ML−Log Ms 0.56=Log ML−1.331og MD 0.02=Log Ms+log MD 0.00=1.5log MD+1.5log MS−log Mq′s The following values for water discharge, length, depth and specific discharge of bed load scales result: ML=51.4450, MQ=3224.83370, MD=7.3396, MD=0.1427, Mqs′=1.0715 2.4 Determination of porosity The porosity η is given by (6) where V is total volume of soil mass, Vv is the volume of voids, Va is the volume of air and Vw is the volume of water. Eqn.6 can be re-written as (7) where ρB is the bulk density, ρs is the density of the solid particle and Msl is the mass of solids. In determining η, V was obtained by core cutter method. A miniature core made of plastic cylinder of specific height and diameter was pressed down into the sediment layer until it was filled with sample material. The ends of the core were the trimmed flat to the ends of the cutter by a straight cutting edge. The volume of the soil collected in the undisturbed state V, was the volume of the core. The sample collected in the core was oven dried at temperature of between 105°C and 110°C for 24 hours and weighed. This gave the mass of solids (Msl). To determine the specific gravity of the solids ρs, complete density bottles were dried at 105 to 110°C and cooled in a desiccator and then weighed (m1). A 5 to 10g sample of the solid which pass through a 2mm sieve was oven dried and put in the density bottle directly from the desiccator. The bottle and its contents were then weighed (m2). Sufficiently air free distillated water was added so that the soil in the bottle was just covered with water. The bottle was then placed in a desiccator to eliminate air completely. More air free distillated water was then added until the bottle was full and the

Water resources of arid areas

434

contents and the bottle are then weighed (m3). Finally the bottle was emptied and cleaned and filled up with air free distilled water. The bottle is weighed (m4). The specified gravity of the soil particles is given by

2.5 Determination of specific yield and specific retention Specific yield (Sy) is the volume ratio of drained water and to the volume of the soil, while specific retention (Sr) is the ratio of volume of water remaining in the soil to the volume of soil (Linsley et al, 1982). An improvised permeater consisting of a PVC cylinder with a permeable material held over one end was used as a core cutter to obtain the undisturbed sample from the volume. Volume of drained water was obtained by collecting water dripping from the improvised permeater for at least six hours and weighing it. The volume of water retained was then found by finding the difference in weight between the fully drained sample and that of the oven dried sample. 2.6 Grain size distribution Sieve analysis was used to determine grain size distribution. To describe the shape of the grain size distribution curve with a single number, two coefficient were used: – the coefficient of uniformity Cu

(9) – the coefficient of curvature Cc

(10) where d10, d30 and d60 are grain sizes for which 10, 30, and 60% of grain particles are smaller than these sizes respectively. A uniformly graded soil will have its Cu less than 3.0 and well graded soil will have Cu greater than 6 and Cc more than 1.0 but less than 3.0. 2.7 Weir type and step heights The types of weirs used were solid and mesh types. The solid weir was made of metal sheet while the mesh weir was made of wire mesh with an opening of 5mm. The weirs were set at specific heights (representing step height increments). This height was applicable for specific slope and discharge rate. When the space behind the weir is completely filled up with sediments, the flow is stopped. The height of the weir is then increased and flow started to trap more sediments. Sampling of the sediments was done

Hydraulic studies in the design of sand dams

435

after the last height increment of the weir. Sampling was done at the inlet, middle and outlet of the channel. 2.8 Statistical analysis The inherent variability of experimental results necessitates the use of statistical theory in the design of experiments (Mead, 1994). The layout of the experiment conducted in this case was a generalized block design and the principles used in the design were replication and blocking. The treatment for the experiments were step height increments of between 5 and 45mm while the blocks consisted of the following on the model scale: (i) Rate of flow of 0.11, 0.25, 0.4 and 0.6l/s (ii) Solid and perforated weir types (iii) Slopes of 1.5, 4 and 6.26%. The response variables were the specific yield, specific retention, porosity, grain size and grain size distribution coefficients. One and two factor effects were compared by running parallel identical experiments with all but the factor in question held constant. Thus the comparison of a single factor effect on the response variables was conducted by comparing the variability of means of observed values of the response variables for different runs with the same setting of constant slope of 6.26%, flow rate of 0.25l/s and a solid weir. The factor in question (single factor) is this case was the treatments consisting of steps height increments. For the two factor tests, apart from the step height increments the other factor that was different in the parallel run were either the slope, the flow rate or weir type. Thus a slope of 4% was compared with that of 6.26%, a flow rate of 0.25l/s was compared with that of 0.11l/s and a solid weir type was compared with a perforated one, while keeping all other factors constant. The second factor here that is different was termed the block. The variability of the various response variables was assessed by the F-statistical at the level of significance of 5%. When the calculated value for F was greater than the one read out from statistic table (F—critical) for 5% level of significance, the null hypothesis was rejected at that level of significance. The null hypothesis in all cases was that the treatments (step height increments) had no significant effect and the blocking (slope, flow rate or weir type) too had no significant effect on the response variables. This suggested that there was no difference in response variable due to changes in step height increment. Similarly the null hypothesis for the blocking effect, suggested that there was no significant difference in the response variable whether there were changes in the slope, flow rate or weir type. 2.9 Summary of experiment Figure 1 shows the flowchart of the experiment. The observations recorded were discharge, slope, weir type, flow depth and depth of sediment layer. Porosity, specific yield, specific retention and grain size distribution curves of the sediments were then determined. All sizes of sediments (d10, d30, d50 and d60) increase very much from the middle of the channel towards the outlet. Figure 2 illustrates the single factor test at 5% level of

Water resources of arid areas

436

significance. The letters O, M and I represent the outlet, middle and inlet sections of the channel respectively. Figure 2 shows that step height increments affect the sizing of sediments at the outlet and middle sections. The effect of moderately high step increment of about 10mm to about 30mm (73 to 220mm on the prototype) have more desirable effect as bigger grain sizes are obtained. However the step height increment has no effect on the coefficient of uniformity and curvature. The effects of step height increment on Sr, Sy and η are shown in Table 1. The results show that step height increment has no effect on Sr, Sy and η. 2.10 Effects of rate of flow The size of the sediment deposited increase with increase in flow. The significance test showed that flow rate significantly affects the sizing of the sediments along the channels. The flow rate has also significant effect on Sr and Sy (see Table 2). From the experiment it is desirable to have higher flow rates of say 0.25l/s as compared to 0.11l/s to get sediments with better properties of storage and yield.

Hydraulic studies in the design of sand dams

437

Figure 1. Flow chart representing the experimental process.

Water resources of arid areas

438

Figure 2. Single factor test for the significance of step height increments at a flow rate of 0.25l/s and slope of 6.26% for solid barrier. Table 1. Single factor test for treatment effect (6.26% slope, 0.25l/s discharge, solid weir) on Sr, Sy and η. Sr

Treatments F F Has Sy crit effect

Inlet 0.48 3.23 no Middle 0.56 3.23 no Outlet 0.29 3.23 no

Treatments F F Has η crit effect

Inlet 0.47 3.23 no Middle 0.32 3.23 no Outlet 0.19 3.23 no

Treatments F F Has crit effect

Inlet 1.09 3.23 no Middle 0.56 3.23 no Outlet 1.61 3.23 no

Table 2. Two-factor test for solid weir, 6.26% slope and 0.11l/s versus 0.25l/s. Treatments F F Has crit effect Sr Inlet 0.93 3.79 Middle 1.39 3.79 Outlet 0.37 3.79 S y Inlet 3.03 3.79 Middle 1.59 3.79 Outlet 3.55 3.79 η Inlet 2.86 3.79 Middle 0.82 3.79 Outlet 1.84 3.79

Blocks F F Has crit effect

no no no

13.17 5.59 yes 3.32 5.59 no 4.46 5.59 no

no no no

20.61 5.59 yes 5.12 5.59 no 15.49 5.59 yes

no no no

3.84 5.59 no 1.09 5.59 no 0.21 5.59 no

Hydraulic studies in the design of sand dams

439

2.11 Effect of slope Slope has significant effect on Sy and η (see Table 3). The size of sediment is found to be much smaller at the outlet for lower channel slope of 4% compared to 6.26% This is an undesirable trend for the lower slope since it implies much less specific yield compared to the higher slope of 6.26%. Moreover the coefficient of uniformity is much smaller for the higher slope than for the lower slope at the middle and inlet sections. From this it can be said that it is desirable to have higher slopes to get better sediments characteristics. 2.12 Effect on weir type Weir type has a significant effect on the Sy and η of the sediments. However, it does not significantly affect Sr. Table 4 shows the two-factor test with conditions of 0.25l/s flow and 6.26% slope between solid versus mesh weir. 2.13 Practical implication for design of sand dams Since the step height increment has no effect on the properties of the sediments, a single stage construction of the weir wall is recommended in construction practice. The advantage of a single stage construction is the savings in cost in mobalizing several times the concerned communities for the construction. Field experiences in Kenya (Thomas, 1998) and Namibia (Burger and Beaumont, 1967) showed that the single stage construction is less expensive than the multi stage construction. The channel slope should be preferably greater than 0.9%. Slopes of between 1% and 2% have been recommended in Kenya (Kibyii, 2001) because higher slopes allows higher specific yield

Table 3. Two factor test for 0.24l/s flow, solid weir and 4% versus 6.26% slope. Treatments F F Has crit effect Sr Inlet 0.61 3.44 Midlle 0.73 3.44 Outlet 0.38 3.44 Sy Inlet 0.87 3.44 Middle 1.01 3.44 Outlet 0.28 3.44 η Inlet 1.82 3.44 Midlle 3.31 3.44 Outlet 2.21 3.44

Blocks F F Has crit effect

no no no

3.75 5.32 no 3.89 5.32 no 0.25 5.32 no

no no no

22.81 5.32 yes 1.01 5.32 no 9.05 5.32 yes

no no no

98.70 5.32 yes 89.20 5.32 yes 53.15 5.32 yes

Water resources of arid areas

440

Table 4. Two-factor test for 0.25l/s flow, 6.26% slope and solid versus mesh weir. Treatments F F Has crit effect Sr Inlet 1.04 3.44 Midlle 0.78 3.79 Outlet 0.26 3.44 Sy Inlet 2.01 3.44 Middle 1.23 3.79 Outlet 1.10 3.44 η Inlet 1.69 3.44 Middle 0.90 3.79 Outlet 2.05 3.44

Blocks F F Has crit effect

no no no

1.79 0.65 0.05

5.32 no 5.59 no 5.32 no

no no no

23.34 5.32 yes 13.32 5.59 yes 13.71 5.32 yes

no no no

29.27 5.32 yes 47.89 5.59 yes 117.01 5.32 yes

and more importantly they give highest storage per unit volume (specific capacity). The discharged should at least be of the order of 65l/s per m width. The upper limit to these values could not be known because of the scale constraints for the model and finally a solid weir is better than a meshed or perforated weir. 3 CONCLUSIONS The sand dams simulation experiments, based on the notions of hydraulic modeling where a geometrically distorted existing laboratory flume model was used, revealed a number of trends in order to achieve the best sediments in terms of storage and yield properties. The following conclusions can be made for dam construction: ● That the step height increment use for design of sand dams should be at least 185mm. A much higher step height increment of up to 350mm is desirable and more economical ● The weir used across the water channel for trapping sediment should be of solid type ● The initial channel slope should be high or at least 0.9% ● The flow rate during sediment deposition should be at least 65l/s per m width of the channel.

REFERENCES Armfield, 1993. Instruction manual: Sediment transport channel. England, Armfields Ltd. Barrow, C. 1987. Water resources and agricultural development in the tropics. United Kingdom, Longman Scientific and Technical Publishers.

Hydraulic studies in the design of sand dams

441

British Standards Institution, 1995. BS 1377: Methods of test for soils in civil engineering purposes. Burger, S.W. & Beaumont, R.D. 1967. Sand storage dams for water conservation. CSIR report R. Meg 329. Kibiiy, J.K. 2001. Runoff harvesting using sand dams in ASAL areas of west Pokot, Kenya: some design aspects. PhD thesis, Moi University, Kenya. Linsley, R.K., Kohler, M.A. & Paulhus, J.L.H. 1982. Hydrology for Engineers. New York, McGraw-Hill. Mead, R. 1994. The design of experiments: Statistical principles for practical applications. New York, Cambridge University Press. Novak, P. & Cabelka, J. 1981. Models in hydraulic engineering—Physical principles and design applications. London, Pitman. Nzaba, S. 2001. Hydraulic similitude studies in the design of sand dams. Msc thesis. Moi University, Kenya Thomas, D.B. 1999. Where there is no water—A story of community water development and sand dams in Kitui district, Kenya. Nairobi, Majestic Printing Works Ltd.

Designing and implementing an aircraft survey mission using high-resolution digital multi-spectral camera for vegetation mapping for upscaling transpiration of Serowe, Botswana Y.A.Hussin1, M.W.Lubczynski1, O.Obakeng1,2 & D.C.Chavarro1 1

International Institute for Geoinformation Science and Earth Observation (ITC), Enschede, The Netherlands 2 Geological Survey of Botswana, Lobatse, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: An aerial survey was designed and implemented to collect images for mapping vegetation for an upscaling transpiration study in Serowe, Botswana. The images were collected using a multi-spectral digital camera (TETRACAM) in three bands in green, red and near infrared. The data is collected in a rectangular frame (e.g. image) of 1280×1024 pixels. The camera was mounted on a small aircraft to collect data in 30, 60 and 100cm spatial resolution. Two survey missions were successfully done in November 2003 and February 2004.

1 INTRODUCTION The existence and presents of people, plants, and animals are very much related to water availability. In arid and semi-arid regions the main source of water is groundwater. It is well known that vegetation cover has an influence on the hydrological cycle through interception, infiltration through to ground and transpiration of water. Thus it influences the amount and quality of ground water. Therefore, vegetation plays an important rule in hydrological cycle. Mapping vegetation cover especially in semi-arid region like Botswana is very important and can help as a good indicator of ground water since water is scarce in this country. The growing human population leads to increasing demand for water, both in terms of quantity and quality, especially for drinking water. Therefore, water for people and for the environment is an important issue to be considered in the framework of sustainable use of the resources.

Designing and implementing an aircraft survey mission

443

Having in mind all what has been said in the above, different types or species of vegetation would deals differently with the issue of influencing the hydrological cycle by intercepting, infiltrating through to ground and transpiring of water. The primary goal of the vegetation mapping effort is to classify vegetation to different ecosystem, communities and species. Vegetation mapping has been developed because of the demand for obtaining more detailed map about species habitat distribution, community mapping and in general its resources. In this process the availability of suitable sensors capable of recording images in a suitable spectral and spatial resolution is an important tool in the analysis of vegetation mapping. The process of vegetation mapping can be greatly aided by interpretation of air and satellite-based images. These images (often captured in multiple spectral bands) can be used to delineate a variety of vegetation types based upon species and/or structure. Further, visual interpretation or digital processing of multi-spectral data can help delineating certain features of interest of vegetation (e.g., vegetation types, species, canopy size and structure, density, biomass, leave area coverage), thus aiding to the mapping process. A flight campaign was planned to collect very high spatial resolution (e.g. 30, 60 and 100cm) aircraft images of an area in Serowe, Kalahari, Botswana using the multi-spectral digital camera TETRACAM for transpiration mapping by up scaling of sap flow measurements. The objective of this paper was to design and implement an aircraft survey mission using digital multi-spectral camera for vegetation mapping for upscaling transpiration of part of Serowe, Botswana. 2 STUDY AREA The study area is located in the Central District, about 275km NE of Gaborone the capital of Botswana. Topography is gentle, which varies from 1060 meters above sea level to approximately 1240. It is characterized to be lower in the east and southeast of the region, and the highest location in the vicinity of the escarpment edge. From these ones the average slope is 5% and it gradually decrease to less than 1% towards the east and southeast. Soils units, which can be found in that region, are related to arenosols, regosols, lixisols, luvisols and vertisols. Arenosols are the most common soil units in the study area. It has low moisture retention capacity than the other soil units. Climate is a semi-arid with a mean annual rainfall of 447mm. Rainfall occurs mainly in the summer followed by a dry winter season. Summer season stretches from October to April and the winter begins in May to September (Tyson, 1986; Obakeng, 2000). Main vegetation type is thought that belong to the Northern Kalahari Tree and Bush Savanna. Trees are mostly of Acacia specie, which are characterized by the marked tendency to occur in cluster, and are normally accompanied by a variety of grass species such as Ariatida and Eragrotis. Vegetation communities are determined by location on either sandveld or hardveld areas. Dense vegetation is found within and along river courses. This suggests that the vegetation density is governed by the availability of water, which may be partly controlled by topography and geomorphology (Obakeng, 2000).

Water resources of arid areas

444

3 AERIAL SURVEY AND DATA COLLECTION The airborne multi-spectral data was collected using TETRACAM multi-spectral digital camera, which collects its data in three spectral bands namely red, green and near infrared. The data is collected in a rectangular frame (e.g. image) of 1280×1024 pixels. The size of the pixel (e.g. ground resolution) is depending on the altitude of the aircraft above the ground. Therefore, the camera was mounted on a small aircraft (e.g. Cessna 210) (Figures 1 and 2). The camera saves the image in DCA format (Digital Camera Format), which is a compressed file that can be

Figure 1. Cessna 210 Aircraft used for the data collection. un-compressed and transferred to a Bitmap format. Bitmap format file can be imported to any image processing software. The Airborne data was collected in three different spatial resolutions 30, 60 and 100cm. An area of 10×10Km was selected as a study site. It is located in the hardveld part of the Serowe terrain, on which two multi-spectral IKONOS satellite scenes of November 2001 and February 2002 were collected. For this study area, two aerial surveys were implemented to collect the multi-spectral digital camera data. The first aerial survey was done in November 2003 and the second one was done in February 2004. These surveys were designed and implemented using Aerial-Photography types of survey (Figure 3). The survey divides the area into flight lines. Within each flight line, images were collected with a front overlap of 20% and a side overlap between flight lines of 20% too. To start imaging in each line the aircraft starts at a certain coordinates till the end of the line. The time lag of the camera, which is the time between two shots that the camera uses it to process the image, was used as the time between two image in any flight line

Designing and implementing an aircraft survey mission

445

(Table 1). The image data is stored in what is known as the flash cards of 256MGB. Then the

Figure 2. The location where the camera was mounted.

Water resources of arid areas

446

Figure 3. The design of the aerial survey of collecting digital multispectral images. Table 1. An example of the flight lines and images collected in 60cm resolution. Line no

Deg east

Start

End

1 2 3 4 5 6 7 8 9 10 11

26:16:20 26:16:40 26:17:00 26:17:20 26:17:40 26:18:00 26:18:20 26:18:40 26:19:00 26:19:20 26:19:40

22:17:20 22:22:45 22:17:20 22:22:45 22:17:20 22:22:45 22:17:20 22:22:45 22:17:20 22:22:45 22:17:20

22:22:45 22:17:20 22:22:45 22:17:20 22:22:45 22:17:20 22:22:45 22:17:20 22:22:45 22:17:20 22:22:45

Designing and implementing an aircraft survey mission

12 13 14 15 16 17 18 19 20 21

26:20:00 26:20:20 26:20:40 26:21:00 26:21:20 26:21:40 26:22:00 26:22:20 26:22:40 26:22:50

22:22:45 22:17:20 22:22:45 22:17:20 22:22:45 22:17:20 22:22:45 22:17:20 22:22:45 22:17:20

447

22:17:20 22:22:45 22:17:20 22:22:45 22:17:20 22:22:45 22:17:20 22:22:45 22:17:20 22:22:45

Figure 4. Example of the data with 30cm spatial resolution image. images are downloaded to a computer disk. One DCA file, which contained 3 images (green, red and NIR) in compressed format, has a size of 1.3MGB. Consequently you can save up to 180 images in each of the 256 flash card. The following data were collected: 1. 30cm spatial resolution: 39 flight lines with a total of 910 images 2. 60cm spatial resolution: 21 flight lines with a total of 333 images 3. One meter spatial resolution: 14 flight lines with a total of 187 images Example of the above 3 spatial resolution images collected can be seen in Figures 4–6.

Water resources of arid areas

448

Figure 5. Example of the data with 60cm spatial resolution image.

Figure 6. Example of the data with 1-m spatial resolution image.

Designing and implementing an aircraft survey mission

449

ACKNOWLEDGMENT This research work was partly supported by the internal research fund of GWFLUX Project at ITC. However, Botswana Geological Survey (BGS) has offered the main financial support of the aerial survey missions, fieldwork logistics and transportation. The authors appreciate and acknowledge the support of Botswana Geological Survey. REFERENCES Obakeng, O.T. 2000. Groundwater recharge and vulnerability: A case study at the margins of the south-east Central Kalahari Sub-basin, Serowe region, Botswana. Unpublished MSc, ITC— International Institute for Geoinformation Science and Earth Observation, Enschede. Tyson, P.D. 1986. Climatic Change & Variability in Southern Africa. Cape Town, South Africa: Oxford University Press.

Relevance of groundwater interaction with surface water to the eco-hydrology of semiarid regions John Y.Diiwu Alberta Research Council Inc., Vegreville, Canada Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The interaction of groundwater with surface water is an important process for maintaining the ecosystem. The process affects the ecology of surface water by sustaining streamflow during periods of low flow, moderates water level fluctuations of groundwater-connected lakes, and maintains wetlands which serve as habitats for a myriad of wildlife. The interaction also helps to stabilize water temperature as well as concentrations of nutrients and other organic/inorganic compounds in water. Thus groundwater interaction with surface water helps to provide thermal refuge for aquatic species in semi-arid regions where temperatures may otherwise rise to levels that may be lethal to these species. With the growing demand for the sustainable management and utilization of natural resources a better understanding of all components of the ecosystem, such as the linkage between groundwater and surface water becomes imperative. This is even more relevant for the semi-arid regions where the impacts of environmental stresses tend to be more pronounced. This paper is the refore intended to review fundamental concepts of the ecohydrology of the interaction of groundwater with surface water, and discuss the relevance of this interaction to the sustainable management of water resources of semi-arid regions.

1 INTRODUCTION Groundwater systems are not isolated from surface water systems, but are in continuous dynamic interaction at local, intermediate and regional scales. The degree of the interaction between groundwater and surface water depends on physiographic and climatic conditions. Irrespective of the degree of the interaction between the two systems, development and/or contamination of one ultimately affects the other, and hence the entire ecosystem (Lamontagne et al., 2003). An understanding of the basic principles of

Relevance of groundwater interaction

451

the interactions is therefore needed for effective management of water resources. This is even more imperative for semi-arid regions where water resource systems are highly vulnerable due to climate change and anthropogenic activities. Interest in the relationship of groundwater with streams, lakes, wetlands and estuaries increased in recent years due to concerns about acid rain, eutrophication, and the disappearance of coastal ecosystems as a result of development (Winter, 1995). In the last two decades, attention has been focused on exchanges between near-channel and inchannel water, which are necessary for evaluating the ecological structure of streams and for designing stream restoration and riparian zone management programmes. The need for a holistic approach to environmental protection has heightened the attention of ecologists, geoscientists, and watershed managers to groundwater interaction with surface water. The partitioning of precipitation into surface runoff, infiltration and potential recharge/ discharge is highly variable in space and time in semi-arid regions. Understanding the spatial and temporal variability of these processes at a range of scales improves our ability to quantify and manage the available water resources. The recharge/discharge component, which links groundwater and surface water systems has received renewed attention in the last few decades. This paper is therefore intended to review fundamental ecological and hydrological concepts useful for understanding groundwater interaction with surface water, discuss the relevance of the interaction to the ecology of semi-arid regions, and provide information for further studies of this important pathway between groundwater and surface water systems. 2 MECHANISMS OF GROUNDWATER INTERACTION WITH SURFACE WATER Surface and subsurface water interactions occur by subsurface lateral flow through the unsaturated soil and by infiltration into or exfiltration from the saturated zones. Also, in the case of karst or fractured terrain, interactions occur through flow in fracture or solution channels. In general, subsurface flow through porous media is sluggish. The mechanisms by which subsurface flow enters streams quickly enough to contribute to streamflow responses to individual rainstorm and snowmelt inputs are discussed in the literature (Beven, 1989; Winter, 1995 and Lamontagne et al., 2003). In particular, Beven (1989) identifies four mechanisms that account for fast subsurface contributions to the storm hydrograph: translatory flow, macropore flow, groundwater ridging, and return flow. Translatory flow, also known as plug flow or piston flow (Hewlett and Hibbert, 1967), is easily observed by allowing a soil column to drain to field capacity and then slowly adding a unit of water at the top. It would be observed that some water flows from the bottom immediately, but this is not the same water that was added at the top. Macropore flow is fast flow through larger noncapillary soil pores, resulting in rapid subsurface responses to storm events (Beven and Germann, 1982). Groundwater ridging describes the large and rapid increases in hydraulic head in groundwater during storm events (Sklash and Farvolden, 1979). As a result, an increase occurs in the net hydraulic gradient toward the stream and/or the size of the seepage face, thus enhancing fluxes to the

Water resources of arid areas

452

stream. The streamflow contribution induced thereby may greatly exceed the quantity of water input that induced it. Return flow is the discharge of subsurface water to the surface. This may result if the water table and capillary fringe are close to the soil surface, such that small amounts of applied water are necessary to saturate the soil surface completely (Dunne and Black, 1970). The response of any particular watershed may be dominated by a single mechanism or by a combination of mechanisms, depending on the magnitude of the storm event, the antecedent soil moisture conditions in the watershed, and/or the heterogeneity in soil hydraulic properties in the watershed (Sklash and Farvolden, 1979). 3 GROUNDWATER INTERACTION WITH STREAMS Large scale exchange of groundwater with surface water is controlled by the distribution and magnitude of hydraulic conductivities, both within the channel and the associated alluvial plain sediments; the relation of stream stage to the adjacent groundwater level; and the geometry and position of the stream channel within the alluvial plain (Woessner, 2000). The direction of the exchange processes varies with hydraulic head, whereas flow depends on sediment hydraulic conductivity. Storm events and seasonal patterns alter the hydraulic head and thereby induce changes in flow direction. Two net directions of flow are: the influent condition where surface water contributes to subsurface flow (losing stream, Figure 1), and the effluent condition where groundwater drains into the stream (gaining stream, Figure 2). On one hand, variable flow regimes could alter the hydraulic conductivity of the sediment via erosion and deposition processes and thus affect the intensity of groundwater interaction with surface water (Brunke and Gonser, 1997). During periods of low precipitation, baseflow in many streams constitutes the discharge. On the other hand, under conditions of high precipitation surface runoff and interflow gradually increase, resulting in higher hydraulic pressures in the lower stream reaches, which cause the river to change from effluent to influent condition, infiltrating its banks and recharging the aquifer. Thus,

Figure 1. Schematic of a losing stream.

Relevance of groundwater interaction

453

Figure 2. Schematic of a gaining stream.

Figure 3. Schematic of a perched stream. successive discharge and recharge of the aquifer has a buffering effect on the runoff regimes of rivers (Brunke and Gonser, 1997). In perennial streams, baseflow is more or less continuous, whereby these streams are primarily effluent and flow continuously throughout the year. Intermittent streams receive water only at certain times of the year and are either influent (losing) or effluent (gaining), depending on the season. In ephemeral streams the groundwater level is always beneath the channel, so they are exclusively influent when they are flowing (Gordon et al., 1992). The streambed of an ephemeral stream is always separated from the aquifer by the unsaturated zone; thus it is also called a perched or discontinuous stream, shown in Figure 3.

Water resources of arid areas

454

3.1 Groundwater interaction with lakes The hydrologic regime of a lake is strongly influenced by the regional groundwater flow system in which it is located. This interaction plays a critical role when the water budget for the lake is being evaluated. Lakes dominated by surface water typically have inflow and outflow streams, while

Figure 4. Illustration of the interaction of the surface water system, the local and regional groundwater systems and the hyporheic zone. seepage lakes are groundwater dominated. The type of interactions between groundwater and lakes are generally similar to interactions with streams. The main difference is that lakes have a much larger surface water and bed area. Furthermore, the slower flowthrough rates in a lake often result in accumulations of low permeability sediments in the lake floor which can affect the distribution of seepage. As a result, the rate of seepage is often greatest around the lake margin where wave action may restrict the deposition of finer sediments (Winter, 1995). The rates of groundwater inflow are controlled by watershed topography and the hydrogeologic environment (Gilbert et al., 1994). 3.2 Groundwater interaction with wetlands Wetlands typically occur in areas where groundwater discharges to the land surface or in areas where ground conditions impede the drainage of water. For situations where impeded drainage occurs, stream depletion effects are unlikely to be significant because the layer impeding drainage is also likely to inhibit the upward transmission of any pumping effects. However, in areas where groundwater springs discharge into wetlands, the pumping from underlying aquifers can affect the amount of groundwater discharge to the wetland (Winter, 1995; Gilbert et al., 1994). 3.3 The hyporheic zone The hyporheic zone, shown in Figure 4, is the region of saturated sediment where surface water and groundwater are actively mixing and exchanged (Gordon et al., 1992). Hyporheic processes occur at a variety of scales, from the small scale exchanges caused by obstacles along the stream bottom to the transit of surface water through buried

Relevance of groundwater interaction

455

paleochannels (Woessner, 2000). The measurement of hyporheic and riparian processes have been widely reported in the literature, even though these processes are often studied separately from groundwater-surface water interaction. Since groundwater and hyporheic processes are not independent of one another, to be able to integrate groundwater-surface water interactions into hydrological and ecological models for application in semi-arid regions, there is a need to integrate studies on hyporheic and riparian processes with those on groundwater-surface water interactions (Jones and Holmes, 1996). 4 IDENTIFICATION AND MEASUREMENT OF THE INTERACTION The methods developed so far for measurement of groundwater-surface water interaction are extremely complex, require specialized knowledge to use them, and are resource intensive. Tools for identification of the presence of groundwater interaction with surface water range from inexpensive to resource intensive, and may be moderate to highly complex to use. First, a topographic map and aerial photo, braided channels, ancient stream channels, and dense vegetation may indicate a groundwater—surface water interaction zone. Next, vegetation type, such as cottonwood, and the presence of algae along shallow edges of waterways, may point to a groundwater—surface water interaction zone (Gordon et al., 1992) Various probes may be used to measure changes within the channel, which may indicate the points of groundwater-surface water interaction. Temperature probes are commonly used to indicate the influence of groundwater on surface water. Hyporheic probes may be used to measure interstitial flow rates and change in gradient. Also, the potential for groundwater and surface water to interact, which is indicated by change in hydraulic head, may be measured using minipiezometers (Brunke and Gonser, 1997). The ability to detect and quantify patterns in groundwater-surface water interaction at nested spatial scales may be enhanced through the use of techniques complimentary to measurements using minipiezometers. In particular, accretion studies of streamflow and thermal mapping can compliment minipiezometer use and yield more complete perspective on valley segment to reach scale patterns of groundwater-surface water interactions at smaller spatial scales (Gordon et al., 1992). This may involve the use of minipiezometers at a high sampling resolution (Baxter and Hauer, 2000), fine scale measurement of streambed temperature (Gordon et al., 1992), use of seepage meters (Lee and Cherry, 1978), digging sampling pits and performing dye injections (Dahm and Valett, 1996), or injection of conservative tracers (Gordon et al., 1992). Any attempt to characterize patterns of groundwater-surface water interaction can benefit from a multiscale approach, as well as the use of multiple, complimentary methods. 5 QUANTITATIVE ANALYSIS The flow of water on the surface, and in the unsaturated and saturated zones is driven by gradients from high to low potentials. The hydraulic connection between the stream and groundwater may be direct, as shown in Figures 1 and 2 above. On the other hand, it may be disconnected by an intervening unsaturated zone, with streams losing water by

Water resources of arid areas

456

seepage through a streambed down to a deep water table, as shown in Figure 3 above. The degree of connection can change over different reaches within any one stream and from time to time over the same reach. For hydraulically connected stream-aquifer systems, the resulting exchange flow is a function of the difference between the river stage and aquifer head. A simple approach to estimate flow is to consider the flow between the river and the aquifer to be controlled by the same mechanism as leakage through a semi-impervious stratum in one dimension (Rushton and Tomlinson, 1979). This mechanism, based on Darcy’s law, where flow is a direct function of the hydraulic conductivity and head difference, can be expressed as: q=k∆h (1) Where, ∆h=ha−hr (ha is aquifer head, and hr is river head/stage); q is flow between the river and the aquifer (positive for baseflow for gaining streams, and negative for river discharge for losing streams); and k is a constant representing the streambed leakage coefficient (hydraulic conductivity of the semi-impervious streambed stratum divided by its thickness). Equation (1) can be used to represent both baseflow and river discharge, even though in practice, the mechanisms representing the two processes can be different. At times of high recharge, the leakage calculated by the linear relationship in Equation (1) is much greater than would occur in practice and takes no account of water as its volume increases. For such increased resistance to flow Rushton and Tomlinson (1979) propose a nonlinear relationship of the form: q=k1[1−exp(−k2∆h)] (2) Where k1 and k2 are constants. In cases where the suggestion of a maximum flow rate is not acceptable, Rushton and Tomlinson (1996) propose a combination of linear and nonlinear relationships of the form: q=k1∆h+k2 [1−exp(−k3∆h)] (3) Where k1, k2 and k3 are constants. In semi-arid regions where the aquifer head is lower than the river head most of the time, an exponential relationship with a maximum flow is more appropriate. Under such conditions, channel seepage is often the largest source of recharge. The magnitude of infiltration depends on a variety of factors, such as hydraulic properties of the unsaturated zone, available storage volume in the unsaturated zone, channel geometry and wetted perimeter, flow duration and depth, antecedent soilwater content, clogging layers on the channel bottom, and water temperature. 6 ECOLOGICAL SIGNIFICANCE In semi-arid regions where intense runoff occurs in a relatively short periods of time, closed topographic depressions of varying sizes are filled by runoff to form ephemeral ponds or wetlands. Playas in arid and semi-arid regions are some examples of such

Relevance of groundwater interaction

457

ephemeral ponds (Gordon et al., 1992; Brunke and Gonser, 1997). As the water level in a pond occupying a depression rises in response to input from overland flow and streamflow, water flows from the pond to groundwater where the adjacent groundwater level is lower than the pond. The period of standing water in the depression affects the species richness of aquatic invertebrates, amphibians, and their predators. From a study of 22 wetlands Snodgrass et al. (2000) found that amphibian species richness increased with increase in the duration of standing water in the wetlands, but found no significant relationship between species richness and wetland size. In semi-arid regions, intense runoff coupled with high evapotranspiration produces wetlands with intermediate duration of interaction with groundwater. This is crucial for biodiversity because such wetlands maintain high productivity by periodic drying, which results in routine recycling of organic materials and nutrients (Gordon et al., 1992; Snodgrass et al., 2000). The hyporheic zone, as shown in Figure 4 above, is a mixture of surface water and groundwater, and so has physical and chemical characteristics considerably different from stream water. The zone is therefore an ecotone between the surface environment characterized by light, high dissolved oxygen, and temperature fluctuation and the groundwater environment characterized by darkness, less oxygen, and stable temperature (Gilbert et al., 1994). Invertebrates living in the hyporheic zone exploit the groundwater environment to varying degrees. Some species spend their entire life cycle in the hyporheic zone, while others spend their egg and laval stage in the zone, and then move to the surface environment to spend their adult life. A third category of species use the hyporheic zone only to seek protection from unfavorable situations (Gilbert et al., 1994). The food web of the hyporheic zone is fueled by the heterotrophic microbial communities which depend on dissolved oxygen provided by surface water exchange, particulate organic carbon, and dissolved organic carbon in nutrient-rich groundwater. The microbes provide food for grazers, which in turn provide food for invertebrate predators. Dissolved organic carbon stored in the hyporheic zone can serve as a food source when it is not readily available in surface water, and therefore has a crucial influence on the metabolism of the fluvial ecosystems (Brunke and Gonser, 1997). The hyporheic zone provides a number of ecologically important services. When surface water recharges groundwater, there is opportunity for organic pollutants and detritus to become trapped in the sediment. The bacteria may then catalyze reactions that could change the chemicals into less toxic forms or into available nutrients. For instance, in contaminated aquifers many bacterial micro-organisms residing in groundwater and sediment interstices can aid in groundwater remediation by degradation and denitrification (Gilbert et al., 1994). During floods, excess water that enters bank storage may percolate to recharge groundwater or may re-emerge at a different location in the watershed and at a different time. These diversions allow the onslaught of water into streams to be delayed by days, weeks, or even months and thus mitigates the effects of flood flows (Winter, 1995; Brunke and Gonser, 1997). The interaction of groundwater with surface water within the hyporheic zone also has a thermal service. Since groundwater temperatures remain relatively constant, the water that discharges tends to be cooler than surface water in semi-arid regions. The hyporheic zone therefore serves as a thermal refuge for fish and other aquatic species in semi-arid regions. The zone also serves as a habitat for micro-organisms, macro-invertebrates, fish and wildlife; provides flow augmentation; refugia for endangered aquatic species under conditions of increased

Water resources of arid areas

458

fragmentation and degradation of aquatic habitat; and food source for fish in surface water ecosystems and organic matter for microbial activity in groundwater ecosystems (Winter, 1995). Surface water moving into groundwater is one of the ways in which microorganisms may colonize groundwater environments. The presence or absence of certain groundwater species may indicate the location of groundwater-surface water interaction zones and a decline in the diversity of groundwater species may indicate a decline in water quality (Gilbert et al., 1994). Groundwater invertebrates and microorganisms are an important food source for fish, and so the interaction of groundwater with surface water, which determines the availability of such organisms, has the potential to affect the viability of native fish populations (Gilbert et al., 1994). 7 ANTHROPOGENIC IMPACTS AND WATER RESOURCE SUSTAINABILITY Valley bottoms in semi-arid regions often serve as desirable areas of grazing and agriculture because of continuous availability of soilwater in the unsaturated zone and hence green pasture throughout the year. While these areas have the ability to introduce the cooling effects of groundwater to surface water and continuously make soilwater available in the unsaturated zone, they are also easily degraded by mismanagement. Grazing and agriculture may cause accelerated erosion and soil compaction in the valley bottoms, thus causing the permanent loss of such vital components of the ecosystem in semi-arid regions (Gilbert et al., 1994). In semi-arid regions, crop production requires consumptive use of large quantities of water. Water, which is already scarce must be shared among several consumptive as well as non-consumptive uses. Consequently, society faces serious water management problems. The decline of groundwater levels due to over-pumping ultimately results in reduced baseflow, which would have discharged into surface water to sustain aquatic life during periods of low flow. At sufficiently large pumping rates, these declines induce flow out of the body of surface water into the aquifer, and this leads to streamflow depletion. As discussed in a previous section, groundwater-surface water interactions are also important in situations of groundwater contamination by polluted surface water, and in situations of degradation of surface water by discharge of saline or other low quality groundwater. An understanding of groundwater-surface water interaction in semi-arid regions is therefore important for the sustainable management of water resources in those regions. 8 RESEARCH NEEDS An understanding of the near-channel and in-channel exchange of water, solutes, and energy is an important key to evaluating the ecological structure of stream systems and their management. Despite the recent increase in research on groundwater-surface water exchange, there are still many related processes that are not well understood. The relative importance of variables affecting the activity of the hyporheic zone at sediment and reach scales over time is unclear, and the spatial and temporal dynamics of groundwater

Relevance of groundwater interaction

459

discharge and recharge along active channels in varying geomorphic settings needs to be further investigated (Winter, 1995). Jones and Holmes (1996) concludes that whereas surface-hyporheic exchanges and water residence times are known to be important regulators of subsurface biochemical transformations, the manner in which these parameters vary across streams and under different climatic conditions, such as semi-arid regions is not yet known. The effect of heterogeneity on water fluxes in general, and specifically between groundwater and surface water is still a major challenge. The hydraulic properties of stream and lake beds control the interactions between groundwater and surface water systems, but these properties are normally difficult to measure directly. The primary limitation has so far been the difficulty of spatially defining the hydraulic properties and heterogeneities of a stream and lake beds. Streambed clogging and stream partial penetration are factors which are equally important as heterogeneity. All these factors need to be considered during analytical treatments of groundwater-surface water interactions (Jones and Holmes, 1996). Moreover, the relative importance of streambed clogging, stream partial penetration and heterogeneity under semi-arid conditions needs to be further investigated. At the current state of research, most techniques and models developed for groundwater-surface water interactions were based on information from humid regions (Winter, 1995). There is therefore a need to revise such techniques and models utilizing both in-situ and remote sensing observations from semi-arid regions. These techniques also need to be coupled with Geographic Information Systems (GIS) technology and statistical analysis to study groundwater-surface water interactions in semi-arid regions in a multidisciplinary and multiscale approach. 9 CONCLUSIONS The realization that hydrological and ecological settings are inter-related has prompted the coining of the term “ecohydrology” to describe this inter-relationship (Wassen and Grootjans, 1996). Baird and Wilby (1999) provide several examples from a range of environments on how exchange between groundwater and surface water affects interface ecology, and how biological communities affect groundwater-surface water exchanges. Several studies investigating the advantages of the inter-relationship have also been reported in the literature. However, there are still many gaps in our understanding of the processes involved in groundwater-surface water interactions, and the environmental implications of the exchanges. The boundaries between hydrological and ecological research are gradually disappearing, yet a need remains for closer collaboration between these traditionally distinct disciplines, and among researchers working in different climatic regions so that research results may be pooled and applied to the benefit of the global environment, such as for the sustainable management and utilization of natural resources.

Water resources of arid areas

460

REFERENCES Baird, A.J. & Wilby, R.L. 1999. Eco-Hydrology: Plants and Water in Terrestrial and Aquatic Environments. New York, Routledge Press. Baxter, C.V. & Hauer, F.R. 2000. Geomorphology, hyporheic exchange and selection of spawning habitat by bull trout (Salvelinus confluentus). Canadian J. Fisheries and Aquatic Sciences 57:1470–1481. Beven, K.J. 1989. Interflow. In: H.J.Morel-Seytoux (ed) Unsaturated flow in hydrologic modeling: theory and practice, Kluwer, Dordrecht: 191–219. Beven, K.J. & Germann, P.F. 1982. Macropores and water flow in soils. Water Resour Res, 18:1311–1325. Brunke, M. & Gonser, T. 1997. The ecological significance of exchange processes between rivers and groundwater. Freshwater Biol. 37:1–33. Carrere, R. 1996. Pulping the South: Brazil’s pulp and paper plantations. Ecologist, 26:206–214. Dahm, C.N. & Valett, H.M. 1996. Hyperheic zones, In: F.R. Hauer & G.A. Lamberti. (eds). Methods in Stream Ecology. San Diego, California, Academic Press: 107–119. Dunne, T. & Black, R. (1970). An experimental investigation of runoff production in permeable soils. Water Resour Res 6:478–490. Gilbert, J., Danielpool, D. & Stanford, J.A. 1994. Groundwater Ecology, San Diego, California, Academic Press. Gordon, N.B., McMahon, T.A. & Finlayson, B.L. 1992. Stream Hydrology: An Introduction for Ecologists, Chichester, Wiley. Hewlett, J.D. & Hibbert, A.R. 1967. Factors affecting the response of small watersheds to precipitation in humid areas. In: W.E.Sopper and H.W.Lull (eds) Proc Int. Symp on Forest Hydrology, Oxford, Pergamon Press: 275–290. Hoehn, E. 1998. Solute exchange between river water and groundwater in headwater environments. In: Proc. Headwater ‘98 Conf Hydrology, Water Resources and Ecology in Headwaters, Meran/Merano, Italy, IAHS Publ 248, Wallingford, 165–171. Jones, J.B. & Holmes, R.M. 1996. Surface-subsurface interactions in stream ecosystems. Trends Ecol Evol, 16:239–242. Lamontagne, S., Herczeg, A.L., Dighton, J.C. & Pritchard, J.L. (2003). Groundwater-surface water interactions between streams and alluvial aquifers: Results from the Wollombi Brook (NSW) Study, Part II-Biogeochemical Processes. CSIRO Land and Water Technical Report 42/03. Lee, D.R. & Cherry, J.A. 1978. “A field exercise on groundwater flow using seepage meters and minipiezometers”. J. Geol. Educ., 27:6–10. Rushton, K.R. & Tomlinson, L.M. 1979. Possible mechanisms for leakage between aquifers and rivers. J. Hydrol 40:49–65. Sklash, M.G. & Farvolden, R.N. 1979. “The role of groundwater in storm runoff’. J. Hydrol, 43:45–65. Snodgrass, J.W., Komoroski, M.J., Bryan, A.L.J. & Burger, J. 2000. Relationships among isolated wetland size, hydroperiod, and amphibian species richness: Implications for wetland regulations. Conserv Biol. 14: 414–419. Wassen, M.J. & Grootjans, A.P. 1996. Ecohydrology: An interdisciplinary approach for wetland management and restoration. Vegetation, 126:1–4. Winter, T.C. 1995. Recent advances in understanding the interaction of groundwater and surface water. Rev Geophys (Suppl):985–994. Woessner, S.M. 2000. Stream and fluvial plain groundwater interactions: rescaling hydrogeologic thought. Groundwater, 38:423–429.

Impacts of climate change in water resources planning and management Alfred Opere Department of Meteorology, University of Nairobi, Kenya Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Freshwater resources are an essential component of the earth’s hydrosphere and an indispensable part of all terrestrial ecosystems. The freshwater environment is characterized by the hydrological cycle, including floods, and droughts, which in some regions have become more extreme and dramatic in their consequences. Global climatic change and atmospheric pollution could also have an impact on freshwater resources and their availability and, through sea-level rise, threaten low-lying coastal areas and small island ecosystems. Water is needed in all aspects of life. Innovate technologies, including the improvement of indigenous technologies, are needed to fully utilize limited water resources and to safeguard those resources against pollution, Rational water utilization schemes for the development of surface and underground water supply sources and other potential sources have to be supported by concurrent water conservation and wastage minimization measures.

1 INTRODUCTION Water is one of our most important natural resources. Without it, there would be no life on earth. The lifestyle we have become accustomed to depend heavily upon having plenty of cheap, clean water available as well as an inexpensive, safe way to dispose of it after use. The supply of water available for our use is limited by nature. Although there is plenty of water on earth, it is not always in the right place, at the right time and in the right quality. Adding to the problem is the increasing evidence that chemical wastes improperly discarded yesterday are showing up in our water supplies today. Freshwater is a crucial resource for sustainable development. Considering the current situation and the multifaceted dimensions of the water crisis, there is acute need for action—no time to waste:

Water resources of arid areas

462

1. More than 1 billion people worldwide lack access to safe drinking water and close to 3 billion are not provided with adequate sanitation. 2. Ecosystems are harmed as a consequence of pollution or of drying out. 3. People and nations are competing for scarce and finite water resources. Moreover the problems of variability of precipitation and climate, changing human settlement and land use changes as well as exploitation of natural resources caused a considerable increase of the number of catastrophes, such as floods and droughts, over the past 50 years. 2 HYDROLOGY AND WATER RESOURCES Of the 19 countries around the world currently classified as water-stressed, more are in Africa than in any other region, and this number is likely to increase, independent of climate change, as a result of increases in demand resulting from population growth, degradation of watersheds caused by land use change and siltation of river basins. A reduction in precipitation projected by some GCMs for the Sahel and southern Africa, if accompanied by high inter-annual variability, could be detrimental to the hydrological balance of the continent and disrupt various water-dependent socio-economic activities. Variable climatic conditions may render the management of water resources more difficult both within and between countries. A drop in water level in dams and rivers could adversely affect the quality of water by increasing the concentrations of sewage waste and industrial effluents, thereby increasing the potential for the outbreak of diseases and reducing the quality and quantity of fresh water available for domestic use. Adaptation options include water harvesting, management of water outflow from dams and more efficient water usage. Rainfall is the key input variable that activates flow and mass transport in hydrological systems, and models for simulating and forecasting rainfall in space and time can play an important role in enhancing our understanding of the hydrological system response, and in the design and operation of water resource systems. Today, we face record consumption, uncertain supplies, and growing demands for protection from flooding and pollution. The health and economic effects of a shortage of clean water are matters of great concern. Hydrology has evolved as a science in response to the need to understand the complex water systems of the earth and help solve water problems. Hydrologists play a vital role in finding solutions to water problems. 2.1 What is hydrology? Hydrology is the science that encompasses the occurrence, distribution, movement and properties of the waters of the earth and their relationship with the environment within each phase of the hydrologic cycle. The hydrologic cycle is a continuous process by which water is purified by evaporation and transported from the earth’s surface (including the oceans) to the atmosphere and back to the land and oceans. All of the physical, chemical and biological processes involving water as it travels its various paths in the atmosphere, over and beneath the earth’s surface and through growing plants, are of interest to those who study the hydrologic cycle.

Impacts of climate change in water resources planning and management

463

There are many pathways the water may take in its continuous cycle of falling as rainfall or snowfall and returning to the atmosphere. It may be captured for millions of years in polar ice caps. It may flow to rivers and finally to the sea. It may soak into the soil to be evaporated directly from the soil surface as it dries or be transpired by growing plants. It may percolate through the soil to groundwater reservoirs (aquifers) to be stored or it may flow to wells or springs or back to streams by seepage. The cycle for water may be short, or it may take millions of years. People tap the water cycle for their own uses. Water is diverted temporarily from one part of the cycle by pumping it from the ground or drawing it from a river or lake. It is used for a variety of activities such as households, businesses and industries; for irrigation of farms and parklands; and for production of electric power. After use, water is returned to another part of the cycle: perhaps discharged downstream or allowed to soak into the ground. Used water normally is lower in quality, even after treatment, which often poses a problem for downstream users. The hydrologist studies the fundamental transport processes to be able to describe the quantity and quality of water as it moves through the cycle (evaporation, precipitation, streamflow, infiltration, groundwater flow, and other components). The engineering hydrologist, or water resources engineer, is involved in the planning, analysis, design, construction and operation of projects for the control, utilization, and management of water resources. Water resources problems are also the concern of meteorologists, oceanographers, geologists, chemists, physicists, biologists, economists, political scientists, specialists in applied mathematics and computer science, and engineers in several fields. 2.2 What do hydrologists do? Hydrologists apply scientific knowledge and mathematical principles to solve waterrelated problems in society: problems of quantity, quality and availability. They may be concerned with finding water supplies for cities or irrigated farms, or controlling river flooding or soil erosion. Or, they may work in environmental protection: preventing or cleaning up pollution or locating sites for safe disposal of hazardous wastes. Persons trained in hydrology may have a wide variety of job titles. Some specialize in the study of water in just one part of the hydrologic cycle: hydrometeorologists (atmosphere); glaciologists (glaciers); geomorphologists (landforms); geochemists (groundwater quality); and hydrogeologists (groundwater). Engineers who study hydrology include those in agricultural, civil, environmental, hydraulic, irrigation and sanitary engineering. 3 SURFACE WATER RESOURCES Hydrologists help cities by collecting and analyzing the data needed to predict how much water is available from local supplies and whether it will be sufficient to meet the city’s projected future needs. To do this, hydrologists study records of climate such as rainfall,

Water resources of arid areas

464

snow-pack depths and river flows that are collected and compiled by hydrologists in various government agencies. Managing reservoirs can be quite complex, because they generally serve many purposes. Reservoirs increase the reliability of local water supplies. Deciding how much water to release and how much to store depends upon the time of year, flow predictions for the next several months, and the need of irrigators and cities as well as downstream water-users that rely on the reservoir. If the reservoir is also used for recreation or for generation of hydroelectric power, those requirements must be considered. Decisions must be coordinated with other reservoir managers along the river. Hydrologists collect the necessary information, enter it into a computer, and run computer models to predict the results under various operating strategies. On the basis of these studies, reservoir managers can make the best decision for those involved. The availability of surface water for swimming, drinking, industrial or other uses sometimes is restricted because of pollution. Pollution can be merely an unsightly and inconvenient nuisance, or it can be an invisible, but deadly, threat to the health of people, plants and animals. Hydrologists assist public health officials in monitoring public water supplies to ensure that health standards are met. When pollution is discovered, environmental engineers work with hydrologists in devising the necessary sampling program. Water quality in estuaries, streams, rivers and lakes must be monitored, and the health of fish, plants and wildlife along their stretches surveyed. Related problem concerns acid rain and its effects on aquatic life, and the behavior of toxic metals and organic chemicals in aquatic environments. Hydrologic and water quality mathematical models are developed and used by hydrologists for planning and management and predicting water quality effects of changed conditions. It would be difficult to think of any human activity or interest that is not in some way affected by weather and climate. And so, it should be a matter of considerable concern that, if we continue to emit carbon dioxide (CO2), methane (CH4), nitrous oxide (N2O) and other radiatively active trace gases (the “greenhouse gases”) at current or even significantly reduced rates, the lower portion of the atmosphere will grow warmer. As a result, patterns of cloudiness, precipitation, humidity, windiness and possibly the frequency and severity of extreme weather events to which we are now accustomed will be different—not necessarily or universally worse, but different. If this be so, then food and water supplies, the health of ecosystems and of humanity will be affected, maybe for the better, more likely for the worse. 4 GCM’S FOR CLIMATE PREDICTION On the basis of computer—intensive general circulation models (GCMs hereafter), the scientific community is reasonably confident that warming will be greatest in the high latitudes and most moderate in the tropics. Because the capacity of the atmosphere to hold water vapor increases exponentially with rising temperature, rates of evaporation will increase wherever there is water. More water in the atmosphere means that precipitation will also increase. Thus, warming will intensify the hydrologic cycle. But the geographic distribution of changes in precipitation

Impacts of climate change in water resources planning and management

465

is quite uncertain. Many of the GCMs indicate that some regions will be drier as a result of global warming. Some GCMs anticipate increased precipitation for other parts of the world, such as sub-Sahelian Africa. As a consequence, the chronic droughts that have long afflicted these regions could become a thing of the past. 5 IMPACTS OF CLIMATE What follows are some possible (but far less certain) impacts of impending climatic change. 5.1 Impact on sea levels If the world warms, sea levels will rise because of thermal expansion of the ocean waters and the melting of mountain glaciers. Half of humanity lives in the world’s coastal regions. Low-lying islands in the Pacific and elsewhere can lose much of their land area to the sea. There will be flooding caused by sea-water surges driven by cyclonic storms. With higher sea levels the surging water would obviously be deeper and penetrate farther inland. Rising sea levels could also cause permanent abandonment of large areas of agricultural land resulting into permanent food shortage. If sea level rises, the salt water would penetrate coastal aquifers. 5.2 Impact on water resources Climatic change can have major impacts on regional water resources throughout the world. Water for household and commercial use, irrigation, hydropower generation, power plant cooling, navigation, in-stream ecosystems and recreation will all be affected. This will be of particular concern to regions already under water stress, such as the arid and semi-arid regions, and those in which there is already considerable competition among users. Changes in natural flow in river basins in response to the climate change droughts means that (i) Much of the irrigation practiced would have to cease. (ii) Hydropower generation would be more difficult to sustain. (iii) Increased competition for water would threaten its use for wildlife habitat and recreation. Prospects of increased frequency and severity of floods and droughts raise difficult issues for water managers around the world. In arid regions, water supplies are managed largely through impoundments. In temperate regions, floods are controlled primarily by means of dams and dikes. Costly changes in both structures and procedures are almost certain. Arid and semi-arid regions are among those experiencing rapid population growth, increased urbanization, industrial development and, in some cases, agricultural expansion. The resultant increases in demand for water will be accompanied by uncertainty about water availability if these regions also experience decreases in runoff.

Water resources of arid areas

466

6 CONCLUSION There is uncertainty with respect to the prediction of climate change at the global level. Although the uncertainties increase greatly at the regional, national, and local levels, it is at the national level that the most important decisions would need to be made. Higher temperatures and decreased precipitation would lead to decreased water-supplies and increased water demands; they might cause deterioration in the quality of freshwater bodies, putting strains on the already fragile balance between supply and demand in many countries. Even where precipitation might increase, there is no guarantee that it would occur at the time of year when it could be used; in addition, there might be a likelihood of increased flooding. Any rise in sea level will often cause the intrusion of salt water into estuaries, small islands and coastal aquifers and the flooding of low-lying coastal areas; this puts low-lying countries at great risk. On a more optimistic note, technological innovations to increase water supply, by evaporation suppression, water reuse and recycling systems, satellite-guided and computer-controlled irrigation systems, can help mitigate some of the negative impacts of climate change. Water demand management through appropriate climate factoring and institutional adaptation will be the key to increasing flexibility of water resource systems in the face of climatic change. REFERENCES Houghton, J. 1997. Global Warming: The Complete Briefing. Cambridge, Cambridge University Press.

x Rosenberg, N.J. & Cooper, C.I. 1982. Likely impacts of a likely global warming (CC’82). Watson, R.T., Zinyowera, M.C., Moss, R.H. & Dokken, D.J. eds. 1997. 1PCC. Special Report on the Regional Impacts of Climate Change: An Assessment of Vulnerability. Summary for policymakers. White, R.M. 1998. “Kyoto and Beyond,” Issues in Science and Technology, 1998. Wuebbles, D.J. & Rosenberg, N.J. 1998. “The Natural Science of Global Climate Change.” Chapter 1 in S Rayner and EL Malone, eds. Human Choice and Climate Change, Vol 2. Resources and Technology Columbus, Battelle Press, OH, 1–143.

Turning a liability into an asset: the case for South African coalmine waters B.H.Usher & F.D.I.Hodgson Institute for Groundwater Studies, University of the Free State, Bloemfontein, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: South Africa is a water-poor country. With increased industrialisation and population growth, the demands on this resource are increasing. This has resulted in costly inter-basin transfer schemes to supplement the supply to the northern and central provinces. South Africa is the fourth largest producer of coal in the world and the 224 million metric tons of coal produced per year directly supports employment for approximately 50000 employees. Unfortunately associated with mining several water-related problems, largely associated with water quality deterioration due to pyrite oxidation, occur. With new regulations pertaining to waste water discharge costs and the potential costs of “raw” water, the use of mine water for various purposes, and the use of mine voids as storage areas is becoming more feasible. Estimated post-closure water make from the Mpumalanga coafields is estimated to be in the order of 360Ml/day (Grobbelaar et al., 2001). An estimated 3100Mm3 of coal will have been removed via underground mining methods by mine closure. The voids created in this way will all eventually fill up with water.The challenge lies in finding ways to utilise these vast volumes of mine water or the voids as alternative storage. The water quality will determine the suitability for future use and management options need to be implemented to ensure optimal future qualities. Options such as appropriate mine planning and enhanced recharge to limit pyrite oxidation have been investigated. Changing the current paradigm to allow harvesting of flood waters to inundate workings and utilise storage will be of great benefit to the country. Research has shown that several of these mine waters are suitable for irrigation of crops such as maise, and current research is geared to determining the effect on the associated groundwater resources.

Water resources of arid areas

468

1 INTRODUCTION South Africa is a water-poor country. With increased industrialisation and population growth, the demands on this resource are increasing. This has resulted in costly interbasin transfer schemes to supplement the supply to the northern and central provinces. South Africa is the fourth largest producer of coal in the world and the 224 million metric tons of coal produced per year directly supports employment for approximately 50000 employees. Unfortunately associated with mining several water-related problems, largely associated with water quality deterioration due to pyrite oxidation, occur. The coal mining industry has been involved with catchment management since the late 1980’s (Salmon, 2003). 1.1 Overview of the Mpumalanga coalfields 1.1.1 Topography and surface drainage Surface drainage occurs to the north, east and south. The topography is of a gentle rolling nature. Steeper slopes are present at sandstone outcrops. In terms of this study, the main concern is the proximity of the Olifants River and Witbank Dam to the mine workings. In the event of mine water spilling into the river, this could have a significant impact on the dam water quality, particularly in the dry season. 1.1.2 Stream water quality Coal mining has been ongoing in the Olifants Catchment for many decades. Very few of the mines have, however, filled up with water to decant freely into surface streams. The stream water quality has nevertheless deteriorated over the past 20 years, due to discharge and seepage of mine water. 1.1.3 Mining methods and extent Mining in the Witbank region has been extensive. The depth of mining ranges from less than 10m below surface to more than 100m. The coal seams generally increase in depth to the south. Mining methods are bord-and-pillar, stooping and opencast. Opencast mining has been introduced during the late seventies. Underground mining on the 2 seam comprises in excess of 100000ha while opencast mining is expected to eventually exceed 40000ha (Grobbelaar et al., 2002). Coal extraction has been ongoing in the Mpumalanga Coalfields for more than 100 years. (Havenga, 2003). Coal is generally mined by opencast- or underground methods in South Africa. (Grobbelaar, 2001). At depths down to 50 metres coal is normally extracted by surface mining, the extraction rate associate with type of mining being currently 85– 90 percent (Hodgson et al., 1995). At depths below 50 metres more conventional mining methods such as bord-and-pillar extraction have been used sine mining began.

Turning a liability into an asset

469

1.1.4 Geology The Karoo Supergroup in the Witbank region comprises the Ecca Group and Dwyka Formation. The total thickness of these sediments ranges from 0–100m. The Ecca sediments consist predominantly of sandstone, siltstone, shale and coal. Combinations of these rock types are often found in the form of interbedded siltstone, mudstone and coarse-grained sandstone. Typically, coarse-grained sandstones are a characteristic of the sediments in the Witbank Area. Of all the unweathered sediments in the Ecca, the coal seams often have the highest hydraulic conductivity. Packer testing of the No. 2 Seam and underlying Dwyka tillite has a hydraulic conductivity. Five coal seams, numbered from bottom to top as No. 1–5, are present. Only two of the seams are mineable over most of the area. These are the No. 2 and 4 Seams, which are usually separated by sediments of a total thickness in the order of 20–30 m. Seams 1 and 5 are, however, mined locally. Dolerite intrusions in the form of dykes and sills are present within the Ecca Group. Faults are rare. However, fractures are common in competent rocks such as sandstone and coal. 2 WATER IN SOUTH AFRICAN COAL MINES 2.1 Sources of water Several sources of water influx are expected in South African Collieries. In opencast areas, much of the influx is dependent on the state of the post-mining rehabilitation, while in underground mining factors such as the mining type, depth and degree of collapse and interconnectivity are important. Figure 1 illustrates the generalised hydrological conditions associated with an opencast environment. Normal groundwater movement still takes place in aquifers. Groundwater flow directions will necessarily be directed toward the pits, due to an artificial change in gradients on a local scale and a higher K-value in spoils (Grobbelaar, 2001). This flow, together with direct recharge into the spoils will create an artificial groundwater level in the heaped spoil until a decant level is reached. Water that decants out of the spoils as well as run-off from the surface of the spoils will follow the natural gradient and will flow to the nearest river or stream.

Water resources of arid areas

470

Figure 1. General geohydrology of opencast pits (Grobbelaar et al., 2001). Table 1. Water recharge characteristics for opencast mining. (Hodgson and Krantz, 1995). Sources which contribute water

Water sources Suggested into opencast pits average values

Rain onto ramps and 20–100% of voids rainfall Rain onto unrehabilitated 30–80% of rainfall spoils (run-off/seepage) Rain onto levelled spoils 3–7% of rainfall (run-off) Rain onto levelled spoils 15–30% of rainfall (seepage) Rain onto rehabilitated 5–15% of rainfall spoils (run-off) Rain onto rehabilitated 5–10% of rainfall spoils (seepage) Surface run-off from 5–15% of total pit surroundings water Groundwater seepage 2–15% of total pit water

70% of rainfall 60% of rainfall

5% of rainfall 20% of rainfall 10% of rainfall 8% of rainfall 6% of total pit water 10% of total pit water

Turning a liability into an asset

471

In terms of expected sources of water, the table above (Table 1) from Hodgson and Krantz summarises, the most important information. The high recharge percentage of around 20% is due to a multitude of factors such as ponding, areas of spoils exposure, restabilisation cracks and influx into ramp areas. Several researchers have confirmed these high recharges through decant measurements (Hodgson, 1999, van Tonder et al., 2003). The high influx is naturally an important driver on the observed water quality in opencast pits. In underground mines the following sources of water could be encountered: – Water encountered in the seam as mining commences. This is fairly low except where fractures or fissures as they are known within the mining industry are encountered. – Recharge through the roof lithologies. The magnitude of this varies depending on mining induced fracturing of the overlying sediments. – Direct recharge where cracks from the collapse of mining areas, usually due to high extraction mining, run through to the higher-yielding transmissive aquifers nearer surface. – Regional groundwater flow, which will usually flow along the coal horizon, due to its higher hydraulic conductivity compared to the surrounding sediments. – Influx through the floor lithologies. This can play an important role in areas where the floor is transmissive, but where the mining floor is close to the Dwyka such as where the No. 1 or No. 2 seam are mined, such influxes are negligible.

Table 2. Key mining and storage statistics for Mpumalanga coalfields. Future volume Future Current area Total Volume coal to Total mined to be area to Extraction volume Storage coal be volume height Mining area mined be Extraction mined mined to be type (ha) (ha) mined (m) rate (Mm3) (Mm3) removed (Mm3) U/G 5 7842 7050 14892 2.5 seam U/G 4 13485 2833 16318 3 seam U/G 2 98550 39695 138245 3 seam U/G 1 2525 0 2525 2.5 seam Opencast 13557 14480 28037 3.5 Totals 135959 64058 200017

0.65

127

115

242

242

0.65

263

55

318

255

0.65

1922

774

2696

1887

0.6

38

0

38

30

0.9

427 2777

456 1400

883 4177

221 2635

The water encountered, as mining continues, will subside over time. In bord-and-pillar areas where mining has been completed, features such as roof bolts, drilled to stabilise the roof during mining, act as local sinks for water to drain to. Where such a roof bolt intersects horizontal or vertical fractures, increased influx is experienced. This is particularly problematic in areas adjacent to water storage compartments, where seals are

Water resources of arid areas

472

installed to accommodate a head in excess of the mining height. As the water level rises, intersection of the naturally occurring bedding plane fractures occurs. These act as more transmissive conduits, which allow the water to flow more freely toward the locally created sink. 2.2 Potential storage volumes Based on the mining areas, expected volumes of coal removed can be determined. Using the coal extraction ratios, the determined storage factors such as 25% for opencast mines and the storage factors determined for collapse in high extraction mining, an indication of storage in the Mpumalanga coalfields can be obtained. The expected storage volume is therefore approximately 20 times greater than that of the Witbank Dam. This stored mine water is a resource that cannot be ignored in this country. 2.3 Water quality impacts Associated with coal mining in South Africa, the phenomenon of Acid Mine Drainage (AMD) occurs. Acid mine drainage occurs when sulphide minerals in rock are oxidised, usually as a result of exposure to moisture and oxygen. This results in the generation of sulphates, metals and acidity that can have manifold environmental consequences. Pyrite (FeS2) an iron disulphide (commonly known as fool’s gold), is the most important sulphide found in South African coalmines. When exposed to water and oxygen, it can react to\ form sulphuric acid (H2SO4). The following oxidation and reduction reactions the pyrite oxidation that leads to acid mine drainage. (1) (2) Fe2++1/4 O2+H+=>Fe3++1/2 H2O (rate limiting step) (3) Fe3++3H2O =>Fe(OH)3 (yellow boy)+3H+ (4) In the coalfields there are co-existing carbonates such as calcite and dolomite, which can neutralise the acidity generated. The reaction with calcite is given by:

The results of this are that many of the mine waters are not necessarily acidic, but often high in dissolved salts. Additionally, the sediments overlying the mines, can be fairly saline, and particularly in the southern portion of the coalfields, are high in sodium.

Turning a liability into an asset

473

Table 3. Key processes in different mining types that affect water quality. Opencast mining Bord-and-pillar High extraction mining mining Recharge 15–25% 3–8% 5–15% Dilution can occur Build up of Rapid inundation oxidation products with some dilution; also often an additional source of alkalinity Relative High Low Medium surface area S-value 25% In line with 20% over the extraction—50– collapsed height 65% Flooding can occur Large amount of Smaller volume of to prevent excessive water-high water distributed salt loads water:rock ratio over greater height Coal 90% 50–65% 90% removal If coal seams are If coal seams are If coal seams are the the most likely to the most likely to most likely to acidify acidify it reduces acidify it has the it reduces the risk the risk greatest probability

2.4 Summary of influx and effect on water quality The different mining types each have conditions, which promote poor quality water, and others that should ameliorate the effects of pyrite oxidation. In all of the mining types, however, the local mineralogical conditions provide the most important driver on water quality. Trenches dug in the spoils clearly indicate the validity of this observation. In underground mining, this is no less important. A case study at an underground compartment shows that the roof sediments have a higher probability of acidification than the coal seam itself, implying that the high extraction areas pose the greatest risk. The balance between the increased recharge, with addition of alkalinity and faster inundation, against the higher surface area and lower net neutralising potential will determine what the final water quality will be. Table 3 summarises some of the key issues from a hydrogeochemical perspective (Usher, 2003). 2.5 Intermine flow and water volumes Previous researchers (for example, Grobbelaar, 2001), have investigated intermine flow. After the closure of mines, water in the mined-out areas will flow along the coal seam floor and accumulate in the lower-lying areas. These man-made voids will fill up with water and hydraulic gradients will be exerted onto peripheral areas (barriers) or

Water resources of arid areas

474

compartments within mines. This results in water flow between mines, or onto the surface. This flow is referred to as intermine flow (Grobbelaar, 2001). Projections for future volumes of water to decant from the mines have been made by Grobbelaar et al. (2000). In total, about 360ML/d will decant from all the mines in combination. On a catchment basis, it relates to the following (ML/d): Wilge/Klip Olifants Klein Vaal Komati Olifants 23

170

45

120 2

3 REGULATIONS AND EXTERNAL PRESSURES 3.1 New Water Act of 1998 and other legislation With the proclamation of The National Water Act, Act 36 of 1998, a new responsibility towards groundwater and groundwater management developed. The establishment of eventual Catchment Management Agencies and other localised water management forums, have all led to a more intensive focus on mine water management. Provisions in the new legislation to allow for polluters to pay for their pollution costs and the shift in ownership of water have all lead to mines being forced to find ways of reusing water, improving final water qualities and minimising potential future liability. 3.2 Waste discharge charges The imminent introduction of waste discharge costs is an important driver to ensuring water quality aspects are adequately dealt with. The polluter-pays principle, which uses the principle of internalising costs to the costs of pollution, is the basis for the implementation of these charges. This principle specifically supports payment for the costs of pollution, which includes the direct economic cost borne by downstream activities impacted by pollution, the environmental cost borne by downstream activities impacted by pollution and the cost of treating waste (DWAF, 2003). Discharges are measured in waste load, where waste load is defined as: Li=Ci×Q where Li is the waste load for pollutant (i), measured in kg Ci is the concentration of pollutant (i) in the effluent, measured in mg/l Q is the volume of water, measured in m3 The greater the load that is released, the greater the charges to be paid by the mine. The implementation of these costs is under review but a multi-tiered approach is likely. Three tiers are costs are proposed namely a Basic/administrative charge (Tier 1), Loadbased charge (Tier 2) and Deterrent charges (Tier 3 and 4). Providing incentives for pollution prevention, rather than mitigation (end-of-pipe solutions), could be achieved through revenue disbursements.

Turning a liability into an asset

475

3.3 Interbasin transfer The coalmines are in closer proximity to the major industrial hub of South Africa, Gauteng, than the other areas being target for interbasin transfer. In 2000, 1310Mm3 of water was transferred to the Upper Vaal catchment. Based on the expected post-closure water make of 360ML/d or 130Mm3/year if the water quality in these mines can be optimised in the long-term, the excess water in the catchment will alleviate the pressure on the Upper Vaal catchment for water. Using the financial incentives built into the waste discharge charges, if the coalmines as an entity can deal with water quality, this will be an invaluable resource for the country. 4 WATER MANAGEMENT TO MAXIMISE THE RESOURCE 4.1 Alternative water uses 4.1.1 Reuse and recycling Large-scale resuse and recycling of mine water is currently used by the coal mining industry. In this way mine water is utilised in processes, for dust suppression and for general washing of plant areas. Mine water is also used to supply cooling water and other water requirements to several power stations. In many cases, this water is treated prior to use. 4.1.2 Mine water irrigation Irrigation provides for a novel approach to the utilisation and disposal of gypsum-rich mine water. Research at scales ranging from experimental to semi-commercial over a period of more than 10 years has demonstrated the potential to successfully use this water for the irrigation of a range of crops and significantly reduce the salt load emanating from mine drainage, by precipitating gypsum within the soil. An evaluation of the feasibility and sustainability of this practice from agricultural and environmental perspectives is the subject of an on-going WRC project at five irrigation pivots (Annandale et al., 2003). Large volumes of water in underground and opencast collieries are currently not being utilised. If irrigation with mine water is proven to be sustainable, excess mine water can be regarded as a national asset, rather than a liability. After collieries close down, it can provide economic advantages for retrenched workers, and also create opportunities for small farmers. Minnaar Colliery is an underground mine located in the Mpumalanga Province between the towns of Ogies and Witbank, where mine water irrigation is actively practiced (Vermeulen, 2003). Two boreholes are located in a surface dam. During the rainy season, water from the dam drains through the boreholes into the mine. The mine water is used by the farmer for irrigation. The surface area above the mine is used for agriculture, with both summer- and winter crops being harvested. Two 40ha center pivots have been in operation for nine years, utilising water from dam, supplemented by mine

Water resources of arid areas

476

water. An additional 40ha center pivot has been in operation since 2002, extracting water from another section of the mine. According to the farmer, he applied approximately 35mm per week from the mine onto the 40ha crop of maise. The ideal mine-water management option is already in place for Minnaar Colliery. Excess water is utilised for irrigation and the mine water is simultaneously flushed. Care should, however, be exercised not to over utilise the source, as a permanent drop in the water level can result in an influx of oxygen and subsequent oxidisation of pyrite (Vermeulen, 2003). 4.2 Management options 4.2.1 Regional management options Several major paleo-drainage channels are present on the No. 2 Coal Seam floor. If no barriers existed, much of the water on this coal seam horison would drain towards the paleo-channels, where it would dam up. By providing for controlled release systems through barriers, much of the mine water could therefore be drained naturally towards centrally located areas, from where abstraction and treatment can be done in a controlled fashion. An excellent example of such an area where this will be possible, would be in the central Witbank Coalfield, where water from Tavistock, Rietspruit, Douglas, New Clydesdale, TNC and SACE mines could centrally be controlled and, if required, treated before flood discharged. This would typically involve some 100ML/d. It is obvious that significant advances in water-quality management can be made by planning the possible transfer of salt to centralised treatment facilities. Here acid water could be blended with alkaline water, for instance. Settlement of iron from the mine water could be allowed, before release of the water during flood events. 4.2.2 Localised water management options 4.2.2.1 Containment of mine water for flooding and neutralisation purposes Mine water has historically been pumped from active workings to allow unhindered coal production. Almost no consideration has been given to the best management strategy for water while mining. Yet, this is simple: Mine from deep to shallow and leave water behind in the mined-out workings. This strategy has, for the past few years, been applied in several of the larger collieries with significant success. The advantage of this mining sequence does not only lie in managing water volumes, but also in water quality management. Mined-out areas are flooded, thus excluding oxygen. Furthermore, the natural alkalinity of the water is not flushed from the rock. This counteracts acidification. 4.2.2.2 Mixing of mine water At most of the larger mines, the opportunity exists for mine water of different qualities to be mixed, thus improving the overall water quality. Typical benefits of doing this would lie in pH adjustment and iron precipitation. For the latter, retention of the mine water in a

Turning a liability into an asset

477

surface-holding facility where aeration is possible, is necessary. Such a facility could also be used for quick release of the water during flood discharge. Very few other chemical benefits would be forthcoming from mine water mixing, because most of the constituents are undersaturated in this water. 4.2.2.3 Minimising salt loads Several options for minimising the salt loads from mines need to be exploited for future use. These typically are: (1) Flooding of mines as soon as possible after closure; (2) Active flushing of flooded mines and (3) Greater utilisation of the natural base potential in the coal and rock for acid water neutralisation. These are all concepts with great potential, but will necessitate a change in the direction of thinking by the controlling authorities. Active flushing would imply the controlled release of salt into a catchment, with the specific purpose of improving the mine water quality to the extent that it would become a useable resource. Considering that many of the mines will have holding capacities far exceeding that of the Witbank Dam, this is a management option that, at the very least, should be pursued in a couple of the mines, thus establishing a number of field trials. 5 DISCUSSION In a water-stressed country like South Africa, all water must be regarded as a potential resource. The proximity of the Mpumalanga coalfields to major water users such as power stations, industries and the Gauteng province which is the most highly populated and industrial area of the country, means that this water is a prime resource. Water can only be used if the quality thereof is fit for its intended use. The water management focus at the coalmines needs to change from a volume-driven focus to a quality driven focus. If water qualities can be dealt with in a way that an array of qualities for different purposes can be provided, the cost-savings or rebates from the waste discharge costs will be such that the current waste becomes a prime asset for the mine and the country. REFERENCES Department of Water Affairs and Forestry. 2000. Blesbokspruit Catchment—Geohydrological Report for Acid Mine Drainage Collection and Conveyance System for Abandoned Mines, WQM/01/00. Department of Water Affairs and Forestry, 2003. Water Quality Management Series. Sub-Series No. MS11. Towards a Strategy for a Waste Discharge Charge System. 1st ed, Pretoria. Grobbelaar, R. 2001. The long-term impact of intermine flow from collieries in the Mpumalanga Coalfields. Unpublished M. Sc Thesis. Univ. of the Free State. Grobbelaar, R., Usher, B.H., Cruywagen, L-M., de Necker E. & Hodgson F.D.I. 2002. The Longterm Impact of Intermine Flow from Collieries in the Mpumalanga Coalfields, Water Research Commission Report.

Water resources of arid areas

478

Hodgson, F.D.I. 1999. Intermine flow between Tavistock and Rietspruit Collieries. Unpublished report. Hodgson, F.D.I. & Krantz, R.M. 1998. Groundwater quality deterioration in the Olifants River Catchment above the Loskop Dam with Specialised investigations in the Witbank Dam SubCatchment. Report to the Water Research Commission by the Institute for Groundwater Studies, Univ. of the Orange Free State. Hough, J.J.H. 2003. Evaluation of Management Options for Intermine Flow and associated Impacts in the Central Witbank Coalfield. Unpublished M.Sc thesis. Univ. of the Free State. Stumm, W. & Morgan, J.J. 1970. Aquatic Chemistry, 2nd Ed. New York, John Wiley & Sons, Inc., New York. Usher, B.H., 2003. Development and Evaluation of Hydrogeochemical Prediction Techniques for South African Coalmines. Unpublished Ph.D thesis, Univ. of the Free State. Van Niekerk, A.M. 2001. Innovative And New Mine Water Treatment Technologies. Proc. Conf. on Environmentally Responsible Mining in Southern Africa, Sept 2001. Johannesburg, South Africa. van Tonder, G.J., Vermeulen, P.D., Kleynhans, J. & Cogho, V. 2003. Prediction Of The Decant Rate And Sulphate Concentration From Rehabilitated Open Cast Coal Mines In South Africa. Paper submitted to the 6th International Conference of Acid Mining Drainage. In press. Vermeulen, P.D. 2003. Investigation of decant water from the underground collieries in Mpumalanga. Unpublished M.Sc thesis, Univ. of the Free State.

Environmental hydrogeology of the dolomite aquifer in Ramotswa, Botswana Michael Staudt Geological Survey of Finland, Espoo Horst Vogel Department of Geological Survey, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Since its independence from Britain, Botswana has experienced enormous economic growth. But Botswana also has its problems. The relationship between accelerated economic growth, a growing population, and environmental limitations such as adverse climatic conditions, present a formidable challenge for ground water management. Hence, identifying those areas where groundwater has been impacted upon by human activities can help planners address Botswana’s looming water-management crisis by focusing on monitoring important but vulnerable areas.

1 INTRODUCTION This study was carried out as part of a technical co-operation project between the Botswana Department of Geological Survey (DGS) and the German Federal Institute for Geosciences and Natural Resources (BGR). The study objective was to establish the quality of the groundwater resources in Ramotswa and to produce environmental hydrogeology maps for regional and urban planners. One vulnerable area is the village of Ramotswa, which experienced one of Botswana’s worst cases of groundwater pollution during the 1990s. The successful promotion of pit latrines and the location of Ramotswa on top of Botswana’s most productive dolomite aquifer meant disaster as human wastewater polluted the shallow aquifer in no time at all. As a result, the entire wellfield had to be abandoned in favour of surface water from the dam in the nearby capital city of Gaborone. In 2001, the Department of Geological Survey (DGS) in Lobatse decided to carry out an environmental hydrogeology investigation in Ramotswa because the aquifer there is possibly the most productive in a country where water resources are scarce. The study results revealed that the most obnoxious groundwater pollution problem was still due to high nitrate concentrations in several boreholes. The recorded maximum

Water resources of arid areas

480

nitrate concentration was 442mg L−1. The latter is attributed to the continued use of pit latrines. The industrial complex in the Ramotswa station area turned out to constitute another area of concern. High nitrate levels and elevated concentrations of sulfate, sodium, and chloride were identified. This was indicative of anthropogenic pollution, as were the observed nickel and possibly the aluminium levels. The study area includes the settlements of Ramotswa, Ramotswa station (Taung), Boatle, and surroundings. The size of the study area was approximately 174km2. The dominant Ngotwane river valley follows mainly N-S trending structures through the study area. River infiltration into the aquifers occurs beneath the Ngotwane river bed, which is approximately 2m thick. The overall hydraulic gradient is 1:300 along the course of the river. Other rivers are the Taung and the Boatle, which are tributaries to the Ngotwane river. All rivers are ephemeral and belong to the Limpopo river basin. The Ngotwane river forms the international boundary with the Republic of South Africa to the east. The rainfall pattern is seasonal, as is generally the case in Botswana. Mean annual rainfall amounts to approximately 475mm. In good years annual rainfall may exceed 1000mm, while in poor years it may be as low as 125mm. Rains fall mainly in short, high intensity events. Occasionally, heavy rainfalls provide the bulk of the annual precipitation. Rainfall is the primary source, which replenishes water resources in the study area. The by far most important form of landuse is livestock grazing. Although the recommended stocking rates are 3 to 6 livestock units per hectar, the present stocking rate in the study area is much higher, around 12 to 14 livestock units per hectar. This results in overgrazing and environmental degradation. 2 HYDROGEOLOGY There are two primary aquifer systems in the study area, namely the Ramotswa dolomite formation, which underlies most of the major village of Ramotswa, and the Lephala formation to the south and south-east. The latter consists of clastic cherts and conglomerates. In between the two major formations are three lesser formations, namely the Maholobota formation, which consists of interlayered dolomite with minor chert, the Magopane formation, which is made up of bedded chert and minor dolomite, and the Ramotswa shale formation, which comprises siltstone and shales. All five formations belong to the Transvaal supergroup. All these aquifers are considered to be in local hydraulic connection via predominately N-S trending fracture zones. The dominant feature of the system is a marked anisotropy associated with high density fracturing. In the dolomites, the active groundwater circulation has favoured local karstification along structural lineaments producing high transmissivity (T) and storativity (S). The Ramotswa dolomite aquifer consists of two different karst zones, a shallow and a deep zone (Institute of Hydrology, 1986). The upper karstic zone has a variable thickness of 20 to 50m and receives recharge from the river and percolating rainwater. Dolomite

Environmental hydrogeology of the dolomite aquifer in Ramotswa, Botswana

481

solution appears preferentially along fractures. The deeper karstic zone has a thickness of beween 25 to 50m and recharge is probably from across the border in South Africa. Within the dolomite aquifer, areas of major linear karst and unfractured dolomite country rock have to be distinguished. While high yielding wells are commonly found along the major linear karst, low yielding boreholes are common in the country rock. During field work the boreholes in this latter zone rapidly depleted at low pumping rates. Away from the river, dry boreholes were common. Hence it came as no surprise that the main production wells are located in E-W direction along the major linear karst in the southern part of the study area. The linear karst there features a dense fracture pattern with different fracture directions. This fracturing, in combination with infiltration of river water and the intersection of minor side valleys in E-W direction, has produced favourable conditions in terms of permeability. Borehole logs perfectly highlighted the 3 subdivisions described above, namely the upper aquifer extending to a depth of about 30m followed by a zone of less water-bearing fissures between 30–50m depth, and a deeper aquifer of variable thickness between 45 and 100m depth. The Lephala formation aquifer outcrops in the southern edge of the project area, where the rocks are faulted against the dolomites. The Lephala aquifer is similar to the Ramotswa dolomite aquifer but is unaffected by karstification. The formation is characterized by two fissured zones which are separated by a less fisured zone, a thickness of the upper zone of 30 to 40m, and a thickness of the lower zone of approximately 30m. Yields of the boreholes in the Lephala formation depend on their proximity to the river, the intersection of the fissured zones, and the extent of the secondary infills. Recharge may be restricted to surface runoff via fractures or infiltration from the river. 3 METHODS AND MATERIALS At the onset of the study, a well census was carried out. The aim was to identify all boreholes in Ramotswa and the surrounding areas. Once located, the site of the boreholes was determined by means of a hand-held GPS 12XL (http://www.garmin.com/). While carrying out the well census, the manually produced preliminary borehole location map was transferred into the ESRI ArcView GIS software environment (http://www.esri.com/). Similarly, all borehole parameters such as construction details, lithologies, water levels, water yields, water strikes, usage, coordinates, and owner data, were entered into the GeODin digital data base (http://www.geodin-system.com/). In the second phase all environmental hazards to groundwater were identified in the field. They were mainly anthropogenic and were differentiated as point hazards, such as, for example, pit latrines, and area hazards, such as, for example, cemeteries. Again, the location of all hazards was determined by GPS so as to produce a digital environmental hazard map. During the third stage of the field survey, groundwater from all accessible boreholes was sampled for chemical analyses. All in all, a total of 31 boreholes were sampled, 21 of which were also sampled for trace elements. It was most unfortunate that a number of

Water resources of arid areas

482

extra boreholes could not be sampled either because of vandalism or inaccessibility due to plentiful rains. Prior to sampling, the groundwater level was determined with the help of a dipper. At the same time the height of the casing and the diameter of the borehole were recorded. Before the pump was lowered into the borehole, the borehole log (whenever available) was checked in order to determine the most appropriate depth level to place the pump. Unfortunately, archive records were often incomplete or nonexistent. Because in such cases one could not know the depth of the water strike, the pump was installed half way down of the overall borehole depth. Most of the boreholes were not equipped with a pump. In these cases a submersible pump (Grundfos MP1) was employed, which was driven by a mobile power generator. To reach the desired depth, 3-m-long PVC tubes were used and put together. For security reasons the pump was never placed at very shallow depth; but the drawdown still had to be observed regularly. Based on the depth of the well and its casing diameter, the volume of the water column was calculated and a minimum pumping time calculated based on pumping yield [L/s]. Once pumping began, a series of field parameters [electrical conductivity (EC), water reaction (pH), water temperature (degrees Celsius), dissolved oxygen concentration (O2), and total dissolved solids (TDS)] were measured employing field electrodes in a flow cell (http://www.wtw.com/). Parameter values were checked every 10 minutes, and from time to time the flow cell was cleaned. As soon as the readings had stabilized for a period of at least 15 minutes, groundwater samples were taken and filled in 250-ml PVC bottles. The samples earmarked for the analysis of the trace elements were filtered through 0.45µm filters and acidified. All samples were immediately put into a cooler box and later stored in a fridge. After the groundwater sampling campaign had been completed, a GPS precision survey was carried out so as to determine the precise elevations of the sampled wells. But because of the high costs involved only 17 sites were surveyed. The elevations of the remaining boreholes were extracted from Geotechnical Consulting Services (2000). Since the difference between the two data sets was only ±0.3m, the accuracy level needed to obtain groundwater flow directions was adequate. 4 RESULTS As expected, magnesium (Mg2+), calcium (Ca2+), and bicarbonate were the most important ions (Fig. 1). Hence, the dominant groundwater type was of Mg-HCO3 and Mg-Ca-HCO3 facies. Only the boreholes in the Ramotswa station industrial area were mainly of the Cl-Na-HCO3 water type. The concentration of bicarbonate ranged from 234 (minimum) to 734mg L−1 (maximum). As expected, the dolomites showed very high levels of HCO3, while the Lephala formation had smaller concentrations.

Environmental hydrogeology of the dolomite aquifer in Ramotswa, Botswana

483

Figure 1. Groundwater types in Ramotswa (Piper diagram). All samples had calcium concentrations, which were more or less ideal (80mgL−1) to acceptable (150mg L−1) according to the Botswana standard for drinking water (BOS, 2000). In the case of magnesium, seven boreholes in Ramotswa village and in the Ramotswa station industrial area exceeded the maximum allowable limit of 100mgL−1 (BOS, 2000). The water reaction of the examined groundwater samples was in the range of pH 6.4 (minimum) to 7.7 (maximum). All values above 7 are indicative of the Ramotswa dolomite aquifer, while the lower values reveal mostly waters from the Lephala aquifer, or else, water from transition zones. The dissolved oxygen contents varied from 0.0 to up to 4.4mg L−1, with the vast bulk of the samples showing low levels of smaller than 2.0mg L−1. Such low oxygen contents are typical for anaerobic (anoxic) groundwater conditions. Electrical conductivity varied from a minimum value of 258 to up to 6070µS/cm. Four boreholes exceeded the maximum allowable limit of 3100µS/cm (BOS, 2000). Such high values, as encountered within the Ramotswa station industrial area, suggested pollution. Similarly, three wells exceeded the maximum allowable limit of 2000mg L−1 of total dissolved solids (TDS). concentrations were mostly below the acceptable limit The observed sulfate of 250mg L−1(BOS, 2000). But three boreholes at Ramotswa station had elevated levels

Water resources of arid areas

484

that were above the maximum allowable limit of 400mg L−1. Again, this was indicative of pollution as were the much higher than acceptable (200mg L−1) chloride [Cl¯] concentration within the Ramotswa station area, which again suggested anthropogenic pollution. Sodium (Na+) did not show elevated levels in the major village. But again three boreholes in the Ramotswa station area revealed high levels. In all three boreholes the levels were higher than the maximum allowable level of 400mg L−1(BOS, 2000). Four boreholes revealed aluminium (Al3+) levels above the maximum allowable level of 200µg L−1 (BOS, 2000).

Figure 2. Nitrate concentrations (mg/L) in Ramotswa in October/November 2001. Two boreholes in the Ramotswa station area had levels that were even twice as high as the maximum allowable level. Three boreholes close to the railway line in the Ramotswa station area also exceeded the maximum allowable limit for nickel (Ni2+) of 20µg L−1 (BOS, 2000). A few wells in the study area also showed elevated concentrations of iron (Fe2+), manganese (Mn2+), and arsenic (As3+, As5+), which were probably of geological origin. Iron concentrations greatly exceeded the maximum allowable level of 2000µg L−1 (BOS, 2000) in four wells. Most likely these high levels were due to ferruginous rocks, which are widely represented in the study area. In addition, they may also have been attributable to anoxic (reducing) groundwater conditions as was indicated by the smell of H2S in places. Three wells exceeded the maximum allowable level for manganese, which is 500µg L−1 (BOS, 2000). All three are located south of Ramotswa village along the border to South Africa. Like in the case of iron, this was probably due to geology. For example at one borehole black liquid was encountered, which was identified as manganese

Environmental hydrogeology of the dolomite aquifer in Ramotswa, Botswana

485

hydroxide. Most likely the manganese hydroxide stemmed from an underlying manganese-bearing slate. Four boreholes along the river Ngotwane revealed arsenic levels above the Botswana standard for drinking water of 10µg L−1 (BOS, 2000). Interesting to note was that three were aligned in a row in the northern part of the project area, and also that the As concentration rose in flow direction from 11.5µg L−1 to 22µg L−1 and finally to 51µg L−1. The fourth borehole (4166) that featured an elevated As concentration was located south of Ramotswa village. Of special interest in the study area was the case of nitrate because Ramotswa had experienced one of Botswana’s worst cases of groundwater pollution in recent history. This had followed the successful campaign by government to encourage villagers to build pit latrines. Available information suggested that by early 1997 the problem had become so bad that the Ramotswa wellfield had to be abandoned as a source of drinking water (Norwebb, 1996). Out of the 31 boreholes sampled in late 2001, 11 revealed nitrate levels that exceeded the Botswana standard of 45mg L−1 for drinking water (BOS, 2000). The maximum value was 442mg L−1 (Fig. 2). The nitrate concentrations in the other boreholes that also exceeded 45mg L−1 ranged from 64 to up to 188mg L−1. Looking at the spatial distribution of the affected boreholes and correlating the spatial distribution pattern to the prevailing groundwater flow direction, it was concluded that unpolluted water entered the study area from the south. The southern well field revealed no nitrate contamination at all. But the first borehole in northerly direction, which featured an elevated nitrate level of 72mg L−1, was located within the village. The prevailing groundwater flow direction in the area is in northwesterly direction. Hence, polluted groundwater is being carried towards the river Ngotwane. The nitrate levels in the boreholes located in the river plain ranged from 65.5 to 188mg L−1. And the peak of 442mg L−1 mentioned above was also located in the river plain. Fractures and cavities in the Ramotswa dolomite and in the upper karst zone are contributing to the transport of polluted groundwater in this direction. The groundwater flow in the Ramotswa station area is separated from the general groundwater flow through a watershed in the form of the Sepitswane hills. These hills, together with the hills at Taung (e.g. Bojanjwe hill), form a small drainage basin and the groundwater flows in a NNW direction parallel to the river Taung, until it comes together with the river Ngotwane north of the study area. In this area pollution due to pit latrines (Taung residential areas) and septic tanks (industrial area at Bolux, Tswana Steel, and White Dove Garments) is evident. Clearly, the hydrogeological conditions (dolomite karst) in the Ramotswa project area are unfavourable for sanitation practices such as pit latrines and septic tanks (cf. Lewis et al., 1980, Tredoux, 1993). In fact, they are considered to be the major source of nitrate pollution in populated eastern Botswana (Hutton, 1976; Lagerstedt, 1992; Lagerstedt et al., 1994; Carling & Hammar, 1995; Jacks et al., 1999). While 146 septic tanks were in use in Ramotswa in 1985 (WLPU, 1985), almost all septic tanks in the study area have been replaced over the last 10 years with connections to the newly built sewage system. In late 2001 only a very few were still in use mainly in the industrial area of Ramotswa station (Bolux, Tswana Steel, White Dove Garments).

Water resources of arid areas

486

Yet, the fact that nitrate levels were still high at the end of 2001 highlights a persistent pollution problem. Most likely the problem emanated from the still large number of pit latrines. In 1991, 2432 pit latrines were in use that is 66% of all households used pit latrines then (Enneco, 1996). Given the fact that the population grew over the last two decades from approximately 14000 inhabitants in 1981 to 20680 in 2001, the number of pit latrines has most likely increased further. Hence, the nitrate pollution problem in Ramotswa will not be solved unless all pit latrines are replaced with flush toilets that are connected to a properly functioning sewage system. But in late 2001 not even the new sewage system was in proper working condition. At several places the new sewers were leaking, ironically because of damages caused by excavators who were still working on the completion of the system. Livestock may be another source of nitrate pollution. In places where livestock is highly concentrated, animal waste may contribute to groundwater pollution. A case in point was the cattle kraal next to a borehole where the watering site was less than 50m away from the well house. Here the nitrate concentration was 148mg L−1 in late 2001. The primary health hazard from drinking water with nitrate-nitrogen (NO3-N) occurs when nitrate (NO3) is transformed to nitrite (NO2) in the digestive system (Vogel, 2002). While NO3 is not very toxic, NO2 is toxic. The nitrite oxidizes iron in the hemoglobin of the red blood cells of humans (and other warm-blooded animals) to produce methemoglobin. Methemoglobin lacks the ability of hemoglobin to transport oxygen to body tissues. This creates the condition known as methemoglobinemia (“blue baby syndrome”), in which red blood cells carry insufficient oxygen to the individual body cells thus causing the veins and skin to appear blue (“internal suffocation”). This condition is especially serious in infants. 5 DISCUSSION AND CONCLUSIONS The results of this study confirmed that groundwater pollution problems still exist in Ramotswa. The most obnoxious pollution problem was again due to high nitrate concentrations in several boreholes. The latter was attributed to the continued use of pit latrines. In fact, at sites where the soil is very thin, and given a mean depth of the pit latrines of 1.7m (WLPU, 1985), human waste may enter directly into the aquifer. Clearly, pit latrines in the study area ought to be replaced with facilities connected to the new sewerage system. But the latter also must be maintained properly. The industrial complex in the Ramotswa station area turned out to constitute another area of concern. In this area five boreholes were found that produced contaminated groundwater. High nitrate levels and elevated concentrations of sulfate, sodium, and chloride were identified. This was indicative of anthropogenic pollution, as were the observed nickel and possibly the aluminium levels. Because of these pollution indicators and also because of the observed elevated but possibly natural arsenic concentrations along the river Ngotwane, it was decided to resample all boreholes starting in late 2003. In addition, several blocked boreholes, which could not be sampled in 2001, will be rehabilitated and a very few new boreholes will be drilled in strategically important places. The objective of this programme is to thoroughly

Environmental hydrogeology of the dolomite aquifer in Ramotswa, Botswana

487

determine groundwater pollution levels in the Ramotswa area, and to contribute towards the reduction of groundwater pollution. This will also help alleviate a further matter of concern that became glaringly obvious while carrying out this study, namely the lack of reliable data for the important Ramotswa wellfield. For example, it was impossible to establish trend lines of certain pollutants over time from the data retrieved from the national borehole archive. Only a very few were possible but even these raised doubt with regards to data reliability. Last but not least, the groundwater protection zones and previously made recommendations (WLPU, 1985; Water Surveys, 1994) ought to be strictly observed. This, and the elimination of the existing pollution problem, may in future allow for the renewed use of Botswana’s most important aquifer. REFERENCES BOS (2000). Water quality—Drinking water—Specification. BOS 32, Botswana Bureau of Standards, Gaborone, Botswana. Carling, M. & Hammar, M. 1995. Nitrogen metabolism and leakage from pit latrines. A minor field study from south-east Botswana. M.Sc. thesis, 54 p. plus appendices, Lulea Univ. of Technology, Dept. of Environmental Planning and Design, Div. of Waste Management and Recycling, Lulea, Sweden. Enneco 1996. Ramotswa Planning Area Development Plan. Report of Survey by Enneco (Pty) Ltd., Ministry of Local Government, Lands and Housing, Dept. of Town and Regional Planning (DTRP), Gaborone, Botswana. Geotechnical Consulting Services 2000. Groundwater Monitoring Project. Final Report, Vol.19, Ramotswa wellfield, Gaborone, Botswana. Hutton, L.G., Lewis, W.J. & Skinner, A.C. 1976. A report on nitrate contamination of groundwaters in some populated areas of Botswana. Report no. BGSD/8/76, DGS, Botswana. Institute of Hydrology 1986. Ramotswa Wellfield, Southeastern Botswana, Digital Model Study and Storage Estimates, Gaborone, Botswana. Jacks, G., Sefe, F., Hammar, M. & Letsamao, P. 1999. Tentative nitrogen budget for pit latrines, Eastern Botswana, Environmental Geology, 38(3):199–203. Lagerstedt, E. 1992. Nitrate contamination in the groundwater and nitrogen circulation in an area of south-east Botswana. M.Sc. thesis, 55 p. plus appendices, Stockholm Univ., Stockholm, Sweden. Lagerstedt, E., Jacks, G. & Sefe, F. (1994). Nitrate in groundwater and N circulation in Eastern Botswana. Environmental Geology, 23:60–64. Lewis, W.J., Farr, J.L. & Foster, S.S.D. 1980. The pollution hazard to village water supplies in eastern Botswana. Proc. Instn, Civ. Engrs 2(69):281–293. Norwebb, B.T. 1996. Nitrate levels—Ramotswa Wellfield. Dept. of Water Affairs (DWA), Gaborone, Botswana. Tredoux, G. 1993. A preliminary investigation of the nitrate content of groundwater and limitation of the nitrate input. Report to the Water Research Commission (WRC), No 368/1/93, 76 pp., Pretoria, R.S.A. Vogel, H. 2002. The soil nitrogen cycle. Report by the Environmental Geology Division, 25 pp., Dept. of Geological Survey (DGS), Lobatse, Botswana. Water Surveys 1994. Groundwater Pollution Vulnerability Map of the Ramotswa and Mogobane Area. Map 3.5 including Report No.4, Scale 1:50000, Water Surveys (Botswana) (Pty) Ltd., Gaborone, Botswana.

Water resources of arid areas

488

WLPU 1985. Ramotswa Wellfield Pollution Study. Watermeyer, Legge, Piesold & Uhlmann (WLPU), Final Report, Dept. of Water Affairs (DWA), Gaborone, Botswana.

Investigation of natural enrichment processes of nitrate in soil and groundwater of semi-arid regions: case study—Botswana S.Stadler, M.von Hoyer, W.H.M.Duijnisveld & T.Himmelsbach Federal Institute for Geosciences and Natural Resources (BGR), Hannover, Germany M.Schwiede & J.Böttcher Department of Soil Ecology, University of Hannover, Germany H.Hötzl Department of Applied Geology, University of Karlsruhe, Germany Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Extraordinary enrichment of nitrate in groundwater is a worldwide phenomenon, mostly due to anthropogenic activities in densely populated areas. In Southern Africa elevated nitrate concentrations are observed in mostly uninhabited semi-arid areas. In the Kalahari of Botswana groundwater locally exhibits concentrations up to 600mg/l, leading to infant mortality and stock losses. There, natural nitrate accumulation processes appear to play an important role. Enrichment processes are not yet fully understood. In our study we use a combined approach of soil science and hydrogeology to identify sources. We found that observed concentrations are a result of a complex interaction between sources and sinks, but also of spatially varied reactive potential in the aquifer.

1 INTRODUCTION Extraordinary nitrate enrichment in groundwater has been described in numerous studies around the world. Elevated concentrations are equally known from Botswana, South Africa and Namibia. Nitrate concentrations in groundwater samples from the Kalahari near Orapa/Botswana generally exceed the WHO-guideline value (WHO, 1998) of 50mg/l (as NO3) (Mokokwe, 1999). The exposure of humans to high doses of nitrate e.g. in drinking water causes severe health effects, e.g. infants can suffer from methemoglobinemia. In Southern Africa the risk of infant mortality has increased as an

Water resources of arid areas

490

immediate result of the AIDS epidemic which requires bottle-feeding of the infants, often with nitrate-rich water (Colvin, 1999). Nitrosamines are harmful to humans of any age. Elevated nitrate concentrations also shorten the lives of cattle, lead to deficient growth and decreased reproduction (Davidson et al., 1964). Plenty of research efforts have been conducted towards anthropogenically caused nitrate accumulation in groundwater, especially relating to spatial nitrate input from fertilizers. However, not much attention has been paid to natural nitrate enrichment in groundwaters of semi-arid and arid regions (Burt et al., 1993; Van der Hoek et al., 1998). Only few studies until now have focused on natural nitrate sources (Edmunds and Wright, 1979; Edmunds and Gaye, 1997; Barnes, 1992; Aranibar et al., 2003; Walvoord et al., 2003). In the Kalahari of Botswana natural nitrate accumulation processes appear to play an important role and were firstly investigated by Heaton, 1983, 1984. Enrichment processes, however, are not yet fully understood. To address this topic a research cooperation between the Federal Institute of Geosciences and Natural Resources (BGR), the Department of Applied Geology of the University of Karlsruhe, the Institute of Soil Science of the University of Hannover and the Department of Geological Survey (DGS) of Botswana was established. In this study we investigate the natural enrichment processes of nitrate in groundwater of the Ntane-Sandstone Aquifer between Serowe and Orapa in the Kalahari of Botswana. The aim is to identify the (major) natural input factors and the fate of nitrate in groundwater. We use a combined approach of soil science and hydrogeology to fully trace the sources. 1.1 Study area The study area is located on the Eastern fringe of the Kalahari between Serowe and Orapa, Botswana and covers 70×250km. It lies in a semi-arid region with mean annual rainfall between 200 and 1100mm/yr (Lubczynski, 2000) and low humidity, restricted to one annual rainy season. However, there are extreme temporal and spatial variations in the distribution of precipitation. Potential evaporation ranges at 900–1200mm/yr (Lubczynski, 2000) and recharge is at the order of 1–5mm/yr (Verhagen, 1990; De Vries et al., 2000). The area is characterized by a flat, slightly undulating topography at about 1200m above sea level; the only topography feature is the Eastern escarpment. 1.2 Concept As the Kalahari region offers much potential for nitrogen input from different compartments (see Fig. 1), it was necessary to validate and investigate the individual sources. Dominant nitrogen input was expected from the following: i) Natural vegetation, especially acacia trees and shrubs, the predominant vegetation in the area, are known for their N-fixing potential. ii) Cattle, though extensive, cause (over)grazing, displacing natural mammals, and leave nitrogen from manure especially in Kraals (cattle posts). iii) Termite mounds, occurring in the area, allow an accumulation of nitrogen in symbiosis with fungi cultures. All these features, coupled with climatic aspects such as downward flushing during major rainfall events in combination with a low seepage rate, and the transport by surface and preferential flow of seeping water in macropores (root channels,

Investigation of natural enrichment processes of nitrate

491

animal burrows), enhance nitrate enrichment in the subsurface. Lithology could provide further nitrogen, e.g. in calcretes and silcretes of the uppermost formation

Figure 1. Main sources and pathways of nitrate in the study area: (i) Atmospheric input by rainfall and dust; (ii) Vegetation and organic matter: nitrogen accumulation processes in the root zone; (iii) Animals: manure from cattle, wild animals, activities of termites, beetles etc.; (iv) Climate: long dry periods alternating with periods of heavy rainfall, leading to downward leaching in soils; (v) Lithology: Nitrate containing minerals or reaction partners, e.g. in fossil pans; (vi) Uprising groundwater: Inflow of nitrate rich groundwater from lower lying aquifers. of the investigated area (acting as source or hydraulic barrier). However, the potential sink capacity of the aquifer also has to be considered, e.g. the presence of reaction partners for nitrate reduction in the aquifer. Furthermore, influence from potentially

Water resources of arid areas

492

nitrate-rich uprising groundwater from the lower lying aquifer (Mosolotsane Fm.) or other allochthonous processes might play a role. For those reasons we used a combined approach of hydrochemical, hydroisotopic, and soil scientific methods. Soil scientific investigations include (i) the quantification of the soil nitrogen balance at typical sites of the Kalahari, and the evaluation of stocks of mineralized nitrogen in soils and unsaturated zone with respect to vegetation and cattle grazing intensity, (ii) the identification of nitrogen dynamics in the root zone, and investigation of leaching processes and the transport and transformation of ammonia and nitrate in the unsaturated zone, and (iii) the development of a model for prognoses of nitrate leaching into groundwater with respect to vegetation and land use (grazing) intensity. Hydrogeological and hydrochemical investigations include (i) the determination of hydrochemical characteristics of the groundwater types occurring in the Serowe-Orapa (Botswana) groundwater system (ii) the determination of groundwater residence times and recharge sources by a wide range of environmental tracers (iii) the investigation of input processes leading to the varied spatial distribution of nitrate in the aquifer, including rock and water analysis, and (iv) the investigation and modeling of the reactive transport and fate of nitrate under confined conditions. 2 SOIL 2.1 Soil scientific settings In the Kalahari region the soil type is mainly Arenosol, a deeply to very deeply developed soil type with a poor differentiation between the different soil horizons. On the Kalahari plateau, the Arenosol is formed in aeolian sediments from palaeo-climatic dune sands. In the investigation area the Arenosol is mostly ferralic, its texture is medium to fine sandy with a yellowish-brown to dark red color, whereas its drainage capacity is moderately well to somewhat excessive (De Wit & Nachtergaele, 1990). A region dominated by Petric Calcisols is found in the North-western part of the investigation area at Letlhakane and Orapa. This soil type region is bordered by the Makgadikgadi Pans where the saline soil type Solonchak is dominating. Throughout the study area “duricrusts”, like calcrete, silcrete and ferricrete crusts, are widespread. These hard layers occur in different soil depths (from several decimeters to decameters) reaching a vertical extension from several decimeters up to a few meters (Martini & Chesworth, 1992). 2.2 Botanical settings and fauna In shrub and tree savannas plant communities on Arenosols are dominated by Acacia and Terminalia species, whereas Mopane trees are dominating the sites of calcareous soils (Thomas & Shaw, 1991; van Wyk, 2001). Various kinds of termites are ubiquitous in the Kalahari. Their high number and biomass is playing an important role in the Kalahari ecosystem as collector and decomposer of wood, grass and organic litter (Lavelle & Spain, 2001).

Investigation of natural enrichment processes of nitrate

493

2.3 Field and laboratory work The field work started in 2001 with a description of soils and vegetation in the study area by mapping. Mapping basis is a Landsat TM scene. During the first sampling campaign in the second half of 2002 soil pits were dug at representative sites to describe the soils in the area. Soil samples were taken and analyzed for soil texture, pH, organic carbon, organic nitrogen, carbonate and cation exchange capacity. In addition, 22 soil profiles up to a depth of 5m were drilled by hand and soil samples were taken over the total depth of these profiles. The soil profiles were situated along transects next to potential nitrogen sources. During the second field campaign carried out at the end of 2003, a heavy hydraulic/percussion drill apparatus (Selaolo, 1998) was used to sample 11 soil profiles up to an average depth of about 20m, in one case down to 65m. To determine concentration-depth profiles of chloride (Cl), nitrate (NO3) and ammonia (NH4) from the sampled soil profiles, the soil samples were extracted with distilled water and the water samples were filtered through a 0.45µm filter in a provisional field laboratory in the study area. The filtered water samples were deep-frozen on site for transport and the hydrochemical analysis was done in a chemical laboratory. For the calculation of the anion concentration in the original soil solution the moisture content of parallel probes was measured. 2.4 First results Figure 2 (left) shows depth profiles of organic carbon and organic nitrogen for two soil profiles in the study area. The values of the organic-C and -N content shown are representative for soil profiles under natural vegetation throughout the study area. The values are rather low, and only a weak accumulation of organic matter takes place in the upper part of the profile. The pH-values (Fig. 2 right) of these two soils under natural vegetation, however, vary from more or less acidic (S51) to almost neutral (S50). Further analysis of soil profiles and soil chemical parameters is necessary to determine the reasons behind these observations. Measured concentration profiles show very high nitrate and chloride concentrations in soils and soil solutions in kraals and next to the cattle watering tanks of the cattle posts. This is shown by the results of a transect at the Makhi test farm (see Fig. 3), where the influence of livestock grazing

Water resources of arid areas

494

Figure 2. Organic carbon and organic nitrogen content (left) and pH-values (right) of two soil profiles in Arenosols (S50 and S51) under natural vegetation in the study area.

Figure 3. Mean values of nitrate-N, chloride and ammonium-N concentrations (mg/kg soil) and moisture content (%-weight) in 3–5m deep soil profiles along a radial transect of Makhi test farm (0=kraal). under controlled conditions is analyzed (Mphinyane, 2001). In kraals and their direct surroundings the measured average concentration is 54mg NO3-N/kg of dry soil, whereas the “natural” background concentration at a distance of more than 1000m from the cattle post is about 1.1mg NO3-N/kg (Fig. 3). Even this lower “natural” level is higher than expected for this semi-arid region with sandy soils. Nitrate concentrations decrease with

Investigation of natural enrichment processes of nitrate

495

increasing distance from the cattle post and thus are also correlated with grazing intensity. Also, at the base of termite mounds extremely high concentrations of up to 750mg NO3-N/kg were measured (Fig. 4). In between these two termite mounds the concentrations are comparable to the values found under natural vegetation at the test farm (Fig. 3). Thus cattle posts and termite mounds are potential sources of nitrate pollution of the groundwater. To get more insight into the spatial and temporal variability and heterogeneity of nitrate and ammonium concentrations, appropriate soil profiles and sites were sampled. Figure 5 shows the results of soil profile sampling below a kraal in October 2002 and again at the same site in October 2003. The overall picture of the nitrate concentration profiles on both dates is more or less the same, but spatial heterogeneity masks the detection of a potential vertical downward movement of the nitrate profile during this period of one year. Very interesting are also the concentration profiles of two soil profiles sampled 50m apart at the same date in November 2003 (Fig. 6). The nitrate profiles A and B look very much the same, which indicates, that the spatial variability seems to be quite low. However, it is currently not possible to explain the increased nitrate concentrations found in 8–12m depth. A more thorough analysis of the climatic conditions during

Figure 4. Mean values of nitrate-N, chloride and ammonium-N concentrations (mg/kg soil) and moisture content (%-weight) in four soil profiles (1m deep) at the base (0 and 105m) and between two termite mounds 105m apart.

Water resources of arid areas

496

Figure 5. Nitrate-N and ammonium-N concentration profiles at the kraal of the Makhi test farm sampled in October 2002 and October 2003.

Figure 6. Two soil profiles sampled 50m apart at the same date in November 2003 under natural vegetation showing only small spatial variability in nitrate-N, resp. ammonium-N concentrations.

Investigation of natural enrichment processes of nitrate

497

the last 30 years might bring better insight into the leaching conditions during this period and might help to explain the measured concentration profiles in Figure 6. 2.5 Soil: First conclusions and further research First results of the soil profile sampling reveal potential sources of nitrate, which especially have high concentrations at cattle posts and termite mounds. But also under natural vegetation the concentrations in the soil and the soil solution are unexpectedly high. A further analysis of the already sampled soil profiles is needed to get a more complete picture of the overall situation. Microbial research is currently done on acacia trees and other leguminous plants to get a better estimate of the role of nitrogen fixation. Incubation experiments are conducted in the laboratory to get insight into Nmineralization potential of the soils. Also grasses, leafs and other parts of the natural vegetation have been sampled to determine the dry biomass and the biomass production per unit area. Modeling of the soil water regime and nitrate transport in the unsaturated zone must be done to get quantitative insight into a potential nitrate leaching to the groundwater. 3 GROUNDWATER 3.1 Hydrogeological setting The studied area is located in the Karoo strata. The main aquifer is the Triassic Ntane sandstone aquifer with an average thickness of 100 m, underlain by the less permeable Mosolotsane layers (mudstone, siltstone, sandstone). It is overlain by the Early Jurassic Stormberg basalt (SGAB, 1988). This basalt is of varying thickness and mostly confines the Ntane aquifer. The former is covered by the Tertiary to recent Kalahari beds, a group of superficial deposits of soil, sand, and clay, with locally extensive formations of calcrete, silcrete and ferricrete that have been formed along former drainage lines and in shallow pan depressions. The hydrogeology of the area is strongly controlled by structural features. The general groundwater flow direction is SSE to WNW. However, the area is dominated by a large number of hydraulically influencing WNW-ESE trending tensional faults, which cut through the entire area, and some of which are intruded by lowly permeable dolerite dykes. Those divide the Karoo units into horst and graben structures of varying size. A similar, but hydraulically less significant set of faults exists in the NE-SW and NW-SE directions (Wellfield Consult, 2000). 3.2 Hydraulic properties The Ntane Aquifer is an aeolian sandstone with a porosity controlled by intergranular voids (primary porosity) and fractures and fissures (secondary porosity). Groundwater levels are at 30 to

Water resources of arid areas

498

Figure 7. Piper diagram of sample of the Serowe study area, showing a correlation of NO3 with a Ca-(Na)HCO3 water type. 100m below the surface. Its transmissivity (T) ranges from 5.0*10−5−1.7*10−4m2/s (Wellfield Consult., 2000), the confined storage coefficient in the range of 1.0*10−4−5.0*10−5 and the average yield of the boreholes ranges over a span of 2– 30m3/h. Those parameters are of course strongly influenced by the location of the borehole in a fracture zone or away from it in a sparsely fractured block. The huge thickness of the unsaturated zone and the low recharge of 1–5mm/yr (Verhagen, 1990; De Vries et al., 2000) have a strong influence on groundwater chemistry, especially on substances that are infiltrated through the soil. 3.3 First results from hydrochemistry 3.3.1 General characterization of the area The groundwater from the Ntane Aquifer is dominated by Ca-HCO3 to Na-HCO3– character, indicating cation exchange processes. It has a low to medium ion content (EC range 500–800µS/cm) and neutral to basic pH of 7.0–9.0, and temperatures of 26–28°C. EC is higher in waters with a Ca-HCO3 signature than in those with a Na-HCO3 signature.

Investigation of natural enrichment processes of nitrate

499

3.3.2 NO3 distribution The nitrate distribution is very complex. Elevated nitrate occurs in a spatially scattered pattern, with localized clusters of elevated concentration. However, in some cases in directly neighboring wells (distance <20m) steep gradients can be observed. Maximum nitrate concentrations in the Serowe area amount to 220mg/l (as NO3). An attempt to verify the pattern with hydraulic and hydrologic information indicated that the distribution of high nitrate cannot singularly be attributed to the flow path. The strong heterogeneity in distribution occurs both in the spatial as well as in the vertical direction. However, water chemistry appears to play an important role. Our investigations show clearly that elevated nitrate is correlated with the Ca-Mg-HCO3-type water, whereas waters with Na-HCO3 signatures have low NO3 concentrations (Fig. 7). This is supported by the 14C-NO3 relationship that indicates that the younger groundwater exhibits higher nitrate concentrations than the older groundwater (data not shown). 3.3.3 Reaction processes influencing the NO3 distribution Despite little variation in general water type we found a spatial variability of the different water component concentrations, e.g. Na, Ca and NO3. This suggests a mixture of waters, where the local character of the water is a picture of the degree of mixing. This may suggest two options: i) The attraction of waters from different hydrological regimes. This could mean the attraction of

Water resources of arid areas

500

Figure 8 (a, b). Example of results from a chemical pump test at BH8480 (pump rate: 15m3/h, test duration: 72h). Different water chemistry is attracted after 1000 min of pumping. water from fissure and fracture zones as opposed to pore water of the Ntane sandstone. Indications for this were found in results from 72-hour chemical pump tests that showed changes in water components as a function of pumping time (Fig. 8 a, b). Also, the presence of water of the (full) exchange series from Ca-Mg-HCO3 to Na-HCO3 could reflect this. ii) The local differences in water chemistry could be a result of the abundance of reaction zones. In depth orientated samples we locally found nitrate variation with

Investigation of natural enrichment processes of nitrate

501

depth and according variation of redox parameters, e.g. oxygen and bicarbonate. This could be a sign for the possibility of autochthonous nitrate reduction (Fig. 9). Due to local heterogeneities a complex interaction of both options is assumed. 3.4 Groundwater: First conclusions and further research We found a complex distribution of nitrate in the groundwater. The occurrence of high nitrate is very heterogeneous and cannot be singularly attributed to the flow path. Nitrate distribution is, however, not only controlled by different input sources, but by a complex interaction of sources and sinks. Their distribution is controlled by two components: The flow system partly dominated by fissures and fractures, as well as by the likely abundance of reaction zones. For both components we observed a high spatial heterogeneity in the investigated study area. This heterogeneity explains the high variability of the observed local nitrate concentrations. Further interpretation of the water analyses and hydroisotopic data will provide more insight into the involved processes and their relation to nitrate occurrence. The investigation of further core samples will give information on the presence or absence of N-bearing minerals or reaction partners within the aquifer. A more detailed look on hydrochemical zonations will be given from multiparameter probe profiles. Different scenarios on the fate of nitrate in the aquifer will be

Water resources of arid areas

502

Figure 9. Example of depth orientated water chemistry in borehole BH8471. Possible nitrate reduction in depths >80m below the water level is suggested by redox indicators oxygen and bicarbonate (assuming calcite precipitation with its according decrease in calcium and bicarbonate).

Investigation of natural enrichment processes of nitrate

503

verified by hydrochemical modeling. Numerical simulations will be used in order to derive detailed flow paths, flow velocities and mean residence times. REFERENCES Aranibar, J.N., Anderson, I.C., Ringrose, S. & Macko, S.A. 2003. Importance of nitrogen fixation in soil crusts of southern African arid ecosystems: acetylene reduction and stable isotope studies. J. Arid Env. 54: 345–358 Barnes, C.J., Jacobsen, G. & Smith, G.D. 1992. The origin of high nitrate ground waters in the Australian arid zone.—J. Hydrol.137:181–197 Burt, T.P., Heathwaite, A.L. et al. 1993. Nitrate: Processes, patterns and management, Chichester, John Wiley & Sons, Chichester, 444pp. Colvin, C. (1999): Increased risk of methemoglobinemia as a result of bottle feeding by HIV positive mothers in South Africa.—IAH Congress, Bratislava. Davidson, K.L. et al. 1964. Nitrate toxicity in dairy heifers. I. Effects on reproduction, growth, lactation and vitamin A nutrition.—J. Dairy Sci., 47:1065–1073. De Vries, J.J., Selaolo, E.T. & Beekman, H.E. 2000. Groundwater recharge in the Kalahari, with reference to paleo-hydrologic conditions. J. Hydrol. 238:110–123 De Wit, P.V. & Nachtergaele, F.O. 1990. Soil Mapping and Advisory Services Botswana: Explanatory Note on the Soil Map of the Republic of Botswana (Typifying Pedons and Soil Analytical Data); Gaborone, December 1990 Edmunds, W.M. & Gaye, C.B. 1997. Naturally high nitrate concentrations in groundwaters from the Sahel. J. Environ. Qual. 26:1231–1239 Edmunds, W.M. & Wright, E.P. 1979. Groundwater recharge and paleoclimate in the Sirte and Kufra Basins, Lybia. J. Hydrol 40:210–241 Heaton, T.H.E, Talma, A.S & Vogel, J.C. 1983. Origin and history of nitrate in confined groundwater in the Western Kalahari, J. Hydrol 62:243–262 Heaton, T.H.E. 1984. Sources of the nitrate in phreatic groundwater in the western Kalahari, J. Hydrol 67: 249–295. Lavelle, P. & Spain, A.V. 2001. Soil Ecology. 514 pp. Dordrecht, Kluwer Academic Publishers. Lubczynski, M. 2000. Groundwater evapotranspiration—Underestimated component of the groundwater balance in a semi-arid environment—Serowe case, Botswana. In: Sililo et al. (eds): Groundwater: Past Achievements and Future Challenges. Rotterdam, Balkema: 199–204 Marrett, D.J., Khattak, R.A., Elseewi, & Page, A.L. 1990. Elevated nitrate levels in soils of the eastern Mojawe desert.—J. Environ. Qual. 19:658–666 Martini, I.P. & Chesworth, W. 1992. Weathering, Soils & Paleosoils—Developments in Earth Surface Processes2:309–377, Department of Land Resource Science, University of Guelph, Ont. N1G 2W1, Canada, Elsevier, Amsterdam-London-New York-Tokyo Mokokwe, K. 1999. Occurrence of groundwater with high nitrate content—Orapa wellfields. Inception report, DGS, Lobatse. Mphinyane, W.N. 2001. Influence of livestock grazing within biospheres under free range and controlled conditions in Botswana. PhD thesis: Department of Plant Production and Soil Science, Faculty of Natural and Agricultural Sciences, University Pretoria Perkins, J.S. & Thomas, S.G. 1993. Environment Responses and Sensitivity to Permanent Cattle Ranching, Semi arid Western Central Botswana In: Landscape Sensitivity: 273–286, edited by D.S.G.Thomas, & R.J.Allison (eds), John Wiley and Sons Ltd Schulze, E.D., Gebauer, G., Ziegler, H. & Lange, O.L. 1991. Estimates of nitrogen fixation by trees on aridity gradient in Namibia. Oecologia 88:451–455, Springer Verlag. Selaolo, E.T. 1998. Tracer studies and groundwater recharge assessment in the eastern fringe of the Botswana Kalahari, the Letlhakeng-Botlhapatlou area. Thamaga, Botswana.

Water resources of arid areas

504

SGAB (Swedish Geological Survey) 1988. Serowe—Groundwater resources evaluation project, Internal report. DGS-Botswana, Lobatse SRK Consulting 1981. Report on the Letlhakane Mine Dewatering System for DeBeers Botswana. Internal Report No. CL.2439/2 Thomas, D.S.G. & Shaw, P.A. 1991. The Kalahari Environment, Cambridge, Cambridge University Press, Science B.V.: 284pp Walvoord, M.A., Phillips, F.M., Stonestrom, D.A., Evans, R., Hartsough, R.D., Newman, P.C., B.D. & Striegl, R.G. 2003: A Reservoir of Nitrate Beneath Desert Soils.—Science, 302:1021– 1024. (in Reports) WHO (1998): Guidelines for drinking water quality. Van der Hoek, K.W. et al. (eds.) 1998: Nitrogen—First Int. Nitrogen Conf. Proceedings. ISBN 008 0432018 Van Wyk, P. (2001): Southern African Trees—A Photographic Guide; Struik Publishers (Pty) Ltd, Cape Town 144 pp. Verhagen, B.T. (1990): Isotope hydrology of the Kalahari: Recharge or no recharge? Palaeoecology of Africa and the surrounding islands. 21, S:143–158 Wellfield Consulting (2000): Serowe Wellfield 2 Extension Project (TB10/3/10/95–96). Internal report for the Department of Water Affairs, Botswana

Hydroclimatological approach to sustainable water resources management in semi arid regions of Africa U.T.Umoh University of Botswana, Gaborone, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Climatic and runoff data in semi arid regions of Africa were used to assess the implications of climatic variability on some water balance parameters. Result shows the coefficient of determination between discharge and rainfall as 0.64 to 0.72, soil moisture capacity of 170mm and infiltration of 608.3mm. Soil moisture deficit and surface runoff are 404.1mm/m2 and 161.6mm/m2 respectively. Responses to climate change include temperature rise of about 1.5°C, incidence of droughts, desertification, aridity, decline in rainfall, decline in runoff and water shortages for plant, animals and man. Water resources management strategy is vital for sustainable use of the region’s meagre water resources. Suggested sustainable water resources management policy will include: effective water pollution control, water re-use, water recycling, water demand management involving restriction/reduction in water use, control of distribution losses and the exploitation of new sources of water such as rain water harvesting and desalination.

1 INTRODUCTION Semi arid regions of Africa has greatly been affected by climate change and climatic variations over the past centuries and decades. Responses to climate change in drylands of Africa include incidence of droughts, desertification, aridity, decline in rainfall, decline in runoff and water shortages for plant, animal and man. During the late sixties and early seventies, for example, the southern border region of Sahara desert, the Sahel, succumbed to prolonged drought, desertification and famine. Climatic variations resulting in disastrous decline in rainfall over semi arid regions of African continent have impacted negatively on various components of water resources. Droughts and disastrous decline in rainfall which are direct consequence of climatic variability have been on the increase since the 1960’s in drylands of Africa. These phenomena have been attributed to complex interplay of sea surface temperature anomalies over the tropical Atlantic Ocean and the attendant latitudinal shifts of circulation systems during the dry years,

Water resources of arid areas

506

exacerbated by anthropogenic factor (Hastenrath, 1990). The El-Nino Southern Oscillation (ENSO) has been identified as the most dominant perturbation responsible for interannual climate variability over eastern and southern Africa (Nicholson and Entekhapi, 1986). Over northern Africa, the North Atlantic Oscillation (NAO) is the key factor that is responsible for interannual variability in the climate (Lamb & Peppler, 1992). Climate change is reflected in decline in rainfall amount and increasing droughts. This in turn affects the various components of surface and ground water resources with serious implications on agriculture, water needs and supply, sanitation, transportation, hydroelectric power generation and environmental hazards such as flooding and erosion. Water Resources Management (WRM) can be defined as a whole set of technical, institutional, managerial, legal and operational activities required to plan, develop, operate and manage water resources. For sustainable water resources management, various components of water balance parameters must be understood. This paper explores hydroclimatological approach to examine the implications of climate change on some water balance parameters for sustainable water resources management. 2 METHOD OF STUDY Climatic data (daily and monthly rainfall, temperature and evaporation), runoff records (discharge and stage), from 1900 to 1999 were collected from different stations in semi arid regions of Africa and used in the analysis. As part of analysis carried out in the study, rainfall totals and means for the study period were taken as series. The series were subjected to a battery of time series analysis procedure including test for variability and Gaussian low pass filter technique for detecting marked fluctuations. Regression, rainfallrunoff and water balance models were also constructed. Water balance parameters of Potential Evapotranspiration (PET), Actual Evapotranspiration (AET), Readily Available Supply in the Soil (RAS), Soil Moisture Deficit (SMD), Surface Runoff (SR), Infiltration and Areal Annual Volume were calculated according to water balance estimation method (Umoh, 1999). 3 RESULTS AND DISCUSSION 3.1 Climate change Climate change includes the extremes of rainfall (i.e. droughts and floods) and differences of monthly, seasonal and annual values from normals. Climate change scenarios for semi arid regions of Africa, based on results from several general circulation models using data collated by the Intergovernmental Panel on Climate Change Data Distribution Centre, indicate future warming across the region ranging from 0.2°C to 0.5°C per decade. This warming is greatest over the interior of semi-arid margins of the Sahara and central southern Africa (Desanker & Magadza, 2001). In order to show the possible long-term fluctuation in rainfall over semi arid regions of Africa, yearly rainfall values for 100-year period were plotted for selected stations within the region.

Hydroclimatological approach to sustainable water resources management

507

The rainfall regimes for the stations are generally in the form of alternating wet and dry epochs (Figure 1). The annual rainfall curve for Kaduna (West Africa) indicates three periods of increased rainfall: 1905–1909; 1925–1929; 1946–1949; and three periods of marked decrease: 1911–1916; 1931–1938; 1952–1999 (Figure 1). In Botswana (Southern Africa) two periods of increased rainfall were 1921–1927 and 1930–1939 and four periods of marked decrease: 1913–1915; 1940–1948, 1951–1958 and 1960–1999. In between these periods, the pattern of rainfall is oscillatory. Descriptive statistics for annual rainfall series in the semi arid zone of West Africa for the study period are shown in Table 1. Izom registered the highest mean annual rainfall of 1513.1mm while Lokoja had the least being 1122.1mm. The standard deviation ranged between 179.9 for Kontagora and 323.3 for Kachia. It can also be seen from Table 1 that annual variability in rainfall differ from station to station. Results of analysis of basic statistical parameters of Kaduna reveals that the mean annual rainfall over 100 years period is 1356.3mm with standard deviation (SD) of 319mm. Of this total, 1254.1mm or 92.5 percent (with a SD of 244.5mm) fell in the months of May to October. Mean annual rainfall and rainy season coefficient of variation (CV) are 23.5% and 19.5% respectively. The low values of annual coefficient of variation for the station suggests that rainfall is less variable from year to year and that its variability is greater in the individual months. The decade 1950–1959 was characterized by above normal precipitation over semi arid regions of Africa, although rainfall deficiencies prevailed over the near equatorial region. Latter, during the period 1960–1969, this rainfall anomaly pattern dramatically reversed with rainfall deficit observed for semi arid regions of Africa while the equatorial region experienced widespread abundance of rainfall. More recently, the pattern has been one of increased aridity throughout the region. Mean rainfall decreased by 30–45% in the Sahel between periods 1931–1960 and 1968–1997 and generally 15–25% across the rest of the semi arid regions of Africa.

Water resources of arid areas

508

Figure 1. Standardised annual rainfall anomalies at three stations in semi arid region of West Africa (1900–1999).

Hydroclimatological approach to sustainable water resources management

509

Table 1. Statistical properties of annual rainfall at six stations in the semi arid zone of West Africa. Station

Mean Standard (%) Coefficient deviation variation of skewness

Izom 1,513.1 Kachia 1,438.2 Kaduna 1,356.3 Kontagora 1,126.2 Lokoja 1,122.1 Minna 1,230.1

181.8 323.3 319.0 176.9 225.4 203.0

12.0 0.45 22.5 0.35 23.5 1.90 15.7 0.58 20.1 0.42 16.5 0.70

3.2 Rainfall-runoff relationship The runoff series was regressed on the estimated areal rainfall for the semi arid region of West Africa at zero, one, two and three months lags. The coefficient of determination (r2) between discharge and rainfall is 0.72 which indicates a satisfactory linear relationship. Rainfal-runoff data for the two drainage basins within the region reveal the following characteristics. In Kaduna basin, surface runoff occurs in the months of March to November. The months of November to February the loss is due to evaporation but where potential evapotranspiration is less than the loss (May to September) the extra losses are making up ground water deficiencies. For Gbako basin runoff occurs from March to November. The dry season months where evapotranspiration is higher than the water loss is from November to April (a period of 6 months). The months that potential evapotranspiration is lower than water loss is between May and October. Rainfall-runoff relations over drainage basins in the semi arid zone of West Africa indicates that areal rainfall is lowest in Kaduna basin and highest in Gurara basin with the mean annual rainfall ranging from 1220mm in Kaduna (Wuya) to 1474mm in Gurara (Jere). Surface runoff in the drainage basin varies between 127mm in Kaduna and 228mm in Gurara. Runoff as percentage of rainfall is between 10.2% and 10.9% for Kaduna basin; 12.1% for Gbako basin and 12.4%–15.4% for Gurara basin. The 30-year average of areal rainfall over the Niger Basin is 1326.4mm; while surface runoff is 163.6mm constituting about 12.2 percent of the rainfall. Rainfall has dominant effect on peak flow regime variations. 3.3 Water-balance parameters The result of annual areal volume of some water balance parameters for three drainage basins in semi arid zone of West Africa are presented on Table 2. In the drainage basin there is a wide difference between PET and AET which indicates a distinct climatic variation in the area. Soil moisture deficit starts in December and end in April. Of the three basins, soil moisture deficit is highest in Kaduna basin and lowest in Gurara. The computation of annual averages of water balance components of parts of the Niger River Basin in the semi arid region of West Africa shows that the mean annual areal rainfall of the basin is 1321.7mm. Potential Evapotranspiration (PET) is 1072.6mm while actual

Water resources of arid areas

510

evapotranspiration (AET) is 635.6mm. The soil moisture capacity (RAS) is 170mm on the basis of 1.0mm effective root zone, while infiltration to groundwater is 678.3mm. Soil moisture deficit (SMD) and surface runoff are 404.3mm and 161.6mm respectively. A consideration of the relationships between rainfall plus infiltration and soil moisture deficit for Kaduna, Gbako and Gurara River Basins reveals that February marks the onset of rains in Kaduna river basin. Infiltration in the three drainage basins lasts for five months (June to October). There is no soil moisture deficit between May and November (a period of seven months) in Kaduna basin and between May and December (a period of eight months) in Gbako and Gurara basins. Maximum actual evapotranspiration (AET) in Kaduna basin occurs in the month of May corresponding to the period of maximum soil moisture accretion. In Gbako and Gurara basins, the

Table 2. Annual areal volume of some water balance parameters in three drainage basins (×106m3/year). Gbako Gurara Kaduna (65,150km2) (7,540km2) (23,730km2) Rainfall Surface runoff Infiltration AET SMD

81,632.9 8,404.3

9,688.9 1,146.1

33,863.6 4,840.9

42,086.9 35,246.1 30,750.8

4,878.3 4,991.4 2,915.6

17,607.6 16,705.9 8,281.7

peak of AET occurs in December. The difference in maximum AET month within the three drainage basin can be attributed to the fact that cessation of rainfall in Gbako and Gurara basins is in November while that of Kaduna is October, thus resulting to one extra month of rain duration over Kaduna basin. This extra rainfall month consequently increase the duration of moisture in the soil, to the extent that, in the month of December the AET is equal to PET, though the moisture availability is relatively low (15mm in Gbako and 33mm in Gurara). 4 IMPLICATIONS FOR SUSTAINABLE WATER RESOURCES MANAGEMENT The only source of recurrent water is rainfall in the region. It supplies water for both surface (overlandflow and streamflow) and ground water runoff. Rainfall is seasonal in the region and most rivers dry up during the dry season. There is a close relationship among rainfall, overlandflow and stream discharge. Drought and desertification disrupt this closeness since they decrease both surface flow and ground water storage. In the semi arid region, a lot of rainfall, that usually comes with high intensity, runs off rapidly. Such conditions impede groundwater recharge and create flash flow on the surface. When

Hydroclimatological approach to sustainable water resources management

511

drought occur in the region, much less amount of water is available for the two processes because the rainfall is less and a high proportion of the water is evapotranspired. Effects of climate change include increasing incidents of drought and aridity, incursion of desert-like conditions and decline in agricultural activity (Olofin, 1985; UNEP, 1987; Issar, 1998). The implications of climatic variability on water resources is demonstrated by the coincidence of low discharge, recharge and runoff with periods of drought. An indication of long term climate changes in the semi arid regions of Africa is depicted by a 90-year long record of flow in the Zambezi River at Victoria Falls shown in Figure 2. Average flow over the study period was 1056m3/s. Annual flows were mostly below the average for 38 years prior to 1945. Lower than average flows was recorded from 1910 to 1940 and 1980 to 1995. Low discharge and recharge affect the surface and the ground water of the region adversely. Under such conditions, there is very little water available on the surface and much less available as groundwater. Most rivers dry up during the dry season because of seasonality in rainfall. Climate change has serious impact on water resources of various semi arid African countries. In the region, the incidence of seasonal flow cessation is on the increase, as shown by some streams in Angola, Botswana, Namibia, Tanzania and Zimbabwe, resulting in water shortages for industrial and urban domestic supplies (Magadza, 2000). Major rivers such as Zambezi, Kafue

Figure 2. Graph showing the annual flow of water in the Zambezi River at Victoria Falls from 1907 to 1995 (after Skofteland 1995). (Zambia) and Niger and Kaduna (Nigeria) had very low flow. Water level in lake Kariba, Shiroro and Kainji were drastically reduced. The low inflows from the rivers put the

Water resources of arid areas

512

reservoirs at a dangerously low level, leading to power rationing and load shedding. Smaller streams, small dams and wetland areas in the valleys dried up early, reducing water for human consumption, livestock use and vegetable production. Compounding the effects of climate change on water resources in semi arid Africa, is the explosive urban growth that is placed at 4% per year for developing countries. Provision of clean water in the big cities is a challenging task of development strategists. It is known that already urban water needs compete with agricultural water requirements. Groundwater levels are known to be dropping rapidly as a result of demands of nearby urban centres. When such urban growth with its concomitant water demand is set against declining water availability due to climate change, one will agree with Nakayama (1998) that water may be the 21st century’s oil. Water resources of semi arid regions of Africa are very vulnerable to climate change through occurrence of droughts and declining rainfall. The most important input that can be made to cushion the adverse effect of climate change on water resources in the region is to adopt a water management option that would stand up to the extremes of climatic variations that is currently plaguing the region. Water conservation is vital for sustainable use of the region’s meagre water resources. Components of such water conservation policy will include: effective water pollution control, water re-use, water recycling, water demand management involving restriction/reduction in water use, control of distribution losses and the exploitation of new sources of water such as rain water harvesting and desalination. 5 CONCLUSION Semi arid regions of Africa has experienced pronounced climatic variations with their accompanying hydroclimatic events and consequences. Rainfall and runoff have fluctuated over the years with a declining trend. The decline in rainfall is reflected by overall decline in annual peak flood and minimum water level. The dominant impact of climate change is reduction in runoff. Current trends in major basins indicate a decrease in runoff of about 17% over the past decade. The severity of the dry season increases during the sequence of months with excessive potential evapotranspiration. Spatial and temporal climatic variability and alterations in water balance parameters have serious implications on human activities in water related sectors. Climate change exerts remarkable effects on river flows, groundwater recharge and other biophysical components of the water resource base, and demands for that resource. The consequences, or impacts, of such changes on risk or resource reliability depend not only on the biophysical changes in stream flow, recharge, sea-level rise and water quality, but also on the characteristics of the water management system. Possible changes in water resources and demand will impact on water supply, flood risk, power generation, navigation, pollution control, recreation, habitats and ecosystems services in the absence of planned adaptation to climate change. Luckily, some suggestions are proffered for effective water resource management in the face of increasing effects of climate change.

Hydroclimatological approach to sustainable water resources management

513

REFERENCES Desanker, P. & Magadza, C. 2001. Africa, In Mc Carthy, J.J. Canziani, O.F., Leary, N.A. Dokken, D.J. & White, K.S. (eds), Climate Change 2001: Impacts, Adaptation and Vulnerability: 489– 531. U.K. Cambridge University Press. Hastenrath, S. 1990. A generalized classification of South African summer rain-bearing synoptic systems. Journal of Climatology 4, 547–560. Issar, A.S. 1998. Climate, history and water. Work in Progress, United Nations University 15(2):16–17. Lamb, P.J. & Peppler R.A. 1992. Further case studies of tropical Atlantic surface atmospheric and oceanic patterns associated with sub-Saharan drought. Journal of Climatology 5:476–488. Magadza, C.H.D. 2000. Climate change impact and human settlements in Africa: prospects for adaptation. Environmental Monitoring, 61:193–205. Nakayama, M. 1998. Water: The 21st Century’s Oil? Work in Progress, United Nations Univ., 15(2):18–19. Nicholson, S.E. & Entekhabi, D. 1986. The quasi-periodic behaviour of rainfall variability in Africa and its relationship to Southern Oscillation. Journal of Climate and Applied Meteorology 34:331–348. Olofin, E.A. 1985. Climatic constraints to water resource development in the Sudano-Sahelian Zone of Nigeria. Water International 10(1):29–37. Skoftland, E. 1995. Water resource management in southern Africa—a vision for the future. Conf. of SADC Ministers responsible for Water Resources Management, 23–24 Nov, 1995. Pretoria, South Africa. Umoh, U.T. 1999. Climatic variability and rainfall-streamflow relationships in parts of the Niger River Basin of Nigeria. Africa Climatological Research Series Vol. 3:10–22. United Nation’s Environment Programme (UNEP) 1987. The changing atmosphere. Environmental Brief No. 1. UNEP.

Impact of cultivation practices on multiple uses of water in the Alemaya catchment, eastern Ethiopia Y.E.Woyessa School of Civil Engineering and Built Environment, Technikon Free State, South Africa A.T.P.Bennie Department of Soil, Crop and Climate Sciences, University of the Free State, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Soil degradation and water pollution are the most important environmental issues facing most of the developing world today. Across Ethiopia, there has been considerable degradation of land, water and biodiversity resources, which has, and will continue to have major impacts on rural communities’ economic production and natural ecosystems. Most of the productive topsoil in the highlands of Ethiopia has been degraded, resulting in chronic food shortages and persistent poverty. Serious erosion is estimated to cause soil degradation on some 25% of the highlands of Ethiopia. As a result of upland erosion, siltation of dams, lakes and reservoirs is becoming a common phenomenon. A field study was conducted in the catchment of Lake Alemaya, eastern Ethiopia, with the objective of evaluating the effect of land cultivation practices on runoff and soil loss, and its possible impact on the siltation of Lake Alemaya. The result showed that it was possible to reduce runoff and soil loss from the cultivated agricultural fields and subsequently reduce sediment load deposited in the Lake by using proper crop residue management on the agricultural fields.

1 INTRODUCTION Soil degradation and water pollution are the most important environmental issues facing most of the developing world today. Soil erosion has long been recognized as a negative attribute of agriculture and pastoralism. It is evident that erosion is partly a natural

Impact of cultivation practices on multiple uses of water

515

process and cannot be avoided entirely. The objective should be to minimize erosion within economic limits rather than trying to eliminate it. It is useful to consider the short and long term impact of erosion. Short-term effects occur within a season, such as damage to infrastructure, loss of crop area, lower yields due to less water. Longer-term effects are generally manifested over decades, such as an accelerated nutrient decline, lower plant available water capacity and siltation of dams, lakes and low lying wetlands. In general, poor catchment management leads to short term effects, while the long term impact of poor catchment management results in downward spiral phenomena, which occurs when longer term degradation effects make management in the short term difficult. For example, higher runoff leads to lower yield and soil cover, which increases erosion, thus accelerating the degradation cycle. The considerable degradation of land, water and biodiversity resources across Ethiopia will continue to have major impacts on rural communities to produce economically and to conserve natural ecosystems. Most of the productive topsoil in the highlands of Ethiopia has been degraded, resulting in poor yields which enhance chronic food shortages and persistent poverty. Serious soil erosion occur on an estimated 25% of the area, and some estimates suggest that 4% of the highland is so seriously eroded that it will not be economically productive again in the foreseeable future (SCRP, 1996). As a consequence of this land degradation, the production capacity of the soils of the Ethiopian highlands is calculated to be declining at a rate of 2–3% annually (Hurni, 1993; cited by Zeleke, 2000). This is a potential threat to national food security if allowed to continue uncontrolled and every effort should be made to reverse the situation. Sonneveld & Keyzer (2003) investigated national agricultural revenues under alternative scenarios of soil conservation, land accessibility and technology. Their results showed that without soil erosion control, agricultural production will stagnate in future resulting in distressing food shortages, while rural incomes will drop dramatically below the poverty line. Even with the adoption of modern technology, soil conservation remains essential, especially over the long term. The vicious cycle of poverty and environmental degradation is the result of interrelationships between the socioeconomic aspects of farmers and land degradation. The socioeconomic situation of farmers determines their capability to implement environmentally viable natural resource conservation measures. These include farming practices and farmers’ attitudes toward rational use of resources. A study conducted by Daba (2003) in the Hararghe highlands of eastern Ethiopia showed that farmers’ perception about the danger of gully erosion is significantly related to aspects like severity of water scarcity, location of the farm within the landscape and the literacy level of farmers. Willingness of farmers to adopt a new or improved soil and water conservation practices is determined by fertilizer availability as an incentive and the education of farmers. Daba (2003) concluded that for the Hararghe highland conditions, design and implementation of soil and water conservation measures should be based on farmers’ priorities such as addressing the alleviation of the water scarcity problem. During the early 1980’s the Government of Ethiopia, through the Soil Conservation Research Project (SCRP), placed a very high priority on stopping land degradation and restoring natural resources. Some of the key aims of the SCRP were to generate information on runoff and erosion processes under different agro-climatic conditions and land use practices for application of proper planning and design of soil and water

Water resources of arid areas

516

conservation practices. It was underlined that such an effort would enhance and promote sustainable development and protect the environment for future generations. Achieving these outcomes is essential to revitalize rural communities and maintain sustainable economic growth, and to ensure that our natural biodiversity is maintained for future generations. It is believed that this can be achieved through integrated catchment management. Catchment management should be based on a partnership between community and Government. Planning and implementation of natural resource management programs should create opportunities for community engagement. Management of natural resources should recognize the linkages between land and water and that the management of one can impact on the other. The catchment of Lake Alemaya is located in the Eastern Hararghe Region of eastern Ethiopia, and is characterized by intensive agriculture and a high population pressure. The average landholding per family varies between 0.5 to 1.7 hectares, which puts a high pressure on the utilization of land resources, such as soil and water. The Lake Alemaya provides the local population with domestic water supply as well as water for livestock and irrigation purposes. However, due to an increase in population pressure and the intensive agriculture practiced in the catchment in general and around Lake Alemaya in particular, the capacity of the Lake Alemaya to store water is decreasing at an alarming rate in the past few years, threatening the water supply to nearby towns. The main reason for this situation is reported to be the accumulation of silt in the Lake due to continuous erosion from the farmlands in the catchment. The aim of this paper is to present results of an experiment conducted in the Alemaya catchment, eastern Ethiopia, with the objective of evaluating the effects of cultivation practices on the catchment hydrology, with especial emphasis on runoff and soil loss from cultivated fields, and its possible impact on the siltation problem of Lake Alemaya. 2 MATERIALS AND METHODS The experimental site was located within the Lake Alemaya catchment on the Alemaya University Campus, which is situated 550km east of Addis Ababa, the capital city of Ethiopia. The climate is characterized as “Dry Weyna Dega” zone (Hurni, 1986), with an altitude of 1960 meter above sea level and a mean annual rainfall of about 800mm. The mean daily minimum and maximum temperatures are 10.1°C and 23.6°C respectively. The rainfall pattern is bi-modal with a short rainy season from March to May and a long rainy season during the months of July, August and September. The soil of the study area is characterized as a Regosol with good internal drainage. The particle size distribution of the topsoil, determined using the standard Pipette procedure as described by Day (1965), is clayey with 45.1% clay, 22.2% silt and 32% sand. The mean bulk density of the soil varies from 1150 to 1280kg m−3. In this experiment, three tillage practices, namely no-tillage (NT), traditional tillage (TT), and conventional tillage (CT) (mouldboard ploughing as a primary tillage) each with four levels of wheat (Triticum aestivum L.) residue cover, namely 0t ha−1, 2t ha−1, 4t ha−1, and 8t ha−1 as treatments were compared. Traditional tillage is the land preparation method with a shallow tillage tool called “Maresha” pulled by a pair of oxen. The

Impact of cultivation practices on multiple uses of water

517

experiment was laid out in a split plot design with tillage practices as main plots and residue cover rates as sub-plots. The size of each plot was 4 meters wide and 6 meters long. All treatments were replicated twice. Following the tillage treatments the calculated amounts of wheat residue were spread evenly over each plot. The experimental plots were laid out on a field with homogenous soil and uniform slope of 5–6%. A corrugated iron sheet border was placed around each plot to measure runoff and soil loss. Pieces of corrugated iron sheet were carefully attached to one another by rivets. The 400mm corrugated iron sheets were inserted 200mm deep into the soil in order to prevent run-on and runoff to and from the plots. At the down slope side of each plot a trough made of iron sheet was installed which collected the runoff in a barrel with a capacity of 0.2m3 (200 liters). During the course of the experiment growth of weeds within the plots was chemically controlled by spraying an herbicide (glyphosate). During the second year (summer 2001) of the experiment, the same experimental plots were used with the same layout and replication. The tillage operation was conducted after carefully removing the corrugated iron sheet borders and the troughs. After completion of the tillage operations the metallic borders and troughs were re-installed and fresh wheat straw was spread on each plot at the required rates. Runoff amount and sediment mass were collected in a barrel for each plot. The volume of runoff in each barrel was measured after every runoff producing rainfall event by inserting a steel tape into the barrel and measuring the depth of water. The depth of runoff in the barrel was converted into volume. The collected runoff was mixed thoroughly after which a sample was taken to determine the sediment concentration, where after the barrel was emptied and cleaned. Rainfall was recorded with an automatic rain gauge situated in the middle of the experimental area. In addition to the controlled experiment on runoff plots, visual observation at selected sites in the Alemaya catchment and informal discussions with small groups of farmers were conducted in order to collect information on farming practices and visible land degradation symptoms, such as erosion gullies. 3 RESULTS AND DISCUSSION 3.1 Runoff The average total runoff for the three tillage and four rates of residue treatments, during the main rainfall seasons of both experimental years (2000 and 2001), are given in Figure 1. Figure 1 illustrates that residue rates of 41 ha−1 and 81 ha−1 had effectively eliminated runoff from all three tillage treatments for both years. Although there was no statistically significant effect of tillage practice on total runoff, the average total runoff from the three tillage practices was in the order of NT>TT>CT. Tillage affected runoff only during the first few storms of the season, where runoff from no-tillage was significantly higher than from the freshly tilled traditional and conventional tillage plots, before the impact of raindrops destroyed the roughness created by tillage operations to form a surface seal or crust. Generally, the effect of no-tillage on runoff is reported to take a long time to

Water resources of arid areas

518

Figure 1. Average total runoff during the main rainfall seasons of (a) 2000 and (b) 2001, from three tillage treatments and four rates of residue cover. Note: R0=0t ha−1, R1=2t ha−1, R3=4t ha−1 and R4=8t ha−1 of wheat residue cover. influence certain physical and hydraulic properties of the soil. For instance, Dickey et al. (1989) reported that between 5 to 6 years is required for changes in soil physical properties under notillage to become measurable, resulting in higher infiltration rates and lower runoff. In another study, Voorhees & Lindstrom (1984) reported that 3 to 4 years are required before conservation tillage developed a more favorable porosity in the upper 0 to 150mm depth of soil.

Impact of cultivation practices on multiple uses of water

519

The residue mass was converted to percentage surface residue cover using a measurement technique as described by Lang & Mallett (1982). Accordingly the residue masses of 2, 4 and 8t ha−1 corresponded to 62%, 76% and 92% of residue cover respectively. The mean total runoff from the three tillage practices were related to the percentage residue cover for 2000 and 2001 rainfall seasons (Fig. 2). Figure 2 illustrates the decrease in runoff with an increase in residue cover. The linear decline in total runoff with increasing residue cover for the first year (Fig. 2a) resulted mainly from the high runoff measured at the 2t ha−1 or 62% residue cover during a high intensity storm of the 12th of August 2000. When the runoff from the 12th of August storm is not included in the curve fitting, the function changes to an exponential form similar to the 2001 data depicted in Figure 2b. It was observed that for most of the storms a residue rate of 2t ha−1 reduced runoff significantly compared with the bare plots. At higher intensity storms, such as the one on the 12th of August

Water resources of arid areas

520

Figure 2. Mean total runoff (Qu) as a function of percentage residue cover during the main rainfall seasons of (a) 2000 and (b) 2001, based on mean runoff from the three tillage practices. 2000, which reached intensities of more than 200mm/hr, a residue rate of 2t ha−1 was insufficient and residue rate of 4t ha−1 was required to sufficiently reduce runoff on this soil and rainfall conditions. Roth et al. (1988) recommended that at least 4–6t ha−1 of mulch is needed to reduce runoff and erosion effectively. However, other reports by Lattanzi et al. (1974) and Unger et al. (1991) indicated that much lower residue amounts could greatly reduce erosion because surface residue reduces soil loss much more than it reduces runoff. In another report, Freebairn et al. (1993) suggested that a cover level of 30% appears to be a critical for erosion control. In this experiment a wheat residue rate of 2t ha−1 was equal to 62% surface cover. The results from this study indicate that residue cover is more important in reducing runoff than the effect of type of tillage practice on this particular soil and under this climatic condition. It has been established by several natural and simulated rainfall studies that surface cover reduces soil erosion more than any other factor in tillage management (Freebairn & Wockner, 1986; Gilley et al., 1986a; Sallaway et al., 1988). Allowing the residue to remain on the soil surface could significantly reduce the amount of surface crusting thereby increasing infiltration (Hillel, 1980). Soil surface characteristics usually govern water entry into the soil during rainfall events. The

Table 1. Soil loss from erosive storms on different dates from three tillage practices and four levels of residue treatments at the Alemaya University experimental site during the main rainfall seasons of 2000 and 2001. Soil loss (t/ha) Residue rate (t/ha) 0 2 4 8 # # Tillage 2000 2001 2000 2001 2000 2001 2000 2001 NT 5.434 2.398 2.058 0.162 0.360 0.012 0.000 0.013 TT 3.813 1.106 1.705 0.334 0.211 0.169 0.000 0.044 CT 4.819 1.703 2.626 0.198 0.111 0.071 0.000 0.016 Mean* 4.689a 1.745a 2.130b 0.231b 0.227c 0.088bc 0.000c 0.025c *Mean numbers for the same year followed by the same letter(s) are not significantly different at 1% probability level. # Experimental years.

surface cover reduces erosion by reducing the runoff volume through stubble protecting the soil surface, thus reducing aggregate breakdown and compaction of the soil surface by raindrop impact (Edwards, 1982; Freebairn et al., 1993).

Impact of cultivation practices on multiple uses of water

521

3.2 Soil loss The total soil losses from the three tillage practices and four rates of residue cover for both experimental years are presented in Table 1. During the first year of the experiment, soil losses occurred mainly from the bare plots of the three tillage practices. There was little soil loss from plots with a residue rate of 2t ha−1 and higher for all the erosive storms, with the exception of the intense storm of the 12th of August 2000, when a residue rate of 2t ha−1 was insufficient to restrict soil losses. Plots with a residue rate of 4t ha−1 had little soil loss except during the last few storms. The ones with residue rate of 8t ha−1 had no soil loss during any of the storms. An intense storm on the 12th of August 2000 caused the highest soil loss from the bare plots of three tillage treatments. When compared to the total soil loss for the main rainfall season of the first year, the single storm of the 12th of August was responsible for 48.7%, 73.7%, and 76.2% of the total soil loss from the bare plots of NT, TT and CT respectively. Rainfall intensity is an important factor controlling soil erosion since interrill soil erosion varies with the square of rainfall intensity (Meyer, 1981). The kinetic energy of a raindrop is also related to the rainfall intensity. For example, the infiltration rate of a soil was reported to decrease sharply with time, as the kinetic energy and intensity of a storm increased, thus increasing runoff and erosion (Karen, 1990). The effect of the different residue rates in general, and the 2t ha−1 rate in particular on soil loss, varied depending on the type of tillage practices and the storm characteristics. From the CT there was no soil loss during the first two storms, probably due to the effect of ploughing that increased infiltrability due to looser topsoil, creation of surface roughness and small depressions. A study by Freebairn et al. (1993) suggested that roughness associated with tilled soils could result in less runoff and erosion compared with NT under certain circumstances. Lindstrom & Onstad (1984) reported that primary tillage operations, such as moldboard ploughing, increased infiltration thereby reducing the danger of erosion by increasing soil porosity and establishing channels or voids in the surface soil layer that conduct water into the soil profile. Kinnel (1996) indicated that differences in tillage practices had no major impact on either sediment concentration or runoff amount, on soils that maintain its tillage-induced surface roughness under rain after cultivation. The soil loss from CT, at residue rates of 0t ha−1 and 2t ha−1, was higher than from TT and NT at a similar residue rates during the last two storms of 2000. This may be the result of tillage induced roughness and depression storage being destroyed by the impact of raindrops of the previous storms, especially the intense storm of the 12th of August that might have created a surface crust on the bare plots. Rao et al. (1998a, b) reported that the effectiveness of soil management practices in reducing runoff and erosion on Alfisols was dependent on the resistance of the soil to the formation of crusts. The mechanical breakup of crusts by tillage is soon lost with the formation of another surface crust after a few rainstorms. During the second year (2001), the soil loss was generally very low from all the tillage and residue cover treatments compared with the first year (Table 1). The highest soil loss from a single storm was from the bare plot of CT, which was 1.06t ha−1 compared to 3.67t ha−1 during the previous year. The rainfall amounts per single storm event during the second year were higher than the previous year but the rainfall intensities were lower.

Water resources of arid areas

522

This was believed to be one of the reasons for the lower soil loss per storm during the second year. For six of the eight storms that caused erosion during 2000 and 2001, there were significant differences in soil loss only between the bare and 2t ha−1 residue rates. Among the different residue rates there were no significant differences. For the other two storms, there were significant differences between 0 and 2t ha−1, and between 2 and 4t ha−1 residue rates. From this it can be concluded that a residue rate of 2t ha−1 will be sufficient to control erosion from this soil, except for the occasional high intensity storms where a higher residue rate of 4 to 6t ha−1 will be preferred. The decrease in the total soil loss per season, representing the mean of the tillage treatments, with increasing percentage residue cover, is presented in Figure 3 for both seasons. The decrease in soil loss was linear for the first year, with a reduction of 0.053t ha−1 for every 1% increase in the percentage residue cover. The decline was non-linear for the second year with a more rapid decline with an increase of residue cover from 0 to 60%. Linear relationships between soil loss and the percentage residue cover have also been reported by some authors (Singer et al., 1981; Kinnel, 1996; Cruse et al., 2001), while others have reported non-linear (inverse exponential or polynomial functions) relationships between the two parameters (Gilley et al., 1986a; Papendick et al., 1990; Freebairn et al., 1993). The mean total soil losses averaged over the four rates of residue cover, for the three tillage treatments, for both years, are given in Figure 4. It shows that the mean seasonal total soil loss was the highest for NT, followed by CT. Traditional tillage had the lowest soil loss for both seasons. These differences, however, were not statistically significant. Despite the higher sediment concentration from CT compared with the other two tillage practices, the soil loss from CT was lower than from NT. The soil loss from a given area is the product of the sediment concentration and runoff amount. Despite the lower sediment concentration, the higher runoff generated from NT resulted in a higher soil loss than from CT. Lindstrom & Onstad (1984) warned that soil erosion could be a serious problem on NT when insufficient residue cover is present to reduce the runoff flow velocity. In another study, Myers & Wagger (1996) reported that residue cover did not substantially reduce runoff from NT treatments but consistently decreased soil loss. Overall, the type of tillage practice had little effect on the total soil loss except on the bare NT plots that had the highest erosion for all the storms. When NT is practiced on this type of soil, care should be taken that a residue cover of at least 2t ha−1 wheat stubble or 62% residue cover should be maintained at all times. The beneficial effect of tillage was generally limited to the first few erosive storms until the tillage induced roughness and depression storage disappeared. It is possible that two years were not long enough to detect the differences due to tillage practices. The effect of a tillage practice is reported to take several years to change certain physical and hydraulic properties of the soil. For instance, Dickey et al. (1989) reported that between 5 and 6 years is required under notillage for changes in soil physical properties, resulting in higher water intake, to become measurable. Voorhees & Lindstrom (1984) reported that 3 to 4 years were required before conservation tillage had a more favorable porosity in the upper 0 to 150mm depth of the soil.

Impact of cultivation practices on multiple uses of water

523

3.3 Impact of erosion on the siltation of Lake Alemaya Lake Alemaya is the main source of water supply for domestic water consumption for the nearby town of Harar which has a population of about 300,000 and other small villages. It is also a source

Figure 3. Mean total soil loss (SL) from three tillage practices as a function of percentage residue cover for the main rainfall season of (a) 2000, and (b) 2001.

Water resources of arid areas

524

of irrigation water. Two factories in the city of Harar (Harar Brewery and Hamaressa Edible Oil Mill) which previously depended on their water supply from Lake Alemaya since their establishment had to seek for alternative ground water supply 30km away due to the unreliability of water supply from the Lake. The whole catchment area drains into Lake Alemaya. Unsustainable intensive agricultural practices coupled with a high human population pressure are among the factors that enhance the degradation of land and water in the catchment. The size of Lake Alemaya is shrinking tremendously both in depth and in surface area. Some of the original Lake area is now grazing land for the nearby livestock owners. The Lake, which used to supply the local population with multiple uses of water, is no longer a dependable source of water. Although there could be other factors contributing to the present condition of the Lake, it is believed that unacceptably high erosion from agricultural lands in the catchment is mainly responsible for the siltation and the eventual dry up of some parts of the Lake. Gullies as deep as 5 meters and more are running into the Lake. In some cases the density of gullies is alarmingly high with more than 10 gullies running into the Lake over a distance of just two kilometers.

Figure 4. Mean total soil loss, averaged over rates of residue cover, from three tillage treatments based on average of the two seasons. The general agricultural practice in the area is based on a traditional tillage method which uses a tool called “Maresha” and pulled by a pair of oxen. The main crop grown in the area is a leafy bushy crop, locally known as chat (Catha edulis) used as a human stimulant, together with cereal (maize and sorghum) as intercropping. Tefera (2003) reported that 54% of the cropland is allocated to chat production in Alemaya District. It was also reported that the majority of irrigated land is allocated to chat production and that it utilizes most of the scarce organic manure in farm households. Chat is mainly

Impact of cultivation practices on multiple uses of water

525

grown for the export market to neighbouring countries such as Somalia, Djibouti and some Arab countries such as Yemen. During the dry season when irrigation is very crucial for growing of this cash crop, one kilogram of good quality chat can sell for up to 400 Ethiopian Birr (1US$=8.5 ET Birr) on the export market, which is a huge income for chat growing farmers. In the year 1999/2000, Ethiopia earned 618.8 million birr in hard currency by exporting 15,684 Metric ton of chat (Tefera, 2003). This market incentive encourages farmers to withdraw more and more water for irrigation from Lake Alemaya. Farmers in the surroundings of the Lake use several small pumps in series arrangements to withdraw water from the Lake to irrigate fields as far as to the top of the catchment boundary. At the end of every harvest season no crop residue is left on the agricultural fields due to the fact that the biomass is harvested and the multiple uses of the remaining crop residue, such as for fuel, construction, animal feed, etc., exposing the bare soil surface to the erratic and sometimes very aggressive rainfall. It is at this time of the season, before establishing crops in the field, that most of the erosion occurs. Extrapolation of the results from this study to a catchment level, with the present land use practices where insignificant amount of crop residue remains in the fields, supports a scenario of continuous sediment buildup in the Lake. The total area of the Alemaya catchment draining into the Lake Alemaya is estimated at 1680 hectares. The average soil loss from the bare traditional tillage plots at the experimental site was 2.46t ha−1. If we assume this to be the average annual soil loss for the catchment, the total amount of soil running into the Lake will be 4133ton. This will be the amount for the main rainfall season of July to September, whereas additional erosion will also occur during the short rainy season for the months of March to May with substantial amounts of rainfall. With an average bulk density of 1260kg m−3 for the soil of the experimental site, the total volume of soil deposited in the Lake will be approximately 3280m3 per annum. This is sufficient to reduce the depth of the Lake annually by about 33mm over an area of 10 hectares. The benefit of crop residue on the surface of the soil in reducing runoff and soil loss was demonstrated in this experiment. The presence of wheat residue at a rate of 2t ha−1 was sufficient to reduce soil loss to a minimum level. This shows that it is technically possible to reduce the quantity of sediment deposited in the Lake with proper management of crop residue on the farm lands in the catchment. However, the high population pressure and the nature of subsistence farming in the area, which includes both crop and livestock production (termed as a mixed farming system), puts a lot of pressure on both land and water resources. Under the present circumstances it appears that the use of crop residue and other practices for soil and water conservation is not a priority issue for the small farmers in the area. 4 CONCLUSION The catchment of Lake Alemaya is characterized by an intensive use of agricultural land and a high population pressure. Land degradation due to water erosion is so serious that the existence of Lake Alemaya is being threatened currently due to siltation of the Lake. Results from this experiment have shown that it is technically possible to reduce runoff

Water resources of arid areas

526

and soil loss from agricultural lands through proper crop residue management by retaining at least 65 to 70% crop residue cover on the agricultural fields. However, this objective is difficult to achieve under subsistence farming conditions in the catchment where the main crop is grown for its biomass yield leaving little residue behind and where there are also other multiple uses for crop residue, such as for fuel, animal feed and construction purposes. These are the most pressing and immediate needs of subsistence farmers in the catchment, whereas erosion control seems by far less important to them. Therefore, investigation into the causes and possible remedies to land and water degradation at a catchment level requires a holistic and an integrated approach taking into account the socioeconomic circumstances of farmers and ensuring active participation of the community in the process of planning and implementation of an integrated catchment management approach. REFERENCES Cruse, R.M., Roberto, M. & Mize, C.W. 2001. Surface residue effect on erosion of thawing soils. Soil Sci. Soc. Am.J. 65:178–184. Daba, S. 2003. An investigation of the physical and socioeconomic determinants of soil erosion in the Hararghe Highlands, eastern Ethiopia. Land Degradation and Development. 14(1):69–81. Day, P.R. 1965. Particle fractionation and particle size analysis. In: C.A.Black (ed.) Methods of soil analysis. Amer. Soc. of Agron., Madison, WI. pp. 545–567. Dickey, E.E., Eckert, D.J., Larson, W.E., Johnson, R., Mannering, J., Kinsella, J., Wickner, I. & Cruse, R.M. 1989. To till or no to till during drought. J. Soil Water Conserv., 44:117–120. Edwards, W.M. 1982. Predicting tillage effects on infiltration. In: P.W.Unger and D.M.Van Doren (eds), Predicting tillage effects on soil physical properties and processes: ASA special publication, 44. Am. So. Agron. and SSSA, Madison, WI. pp. 105–115. Freebairn, D.M. & Wockner, G.H. 1986. A study of soil erosion on vertisols of the eastern Darling Downs, Queensland. I. Effect of surface conditions on soil movement within contour bay catchments. Aus. J. Soil Res. 24:135–158. Freebairn, D.M., Loch, R.J. & Cogle, A.L. 1993. Tillage methods and soil and water conservation in Australia. Soil Tillage Res. 27:303–325. Gilley, J.E., Finker, S.C., Spomer, R.G. & Mielke, L.N. 1986. Runoff and erosion as affected by corn residue: Part I. Total losses. Transactions of the ASAE.Vol. 29(1):157–160. Hillel, D. 1980. Application of soil physics. Academic Press, New York. 385 p. Hillel, D. 1971. Soil water: physical principles and processes. Academic press, New York. Hurni, H. 1986. Guidelines for development agents on soil conservation in Ethiopia. Community Forests and Soil Conservation Development Department, Addis Ababa, Ethiopia. 100 p. Karen, R. 1990. Water-drop kinetic energy effect on infiltration in sodium-calcium-magnesium soils. Soil Sci. Soc. Am. J. 554:983–987. Kinnell, P.I.A. 1996. Runoff and sheet erosion from tillage trials under artificial rainfall at Harden, New South Wales. Aust. J. Soil Res. 34:863–877. Lang, P.M. & Mallett, J.B. 1982. The effects of various methods of primary tillage and seedbed preparation upon the maintenance of surface residues in a maize monoculture system. Crop Production. 11:55–58. Lattanzi, A.R., Meyer, L.D. & Baumgardner, M.F. 1974. Influence of mulch rate and slope steepness on interrill erosion. Soil Sci. Soc. Am. Proc. 38:946–951. Lindstorm, M.J. & Onstad, C.A. 1984. Influence of tillage systems on soil physical parameters and infiltration after planting. J. Soil Water Conserv. 39:64–68.

Impact of cultivation practices on multiple uses of water

527

Meyer, L.D. 1981. How rain intensity affects interrill erosion. Transactions of the ASAE. 24:1472– 1476. Myers, J.L. & Wagger, M.G. 1996. Runoff and sediment loss from three tillage systems under simulated rainfall. Soil Tillage Res. 39:115–129. Papendick, R.I., Parr, J.F. & Meyer, R.E. 1990. Managing crop residues to optimize crop/livestock production systems for dryland agriculture. Advances in Soil Science. 13:253–272. Rao, K.P.C., Steenhuis, T.S., Cogle, A.L., Srinivasan, S.T., Yule, D.F. & Smith, G.D. 1998a. Rainfall infiltration and runoff from an Alfisol in semi-arid tropical India. I. No till systems. Soil Tillage Res. 48: 51–59. Rao, K.P.C., Steenhuis, T.S., Cogle, A.L., Srinivasan, S.T., Yule, D.F. & Smith, G.D. 1998b. Rainfall infiltration and runoff from an Alfisol in semi-arid tropical India. II. Tilled systems. Soil Tillage Res. 48: 61–69. Roth, C.H., Meyer, B., Frede, H.G. & Derpsche, R. 1988. Effect of mulch rates and tillage systems on infiltrability and other soil physical properties of an Oxisol in Parana, Brazil. Soil Tillage Res., 11:81–91. Sallaway, M.M., Lawson, D. & Yule, D.F. 1988. Ground cover during fallow from wheat, sorghum and sunflower stubble under three tillage practices in central Queensland. Soil Tillage Res. 12:347–364. Singer, M.J., Matsuda, Y. & Black, J. 1981. Effects of mulch rate on soil loss by raindrop splash. Soil Sci. Soc. Am. J. 45:107–110. Soil Conservation Research Project, 1996. Data base report (1982–1993). Watershed Development and Land Use Planning Department, Ministry of Agriculture, Addis Ababa, Ethiopia. Sonneveld, B.G.J.S. & Keyzer, M.A. 2003. Land under pressure: soil conservation concerns and opportunities for Ethiopia. Land Degradation and Development. 14(1), 5–23. Tefera, T.L., Krieten, J.F. & Perret, S. 2003. Market incentives, farmers’ response and policy dilemma: A case study of chat production in the eastern Ethiopian highlands. Agrekon, Vol. 42(3), 213–227. Unger, P.W., Stewart, B.A., Parr, J.F. & Sing, R.P. 1991. Crop residue management and tillage methods for conserving soil and water in semi-arid regions. Soil Tillage Res. 20:219–240. Voorhees, W.B. & Lindstrom, M.J. 1984. Long-term effects of tillage method on soil tilth independent of wheel traffic compaction. Soil Sci. Soc. Am. J. 43:152–156.

Geochemical evidence and origin of salinity in the shallow basinal brine from the Makgadikgadi Pans Complexes, northeastern Botswana L.N.Molwalefhe Department of Geology, University of Botswana, Gaborone, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: This study explores the compositional variability and geochemical controls of the stable isotopes of oxygen (δ18O), hydrogen (δ2H), and carbon (δ13C) and the δ18O-δ2H-chloride dynamics in brine to constrain the sources of salinity. The δ18O-δ2H-chloride relationships show that the brine is a mixture of a highly evaporated meteoric water and highly saline fluid of magmatic nature. Somewhat, these observations are at variance with the one-stage climate-driven scenario typically used to explain the high salinities in the Makgadikgadi brine. The conclusion of a deeper source is also consistent with the geology and is in accord with recent seismic studies that indicate the area is structurally active.

1 INTRODUCTION Inland salt lakes are common in southern Africa, and many occur in Namibia, Botswana and South Africa (Lancaster, 1978; Shaw, 1988; Seaman et al., 1991). All inland lakes in southern Africa are shallow, and most are ephemeral with salinities that are less than 50g/L (Seaman et al., 1991). This study was conducted at Sua Pan, part of the Makgadikgadi Pans Complexes of northeastern Botswana (Figure 1). The Makgadikgadi pans are 200km from east to west and 120km from north to south and occupy a depression that forms the center of a basin of internal drainage, the lowest point of which is approximately 890m above sea level (masl) located at the northeastern corner of Sua Pan. There are two major pans, Ntwetwe to the west and Sua to the east, and numerous small pans, making up the Makgadikgadi Pans Complexes. The investigated site is located in the northern part of Sua Pan between latitude 20°22′S to 20°27′S and longitude 25°54′E to 26°06′E. The aim of this research is to conduct a focused stable isotope-based study to investigate the sources of salinity in groundwater brines from Sua Pan. Apart

Geochemical evidence and origin of salinity in the shallow basinal brine

529

from explaining the processes that control the formation of brines in the Makgadikgadi, this study will also contribute towards bridging of the critical gaps in our knowledge regarding the causes of salinity over a wide range of geologic environments. 2 GEOLOGY/HYDROGEOLOGY The geological history of the Makgadikgadi pans is tied to the tectonic and magmatic events that have affected the area since the Jurassic times. The rocks surrounding the Makgadikgadi Pans belong to the Kalahari Beds and consist of aeolian sands, sandstones and various fluviatile and lacustrine deposits. The thickness of the Kalahari deposits varies from zero in the eastern and southern margins of the pans to over 100m north of Sua Pan (Gould, 1986). Similar intrusions that occur in the central parts of Botswana are dated to be coeval (187 Ma) with the major flood basalt that is widespread in southern Africa (Smith, 1984). All the rocks

Figure 1. Locality map of the Makgadikgadi Pans Complex and study area, Botswana.

Water resources of arid areas

530

around the pans are highly fractured by post-Karoo faulting and dyke emplacement (Baillieul, 1979). Historical water level data (Gould, 1986) showed that groundwater flow is largely responsive to the topographic slope of the base of the aquifer rather than to changes in the configuration of the water table. The base of the aquifer is very irregular (Figure 2), probably owing to some structural control, and dips consistently from the southeast toward the northwest (Gould, 1986). Pumping tests carried out by Paulsen (1971) defined the aquifer conditions to be confined to semi-confined. 3 MATERIALS AND METHODS Groundwater samples were collected from the medium grained sand aquifer using a network of high capacity boreholes and were analyzed in the laboratory for their chemical and isotopic content. The main tracers that were analyzed for were the stable hydrogen and oxygen isotopes and the major element chemistry of the brine. The water temperature, specific conductance and pH measurements were determined at the wellhead using appropriate electrodes. Titrimetric determinations of alkalinity were not accomplished in the field due to the high concentration of and in the samples. Samples for anion and cation analyses were collected unfiltered in high-density polyethylene (HDPE) bottles. Samples for the analyses of cations were acidified to pH2 or lower using high purity nitric acid to prevent metals from precipitating during storage. The samples were refrigerated at 4°C during storage. For the stable isotopes of oxygen and hydrogen, water samples were collected in 20ml glass vials with inverted cone closures to minimize headspace and potential evaporation (Gat, 1996).

Geochemical evidence and origin of salinity in the shallow basinal brine

531

Figure 2. Hydrogeological crosssection of the northern part of Sua Pan showing the various hydrostratigraphic units beneath the Pan; the upper confining clays, sand-hosted brine aquifer, and lower confining sandstone/basalt unit. Concentrations of anions (Cl−, Br− and ) and cations (Na+, K+, Ca2+ and Mg2+) were analyzed using a Dionex DX-500 ion chromatograph (Dionex, Sunnyvale, CA, USA) equipped with a conductivity detector and an AS40 auto sampler. All samples were diluted 500-fold using ultra-pure de-ionized water to bring the concentrations within the calibration range of the standards. The stable oxygen isotope (δ18O) was analysed following the widely used and established technique of isotopically equilibrating water with pure CO2 in a constant temperature environment (Epstein and Mayeda, 1953). In this procedure, 2ml of water sample is equilibrated with 0.5 atmospheres of CO2 for 25 h at 25°C. In this study, all groundwater samples were concentrated chloride brines and had to be vacuum distilled at 900°C (Molwalefhe, 2003) before reaction with zinc (Coleman et al., 1982). Chlorides have been shown to inhibit the metal-water reaction (Yang et al., 1996). The results for δD and δ18O isotope ratios are reported on the V-SMOW scale. The analytical precision for δ18O measurements is 0.2‰ and for δD is ±2‰.

Water resources of arid areas

532

4 RESULTS AND DISCUSSIONS The Makgadikgadi brine is mainly of sodium-chloride composition but is also moderately rich in carbonates, sulphate and potassium and distinctly poor in the divalent cationic species i.e. Mg2+ and Ca2+ (Table 1). Models that describe the chemical evolution of dilute waters as they evolve toward concentrated solutions demonstrate the importance of solute fractionation processes that commonly result in brines dominated by just a few chemical species. The final brine composition is determined by the composition of the precursor solution (Hardie and Eugster, 1970) and in-lake geochemical processes (Herczeg and Lyons, 1991) that enrich the water in certain ions relative to others. Herczeg and Lyons (1991) showed that by increasing the carbon dioxide partial pressure (pCO2) in water from 1 to 50 atmospheres delays calcite saturation, which could lead to instances of supersaturation with respect to gypsum (CaSO4). Precipitation of gypsum will suppress the calcium concentrations to low levels and cause the DIC to invariably increase on evaporation

Table 1. Comparisons of DIC concentrations and water chemistries from other alkaline saline lakes and the makgadikgadi system. All concentrations are in mg/L. 2=Jones et al. (1977), 3=Jankowski and Jacobson (1989), 4=Gould (1986) and Current. Mono L. L. Woods3 L. Makgadikgadi4 1 2 3 lake Magadi (Australia) Werowrap Pans (Australia) (USA) (Kenya) HCO3/CO3 15710 44000– 105720 pH – 9.85 Na – 40500– 130000 K – 661– 2690 Cl – 15000– 98700 – SO4 Brine type – Na(K)CO3-ClSO4

28100

10527

20000

– 2600

– 13076

9.6 70000

1500

868

2250

1500

13864

85000

5800 Na(K)CO3-ClSO4

4 Na-ClHCO3

9900–15000 Na(K)-Cl-CO3HCO3(SO4)

before calcite saturation could be reached. This raises the pH to 8 or more. Considering the low pCO2 of the atmosphere, it is difficult to account for high DIC in the Makgadikgadi brine using a single-stage mechanism of evaporative concentration of rain or surface water alone. A combination of occasional flooding of the pans followed by fast evaporation prior to infiltration, and contacting with a high pCO2 environment in the

Geochemical evidence and origin of salinity in the shallow basinal brine

533

subsurface is a more likely scenario to generate the chemical composition of the Makgadikgadi brine. The water compositions of the low DIC lakes is dominated by Na(Mg)Cl or Na(Mg)Cl(SO4) and this shows that not all saline lakes show high DIC concentrations. Saline lakes that are similarly enriched in DIC as the Makgadikgadi brine (Table 1) have associated volcanic bedrocks, suggesting the importance of source water control on lake chemistry. Excess alkalinity may also result due to evaporative enrichment of solutes at or near the surface. This would require relatively high initial concentration of DIC in the inflow water, e.g. rainfall. Such high concentrations may therefore be unique to particular areas. The remains of high DIC in the Makgadikgadi brine may indicate evidence for near surface accumulations of carbon of magmatic origin. Excess alkalinity from high pCO2 may strip calcium from the brine, enhance silicate weathering, production of silicic acid, and precipitation of magnesium as sepiolite. The carbon composition in the Makgadikgadi brine may approximate that of magmatic systems; hence, the DIC is likely to be of magmatic origin and potentially, the underlying basalt is the source. 4.1 Stable oxygen and hydrogen isotopes 18

2

The δ O and δ H are a measure of concentrations of the stable isotopes of oxygen and hydrogen found in water molecules. Brine samples from Sua Pan are compared with the long-term (1962 to 2000) meteoric water line (Figure 3) constructed from weighted mean precipitation values for the IAEA station located in Pretoria, South Africa. The meteoric water line is defined by the relationship δ2H=6.2δ18O+5.3 and brine isotopic compositions form a group below this line. The observed deviation of the brine compositions from the meteoric water line shows that evaporation has occurred between rainwater and groundwater. When water evaporates from an open lake, the amount of evaporation that has taken place can be estimated from water balances using δ18O and δ2H isotopes. The evolution of δ18O and δ2H toward heavier compositions during evaporation of lakes can be described by a Rayleigh-type equation; δ18OGW=ε18Ototal*ln(f)+δ18Oprec δ18Oprec=−3: ε18Ototal (1) @30°C=16: f=0.001 δ18OGW=−19*ln(0.001)=+113‰

Water resources of arid areas

534

Figure 3. Long-term meteoric water line (Pretoria) and brine compositions (Makgadikgadi).

Figure 4. δ2H versus δ18O relationship for the Sua Pan brine. Filled circle is the average rainwater composition, filled diamond is the average brine composition and filled triangles are individual brine compositions.

Geochemical evidence and origin of salinity in the shallow basinal brine

535

By using the average δ18Oprec value for the Pretoria meteoric water (−3‰), average δ18OGW of +3‰ (see Figure 4) and assumed humidity of 50%, the calculated amount of water, f, that remains is 69%. This suggests that about 31% evaporation has taken place to cause the shift in the brine compositions. The above calculation does not provide a realistic conclusion that the salinities in the brine are a result of evaporating surface water by only 31%. If equation 1 is considered, and a concentration factor of 1000 (f=0.001) is applied to surface waters to produce the salinities, the isotopic values for the remaining water can be calculated to determine if isotopic values that could be produced in the brine. A calculated value of +113‰ is obtained which is much different from the isotopic value of the brine (+3‰). Therefore, another source might be contributing water that is isotopically light and more saline. In Figure 5, the meteoric water equation is applied on the actual brine compositions to provide a reconstruction of the expected isotopic composition in the brine. Values that are expected because of evaporating rainwater fall along a line, and indicate heavier compositions compared to the measured values. Assuming the brine represents evaporated meteoric water, then the isotopic compositions should range from +17 to +32‰ for δ2H. It is clear from Figure 5 that

Figure 5. Expected isotopic compositions of the brine (filled circles) reconstructed by superimposing actual values (filled triangles) on the meteoric water equation.

Water resources of arid areas

536

Figure 6. δ18O versus chloride for samples that fall along a line of mixing. the δ2H values have drifted to lighter compositions. Since exchange of hydrogen with the rocks is a less common phenomenon, mixing with an isotopically lighter end-member in the subsurface is a more realistic process. Magmatic waters are usually depleted in hydrogen. Estimated isotopic compositions for magmatic systems fall between −40 to −80‰ for δ2H and +6 to +9‰ for δ18O (White, 1974; Campbell and Larson, 1998) and it is likely that mixing with isotopically lighter magmatic water has taken place. 4.2 Stable oxygen and chloride concentration The relationship between δ18O and chloride for the samples shows two groupings; one falls along a line of δ18O-chloride enrichment (Figure 6) and the other samples form a group that shows mostly isotopically enriched samples (not shown) from the surface that are less concentrated with regard to chloride. This shows that evaporating surface water acquires limited salinity before infiltration into the subsurface environment, then mixing with a more isotopically light groundwater in the subsurface. Such infiltration might have taken place over geologic timescales. The compositions of the incoming water that is isotopically heavier due to evaporation but relatively dilute with respect to chloride might have changed in response to interactions with the subsurface. The fact that the heavier values of hydrogen and oxygen from the surface are not maintained in the aquifer signifies that there are other processes involved in concentrating solutes than just evapoconcentration at the surface. 5 CONCLUSIONS The geo-chemical and isotopic studies of the hyper-alkaline Na-Cl brine from the northern part of Sua Pan in the Makgadikgadi area have indicated that the brine has likely evolved along various geochemical pathways and underwent more than one stage of in-

Geochemical evidence and origin of salinity in the shallow basinal brine

537

lake modifications via a sequence of surface and subsurface processes. The study demonstrated the need for a multiple-stage process to account for the high salinities in the brine. Chemical data also suggest that surface evaporative processes have played an equally significant role in concentrating salts in the brine while DIC concentrations showed that the brine has imprints of a magmatic source. This model indicates that the evolution of the Makgadikgadi brine is partially maintained by underground discharge and partially by meteoric recharge. REFERENCES Baillieul, T.A. 1979: Makgadikgadi Pans Complex of central Botswana. Geological Society of America Bulletin, Part II, v.90 pp. 289–312. Campbell, A.R. and Larson, P.B. 1998: Introduction to stable isotope applications in hydrothermal systems. In: Techniques in hydrothermal ore deposits geology, Society of economic geologists, Reviews in economic geology v.10, pp. 173–193. Coleman, M.L., Shepard, T.J., Durham, J.J., Rouse, J.E. and Moore, G.R. 1982: Reduction of water with zinc for hydrogen isotope analysis. Anal. Chem., 54:993–995. Coplen, T.B., Herczeg, A.L. and Barnes C. 2000: Isotope engineering—Using stable isotopes of the water molecule to solve practical problems: In Environmental Tracers in Subsurface Hydrology. P.Cook and A.L.Herczeg (Eds) Kluwer Academic, Dordrecht, pp. 79–110. Clark, I.D. and Fritz, P. 1997: Environmental Isotopes in Hydrogeology. Lewis Publishers, New York, pp. 112–123. Epstein, S. and Mayeda, T. 1953: Variations of 18O content of waters from natural sources. Geochim. Cosmochim. Acta, 4:213. Gat, J.R. 1996: Oxygen and hydrogen in the hydrologic environment. Ann. Rev. Earth Plan. Sci. 24:225–262. Gonfiantini, R. 1986: Environmental isotopes in lake studies. In: Handbook of environmental isotope geochemistry, Vol 2, eds. P.Fritz and J.-C.Fontes, Elsevier, Amsterdam. pp. 113–168. Gould, D. 1986: Brines of Sowa Pan and Adjacent Areas, Botswana. In Mineral Deposits of Southern Africa Vols I and II, Anhacusser C.R. and Maske S. (Eds) pp. 2289–2299. Lancaster, I.N. 1978: The pans of the southern Kalahari, Botswana. Geographical Journal 144:80– 98. Reeves, C.V. 1972: Rifting in the Kalahari? Nature 237:95–96. White, D.E. 1974: Diverse origins of hydrothermal ore fluids. Economic Geology 69:954–973. Yang, W., Krouse, H.R. and Spencer, R.J. 1996: Stable isotope analyses of microlitre quantities of water from inclusions in halite and concentrated brines. Chem. Geol (Isot. Geosc. Sect.) 130:139–145.

Theme F: Vulnerability and risk

Decision support for optimal water system planning: a Wadessy case study A.A.Ilemobade School of Civil and Environmental Engineering, University of the Witwatersrand, South Africa D.Stephenson Department of Civil Engineering, University of Botswana, Gaborone, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The unsustainability of small water reticulation networks (WRNs) in many Southern African communities continues to be a reality. This is significant in light of the efforts by governments to provide a minimum quantity of potable water to its citizenry at the least possible cost while constrained by suboptimal systems, tight budgets and inadequate skills. Presented herein is a generic WRN design algorithm within a decision support software called Wadessy (an acronym for Water decision support system). Wadessy comprises a suite of computer programs that facilitate the optimal design and operation of small water distribution systems. The validity of Wadessy is proved by comparison with a well-known example from literature and a practical case study in Selebi-Phikwe, Botswana. Wadessy achieved a 1,91% cost saving in comparison to the example, and a 32,52% cost saving (about 1497190,00 Pula) for the Selebi-Phikwe WRN based on October 2001 pipe costs.

1 INTRODUCTION The design of water distribution systems (WDSs) has received a great deal of attention because of its importance to industrial growth and water’s crucial role in society for health, fire-fighting, and quality of life, particularly in light of increased urban development and water use (Sherali et al 1998:1381). WRNs are essential components of all WDSs as they convey potable water from source, pump station or storage to the consumers. The cost of these networks may amount to as much as 60% of the entire water supply scheme (Sarbu & Borza 1997; Stephenson, 1998:49) and as a result, operation and maintenance costs may soar higher if networks are ill designed (Ilemobade

Water resources of arid areas

542

& Stephenson 2002:80). WRNs also account for the largest costs in municipal maintenance budgets (Sherali et al 1998:1381). Despite often scarce resources, national and local governments are obligated to provide this resource. Since WRNs are composed mostly of pipes, pipe-sizing decisions have become critical during the design of cost effective WDSs capable of handling varied demand loadings and satisfying minimum pressure head requirements. Optimisation techniques have been proven to facilitate the best WRN designs. 2 FORMULATION OF Wadessy’S MODEL The basic equations of continuity, conservation of energy and pressure head difference are utilised in modelling WRNs. To arrive at an optimal solution, an iterative simulationoptimisation algorithm is employed. Efficient hydraulic simulation (both static and dynamic) in Wadessy is based on modelling the WRN using the above-mentioned equations and determining the unknown variables using the established Newton-Raphson iterative procedure on simultaneous equations generated using the nodal method (Cornish 1939). Pipe sizes (which are initially assumed for new WRN designs) and other pipe parameters, consumer demands, network layout configuration, pump constants and fixed grade node (FGN) elevations are known prior to simulation. The Choleski Decomposition technique (Stoer & Bulirsch 1993) is employed to generate the matrix for computing node pressures. Based on either the Darcy-Weisbach or Hazen-Williams pipe equations, continuity is checked at each network node. If a violation exists, a correction factor is introduced into the procedure and the entire simulation process is repeated. Output from the simulation include pipe flows and orientation, pipe headlosses, friction factors, node residual pressure heads, draw-off at each source node, pumping heads and valve head losses. Wadessy’s design optimisation procedure is adapted from Featherstone & El-Jumaily’s (1983) model, which is based on the concept that a hypothetical linear hydraulic gradient, So for a balanced WRN exists by which the initial network design can be iteratively corrected to produce optimal pipe sizes and an optimal relation between each pipe. In addition to Featherstone & El-Jumaily’s (1983) model, the effects of hydraulic surfaces in determining optimal designs were previously undertaken by Deb & Sarker (1971), Wu (1975) and Alperovits & Shamir (1977). Ilemobade & Stephenson (2003:3) discuss Featherstone & El-Jumaily’s WRN optimisation concept, the work published by the workers mentioned above and Wadessy’s optimisation model. 3 MODEL VALIDATION 3.1 An example network In validating the LPG model, Alperovits & Shamir (1977) present an example network shown in Figure 1. The operation of the network is studied under peak and night flow conditions. Basic network data and costs (in arbitrary units) are shown in Tables 2 and 1 respectively (Alperovits and Shamir 1977:892). Minimum permissible pressure head for

Decision support for optimal water system planning

543

each demand node is 30m. Since several workers have used the results presented by Alperovits & Shamir (1977) to validate their models, a comparison of the results obtained by Alperovits & Shamir (1977), Farmani et al (1999) and Wadessy are presented in Tables 3 and 4. The optimal solution achieved by Wadessy was obtained after running the optimisation procedure from several starting network designs. Flow distribution is treated as a variable until an optimal solution is reached. A 1, 5% deviation from the minimum residual pressure (30m) was permitted during computations, hence a computed residual pressure head of 29, 62m for Node 6 (the critical node). The global optimum network design cost computed by Wadessy is $262000,00

Figure 1. Example network. compared to $267113,00 and $268000,00 computed by Alperovits & Shamir (1977) and Farmani et al (1999) respectively. Wadessy thus achieves a cost saving of 1,91% from Alperovits & Shamir’s (1977) solution and 2,24% from Farmani et al’s (1999) solution. The cost savings achieved by Wadessy is despite the fact that Wadessy provides only one pipe size per link (discrete links) in contrast to Alperovits & Shamir (1977) who provide multiple pipe sizes per link (segmental links). The tendency towards a branching network may be seen as distance from each FGN increases and as demand at a node, especially during peak flow, is supplied from both sources (Ilemobade & Stephenson 2003:6). In a practical network of this magnitude, a 1 inch pipe, as recommended for Pipe 4, would be meaningless. By replacing with a 2 inch pipe, the WRN costs $265000,00: a saving of 0,79% and 1,12% from the solutions presented by Alperovits and Shamir (1977) and Farmani et al (1999). Changes in residual pressures and flows (Table 4) due to this replacement are insignificant: a pressure of 29, 61m was computed for the critical node (Node 6). 3.2. A case study: Selebi-Phikwe, Botswana Wadessy was employed to analyse the existing Selebi-Phikwe WRN in order to determine the performance of the existing network (Figure 2 and Table 6) on the addition of three

Water resources of arid areas

544

new residential developments. During analysis, Wadessy was also used to hypothetically determine the optimal WRN design for Selebi-Phikwe in comparison to that existing, based on October 2001 consumer demands and pipe costs (Table 5). Average peak and night flows calculated for October 2001 were 0,270m3/s and 0,135m3/s respectively. 20Ml of storage is provided in 3 concrete cylindrical tanks situated at the southeast section of Selebi-Phikwe. Minimum and maximum node pressure heads are 15m and 90m respectively (WUC, 1995). Minimum pipe sizes to connect users to the reticulated mains and on which fire hydrants will be located are 63mm & 75mm respectively (WUC, 1995). Three pumps are installed in the existing pumpstation; one of which acts as standby. The polynomial equation: Hp=AQ2+BQ+C adequately represents pumping head within the WRN. For two pumps in parallel: A=−414, 94; B=−3,50 & C=88, 93. For pumps operating individually: A=−1660, 30; B=−12, 31 & C=90,04.

Table 1. Pipe cost data. Pipe diameter 1 2 3 4 6 8 10 12 14 16 18 20 (in.) Unit cost per m 2 5 8 11 16 24 32 50 60 90 130 170 ($)

Table 2. Pipe and node data. Pipe Length CHW Range of Node Elevation Node Consumption ref. (m) allowable ref. (m) minimum (m3/hr) diameters residual Peak Night (in.) pressure flow flow head (m) 1 1000 130 0–20 1 210 0 −420,00** −300,00** 2 1000 130 0–20 2 150 30 100,00 0,00 3 1000 130 0–20 3 160 30 100,00 0,00 4 1000 130 0–20 4 155 30 120,00 0,00 5 1000 130 0–20 5 150 30 270,00 0,00 6 1000 130 0–20 6 165 30 330,00 0,00 7 1000 130 0–20 7 160 30 200,00 0,00 8 1000 130 0–20 8 195 0 −700,00** 300,00 9 100 130 0–20 **Negative consumption represents supply to the node (i.e source node).

Decision support for optimal water system planning

545

Table 3. Comparison of pipe optimisation output for Figure 1 from three models. Alperovits & Shamir (1977) Wadessy (discrete links) (segmental links) Peak Night Farmani et al Peak flow Night flow flow flow (1999) (discrete links) Pipe St En Len Dia Len Di Head Head Len Di Len Diam. Flow, Head Flow ref. art d gth 1++ m. gth am. loss Loss gth am. (in.) (m3/hr)+ loss (m3/hr)+ ++ No No (m) (m) (in.) 2 (in.) (m) (m) (m) (in.) gth de de (m) (m)

He ad loss (m)

1 1 2 33,89 10,00 966,08 12,00 8,99 4,82 1000 12,00 1000 14,00 748,08 13,77 486,00 6,20 2 2 3 374,46 5,00 625,51 8,00 10,91 3,12 1000 8,00 1000 8,00 214,56 22,60 173,16 15,21 3 2 4 1000,00 10,00 0,00 0,00 5,82 7,65 1000 10,00 1000 12,00 433,44 10,87 312,84 5,94 4 4 5 816,14 6,00 183,85 8,00 5,39 3,96 1000 1,00 1000 1,00 0,72 18,81 0,72 24,49 5 4 6 999,97 10,00 0,00 0,00 0,19 5,08 1000 8,00 1000 10,00 312,84 14,84 312,12 14,75 6 6 7 999,99 14,00 0,00 0,00 2,12 0,99 1000 12,00 1000 10,00 −17,28 0,07 312,12 14,75 7 3 5 929,62 6,00 70,36 4,00 0,30 8,49 1000 10,00 1000 8,00 114,48 7,07 173,16 15,21 8 7 5 825,67 10,00 174,28 8,00 7,32 2,11 1000 12,00 1000 10,00 154,80 4,04 −173,88 5,01 9 7 8 100,00 16,00 0,00 0,00 0,54 0,11 100 14,00 100 14,00 −372,24 0,31 486,00 0,51 WRN pipe cost $267113,00 $268000,00 $262000,00 + Negative flow represents flow in the direction opposite to the initial flow orientation; ++ Total length of pipe link between two nodes is the sum of Length 1 plus Length 2.

Table 4. Comparison of node optimised output for Figure 1 (Farmani et al (1999) did not publish pressure head values for their optimised WRN for Figure 1). Alperovits & Shamir (1977) Minimum Pressure head (m) residual Peak flow Night flow pressure Node ref. head (m) 1 2 3 4 5 6 7 8

0,00 30,00 30,00 30,00 30,00 30,00 30,00 0,00

0,00 Not Available 30,10 40,20 39,80 30,00 37,10 0,00

Not Available Not Available Not Available Not Available Not Available Not Available Not Available Not Available

Wadessy Pressure head (m) Peak flow Night flow 0,00 70,33 37,73 54,46 40,66 29,62 34,69 0,00

0,00 77,90 52,69 66,96 47,48 42,21 32,47 0,00

Water resources of arid areas

546

Figure 2. Existing & Selebi-Phikwe water reticulation network (skeletonised). Table 5. Commercially available pipe sizes and costs for Selebi-Phikwe (October 2001). Diam. 63 75 (mm)

90 100 110 150 160 200 250 300

350 400

450

500

Class – 12+ – 12 12+ 12+ – 12 – 12 12+ 12+ 12+ 12+ a – 161,0 183,0 200,0* 220,0* 230,0* Cost/m – 11,4* – 15,8 25,7* 27,4 – 29,4 uPVC Class 9 9 9 9 9 – 9 9 6 9+ – – – – a – – – – Cost/m 8,0 11,4 2,4 7,2 25,7 – 45,5 43,1 169,0 161,0* aP ipe costs are in Botswana Pula; +Class of pipe assumed; *Pipe cost per meter assumed as actual cost not available; AC represents Asbestos Cement pipes while uPVC represents Unplasticised Poly Vinyl Chloride pipes. AC

Table 6. Selebi-Phikwe WRN pipe and node data. Consumption (m3/hr) Node Length Diameter Pipe Node Elevation Peak Night Min. ref. (m) (mm) Material* ref. (m) flow flow residual pressure head (m) 1 2 3

2950,00 1800,00 310,00

300,00 AC 110,00 uPVC 100,00 AC

1 2 3

842,70 842,70 852,13

−705,96 −486,00 0,00 0,00 0,00 0,00

0,00 15,00 15,00

Decision support for optimal water system planning

547

4 850,00 200,00 uPVC 4 868,90 89,42 0,00 5 600,00 200,00 uPVC 5 868,90 89,42 0,00 6 650,00 110,00 uPVC 6 876,52 0,00 0,00 7 450,00 250,00 uPVC 7 861,13 0,00 0,00 8 800,00 250,00 uPVC 8 862,81 39,37 0,00 9 550,00 110,00 uPVC 9 861,13 0,00 0,00 10 1000,00 200,00 uPVC 10 856,71 44,11 0,00 11 680,00 300,00 AC 11 862,81 77,76 0,00 12 560,00 250,00 uPVC 12 868,90 48,60 0,00 13 880,00 300,00 AC 13 876,52 73,87 0,00 14 4050,00 300,00 AC 14 853,66 40,82 0,00 15 1470,00 300,00 AC 15 856,71 32,04 0,00 16 800,00 300,00 AC 16 862,81 7,78 0,00 17 780,00 300,00 AC 17 852,13 36,00 0,00 18 2450,00 200,00 AC 18 838,42 0,00 0,00 19 1080,00 200,00 uPVC 19 848,17 93,31 0,00 20 250,00 160,00 uPVC 20 852,13 55,40 0,00 21 2330,00 160,00 uPVC 21 853,66 0,00 0,00 22 800,00 200,00 AC 22 895,73 −266,04 486,00 23 2000,00 200,00 uPVC 23 868,90 0,00 0,00 24 200,00 300,00 AC 24 859,76 0,00 0,00 25 180,00 450,00 AC 25 859,76 5,22 0,00 26 600,00 500,00 AC 26 861,28 19,19 0,00 27 680,00 500,00 AC 27 838,41 18,47 0,00 28 930,00 450,00 AC 28 838,41 28,80 0,00 29 460,00 350,00 AC 29 868,92 16,52 0,00 30 1560,00 300,00 AC 30 874,00 0,00 0,00 31 740,00 300,00 AC 31 870,00 16,52 0,00 32 1000,00 300,00 AC 32 876,52 0,00 0,00 33 800,00 300,00 uPVC 33 876,52 35,57 0,00 34 20,00 400,00 AC 34 861,28 38,39 0,00 35 540,00 300,00 AC 35 861,28 64,80 0,00 36 1170,00 300,00 AC 37 820,00 160,00 uPVC 38 950,00 110,00 uPVC 39 700,00 100,00 uPVC 40 800,00 75,00 uPVC 41 2420,00 75,00 uPVC 42 2180,00 75,00 uPVC 43 510,00 160,00 uPVC 44 2270,00 160,00 uPVC 45 510,00 200,00 AC 46 1085,00 200,00 AC *AC and uPVC represent Asbestos Cement and Unplasticised Poly Vinyl Chloride pipes respectively.

15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 0,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00 15,00

Water resources of arid areas

548

Detailed results of the optimisation are presented in Figures 3, 4 and 5. Figures 4 and 5 present node residual pressure head results for the existing and Wadessy optimised WRN based on peak and night flow conditions. It can easily be seen that although both network designs are adequate under night flows, the existing network design is deficient under peak flows where the minimum permissible node pressure head (15m) is violated in nodes 4, 6 and 28 (9, 64m, 13, 42m and

Figure 3. Existing and Wadessy optimised Selebi-Phikwe pipe reticulation network.

Figure 4. Existing and Wadessy optimised node pressure head results for peak flows.

Decision support for optimal water system planning

549

Figure 5. Existing and Wadessy optimised node pressure head results for night flows. −6,64m respectively)—see Figure 4. This is because pipes 2 and 42 in the existing Selebi-Phikwe WRN (Figure 3) generate large headlosses (26,44m and 82,84m respectively) during peak flows, as their capacities are small (110mm and 75mm respectively) in relation to the high demands at their downstream nodes (Node 4, 6 and 28). Pipe 42 supplies the Botswana Defence Force camp, which currently experiences low or no flow during peak demand periods. Better flow conveyance and reduced headlosses in pipe 42 is enhanced in Wadessy’s optimal design by the provision of a 100mm uPVC pipe. Based on October 2001 prices, it is estimated that 1497190,00 Pula (about 32,52%) could have been saved from the 4603914,00 Pula estimated to have been spent on the installation of the existing pipe network alone. 4 CONCLUSIONS Wadessy utilises efficient algorithms to facilitate effective decision-making in the planning, design, analysis, and operation of small WDSs. In this paper, Wadessy’s WRN design algorithm is validated by applying it to a popular design example first proposed by Alperovits & Shamir (1977), and to the existing Selebi-Phikwe WRN. By permitting a 1,5% deviation from the minimum node pressure head of 30 metres, a 1,91% cost saving was achieved using Wadessy’s design algorithm in comparison to that determined by Alperovits & Shamir (1977). A cost saving of 32,52% (1497190,00 Pula) was also achieved by using Wadessy’s design algorithm in comparison to the cost of the existing Selebi-Phikwe WRN based on October 2001 prices. By minimising the changes in node pressure heads during optimisation, the optimal interaction between each network pipe was enhanced.

Water resources of arid areas

550

5 SOFTWARE PROCUREMENT Wadessy’s WRN design software may be obtained from the author at a minimal fee. ACKNOWLEDGEMENT The authors are indebted to the Richard Ward Endowment Fund (fund no. TW CIVN ILEM) for funding this research. REFERENCES Alperovits, E. & Shamir, U. 1997. Design Of Optimal Distribution Systems. Water Resources Research, 13(6):885–900. Cornish, R.J. 1939. The Analysis of Flow in Network of Pipes. Journal Institute of Civil Engineers, 13:147. Deb, A.K. & Sarkar, A.K. 1971. Optimisation in Design of Hydraulic Network. Journal of the Sanitary Engineering Division. Proceedings Paper 8032, ASCE, 97(SA2):141–159. Farmani, R., Matthew, R.G.S. & Javadi, A.A. 1999. Discrete Optimisation of Water Distribution Networks using Genetic Algorithms. In D.A.Savic & G.A.Waters (eds.) Water Industry Systems: modelling and optimisation applications (Vol 2), Water Engineering Management Series, Research Studies Press, England: 427–436. Featherstone, R.E. & El-Jumaily, K.K. 1983. Optimal Diameter Selection for Pipe Networks. Journal of the Hydraulic Division, ASCE, 109(HY2):221–234. Ilemobade, A. & Stephenson, D. 2002. Optimally Upgrading Small Water Reticulation Networks using Wadessy: Case Study—Selebi-Phikwe, Botswana. Proceedings, International Conference on Water—The Lifeblood of Mankind and 5th Biennial Congress of the African Division of the IAHR in collaboration with the IET. Arusha, Tanzania. 11–13 December:79–88. Ilemobade, A.A. & Stephenson, D. 2003. Generic Optimisation of Small Water Reticulation Networks using Wadessy. Journal of the South African Institution of Civil Engineers. Paper 548. 45(4):2–9. Sarbu, I. & Borza, I. 1997. Optimal Design of Water Distribution Networks. J. of Hydraulic Research, 35(1): 63–79. Sherali, H.D., Totlani, R. & Loganathan, G.V. 1998. Enhanced Lower Bounds for the Global Optimisation of Water Distribution Networks. Water Resources Research, 34(7):1831–1841. Stephenson, D. 1998. Water Supply Management. Water Science and Technology Library. Kluwer Academic Publishers. Stoer, J. & Bulirsch, R. 1993. Introduction to Numerical Analysis. Texts in Applied Mathematics 12. 2nd Ed., Springer-Verlag Inc, New York: 180. Wu, I.P. 1975. Design of Drip Irrigation Main Lines. Journal of the Irrigation Drainage Division. Proceedings Paper 11803, ASCE, 101(IR$):265–278. WUC—Botswana Water Utilities Corporation, 1995. Design Standards, DS 0295. Gaborone, 22 February.

The importance of constructing a correct conceptual model for an aquifer G.van Tonder, I.Dennis & D.Vermeulen Institute for Groundwater Studies, University of the Free State, Bloemfontein, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Intensive groundwater investigations in Southern Africa are usually limited by budgets and the complex fractured nature of the majority of the aquifers. Furthermore groundwater is usually ‘invisible’ to humans except for springs and seepage faces. The paper focuses on the problems associated with the development of an inappropriate conceptual model when solving groundwater problems. A number of other problems will be highlighted for example problems with slug-test interpretation in a fractured rock environment; incorrect contouring of water levels; incorrect monitoring of groundwater pollution and incorrect calculation of aquifer parameters. Illustrations will be presented where incorrect conceptual models have led to incorrect groundwater management.

1 INTRODUCTION Intensive groundwater investigations in Southern Africa are usually limited by budgets and the complex fractured nature of the majority of the aquifers. In every study the natural system is represented by a conceptual model. A conceptual model includes designing and constructing equivalent but simplified conditions for realworld problems that are acceptable in view of the objectives of the study and the associated management problems. The process of conceptualising and transferring the real-world situation into an equivalent system, which can then be used to solve groundwater problems is crucial. The following is included in a conceptual model: 1. The known geological and geohydrological features and characteristics of the area. 2. The static water levels/piezometric heads of the study area. 3. The interaction of the geology and geohydrology in the study area. 4. The boundaries of the study area.

Water resources of arid areas

552

5. A description of the processes and interactions taking place within the study area that will influence the movement of groundwater including aquifer parameters and recharge calculations. The paper focuses on a few of the problems associated with the development of an inappropriate conceptual model when solving groundwater problems and present some solutions to these problems. 2 SOME PROBLEMS ASSOCIATED WITH DEVELOPING CONCEPTUAL MODELS 2.1 General Problems associated with conceptual models can be the result of: ● Incorrect analysis of aquifer tests. ● Misunderstanding the results of aquifer tests.

Table 1. Hydraulic parameters for borehole UO5. Parameter

Value

T of formation* (m2/d) K of facture zone (m/d) T of fracture zone (m2/d) K of matrix (m/d) T of matrix (m2/d) * Average for fracture & matrix.

19 3600 576 0.17 3

● Water levels being incorrect as a result of interpolation techniques. ● Errors in monitoring of groundwater systems. These problems are normally magnified in the complex fractured rock aquifer systems found in South Africa and will be discussed in more detail in the following sections. 2.2 Determination of aquifer parameters 2.2.1 Transmissivity/hydraulic conductivity Using available slug-test interpretation methods to analyse a slug test in a fractured rock aquifer to estimate a transmissivity (T) or hydraulic conductivity value (K) is problematic. The estimated value is dependent on the length of flow (part of the aquifer in which flow occurs due to the slug input). If this length of flow is known, the estimated K-value is more representative of that of the fracture zone. By using the total thickness of the formation for the estimation, the K-value, and thus T-value, is not representative of the formation. This has been illustrated at the Campus Test Site at the University of the Free State, South Africa.

The importance of constructing a correct conceptual model for an aquifer

553

The Campus Test Site is underlain by a series of mudstones and sandstones from the Adelaide Subgroup of the Beaufort Group of formations in the Karoo Supergroup. There are three aquifers present on the Site: ● The first, a phreatic aquifer, which occurs within the upper mudstone layers on the Site. A confining layer of carbonaceous shale with a thickness of 0.5 to 4m separates this aquifer from the second aquifer. ● The second and main aquifer, occurs in a sandstone layer is between 8 and 10m thick, a bedding plane fracture occurs in this aquifer at approximately 21m bgl. ● The third aquifer occurs in the mudstone layers (more than 100m thick) that underlie the sandstone unit. The known parameters of a borehole UO5 at the Campus Test Site are listed in Table 1. By performing and analysing a slug test, a K of 17m/d (T=330m2/d) was obtained, which does not compare with that of the fracture, the matrix or the average for the fracture and the matrix. Performing constant rate discharge tests at different abstraction rates result in different yield predictions. Borehole M11 at the Meadhurst Test Site, Bloemfontein, South Africa is used to demonstrate this effect. Borehole M11 was drilled along a dolerite dyke with the main water strike 30m below the rest water level, which is situated at 22m below surface. The dolerite dyke was intersected at 28m below the rest water level. At 30m below the rest water level a water strike of 4L/s was encountered in the dolerite. Two constant rate tests were performed on M11—one at 3L/s and the other at 7L/s. Figure 1 shows the pumping test results, as well as the information on the water strikes. It is clear from Figure 1 that the results from the two abstraction rates produced different drawdown curves. The estimated T-value with the 3L/s abstraction rate is 170m2/d, while the estimated T-value for the 7L/s second abstraction rate is 20m2/d. At a rate of 7L/s, the fractures could not sustain the abstraction rate with the result that the estimated T-value of 20m2/d is the formation T-value, while the lower rate gives a Tvalue more representative of the fractures.

Water resources of arid areas

554

Figure 1. Pumping test results and water strike information of borehole M11. 2.2.2 Storativity values The estimated storativity (S)-value in a fractured-rock aquifer from a pumping test, if analysed with any analytical method, will be incorect. The reason for this incorrect estimate is that inappropriate conceptual model (mainly the Theis, Gringarten or Kazemi model) is used. These models generate S-values that are a function of the distance

The importance of constructing a correct conceptual model for an aquifer

555

between the abstraction and observation boreholes (the larger the distance, the smaller the estimated S-value). Neuman (1994, personal communication) gave the following possible explanation: Consider the rock to consist of nested storage ‘reservoirs’ comprising different scale fractures. At one end of the spectrum are a few large, permeable fractures occupying a small relative rock volume which therefore has small porosity and storativity. On the other end are many small, low-permeable fractures occupying a relatively large rock volume which therefore has large porosity and storativity. Close to the pumping well, pressure in the large fractures declines rapidly relative to its rate of decline in the small fractures. The latter therefore release a relatively large amount of water into the large conductive fractures due to a sizeable local pressure gradient between the small and large fracture reservoirs. Hence S is large. Far from the pumping well, the pressure gradient between the small and large fractures is relatively small. Therefore, water release from the small to the large fractures occurs very slowly. Most of the initial drawdown (in the large fractures) at a great distance is associated with water release from storage in the large fractures. Hence S is small. With time, local pressure differentials between the reservoirs stabilise and flow everywhere within a given radius approaches a steady radial pattern. Therefore, it could be expected that S should approach a uniform value representing both reservoirs. However, as the flow pattern is now essentially stabilised and close to steady state (even though absolute pressures may continue to decline), standard pumping tests may not reveal this fact: the flow is sensitive to S only at early times. If there were only two reservoirs with very different S-values, log-log time-drawdown curves close to the pumping well would exhibit a familiar dual-porosity time inflection (of the kind analysed by Neuman for unconfined aquifers). However, if there is a continuous hierarchy of such reservoirs with a more or less continuous local range of T- and S-values, such inflections cannot be seen. The early log-log time-drawdown behaviour would then just look like a regular Theis curve. Only long pumping tests would reveal deviations from this curve, but unfortunately, storage effects during late behaviour are usually masked by large-scale heterogeneities and boundary effects. An example of this incorrect estimate was illustrated by Brook (1990) for the Jwaneng Mine in Botswana. Initial S-estimates from pumping test analyses were in the order of 1×10−4 while model calibration after a number of years of monitoring showed a reliable S-value in the order of 1×10−2. 2.3 Contouring of groundwater levels Environmental phenomena (e.g. rainfall and the occurrence of groundwater) cover such vast areas, that it is not always possible to measure their associated variables at all relevant points in space and time. Interpolation is a method to obtain values for these variables at points where no measurements were taken. However, different interpolation methods (e.g. distance-weighted and Kriging) produce different contour maps. There are many situations in the environmental sciences where a given variable correlates with another one. For example, groundwater levels often follow the surface topography of the aquifer. If the latter variable (topography) can be sampled more frequently than the first one (groundwater levels), then one can use this information to

Water resources of arid areas

556

improve estimates of the first variable (groundwater). In other words, topography can be sampled more frequently than groundwater levels, thus this information can be used to improve estimates of groundwater levels. Bayesian interpolation is a method that uses this principle. A groundwater investigation of the Hendrina Power Station, South Africa, was to be conducted. For this study groundwater levels had to be generated. The distribution of available data for the study area is shown in Figure 2. Initially Kriging was used to generate water levels (see Figure 3). However, there is a correlation of 87% between topography and groundwater levels. Therefore Bayesian interpolation can be used to generate groundwater levels (Figure 3). Field studies have verified that Bayesian interpolation produced a more accurate water level map. 2.4 Groundwater monitoring Due to the fractured nature of South African aquifers, strict and correct groundwater monitoring needs to be implemented to ensure sustainable use of these aquifers. In addition the aquifer systems normally consist of numerous layers, and piezometers must be installed in each of these layers to monitor the piezometeric levels. Single samples from such boreholes do not indicate the nature of the water quality, therefore samples at desired depths must be taken. An example of the impacts of incorrect sampling is demonstrated from one of the coal mines situated in the Free State Province, South Africa. A hydrogeologist built a conceptual model based on incorrect measuring of water levels. He came to the following conclusions concerning the closure of the mine: ● The underground coal mine will flood and once the mine is flooded groundwater will discharge onto surface (decant) at between 7–3Ml/d.

The importance of constructing a correct conceptual model for an aquifer

Figure 2. Distribution of observed water levels.

Figure 3. Groundwater levels (mamsl) generated using (A) Kriging and (B) Baysian interpolation.

557

Water resources of arid areas

558

Later a second hydrogeologists conducted a similar study, however in this investigation the piezometric levels in the geological layers below and above the mine were recorded. In addition the water levels within various compartments of the mine were measured, based on the recorded levels, the hydraulic within the system could be determined and modelled. The results indicate that the mine is not going to decant. Similar discrepancies were noted in the water quality sampling. The management of the latter post-closure scenario is less costly than that of the former. The mine therefore saved millions by conducting accurate sampling. 3 CONCLUSIONS AND RECOMMENDATIONS Problems associated with conceptual models can be the result of: ● Incorrect analysis of aquifer tests ● Misunderstanding the results of aquifer tests ● Water levels being incorrect as a result of interpolation techniques ● Errors in monitoring groundwater systems. These problems are normally magnified in the complex fractured rock aquifer systems found in South Africa. Incorrect conceptual models can lead to poor management of aquifer systems. Therefore emphasis must be placed on accurately understanding groundwater systems. REFERENCES Brooke, M.C. 1990. The reliability of a groundwater model: The history of modeling the Jwaneng Northern Wellfield in Botswana. In: Karel Kovar (ed.). ModelCARE 90: Calibration and reliability in Groundwater Modelling. Proc. Intern. Conference, The Hague, 3–6 September 1990, IAHS Press. Neuman, S.P. 1994. Personal Communication. University of Arizona, Tuscon, USA.

Water resources development and risk assessment in mountain regions of Africa Helmut Scheuerlein University of Innsbruck, Austria Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Water resources development in Africa deserves and receives high priority, particularly in arid and semi arid zones and in areas of extreme vulnerability like mountain regions. The access to water is often the main limiting factor of economical development. At the same time water may also become one of the most serious threats to man and his property. Water resources development and risk assessment should always be observed together. In the paper, several examples from various regions of Africa are introduced and discussed, such as river engineering and expanding tourism in the High Atlas in Northern Africa, population expansion in the Usambara mountains in Tanzania causing top soil erosion and the unforeseen side-effects of reservoir construction in the semi arid hilly regions of the Tell Atlas.

1 INTRODUCTION At the Summit on Sustainable Development in Johannesburg water has been declared as the raw material of the 21st century and the access to safe drinking water as a basic human right. At the same conference special emphasis has been put on five key areas for action which have come to known as the WEHAB acronym standing for – Water and sanitation – Energy – Health – Agricultural productivity – Biodiversity and ecosystem management All of these five key areas are somehow related to water. Although the Johannesburg Summit had to deal with the whole world, in the list of the eight key outcomes of the conference one was specifically dedicated to Africa: Africa and NEPAD (New Partnership of Africa’s Development) were identified for special attention and support by the international community to better focus efforts to address the development need to Africa.

Water resources of arid areas

560

2 WATER RESOURCES DEVELOPMENT IN AFRICA Regardless whether African engineers in charge of water works are dealing with water resources exploitation or with protection against floods or droughts, their work will be governed by the given natural, social and economical conditions of their environment. Whatever their particular project is aiming at, the solution has to be oriented towards sustainability. Impediments on the way towards a sustainable solution may become effective through budgetary constraints, social requirements, insufficient expertise, unconsciousness of ecological coherence, ignorance of environmental aspects, etc. Very often, budgetary constraints make it necessary to apply for external financial support and the donors dictate conditions which may be justified in economically saturated societies in Europe or North America but may be questionable for the African country in question. Another problem which is also related to the notoriously limited budgets in Africa is the competition between water resources development and disaster prevention. Investments spent for a water resources development project can be expected to pay back as soon as the project is operational whereas money paid for mitigation measures with respect to disaster prevention never produces direct revenues. The benefit of this type of costs lays in risk reduction concerning potential hazards without any visible income. 3 HAZARD POTENTIAL AND RISK ASSESSMENT IN MOUNTAIN REGIONS Mountain water is world-wide estimated as an extremely valuable resource: It is available over long time sections of the year, it is less polluted than other water resources, and it “come on its own feet” as the oriental people say which means that no external energy is required to bring it to the place where it is needed. On the other hand, mountain water may also be a serious threat, a potential hazard, apt to cause disaster, to carry away property, destroy houses, devastate farmland, even to take life. Mountainous regions are zones of specific natural conditions concerning topography, geology and climate. Mountainous ecosystems are strongly specialized and extremely vulnerable as they have to survive in a harsh environment with limited supply of nutrients. Living conditions are not easy for human beings, too. In the past, mountain areas have been occupied by man rather slowly and with careful observation of the given natural conditions considering the potential hazards connected with settling in such areas. As long as the mountain valleys were not too densely populated and settling was restricted to safe places, land-use was restricted to pastures for cattle and to modest use of the forest for firewood and construction material for housing. At that time the ecological balance of use and regeneration of natural resources was more or less in equilibrium. Exploitation of natural resources and protection against natural disasters are part of the responsibility of water resources engineers. By trying to meet both demands, exploitation on one hand and protection measures against hazards on the other, conflicts are unavoidable. For the future, concepts have to be developed to find a balance between sustainable use of the natural resources and effective risk management.

Water resources development and risk assessment in mountain regions of Africa

561

4 EXAMPLES OF WATER-RELATED HAZARD POTENTIALS IN AFRICA 4.1 Watershed management in the High Atlas mountains (Morocco) The river systems at the northern slopes of the High Atlas (Tensift) are characterized by two significant specifications: – extraordinary climatic condition: extremely low flow during dry season (droughts) extremely high peaks at flood season (flash floods causing severe damages, Fig. 1) high erosivity of top soil in the catchment (sheet and gully erosion) – considerable human activities in the flood plains: increasing settling pressure (housing, tourism) extended road construction (using parts of the river bed, Fig. 2) inappropriate land-use (deforestation, erosion promoting crops) The consequences are events like torrential floods, debris flow and landslides, all of them increasing in number and intensity. The responsible water authorities try to cope with this situation the best they can by constructing river training works in the main river, check dams in the tributaries and by local attempts of soil conservation in the watersheds but their means (financial and manpower) as well as their political influence are limited compared with the tourism and road construction lobbies. Regarding the limited means, the efficiency of the

Figure 1. Flood damages at Oued Ourica, High Atlas, Morocco.

Water resources of arid areas

562

Figure 2. Reduction of river width due to human activities (Oued Ourika). measures necessary to cope with the existing problems must be increased through optimization wherever possible. The first step in this respect should be seen in careful analysis of the given situation and in genuine risk assessment. The second step should be to investigate on suitable measures concerning risk reduction. The third step then would be to decide upon priorities among the necessary measures to meet the available means (Scheuerlein, 2000a). 4.2 Watershed management and risk assessment in the West Usambara mountains (Tanzania) The Usambara mountains are situated in the north of Tanzania at about 100km distance from the Indian Ocean close to the Kenyan border. The Lwengera river divedes Usambarea in an eastern and a western part. West Usambara covers an area of roughly 2500km2 of steep mountainous catchment rising up to 2300m above the Massai Steppe. 65% of the territory is agricultural land, 20% is still covered with forest (up to colonial age almost all of the Usambaras was covered with forest). The climate of the Usambaras is determined by its high altitude and the Trade Wind Belt resultung in a main rainy season during October to January and a small rainy season in April and June. West Usambara is one of the most densely populated and most intensively used agricultural areas in Tanzania. The rapid population growth after liberation combined with landuse changes

Water resources development and risk assessment in mountain regions of Africa

563

Figure 3. Devastated cultivated land due to flash flood and debris flow in Usambara, Tanzania. pushed the forest boundary higher and higher up the mountains. Progressive use of steep slopes with unfeasible crops promoted soil erosion and degradation. Disastrous floods like one in January 1993, leaving 70 people dead and many hectares of agricultural land deteriorated must be seen in this context. The water resources situation in the Usambaras is determined by the relief of the mountain range characterized by: – cliff-like steep slopes all around the mountain range, rising up to 1000m high above the surrounding steppe (ca. 300msl) – moderately sloped plateaus at about 1300 to 1500msl, providing suitable conditions for agricultural activities and human settlement – steeply sloped catchments surrounding the plateaus rising up to peaks of more than 2000msl, partly covered with residual natural forest but gradually also subject to agricultural exploitation. Water resources development in West Usambara is affected by overgrazing and deforestation. Consequently, agricultural activity is gradually pushed further up to higher locations in the catchment towards the still existing forest using more and more steeply sloped areas for unsuitable crops with unfavorable cultivation methods (maize, up and down slope cultivation, etc.). These conditions result in high erosion rates in the steeper parts of the catchment. Together with the unevenly distributed precipitation this leads to disastrous consequences like flash floods debris flow as the one of 1993 which was caused by one single local storm event (Fig. 3).

Water resources of arid areas

564

In the Usambara mountains hazardous floods can not be avoided nor can the population pressure upon the land be released. Sustainability of water resources development can only be reached with a joint effort of various measures on varying levels, namely – on legal level: prohibition of deforestation and up and down slope cultivation at slopes steeper than an agreed value – on educational level: raining of local farmers with respect to soil conservation methods and appropriate farming practices – on technical level: stabilization of mountain streams in steep catchments by means of check dams, gully plugs, drop structures, sediment retention basins before the flow enters the plains, etc.

Figure 4. Erosion control through SECAP activities in Usambara. The measures dedicated to the legal level are partially already executed by respective laws (f.i. deforestation). In practice, however, these laws are frequently violated. Prohibition of up and down cultivation on steep slopes might have the same fate. However, in this case, contrary to deforestation any contempt of the law could easily be traced back to the responsible party.

Water resources development and risk assessment in mountain regions of Africa

565

The measures on educational level are certainly the most difficult and time consuming part. Fortunately enough this part is already well in progress since the early eighties of the past century. The respective conservation programme is called SECAP (Soil Erosion Control and Agroforestry Project), an initiative of the German foreign aid company GTZ (Scheuerlein et al, 1998). Meanwhile, the project has been handed over to the local authorities and is carried on by them with full responsibility (Fig. 4). The overall goal of SECAP is stabilization and ecological balance of all West Usambara watersheds. The project was introduced to the local people from the beginning as Farmers in the Western Usambaras apply ecologically adapted and economically sustainable farming systems. The meritorious activities of SECAP are a successful approach towards soil conservation and erosion control in the catchment area, however, the water courses draining the catchment are not included in this programme. Being aware of the fact that all sediment eroded in the catchment sooner or later ends up in one of the creeks, streams, or rivers of the watershed, it becomes obvious that erosion control can not ignore the necessity to rehabilitate and stabilize the water courses, too, particularly those incised in the steep slopes. Stabilization of water courses, however, requires engineering expertise. The problems to be solved and the mitigation measures which should be applied have been identified by Scheuerlein et al (1998) as − Stabilization of steep torrential streams to withstand destruction even when charged with highly sediment-laden flow Measures: Strengthening of existing step-pool systems or construction of new ones, preferably with natural local material; at very steep slopes, construction of consecutive gully plugs or drop structures. – Stabilization of riparian slopes potentially endangered of sliding into the streams Measures: Drainage of the slopes; construction of retaining walls; strengthening of the stream banks. – Protection of the plains of the plateau against devastation through debris and mud flow Construction of sediment retention basins by means of check dams or selfregulating barriers capable of automatic grain sorting. As the above mentioned measures are time consuming and costly, careful risk assessment must be carried out beforehand to allow for optimized selection and ranking of the measures.

Water resources of arid areas

566

Figure 5. Dry farming in the plateau areas of the Tell Atlas (Algeria). 4.3 Watershed management in the catchment of large reservoirs in the Tell Atlas mountains (Algeria) In the semi arid regions in northern Africa water resources development depends strongly on the construction of large reservoirs which are usually determined to serve the people downstream of the dam. They are attracted by the fact that unlimited water for drinking and irrigation purposes will be provided by the authorities responsible for the project. They expect that prosperous cities and industrial centers will develop and that the living conditions will reach high economic standards. The people living on the plateaus upstream of the reservoirs are since centuries used to dry farming, the traditional agricultural technique of the mountainous watersheds (Fig. 5). This technique is considered by the reservoir owners to stimulate erosion of the top soil particularly at sequences of consecutive dry years. The solution was seen in the consequent change of the land-use practice of the local farmers. The idea was to attract the local people to give up dry farming on the plateau and to shift towards irrigation techniques in the small valleys of the tributaries discharging into the large reservoirs. With this goal in mind thousands of small diversion dams have been built during the last decades of the past century to attract the people of the plateau to accept the new life conditions. During the nineties, however, it became visible that this concept did not function as desired. The consequences of the concept can be summarized as follows. Many farmers gave up their traditional dry farming practices, left their homes on the plateau and moved to the small valleys where they tried to start irrigation-based agriculture by taking advantage of the small diversion dams provided for their disposal. However, it turned out that they were not able to handle the new techniques successfully. (Experiences of this kind were also made with other climates and cultures around the

Water resources development and risk assessment in mountain regions of Africa

567

world, f.i. in the Sahel region south of the Sahara desert). In northern Africa, the people finally also gave up irrigation-based agriculture and started to move further downhill to the attractive centers downstream of the large reservoirs where they expected more prosperous living conditions. Meanwhile, the abandoned soils on the high plateaus had deteriorated within short time due to sheet, rill and gully erosion and the land was turned to badlands with extremely high erosion rates. As a consequence, the small irrigation ponds in the valleys also experienced rapid sedimentation and—as the inexperienced dry farming trained people were not used to operation and maintenance of diversion dams—most of these dams became either useless due to complete sedimentation or were destroyed by excessive floods due to inefficient flood control operation (Fig. 6, and Scheuerlein, 2002). Deterioration of the uncultivated plateau and failing to convert the dry farming people to irrigation technicians led to considerably faster sedimentation of the large reservoirs than

Figure 6. Destruction of diversion dam due to poor maintenance at Tell Atlas. anticipated. In the Oued Mina watershed in Algeria a concept was developed in cooperation of the Algerian government with the German GTZ which tried to combine soil conservation and erosion control on the plateaus on the basis of traditional land-use practices with socio-economic aspects (Gorner, 1993). In addition, a programme was started to rehabilitate the small diversion dams by supplying them with more efficient flood control structures and by improving the erosion control measures in the affluents by means of bed and bank protection.

Water resources of arid areas

568

5 COMPARISON OF THE PRESENTED AFRICAN EXAMPLES OF WATER RESOURCES DEVELOPMENT IN MOUNTAIN REGIONS WITH RESPECT TO RISK ANALYSIS APPROACHES Water-related hazardous events in African mountains can be associated with at least one of the following phenomena: – torrential floods – debris flow – landslides In the example dealing with the High Atlas (Morocco) torrential floods must be considered the prime risk but debris flow events and landslides may occur as well. In Usambara (Tanzania) debris flow is the most dangerous event. It occurs together with floods but the highest hazard potential lays in the devastating consequences of debris flow. Recently, also landslide events contribute to debris flow, often caused by inappropriate land-use practices. As far as the third example (Tell Atlas, Algeria) is concerned the hazard can not be related to one single event like in the other examples, it rather must be seen as an accumulative process caused by series of events like floods, debris flows and landslides in the whole catchment of a reservoir. Mitigation measures to minimize the risk of damages or failing of water resources development works can be divided in structural and non-structural measures (Scheuerlein, 2000b): – Structural measures Structural measures usually comprise erosion control, river training works (f.i. bed and bank protection). In lower regions of the catchment sediment retention basins may also be considered. – Non-structural measures Non-structural measures comprise hazard zone mapping as the main component. Hazard zone mapping is a rather new method, developed mainly in Alpine countries like Austria and Switzerland. In Austria hazard zone mapping is instrumental since more than 35 years

Water resources development and risk assessment in mountain regions of Africa

569

Figure 7. Hazard zone mapping in Austria. (Scheuerlein, 2001). The reason why it was developed was the economic upswing in the sixties of the past century which resulted in extensive settling activities in formerly sparsely populated mountain regions. The exploding costs for the control of torrential floods, debris flows and avalanches made it necessary to

Water resources of arid areas

570

evaluate and compare risk potential and costs for protective measures. In Austria a hazard zone map consists of two components, one cartographic part and one report. The cartographic part comprises a topographical map of scale 1:50000, 1:25000 or 1:20000 which covers the area to be evaluated plus the relevant catchment. The endangered zones (hazard zones) are indicated in a separate map scale 1:5000 (Fig. 7), with colours indicating the degree of the hazard potential (red and yellow as the extremely dangerous zones and brown for landslides, blue and violet for additional information concerning special restrictions). The classical criteria for the evaluation is an event of 150 years of reoccurrence. At present, however, there are plans to develop a more scientific concept combining a probability approach with systems analysis and causality chains and decision trees instead of working with one single event only. The risk potential of disastrous events can be significantly reduced by applying appropriate structural measures. However, as structural measures need time and money, the feasibility of protective measures has to be investigated, too. The respective evaluation is in Austria always done on the basis of hazard zone maps. This tool is also helpful for decision making purposes (i.e. for selection and ranking of structural measures). 6 CONCLUSIONS The Summit on Sustainable Development in Johannesburg has declared water as the raw material of the 21st century and the access to safe drinking water as a basic human right. Furthermore, Africa was identified for special attention and support by the international community to better focus efforts to address the development needs of Africa. Water resources development in Africa has to be oriented towards sustainability, particularly in mountain regions. Impediments are budgetary constraints, social requirements, often combined with insufficient expertise, unconsciousness of ecological coherence and ignorance of environmental aspects. Examples from various regions of Africa (Morocco, Tanzania, Algeria) show that the main problems to be dealt with in mountain areas are torrential floods, debris flow and landslides. Measures to tackle the problems may be structural and/or non-structural. Structural measures comprise erosion control, river training works and sediment retention basins. Non-structural measures are risk assessment with hazard zone mapping as the main component (Austria as example). The risk potential of disastrous events can be reduced by applying appropriate structural measures. Assessment of the risk potential can be carried out on the basis of hazard zone mapping. REFERENCES Gorner, D. 1993. Ecoulement et erosion dans des petits bassis-versants à sols marneux sous climat semi-arid mediterraneen, Deutsche Gesellschaft für Technische Zusammenarbeit (GTZ), 207pp.

Water resources development and risk assessment in mountain regions of Africa

571

Scheuerlein, H. 2000a. Determination des zones sensibles et des mesures de protection en bassins versants de haute montagne, Invited Paper, Séminaire Etude préparatoires a la creation de L’Agence du Bassin Hydraulique de Tensift (ABHT), Marrakech, Maroc, Rapport final, 23 pages. Scheuerlein, H. 2000b. Risk assessment and minimization in steep catchments, Invited Paper, 8th International Symposium on Stochastic Hydraulics, Beijing, China, Proceedings, 413–423. Scheuerlein, H. 2001. Hazard zone mapping in Austria, XXIX IAHR Congress, Technical Workshop 6, Risk Management in Mountainous Areas, Beijing, China. Scheuerlein, H. 2002. Sediment problems initiated through unexpected reactions of the public upon reservoir operation strategies, UNESCO/ICCORES Workshop on Ecological, Sociological and Economic Implications of Sediment Management, Paestum, Italy, Proceedings: 178–198. Scheuerlein, H. & Kommes, Ch. 1998. Sustainable water resources development in the tropical watershed of Western Usambara, Tanzania, Conference of IAHR African Division on Coping with Water Scarcity, Hurghada, Egypt, Proceedings: 3.4–1 to 3.4–8.

Reliability, resilience and vulnerability for reservoir sizing and operation J.G.Ndiritu School of Civil and Environmental Engineering, University of the Witwatersrand, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Reliability, defined as the proportion of time that the reservoir supplies the required demand is traditionally used as the indicator of reservoir system performance. Reliability however, does not indicate how long and how severe restrictions are likely to be when they occur. Researchers have developed two measures: resilience and vulnerability to quantify these factors. A study of the relationship between reliability, resilience and vulnerability is carried out using a realistic simulation of a system of two reservoirs located in a semi arid area of South Africa. The system is simulated 100 times using stochastically generated monthly streamflow and rainfall sequences each 77 years long. Two definitions of the two measures are used; resilience as the maximum number of periods of continuous supply shortfalls and the average length of shortfalls; and vulnerability as the maximum cumulative supply deficit and the average cumulative supply deficit. The obtained relationships exhibit large scatter and indicate that the higher reliability, the shorter the length of shortfalls is likely to be and the less severe the cumulative supply deficits are likely to be. Generalized relationships between the three performance indicators may not be practically obtainable and it recommended that measures of resilience and vulnerability should be integral components of reservoir system sizing and operation analyses.

1 INTRODUCTION Reliability, defined as the proportion of time out of the total simulation period that there are no shortfalls to supply has been traditionally used as the criteria for reservoir sizing and also in stochastic optimization of reservoir operation. Reliability is definitely an important indicator of system performance but it fails to capture some vital aspects of performance. These include the severity of shortfalls once they occur and the ability of

Reliability, resilience and vulnerability for reservoir sizing and operation

573

the system to recover from shortfalls. As Srinivasan et al. (1999) point out; continuous shortfalls of supply generally result in greater negative impacts than intermittent shortfalls. Continuous shortfalls may for example cause farmers to alter their irrigating habits permanently. Long shortfalls may also stress out the alternative water source (e.g. groundwater) meant to supplement supply from reservoirs during droughts. The term resilience has been used for measures quantifying the ability of the system to recover from supply shortfalls and the term vulnerability for measures that quantify the severity of shortfalls. Several studies on resilience and vulnerability have been carried out (Hashimoto et al., 1982, Srinivasan et al., 1999, Moy et al., 1986, Vogel et al., 1999). Hashimoto et al. (1982) applied a simple seasonal model in the analysis and Moy et al. (1986) analyzed a reservoir using a short period of 18 monthly inflows that Srinivasan et al. (1999) also applied. Vogel et al. (1999) applied a simulation of an annual lag 1 auto regressive model and used a yearly time scale. Arid and semiarid regions are prone to prolonged droughts and many parts of Southern Africa are currently facing severe drought conditions. Incorporation of resilience and vulnerability into system design and operation is therefore even more important in such regions. This analysis of the relationships provides a sense of the inadequacy of using reliability as the sole performance criterion and the need to incorporate resiliency and vulnerability at the sizing and operating stages of water resource systems. This study aimed to obtain relationships between reliability, resilience and vulnerability for a reservoir system in located in a semi-arid area. A realistic simulation of a system of two reservoirs incorporating reservoir operating shortage rules is used. A monthly time interval is applied and the system is simulated 100 times using synthetic streamflow and rainfall sequences, each 77 years long. The possibility of applying volumetric reliability, defined as the proportion of the volume of water supplied to the volume demanded, in place of the traditional time-based definition of reliability is investigated. An analysis of alternative ways of defining resilience and vulnerability is also carried out. 2 DEFINITION OF SYSTEM PERFORMANCE MEASURES: RESILIENCE Resilience is the ability of a system to recover from droughts has been quantified in a variety of ways. Hashimoto et al. (1982) defined resilience as the average probability of a recovery from the failure set in a single time step. (1) where γ is the resilience; ρ is number of times the reservoir shifts between full supply and restricted supply as a ratio of the total number of periods of analysis and 1−α is the number of periods of restricted supply expressed as a ratio of the total number of periods of analysis. According to equation 1, a larger value of γ means a more resilient system. Moy et al. (1986) defined resilience as the longest number of consecutive supply restrictions that occur with a lower value meaning a more resilient system. Srinivasan et

Water resources of arid areas

574

al. (1999) extended this definition further to incorporate the ability to recover from all the droughts and not the worst drought alone. This definition has implicit equivalence with that of Hashimoto et al. (1982) defined in equation 1. The current analysis uses two definitions; the longest period of consecutive restrictions; and the average time taken for the system to recover from restrictions or the average number of months that a failure is expected to occur. The reciprocal of this definition gives the definition of equation 1. This definition is considered more practical as it relates to performance more directly than equation 1. The lower the number of months to recovery, the more resilient the system. 3 DEFINITION OF SYSTEM PERFORMANCE MEASURES: VULNERABILITY Hashimoto et al. (1982) defined vulnerability as the summation of the product of the maximum severity of a given failure and probability of its occurrence in the failure event. (2) where ν is the vulnerability; j is the period of the severest failure in failure event F; sj is a measure of this severity; and ej is the probability that the failure that corresponds to sj is the most severe outcome in failure event F. Moy et al. (1986) and Srinivasan et al. (1999) defined vulnerability as the largest deficit during the period of simulation. The simulation herein includes restriction rules that reduce the supply by specified percentages depending on the storage state of the reservoirs and the month of the year. The definition of vulnerability as the single most severe shortage is therefore unsuitable as it is likely to be mostly confined to these percentage reductions. A definition that considers only the worst single period may be unrealistic especially when working with time intervals much shorter than the probable lengths of droughts. The following two definitions of vulnerability are therefore applied herein; the worst cumulative deficit during a drought event; and the average of the cumulative deficits during the simulation period. 4 SYSTEM SIMULATION AND NUMERICAL EXPERIMENTS The configuration and historical data from a system of two reservoirs located in the Elands river catchment in South Africa was selected for analysis. The upper dam, Rust de Winter has a catchment area of 1145km2 and a mean annual runoff of 19.8mm. The incremental area for the downstream reservoir, Mkombo is 2578km2 and the mean annual runoff from the incremental area is 3.9mm. The mean historical mean point rainfall at Rust de Winter and Mkombo is 605 and 243mm respectively. Seventy seven years of monthly streamflows and point rainfalls at the reservoir site was available. Monthly average Symon’s pan evaporation rates were also available. This system had been applied in another study (Ndiritu, 2004a), which aimed to obtain reservoir sizes and operating rules that maximize yield whilst dealing with multiple constraints of supply reliability.

Reliability, resilience and vulnerability for reservoir sizing and operation

575

One of the optimized solutions obtained from this study was selected as the basis of the numerical analysis carried out here. This solution included the optimized live storages of the two reservoirs and the monthly operating rule curves for each reservoir and for the total system storage. The operating rules allowed for 4 levels of supply (100%, 80%, 60% and 30%) each month depending on the volume in storage in each reservoir and the total system storage. Allowance for lower percentages of supply (<30%) and for regulated flow from the upper to the lower reservoir was also incorporated. Figure 1 presents a schematic of the system. To achieve the aims of the study, it was necessary to perform several system simulations each providing corresponding values of reliability, resilience and vulnerability for each reservoir. The historical streamflow and rainfall data was used to generate 100 synthetic sequences of the same length (77 years) using a nonparametric stochastic data generation method. This method, developed by the author, is based on the premise that the historical data can be rearranged a large number times to generate series that capture the important characteristics without having to ‘decompose’ the historical data into statistics and then build it up again. This method captures the droughts in the historical flow and produces a synthetic flow that has have the same number and lengths of droughts but assigns these randomly over time. The generated flow series thus maintains the important aspects of the annual serial correlation. The cross correlation between streamflows is captured by noting the common drought years between the historical streamflows and generating synthetic flows that have a similar number of common droughts lengths and periods. The historical monthly distributions are used to disaggregate the annual flows in a manner that maintains the serial correlation at the end of one year and the beginning of the next. Tests on this approach (Ndiritu, 2004b) indicate it is effective and efficient.

Figure 1. Main components of system simulation.

Water resources of arid areas

576

Figure 2. Relationship between reliability and volumetric reliability.

Figure 3. Relationship between two definitions of resilience.

Figure 4. Relationship between two definitions of vulnerability. 5 RESULTS AND DISCUSSION Figures 2 to 7 present the results graphically. Rust de Winter dam is denoted by ‘RdW’ and Mkombo dam by ‘Mko’ in the Figures. The following observations are made.

Reliability, resilience and vulnerability for reservoir sizing and operation

577

– There are high correlations between the two definitions of each performance indicator (Figures 2, 3 and 4). – A correlation exists between reliability and resilience but with a large scatter (Figure 5). – A correlation with a large scatter also exists between reliability and vulnerability (Figure 6). – A high correlation exists between resilience and reliability (Figure 7). – Rust de Winter gives a greater scatter that Mkombo in all the relationships. The high correlation between the two definitions of reliability implies no advantage is likely to be obtained by replacing one with the other. Relationships between volumetric reliability and the two other performance measures were of similar quality to those obtained with the traditional definition of reliability. The significant difference in the relationships for Rust de Winter and for Mkombo dam is an indication that generalization of these relationships may be difficult except for

Figure 5. Relationship between reliability and resilience.

Figure 6. Relationship between reliability and vulnerability.

Water resources of arid areas

578

Figure 7. Relationship between resilience and vulnerability. simplified screening studies. The large scatter in the relationships between reliability and the other two performance indicators supports further the need to include these indicators (resilience and vulnerability) in sizing and operating studies. If they are not, the expected behaviour of the system should be analyzed to obtain the same information as the two indicators i.e. how severe and how long shortages to supply likely to be? Reliability gives a sense of the expected average performance of a reservoir and so do the resilience and vulnerability parameters defined as average failure lengths and average cumulative deficits. The definitions of resilience and vulnerability defined as the longest length of deficit and the most severe cumulative deficit respectively are therefore likely to be more useful to the cautious water resource systems analyst. If a simulation-optimization approach is adopted in system sizing and operation, it is easy to incorporate any of the definitions of reliability, resilience and vulnerability. If other optimization approaches such as linear or dynamic programming are applied, incorporating all the performance measures is likely to be more challenging. Pioneering work regarding this has however been carried out (Moy et al. (1986), Srinivasan et al. (1999)). 6 CONCLUSIONS AND RECOMMENDATIONS The relationships between reliability and two other reservoir performance measures: resilience and reliability have been studied using data and the configuration of a system of 2 reservoirs located in a semi-arid region of South Africa. The relationships obtained indicate that in general, a more reliable system will have a higher resilience and a lower vulnerability. These relationships however exhibited a large scatter and generalized relationships that provide the resilience and vulnerability for a given reliability may not be obtainable. It is therefore recommended that measures of resilience and vulnerability should be incorporated into reservoir system sizing and operation studies. Water resources planning and management is shifting towards greater stakeholder participation. For this reason, it is desirable that measures of resilience and vulnerability should be defined in a practical way that is easily understandable by the stakeholders and not only the systems analysts.

Reliability, resilience and vulnerability for reservoir sizing and operation

579

REFERENCES Hashimoto, T., Stedinger, J.R. & Loucks, D.P. 1982. Reliability, resiliency and vulnerability criteria for water resource system performance evaluation. Water Resour. Res., 18:14–20. Moy, W.S., Cohon, J.L. & Re Velle, C.S. 1986. A programming model for analysis of the reliability, resilience and vulnerability of a water supply reservoir. Water Resour. Res., 22:489– 498. Ndiritu, J.G. 2004a. Optimizing Water Supply System Yield Subject to Multiple Reliability Constraints using Genetic Algorithms. Submitted to J. Water Resour. Plann. and Manag., ASCE. Ndiritu, J.G. 2004b. ‘A pragmatic nonparametric approach for monthly multisite streamflow generation’. Paper in preparation. Srinivasan, K., Neelakantan, T.R., Shyam Narayan, P. & Nagarajukumar, C. 1999. Mixed-integer Programming Model for Reservoir Performance Optimization. J. Water Resour. Plann. and Manag., ASCE, 125(5):298–301. Vogel, R.M., Lane, M., Ravindiran, R.S. & Kirshen, P. 1999. Storage Reservoir Behaviour in the United States. J. Water Resour. Plann. and Manag., ASCE, 125(5):245–254.

Hydrological impact of dam construction in an arid area D.Stephenson University of Botswana Z.Chengeta Mantswe Natural Resources Consultants, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Hydrological investigations were carried out for a possible dam on the Thune river in Eastern Botswana. The main motivation for constructing the dam was to provide an alternative to depleting groundwater resources to surrounding villages. The space boundary of the project was confined to the project area and within Botswana. Use of reservoirs in arid areas increases the risk of drought compared with temperate areas. This is because flows are more variable and water has to be stored longer. Risks also arise due to siltation, cracking, decommissioning, poor hydrological data and economic viability.

1 INTRODUCTION During investigations on the hydrological impacts of constructing a dam on the Thune River in Botswana, the following factors emerged: A dam would reduce low flows considerably downstream, unless releases were made. Instream and riparian users would suffer but adjacent villages and farming could benefit considerably by making water available from the dam. Downstream floods would be reduced due to the routing effect of the dam. This applies particularly to smaller floods but larger floods may not be so affected. A large volume of water would be lost due to evaporation from the reservoir surface. This is due to high temperatures, low rainfall and large surface area of reservoir. The downstream flow spectrum would change. In particular there would be less total flow, and less low flow. The larger the dam the higher the evaporation and the marginal benefit is therefore diminished. Soil erosion and sedimentation problems arise. Increased farming and construction increase soil erosion and sediment is deposited in the reservoir particularly in the upper reaches. This increases surface area for any capacity and therefore evaporation increases. There is less flood bank silt deposition reducing bank fertility. Downstream of the spillway, erosion is increased. The sediment will over years reduce the capacity of the

Hydrological impact of dam construction in an arid area

581

reservoir to useless proportions. The sedimented reservoir could be used as a sand reservoir with reduced evaporation, but capacity is less and abstraction is difficult. Groundwater levels and bed water decrease which is a problem for animals and humans used to digging wells in the river bed. 2 RELEVANT INTERNATIONAL OBLIGATIONS, TREATIES AND PROTOCOLS – The SADC Protocol on Shared Water Course Systems; signed by SADC countries on 28 August 1995. The SADC Protocol requires co-operation between states within a particular basin on issues regarding the river. – The Helsinki Rules on the Uses of Water of International Rivers; adopted at the 52nd Conference of the International Law association in Helsinki in 1966. An important principle of the Helsinki Rules is that each basin is entitled, within its own territory, to a ‘reasonable and equitable share’ in the beneficial uses of water of an international drainage basin.

3 ASSESSMENT OF HYDROLOGICAL DATA A gauging weir on the lower Thune was found to be well designed with gauging staff and a recorder house. However, the bed of the river is alluvial which is likely to suspend during floods and there is likely to be a backwater effect from the Motloutse confluence a few kilometers downstream. By comparing with other records on a longer term basis, we are able to check the reliability of the records and extrapolate them to the proposed dam site. The record is however so short that it had to be extrapolated using rainfall-runoff modeling methods. The catchment was inspected from the point of view of estimating hydrological runoff as well as topographical features which will be incorporated in the catchment models. Vegetation and soil types were inspected from the point of view of erosion and agriculture. Wells on the banks of river were inspected and surveyors traveled down the riverbed to investigate the use of ground water from the riverbed. The cross sections of the river and flood plains were surveyed for use in the flood study. Rainfall data was obtained from the Department of Meteorological Services, i.e. monthly and daily records from surrounding rain stations, of which Francistown is the most comprehensive. Flow records were obtained from the Thune river for the limited time available as well as Motloutse, Lotsane and Shashe rivers. These are largely on a monthly basis but the water resources study is done on a monthly basis for mass balance purposes.

Water resources of arid areas

582

4 RESERVOIR YIELD The effect of the reservoir on the yield of the river and in particular downstream flows was investigated. A generalized model of the catchment was used whereby the flow records were extended using rainfall-runoff modeling, which is relatively simple using the Rafler model. The model was calibrated by using the existing stream gauge data available lower down on the Thune river. The model was able to extend the monthly flows down the river to the reservoir to 50 years which enabled a better statistical comparison to be made and better risk analysis to be done than the existing short period records. The results were modeled to the proposed reservoir to obtain the net downstream flow after allowing for draft and evaporation and other losses. The results indicated a considerable change in the flow spectrum downstream. That is, low flows, particularly when the reservoir is low, will be absent, whereas they were previously used to replenish the water in the sandy bed of the river. Arrangements will therefore have to be made to release water at a programmed rate in order to maintain the regime downstream. Water is abstracted from the bed of the river downstream, particularly for stock and domestic use. The effect of soil erosion upstream in the Thune catchment was examined and the variation of erosion rate with runoff and rainfall rate was investigated. It was found that the catchment is particularly sensitive to the rate of rain and the highest erosion rate occurred during the more intense storms. These may not necessarily be those which result in the maximum storm runoff. It was found that soil erosion will significantly reduce the yield of the reservoir even if it is constructed to many times the average annual flow of the river. The yield of the reservoir will reduce by more than 50% over 50 years after construction owing to sedimentation. The reservoir is also found to significantly increase evaporation losses such that the downstream net flows are even less than indicated for just draft from the reservoir. In fact, the dam would spill rarely unless simulated large releases are made at intervals. More water is going to be lost by evaporation than will be consumed using a steady draft operating system. The construction of a dam across a river results in the capturing of most low flows and the only uncontrolled release is the flood which over-tops the spillway on occasions when the dam is full. The following changes to the hydrology of the river downstream therefore occur. Without releases from the dam, the river dries up for months a year and the time span between discharges increases considerably. There may only be one or even no flows down the river in any year. The result is that communities based along the river banks become accustomed to a dry river and can suffer when a flood, which is largely unexpected, does occur. Communities are also dependent on low river bed flow even during dry seasons. Water in the sandy bed of the river can be abstracted by people and animals digging shallow wells, and wells in the banks of the river also draw water from the bed of the

Hydrological impact of dam construction in an arid area

583

river. The average release from the dam would have to be approximately 25 litres/second to meet these requirements downstream to the Motloutse river beyond which the Motloutse flow will contribute to the bed flow. The reservoir behind the dam will reabsorb some of the floods, but it appears from the calculations that the reduction in peak floods would only be of the order of 5% which is small and this only happens on occasions. The reduced flood flow will result in slightly lower water levels and this will enable an additional 160 hectares downstream to be less flooded. 4.1 Mass balance of runoff Although the average flow in the Thune river is 26 million cubic litres a year, the variation from year to year is very wide. The result is that a large storage capacity is needed to regulate the river, i.e. to provide any reliability in flow. This effect means that relatively more storage than on average in Southern Africa would be required. A reservoir with the capacity of twice the average flow of the river would nevertheless only make a reliable flow of 5–10 million cubic litres available a year. This assumes some sedimentation has occurred to reduce the capacity of the reservoir. In general then the flow of the river into the reservoir is largely caught but discharges over the spillway will still account for approximately 1/3rd of the inflow. Another approximately 1/3rd of the inflow will be lost in evaporation, leaving only approximately 1/3rd of the inflow available for drawoff. Part of this in turn will have to be released downstream for instream requirements. 5 FLOOD ANALYSIS Floods were calculated in the Thune river for various recurrence intervals. This is to study the routing effect of the dam on the flows downstream and also to study the changing high flow frequencies due to retention by the dam. We used the standard methods of flood estimating, namely the rational method, the unit hydrograph method with data we have obtained for other flood studies in the region, and catchment models. An average was selected from the results and flood frequencies ranging from 20 years to a 1000 years could readily be extrapolated and the hydrographs resulting, routed through the proposed dam. It appears there will be an attenuation of the floods even if the dam was relatively full and therefore the downstream flows would mean the resulting floods downstream would be routed down the river using our hydraulic water profile program. The resulting flood plain with and without the dam were compared. The dam has a routing effect on floods and downstream the flood inundation is less if the dam is constructed to the amount of 160ha. The hydrological report compares the flood spectra with and without the dam. The entire flood spectrum downstream of the dam changes and there is not much effect on the larger floods, but the smaller floods and low flows will not happen unless positive releases are made from the dam. A proper operating rule is therefore required and the report indicates the minimum flow required for instream requirements.

Water resources of arid areas

584

The construction of the dam increases evaporation losses considerably and the graph shows the relative effect of abstraction from the dam and evaporation for the selected dam size. The evaporation loss can be significantly greater than the release and once the downstream instream requirements are also released, the yield of the dam will be very small compared with the mean annual flow of the river. This means that most of the water is lost by evaporation and this has a serious adverse effect of the dam. 6 SEDIMENTATION The catchment is one of fairly high erodibility even compared with other Botswana rivers. The soil erosion upstream of the dam which occurs, irrespective of whether the dam is built or not, is approximately 700000 cubic metres a year. A reservoir capacity of 50 million cubic litres would therefore be 2/3rds full of sediment within 50 years, reducing the reliability of the drawoff considerably. Whereas the bulk of the sediment transported into the reservoir is deposited, flows over the spillway during floods will be clean of sediment and have a greater potential for erosion of the river bed and banks downstream of the dam wall. The sediment regime of the river is considerably affected by the presence of the dam. It is unlikely that any operating rule of the dam will change this regime much and the following is the main impact. The reservoir behind the dam will catch practically all the sediment as indicated in our report. This will cause the upper reaches of the dam initially to become blocked with sediments including silt and sand and push the flood plain further back upstream. Within 30 years the dam will be nearly half full of sediment and its reliable yield will be much less after 20 to 50 years. Downstream of the dam there will be less sediment released by the dam and river. There will therefore gradually be erosion of the existing sand bed and banks of the river, more than the present average rate, and this sediment will be transported to the Limpopo River via South Africa, Zimbabwe and Mozambique to the Indian Ocean. This is because the river is impounded and it loses the carried sediment in the reservoir and has to pick up new sediment downstream to reach its equilibrium of concentration. In addition, the energy dissipation effect of discharge over the spillway will cause erosion downstream even if there is construction of a stilling basin. This is because cleaner water has a higher energy and there is a considerable energy released over the dam spillway which, without the dam, would have been dissipated in the form of channel friction along the length of the reservoir. Sediment yields over all the catchment indicates a potential annual average erosion rate of 2.5×106m3 per annum. However, the erosion modeling which allowed for reasonable ground cover and limitations on the carrying capacity of the runoff indicated an average sediment yield of only 6,700000m3 per year. This compares well with the yield at times for the Shashe dam and others in the area on a per km2 basis. It should be noted that even with the reservoir capacity of 50 million cubic metres initially, the reservoir volume would be reduced to 20 million cubic metres after 50 years. The corresponding yield of the reservoir would therefore decrease from about 5 million cubic metres per annum with an initial capacity of 50 million cubic metres, or 4 million cubic metres per annum with an initial capacity of 40 million cubic metres, down

Hydrological impact of dam construction in an arid area

585

to less than half after 50 years. When the releases required downstream are taken into consideration, the yield is even smaller. The yield is based on 98% reliability and would be higher if a higher risk of failure is acceptable. It is anticipated that a concrete causeway would be constructed to enable trucks and other vehicles to cross the river for construction purposes. Such causeways can block river bed flow unless adequate discharge pipes are allowed through the causeway even below the bed surface. The causeways should be built not much more than half a metre above the sand level to enable floods to pass sand downstream. The river is likely to be temporarily diverted to one side of the channel while construction of the embankment occurs on the other side and a bypass tunnel may be constructed under the first half of the embankment so that the second half of the embankment can be constructed subsequently. During these diversion works, the floods will be concentrated and cause additional erosion downstream of both the bed and the bank of the river. Erosion protection should therefore be specified as part of the temporary diversion works. Any outlet gates from the proposed dam would be insufficient to discharge the full sediment load through the dam wall. Only fine sediment in the vicinity of the gate would scour however long the gate was opened for. Releases downstream should therefore be primarily aimed at maintaining the ecosystem along the bank. The maximum rate of release will depend on flood rate which may be tolerable and less losses would occur in the form of evaporation if the releases were made for example, once a week over a few hours, instead of continuously over 24 hours a day and 365 days a year. The downstream instream requirements should be released in gulps to minimize evaporation loss and to ensure a flow rate never exceeding 2m3/second to avoid endangering livestock or people. This would mean that releases should only occur for 1% of the time, i.e. say 1 hour in 100, or approximately 2 hours a week. The rate of the downstream release will depend on the sizes of the outlet valves and pipework. 7 MACRO REGIONAL EFFECTS The presence of the reservoir behind the dam wall would increase evaporation losses from the Thune river. The area of the reservoir is however relatively small and the effect on the climate both locally and nationally would not be noticeable. The effects on downstream rivers would be noticeable, particularly for the low flow. The Motloutse river into which the Thune flows and further down, the Limpopo river, would receive less base flow from the Thune catchment. In fact, the entire base flow would be eliminated. On the other hand, the peak floods are largely unattenuated and therefore the dam would have negligible effect in managing floods for downstream rivers. There is also a potential salinization problem. Salinization is a major problem in irrigated land globally and has severe, long-term and often permanent impacts on land, agriculture and livelihoods where rehabilitation is not undertaken.

Water resources of arid areas

586

8 POSSIBLE NEGATIVE CUMULATIVE AND REGIONAL IMPACTS OF THE DAM – Increased risks of flooding in downstream areas, e.g. Mozambique – Long term deteriorating water quality and possible reduced species composition. – Increased erosion downstream because of reduced sediment load in the water (clean water has more energy). – Less water to downstream users not only in Botswana but also in South Africa, Zimbabwe and Mozambique. – Less water to sustain natural systems downstream. – Contribution to emission of greenhouse gases due to the rotting of vegetation and carbon inflows from the catchment.

9 DECOMMISSIONING DAM AFTER ITS USEFUL LIFE The dam will reach the end of its economic life within a century. This will largely be due to sedimentation. There may also be changes in water requirements and value and the possibility of abandoning the dam in the future should be considered. The sediment deposited in the dam would be fine silt as opposed to the coarse sandy bed material which is present in the river at the moment. It could therefore theoretically be used for

Figure 1. Effect of dam of various sizes on evaporation and spill losses.

Hydrological impact of dam construction in an arid area

587

Figure 2. Effect of reservoir on downstream flow spectrum. growing crops, but in the reservoir on its own it would be subject to flooding, so it would therefore have to be dredged or dug out of the reservoir and spread in other areas. It appears that the dam wall will be an embankment-type wall which is fairly stable but it should be borne in mind that the sediment caught behind the dam wall will be wet and therefore continuous seepage through the dam wall could be expected as long as the dam wall were present. Safety checks for slope stability and piping (erosion due to underground flow) would need to be monitored if the dam wall were abandoned. The collapse of the dam wall, if it were left unattended, will have a considerable effect because a large sediment flow and possibly water flow would proceed down the river channel. The reliability of drainage under the embankment should be considered by the design engineers. Drainage should also be provided to dry the sediment in the reservoir, to make it safe to walk on. 10 RESERVOIR LEVEL MANAGEMENT To optimize the releases or drawoffs from the reservoir, the reservoir level will have to be held as high as possible. There are however a few reasons why the water level should be fluctuated over short periods of time. – The releases for downstream instream requirements should preferably be made over short periods to minimize evaporation losses in the downstream channel and enable the release to travel as far as possible down the Thune river – Water based insect vectors can be controlled by varying the level in a tropical reservoir. Bilharzia snails can be left on the dry banks and killed if the water level is fluctuated, i.e. dropped, for a few weeks over the growing seasons of the snails. Mosquitoes will also reduce breeding if the shallow water on the banks is occasionally dropped and banks dried. – Benthic deposits and accumulations in the reservoir can be minimized by discharging the heavy sediment-laden waters from the bottom of the reservoir occasionally. Heavy

Water resources of arid areas

588

metals and even nitrates can be discharged in this way. It will also provide organic matter to the downstream channel.

11 CONCLUSIONS A dam constructed in an arid area has many more pitfalls than one constructed in a temperate area. The river hydrology and regional water balance are affected. There is large evaporation loss and downstream flows are reduced during drought. There are also probable cumulative effects that may be national and regional.

The geochemistry of fresh water supplies in Botswana L.Molwalefhe Department of Geology, University of Botswana, Gaborone, Botswana S.Vriend Department of Earth Sciences—Geochemistry, Faculty of Geosciences, Utrecht University, The Netherlands Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Fresh water chemistry from various parts of Botswana was studied to evaluate the factors that control the water quality of fresh waters across the country. Evaporative concentration is a significant process that determines the chemical compositions of the fresh water supplies. At high salinity, most waters precipitate calcite, dissolve gypsum and acquire more chloride content. This is observed for desert environment in the western part. At low salinity, i.e. less evapotranspiration, dissolution of calcite is dominant. Na is added to the water from feldspar dissolution and cation exchange processes. Three clusters were defined; dissolution, transition and evaporated, where calcite is either dissolving or precipitated.

1 INTRODUCTION Botswana has an arid to semi arid climate, characterized by erratic rainfall and frequent severe droughts which gives rise to relatively limited fresh water resources. Water supply in the major cities is largely dependent on surface water. Several major dams have been built for this purpose in the eastern part of the country where 80% of the population lives. In the rural areas 90% of the water supply is based on groundwater. In the northern regions several perennial rivers provide an ample source of fresh water supply. In a country like Botswana with limited water resources it is importance to have a good understanding of the factors that determine the water composition. The Department of Water Affairs routinely monitors the major chemical elements in the fresh water supplies. This database is used in this paper to study the hydrochemistry of fresh waters in Botswana. Through statistical and hydrochemical approaches, a survey is made of the

Water resources of arid areas

590

water types that are present in the fresh waters and the processes that control the chemical. 2 CLIMATE The climate is continental semi-arid. Due to the absence of major topographical barriers there is a low spatial variability in the climate. The average of the erratic yearly rainfall is 450mm and varies from 230mm in the southwest to 650 in the northeast (Figure 1). The dry winter lasts from May to September and the rainy summer from September to March. Temperatures are high (38°C) in summer (Bhallotra, 1984). Daily variations in temperature of about 20°C in winter are common. Evapotranspiration is generally high due to the high temperatures and the low humidity. Evaporation from open water is in the order of 2m per year. 3 DATA BASE AND METHODS We only have waters in the database that have some economic value i.e. are potentially useful for drinking water or cattle drinking purposes.

Figure 1. Precipitation map.

The geochemistry of fresh water supplies in Botswana

591

Figure 2. Boxplots of the saturation indices for calcite and gypsum. The different groups refer to the fuzzy cmeans cluster analysis. 3.1 General hydrochemical interpretation of the results The water composition is influenced by many factors. A meaningful interpretation can therefore only be attained through a thorough analysis of the data. For this purpose univariate and multivariate statistics as well as thermodynamic considerations have been used. 3.2 Thermodynamic calculations Saturation indices for calcite and gypsum were calculated using PHREEQC (Parkhurst, and Appelo, 1999). The results are presented in Figure 2 and show a major part of the dataset is saturated or close to saturation with respect to calcite. The waters saturated in gypsum are rare. 3.3 Histograms Histograms of the logtransform of a selection of constituents are given in Figures 3 and 4. 4 SCATTERGRAMS In Figure 5, scattergrams of the different elements are plotted against chlorine. Chloride is assumed to be entirely derived from rainwater and is a conservative element (Selaolo, 1998).

Water resources of arid areas

592

Figure 3. Cl is similar to Na, and HCO3 to Ca. Except for the pH (intrinsically logarithmic) all variables show lognormal characteristics. Also some multimodality may be present. The histograms of Ca and HCO3 is somewhat skewed to the left, which is explained by the saturation of calcite in many waters, as is indicated by the thermodynamic calculations, thus limiting the upper tail of the distribution. This phenomenon is also expressed by the smaller range of values for Ca and HCO3 than for the other measured ions. In the further data treatment of the logtransform of the variables is used with the exception of the pH.

Figure 4. Histograms of a selection of constituents. Na is similar to Cl, HCO3 to Ca. The X-axis gives the log value of the concentrations in mg/ℓl.

The geochemistry of fresh water supplies in Botswana

593

Figure 5. Scatter grams against chloride. No other mechanisms, except precipitation as salt at very high concentrations, are available to remove chloride from the waters. It appears that the seawater dilution line is an adequate starting point for the description of these ions. A similar conclusion was reached by Gieske (1992) for rain water analyses of Botswana. He noticed a gradual increase in relative chloride concentrations in the western parts of the country. He postulated that this phenomenon might be due to dust that is derived from salt pans. At lower concentrations, a surplus of sodium is observed. This is explained by the release of sodium during feldspar weathering, a common rock-forming mineral in eastern Botswana (Carney et al., 1994). At higher chloride concentrations, evapotranspiration is the predominant process. The Cl vs K plot shows similarities with Na vs Cl. However, many samples appear to be depleted in K according to the seawater dilution line, an indication that these samples have lost potassium probably via cation exchange processes in the aquifers. The Ca vs. Cl pattern is caused by different phenomena. At low chloride concentrations Ca is dissolving from calcareous rocks or from the calcrete deposits. The more evaporated waters have lost Ca through precipitation as calcite. Somewhat the Ca

Water resources of arid areas

594

concentrations do increase with the Cl concentration. This is due to the increase in ionic strength which allows that more Ca can remain in solution in a saturated environment. SO4 is often enriched in relation to the seawater dilution line. Only a few waters are supersaturated in gypsum. Also a number of waters are depleted in sulphate, which is probably caused by sulphate reduction in the aquifers. 5 PIPER DIAGRAM Relative amounts of meq% of major cations and anions are expressed in Piper diagrams as shown in Figure 6. In the cation triangle the waters with the higher salinity converge at the Na+K apices. The more dilute waters contain carbonates as major anion species, while at higher concentrations the major anion is Cl with sometimes considerable quantities of SO4. 6 FUZZY C-MEANS CLUSTERING AND FACTOR ANALYSIS 6.1 Cluster analysis Cluster analysis was applied to separate the datasets in homogeneous groups. Instead of conventional hard clustering a technique that uses a fuzzy algorithm is used. Conventional

The geochemistry of fresh water supplies in Botswana

595

Figure 6. Pipe diagram for compositional representation of the water chemistries. clustering techniques often are too rigid and do not consider compositional overlap between clusters. In this dataset considerable overlap is expected as the sampled environment probably exhibits wide ranging compositions that may be subject to several gradual changes. A technique that has successfully been applied on similar datasets is fuzzy c-means cluster analysis (FCM). In FCM, the similarity between samples and clusters is described with a membership, which varies between 1 (exactly similar) and zero (completely different). Samples that have a composition intermediate between two or more clusters can be recognized by the similarity of the assigned memberships. One of the strong points of FCM versus conventional hard clustering is that outliers and intermediate samples have little or no effect on the outcome of the clustering procedure and that they are easily identified as they have more or less equal memberships to all clusters. The number of clusters that is actually present in a dataset is not always easy to determine. Two functionals, the classification entropy H and the partition coefficient F, are available to resolve this, but still these functionals may not give conclusive evidence. Therefore, the evaluation of the FCM results with a different technique, where a visualization of the FCM results is possible becomes helpful. Here a non-linear mapping technique (NLM) was used for this purpose. Thus if the FCM gives a compact grouping in a NLM plot, this

Water resources of arid areas

596

is strong argument for the validity of the used cluster model. For this study, following the above, a four cluster model has been adopted. The cluster centers are given in Table 1. A representation of the correspondence between the FCM and NLM result is given in Figure 7. In this interpretation the fuzziness has been removed from the classification and the memberships have been hardened to either 0 or 1. This resulted in the identification of some 14% as intermediate samples. A characteristic label is assigned to each cluster. This label is used in the evaluation and discussion below. 6.1.1 Dilute cluster This cluster is characterized by very low concentrations in all components. However, if the samples are studied in detail, it appears that there is a great variety in compositions. If a separate clustering is carried out on this dilute group of samples an useful subdivision can be made (Table 2). The first cluster could be called the Okavango clusters as it mainly consists of samples taken from the Okavango river and Delta. Rain water mostly contains less than a ppm of Cl (Selaolo, 1998). Concentrations are extremely low for all components. The second cluster contains especially samples collected within the sandstone area east of Thamaga. Given the low concentrations and a pH that is somewhat lower than the pH of rainwater, it seems that the aquifers in the sandstone area are replenished with rainwater that either has had minimal interactions with rock or that passed through sandstone that contains solely intergrown quartz grains. The third cluster is extremely low in calcium and magnesium and is mainly found in the Tsabong area. It looks like rain water that has only been in contact with a layer of a pan?

Table 1. Cluster centers of the four cluster model for the entire dataset. An L in the second column indicates that the log transform is used in the analysis. Variable Transform Dilute Dissolution Transition Evaporated No. samples pH Ec Ca Mg Na K Cl SO4 HCO3

L L L L L L L L

35

138

169

98

6.3 9 2.7 .9 4.5 .75 3.9 1.1 24

7.6 52 33 12 35 1.5 23 6.4 240

7.7 113 58 29 100 3.7 90 31 370

7.7 590 79 42 1000 16 1200 510 305

The geochemistry of fresh water supplies in Botswana

597

Figure 7. NLM-plot wherein the cluster configuration is indicated. Note that the “intermediate” are located in the transition zones between clusters. Table 2. Cluster centers of a three cluster model for the dilute cluster. An L behind a variable indicates that the log transform is used in the analysis. Compositional ranges for rain for Botswana are from Gieske (1992). Variable Transform River pH Ec Ca Mg Na K Cl SO4 HCO3

L L L L L L L L

Sandstone Tsabong Rain 6.9 4.7 3.8 .8 1.6 .8 .7 .6 25

5.4 5.3 2.0 1.1 3.3 .4 5.6 .8 12

6.4 26 .5 .5–1.8 .5 .14–.8 44 .3–.9 2.7 .3–.9 47 .35–1.25 3.7 .8–3.8 35

Water resources of arid areas

598

7 CONCLUSIONS 7.1 Dissolution cluster The concentrations in this cluster are relatively low. Calcite saturation is common, and some water rock interactions have occurred judging from the surplus of Na and K. Processes that have been operative in these waters may also be noticed through factor analysis. Table 3 gives the varimax rotated two factor model that was calculated for these waters. The first factors has high positive loadings of EC, Na, Cl and SO4 and thus points towards evapotranspiration. The second factor is highly positively related in the EC, Ca, Mg, and HCO3, pointing towards dissolution of limestone. 7.2 Transition cluster With the exception of pH, Ca and HCO3 concentrations are about three times higher than those in the previous cluster. The factor analysis is similar to the one for the dissolution cluster (Table 4). However, the loading of HCO3 on the second factor is considerably lower. The cluster can be interpreted as a transition between the dissolution and the next so-called evaporation cluster.

Table 3. Varimax rotated factor model for the dissolution cluster. Factor

1

pH Ec Ca Mg Na K Cl SO4 HCO3

2 .49 .69 −.28 −.40 .89 −.13 .73 .69 .25

−.49 .63 .81 .77 −.15 −.30 .18 .79

Table 4. Varimax rotated factor model for the transition cluster. Factor pH Ec Ca Mg Na K Cl SO4

1

2 −.29 .80

.79 .83 .65

−.50 .34 .81 .86 −.42 −.49 .19 −.25

The geochemistry of fresh water supplies in Botswana

HCO3

−.18

599

.33

Table 5. Varimax rotated factor model for the evaporated cluster. Factor pH Ec Ca Mg Na K Cl SO4 HCO3

1

2 .96 .19 .23 .97 .21 .91 .78 .17

−.84 .15 .83 .74 −.10 .18 .27 −.75

7.3 The evaporated cluster The contents are highest for Na, K, Cl, and SO4. For limestone related elements the concentrations are similar, though somewhat higher, to the transition cluster. Factor analysis shows that evaporation is now the most important process (Table 5) expressed by the very high loadings for these elements. The second factor loads Ca and Mg positively, and HCO3 negatively. This is caused by precipitation of calcite through evaporation and follows the principle described in Drever(1997). REFERENCES Bhalotra, Y.P.R. 1987. Climate of Botswana Part 2. Elements of Climate. Department of Meteorological Services, Botswana. Carney, J.N., Aldiss, G.T. & Lock, N.P. 1994. Geology of Botswana. Geological Survey Department, Botswana, Bulletin 37. Drever, J.I. 1997. The Geochemistry of Natural Waters, Surface and Groundwater Environments. 3rd ed, Prentice Hall, New Jersey, Prentice Hall: 436pp. Gieske, A. 1992. Dynamics of groundwater recharge: a case study in semi-arid eastern Botswana. Unpublished PhD Thesis, Free University Amsterdam. Parkhurst, D.L. & Appelo, C.A.J. 1999. User’s guide to PHREEQC (version 2)—A computer program for speciation, batch-reaction, one-dimensional transport, and inverse geochemical calculations: U.S. Geological Survey. Water-Resources Investigations Report 99–4259:312pp. Selaolo, E.T. 1998. Tracer studies and groundwater recharge assessment in the eastern fringes of the Botswana Kalahari, Letlhakeng Botlhapatlou area. Unpubl. PhD Thesis, Free University Amsterdam.

Groundwater modelling with limited data: a case study of Yobe River Basin, North East Nigeria M.Hassan1, R.C.Carter2 & K.R.Rushton2 1

Department of Physics, University of Maiduguri, Nigeria 2 Cranfield University, Silsoe, UK Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: An exploratory numerical groundwater model of a shallow aquifer interacting with a river in a semi-arid zone is developed using MODFLOW. The model has served as interpretation of fieldwork data and the physical processes presented in its conceptual models. Limited data did not constrain the model’s plausibility in the representing to a considerable extent the understandings incorporated in its concept. The behaviour of the aquifer to exist under both confined and unconfined conditions has been captured by the modelling results. Water balance from the model shows that river to aquifer flow dominates other sources of recharge, and its magnitude is limited not only by relative head difference, but by the transmissivity and hydraulic gradient of the aquifer.

1 INTRODUCTION The area under study is located in the semi-arid zone of north-east Nigeria, which is characterised by low rainfall and reduced river flow, Figure 1. In the last three decades rainfall has decrease by about 30% (Hess et al. 1995) and annual discharge by the major headwaters of Rivers Hadejia and Jama’ are has decreased by almost 60%. The reduction in the discharge from these rivers to the

Groundwater modelling with limited data

601

Figure 1. Study area map. Yobe Basin is due to construction of dams across them in addition to low rainfall. As a result of these changes in the hydrology of the area, Federal and State governments in the North East Arid Zone (NEAZ) made tremendous efforts towards developing and managing the existing water resources. Professional water consultants were engaged over the last two decades (1975–1995) with various terms of reference to develop the groundwater in the river basin (Shultz, 1975; IWACO, 1985; Water Surveys, 1986; Diyam, 1987; NEAZDP, 1990; Water Surveys, 1994). Four natural features that limit opportunities for the development of the water resources in the area were identified from these works. They include aridity, highly seasonal climate, climates that have shown major changes or trends in recent decades and poor aquifer (Carter, 1998). These studies did not address the state of the groundwater in the area or even try to combine its availability with strategies for its management. A study by Alkali (1995) and in a paper by Carter and Alkali (1996) it was suggested that the shallow aquifer in Yobe Basin has complex hydrogeologic features. For example, it was discovered that the aquifer is covered extensively with low permeability material such as clay that hinders vertical infiltration of water. According to them, the dominant factor in the recharge of the aquifer is the river and that the aquifer is capable of converting from unconfined to confined conditions. A desk study that consists of photointerpretation of the geomophological features of the Yobe floodplain by Marinof-Petkoff (1994) and hydrogeological and geophysical studies by Hassan (2002) have shown that there exist some areas that are covered by permeable deposits. In all the studies carried out in the area, there was difficulty in assessing the aquifer potential based on the existing field data. There was complete lack of historical data in some cases. In the cases where some data was available, there were problems of missing records. Some or all of the problems discussed above needed to be addressed in the context of existing data. The area where the study is conducted is characterised also with inadequate

Water resources of arid areas

602

data needed to carryout ‘conventional’ modelling. This limitation of data has constrained the development of a ‘full-scale’ model where calibration, verification and prediction are possible. It was against this background that an exploratory groundwater model was developed using MODFLOW to get an insight into the possible interactions highlighted in the conceptual model of Hassan (2002). 2 A SUMMARY OF THE CONCEPTUAL MODEL The River Yobe system as conceptualised in Hassan (2002) is shown in Figure 2. It consists of the followings: ● The aquifer geometry and the boundary show that the aquifer is 10m thick and 4km wide with the river almost in the middle. It has clay cover in some places whose thickness varies from 0.5

Figure 2. Conceptual model of the River Yobe system. to 3m. A no flow boundary condition in the north and a constant head in the south that allows small seepage to the upland bound it. ● The landforms show that the Yobe floodplain consists of areas that could allow vertical recharge. The flow processes in the aquifer are vertical recharge from rainfall and overland flooding and river to aquifer flow. ● Aquifer parameters such as storage coefficient and hydraulic conductivity cover a range of values. ● The river-aquifer interaction is represented with a varying river coefficient; and the magnitude of flow between the river and the aquifer depending on relative head difference between water in the river and groundwater in the aquifer.

Groundwater modelling with limited data

603

3 METHODOLOGY AND MODEL PREPARATION The idealized conceptual model was used to design and simulate various scenarios using the MODFLOW model (McDonald and Harbaugh, 1988). A combined pre and postprocessor model independent graphical interface called Groundwater Vistas was used for data input and interactive modelling with MODFLOW. The model area was discretized three-dimensionally. In the x-direction the size of the grid blocks were 500m, and in the y-direction it was variable with the smallest being 35m and the largest 150m. Each grid consists of 3 columns and 50 rows. In the vertical direction the model consists of one layer with a thickness of 16m. Row 25 contains the river. Input data to model consists of the three major external stresses river stage time series with a varying river coefficient, vertical recharge and ‘leakage’. Figures 3, 4 and 5 show the starting conditions and time series of the input data respectively. The input for one year consists of 36 stress periods each with a length of 10 days and a single time step. This was repeated for 2.4 years (86 stress periods) with stress period one starting from 30th October. The choice of the number and length of stress periods and time steps was dictated by the rapid change in the river stage. Sensitivity analysis was carried out on the wide range of aquifer parameters to arrived at acceptable values and were also assigned as follows: Unconfined region the horizontal hydraulic conductivity, Kh=15m/d reducing to 0.1m/d towards the southern boundary. Confined region: Kh=15m/d Confined storage coefficient, Sc=0.001 Unconfined storage coefficient, Su=0.05 The clay cover was modelled as a ‘leakage’ factor that allows water to seep continuously into the aquifer. A vertical leakage of 1.5×10−5m/d, equivalent to vertical hydraulic conductivity of clay, is used.

Figure 3. Starting conditions. A recharge value of 1.25mm/d was estimated using a water balance model (Hess, 1997), this is equivalent to 50mm of recharge per annum. This amount is consistent with

Water resources of arid areas

604

independent estimates by Carter et al (1994) and Edmonds et al (2002). The recharge was applied to rows 29 to 33 and in stress periods 32 to 35 inclusive. These stress periods corresponds to 10th–19th September to 10th–19th October respectively when recharge is believed to occur. The outputs from the model consisting of groundwater heads for each of the 86 stress periods were used for calculation of the various flow processes. The river to aquifer flow is calculated

Figure 4. River stage data.

Figure 5. Recharge and ‘leakage’.

Groundwater modelling with limited data

605

using the equation Qriv=Criv x HDIFF (Rushton and Tomlinson, 1979). Where Criv is the river coefficient and HDIFF is the relative head difference between water level in the river and the groundwater head in the aquifer. 4 RESULTS 4.1 The groundwater hydrographs 4.1.1 South of the river Figure 6 shows the modelled groundwater heads for node (27,2) located near the river compared with a field observed heads from piezometer P7. The figure suggests a strong influence of the river compared with heads far away from the river node. The plot suggests a good measure of representation of the processes taking place in the vicinity of the river. A side-by-side comparison of the modelled heads from the river with piezometers located at similar distances is also indicated in Figure 7. North of the river: in the largely confined north, the modelled groundwater heads show little variation near or far from the river.

Figure 6. Observed and modelled groundwater heads for node (27,2).

Water resources of arid areas

606

Figure 7. Groundwater heads at 600m south of the river: (a) modelled, (b) observed.

Figure 8. Groundwater flow from the aquifer to the north and south at the river node (25, 2).

Groundwater modelling with limited data

607

Figure 9. Groundwater flow from the aquifer to nodes (17, 2) and (32, 2), 637.5m from the river. 4.2 Groundwater flow in the aquifer The various groundwater flows in the aquifer were calculated using the groundwater heads, Darcy and the continuity equations. Figure 8 shows the flow from the aquifer node (25, 2) beneath the river to adjacent nodes north and south of the river. The results show that the flow to the confined north is much smaller than the flow to unconfined south. Similar results are shown in Figure 9 where the nodes are located more than 637.5m from the river in both directions. The figure suggests that the model has the ability to exhibit the rapidity and inertia of the confined and unconfined conditions obtained in the north and south respectively.

Water resources of arid areas

608

Figure 10. River to aquifer flow. 4.3 River-aquifer flow The modelled groundwater heads at the river node together with the input river stage and river coefficient were used to calculate the river to aquifer flows. Figure 10 shows the flow with positive values indicating flow from river to aquifer. The result indicated that the river is adequately represented because the aquifer responded to changes in the river level. It also shows that during recharge, the flow to the aquifer decreases even at high river level. This suggests that Qriv is limited both by the ability of the aquifer to transmit water and the magnitude of HDIFF. 4.4 The water balance The water balance consists of the difference between the total water flowing into the aquifer and the total water coming out of it. This in turn is equal to the change in storage. The inflow model consists of recharge, the ‘leakage’ through low permeable surfaces and the river to aquifer flow. The outflow consists of the aquifer to river flow during low river stage and boundary outflow. Figure 11 shows the time series plot of the water balance for one year. It indicates that the river-aquifer flow dominates all other inflow to the model; about 70% of the total inflow to the aquifer is from the river-aquifer interaction. 5 DISCUSSION AND CONCLUSIONS The basic and exploratory single layer model has demonstrated an ability to serve as an interpretation of observed field data. It has also reflected the physical processes presented in its conceptual model. The process of generating the conceptual model and fitting field

Groundwater modelling with limited data

609

data into the numerical model has provided the ‘feel’ or insight into the magnitudes of the system’s processes and interactions. Despite uncertainty in the estimates of some parameters such as river coefficients and aquifer parameters, the similarity between the model results and observed data is encouraging. For example the model has also established that in confined areas less water enters the aquifer and there is immediate response to changes in the application of stress when compared to the unconfined areas. The results from the model are plausible and that it has represented to some extent the understanding incorporated in its conception.

Figure 11. Water balance: Str=change in storage; Qriv=river to aquifer flow; Reah+leak=recharge and leakage. REFERENCES Alkali, A.G. 1995. River-Aquifer Interaction in the middle Yobe River Basin, Northeast Nigeria. Unpublished PhD thesis. Silsoe College. Carter, R.C. 1998. Prospects for Sustainable Water Management Policy in Sub-Saharan Africa, with Special Reference to the Northeast Arid Zone of Nigeria. In Water Resources Management, A Comparative Perspective. Ed. Dhirendra K.Vajpeyi. Carter, R.C., and Alkali, A.G. 1996. Shallow Groundwater in the Northeast Arid Zone of Nigeria. Quarterly Journal of Engineering Geology, Vol. 29, 341–355. Diyam Consultants. 1987. Kano State Shallow Aquifer Study. Final reports Vol. 1. P.O. Box 701, Kano, Nigeria. Edmunds, W.M., Fellman, E., Goni, I.B., and Prudhomme, C. 2002. Spatial and temporal distribution of groundwater recharge in northern Nigeria. Hydrogeology Journal (2000) Vol. 10. pp 205–215. Hassan, M. 2002. Exploratory groundwater modelling in data-scarce environments: The shallow alluvial aquifer of River Yobe Basin, North East Nigeria. Unpublished PhD Thesis Cranfield University UK.

Water resources of arid areas

610

Hess, T.M. 1997. BALANCE—A soil water balance model. Unpublished, Cranfield University, UK. Hess, T.M., Stephens, W. and Maryah, U.M. 1995. Rainfall Trends in the North East Arid Zone of Nigeria, 1961–1990. Agricultural and Forest Meteorology, 74, 87–97. IWACO B.V. International Water Supply Consultants. 1985. Study of the Water Resources in the Komadougou Yobe Basin. Report No. 5. Groundwater Resources. Rotterdam, Netherlands. Marinof-Petkoff, M.N. 1994. A Geomorphological study of the Yobe River Floodplain: Implication for Groundwater Recharge. Unpublished MSc. Thesis. Cranfield University. McDonald, M.G., and Harbaugh, A.W. 1988. A modular three-dimensional finite-difference groundwater flow model. Techniques of water resources investigations 06-A1, USGS, 576p. NEAZDP. 1990. Northeast Arid Zone Irrigation Project: Groundwater Resources Report. PMU, Garin Alkali, P.M.B 18. Gashua. Yobe State, N.E. Nigeria. Rushton, K.R. and Tomlinson, L.M. 1979. Possible mechanisms for leakage between aquifers and rivers. J. Hydrology, 40:49–65.

Theme G: Water resources management

Apple and grape vinegar application as csource in water denitrification Ş.Aslan Cumhuriyet University Department of Environmental Engineering, Sivas, Turkey A.Türkman Dokuz Eylül University, Department of Environmental Engineering, Izmir, Turkey Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: In this study, microbial denitrification of drinking water was studied. Treatability studies were conducted in a fixed bed reactor with synthetic water samples using locally available chemicals like apple and grape vinegar applying various C/N ratios. The effluent of the reactor is send to a slow sand filtration unit for the improvement of water quality.

1 INTRODUCTION Nitrate contamination in drinking waters is a growing problem for all over the world because of its harmful effects on human health. Nitrates cause methemoglobinemia in infants and pregnant women, which is also known as “blue baby syndrome”. Nitrates also cause the formation of n-nitroso compounds with amines and amides in the human body. These n-nitroso compounds are known to be carcinogens. To protect human health, public health agencies set nitrate limit in drinking water standards. There are various nitrate removal methods such as ion exchange, reverse osmosis, electrodialysis, distillation, chemical denitrification, mebrane biorector and biodenitrification (AWWA, 1989; Delanghe, 1994; Barreiros, 1998; Flora, 1994). Biological process has been shown to be practical, efficient and cost-effective. The majority of microbial denitrification treatment relies on heterotrophic bacteria which require an organic carbon source to reduce nitrogen oxides to nitrogen gas; but drinking water has low carbon content. Therefore an external carbon has to be supplied for microbial growth. If the organic substances in the nitrate contaminated water are below the stoichometric requirement for denitrification, they must be added in the form of acetic acid or ethanol resulting in increase in treatment costs. Several types of organic compounds have been used like methanol (Wasik et al., 2001; Gomez et al., 2000; Lee et

Water resources of arid areas

614

al., 2001; Hoek and Klapwizk, 1987), and acetic acid (Dahab and Kalagari, 1996; Bandpi et al., 1999). Though methanol assures the highest denitrification rate (Mansell and Schroder, 1999), it can constitute certain risk if the treated water is used for drinking purpose (Adriaan, 1992). The main disadvantage of acetic acid as compared to other carbon sources was its high consumption ratio and high cost (Bandpi et al., 1999). Balszczyk et al., (1981) pointed out that using acetic acid could have a significant effect on the production of nitrite in the reactor. Consequently, the use of ethanol as alternative is becoming more popular (Green et al., 1994; Gomez et al., 2000; Bandpi et al., 1999; Delanghe et al., 1994; Fonseca et al., 2000; Dahab and Sirigina, 1994; Richard, 1989, and Gomez et al., 2000). The possibility of using alternate substances such as volatile fatty acid (Yatong, 1996), shredded newspaper (Volokita et al., 1996a), wheat straw (Soares and Abeliovich, 1998, Aslan and Türkman, 2003), unprocessed short fibre cotton (Volokita et al., 1996b), atrazine (Stucki et al.,

Table 1. Composition of medium. KNO3 KH2PO4 NaHCO3 FeSO4*7H2O NaMoO4 MnSO4*7H20 CoCl2*6H2O H2O (pure)

1 L Medium 361mg (50mg/l as N) 150mg 32.5mg 0.816mg 0.2365mg 0.1565g 0.526g →1L

2000), natural gas methane (Rajapakse and Scutt, 1999), elemental sulfur (Eisentraeger et al., 2001; Soares, 2002), sugar or glucose syrup (Nurizzo and Mezzanatte, 1992) and sugar cane (INCO-DC, 2000) have also been studied in the biological denitrification processes. The aim of this study is to determine the applicability of the biological denitrification of drinking water using locally available chemicals as carbon source, apple and grape vinegar. Denitrification efficiencies have been determined by applying different carbon sources in up-flow packed column. Second step of the treatment is sand filtration, which is not used commonly in Turkey, even though it is very suitable because of its simplicity of operation, availability of sand and the advantage of removing microorganisms that is formed in denitrification phase. 2 MATERIALS AND METHODS 2.1 Analytical techniques Effluent samples from the up-flow fixed film reactor were collected on daily basis and tested for temperature, pH, and turbidity, NO3-N, NO2-N and COD. Prior to analyse, samples withdrawn from the reactor were filtered by 0.45µm, white 47mm radius filter to

Apple and grape vinegar application as c-source in water denitrification

615

hinder the interference of colour and turbidity. COD analyses were performed with the clear samples according to APHA (1984). NO2-N (14776) and NO3-N (14773) were determined using analytical kits and a photometer Merck SQ 300. Dissolved oxygen (DO) measurements were carried out using WTW oxygen meter. Turbidity was measured by the turbidity meter in the unit of JTU (Heck Chemical Company). 2.2 Synthetic medium composition The liquid medium used, consisted of a mineral base media supplemented with nitrate as sole electron acceptor and ethanol as donor. The composition of the medium is given in Table 1. CoCl2 was added to get rid of dissolved oxygen in water and also the dihydrogen potassium phosphate was added as a P source as well as to provide buffering capacity. The final pH is adjusted to 6–8 intervals. The amount of carbon added was calculated by stoichiometric relationship with nitrate. 3 START-UP Inoculum was taken from a denitrification reactor used in the laboratory. For the first three days, the reactor was fed daily in a fill and draw mode, with recirculation. From the third day the reactor was fed continuously. It was considered that the reactor had reached state conditions when the NO3-N removal efficiency reached to the higher than 90%.

Water resources of arid areas

616

Figure 1. Experimental set-up. 4 EXPERIMENTAL SET UP Experiment was carried out in an up-flow fixed film reactor operated in continuous mode. The packed column was filled with 10mm pieces of plastic coils materials, which supported bacterial growth. The biodenitrification reactor consisted of a cylindrical glass reactor, 5.4cm in diameter and 50cm in height, completely submerged an effective volume of 0.841. The sand filter column, which had 8cm diameter and 30cm height, was filled with filter sand of an effective diameter of 0.4mm and uniformity coefficient of 0.89 (Figure 1).

Apple and grape vinegar application as c-source in water denitrification

617

5 RESULTS AND DISCUSSION 5.1 Apple vinegar experiment In this series, apple vinegar was used as carbon source, which has 61600mg COD/1. Medium solution was pumped through the reactor with the hydraulic retention time (HRT) of 0.9 hour. NO3-N concentration was adjusted to 50mg/l. 85% NO3-N removal was achieved during the first two days and in this time period nitrite accumulation was observed. This can be explained by high reduction rate of nitrate (Figure 2). The biggest problem determined in this series is the clogging in denitrification reactor because of high biomass formation. At the end of five days, the COD concentration in the effluent was 24mg/l. It was decided that high carbon concentration was causing the clogging problem. NO3-N concentration decreased gradually in four days and more stable conditions were reached. At the end of experimental period, high biomass formation was observed. As a result, turbidity has also increased during this four days period. In the second step, experiment was performed by lower carbon addition. C/N ratio was kept as 3.0. In spite of decrease in influent COD concentration, high biomass formation caused clogging in the reactor.

Figure 2. NO3-N and NO2-N removal in the reactor using apple vinegar as Csource.

Water resources of arid areas

618

Figure 3. NO3-N and NO2-N removal in the sand filter. Figure 3 illustrates the variation of NO3-N and NO2-N concentrations in the sand filter effluent. As can be seen from Figure 3, sand filter acts as denitrification reactor. In the sand filter about 64% NO3-N removal was determined. Although NO3-N concentration in the effluent is low (below 10mg/l NO3-N) most of the time, nitrite concentrations were high and cannot be considered acceptable considering nitrite standard (0.1mg/l). Denitrification brought about by anaerobic bacteria results in a significant loss (up to 45%) of nitrogen. The removal was achieved by biomass formed in the bed. The sand filtration unit behaved like additional biodenitrification reactor. Turbidity measured in the effluent was 10 JTU units. 5.2 Grape vinegar experiment Grape vinegar has the advantage of availability. Also its residual will not cause health problems. Experiments carried out various C/N ratios by applying nitrate-nitrogen concentration were 50mg/l. The biodenitrification microorganisms were acclimated to the grape vinegar for two days. Nitrate nitrogen removal efficiency in the reactor was 91% on the average (Figure 4). Although the nitrate concentration in the treated water below the standards, COD concentration remains above the standards (Figure 5). Decreasing the C/N ratio to 1.25 did not improve the effluent water quality but resulted in higher nitrite accumulation due to the insufficient carbon concentration in the influent. Because of the high C/N ratio excess amount of carbon remain in the effluent of the reactor and high nitrogen removal (about 60%) occurred in the sand filter unit (Figure 6).

Apple and grape vinegar application as c-source in water denitrification

Figure 4. NO3-N and NO2-N removal in the reactor using grape vinegar as Csource.

Figure 5. NO3-N removal and residual C concentration in the effluent water using grape vinegar as C-source.

619

Water resources of arid areas

620

Figure 6. NO3-N and NO2-N removal in the sand filter using grape vinegar as C-source. 6 CONCLUSIONS Based on the results, the following conclusions may be drawn; 1. Bio-denitrification seems to be a very effective method of removing nitrates from contaminated drinking water supplies. Using fixed film bio-denitrification, high nitrate-N removal efficiencies were achieved despite very short detention times (as 1.19 and 0.9 hours). 2. The combined sand filter system was very effective polishing denitrified effluent. The denitrified effluent met drinking water criteria with respect to nitrate and nitrite. However, post treatment was needed to ensure the bacterial and COD requirements. 3. The nitrate-N and nitrite-N concentrations decreased to limit value after a few days of acclimation in sand filter unit. 4. Application of vinegar as compared to the other chemicals has the cost advantage in Turkey. 5. Commonly used denitrification systems are ion exchange, reverse osmosis, activated carbon and biodenitrification. But biodenitrification systems have additional advantage of lower installation, operation and maintenance cost, besides its high nitrate removal efficiency.

ACKNOWLEDGEMENT The authors are thankful to INCO-DC Project (Contract ERBIC 18 CT 970167) for financial support.

Apple and grape vinegar application as c-source in water denitrification

621

REFERENCES Adriaan, H.S. 1992. Metabolic Pathways in Paracoccus Denitrificans and Closely Related Bacteria in Relation to the Phylogeny of Procaryotes. Antonie van Leeuwenhoek 61:1–33. APHA Standart Methods for the Examination of Water, Sewage and Industrial Wastes. 1984. 16th edn, American Public Health Association/American Water Works Association/Water Environment Federation, Washington DC, USA. Aslan, Ş. & Türkman, A. 2003. Biological Denitrification of Drinking Water by Using Various Natural Organic Solid Substrates. Water Science and Technology. 48. 11–12:489–495. AWWA. 1989. American Water Works Association Water Desalting and Reuse Committee, Membrane Desalting Technologies. Journal, AWWA, 81, 30. Barreiros, A.M., Rodrigues, C.M., Crespo, J.P.S.G. & Reis, M.A.M. 1998. Membrane Bioreactor For Drinking Water Denitrification. Bioprocess Engineering 18. Bandpi, M.A., Elliott D.J. & Memeny-Mazdek A. 1999. Denitrification of groundwater using acetic acid as a carbon source. Water Science and Technology. 40(2):53–59. Balszczyk, M, Przytocka, M, & Mycielski, R. 1981. Denitrification of High Concentrations of Nitrites and Nitrates in Synthetic Medium with Different Sources of Organic Carbon. Acta Microbiologica Polonica. 30:49–58. Dahab, M.F. & Sirigina, S. 1994. Nitrate removal from water supplies using biodenitrification and gac-sand filter system. Water Science and Technology. 30(9):133–139. Dahab, M.F. & Kalagari, J. 1996. Nitrate removal from water using cyclically operated fixed film biodenitrification reactors. Water Science and Technology. 34(1–2):331–338. Delanghe, B., Nakaruma, F., Myoga, H., Magarat, Y. & Guibal, E. 1994. Drinking water denitrification in a membrane bioreactor. Water Science and Technology. 30(6):157–160. Eisentraeger, A., Klag, P., Vansbotter, B., Heymann, E. & Dott, W. 2001. Denitrification of groundwater with metan as sole hydrogen donor. Water Research. 35(9):2261–2267. Flora, J.R.V., Suidan, M.T., Islam, S., Biswas, P. & Sakakibara, Y. 1994. Numerical Modeling Of a Biofilm-Electrode Reactor Used For Enhanced Denitrification. Water Science and Technology. 29:10–11. Fonseca A.D., Crespo J.G., Almedia I.S. & Reis M.A. 2000. Drinking water denitrification using a novel ionexchange membrane bioreactor. Environmental Science Technolgy. 34:1557–1562. Gomez, M.A., Gonzales-Lopez, J. & Hantorie-Garcia, E. 2000. Influnce of carbon source on nitrate removal of contaminated groundwater in a denitrifying submerged filter. Journal of Hazardous Materials. B80:69–80. Green, M., Schnizer, M., Tarre, S., Bogdan, B., Shelef, G. & Sorden, C.J. 1994. Groundwater denitrification using an upflow sludge blanket reactor. Water Research. 28(3):631–637. Hoek, J.P. & Klapwizk, A. 1987. Nitrate removal from groundwater. Water Research. 21(8):989– 997. INCO-DC. Project 2000. Development of a simple technology in drinking water treatment for nitrate and pesticide removal, 7. Meetings, Rabat, Morocco. Lee, D.U., Lee, S., Choi, D. & Bae, J. 2001. Effects of external carbon source an empty bed contact time on simultaneous heterotrophic and sulfur-utilizing autotrophic denitrification. Process Biochemistry. 36: 1215–1224. Mansell, B.O., & Schroeder, E.D. 1999. Biological Denitrification in a Continuous Flow Membrane Reactor. Water Research. 33(8):1845–1850. Nurizzo, C. & Mezzanatte, V. 1992. Groundwater biodenitrification on sand fixed film reactor using sugars as organic carbon source. Water Science and Technology. 26:827–834. Rajapakse, J.P. & Scutt, J.E. 1999. Denitrification with natural gas and various new growth media. Water Research. 33(18):3723–3734.

Water resources of arid areas

622

Richard, Y.R. 1989. Operating Experience of Full-scale Biological and Ion-exchange Denitrification Plants in France. Journal Inst. Water Environmental Management. 3:154–165. Soares, M.I.M. & Abeliovich, A. 1998. Wheat straw as substrate for water denitrification. Water Research. 32(12):3790–3794. Stucki, G., Yu, C.W., Baumgartner, T. & Gonzales-Valero, J.F. 1995. Microbial atrazine mineralization under carbon limited and denitrifiying conditions. Water Research. 29(1):379– 385. Volokita, M, Belkýn, S., Abeliovich, A. & Soares, M.I.M. 1996a. Biological denitrification of drinking water using newspaper. Water Research. 30(4):965–971. Volokita, M., Abeliovich, A. & Soares, M.I.M. 1996b. Denitrification of groundwater using cotton as energy source. Water Science and Technology. 34(1–2):379–385. Wasik, E., Bahdziewicz, J. & Blasszczyk, M. 2001. Removal of nitrates from groundwater by a hybrid process of biological denitrification and microfiltration membrane. Process Biochemsitry. 37:57–64. Yatong, X. 1996. Volatile fatty acids carbon source for biological denitrification. Journal Environmental Science. September, 8(3):257–269.

Water resources management in the National Park, central Australia E.R.Rooke Consultant Hydrogeologist, Australian Groundwater Technologies Pty Ltd., Adelaide, Australia Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Permanent surface-water occurs only beneath rock domes within National Park (UKTNP). Spasmodic, extreme rainfall causes floods. Marginally fresh to saline, high nitrate, groundwater occurs within Cainozoic (Cz) sediments and weathered/ fractured bedrock in the Dune Plains Aquifer (DPA) and Southern Aquifer (SAQ). The Cz acts as a reservoir whilst bedrock transmits groundwater to water-supply bores. Groundwater flows north-eastwards, along a palaeovalley, towards a groundwater sink (Lake Amadeus saltlake). Recharge occurs after exceptional rains.

1 LOCATION, PHYSIOGRAPHY, VEGETATION, WATER SUPPLY AND WATER POLICY UKTNP occupies vast, flat, sandy plains and dune fields (the ‘Dune Plains’), and steep rock domes of (‘Ayers Rock’) and (The Olgas’) and lesser rock outcrops. Stable, parallel, reticulate and irregular dunes rise an average of 10m above the plain. The plains slopes gently SW to NE, descending about 2.5m every 1km to Lake Amadeus, a large salt lake nearly 60km away, and a groundwater discharge zone for UKTNP aquifers’ (and for other aquifers, north of Lake Amadeus).

Water resources of arid areas

624

Figure 1. Location of UKTNP and Yulara (source: Department of the Environment & Heritage, Australian Government). UKTNP is home to approximately 450 people. They rely on potable water drawn from the nearby local aquifer, known as the Southern Aquifer (SAQ). 2 GROUNDWATER The groundwater resource is two layered; a Cz sedimentary aquifer overlies fractured bedrock aquifers of the sedimentary Amadeus Basin. Groundwater flows at a rate as little as 1m/year from SW to NE along the Dune Plains palaeovalley to Lake Amadeus, where groundwater intersects the surface/sub-surface and evaporates. Lake Amadeus is a saltlake or salina and a major evaporative sink for central Australia. Many water supply production bores (PBs) are screened against weathered bedrock/bedrock that, essentially, is used as a collector system draining the Cz that provides storage, although PBs have been constructed at depths less then 70m within sediments. To predict long-term aquifer performance, it is important to map the base of the Cz, to enable determination of the volume of water that can be withdrawn from

Water resources management

625

storage (Read, 1999; File Note). However, the position of the contact between Cz and extremely weathered bedrock is uncertain. Near the surface massive, karstic, calcrete with moderate water-bearing capacity occurs. In other places, calcrete is impermeable and perched water tables may develop; e.g. near the New Wellfield at Yulara, a layer of fresh water in the aquifer (modern recharge) overlies more saline water. Calcrete dissolution may directly contribute to higher TDS of groundwater; where cavities are silicified this phenomenon is mitigated. Aquifer recharge has occurred three or four times in the past 37 years (of recordkeeping). After each recharge event there has been prolonged recessions of the water table. 2.1 The Southern Aquifer The SAQ is situated immediately south of . Its geometry is poorly-defined; from drilling its known area is about 6km2 Drill-holes are capable of airlift yields >2L/s. PBs in the eastern part tap Cz sediments, whilst in the west, PBs tap bedrock. The central and western parts of the SAQ manifest substantial undulation of the bedrock/Cz interface with the thickness of Cz sediments varying from about 30m to 160m (Wischusen, 1999).

Table 1. Geohydro-stratigraphy (after Read, 1999; file note). Ground level to a depth 10 and 90m— Cainozoic (Cz) clay, sand and silt; some silcrete.

Where saturated, the volume of water that can be extracted from the sands and silts has been estimated to be between 5 and 10% of the volume the sands and silts occupy. Some permeable sands occasionally occur towards the bottom of the sequence. Below the Cainozoic The volume of water that strata occurs between can be extracted from the 10 and 50m of extremely weathered extremely weathered Proterozoic rocks has been rock of Proterozoic estimated to be between 1 age (Pinyinna and 3% of the volume the beds)—has the sands and silts occupy. appearance of white Low permeability. clay, silt or sand when drilled with a rotary dril. The volume of water that Moderately can be extracted from these weathered rocks has been estimated to Proterozoic aged be less than 0.5% of the strata (Pinyinna volume the rocks occupy. beds) (‘bedrock’) The permeability is high comprised of

Water resources of arid areas

interbedded shale, dolomite, chert and sandstone occurs beneath the extremely weathered rock.

626

when the rocks are dolomite, chert or sandstone and low when shale is intersected. Most production bores are in this sequence.

Transmissivities and SWL responses in the Cz or bedrock aquifer are similar in the centre and west whereas, in the south and east, water chemistry, SWLs, and drilling suggest a separate aquifer in the bedrock. A groundwater hydrograph against local rainfall is shown in Figure 8. After recharge events in 1974, 1982, 1989 and 2000, prolonged recessions of the water table occurred. At RN 11547, the SWL is some 3m higher at present than it was 20 plus years ago. TDS increases towards the SE, i.e. along the flow path of the aquifer (Wischusen, 1999) and in bedrock is variable (TDS >3,000mg/L in places). The groundwater is at or near saturation with respect to bicarbonate (300mg/L HCO3) (ibid). The shallow aquifer zone within the Cz has low chloride and low sulphate waters; the intermediate aquifer zone (weathered bedrock?) has relatively low chloride and high sulphate and the deepest aquifer zone (bedrock?) has high chloride and relatively high sulphate. Nitrate is generally low in the main aquifer zone. A high bicarbonate to chloride ratio in the NW part of the SAQ is likely to indicate a recharge zone (Read, 1977 and Jolly, 1979). The north-central part is associated with floodouts emanating from the southern face of that cross the main Ayers Rock Fault, so indirect recharge down the associated fracture-zone may be possible. Isotope chemistry suggests that the bulk of stored water entered the aquifer in the geologically-recent past (last interglacial period with a wetter climatic regime) and relatively insignificant recharge occurs under present hydrological conditions. Tritium (3H) indicate component of modern water in the saturated zone (Jacobson, 2000), including a bore close to (<50 years). Direct recharge of the SAQ, is indicated by the relative enrichment of the Oxygen 18 isotope concentration in groundwater (Jacobson et al., 1989b). This suggested that a high degree of evaporation had occurred during recharge. Some samples were ‘ancient’ groundwaters, recharged more than 30,000 years ago. 2.2 Sustainable yield As the aquifer zone is poorly-defined, aquifer throughflow is difficult to calculate. Wischusen (1999) examined monitoring data, short-term test pumping and sustainable PBs’ production rates. This gave a sustainable yield of 200ML/yr with 50 to 75ML/yr cited as a sustainable extraction rate from a typical PB. A specific yield (Sy) of 0.018 was calculated by volume of water yielded per metre of drawdown/aquifer area dewatered (Wischusen, 1999); and in combination with the saturated depth to bedrock as mapped by Wischusen (1999), an aquifer storage volume of about 6,000ML was estimated. This is a conservative estimate and does not include groundwater stored in the bedrock in the north-west of the SAQ or groundwater stored below Cz in the bedrock (ibid).

Water resources management

627

The indicated sustainable yield of 200ML/yr represents 3% of the estimated storage of 6,000ML. The storage to yearly extraction ratio (30:1) is large ensuring that there is a buffer of large amounts of groundwater in storage within the SAQ, i.e. should recharge diminish to zero, about 30 years of production could still be sustained (Wischusen, 1999). 2.3 The Dune Plains Aquifer The decision to locate a new tourist village, Yulara, outside UKTNP, led the Northern Territory Government to fund regional groundwater investigations commencing 1978 to source a reliable, potable supply of water. Read (1978) produced the first conceptual hydrogeological model of the DPA, updated by English (1998). The distribution, continuity, and flow mechanisms of the DPA are not well understood. Generally, SWLs range from about 40m bgl between and Kata to 12m bgl north of Yulara. Groundwater flow is towards the NE. The major features of interest interpreted from this follow: ● SWLs in 2001/02 generally 3m to 4m higher than at the commencement of monitoring in the late 1970’s; ● a marked positive response to the 1989 flood and a more subdued (currently rising) positive response to the 2000 flood; and,

Figure 2. Water balance for Yulara and the Dune Plains Aquifer (DPA). ● a dampening of positive response to rainfall-recharge events with increasing depth to the water table at each MB.

Water resources of arid areas

628

The potentiometric surface showed an 8m fall in SWLs from the SW end of the wellfield to the NE end Coffey (1999). The DPA has a somewhat higher groundwater salinity than the SAQ and is chloride-rich rather than bicarbonate-rich. When SWLs decline, due to long periods of limited or no recharge, groundwater quality appears to deteriorate. Isotope dating (14C, 2H and 18O) of groundwaters in the DPA (Jacobson et. al., 1989 and 2000) indicated modern groundwater ages between less than 500 years and 1,200 years with two dates greater than 5,000 years from mixing of young, local recharge (probably less than 50 years old) and older groundwaters flowing along the palaeovalley. 2.4 Sustainable yield The uncertainty regarding throughflow contribution from weathered bedrock/bedrock, the complex nature of flow patterns (braided sand horizons with enhanced permeability; regional flow along the palaeovalley and W to E cross-aquifer flow from the Sedimentaries runoff-recharge) makes throughflow calculation problematic. DIPE (File 58.3.P3) quote a throughflow of 200ML/ year. In addition to this estimated substantial storage volume of marginally potable to brackish groundwater of the DPA, there are potentially useable groundwater resources to the N and E of Yulara. 3 WATER USE AND PRODUCTION WITHIN THE UKTNP Within the UKTNP water, sewerage and waste disposal services are provided by Parks Australia to the communities of Mutitjulu, Maruku and Rangerville, the Cultural Centre and Ranger Station, the ‘Base of the Climb’ at , and the Park Entry Station. Three bores situated in the SAQ west of Mutitjulu supply water to Mutitjulu/Maruku, the Cultural Centre and Rangerville. The safe yield of the wellfield (approximately 6.5L/s) (Wischusen, 1999) will be exceeded by demand in approximately 20 years at 4% annual demand growth (QANTEC, 2002). 4 WATER USE AND PRODUCTION AT YULARA The Yulara wellfield situated in the DPA consists of five PBs each producing between 5L/s and 25L/s, including one acting in a standby capacity. The hardness, salinity (TDS about 1500mg/L) and high nitrate content of the groundwater requires treatment for potable supply. Approximately 1.1ML/day is treated to 200mg/L TDS by reverse osmosis (RO) desalination. Some 25% (i.e. about 0.3ML/day) of the total water put through the RO issues as reject brine and is discharged to an inter-dunal area between the township and Yulara airport. 30% of wellfield production (0.5ML/day; up to 2,100mg/L TDS), raw water is supplied directly to Yulara township for irrigation, sanitation, and fire fighting use. There has been an average increase in water consumption of approximately 10% per annum. In 2002, production was 707ML. Historically, larger demands occurred from September to January. This demand cycle does not directly correlate with the temperature

Water resources management

629

variation for Yulara, and is probably due to a combination of tourist numbers and higher temperatures (Coffey, 1999). From 1998 to present there is a noticeable decrease in water supply seasonality, with the wellfield operating at historically higher outputs throughout the entire year, consistent with visitor numbers increasing over the summer period (AGT, 2003). The present water demand at Mutitjulu, Maruku and Rangerville is nearly 55ML/yr (QANTEC, 2002). However the actual water production in year, 2001 was more than 70ML. This infers that there is unaccounted for water (system leakage and wastage) and/or irrigation is being applied (possibly to the sports oval?). Taking the higher population growth rate, in year 2012, a demand of, say, 120ML/yr is considered sustainable (AGT, 2003). It is less than the estimated aquifer throughflow of 200ML/yr and represents only 2% of the estimated aquifer storage of 6,000ML. 5 AQUIFER RISK ASSESSMENT TO CONTAMINATION The BOD (oxygen removing capability) of raw effluent from Mutitjulu and Yulara sewerage discharge is similar to that elsewhere (150–200mg/L) but after treatment this has been reduced to 9mg/L, over 90% reduction at Mutitjulu (Allen & Assoc., 1999), and likely similar level of treatment at Yulara. This concentration of BOD is quite low, and the level of treatment consequently is considered effective. The main problems associated with high nutrient levels are production of nuisance algal blooms in surface waters. Apart from algal growth in the stabilization ponds (Allen & Assoc., 1999), this is unlikely to be a problem at Mutitjulu, and as effluent is reused for irrigation at Yulara, this is considered to be beneficial to vegetation being irrigated. An estimated 5ML of leachate could be expected to be produced each year from a 12ha landfill area under conditions within the Yulara area (GHD, 1993). This was based on an anticipated infiltration rate of rainfall of 42mm/yr, given local climatic conditions. As the landfill serving Mutitjulu is unlined, leachate generated within the landfill will percolate into the unsaturated zone. Natural processes of geo-purification will assist in reducing concentrations of contaminants in percolating leachate. The high concentrations of contaminants in leachate could impact groundwater if there is a hydraulic connection with regional groundwater beneath the clay sediments within the palaeochannel. Flow paths within the palaeochannel are likely to be complex, and the time taken for leachate to travel from the landfill to groundwater would be long, unless channelling of leachate occurred. Under conditions where no channelling took place, bacteria would remove organic contaminants, ammonium would be adsorbed on aquifer minerals and transport of this through the unsaturated zone would be much slower than that of leachate (i.e. ammonium would be attenuated). Groundwater quality protection guidelines recommend against locating fuel storages within capture (protection) zones of water supply wells to avoid contamination, as cleanup of soil once contaminated with these is extremely difficult and very expensive. The only known potential risk is at the Old Wellfield (DPA), where diesel fuel for the PBs is stored in elevated steel tanks which are bunded to reduce the possibility of bulk fuel contamination of the DPA.

Water resources of arid areas

630

Reid et al. (1993), Baker & Jarman (1995) suggested that the Mulgara (a marsupial rat) habitat and the UKTNP’s aquifers were, in some way, ecologically connected. Jacobson (1996) reviewed the palaeodrainage and proposed that the depth to the water table at the Mulgara habitat might be less important than its proximity to run-off from the Sedimentaries through the adjacent extensive Mulga shrub-land. The review emphasised the significance of this sheetwash zone near the northern boundary of the Park may have in concentrating surface water and nutrients in the transitional zone between Mulga and sand-plain, the core Mulgara habitat. Several rare and endangered fauna are associated to some extent with these run-on mulga groves, including the Mulgara Dasycercus cristicauda, Great Desert Skink Egernia kintorei and Hairy-Footed Dunnart Sminthopsis hirtipes (Reid & Hobbs, 1996). Earlier assessments on the distribution of Mulgara in the area suggested that the distribution of this endangered mammal coincided with the bands of mulga (e.g. Reid et al., 1993; Baker & Jarman, 1995). Subsequent work suggests Mulgara are linked more closely to hummock-forming spinifex Triodia basedowii (P.Masters, pers. comm.; J.Reid, pers. comm.; Reid & Hobbs, 1996). T.basedowii grows in or near run-on areas. T.basedowii, however, is unlikely to have deep roots that tap into the groundwater, since these plants show significant dieback during extended droughts (P.Masters, pers. comm., J.Reid, pers. comm.). This dieback would not occur if these plants were permanently tapping groundwater. 6 CONCLUSIONS WITH RESPECT TO AQUIFER SUSTAINABILITY The aquifers of UKTNP and Yulara have a huge storage of groundwater that is large in proportion to throughflow, recharge and usage. Wellfield management may, inevitably, require the resource to be mined due to the rare, but significant recharge events experienced in this semi-arid environment. Because of the extremely variable nature of arid zone rainfall amounts/intensities, prediction of rainfall-recharge return period is problematic for aquifer replenishment. Hydrographs/rainfall relationships empirically suggest that rainfall events would need to be in excess of 180mm in any given month to induce significant aquifer recharge, although this would depend on a number of variables including rainfall intensity/duration/area, antecedent soil moisture conditions, etc. SWL monitoring data for bores has shown a number of recharge events during the past 30 years, but most of the water in the aquifer was emplaced in earlier events over a long period of time. 7 RECOMMENDATIONS A water conservation and re-use strategy should be enhanced and formalised for Yulara (ARR) and

; to be re-viewed every ten years.

Water resources management

631

● core drilling, and petrophysical analysis thereof, to provide stratigraphic control of the aquifers to enable the determination of aquifer geometry, storage capacity (permeability and specific yield); ● review of pre-existing test pumping data to confirm hydraulic parameters; ● hydrochemical analysis and facies differentiation to determine flow paths and aquifer recharge and discharge zones; ● samples taken for isotopic dating at each level of the aquifer to determine residence times and the nature of recharge, i.e. diffuse (‘direct’) or flood-out (‘indirect’); ● suite of time-constant reduced SWL measurements recorded to construct a potentiometric surface and ‘flow-net’ to calculate throughflow; ● ‘infill’ of drilling with surface geophysics (gravity, microgravity, transient electromagnetic to comprehensively map the DPA and the SAQ); ● aquifer numerical modelling to examine sustainability and optimise operational pumping publish a hydrogeological map with accompanying notes; ● risk assessment of the vulnerability of the aquifer to extended periods between critical rainfall/ recharge events; ● observe and record flood events to assist in determining recharge return periods and volumes potentially available for recharge to aquifers; and, ● an aquifer vulnerability study (accounting for planning proposals) to mitigate against contamination risk; ● investigate plant physiology by measuring plant water potential at various times of day to assess the likely effects of changes in water levels and composition of the aquifer(s) on the terrestrial flora and fauna isotopic results from possible water sources tapped by plants should be compared with the isotopic composition of their shoots or twigs. This would characterise and identify the water source from which they are extracting water; ● research groundwater dependent ecosystems at UKTNP (little is known about subterranean ecosystems in the semi-arid zone).

REFERENCES Allen, N. & Associates, 1999. Mutitjulu Community: Wastewater Management and Proposed Sporting Oval. Report prepared for Parks Australia. Anderson, V.J. & Hodgkinson, K.C. 1997. Grass-mediated capture of resource flows and the maintenance of banded mulga in a semi-arid woodland. Aust. J. Bot. 45:331–342. Australian Groundwater Technologies, 2003. Aquifer Review, 2002— National Park. Report prepared for Parks Australia North, Environment Australia. Ayers Rock Resort, 2000. Ayers Rock Resort Five Year Plan 21/12/00. Baker, L. & Jarman, P. 1995. A conservation strategy for the Mulgara, Dasycercus cristicauda, at National Park, N.T. Report #76 for the Australian Nature Conservancy, Canberra. Coffey Geosciences, 1999. Yulara Borefield Operational Review. Report to the Northern Territory Power and Water Authority. DIPE (Department of Infrastructure, Planning and Environment, Northern Territory Government) File 58.3P3, Alice Springs, NT.

Water resources of arid areas

632

English, P. 1997a. Palaeodrainage Mapping and Hydrodynamics, National Park. Australian Geological Survey Organisation, Department of Primary Industries and Energy, 1997. National Park,

English, P. 1997b. Review of the Palaeodrainage Model for Northern Territory; report to accompany GIS Coverages for Park, NT. Australian Geological Survey Organisation.

National

English, P. 1998. Cainozoic Geology and Hydrogeology of National Park, Northern Territory. Australian Geological Survey Organisation, Canberra. Gutteridge Haskins & Davey Pty Ltd (GHD), 1993. Proposed Yulara Landfill Hydrogeological Assessment. Report prepared for the Ayers Rock Resort. Jacobson, G., Calf, G.E., Jankowski, J. & McDonald, P.S. 1989a. Groundwater Chemistry and Palaeo-recharge in the Amadeus Basin, Central Australia. Journal of Hydrology. 109(3–4):237– 266. Jacobson, G., Lau, G.C., McDonald, P.S. & Jankowski, J. 1989b. Hydrogeology and Groundwater Resources of the Lake Amadeus and Ayers Rock Region, Northern Territory. Bulletin 230. Department of Primary Industries and Energy, Bureau of Mineral Resources, Geology and Geophysics, Canberra. Jacobson, G. 1996. The Interrelationship of Hydrogeology and Landform in Central Australia. In S.R.Morton and D.J.Mulvaney (eds). Exploring Central Australia: Society, the Environment and the 1894 Horn Expedition. Surrey Beatty and Sons, Chipping Norton. , Central

Jacobson, G., Cresswell, R. & Hostetler, S. 2000. The Age of Groundwaters at

Australia. Report to National Park, April 2000. Bureau of Rural Sciences Report (Unpublished). Jolly, P. 1979. Ayers Rock Water Supply Investigation of Southern Aquifer System; Project 45, 1977. Water Division, Northern Territory Department of Transport and Works, Darwin. Knott, G.G. 1980. Yulara Village Water Resources 1978–1980 Investigation; Report 79; NT Department of Mines and Energy. Parks Australia, 2000. Plan of Management, QANTEC, 2002.

National Park, 2000.

Borefield—Development Strategies. Strategy for Future Upgrading of

Park Bores and Bulk Water Supply for Mutitjulu Communities, Cultural Centre, Headquarters. Report for Parks Australia North. Read, R.E. 1978. The Geology and Hydrogeology of the Ayers Rock Area. NT Department of Mines and Energy. Read, R.E. 1999. File Note. Department of Infrastructure, Planning and Environment, Northern Territory Government, Alice Springs, NT. Reid, J.R.W. & T.J.Hobbs (eds) 1996. Monitoring the Vertebrate Fauna of

—

National Park, Phase II-Final Report. Consultancy report to the Australian Nature Conservation Agency. CSIRO, Alice Springs. 380 pp. Reid, J.R.W., Kerle, J.A. & Morton S.R. 1993.

fauna: the distribution and abundance of

vertebrate fauna of (Ayers Rock-Mount Olga) National Park, N.T., Kowari, vol 4, Australian National Parks and Wildlife Service, Canberra. Slatyer, R.O. 1961. Methodology of a water balance study conducted on a desert woodland (Acacia aneura) community. Arid Zone Research 16:15–26.

Water resources management

Sweet, I.P. & Crick, I.H. 1992. Geological Survey Organisation.

and

633

—A Geological History. Australian

Wischusen, J. 1999. Hydrogeology of the Southern Aquifer,

: A Preliminary Assessment of

the Mutitjulu Groundwater Resource. A report to the National Park. Young, D.N., Duncan, N., Camacho, A., Ferenczi, P.A. & Madigan, T.L.A. 2002. AYERS ROCK SG 52–8. 1:250,000 Geological Series. Edition 2. Northern Territory Geological Survey. Young, D.N., Duncan, N., Camacho, A., Ferenczi, P.A. & Madigan, T.L.A. 2002. AYERS ROCK SG 52–8. 1:250,000 Geological Series. Explanatory Notes. Northern Territory Geological Survey.

Integrated water resources management and agriculture in southern Africa M.McCartney & H.Sally International Water Management Institute, Pretoria, South Africa A.Senzanje University of Zimbabwe, Mt Pleasant, Harare, Zimbabwe Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: More than 80% of annual water withdrawal in southern Africa is for agriculture. Much of the region experiences semi-arid to arid climates and in many countries water management is complicated by the frequent recurrence of droughts. This paper presents the concept of Integrated Water Resources Management (IWRM) which has been widely proffered as a modern approach to water management that will ensure long-term benefits whilst simultaneously protecting the environment and ensuring sustainability. The implications of IWRM for agriculture in southern Africa are assessed. Benefits for the agricultural sector include reduced vulnerability to climatic variability, improved productivity and a more rational and transparent approach to decision-making. However, there remain many constraints to the application of IWRM including lack of quantitative information pertaining to water use and a wide range of institutional and other socio-economic barriers. An evaluation of changes required to introduce IWRM within the agricultural sector is presented.

1 INTRODUCTION For many years experts have debated the capacity of world’s agricultural systems to produce enough food and fibre for an expanding population. It is estimated that at present there are approximately 840 million undernourished people in the World of whom some 210 million live in sub-Saharan Africa (FAO, 2003). This situation led the 1996 World Food summit to set a goal, reaffirmed at the 2000 Millenium Summit, of halving the number of hungry people by 2015. Fulfilling these objectives will have significant implications for water use in the region. At present, irrigated agriculture, which covers approximately 275 million hectares globally, produces 40% of all the world’s food crops. However, less than 5% (i.e., just 13 million hectares) of the total irrigated area is in Africa (Weligamage et al., 2002). To meet the millennium goal the World Bank estimates that globally irrigated agriculture

Integrated water resources management and agriculture in southern Africa

635

will have to grow at a rate of about 4% per year. The largest percentage rise is required in sub-Saharan Africa (FAO, 2003). Although, depending on efficiency, there need not be a proportional rise, obviously any increase in irrigated area will result in greater water consumption. However, water is increasingly viewed as a scarce (i.e., limited and finite) resource and for many countries in sub-Saharan Africa water scarcity, either physical or economic, is an increasing constraint to economic growth. There is growing concern about the environmental consequences and the implications for other sectors of using more water for agriculture and a growing realisation that future land and water management must change significantly to support human population growth in a sustainable manner. Against this background, Integrated Water Resources Management (IWRM) is promoted by its advocates as the modern and preferred way of managing land and water resources (GWP, 2000). It is proffered as an approach that will simultaneously enable development, protect the environment and ensure sustainability. This paper investigates the implications of IWRM for agriculture in southern Africa. 2 AGRICULTURE IN SOUTHERN AFRICA Agricultural potential in Africa is huge. It is estimated that only 24% (i.e., 28.2 million km2) of arable land is currently under cultivation. In southern Africa, approximately 80% of the poor live in rural areas and are dependent on agriculture for their livelihoods. Furthermore, agriculture is the most important sector of the economy for many countries, contributing an average of 16% of total gross domestic product (Table 1) and accounting for 67% of employment in the region. It is an important supplier of raw materials, food and labour and is also important as the home market for much local industrial output (Tiffen, 2003). The population of sub-Saharan Africa is expected to increase by nearly 3% annually to over one billion in 2025. The Forum for Agricultural Research in Africa (FARA) estimates that, to keep up with this increase and achieve food security by 2015, agricultural production must increase at an annual rate of 6% (FARA, 2003). However, since the mid-1960s, per capita food production, in sub-Saharan Africa, has fallen by about 20% (Pretty, 1999). Substantial new investments in agriculture are needed to meet targets for poverty alleviation and food security. The United Nations Food and Agricultural Organisation (FAO) estimates that about 75% of the growth in crop production in sub-Saharan Africa required by 2030 will have to come from intensification in the form of yield increases (62%) and higher cropping intensities (13%), with the remaining 25% coming from arable land expansion (FAO, 2002). Currently, by far the greatest part of cultivation in southern Africa is rainfed (Table 2). Obviously, rainfed agriculture is highly dependent on the quantity and temporal distribution of rainfall. As a general rule, an absolute minimum of 300–400mm of precipitation is required per year to make rainfed arable farming possible. In many southern African countries relatively low annual totals and high rainfall variability, not only make decisions about crop choice and planting dates extremely difficult, but also significantly reduce productivity. Irrigation, if properly designed and managed, helps overcome many of the disadvantages inherent in rainfed agriculture. It overcomes the

Water resources of arid areas

636

need for shifting cultivation and reduces the pressure on fragile environments. The risk of crop failure is minimized and farmers can hope for higher and more reliable agricultural production and better levels of income (Sally, 2003).

Table 1. GDP and percentage accounted for by agriculture in southern African countries in 1998 & 2001.

Country

Gross Domestic Product (millions US$) 1998 2001

Percentage of GDP from agriculture 1998 2001

Angola 6,445 9,471 13 8 Botswana 4,932 5,196 3 2 Lesotho 890 797 18 16 Malawi 1,736 1,749 34 34 Mauritius 4,146 4,526 9 6 Mozambique 3,873 3,569 32 23 Namibia 3,399 3,100 11 11 South Africa 133,663 114,174 4 3 Swaziland 1,359 1,255 17 17 Tanzania 8,383 9,341 45 45 Zambia 3,238 3,639 21 22 Zimbabwe 5,732 9,057 22 18 (Source: World Bank Development Indicators, http://devdata.worldbank.org/data-query).

3 AGRICULTURE WITHIN THE CONTEXT OF IWRM The water resources of a country are usually assessed in terms of the proportion of rainfall that enters streams and recharges groundwater and so can potentially be abstracted for human use. In most countries in southern Africa, because a high proportion of rainfall is evaporated, the annual renewable water resource represents only a relatively small fraction of the total rainfall. Nevertheless, throughout the region, current human water demand is only a small proportion of the total resource (Table 3). However, the resource situation is a lot more critical than these figures

Table 2. Area of cultivated and irrigated land in countries of southern Africa.

Country

Total area Total (km2) cultivable

Total Total irrigated cultivated area area % (km2) % (km2) cultivated

Integrated water resources management and agriculture in southern Africa

637

area (km2) Angola 1,246,700 – 29,000 – 750 2.6 Botswana 581,730 62,000 32,420 52.2 13.8 0.1 Lesotho 30,350 – 2,093 – 27.2 1.3 Malawi 118,480 36,000 21,055 58.5 280 1.3 Mauritius 2,040 – 854 175 20.5 Mozambique 801,590 360,000 36,000 10 1,067 3.0 Namibia 824,900 250,000 2,052 1 61.4 3.0 South Africa 1,221,040 183,200 123,560 67.4 12,700 10.3 Swaziland 17,364 – 1,915 67.4 3.5 Tanzania 945,090 400,000 63,000 15.8 1,500 2.4 Zambia 752,610 163,500 10,298 6.3 464 4.5 Zimbabwe 390,760 – 27,500 – 1,166 4.2 (Source: FAO Aquastat database, http://www.fao.org/waicent/faoinfo/agricult/agl/aglw/aquastat/main/index.stm).

Table 3. Water resources and withdrawals in the countries of southern Africa.

Country

Water withdrawals Mean annual Annual Agriculture Domestic Industrial rainfall renewable Total Total Total 3 3 2 (Mm3) 2 (Mm3) water ) (Mm % % %2 mm Mm resource1 (Mm3)

Angola 1,052 1,311,200 184,000 365.0 76 67.0 14 48.0 10 Botswana 401 233,200 147,000 54.0 48 36.0 32 23.0 20 Lesotho 760 23,100 5,200 28.0 56 11.0 22 11.0 22 Malawi 1,014 1,201,000 187,000 809.0 86 95.0 10 32.0 4 Mauritius 2,180 4,400 2,200 276.8 77 58.4 16 24.8 7 Mozambique 969 776,700 216,000 540.0 89 53.0 9 12.0 2 Namibia 284 233,900 45,500 170.0 68 71.0 29 8.0 3 South Africa 451 550,500 50,000 9,580.0 72 2,281.0 17 1,448.0 11 Swaziland 778 13,500 4,500 629.8 96 10.5 2 15.7 2 Tanzania 937 885,500 89,000 1,040.0 89 101.0 9 24.0 2 Zambia 1,011 760,700 116,000 1,318.0 82 270.0 17 18.0 1 Zimbabwe 652 254,900 20,000 963.8 79 170.8 14 85.4 7 (Source: FAO Aquastat database, http://www.fao.org/waicent/faoinfo/agricult/agl/aglw/aquastat/main/index.stm). 1 Including water flowing into the country in rivers originating outside the country. 2 Percentage of total withdrawals.

indicate because these averages mask large spatial and temporal variance in freshwater resources and patterns of requirement. With the exception of South Africa, a lack of investment in infrastructure for water storage means that a large proportion of the total

Water resources of arid areas

638

runoff is inaccessible to people. Consequently, episodic water scarcity is a phenomenon that occurs throughout the region and devastating droughts are recurrent events (Houghton-Carr et al., 2002). There is competition for water between agriculture and other sectors such as domestic supply, industry and mining. These other sectors consume water that could be used for agriculture and conversely agriculture consumes water so that there is less available for other sectors. Non-consumptive uses of water by other sectors (e.g. for washing and cleaning industrial products) can also adversely affect agriculture because the water chemistry is changed. Industrial effluents and untreated domestic waste flowing into rivers, lakes and aquifers can pollute the water to such an extent that it becomes unsuitable for agricultural purposes. Similarly although on average 30% to 60% of the water abstracted for agriculture is returned to rivers (Wood et al., 2000), in some instances it is polluted with salts, fertilizers and pesticides which limits it re-use by other sectors. Leaching of excess nutrients from farms and the presence of agro-chemicals in drinking water supplies are currently problems in only relatively small areas of southern Africa but, without careful management, may increase in the future. The industrial sector may also affect agriculture indirectly through the production of air pollution. It is increasingly evident that human emissions of greenhouse gases (a significant proportion derived from thermal energy plants in developed countries) are changing global and regional climates. The regional impacts, magnitude and rate of change remain unclear. However, by modifying rainfall, temperature and evaporation anticipated global warming will affect not only water availability and the water demand of rainfed crops but also water resources. It is now widely accepted that adverse impacts are likely to be exacerbated in arid and semi arid regions, including several countries in southern Africa, where relatively small changes in climate could have significant effects on groundwater recharge and river flows so greatly affecting the feasibility of both rainfed and irrigated agriculture (Hulme, 1996). Although there will be large variations between countries it is estimated that up to 40% of sub-Saharan countries (including Mozambique and Zimbabwe) could lose a substantial part of their agricultural production without careful planning to adapt to changing conditions (Appleton, 2003). IWRM is a holistic approach to water management that links land and water development within a catchment and links social and economic development with protection of natural ecosystems. It is a concept that attempts to coordinate and balance competing demands for water (i.e., domestic, municipal, agricultural, industrial and environmental) in a way that optimizes benefits and enhances equity. Because in southern Africa agriculture plays such a significant role in human utilization of water, application of the principles of IWRM is likely to have a significant impact on the sector. Water management within the agricultural sector needs to address a wide range of issues that are exacerbated, and so of particular importance, in arid and semi-arid areas. These include: ● Management of supplies (i.e., improving water availability in space and time) ● Management of demands (i.e., increasing efficiency of water use) ● Balancing competing demands (i.e., upstream versus downstream and smallholder versus large water users) ● Sustainability of agro-ecosystems and other water-dependent ecosystems.

Integrated water resources management and agriculture in southern Africa

639

IWRM calls for integrated planning so that water, land and other resources are utilised in a sustainable manner. For the agricultural sector IWRM seeks to increase water productivity within the constraints imposed by the economic, social and ecological context of a particular region or country. A major shift in focus under IWRM is the concept of demand management (i.e., managing water demand rather than simply looking for ways to increase supply). This will have to be brought about in a range of ways including: ● Improvement in crop varieties (i.e., crops that yield more mass per unit of water consumed, perhaps as a result of genetic modification) ● Crop substitution (i.e., switching from more to less water consuming crops) ● Improved land husbandry (i.e., soil management and pest and weed control) ● Improved within-field water management (i.e., better timing and precision in the application of water).

4 ANTICIPATED BENEFITS OF IWRM FOR THE AGRICULTURAL SECTOR Properly implemented IWRM could benefit the agricultural sector in southern Africa through more effective and equitable water use, the development of more sustainable practices and reduced disputes over water supplies. Specifically, if implemented appropriately it would enable: ● The implications of water use by other sectors on all forms of agriculture to be considered in the management process and vice-versa ● A more rational decision-making process in which costs and benefits (not just monetary) of different options for water-use are considered in a transparent manner ● More adaptable and effective utilisation of limited water resources that facilitates the changes required to mitigate the potential negative impacts of climate change ● More effective use of water within the sector and hence increased economic returns. All these aspects are likely to be increasingly important in the future as the “price” for water rises and environmental controls (and financial penalties for non-compliance) become increasingly strict. Overall the agriculture sector will benefit through more rational use of water and clear understanding and assessment of trade-offs associated with different ways of using water both within and outside the sector. In addition farmers will, like everyone else, benefit from more equitable utilization of water and an improved environment. 5 CONSTRAINTS TO IWRM IN THE AGRICULTURAL SECTOR Although there is a growing consensus on the need to manage land and water resources in an integrated way, there is no universally accepted method for applying the principles and there are a large number of impediments to practical application. Barriers to successful implementation of IWRM within the agricultural sector arise for a variety of reasons.

Water resources of arid areas

640

These are not limited to technical issues, but also occur because of complex socioeconomic pressures and limitations in human, financial and institutional capacity. Principal constraints include: ● Perceived “low value” and “wastefulness” of agricultural use of water. This perception obviously ignores the importance of agriculture for the livelihoods of most people in rural Africa and the need to make water available to enable sustainable intensified agriculture. ● Incompleteness in water management policy and legal and regulatory frameworks. This is particularly the case in some southern African countries where water policies are non-existent or at best rudimentary. In other cases they may be too fragmented to effectively address the aspirations and priorities of different categories of users at local, national and regional levels. ● Demographic pressures. Population growth linked to poverty is a significant problem throughout southern Africa, which in turn exerts a lot of pressure on arable land and drives inappropriate and non-sustainable agricultural practices and associated water utilization. Communities, even when they understand the long-term consequences of their actions, often feel that they have no alternative. ● Lack of understanding of IWRM principles and practices. In many instances, only a few people in the hierarchy of water management know and understand IWRM and often there is insufficient technical support to operationalise IWRM within the agricultural sector. ● Lack of reliable data, information and knowledge. Quite often, the data required for detailed analysis of water use trends (e.g. temporal and spatial variations in quantities of water diverted and return flows), cause and effect linkages between land-use patterns and hydrological regime, and the impact of changes on downstream users and ecosystems is not available. ● Inadequate understanding of the inter-relationships between biophysical and socioeconomic aspects of a system. Successful IWRM requires the integration of environmental, social and economic factors, but in any specific situation the relationships between biophysical and socioeconomic systems are not well understood. Consequently, the social implications of management decisions are often impossible to predict. ● Lack of incentives for change. In many places water is provided to the agricultural sector at subsidized rates, partly because of the perceived need by many governments for nations to be self-sufficient in food production. However, the result is that there is little economic incentive for farmers to change long established agricultural practices that fail to improve the productivity of water. ● Entrenched agricultural practices. Very often farmers, like other groups, are unwilling to change practices, if they believe that others will simply continue doing what they have always done. Clearly there is a wide diversity of constraints to implementing sustainable water management practices. Successful IWRM requires consideration of these diverse issues across wide range of scales.

Integrated water resources management and agriculture in southern Africa

641

6 REQUIREMENTS FOR SUCCESSFUL IWRM IN THE AGRICULTURAL SECTOR IN SOUTHERN AFRICA In order for IWRM to succeed in the agricultural sector, it is imperative that a number of requirements, which straddle the socio-economic, technical, and institutional domains, are in place. Some of the key requirements include: ● Participatory management: Water must be managed at the lowest possible level with the effective and meaningful involvement of all users of agricultural water. ● Balancing supply and demand management options: While water demand management can help increase the economic efficiency of water use, improve water quality, and promote sustainable water management practices, the parallel pathway of improving water availability through the provision of improved storage and conveyance facilities should not be neglected. ● Capacity enhancement in IWRM: To date, most IWRM training has been directed at civil service employees, academics, researchers and water management specialists, people who are not directly involved in the actual use of water in agriculture. In future IWRM training must be made accessible to agricultural water users. It should also be included in the appropriate educational curricula of universities and agricultural extension services. ● Incentives for water saving: To encourage more efficient utilization of water resources in the agricultural sector, incentives must be offered to those that save water and/or maximise water productivity. ● Improved evaluation of externalities’. Future water allocation requires much greater consideration of both immediate and long-term environmental and health impacts. At present these issues are not consistently assessed in water planning. ● Application of appropriate decision-support tools: IWRM requires consideration of a large number of complex and inter-related issues. Contemporary decision support tools can help structure decision processes, promote understanding of system dynamics, support analysis of possible choices and facilitate the communication of information between people of different technical understanding.

7 CONCLUDING REMARKS Improving land and water productivity, and increasing poor people’s access to water for domestic and productive purposes are critical elements for the development of southern Africa. Ensuring that such developments occur in a balanced and harmonious way requires an integrated approach to land and water resources management. IWRM offers an opportunity for holistic water management. It allows decision-making from a multidisciplinary perspective involving all uses and users of water taking into account the interactions between them and the impacts of water use by a particular sub-sector, or at a particular location, on other sub-sectors or locations. However, it is apparent that there is no automatic or wholesale adoption and practice of IWRM in agriculture in southern Africa. Several factors that militate against it must be overcome and a number of requirements have to be in place to allow effective implementation of IWRM in agriculture.

Water resources of arid areas

642

REFERENCES Appleton, B. 2003. Climate changes the water rules. The Netherlands: Dialogue on Water and Climate. Forum for Agricultural Research in Africa (FARA) 2003. Building the future for Africa’s children: building sustainable livelihoods through integrated agricultural research for development. SubSaharan Africa, Challenge Program Proposal. Food and Agriculture Organisation (FAO) 2003. Unlocking the water potential of agriculture Rome: FAO. Food and Agriculture Organisation (FAO) 2002. World Agriculture: towards 2015/2030 Rome: FAO. Global Water Partnership (GWP) 2000. Towards water security: a framework for action London: GWP. Houghton-Carr, H., Fry, M., McCartney, M. & Folwell, S. 2002. Drought and drought management in southern Africa. Proceedings of the BHS Eighth National Hydrology Symposium University of Birmingham 8–11 September 2002. London: BHS. Hulme, M. 1996. Climate change and Southern Africa: an exploration of some potential impacts and implications in the SADC region. Climate Research Unit, Norwich: University of East Anglia. Pretty, J. 1999. Can sustainable agriculture feed Africa? New evidence on progress, process and impacts. Environment, Development and Sustainability 1 253–274. Sally, H. 2003. Advances in integrated water resources management research in agriculture. In McCornick, P.G., Kamara, A.B. & Tadasse, G. (eds.) Integrated Water & Land Management Research and Capacity Building Priorities for Ethiopia. Proceedings of a MWR/EARO/IWMI/ILRI international workshop, Addis Ababa, 2–4 December 2002. ILRI: Addis Ababa. Tiffen, M. 2003. Transition in Sub-Saharan Africa: Agriculture, Urbanization and Income Growth. World Development 31(8) 1343–1366. Weligamage, P., Barker, R., Hussain, I., Amarasinghe, U. & Samad, M. 2002. World Irrigation and water statistics 2002 Colombo: International Water Management Institute. Wood, S., Sebastian, K. & Scherr, S.J. 2000. Pilot analysis of global ecosystems: agroecosystems. Washington D.C.: World Resources Institute.

Challenges for managing water resources in semi-arid areas: a case study from two rural communities in Zimbabwe F.T.Mugabe Department of Land and Water Resources Management, Midlands State University, Gweru, Zimbabwe A.Senzanje Department of Soil Science and Agricultural, University of Zimbabwe, Harare, Zimbabwe Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Rainfall is received during the wet season (November to April) in Zimbabwe and the remainder of the months are dry. The streams and shallow wells that are the major sources of water during the wet season dry up during the dry months and years, forcing communities farming and living in the semi-arid areas to depend on water captured by dams and deep wells. However, the quantity of the water resources in surface reservoirs and wells is not known in most cases and this presents a challenge in that: ‘how can one manage a resource whose quantity is not known?’. This problem is further compounded by the fact that the quality of the next season is not predictable given that three out of five years receive below average rainfall that might not produce enough runoff to fill up the reservoirs or recharge groundwater depending on its distribution. Despite the uncertainties in the quantity of the resources at any given time, the management of such rural water resources is further complicated by ownership of the water bearing bodies that is either individually or communally owned depending on whether the resource was privately or government or NGO developed or is in a private or common property. This paper examines how two rural communities (in semi-arid southern Zimbabwe) manage their water resources and the challenges they are facing for improving water resources management. There is no explicit strategy for water resources management in both sites. Management is by crisis when water resources shortages are looming as indicated by failing of wells or when water in surface reservoirs has gone very low or when the wet season has advanced with no signs of rainfall.

Water resources of arid areas

644

1 BACKGROUND About 160 million people farm and live in arid and semi-arid areas of Africa. These areas receive below 600mm of annual rainfall and together with increasing population and lack of infrastructure this means that many people have inadequate access to water. Water scarcity and irregularity in rainfall are increasing due to the effects of EL Nino and possibly to the impacts of global warming. Water is therefore the most valuable resources in the arid and semi-arid areas hence the need for a ‘water resource audit’ (KAWAD, 2001) if sustainable water resources management is to be achieved. The variability and unreliability of rainfall makes the sustainable development of water resources difficult (Griffiths, 1972) hence the need for rural communities to come up with clear management strategies given that their relationship with their resources is key to their survival and prospects (Soussan, 1998). Good management of and secure rights to water resources are crucial to livelihoods and particularly to people’s capacity to cope with variability (Soussan, 1998). Water also provides a means for the diversification of livelihoods. It is also important for addressing poverty and rural development since it is used for food production (Chaturvedi, 2000).

Figure 1. Location of Romwe and Mutangi communal areas. Proper water resources management can reduce/diminish water scarcity. Water scarcity is the underlying factor behind water resources problems within a society or between nations. Water resources scarcity can lead to changes in access rights, changes in property rights, changes in property relations, greater conflicts, livelihood changes, loss and disposal of other assets, over-exploitation of resources (Soussan, 1998). Water is already in short supply in Sub-Saharan Africa (Cleaver and Schreiber, 1994) and since

Challenges for managing water resources in semi-arid areas

645

climate change is likely to increase the water stress, there is now a need to investigate more thoroughly the links between water potential resource base and how it is managed during the season and dry years and see if there are any opportunities for reducing water scarcity. This paper explores how two communities have managed their water resources amid insufficient knowledge of the quantities of their water resources base, source of water, ownership and shifting needs during the season. The paper further outlines the challenges these communities are facing to better manage their water resources. 2 THE STUDY AREAS Romwe and Mutangi are both micro-catchments in Chivi (Figure 1) district that is in the semi-arid parts of Zimbabwe. Mutangi lies about 55km southwest of Masvingo while Romwe lies some 80km due south. The two catchments have contrasting resource endowments (Sullivan et al., 2000) and climatic pattern, with Romwe receiving more rainfall than Mutangi in most years. The annual average rainfall for Romwe and Mutangi are 581 and 500mm respectively and the deviation from the long-term mean for Chivi growth point is depicted in Figure 2. They are both characterized by water shortages during dry seasons and dry years (Moriarty and Lovell, 1999). Water resources in Romwe are principally groundwater while Mutangi is surface water (Sullivan et al.,

Figure 2. Deviation from the long-term mean rainfall at Chivi in semi-arid southern Zimbabwe.

Water resources of arid areas

646

submitted) but there are 18 unlined shallow wells that tap the ephemeral soil/weathered zone aquifer and dry soon up after the rainy season (Mugabe et al., 2003). The annual per capita level of domestic water use per day is reported to be 30.7 and 14 litres for Romwe and Mutangi respectively (Sullivan et al., submitted). Water is also used for washing clothes, bathing and building. 3 ROMWE WATER MANAGEMENT There are 26 privately owned wells in Romwe and most of them fail in those extremely dry years like the 1992/3 season (Moriarty and Lovell, 1999) forcing the farmers to rely on the collector well that was constructed in 1989 by CEH (former IH) and BGS (Waughray et al., 1889). The privately owned wells are dug by the owners and are constructed at the homesteads or in the owners’ fields and are used for garden irrigation and domestic purposes. Water from these wells is accessible to anyone for domestic purposes only. Access to domestic water use from the collector well is not limited to anyone, but productive water use is limited to the families that have vegetable beds in the 1ha garden. The group that uses water from the collector well productively has a committee that is chaired by a chairman and this committee decides on how to use the water. Each member contributes to a fund every month and the fund is used to buy spares whenever the pump breaks. With the assistance from the Institute of Hydrology (IH) some of the garden members were trained in pump maintenance and repair and they therefore do not need outside assistance in the case of a breakdown. Unlike the privately owned wells, the collector well is fitted with a munro chart recorder and water levels is monitored at any one time. Most of the privately owned wells failed in 1992/3 and the collector well was the only source of water (Lovell et al., 1998). The reaction of the garden members was to reduce the area under cultivation by half so that even non-members could have access to domestic water—this demonstrated a shift in the nature of use, patterns of access to and rights over the resources, and the intensity of use decreased. 4 MUTANGI WATER MANAGEMENT There are 18 privately owned wells in Mutangi and three boreholes of which only one is functional. The functional one and the non-functional boreholes were constructed by DDF while World Vision constructed the third one. Productive water use from the shallow wells is done during the harvesting period up to when the wells dry up around June and then most of the gardening takes place in the community garden that is irrigated with water from the dam. Domestic water is from the shallow wells, river and borehole during the wet period and is obtained from the borehole during the dry season (Mugabe et al., 2003). There are a number of problems associated with water use from this dam and these include: ● The dam has a lot of silt and siltation continues during the rainy season

Challenges for managing water resources in semi-arid areas

647

● Poor farming practices in the catchment area resulting in increased siltation ● Animals drink in the dam and leave their droppings resulting in poor quality water ● Fencing material is being stolen. Catchment and water resources management was introduced by CARE who worked on a dam rehabilitation project that covers a number of small dams in Masvingo. In Mutangi, a systems approach to managing the reservoir has been adopted by forming committees to look into minimizing silt from the catchment (nine siltation traps has been built so far in channel 1), rehabilitation of the water source, increasing productivity and conserving and using water resources efficiently. Three major committees were formed with the assistance of CARE and these are Dam rehabilitation, Irrigation and Agronomy. The functions of the Dam rehabilitation committee are rehabilitation of the dam, drafting social contracts, mobilization of materials and human resources for effective implementation of the dam project and linking with RDC. The duties of the Irrigation committee are fencing of the garden, plot allocation, ensuring payment of the user’s fees, marketing and drafting of garden constitution. The Agronomy committee co-ordinates micro-catchment conservation works, extension services, pegging contours, infield demonstration, farmer-to-farmer training, coordination of farmers’ experimentation and field days. However, effective rehabilitation is hindered by the fact that most of the farmers who farm in the upper catchment (where most of the silt comes from) do not use the catchment dam but use another dam that is outside the catchment hence are not interested in siltation control programmes in their fields. 5 THE CHALLENGES FACED IN MANAGING RURAL WATER RESOURCES Unlike the bigger water users, there are no active water allocation mechanisms in the rural areas. The new water act (1998) has left the management of rural water resources to the users. ZINWA should come up with a clear-policy on how water resources should be managed in the rural areas—mostly interested in the distributing water to the bigger users who pay in given communal area-municipalities, farmers, mines etc. Other studies (Butterworth et al., 2001) have shown that regulatory approaches to water resources allocation will require high levels of capacity in catchment management bodies (which looks impossible for the Zimbabwean rural communities together with the current legislative measures in the Zimbabwean water act, 1998) The challenges for managing water resources can be grouped into four, viz: insufficient knowledge of the quantity of water, ownership, pessimism about the quality of the following season, lack of projection of their annual water requirements. The comment by Mukherjee (KAWAD, 2001) ‘Underground water reserves are like bank accounts which can be thoughtlessly depleted by the farmer by resorting to heavy irrigation which can be likened to issuing a series of cheques without depositing anything’ is correct for these two rural communities. Rural communities have insufficient knowledge of both their groundwater and surface water resource base (Mugabe and Hodnett, 2001) and they keep on using water till ‘they are told by the bank manager that they no longer have any more money in their accounts’ i.e. when the wells fail to give

Water resources of arid areas

648

them water or when the levels of their surface water resources are so low. Rural communities do not know what is ‘added to their bank accounts’ i.e. groundwater recharge and run-in the dams and they do not know the amount of water available to them for that given season. Apart from insufficient knowledge of their water requirements they do not project their annual water requirements. Ownership of water resources in communal areas depends on where the resource is found in most cases. Wells dug by individuals in their fields belongs to them while boreholes or wells dug by organizations such as DDF, CARE, the government etc. and river water belongs to the community. Use for water in communal areas is guided by ownership; hence the resource is communally owned in some cases while it is not in other cases. Access to water from individually owned wells is mainly restricted to domestic purposes and not productive water use (Moriarty and Lovell, 1998). In most cases communities view water sources constructed by outside organizations as belonging to these organizations and as such they are hesitant/unwillingly to repair them in cases of breakdown. There are weaknesses in managing communal area water resources because of the prevailing notions that any shortages are temporary and limited to dry seasons. People are pessimistic that the following rainy season will bring enough rains to ‘recharge’ the dams, wells, rivers for use during the dry season, hence water resources are rarely managed except during periods of water scarcity (e.g. when the borehole fails or when people see dam water diminishing fast) when ‘crisis’ management is employed. This was shown at Romwe during the 1992 dry year when garden owners reduced the garden area by half in order to provide domestic water for everybody even the non-garden members. This is further complicated by lack of knowledge of the quantity of the resources in the water bearing body. 6 CONCLUSIONS The points highlighted in this paper that are worth emphasising are: ● Water resources management is only limited to the dry season and dry years; there is plenty of the resource during the wet season. ● Outsiders in both cases, IH/DR&SS in Romwe and CARE in Mutangi have facilitated water resources management. ● Water resources management is tricky in the communal areas given that the quantity of the resources is not known (especially groundwater resources). Need to come up with an inventory of water resources in this rural areas—audit. ● Water point management has been different in the two case studies owing to the differences in the nature of the water resource. Romwe is groundwater based (effect of land use practices is not apparent), hence management is mostly restricted to pump maintenance and water allocation in times of drought. In Mutangi, a catchment approach has been taken since it is surface water (the effect of land use practices is apparent). ● Ownership and access to water resources distorts whatever communal management system is in place.

Challenges for managing water resources in semi-arid areas

649

ACKNOWLEDGEMENT The research was partially supported by an African Doctoral Fellowship provided by START and the Pan-African Committee for START. The authors would particularly like to thank the people of Romwe and Mutangi for their assistance and participation in the research. REFERENCES Butterworth, J., Mogkope, K. & Polland, S. 2001. Water resources and water supply for rural communities in the Sand river catchment, South Africa. 2nd WARFSA/WaterNet Symposium: Integrated Water Resources Management: Theory, Practice, Cases; Cape Town, 30–31 2001. Chaturvedi, M.C. 2000. Water for food and rural development. Water International, 25(1), 40–53. Cleaver, K. & Schreiber, G. 1994. Reversing the spiral: The population, Agriculture and Environment Nexus in Sub Saharan Africa. World Bank, Washington DC. Griffiths, J.F. 1972. Precipitation. In J.F.Griffiths (ed.) World Survey of Climatology Vol 10: Climates of Africa: 24–28. Amsterdam: Elsevier Publishing. KAWAD. 2001. A fine balance: Managing Karnataka’s scarce water resources. Karnataka Watershed Development Society, No 250 1st Main Indiranagar, Bangalore 560038. Lovell, C.J., Butterworth, J.A, Moriarty, P.B., Bromley, J., Batchelor, C.H., Mharapara, I., Mugabe, F.T., Mtetwa, G., Dube, T. & Simmonds, L. 1998. The effects of changing rainfall and land use on recharge to crystalline basement aquifers, and the implications for rural water supply and small-scale irrigation. DFID report 98/3. Moriarty, P.B. & Lovell, C.J. 1998. Water resources development in Chivi: Results of a village mapping exercise. DFID report 98/10. Mugabe, F.T., Hodnett, M.G. & Senzanje, A. 2003. Opportunities for increasing productive water use from dam water: a case study from semi-arid Zimbabwe. Agricultural Water Management, 62:149–63. Mugabe, F.T. & Hodnett, M. 2001. Micro-catchment management and common property resources project: A report on Mutangi hydrological and water resources study in semi-arid Zimbabwe (1999–2001). Institute of Environmental Studies, UZ, (unpubl.) 26 pp. Soussan, J. 1998. Water/Irrigation and sustainable rural livelihoods. in Carney (ed) Sustainable rural livelihoods: what contribution can we make? DFID-ISBN 1 86192 082 2. Sullivan, C., Mutamba, M. & Kozanayi, W. Water use and livelihood security: A study of rural households in Southern Zimbabwe (submitted). Waughray, D.K., Lovell, C.J. & Mazhangara, E. 1998. ‘Developing basement aquifers to generate economic benefits: A case study from Southeast Zimbabwe’. World Development, 26 (10):1903–1912.

An Integrated Water Resources Management tool for Southern Africa allowing low flow estimation at ungauged sites M.J.Fry, S.S.Folwell & H.A.Houghton-Carr CEH Wallingford, Wallingford, UK Z.B.Uka Ministry of Water Development, Lilongwe, Malawi Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: In order to make necessary surface water allocation decisions, water resources managers require detailed information about both water availability and water use. Tools are available to integrate and display existing data, often within a GIS interface, but are less useful where data are unavailable. The ‘Low Flows—Southern Africa’ software described in this paper is a GIS-based water resources management tool that allows low flow statistics to be estimated at ungauged sites and displayed alongside gauging station, water use and catchment characteristic data. The software allows users to estimate the natural flow duration curves and low flow statistics at any point on the river network using relationships between catchment characteristics and flow regimes. The development of the software has been funded by the UK Department for International Development as part of the UNESCO Southern Africa FRIEND project. The software has recently been installed for use in Malawi.

1 INTRODUCTION 1.1 Integrated Water Resources Management It is now widely recognized that Integrated Water Resources Management (IWRM) is essential to sustainable development in Southern Africa. It is particularly relevant to efforts to promote equity and improve opportunities for the poor. This is illustrated by a recent review by the Global Water Partnership (GWP, 2003), which states ‘Since IWRM contains prospects for the equitable allocation of benefits from water and services dependent on it, it is important that these opportunities for healthier and more productive

An Integrated Water Resources Management tool

651

lives among the most at risk and disadvantaged population groups are not lost, but transformed into reality.’ As water use and water stress increase, the margin for error in assessing both water resources and the impacts of water use becomes smaller. At every level, from small catchments to international basins, water management needs to be carried out with the consideration of all stakeholders and with all possible awareness of what the future may bring. In practice, this means that water allocation decisions need to be made by water resource managers with as much information as possible on the available natural resource, the effects of existing water use on this resource, the water needs of stakeholders and the predicted change in these factors in the future. 1.2 Tools for water resource estimation and allocation The tools providing this information should form an interface between water availability and water use, allowing their impacts and interactions to be visualised. For instance, when assessing a planned irrigation abstraction from a river, the decision-maker should be able to ‘see’ the catchment above the point on the river from which the abstraction is to be made, and to understand the influences on the resource available at that point under natural conditions, as well as the variability of the resource. The decision-maker needs to be able to quantify the water use within the catchment, in order to estimate its impact and the effects of possible change in this water use. He or she also needs to be able to ‘see’ the water use downstream from this point in order to determine the effects of the new allocation on the resource for these users. 1.3 Southern Africa FRIEND FRIEND (Flow Regimes from International Experimental and Network Data) is ‘an international framework for the implementation of hydrological research’ (UNESCO, 1997), aiming to improve the scientific and technological basis for the development of regional methods for the management and development of water resources. Within the Southern Africa region of the FRIEND program work is being implemented by the Centre for Ecology and Hydrology (CEH) Wallingford to strengthen the technical capacity of national and regional water institutions, with much of the activity being funded by the UK Department for International Development (DFID). One central concern is the assessment, planning and management of water resources, with particular attention to low flow conditions. CEH Wallingford has developed methods and software for a prototype water resources management tool for use in Southern Africa, intended to strengthen the abilities of countries to assess, plan and manage surface water resources using contemporary GIS techniques. Malawi was selected as a suitable area for this development, with water resource issues typical of the region. This prototype provides both valid insight into the development of suitable hydrological methods and shows the technological possibilities for water resources management tools in these regions.

Water resources of arid areas

652

2 INTEGRATED WATER RESOURCES MANAGEMENT IN MALAWI Water Resources Management in Malawi involves numerous stakeholders including public and private sector bodies and commercial and domestic water users. Scales of water use vary from the localised large commercial agriculture and reservoir fed civic supply schemes to small scale domestic agriculture and gravitationally fed rural village supplies. Surface water abstraction is a major issue within Malawi as it is within many Southern African countries. Potential for expansion of surface water use for both public supply and agriculture exists and is likely to be exploited over coming years. However, increased abstraction must be managed to minimise adverse effects upon ecosystems and existing agriculture and water supply schemes. Generally, larger private bodies and public supplies, within catchments of all sizes, are licensed, whereas many of the numerous smaller domestic and agricultural abstractions are unrecorded. Currently water use licensing is performed by the Water Resources Board, a public sector body attached to the Ministry of Water Development. The decision-making process consists, in general terms, of consideration of the need for the requested water use, assessments of the available resource and the impact of increased abstraction, and an element of public participation. The accuracy of the assessment is limited by the information available. Currently, stream flow is measured consistently within less than 120 catchments within Malawi, a country with an area of 118,500km. Within this area there are approximately 270 raingauges operated by the Department of Meteorology. Some current methods exist for measurement of water availability but are generally only applicable to selected localities. Previous work by Drayton et al. (1980) established a method of flow estimation for the country, but this is only applicable through the manual derivation of catchment boundaries and extraction of statistics from pre-processed maps of runoff. There were previously no available systems to manage abstraction information or to integrate this with water availability. There is clearly a need for both water use data management tools and updated methods for flow estimation away from gauged catchments. Discussions at regional workshops have revealed that this is a situation that many countries of the region currently find themselves in. 3 DEVELOPMENT OF LOW FLOW ESTIMATION TECHNIQUES FOR MALAWI 3.1 Description of the study area Malawi is surrounded by Tanzania, Zambia and Mozambique. The country is located at the southern end of the East African rift valley, which dominates the topography. Lake Malawi, the third largest lake in Africa, occupies the northern two thirds of this section of the rift. The lake lies at an altitude of 470m, and its only outlet, the River Shire, drains southwards to the lower rift valley at 90m altitude, before joining the River Zambezi. All rivers in Malawi eventually drain into the Shire except for a small area in the east of the

An Integrated Water Resources Management tool

653

country, where there is the catchment of Lake Chilwa, which has no outlet, and a smaller area draining eastwards into Mozambique. The topography of Malawi is varied, and the country may be divided into four broad hydrological zones: highlands, plateau, escarpment and rift valley. The plateau is at an altitude of between 900 and 1200m and features broad, undulating plains. The climate is temperate, and these are the most densely cultivated regions of the country, where little of the original woodland remains. The highlands rise abruptly from the plateau reaching altitudes of between 2100m and 3000m. The climate is cool, and the vegetation is forest relicts and open grasslands. These areas are now either forest reserves or game reserves, partly covered with exotic trees. The escarpment marks the boundary between the plateau and the rift valley. It drops down in a series of shelves and is an area of major faulting. Considerable portions of the escarpment are protected by forest or game reserves, but pressure to secure arable land results in some areas of steep land being cultivated outside these protected areas. The rift valley is mainly covered by alluvial deposits of the Quaternary age. The climate is tropical, and the original vegetation is mixed savannah woodland. The most favourable soils in this zone have been developed into irrigated rice and sugar schemes. 3.2 Low flow estimation The main previous study of low flow hydrology and flow estimation in Malawi was that of Drayton et al. (1980) who developed methods for flow duration curve estimation using relationships between mean flow and rainfall, and mapped values of the low flow statistic Q95 for the country. The method enabled the estimation of low flow statistics and flow duration curves at any point in the country, providing the facility to obtain flow estimates at points on the river network where gauged flow data are unavailable. However, the time-consuming and technically challenging nature of this method, which had to be applied manually, has meant that it has not been regularly used. Recent international low flow studies have often concentrated on relationships between flow statistics and the hydrological response of soil types within catchments (Gustard et al., 1992), assisted by the availability of high quality soil coverage data through the work of the United Nations’ Food and Agriculture Organisation (FAO, 1996) and others. Another element of recent low flow estimation studies has been the creation of flow duration ‘type curves’, whereby a set of standardised flow duration curves are created using data from a number of stations representing the variety of flow regimes within a region, and the linking of these standard curves to a low flow statistic, typically Q70 or Q95 (Gustard et al., 1992; UNESCO, 1997). Through estimation of this statistic from observed data, flow duration curves can be predicted. This study aimed to further the work of previous studies in Malawi, and Southern Africa generally, by investigating the application of these methods in Malawi in order to assess their feasibility for flow estimation as part of a water resources management tool. The general approach can be outlined as: (a) Selection of suitable gauged catchments; (b) Detailed scrutiny to assess the reliability of the records;

Water resources of arid areas

654

(c) In cases where insufficient natural flow records are available, naturalisation of the observed flows (i.e., removal of the effects of artificial influences on the pattern of river flows); (d) Extraction of low flow measures from data records for these catchments;

Figure 1. Schematic of stages in flow estimation procedure. (e) Extraction of catchment characteristics from digital maps; (f) Development of relationships between extracted flow measures and catchment characteristics, or country-wide mapping of flow measures. This research resulted in a method for the estimation of ungauged sites at any point on the river network within the country. The method allows selection of a standardised regional flow duration type curve, developed during an earlier phase of the SA FRIEND project (UNESCO, 1997), on the basis of low flow statistics derived from catchment characteristics, namely soil types. This curve is then re-scaled using an estimate of mean flow derived from the catchment rainfall. Monthly artificial influence data (abstraction and discharge volumes) can then be applied to estimate the impact at key flow percentiles. Figure 1 shows a schematic of the stages in the flow estimation procedure.

An Integrated Water Resources Management tool

655

4 SOFTWARE FOR WATER RESOURCES MANAGEMENT The method for estimating flows at ungauged sites is practically applicable within licensing and water resources management operations. The flow estimation method requires information on the extent of different soil types and on mean annual rainfall for each catchment. These data are digitised and available within GIS systems, but can be time-consuming to derive for a particular location. A bespoke software tool can carry out this and other functions of the estimation process, and also allow additional information to be displayed to maximum effect. Software, incorporating these tools, was desired to exhibit a number of features: – Visualisation of rivers, artificial influences, gauging stations and other spatial features – Storage of artificial influence information – Storage of catchment characteristics—rainfall and soil types – Derivation of catchments at any point on the river network – Retrieval of characteristics (rainfall and soil type extents) for given catchment – Production and visualisation of natural flow duration curve – Visualisation of the impacts of existing artificial influences upon natural flow regimes – Creation of ‘scenario’ flow statistics under predicted future influences

Figure 2. Main GIS user interface window of the software, showing detail of western Malawi.

Water resources of arid areas

656

Other important aspects considered within the software design included: – Ease of use for potential users of all levels – Consistency of flow estimates – Usability of outputs within existing procedures – Maximising the usefulness of data management tools – Consistency of interface with common GIS applications and software within MWD The software was based upon the Low Flows 2000 software developed by CEH for the Environment Agency of England and Wales. This system has proved to be a large step forward for water resources management information systems in the UK. Much of the functionality has been carried across from this original version, but it has been adapted to the needs of Malawi, and many new features have been developed. The software is designed to be flexible in terms of both data stored against features within the database and allowing different hydrological models to be ‘snapped on’. Because of this, the extension of the system to other areas of Southern Africa, which may have different hydrological models for flow estimation, would be fairly simple. The system is a fairly radical departure from the water resources and data management procedures currently used in Malawi. The system and its functionality are based around a GIS user interface. This displays geographical features in a map style allowing the user to zoom in and out to view different areas of the map at different scales. Data are stored and displayed as different ‘layers’ within the map, where a layer constitutes a map of a single type of data. For example, Figure 2 shows the software user interface displaying a map of an area of Malawi consisting of separate layers of rivers, abstraction points, lakes and the boundary of the country. These different types of data are called feature types. Within the software the layers can be turned off and on in order to create an appropriate map of the data of interest. Different layers can be stored in different ways, appropriate to the geographical representation of the type of data. For example, the boundary of Malawi is stored as a series of lines, the lakes of Malawi are stored as a series of polygons representing the lakes, the abstractions are stored as points, and the rainfall and soils data is stored as grids, with each grid square containing a different data value. The software uses the ESRI MapObjects component to draw and manipulate the maps. The data within the software is in ESRI’s standard ‘Shapefile’ format (ESRI, 1998) meaning that the software is compatible with many other widely-used GIS products, including ESRI’s ArcView, ArcInfo and ArcGIS systems. Additional data in this format can be added to the software to provide contextual information, for example, roads, villages, industrial areas and so on. These can assist the user in the siting of water use points and other features. The addition of familiar features can also help less confident users become more active in using the software. 4.1 Software features The hydrologically important features within the software can be divided into three groups: – Artificial influences—abstractions, discharges, impoundments

An Integrated Water Resources Management tool

657

– Gauging stations – Spot gaugings or instantaneous flow measurements All of these features are stored within the database, each having different data stored against them in a varying number of fields, or attributes. 4.1.1 Artificial influences The artificial influences interact with the model, providing estimates of water use. The software is designed to store these as licensed influences, but users can also enter influences that reflect estimated water use within the catchment. The structure of a licence can be quite complex, and consists of a number of sites on the river network. Each of these sites can have one or more purposes, the reason for the water use, such as irrigation, domestic use, etc. The structure of the licences stored within the database, as well as the attributes against which data can be stored for a licence, are designed to complement the nature of abstraction and discharge licences in Malawi. The flexibility of the database structure would assist in quickly applying the software in a different country with different data storage requirements. The principal data field within a licence is the monthly volumes of water abstracted from or discharged into a river, and it is this data that is used to modify natural flow estimates to model the effects of artificial influences on flows. 4.1.2 Gauging stations Gauging stations within the database store flow data measured at a particular point on the river network. These data can provide additional information about the flow regime, and real flow data can be used to underpin flow estimates. 4.1.3 Spot gaugings Spot gaugings are occasional measurements of flow at a point in the river where there may not necessarily be a gauging station. These measurements can also provide information as to the validity of flow estimates. 4.2 Software functionality 4.2.1 Catchment definitions One of the features of the software is that it allows users to produce catchment definitions for any point on the river network. This exercise had previously to be completed by hand using paper maps. Catchment definitions are useful for a number of purposes, and in addition, the software retrieves data from the underlying catchment characteristic datasets to provide statistics for that catchment. The user simply either enters an X-Y coordinate of the catchment outlet or clicks on the river network to select the point from which to define the catchment. The catchment boundary is then calculated and relevant statistics (e.g., catchment area, mean annual rainfall, monthly rainfalls, soil types) are displayed

Water resources of arid areas

658

(Fig. 3). Defined catchments can be saved for later use, which is useful when a catchment is under regular scrutiny.

Figure 3. Software windows displaying catchment definition process (left) and defined catchment and details (right).

Figure 4. Flow estimation windows showing summary of artificial influences and natural and influenced flow duration curves. 4.2.2 Natural flow estimates Catchment characteristics retrieved from the underlying datasets are used within the model to estimate flow statistics using the method outlined above. The user runs the catchment definition as before, selecting the starting position required but this time choosing to run the model. The outputs are the flow duration curves and monthly flow statistics. The natural flow estimates window consists of panels showing the catchment area and contributing streams, summarising the catchment statistics and displaying the natural monthly and annual flow duration curves. This information gives water resources

An Integrated Water Resources Management tool

659

managers an idea of the hydrological characteristics of the natural, or uninfluenced, catchment: estimates of area, monthly rainfall, baseflow index and monthly mean- and low-flow statistics. The information is essential for estimating the impact of water use at the catchment outlet. The flow duration curve allows further statistics to be derived graphically; they are displayed by clicking on the name of each flow duration curve, a new form then appears providing the flows at a various percentile intervals. 4.2.3 Influenced flow estimates Once the natural flow estimates have been made, these can be modified to show the impact of artificial influences within the catchment. The software retrieves the influences within the catchment and updates the panels described above with the influenced flow statistics (Fig. 4). The influences are listed, showing the name of the influence and the monthly volumes abstracted or discharged; the form allows users to find individual influences through the map on the right-hand side. The basin-details are updated to show the flow statistics before and after the impact of influences, and the flow duration curve form then allows users to plot monthly and annual curves for both natural and influenced regimes. 4.2.4 Comparing estimates with gauged data The flow duration curves displayed are estimates made using the model described above. In order to make the system more appropriate to Southern Africa, where flow estimates may not always be reliable due to inaccuracies in the underlying datasets, a feature was added to allow observed flow duration curves to be viewed alongside the estimates. A form is shown listing all gauging stations on the database, with a map enabling users to see where the stations are situated relative to the catchment in question. A nearby gauging station or one with a similar catchment is selected and its data is displayed on the original graph. Results can be viewed as a percentage of the mean flow in order to compare data from catchments of different sizes. 4.2.5 Creating influence scenarios The user can observe the effect of a planned, estimated or hypothetical influence through use of the ‘Define scenario’ form. This allows existing monthly abstraction, discharge or impoundment volumes to be increased, either as a percentage of the existing volume, or by a fixed amount. The resulting total influence is re-applied to the flow statistics and three sets of annual and monthly flow duration curves are then visible—natural, influenced and scenario. 5 SUMMARY AND FUTURE WORK Movement towards Integrated Water Resources Management is an essential step to the sustainable management of water resources in Southern Africa. An important part of this movement will be to provide water resource managers with methods to estimate low flow

Water resources of arid areas

660

statistics at ungauged sites, and tools to allow these methods to be easily and consistently applied and to integrate spatial and temporal data on catchments, water availability, water use and the effects on flow regimes of this use. CEH Wallingford, as part of phase II of the Southern Africa FRIEND project, has created a prototype method for low-flow estimation in Malawi, and GIS software tools for its application, appropriate to the requirements of that country. These tools have been installed in the Ministry of Water Development in Malawi to allow water resource managers to visualise water use at a catchment scale and to effectively manage water by modelling current and predicted catchment water use and observing the effects on streamflow statistics. Such tools were previously unavailable in the country have the potential to provide a huge leap forward in the quantitative management of water within most countries in the Southern Africa region. REFERENCES Drayton, R.S., Kidd, C.H.R., Mandeville, A.N. & Miller, J.B. 1980. A regional analysis of river floods and low flows in Malawi. Report No. 72. Institute of Hydrology, Wallingford, UK. ESRI. 1998. Shapefile technical description. ESRI White paper. Environmental Science Research Institute, Redlands, California, USA. FAO. 1996. Digital soil map of the world and derived soil properties. CD-ROM, Food and Agriculture Organisation of the United Nations, Rome. GWP. 2003. Poverty Reduction and IWRM. TEC Background paper no. 8. Global Water Partnership. Gustard, A., Bullock, A. & Dixon, J.M. 1992. Low flow estimation in the United Kingdom. Report No. 108. Institute of Hydrology, Wallingford, UK. UNESCO. 1997. Southern Africa FRIEND. Technical Documents in Hydrology No. 15. United Nations Educational, Scientific and Cultural Organization, Paris.

Organization of water services in Malawi and strengths and weaknesses in implementing Integrated Water Resources Management (IWRM) Milward Selemani Blantyre Water Board, Malawi Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The overall charge for controlling and regulating water abstraction in Malawi lies with the Water Resources Board, which falls under the Government Ministry of Water Development, which is also responsible for issuing water right certificates. There are presently two autonomous water boards supplying potable water to the cities of Blantyre and Lilongwe and three other regional boards service the smaller towns in the north, center and southern regions. The Ministry of Water Development, however, is responsible for providing water services to rural communities throughout the country. Some non-governmental organizations also play an active role in the provision of services to periurban and rural communities. Also within the cities, sanitation services are presently under the jurisdiction of City Councils. These arrangements present problems in efficient management of water services and this paper intends to discuss these issues and solicit comments and suggestions for improvement from symposium participants.

1 INTRODUCTION Integrated Water Resources Management (IWRM) is important for Malawi even though 20% of the country’s 118,500 square kilometre total area is covered by water. Being mountainous some areas lie on high altitudes far from any reliable water resources. On the other hand there are some areas, which as a result of their low-lying topography, they have more than adequate water and some of these low-lying areas are prone to flooding every rainy season. Lake Malawi is the third largest lake in Africa and the Shire river which drains the lake to the Zambezi at annual average flow of 395 cubic meters per second is regarded as the life line for Malawi because of the major activities taking place throughout its course: There are fisheries and irrigation on its upper course, electric power generation and water supply abstraction in the middle course and irrigation of the

Water resources of arid areas

662

sugar estates in the lower or last course of the river. There are also other smaller lakes and a network of rivers and streams throughout the country. 2 WATER SERVICES STRUCTURE IN MALAWI At national level the Ministry of Water Development has the overall responsibility for management of water resources throughout the country: The National Water Resources Board, a unit within the ministry is responsible for approving and issuing of water rights certificates whilst a Water Supply Unit, also within the Ministry of Water Development looks after the provision of potable water to the rural communities by provision and construction of boreholes, protected wells and piped water schemes though out the country. The two cities of Blantyre and Lilongwe are under Blantyre and Lilongwe Water Boards. The two water boards are autonomous and their responsibilities are limited to abstraction, treatment and distribution of potable water within the city boundary and some designated outside the city boundary areas. This responsibility does not include sanitation and sewerage system. Therefore these responsibilities lie under the jurisdiction of City Councils. There are a number of NGOs providing assistance to small communal water projects in collaboration with the Government Water Supply Unit. A notable example is the Malawi Social Action Fund (MASAF) which funds and supervises various self-help development projects. Some of such projects are carried out with high-density peri-urban areas, which are under the jurisdiction of the respective Water Boards and are by nature of their soil structure and due to poor sanitation, pollution risky. Water supply ranks high among these projects as it is regarded as basis for any meaningful and sustainable development. MASAF is known at times to provide assistance to water projects at local level without approval of the controlling ministry. This can be also cause for confusion. The three regional water boards and the two water boards for Blantyre and Lilongwe all report to the Ministry of Water Development on technical issues and Ministry of Statutory Corporations on administrative issues. Here again reporting to two ministries can be regarded as an organisational problem that can cause confusion. At community level villagers sometimes elect leaders who form committees to run and manage various development projects. Each type of development project may have its own committee. Water supply committees may be selected for each water source or catchment and such committees are responsible for formulation of rules and procedures for managing the resource sustainably. 3 WATER SERVICES STRUCTURE AND IMPLEMENTATION OF IWRM The Ministry of Water Development’s Water Resources Board is responsible for monitoring levels in the lake to ensure sufficient water for lake services and controlling flow in the Shire river to ensure sufficient water for power generation and water supply abstraction for the City of Blantyre along the middle course of the Shire and irrigation of

Organization of water services

663

the country’s major sugar estates in the lower Shire. The control of lake levels and flow in the Shire River is done through opening and closing gates at Liwonde barrage. The barrage also has another function of controlling flooding in lower Shire valley. The ministry is also responsible for gauging flows in other main rivers throughout the country for purposes of regulating and managing flows for various purposes e.g. irrigation. The board also approves and issues water rights and issue licences for drilling boreholes throughout the country. It is however doubtful if the Ministry carries out regular inspections for the purpose of monitoring these water allocations to ensure that they are being followed. The ministry of Water Development and the Department of Environmental Affairs’ legist ration requires both to regulate and manage water resources by ensuring proper maintenance of the catchments and the environment in general. However this requirement is not given much prominence and there is sometimes confusion and duplication of functions particularly in regards to issuance of water rights and control of water pollution. There is also insufficient liaison between the Department of Forestry and other stakeholder ministries in issues of conservation of forests particularly in river catchments. There are times when catchments are laid bare without regard to its long-term effect on the environment the water resources and with no plans for re-forestation of the affected areas in the immediate future. At regional and district level the five water board are mainly responsible for water supply in their respective areas of jurisdiction. For example in Blantyre the water board has no powers over other water sources apart from its dams: The City council is the one which has the overall responsibility over the environment and water quality throughout the city including the collection and treatment of sewage and monitoring discharge to rivers from sewage treatment works and industries. The City council formulates, maintains and enforces by-laws to this effect. The City councils’ performance in this area has of late however been unsatisfactory. There are therefore plans to transfer these functions to Blantyre and Lilongwe water boards. At community level, in the rural areas, there is very little IWRM worth writing about although irrigation farmers sometimes organise themselves into small area clubs which can formulate their own water conservation by-laws: There are sometimes cases whereby farmers are barred from

Water resources of arid areas

664

Figure 1. Levels for managing water in Malawi. using irrigation water for watering their animals. However such by-laws are very difficult to enforce. Consequently they are rarely followed. The NGOs, which provide water services to rural communities, also, rarely practice IWRM. Most of them blame lack of specific policy for IWRM. Their immediate concern is usually to ensure sufficient water to the community. The issues of conservation of the resource seem to come secondary, if ever.

Organization of water services

665

4 CONSTRAINTS TO IWRM IN MALAWI There are a number of reasons constraining implementation of IWRM in Malawi. The main ones are: ● Organisational. There is some conflict and overlap of functions among the main stakeholders. There is no clear and distinctive boundary of responsibility. ● Capacity. Low staffing levels and poor or lack of skills in relevant government ministries and water boards impedes the implementation of IWRM policy and strategy. ● Awareness. In Malawi there is a general lack of awareness concerning the values and importance and the procedures relating to allocation and use of water rights. ● Policy and legal environment. Laws and policies and strategies do not explicitly provide for the implementation of IWRM. ● Economic. The economy does not allow for proper management of IWRM. ● Attitude. Water is regarded as a public good to which all people has right without limitations imposed by regulations or by-laws. With this attitude there is very little regard for other uses of the resource.

5 CONCLUSIONS 1. There is great need for promotion of IWRM at all levels i.e. national, regional and community level through education and awareness campaigns. 2. Government should make deliberate policy and create enabling environment for the implementation of IWRM. 3. There is need for better coordination and collaboration among various stakeholder ministries and water boards as well as NGOS in the strategising of the promotion of IWRM in the country.

Towards best water resources management practice in small town water supply system in Tanzania A.Mvungi Water and Wastewater Engineer, University of Dar es Salaam, Dar es Salaam, Tanzania M.Makuya Dar es Salaar, Tanzania Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The water policy in Tanzania is clear on strategies for service delivery to urban and rural communities. While service delivery in urban community is purely commercial, service delivery for rural setting is required to be affordable. Community involvement in management of the service is strongly encouraged not only to create a sense of ownership, but also to ensure that they choose a service that they can afford. This paper gives practical experience on how an effective management arrangement can be put at the local level in small towns to achieve sustainable water resources management. The study took account of the current status of private sector involvement in managing water supply in Tanzania.

1 INTRODUCTION Sustainable use of water resources requires the integration of demand management with effective and efficient institutional arrangement. Therefore, the integration of demand management interventions and policies within the broader water resources management policy are essential in achieving efficient use of the scarce water resource available. Tanzania like many African countries have their small town’s inhabitants less educated, less wealth than urban dwellers, face an uncertain future due to increased populations, limited water resources, and ineffective water supply and sanitation systems. They lack possibilities to make economies of scale, and to cross subsidize the poorer members of the community. In general unable to undertake community water supply and sanitation management. The water policy in Tanzania is clear on strategies for service delivery to urban and rural communities. While service delivery in urban community is purely commercial,

Towards best water resources management practice in small town water supply system

667

service delivery for rural setting is required to be affordable. Community involvement in management of the service is strongly encouraged not only to create a sense of ownership, but also to ensure that they choose a service that they can afford. In Tanzania until recent years almost 80–90% of the rural population did not have access to clean and safe water supply. Through government initiations and donor support countries the situation have drastically changed. Currently there is water supply project going on in almost every district in the country, with emphasis given to provision of safe, adequate and reliable water supply and good sanitation. There is a great concern now however, on how best these water supply systems could be managed in a sustainable manner. This paper gives a practical experience, which dwells on how effective management arrangement can be put at the local level in small towns water supply projects to achieve a sustainable water resources management. The study focused on two small town called Mwanhuzi, and Bariadi in Shinyanga region. The two towns are situated in the northern part of Tanzania. The study aimed at investigating the social economic status of the people, their preferred service level assessment of the willingness and ability to pay for such services. Including collecting their opinion on how best they would manage their water supply system. The study took full account of the current status of private sector involvement in management of the water supply system in Tanzania and the performance of most of the utilities under different management set up in both rural and urban areas. It also reviewed the existing models of private sector involvement in management of water supply projects within the framework of the policy, and in the context of development of Mwanhuzi and Bariadi water supply projects. Looking upon the ownership of assets, roles and responsibilities of the different stakeholders, duration of the contract part of the investment in charge of operator and regulatory issues that have to be made. Accordingly, recommendations on the best appropriate options for small town water supply system have been put forward. They include the establishment of independent water board, which would be entrusted with ownership of the facility and empowered to delegate day to day operations. The board is further insisted to be free from the central government and its board members to be elected from council through subsidiary legislation to be made under the local government act. 2 OBJECTIVES OF THE STUDY The Main objectives of this study was: ● Establishment of average specific consumption/demand for water in the project areas; ● To assess the willingness and ability of the people to pay for Water supply services; ● To establish the preferred level of service and existing management arrangement; ● To propose the best Management practices for the small town water supply system.

Water resources of arid areas

668

3 METHODOLOGY 3.1 Physical observation This facilitated the preparation of plans for administering questionnaires and subsequent semi-structured interviews carried out in the towns. 3.2 House to house surveys This was carried out using pre-prepared questionnaires. Approximately 5–10% of the existing households in each town were interviewed. 3.3 Interviews of the stakeholders This was conducted with street leaders. Members of the water board (WB), Water User Groups (WUGs), NGO’s and the water vendors including informal interviews with randomly selected members of the community. 4 INTRODUCTION TO THE PROJECT AREA 4.1 Mwanhuzi town Mwanhuzi town is the administration and commercial center for Meatu district. It was established in 1988. The town is situated about 140km of Shinyanga and accessed by an all weather road from Shinyanga town. Estimated Population is about 15,000 people. In 1994 the dam that was built in 1962 collapsed and the people had to rely on shallows wells since the possibility of underground water was far from reality since its quality had higher fluoride and salinity above Tanzania standard. In 1998 a new dam was constructed with a full supply level (FSL) of 1,760,000m3 some 6km from town center. Again it collapsed in 2002 and was repaired in early 2002. The town is now assured of reliable water source. The Mwanhuzi town water supply shall therefore be run as a not for project enterprise and on a self financing basis as regard to operational and maintenance costs by changing all consumers for water supplied. 4.2 Bariadi town Bariadi town is a headquarters of Bariadi district in Shinyanga region. The town is located approximately on latitude 2°45′ South and longitude 34° East about 160km North West of Shinyanga. An all weather earth road accesses the town. The mean elevation of Bariadi is 1310m above see level.

Towards best water resources management practice in small town water supply system

669

The town is subdivided into two by Bariadi River, which flows generally westward into Duma River. The Bariadi Dam which is the proposed source of water for the township is constructed about 400m upstream of the bridge connecting the two towns. The reservoir on the Bariadi River completed around 1990 forms a major water source for the proposed water Supply system. Based on existing 1:50,000 topographical maps the catchments area above the reservoir is about 512km2 consisting of moderate catchments slopes. The current capacity of the dam is limited to about 1.5 million cubic metres, unable to meet the present population of over 60,000 people. Local people therefore, relied heavily on shallow wells and rainwater harvesting shallow wells due normally dug into the back yard of house units. Now with water supply in place different option of selling water in bulk to them so that they continue to operate as water user groups shall be looked into assessments shall also involve major sketch holders. 5 RESULTS 5.1 Specific consumption pattern It is very difficult to assess accurately the daily average consumption when the consumers are not metered, and when we do not have a reliable data about production of water. The general assessment in the two (2) towns i.e. Bariadi and Mwanhuzi was carried out by asking 10% of the total head of households the average member of buckets (20lts) that they think they use for domestic consumption. Bariadi despite being an old developed town with a sketchy water supply system the average per capital, consumption assessed through different modalities including consumption based on housing standards, was fairly low 16.7 λ/c/d when the assessment of the actual expenditure on water was done, they tied fairly well with the supposedly water collected per day per household. Mwanhuzi town on the other hand is a new developing town. The average per capital consumption through interview of sample household (10%) is 27 λ/c/d. This was also found to be fairly low specific consumption however, for all having standard is also to the average consumption (28 λ/c/d) for permanent houses and 26 λ/c/d for temporary houses. This was probably due to the uniqueness of the current modality of supply of water each household interviewed however, responded to the effect that the current water was expensive and they could do with little money water if the price was less. It was further noticed that, users have the tendency of underestimating their consumption (which can be 10% or 20% higher in reality). The analysis shows that the actual demand is completely unsatisfied by the existing economic scale of the local people. More than 80% of the surveyed u sers said they would be interested in getting more water. This observation further implies reduction in the price of water for the people to be able to afford more water.

Water resources of arid areas

670

5.2 Service levels The current predominant mode of sanitation is pit latrines. In order to introduce a different level of service of water supply the sanitation situation will need to improve accordingly. Interviews on preferred level of service took these facts into consideration by making community members aware of what it will cost to have a yard tap or multi-tap house connection. Irrespective of this awareness the preferred service level according to the house hold interviews was 61% public tap 14% yard tap and 24% multi-tap house connection. The specific consumption for the different service levels was adopted according to the Ministry of water design criteria for the lower service level but reduced for the higher service level to match the average consumption and living standards in Mwanhuzi and Bariadi. The spread of yard tap and multi-tap level of service was assumed to be 50/50 according to the spread of permanent and temporary housing standards in towns. The level of service is assumed to increase due to efficient supply situation to reach a high level observed in well-supplied municipalities like Tanga with about 34% of the population relying on public taps. The above service level average to about 32.5 λ/c/d is assumed to be a suppressed demand. Based on the average consumption derived from reduction in the price of water the consumption of 421 λ/c/day was adopted as the current water consumption for the towns. One can actually observe a strong demand and a strong willingness to pay for an improved water service. Care must be taken, as willingness does not mean a high ability to pay; as the level of poverty in most of small towns remains quite high. It is however, within the framework of the project not only to improve the service itself, but also to restore the trust of users in the public service and it’s ability to meet the real demand of these users. 6 MANAGEMENT ARRANGEMENT 6.1 Review of existing models for small towns water supply and sanitation services management The preparatory study made by BCEOM in Tanzania identify four basic modes for management of small town water supply and sanitation in terms of ownership of the utility infrastructure, the identity of the system operator, the legal status of the system operator, and the ownership of shares of the operating company where applicable. These modes are briefly described below. 6.1.1 Water users committee direct management Water service is direct managed by a water committee, who is in charge of the scheme, with the support of the District Water Engineer. In some cases the water committee decides to delegate a large part of the operating functions (revenue collection, fuel or gas

Towards best water resources management practice in small town water supply system

671

oil supply, day to day operation, small repairs on the scheme, spare parts supply etc) to a private operator actually it’s just an individual. In this kind of arrangement, the private operator collects money from users, pays all running costs, and gives a determined amount of money to the water committee. 6.1.1.1 Advantages of this model ● Water committee is running the schemes without any public subvention ● Water committee find themselves responsible for a certain kind of social equity in the community, people decide who is exempted of payment for water Water committee are elected as community based entities women have representatives in water committee. At least, water committee are accountable of the water supply and sanitation management to the population. 6.1.1.2 Limitation of this model

● Usually, water committee can’t manage facilities which require technical skills or they have to rely on the District Water Engineer for support ● In most of the cases water committee have no legal status, and their ownership of the system they manage is not official recognised ● Water committee members are not always democratically elected. The representation of women and young people is rather poor. 6.1.2 District operated water supply scheme In this case a team placed under the authority of the District Urban Water Engineer directly operates the scheme. All the main employees are civil servants, with some additional staff with time-limited contract. The district does all the accountancy of the water service. 6.1.2.1 Limitation of this model ● Users are not represented in this management, and there is no way for them to complain about the quality of the water service. ● The water service depends strongly on the central government for all the investments costs, and a large part of the running costs. ● The management team does not have any mean nor incentive to increase the coverage of service by extending the piped scheme or connecting new customers. 6.1.3 The Water Supply Companies model Water Supply Company is a private company limited by guarantee. The Board of Directors is formed by Water User Groups (WUG) each WUG representing a domestic

Water resources of arid areas

672

water point, considered as the ‘lowest level of representation’. According to the size of the system, the water supply company can appoint a few remunerated people to ensure the day-to-day management of the scheme. 6.1.3.1 Advantages of this model ● The Water Supply Company is accountable to the general assembly of the members, who are direct representatives of the users, this maintains a kind of social pressure and promotes transparency ● Water Supply Company have a proper accounting system 6.1.3.2 Limitations of this model ● Individual domestic customers are not represented as a specific category of users ● The external support, which is really a basic need for a small water supply company, remains an option, as well as the involvement of the private sector ● It is quite impossible for a water supply company to have enough funds to recruit qualified personnel, the customer’s base is too weak to generate enough incomes to allow the water supply company. 6.1.4 The urban water and sewerage authority (UWSA) model Authority is a semi autonomous entity created under the auspices of the Ministry of Water. From the legal point of view, these Authorities can be a public, private, cooperative entity. Each Authority is independent in terms of staff management. Each Authority is allowed to delegate outsource some functions to other operators. The Authority signs a ‘memorandum of understanding’ directly with the ministry, which specifies the roles and responsibilities of each part, as well as the performance indicators and the obligations of reporting to the Ministry (weekly and monthly report). 6.1.4.1 Advantages of this model ● All the categories of users are represented in the Board ● Authorities have quite incentives to improve the management of the water service, especially from a “commercial” point of view 6.1.4.2 Limitations of this model ● Authorities remain entities subsidised by the central government, even if they are supposed to acquire gradually their financial autonomy ● The independence of the Authority is very relative, because the Ministry directly appoints its directly accountable to the Board, and at least two members of the Board.

Towards best water resources management practice in small town water supply system

673

7 EXISTING MANAGEMENT ARRANGEMENT IN THE TOWNS There is currently some form of a Water Board in existence in the towns researched. This board, which was formed under the initiative of the District Council (DC) and the District Rural development Program (DRDP), is initially responsible for organizing and raising funds for financing the water supply project. The funds are at the moment very limited. The board consists of the Board Chairman, Secretary other 7 members making a total of 9 members. Following approval of financing of the water supply project, the DC and the Water Board organized training with the following objectives: ● To formalize structure of the Board ● Establish role and responsibility of the Board ● To prepare the board for participation in the construction of facilities. ● To prepare the board for the management of the water supply project. The workshop concluded that the current board was representatives of the water users and that they preferred to form a water user association. This formation of management however is more suited for rural water supply projects than for small town water supply systems. 8 DISCUSSIONS 8.1 Options for management of Small Town Water Supply systems A number of documents reviewed established an option for management of the water supply system. The water policy however provided guideline for appropriate management options. The Small Town Water Supply and Sanitation Project Preparation Study provided experience from other countries, and Uganda in particular. Some of the recommendations made in the document have been endorsed. In general, the exposure to the current status of private sector involvement in management of water supply systems in Tanzania and performance of most of the utilities under different management setups in both rural and urban areas has also been utilized as basis to recommend an option. The following can be concluded from the above review: (1) The water policy is clear on strategies for delivery of services to Rural and Urban communities. While services delivery for urban areas is purely commercial, service delivery for rural areas is required to be affordable. Communities’ involvement in management of the services is strongly encouraged not only to create a sense of ownership, but also to ensure that they choose a service that they can afford. Although small towns, under the current Ministry setup fall under the Directorate of Rural Water Supply, the real policy guideline on future management to ensure

Water resources of arid areas

674

sustainability of service delivery to small urban centers is provided under the Urban Water Supply and Sewerage policy section 4.9. The policy elaborates that in order to have improved water supply and sewerage services in small urban centers: (a) Emphasis will be on privatizing water supply and sanitation services. (b) Local Private Sector institutions shall be promoted and strengthened. Their access to credit facilities will be enhanced. Both components of the policy (Rural and Urban) insist on cost recovery as the basis for sustainable service delivery. The tariff to be developed has to pay for full cost recovery of the water supply system. The Small Town Water Supply and Sanitation Project Preparation Study reviews the existing models of private sector involvement in management of water supply projects. Within the framework of the policy, and the context of development of Mwanhuzi water supply the following models are possible: ● Management Contract ● Leasing Contract ● Concession to private enterprise Whatever model is adopted issues of ownership of assets, roles and responsibilities of the different players, duration of contract, part of the investment in charge of the operator and regulatory issues have to be made clear. The study insists on the board being totally free from the central government and the board members to be elected by the council through a subsidiary legislation to be made under the local government act. (2) A review of the existing institutional setup of some of the water supply facilities has also been made. The Urban water Supply and sewerage Authority set up seem to be OK. The role of the Ministry in backstopping (regulation) is fairly strong which is good. The powers given to the boards however seem to be overshadowed by the involvement of the Ministry (boards not dynamic enough) and the ability of the entities to become full-fledged commercial entities, which is the ultimate objective, is still far fetched. Utilities, which have been operating under boards of trustees like the Hai water supply projects in Kilimanjaro region, have proved to be performing fairly well. These utilities are already charging tariffs, which ensure full cost recovery for the water supply system. The board of trustees in this respect acts as a regulator, and the success could be attributed to the effectiveness of the regulatory mechanism put in place. Experience from operating limited liability companies and in particular KILIWATER, are not encouraging. The company was established as a limited liability company under the company ordinance cap 212. The company is owned by users who acquire shares from the company, and is totally independent from central government and the local government. A board of directors appointed by the users oversees the operation of the water supply system. The company however has not been operating very well due to among other things internal problems, which require external backstopping mechanisms to resolve. Experience from this company also points out to the need of an effective and efficient regulatory mechanism to ensure success.

Towards best water resources management practice in small town water supply system

675

The ongoing reform and restructuring of the Government and developments in private sector involvement in the management of the major utility companies in the country in particular TANESCO and DAWASA has led to establishment of Energy and Water Utility Regulatory Authority (EWURA) for the purpose of regulating delivery of services by the private sector. This regulatory board has not yet been extended to the district level. The mechanisms of operations of the regulatory board need to be reviewed to confirm that it will address the needs as briefly highlighted above. 9 RECOMMENDATIONS 9.1 The proposed institutional arrangement for Mwanhuzi water supply The overall arrangement envisages setting up of an autonomous water board in small towns, which will be entrusted with the management of the facility. The legality of this board will be in accordance with the revised Waterworks Ordinance, which empowers the Minister responsible for water to declare a township an Urban Water and Sewerage Authority. The custodian of the initial investments on the facility is the Local Government. The Central Government through the Ministry of Water and Livestock Development however, is also deeply indirectly involved through soliciting and acting as a guarantor of the necessary financing for construction of the facility. These are the primary stakeholders who are interested in ensuring that the management option set up is successful. These should together provide the regulatory functions in the absence of an EWURA. A strong regulatory framework is considered necessary for the success of the management of the water supply system. The regulator of the authority has to be close to the area and easily reachable. The Regional Water Engineer is in this respect expected to represent the Ministry of Water and Livestock Development, while the District executive Director represents the District. The selected board will go into contract with the District council under the witness of the representative of the Ministry responsible for water and be entrusted with the ownership of the facility and the subsequent successful management thereof. Board members will be selected from the water users of Mwanhuzi town. Tariffs will be set from the onset with the objective of full cost recovery. The expectations and responsibilities of the board will need to be defined from the onset. The board shall have the powers to delegate the day-to-day operations and management of the water supply system to a private operator. The study recommends a management contract as the most appropriate management option for small town water supply systems. An independent water board to be established which would be entrusted with the ownership of the facility and empowered to delegate the day-to-day operations and management of the water supply to a private operator (management contract).

Water resources of arid areas

676

9.2 Operations of the institutional structure: Roles and responsibilities of stakeholders 9.2.1 The District Council The District Council (DC) should be the owner of the water supply system. It is the entity entrusted by the central Government to oversee development at the district level. The Council represented by the district executive director (DED causes the establishment of a Water authority through application to the responsible Ministry. The DC makes the water users aware of the intent and streamlines/defines requirements such selection procedure for members of the water board. A strong dialog should be established between the Ministry of water represented by the Regional water Engineer and the DC to ensure that all conditions necessary for establishment of a Water Authority according to the water policy are made. 9.3 Water Board The water users elect this Board. It considers the interests of all water users in the small towns. The DC may assist the water users on selection of representatives to the Water Board. This board shall among other update demands and service levels in line with what may be possible and look into their own ranks incase of outstanding payments. 9.4 Private Operator The Private operator will be selected through tendering procedure for the provision of services. The by DC and Water Board will prepare the terms of reference (ToR) under assistance of the regional water engineer. The principles of awarding the contract to the winning Private Operator are: Responsiveness to the ToR including availability of business plan, and Competence. 9.5 Water users The beneficiaries of the system will be provided with their demands according to the willingness and ability to pay established in the beginning and as updated through debates in the water board. ACKNOWLEDGEMENTS The authors would like to thank the Managing Director of SERVICEPLAN LTD for the financial support offered during this study.—Thanks to all the District Councils, for their assistance during the entire filed works.

Towards best water resources management practice in small town water supply system

677

REFERENCES MoWLD (Ministry of Water and Livestock Development) 2002. National Water Policy Tanzania Printing Office, Government of Tanzania. MoWLD (Ministry of Water and Livestock Development) 1997. Waste Supply and Waste Water Disposal Design Manual 2nd Draft Vol. 1 of 1997 United Republic of Tanzania. MoWLD (Ministry of Water and Livestock Development) 2001. Annual Report for Urban Water Supply and Sewerage Authorities, Government Printers, United Republic of Tanzania. MoWLD (Ministry of Water and Livestock Development) 2002. Small Towns Water Supply and Sanitation Project Preparatory Study, Vol. 1 of 2002, Government Printers, United Republic of Tanzania. Van der Zaag, P 2000. Water law notes, Department of Civil Engineering, University of Zimbabwe, Harare.

Water management in the Mauritian textile wet processing industry N.Kistamah & S.Roseunee Department of Textile Technology, Faculty of Engineering, University of Mauritius, Reduit, Mauritius Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Water is a universal medium for the wet processing of textile substrates. In Mauritius, water supply is ensured by water collected during rainfall which is seasonal. It is therefore important that water is used efficiently. The use of textile wet processing equipment with low liquor ratios, reliable controllers, the application of improved processes & recipes, the use of better performing dyes and auxiliaries, and the management of water consumption through water reuse strategies and its control through the use of flow meters, are some of the steps which have been taken by a few wet processors to manage and control water usage. Besides, the cost of wastewater discharge to the industrial sewer has increased by 600% over the last five years and this could become a significant driver for the optimisation of water consumption.

1 INTRODUCTION Over the last 25 years, with industrialisation and population growth there has been an increasing demand for water. Enormous pressure has been exerted on water bodies for the continuous supply of fresh water for industrial, commercial, and domestic usage. In Mauritius, water supply is ensured by water collected during rainfall. The island receives on average 2100mm of rainfall with the higher elevation regions receiving up to 4500mm (Water: Resources, Uses & Pollution, 1999). According to the Falkenmark index of water scarcity, between 1990 and 2004, Mauritius has moved from a water-abundant to a water stressed country. Its annual per capita of renewable fresh water has dropped by 17%, from 1750 to 1450m3 (Allybokus et al., 1996). In the past, most of the surface water resources were exploited but with recent industrial development, groundwater resources are being continuously developed to supplement the increasing demand in water. The annual volume of underground water used by the various sectors (domestic, commercial, industrial and tourism) of the country

Water management in the Mauritian textile wet processing industry

679

is 145 million m3 out of a potential usable volume of 160 million m3. The remaining required annual volume of 800 million m3, predominantly for agricultural and hydroelectric power production, is supplied by the surface water sources. In the industrial sector, textile mills are notorious for its large consumption of water, in particular the textile wet processing industry. Water is an excellent medium, relatively cheap and safe to use for textile preparation, colouration and chemical finishing. The textile and apparel industry uses both surface water from rivers, government pipe supply and groundwater from boreholes as per Table 1. In dyehouses alone, water consumption has increased from 5000m3/day to about 30,000m3 per day with an increase in the number of dyehouses from 6 in 1983 to 32 in 2003 (Ramgulam et al., 2000).

Table 1. Consumption of water in m3 per month in the textile industry as per water source. Surface river Groundwater Government water supply 80,000

490,000

340,000

Table 2. Cost of water supply, on average, per m3 in US$ for the industrial sector. Surface river Groundwater Government water supply 0.07

0,11

0,59

The 600% increase in water demand requires attention otherwise water availability can be a limiting factor in the development of the sector. Therefore, comprehensive programmes of water management may be required to ease off this situation. Integrated water management in the textile wet processing industry, in the current context, is not just an environmental initiative. One of its most basic premises is that it improves efficiency and productivity for the industry. These improvements are seen in lower expenditure on resources such as energy and water, increased efficiency in production, fewer risks associated with environmental impacts, and decreased waste-water generation that leads to savings in water treatment costs. In addition, the textile wet processing industry in Mauritius has for too long relied on end-of-pipe solutions without seriously developing and implementing a strategy of minimizing wastewater. Water scarcity and cost is forcing the industry to recognise and develop a relationship between business, water resources and the environment. 2 FRESH WATER AND WASTEWATER CHARGES It is interesting to note that both the fresh water charges from government supply lines and wastewater charges have increased over the last few years. The comparative fresh water charges are given in Table 2. The cost of wastewater charges for factories discharging in industrial sewerage system have increased by more than 600% over the last five years; from $0.07 to $0.52 per m3.

Water resources of arid areas

680

Therefore, water management not only bring relief on the water supply but economic benefits to consumers through a reduction in wastewater charges, water, chemical and energy savings. 3 DYEING MACHINERY 3.1 Liquor ratio Liquor ratio is the volume of liquor consumed in litres per kg of fabric wet processed. Typically, each kilogram of finished textile product requires between 70–250 litres of water. But the growing demand in environmentally friendly processes and the increasing cost of freshwater have forced dyeing machine manufacturers to design and develop more water and energy efficient machinery. For example, the use of air-jet dyeing technology in exhaust dyeing has considerably reduced the volume of water consumption per kg of fabric. In certain cases, the liquor ratio has dropped from 110L/kg to about 65L/kg of fabric. For a medium size dyehouse processing 10 tons of cotton fabrics per day, this may represent a saving of 450m3 of fresh water per day and US$ 175,00 per year for fresh water and wastewater charges. The low to ultra-low liquor ratio not only saves water but also allows for shorter dye cycles and saves chemicals and energy. 3.2 Rinsing techniques In the dyeing of cellulosic materials such as cotton with reactive dyes, a significant proportion of the wet processing cycle is taken up by rinsing. It is estimated that more than 50% of the water consumption for a wet process is for rinsing purposes. Over the years, a lot of attention has been given to the technology of rinsing which has evolved greatly from the conventional drain/fill or overflow to smart rinsing technique. The new technology offers significant savings in water especially on low liquor ratio dyeing machines (Bradbury et al., 2000). 4 OTHER MACHINE FEATURES FOR BETTER WATER MANAGEMENT ● Flow control valves and meters for pre-set volume of water during processing or rinsing ● Efficient sensors for the monitoring of critical dyeing parameters such as pH, temperature and electrolyte content ● Automatic chemical and dispensing systems for minimising spills, precise dispensing and more reproducible results ● Sophisticated adaptive controllers for better and user-friendly control of process The features listed above serve to directly or indirectly achieve water savings without negatively affecting quality of the finished product. The overall objective is to maximise Right-First-Time production with optimum water usage.

Water management in the Mauritian textile wet processing industry

681

5 WATER REUSE Technologies for reuse of water also hold possibilities for the industry to reduce water consumption. Counter-current washing in continuous processing where clean water enters the final wash box and flows counter to the movement of the fabric through the wash boxes and the reuse of not too contaminated rinsing water may be used for processes which do not require very high quality water. For example, in the dyeing of cotton with reactive dyeing on jet machines, water from the last cold rinse may be used for the preparation of the scouring and bleaching baths. 5.1 Case study A dyehouse produces 10 tons of reactive dyed cotton knitted fabrics per day using air-jet dyeing technology. The liquor ratio is 70L for every kg of dyed fabric. Its daily water consumption is 700m3. According to a recent study (Tulsi, 2003), the last rinse of the reactive dyeing process may be segregated for reuse. The water parameters such as total hardness, pH, conductivity, colour content of the rinsed water are acceptable for safe reuse for fabric preparation. The whiteness index of the prepared fabric, prior to dyeing, is comparable to that when using fresh water. Assuming that the rinsing time, using continuous rinsing technique, is 5 mins with a flow rate of 120L/min/100kg, the volume of water saved per day is 60m3. This represents a saving of more than $20,000 per year in fresh water and wastewater charges. 6 PROCESS MODIFICATION Process modification as a means to save water in textile wet processing has been widely investigated (Vigo, 1994). The focus has been on combining textile preparatory processes which traditionally consume large volumes of water. Desizing, scouring and bleaching of woven textiles are mandatory when dyeing pale to medium shades. Traditional methods of processing involve carrying out each process followed by extensive rinsing to wash out processing chemicals, degraded starch, cellulose, and lignin-based by-products. Nowadays, the technology of one-stage process has gained more importance with water, chemical, energy and time savings as very attractive benefits. The use of better performing dyes and auxiliaries has also helped to save water. The first generation reactive dyes for natural fibres suffered from the serious drawback of low fixation efficiency with figures in the range of 60–70%. Now, high fixation reactive dyes with fixation levels of the order of 80–90% may be attained. Besides, following dye application in the presence of large amounts of electrolyte, the substrate has to be subjected to 5–6 rinsing and washing cycles to remove electrolyte, alkali and unfixed dyes. This procedure is both time and water consuming, and generates large volumes of coloured effluents. New generation reactive dyes are mostly polyfunctional and their fixation efficiencies are vastly superior due to the synthesis of more stable fibre-reactive groups. They also confer the advantages of low salt requirement and the removal of the unfixed dyes requires less water.

Water resources of arid areas

682

7 SUPERCRITICAL FLUIDS Supercritical fluids (SCF) such as supercritical carbon dioxide (scCO2) have been considered as solvents for chemical processing of textile materials (Ozcan, 1998). The underlying physics and chemistry of SCF-polymer and SCF-solute interactions are now well understood (Johnston, 1989). Densities and viscosities in supercritical fluids are less and diffusion into the polymer is more rapid than in liquids, shortening the process time and improving productivity. scCO2 itself as a fluid is deemed to be low-cost, environmentally friendly, non-flammable and potentially avoids water usage. It has low critical parameters (31°C, 73.8 bar) and can be recycled. In textiles, it has been used for a range of processes that include scouring, dyeing, dry-cleaning [D] and impregnation of functional finishes. The dyeing of polyester fibres with disperse dyes in scCO2 has been widely investigated and was found to offer a number of advantages when compared to aqueous dyeing (Saus, 1995). For example, no reduction clear (removal of surface deposited disperse dye) was required. This process usually consumes large volumes of water in order to achieve commercially acceptable wet fastness and gives rise to coloured effluents. The colouration of polyester-type fibres in SCF is, therefore, an attractive alternative to aqueous dyeing. The colouration of natural fibres in scCO2 is still at the pilot scale but new knowledge in this area indicates that “dry textile dyeing” may be a possibility in the near future. 8 AWARENESS CAMPAIGN 8.1 Management and employee commitment Some companies have developed and adopted a comprehensive policy that definitively states their commitment to water management. Experience has shown that employees are extremely knowledgeable about sources of waste in their facility and are an excellent source of ideas for reducing water consumption. Savings can often be achieved with little or no capital expenditure by merely changing management practices. This is one of the keys to cleaner production-it need not cost money and often requires little additional resources and time. Training of new staff members about water management practices ensures continuity of the policy. 8.2 Right-first-time production One very important principle in textiles is right-first-time production, which reduces water consumption by avoiding reprocessing of orders. The amount of off quality production runs and reworks in Mauritian dyehouses varies between 10–20%. Planning of orders is also a critical factor since in many instances dyeing machine filling capacities are not optimised, with figures ranging from 50% to a potential of 90%. The amount of water consumed per kg of fabric increases significantly.

Water management in the Mauritian textile wet processing industry

683

8.3 Education A more long-term approach to water management can be taken through formalized employee education. Education programs are more general and less job-oriented than training programs. There is a need for an in-depth understanding, for example, of the chemistry & theory of dyeing and the design & components of the dyeing equipment. This knowledge is essential for water management and long-term improvements. In general, the training is best conducted internally because job-related issues are very site-specific. On the other hand, general education can be conducted either internally or externally. The University of Mauritius has conducted a number of conferences and seminars, in-plant courses by experts in the area of water conservation to help the industry in this respect. 8.4 Equipment maintenance and operations audit Poorly maintained equipment leads not only to bad work, off-quality production runs, high reworks, and poor employee attitudes, but also to increased water consumption. Faulty seals should be repaired. Housekeeping and maintenance are essential for leak control. Preventive maintenance is the solution to these problems and can be accomplished through proper audits. 9 CONCLUSION Mauritius is potentially a water-stressed country, especially during low seasonal rainfalls. Proactive action by water authorities have ensured timely introduction of new legislation to sustain water quality and maintain a fairly good standard of water management practices. However, the exploitation of ground water resources to the level of 90% of available stock is alarming and exposes the network to serious risks of contamination. The Mauritian textile industry, confronted with a number of internal and external challenges, has adapted itself quickly to new standards of clean production and management, thereby reaping both economic and financial benefits. Leading textile buyers now measure the competence of a producer not only in terms of cost of production and delivery schedules, but also environmental parameters such as water efficiency. The textile wet processing sector is on-course to adopt “dryer” technologies as quality problems are addressed through sustained research and development. REFERENCES Allybokus M.E. & Ramjeawon T. 1996. Efficient uses of water in industry and agriculture. World Day For Water Conf., Civil Engineering Department, Faculty of Engineering, University of Mauritius. Bradbury M.J., Collishaw P.S. & Moorhouse S. 2000. Controlled rinsing: A step change in reactive dye application technology. Colourage Journal India, Annual 2000, 73. Johnston K.P. & Penninger J.M.L. 1989. American Chemical Society: Washington DC, 406, 207.

Water resources of arid areas

684

Ozcan A.S., Clifford A.A., Bartle K.D., Broadbent P.J. & Lewis D.M. 1998. J. Society of Dyers and Colourists, 114, 169. Ramgulam R.B., Kistamah N. & Rosunee S. 2000. Study of dyehouse effluent treatment in Mauritius. Mauritius Research Council Project Report. Saus W. & Jasper J. 1995. Textile Technology International, 145. Tulsi S. 2003. Potential reuse of rinsing water in dyeing. BSc. Thesis, Department of Textile Technology, University of Mauritius, Mauritius. Vigo T.L. 1994. Preparatory processes. Textile processing and properties, Textile science andtechnology series 11, T.L.Vigo, 1st ed., Elsevier Science B.V., The Netherlands, 4–31. Water: Resources, Uses & Pollution, (1999), Ministry of Environment, Mauritius, http://www.intnet.mu/iels/%20water_mau.htm

Analysis of the microbiological situation of the quality of domestic water sources and identification of the microorganisms in them, located in the semi-arid regions of the Eastern Cape, South Africa M.Zamxaka, G.Pironcheva and N.Y.O.Muyima Environmental and Natural Products Biotechnology Research Group, Department of Biochemistry and Microbiology, University of Fort Hare, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The water in Alice and Gogogo, which belong to the semiarid areas of the Eastern Cape province of South Africa, was characterized by using both standard microbiological methods and the standard physical methods to investigate its present quality in the sampling sites. For microbial analysis, indicator bacteria, namely, heterotrophic, total and faecal coliforms and physical parameters, such as pH, turbidity and temperature were assessed to check whether the water from dams, rivers, wells, etc. is safe for drinking. Almost all the indicator bacteria counts were above the South African standards. The physical parameters, such as alkalinity of water and high turbidity, proved favourable for bacterial growth. In another series of experiments we identified, using the API20E Assay kit, 54 different species of microorganisms, about 77.5% of them being human pathogens, 53.2% belonging to the family of Enterobacteriacae and only 22.5% being non pathogenic. Our investigations prove that the water in the domestic water sources of Alice and Gogogo is of poor quality and needs further purification.

1 INTRODUCTION The lack of safe drinking water and adequate sanitation measures leads to a number of diseases, such as cholera, dysentery, salmonellosis, typhoid, etc, which claims the life of 2 million people each year in the developing countries (WHO, 1993). The primary goal

Water resources of arid areas

686

of water quality management from a health perspective is to ensure that consumers are not exposed to doses of pathogens that are likely to cause infectious diseases (Pegram et al., 1998). Protection of water sources and treatment of water supplies have greatly reduced the incidence of these diseases in developed countries (Craun, 1986). One of the difficulties in evaluating the impact of the drinking water supply on health is the lack of local demographic statistics, particularly in rural communities. Therefore, it is important to know the incidences, occurring in the rural semi-arid areas, due to polluted water. This will give the opportunity to compare the incidence of water borne diseases between communities that have drinking water and those that do not have. Detection of bacteria, potentially toxic substances and other contaminants usually requires laboratory conducted tests. Detection and enumeration of indicator organisms, is the basic technique, used in water quality monitoring. Coliform group of bacteria can be defined as the principal indicators of purity of water for domestic, industrial or other uses. High faecal and total coliform counts in water are usually manifested in the form of diarrhea, fever and other secondary complications (Fatoki et al., 2001). In the Eastern Cape Province, belonging to the semi-arid areas of South Africa, nearly 80% of the population relies on surface water as the main water source. Almost 30% of the population is without proper water supply services, which implies that many of the people still utilize untreated surface water for domestic purposes. The incidence and prevalence of water-borne pathogens is subject to geographical factors. Most of the pathogens are distributed worldwide, but outbreaks of some diseases for instance cholera, shigellosis and typhoid tend to be regional (Grabow et al., 1994). Meteorological events and pollution are a few of the external factors, which affect physicochemical parameters, such as temperature, pH, and turbidity of water. They have major influence on biochemical reactions that occur within water. The purpose of the present study was to determine the status of domestic water sources, used by Alice and Gogogo rural communities. 1.1 Materials and methods 1.1.1 Growth media ● Bacteriological agar ● m-Endo Les agar ● R2A agar, oxoid, low nutrient medium All growth media and chemicals used in this study were purchased from Merck, Saarchem, Gauteng, South Africa. 1.1.2 Chemicals Sodium thiosulfate API20E kit (Bio Me’rieux, Lyon, France). Filter membranes (0,45mkm, pore size, 47mm diameter) (Microsept, Cape Town, South Africa).

Analysis of the microbiological situation

687

1.2 Water sampling For routine sampling a map 1:10000 of the population sites of interest was designed using geographic information system technology. The sampling points were marked with a circle. Water samples were collected into 1-liter and 2-liter containers and transported on ice in cooler boxes to the laboratory. 1.3 Physico-chemical analysis Temperature was determined in situ at the sampling point, using a mercury thermometer. The pH was measured using a pH meter. Turbidity was measured using turbidimeter. 1.4 Microbiological analysis Untreated water, i.e. water directly from rivers, dams, wells, etc. was directly analysed. Heterotrophic bacteria (HPC). Samples, containing heterotrophic bacteria were analysed after serial dilutions. HPC bacteria were cultured on R2A low nutrient medium oxoid agar, incubated at 28°C for 3–7 days. All the colonies were counted and counts plotted into a graph. 1.5 Total coliforms The counts of total coliform bacteria were determined by membrane filtration method. Five to ten ml of untreated water was filtered using 0.45mkm pore size, 47mm diameter filter membranes. The membranes were placed on m-Endo Les agar and plates incubated at 37°C for 24hrs. All the metallic sheen colonies on the filter membranes were counted and processed as mentioned above. 1.6 Faecal coliforms After water filtration the membrane filters were plated on m-FC agar and the plates incubated at 44.5°C for 24 hrs. Only the blue colonies that appear on the filters were counted and processed as described above. 1.7 Statistical analysis An interactive statistical software, namely statistica was used both in the analysis of the data and in drawing the graphs. For the software factor levels were constructed for the sites and time. Identification of bacteria using the API20E assay. The API20E kit was used according to the manufactures’ instructions. It is an identification system for Enterobacteriacae and other non-fastidious Gram-negative rods. The principle is the use of 21 standardized and miniaturized biochemical tests and database for reading the tests.

Water resources of arid areas

688

2 RESULTS The results of the physical, chemical and microbial analysis (pH, Temperature, Turbidity, HPC, TC and FC) analysis of Alice and Gogogo domestic waters show that both microbiological and physical parameters are above the South African standards and there is a necessity of good purification systems of the water investigated. The pH of 17 different sites of surface water has been measured in the semi-arid areas of Alice and Gogogo as seen from Fig. 1. The optimum pH values for indicator microorganisms, especially coliform bacteria are from pH 3 to pH 10.5. The overall picture of the pH graph shows that Gogogo village has sites with lower pH values as compared to Alice. Most of the sites studied in Alice have high pH values of water. This alkalinity of Alice waters plays a very important role inhibiting the growth of pathogenic microorganisms such Vibrios, Salmonellas, Shigellas, etc. The temperature of the surface water was also investigated at different sites of Alice and Gogogo. In overall, Alice Temperatures of water are higher than Gogogo temperatures, which favour the growth of Salmonella species, Shigella species and E.coli. in Gogogo areas, on the other hand, the water temperature favours the growth of Vibrio species.

Figure 1. Box plots of 17 selected sites in Alice and Gogogo against pH of water for the period starting from 3 July 2001 to 16 July 2002.

Analysis of the microbiological situation

689

Figure 2. Box plots of 17 selected sites in Alice and Gogogo against heterotrophic plate counts (CFU/ 100ml) of water for the period starting from 3 July 2001 to 16 July 2002. The effect of the Turbidity factor was investigated as well on bacterial growth and it has high values both for Gogogo waters and Alice waters which make them not appropriate for water supply. In another series of experiments we have studied the heterotrophic plate counts of bacteria from the water sources of Alice and Gogogo. Our results show that both Alice and Gogogo water sources have high HPC counts (Fig. 2). The reason for this is that the dams in Alice are not protected and are used by domestic animals. On the other hand Enkolweni spring which is the main site used by most people in Gogogo shows high HPC counts due to children play and bathing in it thus introducing faecal contamination in to the drinking water. The quantity of the Total coliforms was investigated in domestic waters in the rural communities of Alice and Gogogo (Fig. 3). The overall pattern of the graph shows that the counts decrease from Gogogo to Alice. Gogogo has higher TC counts as compared to Alice. Gogogo areas have been known to have a number of infectious diseases such as cholera, typhoid fever, etc.

Water resources of arid areas

690

The number of faecal coliforms was also investigated from different water sampling sites of Alice and Gogogo (Fig. 4). Our results show that the most contaminated sites appear to be those, corresponding to Gogogo area. The lowest values of contamination with faecal coliforms are observed in the water sources belonging to the Alice area. We have checked a total of 338 isolates from both Alice and Gogogo water supplies— 183 isolates from Alice area and 155 isolates from Gogogo water sources. Using both standard microbiological methods and the API20E commercial kit, we were able to identify 54 different species of microorganisms, including both pathogenic and non-pathogenic ones. Our results indicate that 77.5% of them are human pathogens and 53.2% of them belong to the family Enterobacteriacae. Only 22.5% of them proved to be non pathogenic to humans. Comparing the distribution of species in the drinking waters of Alice and Gogogo, it is observed that the percentage of species belonging to Enterobacteriacae family is 7.84% higher in the Gogogo water sources. The high Shannon Weaver Index (H) for both areas—Alice—2.84 and Gogogo—2.24 also indicates high biodiversity of microorganisms from sources in both areas.

Figure 3. Box plots of 17 selected sites in Alice and Gogogo against total coliform counts (CFU/100ml) of water for the period starting from 3 July 2001 to 16 July 2002.

Analysis of the microbiological situation

691

Figure 4. Box plots of 17 selected sites in Alice and Gogogo against faecal coliform counts (CFU/100ml) of water for the period starting from 3 July 2001 to 16 July 2002. 3 DISCUSSION The purpose of the present study was to determine the microbiological quality of the domestic water sources used by the rural communities of Alice and Gogogo. Our experiments show that both the total and faecal coliform counts are above the South African recommended standards for drinking water in almost all the sites studied. The high total and faecal coliform counts, TC—3162 colonies per 100ml and FC—10000 colonies per 100ml are the results of contamination especially in dams where the water is used by domestic animals, clothe washing and is exposed to heavy rainy falls. Another possible contamination of water sources might be the presence of pit latrines close to them, little environmental protection and poor catchment points management (Muyima & Ngcakani, 1998). Microbial growth and physical quality of water are considered as priority parameters to be monitored in the rivers, dams and boreholes catchments (Fatoki et al., 2001). The presence of faecal coliform bacteria indicates that the water is contaminated with faecal of humans or animal waste while the total coliforms counts

Water resources of arid areas

692

indicate that the water is contaminated with both faecal waste and other bacteria from the soil. Our study on the identification of the microorganisms in these domestic water sources further shows the presence of 54 different species of microorganisms, out of which 53.2% of them belonging to the family Enterobacteriacea, responsible for serious enteric diseases. The high number of species isolated from the drinking water of the two areas, as well as high Shannon Weaver Indexes are an indication of high water contamination. The water sources in both areas of Alice and Gogogo show high levels of contamination, predominantly by species of the Enterobacteriacea family. All the water supplies, checked need extensive and efficient purification so that they can be used for drinking water. REFERENCES Craun G.F. 1986. Water borne diseases in the United States. 295–302. CRC Press, Inc. Boca Raton, Florida Fatoki O.S., Muyima N.Y.O & Lujiza N. 2001. Situation analysis of water quality in the Umtata River catchment. Water SA. 2(4), 467–474. Grabow W.O.K., Favorov M.O., Khudyakovan S., Taylor M.B. & Fields H.A. 1994. Hepatitis E seroprevalence in selected individuals in South Africa. J. Med. Virol. 44, 384–388. Muyima N.Y.O. & Ngcakani F. 1998. Indicator bacteria and regrowth potential of the drinking water in Alice Eastern Cape. Water SA. 24, 29–34. Pegram G.C., Rollins N. & Espey Q. 1998. Estimating the cost of diarrhea and epidemic dysentery in KwaZulu Natal and South Africa. Water SA. 24, 11–21. WHO. 1993. Guideline for drinking water quality. 1. World Health Organization. Geneva, Switzerland.

Dry season Kalahari sap flow measurements for tree transpiration mapping—Serowe study case, Botswana M.W.Lubczynski1, A.Fregoso2, W.Mapanda1, C.Ziwa1, M.Keeletsang1,3, D.C.Chavarro1 & O.Obakeng1,3 1

International Institute of Geoinformation and Earth Observations, ITC, Enschede, The Netherlands 2 Instituto Nacional de Ecologia,, Mexico D.C., Mexico 3 Geological Survey of Botswana, Lobatse, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Tree transpiration (T) is an important component in assessment of groundwater balance on Kalahari. T can be well estimated by sap flow measurements consisting of measurements of sap velocity and sapwood area. In the last three years (2001–2003) the extensive, dry season field campaigns in the Serowe study area were carried out to characterize sap flow on Kalahari. The sap velocity was measured using Granier’s thermal dissipation probe technique while sapwood area by tracer tests over the cut sections of tree stems. All those measurements, after statistical analysis, resulted in transpiration upscaling functions i.e. species-specific linear relations between sapwood area and stem area and between sapwood area and canopy area and species-specific mean sap velocities. The derived upscaling functions were applied for the dry season transpiration assessment over the selected 10×10km2 Kalahari study area in 78 plots of 35×35m according to the regular grid of 1×1km. That assessment indicated substantial variability of dry season transpiration flux ranging from 0 to 0.15mm/d and reflected large tree biodiversity on Kalahari.

1 INTRODUCTION The presence of green, well adapted to dry season water stress condition vegetation on Kalahari, the recent information about extremely deep tree rooting depths on Kalahari (Canadel et al. 1996, La Maitre et al. 2000) and the particular importance of transpiration in environments with extremely low recharge (Lubczynski and Obakeng 2004, in the same issue), initiated the idea of quantifying transpiration on Kalahari (Lubczynski 2000). This idea has been materialized in the framework of the Botswana Kalahari

Water resources of arid areas

694

Monitoring project (Lubczynski and Obakeng 2004—in the same conference issue), in which this transpiration study was realized addressing the research question on what is the spatial variability of transpiration in semi-arid and bio-diverse Kalahari environment. The answer to this question is expected to support groundwater modeling and management’s problem of spatial variability of groundwater evapotranspiration (Lubczynski 2000). The first contribution to spatial assessment of groundwater evapotranspiration in the study area was made by Timmermans and Meijerink (2000), who used remote sensing solution of energy balance to derive actual evapotranspiration. The application of this technique, particularly with regard to dry season evapotranspiration fluxes, indicated that the method, which proved to be suitable in the irrigated areas, does not provide sufficient accuracy to deal with very low fluxes in dry semi-arid conditions such as Kalahari, substantially overestimating them. Considering the important role of transpiration in groundwater evapotranspiration (Lubczynski and Gurwin 2004), the next attempt discussed in this study, was focused on spatial transpiration assessment through remote sensing upscaling of field sap flow measurements. The methodology applied consists of two steps, first related to field sap flow data acquisition and the second to RS upscaling of field measurements. This paper presents mainly the first step, discussing methodology of the Kalahari sap flow measurements and the results of dry season plot transpiration upscaling in the selected, representative, 10×10km study area (see Fig.1 in Lubczynski and Obakeng, 2004, in the same conference issue). In the second assessment step, which is still in the development stage and therefore is not discussed here, high-resolution IKONOS image (Keelesang 2004) and aircraft based, multiband camera image are used for transpiration upscaling (Hussin et al. 2004a, b). 2 METHODOLOGY OF SAP FLOW MEASUREMENTS Tree transpiration can be well evaluated by direct field measurements of sap flow. Sap flow (Qs) is a product of sap velocity (ν) and sap wood (xylem) area (Ax). Unfortunately so far there is no method being able to directly integrate Qs, therefore Qs is typically defined by separate measurements of ν and Ax. So far, there were in the study area three dry season sap flow measurement campaigns (mostly in Septembers): in 2001 when tree species such as Acacia fleckii, Boscia albitrunca and Lonchocarpus nelsii were measured (Fregoso 2002), in 2002 when Terminalia Sericea, Burkea Africana, Acacia erioloba, Ochna pulchra, Dischrostachys cineria and Acacia luederitzii were measured (Mapanda 2003) and in 2003 when Ziziphus macronata and Acacia karoo were measured (Keeletsang 2004). The eleven above mentioned species cover the tree variety in the study area. 2.1 Sap velocity measurements For sap velocity measurements the Granier’s (Granier 1987) Thermal Dissipation Probes (TDP) manufactured by UP were used. The selection of the monitoring sites was purposive focusing on locations with many trees of the same species but different biometric characteristics, in access of

Dry season Kalahari sap flow measurements for tree transpiration mapping

695

Figure 1. Acacia Erioloba with sap velocity sensor probe inserted. less than 100m distance constrained by the unwanted voltage drop on the cables powering the sensors. Non-healthy trees, trees with irregular stems and trees used as birds’ habitats were excluded from investigation. At each selected site the sap velocity measurements were carried out by simultaneous use of 18, 24 or 30 TDPs’ connected to four DataHog2 Skye loggers. At each of the investigated sites, measurements were carried out for at least 3 days and after that the monitoring system was moved again to another purposively selected site. In total 198 trees were measured this way, some of them by more than one TDP probe. In each selected for sap velocity measurement tree, TDP sensors were always installed on the southern side of the stem, at the height of 0.5m above ground level to avoid thermal effects of sunshine. At each selected tree, first loose bark, small branches and sprouting leaves were removed from the stem. Next, two, 2cm deep holes, 10cm apart one above the other, were drilled in the sap wood and two, 2cm aluminum tubes (provided by UP supplier) were pushed into those holes with the special insertation tool into the structure of the sapwood (xylem). Finally the two TDP probes covered with a thin layer of silicon-grease (to improve thermal contact and prevent moisture entry into the space between sensors and aluminum tubes) were inserted into the aluminum tubes (Fig. 1). The TDP method is generally sensitive to the influence of external atmospheric conditions, therefore radiation-protection-shields and terostat paste were applied to isolate the measuring probes from the impact of the external conditions. The UP sensors are manufactured in systems of three TDP measuring units connected in series. One system of 3 measuring units consumes 0.125 A, therefore the use of up to 10 of such systems in this study required reliable powering system. For that purpose, a solar system consisting of 75W solar panel, 130 Ah sealed lead-acid battery and solar

Water resources of arid areas

696

charge regulator were used. All the measurements, were sampled every 30 seconds and stored every 30 minutes by four synchronized DataHog2 loggers. 2.2 Determination of conductive xylem area In order to derive thickness of the sapwood area, first the increment borer was used. This method however turned to be not applicable for hard wood of Kalahari trees so it was abandoned. Finally, tree xylem structure, was assessed by tree cutting and dye tracing method. In line with this method, trees were cut and put in a bucket with eosin B dye solution for about 3 hours (Fig. 2). After that time, cross sectional stem disks were cut at the level where sap velocity measurements were made and the obtained discs were analyzed using magnifier lens for determination of the conductive xylem area. For each tree also the complete estimates of biometric variables such as two perpendicular stem and canopy diameters were measured by caliper and calibrated tape respectively (Fig. 3).

Figure 2. Tree trunk inside a bucket with eosin solution.

Dry season Kalahari sap flow measurements for tree transpiration mapping

697

Figure 3. Xylem area, Acacia luederitzii. The obtained xylem area measurements were regressed against crown area and the stem area. The resulting regression equations were used to estimate xylem area of individual trees in plot and map transpiration upscaling. 2.3 Sap flow Sap flow (Qs) of individual tree species, normally expressed in l/d can be calculated as: Qs=v×Ax (1) where Ax is the sapwood (xylem) area and ν the sap velocity. According to Granier (1987), ν, which is typically expressed in cm/h, can be estimated from the continuously measured temperature difference (∆T—higher in the night and lower in the day when heat dissipating sap flow occurs) between the upper heated and the lower non-heated TDP probes. In calculation of ν the reference is made (Equation 2) to ∆Tmax, which is the maximum night temperature difference between the two probes when stomata’s are closed so no sap flow occurs. (2) The normalized sap flow QN of individual trees, typically expressed in l/d/m2, can be estimated as: QN=Qs/Ac (3)

Water resources of arid areas

698

where Ac is the projected ground area of the tree crown. 2.4 Systematic plot sampling and plot upscaling In order to analyze spatial distribution of tree transpiration in the study area and provide verification baseline for remote sensing upscaling, 78 systematic, quadratic plots of 35×35m, each at the spacing of 1km were located and sampled by Mapanda (2003). In these plots, a total of 1385 trees were inventoried with regard to identifying tree species, measuring their geographical locations using a Garmin 12XL Geographical Positioning System (GPS) and measuring standard biometric characteristics. Plot transpiration, Tp typically expressed in mm/day, can be estimated as: (4) where ΣQS is the total sap flow of all the trees present in the investigated plot characterized by area A, in this study case 35×35m=1225m2. 3 PLOT TRANSPIRATION In assessment of ΣQs per plot, species-specific correlations of xylem area vs stem area and xylem area vs crown area, combined with the mean sap velocities obtained per species were used and provided spatial, plot transpiration variability (Figure 4). The analyze of dry season transpiration fluxes in 35×35m plots indicates large variability of fluxes from nearly 0 to 0.15mm/d (Fig. 4). It can be noticed that the higher and the lower transpiration flux values are clustered which is mainly resulted by variable density of vegetation in the area of concern. 4 CONCLUSIONS The sap flow measurement is a useful but tedious and quite expensive method for transpiration assessment. However once in the certain area such as Kalahari the background research defining

Dry season Kalahari sap flow measurements for tree transpiration mapping

699

Figure 4. Plot transpiration in mm/day (Mapanda 2003). species-specific mean velocities and species-specific biometric relations (between canopy area and xylem area and between stem area and xylem) are developed, then such research can further be used in cost effective, automated way of transpiration mapping. The mean dry season sap velocities vary among the thirteen investigated Kalahari species from the lowest ~0.6cm/h as for the Acacia laudertici to the highest ~3.8cm/h for Boscia albitrunca. The measured velocities indicated no correlation with the biometric characteristics of the trees. The xylem area is well correlated with the stem area and with the canopy area. The correlation with the stem area is generally better than with the canopy area. For most of the tree species sap flow depends linearly upon the canopy area. The normalized sap flow (QN), which is the sap flow per unit area taken usually as per 1m2 of the canopy, provides species-specific and independent from biometric characteristics measure of transpiration. The lowest mean dry season QN in the study area was defined for Dichrostachys cineria ~0.061/d/m2 and the highest for Boscia albitrunca ~1.551/d/m2. Transpiration flux depends not only on what sort of tree species are present in the analyzed area but also on the density of the trees in that area. Such density has a critical importance with regard to transpiration upscaling. The assessment of dry season transpiration in 35×35m plots over the systematic 1×1km grid of 10×10km study area indicated substantial spatial variability of transpiration varying from 0 to 0.15mm/d. This variability was resulted by relatively

Water resources of arid areas

700

small size of the plot assessment, very sensitive to the local anomalies in the tree density and species type and also to the generally large biodiversity of the Kalahari vegetation. The large biodiversity in the Kalahari study area makes the RS upscaling of sap flow measurements particularly challenging, mainly due to the difficulty in RS classification of the individual tree species. ACKNOWLEDGMENTS We acknowledge Geological Survey of Botswana for financial support and extensive help in sap flow field campaigns. In particular we would like to thank Mr. Phofuetsile for his support to the project and Mr. Ramatsoko and his field crew for the extensive professional and logistical help in the field. REFERENCES Canadell, J., Jackson, R.B., Ehleringer, J.R., Mooney, H.A., Sala, O.E.T. & Schultze, ET.D. 1996. Maximum rooting depth of vegetation types at global scale. Oecologia 108:583–595. Fregoso, A. 2002. Dry-season transpiration of savanna vegetation. Assessment of tree transpiration and its spatial distribution in Serowe, Botswana. MSc thesis, Library of ITC— International Institute for Geoinformation Science and Earth Observation, Enschede, The Netherlands. Granier, A. 1987. Evaluation of transpiration in Douglas-fir stand by means of sap flow measurements. Tree Physiology, 3:309–320. Hussin, Y.A., Chavarro, D.C., Lubczynski, M.W. & Obakeng, O. 2004a. Mapping vegetation for up-scaling evapo-transpiration using high-resolution optical satellite and aircraft images in Serowe, Botswana. Proc. WRASRA conf. Gaborone 3–7 August 2004, Rotterdam, Balkema. Hussin, Y.A., Lubczynski, M.W., Obakeng, O & Chavarro, D.C. 2004b. Designing and implementing an aircraft survey mission using high-resolution digital multi-spectral camera for vegetation mapping for up-scaling evapo-transpiration of Serowe, Botswana. Proc. WRASRA conf. Gaborone 3–7 August 2004, Rotterdam, Balkema. Keeletsang, M. 2004. Assessment of dry season transpiration using IKONOS images, Serowe case study, Botswana. MSc thesis, Library of ITC—International Institute for Geoinformation Science and Earth Observation, Enschede, The Netherlands. Le Maitre, D.C., Scott, D.F. & Colvin, C., 2000. Information on interactions between groundwater and vegetation relevant to South African conditions: A review. In: Past Achievements and Future Challenges. ISBN 9058091597, Rotterdam, Balkema: 959–961. Lubczynski, M.W. 2000. Ground water evapotranspiration—underestimated component of groundwater balance in a semi-arid environment—Serowe case Botswana. In: Past Achievements and Future Challenges, ISBN 9058091597, Rotterdam, Balkema: 199–204. Lubczynski, M.W. & Gurwin, J. 2004. Integration of various data sources for transient groundwater modeling - Sardon study case, Spain. Journal of Hydrology—in revision. Lubczynski, M.W. & Obakeng, O.T. 2004. Monitoring and modeling of fluxes on Kalahari—setup and strategy of the Kalahari Monitoring project. Serowe study case Botswana. Mapanda, W. 2003. Scalling-up tree transpiration of eastern Kalahari sandveld of Botswana using remote sensing and geographical information system.

Dry season Kalahari sap flow measurements for tree transpiration mapping

701

Timmermans, W. & Meijerink, A. 2000. Remotely sensed actual evapotranspiration; implications for groundwater management in Botswana. In: JAG—International Journal of Applied Earth Observation and Geoinformation, 1(1999)3/4:222–233.

Heavy metals and radioactivity in the groundwater of Khartoum State, Sudan Abdelatif Mokhtar Ahmed College of Water Science and Technology, Sudan University of Science and Technology, Khartoum North Sudan Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Groundwater samples were collected from 28 producing wells in the area east of the River Nile and the Blue Nile, Khartoum State, Sudan. A Detailed study of chemical and physical analyses was conducted. This paper mainly deals with the investigation of five heavy metals, namely Pb, Zn, Ni, Cu and Mn, beside investigation of the groundwater radiochemistry of the Nubian aquifer. The chemical analysis shtowed that the groundwater of the studied area and hence of the Nubian Aquifer is free from harmful heavy metals which were below the permissible standard of the WHO (1984). The study also indicated that the aquifer is free from any radioactive pollution.

1 INTRODUCTION People of Khartoum State depend in their drinking water on the fresh surface water of the River Nile and its main tributaries (the Blue Nile and the White Nile) which meet at Mugran area, Khartoum State). The surface water is being treated and distributed to the three main cities comprising the State through the pipelines, although there are many producing ground water wells inside the three main cities. With the tremendous increase in population, and growing demand for groundwater quantity the groundwater quality is now should be taken into account. There are many industrial areas now lacated within the living area which expected to be a source of pollution to the groundwater, this beside the daily municipal waste generated. Agricultural purposes since there are vast areas around Khartoum, as in the case of all Sudan, suitable for agricultural practices. Both governmental and private sector efforts in the field of groundwater in Khartoum area mainly focus on searching, exploration and exploitation of the groundwater. Information with respect to groundwater quality, especially heavy metals, is generally lacking. This paper is an attempt to throw light on certain heavy metals content in the groundwater of Eastern Khartoum State, which will give an idea about heavy metals content of the Nubian Aquifer which extend even to the north in Egypt and Libya comprising the Great Nubian Basin of North Africa.

Heavy metals and radioactivity in the groundwater of Khartoum State, Sudan

703

2 LOCATION AND GEOLOGY The study area entirely lies in the eastern part of Khartoum State, mainly occupies vast areas of Eastern Nile Province. The area is bounded between latitudes 15.42° and 15.78°N and longitudes 32.45° and 32.93°E, it is bounded by the River Nile and the Blue Nile from its western side, and is about 400 square kilometers in area (Fig.1). The area lies within the arid zone with an average annual rainfall of about 167mm during the summer season (July–September), and evaporation is about 10mm/day (Haggaz and Khairalla,

Figure 1. Showing locations of the boreholes in the study area. 1988). It has a hot summer (April-October) with little rainfall and cold dry winter. The mean annual daily maximum air temperature is 37.1°C with extremes of 47.2°C. The lowest temperature is 6°C recorded in 24th December 1951. The area is densely populated and people earn their living by working in towns. The area is flat with some scattered sandunes drained by small wadies and crossed from the eastern side by Khor Soba draining towards the Blue Nile. The area is entirely confined to the Nubian Sandstone Formation which was suggested of upper cretaceous age (Whiteman, 1971). The Nubian Sandstone Formation here belong to the Great Nubian Aquifer of North Africa which mainly lies in north Sudan and extends to the north into Egypt and Libya. The oldest rocks in the study area belong to the Basement Complex System (Whiteman, 1971). They consist of granites, gneisses and schists which crop at the surface outside the

Water resources of arid areas

704

area, mainly to the north and the east. The Basement Complex Rocks are succeeded unconformably by the Nubian Sandstone, which consists mainly of flat-lying or gently dipping rocks made up of continental sediments which include sandstones, grits, mudstones, extra-formational and intra-formational conglomerates (Khairalla, 1966; Whiteman, 1971). The Nubian formation are overlained by alluvial deposits of the Blue Nile and the main Nile. The alluvial deposits consist of ill-sorted clays and silts with sandy and gravelly lenses and believed to be quaternary in age (Whiteman, 1971). In some places the alluvial deposits and the Nubian Formation are covered by wind-blown sands and the recent Nile silts. The groundwater occurs mainly in the Nubian Sandstone Formation and the alluvial deposits of the Niles, and the aquifers of the Nubian Sandstone and the alluvial deposits are believed to be hydraulically interconnected (Khairalla, 1966). Two aquifers had been recognized in the Nubian Formation, an upper aquifer of variable thickness (10–300m) and lower one of more than 400m with higher values of transmissibility and permeability (Bureau of Geological Research, 1979). The depth to the

Table 1. Showing trace elements analyses in the study area. Well Locality no.

Pb Zn Ni Cu Mn (ppm) (ppm) (ppm) (ppm) (ppm)

1 2 3 4 5 6 7 8 9

ND ND ND ND – ND 0.02 0.02 ND

0.01 0.01 0.01 0.01 – 0.01 0.02 0.01 0.01

0.01 0.02 ND ND – ND 0.01 ND ND

0.03 0.01 0.01 0.01 – ND ND 0.02 0.02

ND 0.01 ND ND – ND 0.63 0.01 ND

ND 0.02 ND ND ND ND 0.03 0.03 0.03 0.01 ND ND ND ND ND

0.01 ND 0.01 0.02 0.02 0.01 0.01 0.01 0.01 ND ND ND ND ND ND

0.01 0.01 0.04 0.03 0.02 0.01 0.02 0.07 0.03 0.03 0.03 0.05 0.02 0.02 ND

ND 0.01 ND ND ND 0.01 0.01 0.01 ND ND 0.01 0.01 0.01 0.01 ND

0.01 ND ND ND ND ND ND ND ND ND 0.01 ND ND 0.05 ND

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Um Dureiwa Daroshab South Daroshab North Samrab East Halfaya Um Dawan Ban Abu Groon Eid Sharoom Sheikh Mustafa El Fadni Hillat Kuku Direisab Idd Um Dom Es Sutra El Hur Um Usheira Yafa Idd Babikir North Galaat Dugo Sheikh A.Rahman Gabarona Hattab EL Shafeiab Allogab Kadaro

Heavy metals and radioactivity in the groundwater of Khartoum State, Sudan

25 26

Umel Qura Kadaro Military Camp 27 Sh. Elamin—Bala 28 Saeed Factory ND: not detected.

ND ND

0.01 0.02

0.02 ND

0.01 0.01

ND 0.01

0.01 –

ND –

ND –

0.01 –

0.01 –

705

saturated zone is variable ranging from 5m near the rivers up to 10m at a distance of about 2.3km east of the Blue Nile. 3 MATERIALS AND METHODS Groundwater samples were collected from 28 producing wells located in the study area (Fig. 1) for analyses. The water samples were collected in clean 1 liter polyethylene plastic bottles and stored in a cooler for 24 hours. Electrical conductivity and pH were determined in the field. Analysis of heavy metals was carried out using a Perkin Elmer Atomic Absorption Spectrophotometer (model 1100). Radiochemical analysis was carried out in the Sudan Atomic Energy Commission Laboratories, and the following procedure was applied to determine radioactivities in the samples: About one litre of the sample is weighted in special container made of plastic, and then loaded on the system. The low level counting of the radiation is performed using gamma spectroscopic system. A highly pure germanium detector (HPGe) is used. It is connected to a pre-amplifier and an amplifier. The multichannel analyzer (MCA) as a new version of personal computer analyzer (PCA) software (Gamma Data Reduction ‘GDR’) are used for the analyses and peak identification. The time used for collection is between 4000 to 10000 seconds. The system is calibrated using standard source of known energies and activities. 4 RESULTS AND DISCUSSION The results of heavy metals which are shown on Table 1, indicated that the concentrations of these ions are very low and in most cases even not detected. This indicated that the groundwater aquifer is free from harmful elements and free from heavy metals pollution. Heavy metals values are far below the permissible ones set by the WHO (1984) guidelines for drinking water. Since the groundwater of the study area belonging to the Nubian aquifer, which is part of the Great Nubian Aquifer of north Africa, it can be suggested that the groundwater of this aquifer is free from trace elements unless there are certain local reasons.

Water resources of arid areas

706

5 RADIOACTIVITY The most abundant radioactive element which could be found in water is 226R isotope (IAEA, 1990). The reason for that is relative high solubility when compared to the other radionuclei. Although it dissociate to give 226Ru isotope, but 222Ru is a gas and could not be found in water specially when it is exposed to air. From results of all samples the concentration of 226R is nil which indicate that the whole study area is free from any source of radioactive isotope. The presence of 226Ru and K40 which are showed in each spectrum are the normal radioactive isotopes which are present in the atmosphere and the values are even less than the standard lower limit of the natural radioactivity. Finally the water samples collected from the study area are free from radioactive pollution Similarly it can be suggested that the Nubian Aquifer is free from radioactivity. 6 CONCLUSION From the results of trace elements and radiochemical analyses it can be concluded that the groundwater is free from heavy metals and radioactive pollution. This phenomenon can be applied to the groundwater of the Nubian Aquifer covering fast areas in North Sudan and other areas of the country, and if there is any pollution within this aquifer can related to localized phenomena. REFERENCES Bureau of Geological Research 1979. Groundwater resources in Khartoum Province, Part II, Fed. Inst. Geosci., Nat. Resources, Hannover. Haggaz, Y.A.S. and Khairallah, M.K. 1988. Paleohydrology of the Nubian Aquifer North East of the Blue Nile, near Khartoum, Sudan, Jnl. of Hydrology, 99:117–125. Khairallah, M.K. 1966. A study of the Nubian Sandstone Formation of the Nile Valley between latitude 14°N and 17° 42′ N with reference to groundwater geology, MSc. Thesis, University of Khartoum. Whiteman, A.J. 1971. The Geology of the Sudan Republic, Clarendon, Oxford. WHO 1984. World Health Organization. Guidelines for drinking water standards, vol 1, Recommendations, Geneva; WHO.

Impediments to the effective implementation of a groundwater quality protection strategy in Botswana T.R.Chaoka,1 E.M.Shemang,1 B.F.Alemaw1 & O.Totolo2 1

Department of Geology, University of Botswana, Gaborone, Botswana 2 Faculty of Science, University of Botswana, Gaborone, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Protection of water resources from contamination is one of the key prerequisites for sustainable water resources management. In spite of its limited water resources, however, Botswana did not have a comprehensive groundwater quality protection program until 1992. This paper, therefore, examines the groundwater quality protection program that was developed since the early 1990s, its contribution in reducing pollution and addressing groundwater quality protection, and the challenges in implementing the strategy. An evaluation of the efforts made and the merits of setting the program and its effectiveness is undertaken by taking those areas with pollution sources that post date the establishment and implementation of the program. The existing status of groundwater quality in the various wellfields suggests that the implementation of the groundwater quality protection strategy is not as effective as was expected. The study concludes that the non-effective implementation of the groundwater quality protection strategies could be attributed to several factors: (1) Lack of complete information in the designation of protection zones of the aquifer systems, (2) lack of coordination of the various groundwater quality-related responsibilities assigned to different local and central government agencies, (3) lack of enforcement and (4) lack of public participation in the conceptualization and implementation of groundwater protection strategies. In order for an effective groundwater pollution protection strategy to be put in place in Botswana, the paper recommends plans to address the above impediments.

1 INTRODUCTION For a water scare nation like Botswana, protection of water resources from contamination is one of the key prerequisites for sustainable water resources management. Botswana,

Water resources of arid areas

708

however, did not have a comprehensive groundwater quality protection program/policy/strategy until 1992. The late development of a groundwater protection program was due in part to the lack of information on the risk posed to groundwater quality by various industrial, agricultural, and domestic activities. By the mid 1980’s, however, it had become very clear from the limited water quality monitoring data available that groundwater contamination was occurring in some areas of the country. More confirmation of the increased risk of groundwater contamination was provided by the results of three studies carried out between 1985 and 1991. The groundwater quality protection program that was developed following these findings employs two approaches: vulnerability assessment and wellhead protection. The first approach is aimed at protecting entire aquifers and involves the assessment of the vulnerability of groundwater resources to contamination on the basis of soil and geological conditions (aquifer type). Three vulnerability zones are defined; both soils and aquifers are ranked into three classes according to their vulnerability potential. The second approach is designed to protect individual wells and wellfields. It involves the delineation of a wellhead protection area, that is, the area that contributes water to a water supply well. The WHPA is divided into three zones depending (on types of contaminant under consideration and their transport behavior and possible fate, time of travel, persistence of contaminants in groundwater. Implementation of the Groundwater Quality Protection Program started in 1993 involving more than ten wellfields (and aquifers). Since then it has been extended to other wellfields and aquifers. In spite of this effort, available water quality data indicates that groundwater contamination is still taking place in some areas in which the Groundwater Quality Protection Program has been implemented. These areas can be divided into two groups: those areas in which pollution sources predate the establishment and implementation of the program and those in which the sources post date the establishment and implementation of the program. This paper is concerned with the second group. What is happening in these areas suggests that the groundwater quality protection strategy is not as effective as was originally expected. The objective of this paper is to examine the effectiveness (both potential and otherwise) of the groundwater quality protection program in relation to the following factors: coordination of the various groundwater quality-related responsibilities assigned to different local and central government agencies, enforcement of the program, and public awareness. 2 GROUNDWATER OCCURRENCE IN BOTSWANA Botswana has very limited surface water and groundwater resources. This is due to its arid climate, high rates of potential evapotranspiration (2000mm/a) and low rates of groundwater recharge. Average annual rainfall ranges from 250mm in the southwest to 650mm in the northwest. Most of the rainfall falls in summer from October to March and varies considerably from year to year as well as spatially. In addition to the low average annual rainfall, Botswana is prone to frequent and long periods of drought. Most of Botswana is covered by thick sandy soils. The thick sandy soils together with the high evapotranspiration rates and low topography severely limit the amount of runoff. It is estimated that the amount of surface runoff that originates in Botswana represents

Impediments to the effective implementation of a groundwater quality protection

709

about 1% of the total rainfall that reaches the land surface. Estimated recharge to aquifers over most of the country is in the order of 1mm per year (DWA, 1991b; Gabaake, 1998). Figure 1 shows the different types of aquifer formations found in Botswana and their areal distribution. The different types of aquifers and their properties are summarized as shown in Table 1. From Figure 1, it appears as if the whole of Botswana is covered with aquifers. However, the amount of extractable groundwater resources in most aquifers is very small due to their generally low transmissivities, storativities, and recharge rates. Some of the extractable groundwater resources are unsuitable for human and livestock consumption due to their high TDS, nitrate, iron, manganese, and fluoride content. Pollution of groundwater resources has also occurred in some parts of Botswana and is still occurring in others. For example, in the late 1970’s a number of production boreholes in the Mochudi Wellfield were abandoned due to nitrate and bacteriological contamination from anthropogenic sources (DWA, 2000). 3 GROUNDWATER QUALITY PROTECTION STRATEGY IN BOTSWANA The Government of Botswana initiated three studies that were conducted in 1985 and 1991 to assess the magnitude of and identify sources of water pollution in Botswana (DWA, 1991). The results of these studies indicated that there was a severe threat to groundwater pollution associated with the disposal of sanitary waste. Although the threat from mining, industrial, and agricultural activities appeared insignificant, the studies showed that it was growing. Legislation related to water pollution was also reviewed and was found inadequate and in need of revision. Following these studies, the Department of Water Affairs (DWA) initiated a project on protection zones and guidelines for major wellfields, aquifers and dams in Botswana that was concluded in 1993 (DWA, 1993a, b). The main objectives of the study were to identify existing

Water resources of arid areas

710

Figure 1. Main types of aquifers in Botswana (Source: DWA, 1991). and potential sources of water pollution; develop a water quality protection program for underground and surface waters; and draw up guidelines and regulations for the protection of wellfields and aquifers. A number of wellfields and several aquifers were included in the 1993 project. The guidelines for regulations for wellhead protection areas and vulnerability zones include specific recommendations on what types of activities should be prohibited or restricted in different protection zones. The Botswana Groundwater Quality Protection Strategy is designed to protect individual water supply wells and wellfields as well as entire aquifers from contamination. Protection of wells and wellfields involves the delineation of Wellhead Protection Areas (WHPAs), whereas the levels of protection accorded an entire aquifer depend on its vulnerability to contamination. A WHPA is defined as the area contributing

Impediments to the effective implementation of a groundwater quality protection

711

water to a production well. Implicit in this definition is the assumption that contaminants can only reach a production well if they are released within the area contributing water to the well. 3.1 Delineation of wellhead protection areas (zones) In Botswana, two methods are used to delineate WHP areas: (1) the time-of-travel (TOT) method and (2) the arbitrary or fixed radius approach (DWA, 1993a, b). In the first method, for example, a one-year TOT is defined as the maximum distance that a groundwater pollutant could travel

Table 1. Major aquifer formations in Botswana (after Khupe et al. 1997). Age

LithoLithology stratigraphic unit

Aquifer Regional area Transmissivity km2/Type m2/d

Cenozoic Alluvium in Gravel, sand P Sand Rivers of the E Botswana Okavango and Alluvium delta (Limpopo) Kalahari Beds Gravel, sand Sand, 40 P sandstone, calcrete, 201000 P silcrete, clay Mesozoic Karoo Supergroup Stormberg Basalt 6000 F Lebung Sandstone 138000 FP Ecca, sandstone, 94000 F-P Beaufort Arkose, siltstone, mudstone, F coal Dwyka Tillite, mudstone Proterozoic E and S Botswana Waterberg Conglomerate, 28000 F Supergroup siltstone, quartzite, 21000 F

400–1500

350 0.7–35

1.5–2.5 1–40 0.9–70

0.8–2

1–8

Water resources of arid areas

Waterberg Supergroup Transvaal Supergroup W, NW and N Botswana Olifantshoek Sequence

shale Quartzite, shale Dolomite

712

40000 F-K 2–5

400 F

0.2

Quartzite, chist, porphyry 33000 F 0.6–2 Damara Arkose, Sequence quartzite, schist, 1000 F-K mudstone, shale Dolomite Archean Ventersdorp Porphyry, 50 F 0.7 Supergroup felsite, greywacke Archean Granite, 61000 F 0.5–2 basement of gneiss, SE and E migmatite, Botswana Amphibolite F=fractured; K=karstified; P=porous; Successful borehole rate >0.01m3/h.

Table 2. Aquifer protection zones. Zone

Basis

Operational courtyard Zone I

Distance of 30 to 40m around the borehole 100 day travel time or minimum 100m radius 1000m fixed radius 100 year travel time

Zone II Zone III

towards a pumping well within one year (Bates and Evans, 1996). The TOT is estimated with the aid of numerical models. In the fixed radius method protection zones are demarcated by circles of fixed radii from the production borehole. Delineation of TOT protection zones requires reliable estimates of hydraulic properties of aquifers and wellfields, which are rarely known with any certainty. The shapes of TOT protection zones depend on hydraulic properties. The Wellhead Protection areas are divided into four zones based on groundwater travel time and distance from the production well (Table 2): an operational courtyard, inner (Zone I), intermediate (Zone II), and outer (Zone III). As Table 2 shows, WHP areas are rather restricted in

Impediments to the effective implementation of a groundwater quality protection

713

Figure 2. Combined wellfield protection zones and aquifer vulnerability maps (Mclaren et al., 1996). areal extent compared to aquifers. Unless the WHPA covers an entire aquifer, groundwater resources outside a WHPA are not protected. Aquifer vulnerability mapping is designed to protect such resources. Establishment of levels of protection for groundwater resources outside WHP areas depends on their vulnerability to contamination. 3.2 Groundwater quality vulnerability assessment In Botswana the assessment of the vulnerability of aquifers (groundwater resources) to contamination is based on the physical, biological, and chemical characteristics of soils and geologic formations overlying an aquifer as well as the thickness of the vadose zone and the type of contaminant under consideration (DWA, 1993). A three-class vulnerability rating scheme is used: soils with little or no ability to attenuate non-point source contaminants; soils with moderate ability to attenuate non-point source contaminants; and soils with a large capacity to attenuate contaminants. Similarly, a three-class vulnerability rating scheme has been developed on the basis of the type of geologic formation overlying an aquifer: highly permeable and densely fractured strata; less fractured geologic formations; and geologic formations with extremely low permeabilities or low groundwater potential. Two maps are generally required during the implementation of the groundwater quality protection program: vulnerability maps and maps showing aquifer protection zones. These two maps may be combined as shown in Figure 2.

Water resources of arid areas

714

4 IMPLEMENTATION OF THE GROUNDWATER QUALITY PROTECTION PROGRAM: MAJOR IMPEDIMENTS AND CONSTRAINTS The effectiveness of a groundwater quality program depends on many factors which include program implementation, coordination of the functions and responsibilities assigned to different agencies in relation to groundwater quality protection, groundwater quality monitoring, legal enforcement of the program, and public education (MacDonnell and Guy, 1991). The Botswana Groundwater Quality Protection Program can be analyzed in relation to these elements in order to evaluate its effectiveness both potential and otherwise. As pointed out earlier, the Botswana Groundwater Quality Protection Strategy employs two approaches: (1) delineation of wellhead protection areas (zones) and (2) groundwater quality vulnerability assessment. The Department of Water Affairs (DWA) is responsible for the delineation of wellhead protection areas, vulnerability assessment, and construction of aquifer vulnerability maps. To date maps of wellfield protection zones have been produced for many wellfields (DWA, 1993; 2000). In addition, the terms of reference for all new groundwater developmental projects commissioned by DWA include the delineation of wellhead protection zones as a requirement. While this is a significant effort on the part of the government, it must be recognized that wellhead protection areas can only be defined in areas where there are wells. Large areas of major aquifers in Botswana do not have boreholes. It is for such areas that groundwater pollution vulnerability maps and land use maps are essential. Unfortunately, however, vulnerability maps have been prepared for only small portions of major aquifers up to now. In the other instance, the Department of Town and Regional Planning (DTRP) is responsible for developing land use plans or development plans. Such plans are intended to provide a framework for, among others, environmental protection (including water quality protection) and monitoring. Like vulnerability maps, planning area development plans have only been prepared for a few settlements. The lack of vulnerability maps and development plans in many parts of the country leaves groundwater resources in those areas unprotected and vulnerable to contamination as a result of uncontrolled land use practices in those areas. A draft Water Act was prepared in 1991 as part of the National Water Master Plan Study (NWMP, volume 11). Subsequent to that, the Department of Water Affairs’ 1993 project on “Protection Zones and Guidelines for Major Wellfields, Aquifers and Dams in Botswana” produced guidelines for regulations for wellhead protection areas and groundwater quality vulnerability zones. Thirteen years after the National Water Master Plan Study was completed, enactment of the proposed water legislation is still pending. This means that the Groundwater Quality Protection Program cannot be legally enforced. As the preceding discussion shows, functions directly and indirectly related to groundwater quality protection are spread out among several agencies in Botswana. There is therefore a need to coordinate the functions and responsibilities assigned to different agencies in relation to groundwater quality protection. At present, there is no single agency with the overall responsibility for coordinating these activities. The

Impediments to the effective implementation of a groundwater quality protection

715

National Conservation Strategy Coordinating Agency (NCSA) seems to be the most logical choice for such a role. The NCSA is supposed to be responsible for coordinating the formulation, and implementation of environment friendly policies, legislation, programmes and projects. It is also responsible for promoting environmental education through public awareness programmes and other relevant means. It should also be responsible for the legal enforcement of the groundwater quality protection program. For instance, the NCSA should ensure that the groundwater quality vulnerability maps and wellhead protection zones are compatible with the land use plans produced by DTRP. Sadly, the NCSA does not have sufficient financial resources to conduct public awareness campaigns in relation to groundwater quality protection. If the NCSA wishes to launch a public education programme related to groundwater quality protection, it has to ask DWA for funding. This inhibiting and has the capacity to compromise the agency’s coordination and oversight role. Public participation in the conceptualization and implementation of groundwater protection strategies is another major impediment to the implementation of the successful implementation of groundwater protection programs. Public participation in natural resources management is not a new concept. It is a part of Botswana culture and it has also been recognized as very important in Agenda 21 publication of the Rio De Janeiro Earth Summit. Essentially the local communities as well as other interested and affected parties should be involved at the early stages of land use planning. They can then indicate areas of socio-cultural significance as well as traditional watering points that could be accommodated in the landuse plan. At the minimum, before landuse plans are finalized they need to be presented to the public as proposed decisions for their inputs. Intertwined with the participatory approach is the integrated resources management. In the context of this study it is identified that the integrated management and protection of natural resources would be a more proactive way of making decisions. Basically this would allow for aquifer-wide plan that shows agreed sensitive areas that would be protected by the community in order to ensure sustainable water supply. It is not evident that community participation was envisaged necessity in the formulation as well as implementation plan of the groundwater protection Strategies in Botswana. Another important aspect for the effective implementation of the strategy is the monitoring network of groundwater quality. Currently with the exception of private boreholes and a few wellfields, the Departments of Water Affairs (DWA) and Geological Survey (DGS) are responsible for groundwater quality monitoring in most of the wellfields in Botswana. However, groundwater quality monitoring on a regular (monthly) basis is only performed in production boreholes. This is grossly inadequate. In most cases, the data is insufficient to evaluate trends or changes in water quality and identify areas of special concern, where, for example, poor quality threatens existing uses. The 2000 Groundwater Monitoring Study (DWA, 2000) recommended that a programme to sample observation boreholes be implemented to expand present groundwater quality monitoring network.

Water resources of arid areas

716

5 CONCLUSIONS AND RECOMMENDATIONS The preceding situation analyses show that the Botswana groundwater quality protection program is similar to that of many countries, yet the implementation of the program is not properly coordinated in order for it to be effective. In order for an effective groundwater quality protection strategy to be put in place in Botswana, the paper recommends the following: ● Production of aquifer-wide vulnerability maps for all aquifers in Botswana:- These maps should be used as inputs in the preparation of land use maps. ● Establishment of a unit within the National Conservation Strategy Coordinating Agency to coordinate land use planning, groundwater development, and other related activities that may impact negatively on groundwater resources:- Some of the functions of this unit should include public education on water use, water conservation, and groundwater quality protection. The unit should also have the authority to enforce all water related legislations. ● Public participation in the conceptualization, preparation and implementation of landuse plans:- This will ensure that sensitive areas that may impact on water quality and groundwater resources are demarcated and protected by the community in order to ensure sustainable water supply. ● Expansion of the monitoring network of groundwater quality:- The monitoring of groundwater quality should be performed in all boreholes (both production and observation wells) for adequate planning, development and protection of the groundwater resources of Botswana. It is the opinion of the authors that the above recommendations if implemented, would immensely strengthen the groundwater quality and protection in Botswana. REFERENCES Bates, J.K. and Evans, J.E. (1996). Evaluation of Wellhead Protection Area Delineation Methods, Applied to the Municipal Well Field at Elmore, Ottawa County, Ohio. Ohio Journal of Science, 96 (1), 13–22. DWA (1985) Ramotswa Wellfield Pollution Study. DWA (1991a) Magnitude and Sources of Pollution in Botswana. DWA (1991b). Botswana National Water Master Plan Study. Final Report. Volume 5. DWA (1991c). Botswana National Water Master Plan Study. Final Report. Volume 11. pp. 55. DWA (2000) Groundwater Monitoring Study, volume 1. DWA (1993a) Protection Zones and Guidelines for Major Wellfields, Aquifers and Dams in Botswana, Volume 1. DWA, (1993b) Protection Zones and Guidelines for Major Wellfields, Aquifers and Dams in Botswana, Volume 2. Gabaake, G.G. (1998). International Conference on the Role of National Geological Survey in Sustainable Development. Abstract Volume, 15–19. MacDonnell, L.J. and Guy, D.J. (1991). Approaches to Groundwater Quality Protection in the Western United States. Water Resources Research, 27(3), 259–265.

Impediments to the effective implementation of a groundwater quality protection

717

Mclaren, D.A., Hazell, R.T. and Gyopari, M.C. (1996). Water Quality Protection Strategy in Botswana. Botswana Journal of Earth Sciences, Volume 3, pp. 25–28.

Spatial assessment of groundwater pollution vulnerability of the Kanye wellfield in SE Botswana B.F.Alemaw, E.M.Shemang & T.R.Chaoka Department of Geology, University of Botswana, Gaborone, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Botswana depends mainly on groundwater, a major source of water, for domestic, industrial and agricultural purposes. These groundwater resources are of limited quantity, mainly because of low recharge rates. Therefore there is a need to ensure that the groundwater resources are of drinking quality, by protecting them against pollution by delineating groundwater protection zones to protect the groundwater aquifers from pollution from sources such as landfills, pit latrines, industries, agricultural pesticides and waste. This study is an assessment and evaluation of vulnerability of groundwater to pollution in the Kanye well field in SE Botswana is undertaken using DRASTIC approach, and soil & geology-based vulnerability mapping (SGV method). In the first method the aquifer media properties from 82 boreholes are used to derive the DRASTIC parameters and in the later case the soils types and mapped geology in the vicinity of these boreholes are used to determine SGV classes. A GIS system was employed to generate pollution vulnerability maps using Theissen polygons approach. Both of the approaches yielded comparable results. Results indicate that 58% of the well field area is very high to highly vulnerable to pollution, 34% is moderately vulnerable to pollution, and only 9% has low vulnerability.

1 INTRODUCTION As a semi-arid country, Botswana has little recharge to groundwater resources. Studies have shown that groundwater replenishment rate is very low (Beekman et al., 1996). Groundwater even though limited is a major source of water for domestic, industrial and agricultural uses. Groundwater contamination is a major concern in Botswana. Two wellfields, the Ramotswa and the Mochudi wellfields in the southeastern part of Botswana are no longer used for drinking purposes as a result of bacterial and nitrate pollutions caused by poor disposal sanitary waste (Department of Water Affairs, 1993). Other wellfields are

Spatial assessment of groundwater pollution vulnerability

719

showing evidence of similar pollution. These include the Molepolole wellfield in the central part of the country and the Serowe and Palapye wellfields in the northeast. There is therefore the urgent need to ensure that the groundwater resources are of drinking quality, by protecting them against pollution sources such as landfills, pit latrines, industries, agricultural pesticides and waste. The vulnerability potential of an aquifer to pollution depends to a large extent on the susceptibility of its recharge area, as areas with high replenishment rate are potentially more vulnerable to pollution than others (Bekesi and McConchie, 2002). Unconfined aquifers that do not have a cover of impermeable material are also highly susceptible to contamination. Soils overlying the water table provide the primary protection against ground water pollution. The objective of the study is to investigate the potential to pollution vulnerability of Kanye wellfield based on two methods vis-à-vis soil and geology-based vulnerability mapping (SGV method) and the DRASTIC method. 2 THE STUDY AREA The present study was undertaken on the Kanye wellfield (located between longitudes 25°–25.4°E and latitudes 24.8°–25.5°S Fig. 1), south-eastern Botswana with the aim of assessing the vulnerability of this well field to pollution. The Kanye wellfield comprises three sub-areas where groundwater is abstracted by the Water Utilities Corporation: Northwest, Kgwakgwe and Rammonedi. The main potential source of pollution of this well field is agricultural activities (mainly livestock farming) and domestic waste disposal. The aquifers in the Kanye wellfield consist essentially of carbonate rocks (dolomites) that are highly fractured and fissured at some places. Karst phenomenon is also prevalent within these aquifers. 3 METHODOLOGY Two methods have been employed to assess the vulnerability to pollution of the kanye aquifer system. The first method is geology-based vulnerability mapping (SGV method). The DRASTIC method, which considers other factors in addition to soil and geology, was also used to compare the various vulnerability classes. 3.1 Soil and geology-based vulnerability mapping (SGV method) A detailed description of the first method, SGV method, can be found in previous studies reported in Alemaw et al. (2004). In order to assess the vulnerability of this aquifer to pollution, the types of

Water resources of arid areas

720

Figure 1. Kanye Wellfield Location Map. soils overlying the area were divided into 3 classes, based on their ability to permit flow of fluids. Class one included very coarse-grained soils (sands, gravels, dolomites and calcretes), chert breccias, class two included medium to fine-grained soils (sandy clays, clayey sands, weathered BIF (lateritic soils) and class three included very fine-grained soils (clays, shales and ferruginous clays). The geologic materials overlying the aquifer and their corresponding thicknesses were used in conjunction with the soils vulnerability maps to classify geologic vulnerability into 3 classes; class one, highly fractured rocks, dolomite with karsitic structures very coarse grained sandstone, conglomerates and gravelly formations; class two includes medium to fine grained sandstones, dolomites with less fractures/fissures, clayey sands and sandy clay formations. Class three, included clays, shales, dolerites, granites, fresh dolomite, cherts, quartzites. Two maps are produced, one based on soil vulnerability and the other on geological vulnerability (Figs. 2a, 2b). The soil and geological vulnerability maps are then overlain to produce a single vulnerability map with five possible vulnerability classes. Table 1 shows matrix of groundwater vulnerability class based on combination of soil and geologic vulnerability. The data were integrated in a GIS environment in order to spatially assess the soil, geological and groundwater vulnerability of the well field and the results are summarised in Figure 2 below. Previous studies by Bekesi & McConchie (2002) show that GIS was used in integrating spatial vulnerability data with point borehole log information. The

Spatial assessment of groundwater pollution vulnerability

721

borehole log information at 82 boreholes was used to retrieve and spatially portray the various vulnerability classes in the well field by means of Theissen polygons.

Figure 2a,b,c. Soil, geology and groundwater vulnerability maps (SGV method) (source: Alemaw et al., 2004). Table 1. Five aquifer vulnerability classes based on soil types and host geology (SGV method). Geologic vulnerability Soil G1— G2— G3— vulnerability High Medium Low S1—High S2—Medium S3—Low

Very high High High Moderate Moderate Low

Moderate Low Very low

4 THE DRASTIC APPROACH One of the most widely used groundwater vulnerability mapping method is the DRASTIC method. Developed by the United States Environmental Protection Agency (EPA) as a method for assessing groundwater pollution potential (Aller et al., 1987). The DRASTIC index, designated by Di is calculated as: Di=DrDw+RrRw+ArAw+SrSw+TrTw+IrIw+CrCw where r and w refer to the DRASTIC ratings and DRASTIC weightings assigned to each of the following hydrogeological settings: D=depth to water table R=aquifer recharge A=aquifer media S=soil media

Water resources of arid areas

722

Table 2. DRASTIC ratings and weights adopted for the various hydrogeologic conditions. Depth to Rating Recharge Rating Topography Rating Conductivity Rating GW(m) (mm) (Slope %) (m/d) 0–50 50–100 100–150 150–200 200–250

10 9 7 5 3

0–2 2–4 4–6 6–8 8–10

1 2 4 6 8

>250

>10

9

Drastic weight, Dw=5 Aquifer media

Drastic weight, Rw=4 Vadose zone material BIF gravel

Gravel, 10 coarse sand Dolomite 9 Dolomitic 8 chert

Sandstone, 7 silstone Quartzite 6 Fractured 5 shales BIF Cherts

4 3

Syenite, granite, rhyolite,

2

10 9 7 5 3 1

Drastic weight, Tw=2 Soil media

10

Chert 9 gravel Gravel, 8 sand, sandstone, sandy gravel Dolomite 7

Weather dolomites Sands

9

Calcretes

7

Chert, 6 dolomite Shale, 5 sandyclay, clay Quartzite 4 Sandy 3 chert

Chert 7 breccias Sandy clays 6

Calcrete

10

2

Dolerite, 1 granite

Drastic

8

Clayey sands 5 Weathered 4 BIF (lateritic soils) Clays 3 Shales

0–0.5 0.5–1.0 1.0–1.5 1.5–2.0 2.0–2.5

1 3 5 7 9

>2.5

10

Drastic weight, Cw=3

Gravels

dolorite Drastic

0–2 2–4 4–6 6–8 8–10 >10

2

Ferruginous Clays 1 Drastic

Spatial assessment of groundwater pollution vulnerability

weight, Aw=3

weight, Vw=5

723

weight, Sw=3

T=topography (slope) I=vadose zone C=hydraulic conductivity The DRASTIC score Di is a relative value, with no specific units. A speed sheet program was written to derive the DRASTIC indices for the various boreholes studied. The following table shows the DRASTIC ratings and weights adopted in the study. 5 RESULTS AND CONCLUSION The results indicate that about 47% of the Kanye wellfield area is overlain by soils with high vulnerability to pollution (class 1 soils) Fig. 2a; 44% of the area is overlain by soils of intermediate vulnerability (class 2 soils) and 9% of the area is overlain by soils of low vulnerability (class 3 soils). The geological vulnerability maps (Fig. 2b) show that 40% of the well field area is underlain by rocks with high permeability and significant fracturing thus of high vulnerability (class 1 rocks), 49% of the area is underlain by rocks of medium vulnerability (class 2 rocks) and 11% of the area is underlain by rocks of low vulnerability (class 3 rocks). Considering the overall study area bounded between 25–25.4°E and 22.8–23.5°S, the ground water vulnerability map (Fig. 2c) indicates that 22% of the well field area is very highly vulnerable to pollution, 35% of the area is highly vulnerable to pollution, 34% is moderately vulnerable to pollution, 4% has a low vulnerability and 5% of the wellfield area has a very low vulnerability. The above percentage areas portraying the above levels of vulnerability were investigated in terms of the DRASTIC coefficients. Drastic coefficients vary in the range of 110 to 190. The results of this study indicate that the Kanye wellfield may be highly vulnerable to pollution if not properly managed. However, one should be cautious in the interpretation of these results due to the sparse and uncertain nature of the basic data used for this interpretation, and the inherent variability of the aquifer parameters. The use of the polynomial estimator enabled consideration of simple uncertainty for the aquifer media characteristics modelling process, uncertainty being proportional to the size of the polygons. Estimates of the aquifer media characteristics in areas with only a few observations or few geological logs were considered to have higher uncertainty than areas with many observations. In general, there is a high uncertainty of the groundwater pollution vulnerability classes in the Kanye aquifer system. The Northwest and Rammonedi wellfields have high uncertainty in the classes, while Kgwakgwe wellfield portrays a low level of uncertainty. From the aquifer protection and management point of view, the question is how far one should establish a protection zone around the wellfield so as to protect the entire aquifer from pollution. This needs more study and field checking of the aquifer media characteristics maps. The results presented here are preliminary findings on the use of spatially distributed point data to assess the regional variation of aquifer media characteristics and vulnerability to groundwater pollution.

Water resources of arid areas

724

The present study shows how vulnerable the Kanye wellfield is to pollution. It therefore underscores the need enforcement of the recommendations by the Department of Water Affairs (1993) to establish groundwater protection zones over the most vulnerable aquifer. REFERENCES Aller, L.T., Bennett, J.H. and Hackett, G., 1987.“DRASTIC: a standardized system for Evaluating Groundwater pollution Potential using Hydrogeoloic setting”. US Environmental Protection Agency Report EPA/600/2–87/035, 622 pp. Beekman, H.E., Gieske, A. and Selaolo, E.T., 1996. GRES: Groundwater Recharge Studies in Botswana, 1987–1996. Botswana Journal of Earth Sciences 1:1–17. Bekesi, G. and McConchie, J., 2002. The use of aquifer media characteristics to model vulnerability to contamination, Manawatu Region, New Zealand. Hydrogeology Journal, 10:322–331. DWA, Department of Water Affairs, 1993. Protection Zones and Guidelines for Major Wellfield, Aquifers and Dams in Botswana. Report prepared by Water Surveys, Botswana. Alemaw, B.F., Shemang, E.M. and Chaoka, T.R., 2004. Assessment of Groundwater Pollution Vulnerability of the Kanye Wellfield, Southeastern Botswana—A GIS Approach, B. Journal of Physics and Chemistry of the Earth, Elsevier Science Publishers (accepted for publication).

The effect of socio-economic activities on watershed management: the case study of Gaborone Dam catchment in Botswana George S.Thabeng & Daniel B.Kemiso Department of Water Affairs, Gaborone Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Watershed management entails integrated resource use planning, conservation as well as institutions and organizations involved therein. Issues like environmental concerns, livestock production, fishery, piggery, chicken farming or mineral processing, social and cultural concerns, infrastructure planning as well as involvement of all stakeholders in decision making during planning and implementation stages are of great importance. The most important factors in the whole planning process and implementation of activities in the logistic integrated resources management in which community (owners) involvement and accountability of implementers at all stages from formulation to programmes to evaluation and monitoring are of great importance. This case study looks at one of the socio-economically important river basins in Botswana namely; Gaborone Dam catchment to illustrate watershed management approach. The Gaborone catchment area is very important to our City in terms of agriculture, fishing, poultry, piggery, flour milling, Wildlife management and habitat for important flora and fauna. It covers parts of four administrative regions namely; Kweneng, Ngwaketse, Lobatse and Balete Districts. Since the resource in this catchment transcend administrative and sectoral planning borders of one region, resource use conflicts have emerged over years in the use of land, water and other natural resources. This case study looks at the socio-economic activities in the Gaborone catchment and the emerging trends and conflicts as well as environmental degradation as a result of sectoral planning. Finally the case study analyses the efforts of the Government, interested parties and stakeholders to implement integrated natural resources management by involving local communities in the catchment as well as resolving emerging resources use conflicts.

Water resources of arid areas

726

1 INTRODUCTION Gaborone Dam was first built in 1963 to supply water to the urban centre of Gaborone when demand for water by the town exceeded the supply from the older Notwane Dam and from boreholes that had been drilled in wellfields in the vicinity of the City. As the City of Gaborone continued to grow and, with it, the demand for water by the City, the Gaborone Dam wall was raised by 8 metres to its present height in 1986. The main technical characteristics of the Dam are as follows: Catchment Area: CapaCity of Dam at f.s.l: Type of Embankment: Crest Length: Maximum Surface Area: Maximum Depth: Average Depth of Water at f.s.l: Potential Evaporation:

14,300km2 141.4×106m3 Rock Earthfill 3km 1900ha (19km2) 20 metres 7 metres

2000mm per year

The catchment area covers the southern portion of southeastern Botswana, and also falls within the neighbouring country of South Africa (Figure 1). It falls within 3 different districts and apart

The effect of socio-economic activities on watershed management

727

Figure 1. Location of Gaborone Dam catchment area. from the southern part of Gaborone City, covers two main primary settlements (urban areas), ie Lobatse in the south and Kanye in the west. The 3 districts are South East District covering approximately 45% of the study area, Southern District covering approximately 47% of the District, and Kweneng which accounts for about 1%, whilst the remaining 7% falls within South Africa. The study area (falling within Botswana) measures approximately 14,300km2 in extent. The southern and western parts of Gaborone are considered as falling within the catchment, and it is the activities in the Gaborone area, that is a major driving factor in the land use problems encountered in the catchment area. 2 THE PHYSICAL ENVIRONMENT 2.1 Physiography and landform The area is generally undulating with slopes tending generally SW-NE. A few Hills— notably Kgale Hill and the Polokwe Hill dominate the remnant of the African erosion surface. The Notwane River and its tributaries dominate the drainage of the area.

Water resources of arid areas

728

2.2 Climate The climate is generally hot with a slight drop in the temperature during the winter months of May to August. Mean temperatures range from 12°C (July) to 28°C (January). Considerable local variations occur so that the winter temperatures in Lobatse are lower than in Kanye or Mogobane. Winter frost is common in places like Lobatse, Kanye and surrounding areas. Annual average rainfall (1925 to 1997) is 509.6mm in Kanye, and 555mm in Lobatse. Maximum annual totals recorded in Kanye and Lobatse are 970.6mm and 1070.4mm compared to minimum rainfalls of 104.7 and 262.1 respectively. 2.3 Geology and Geomorphology 2.3.1 Geology The geology of the area is complex. The catchment overlies part of the Kaapvaal Craton, an ancient stable segment of acidic crustal rocks. These are Archaean metamorphics, profoundly altered by heat and pressure, as well as igneous rocks. These form the floor above which Proterozoic and younger non-metamorphic sedimentary and volcanic rocks occur. Of the supercrustal rocks, the Lobatse Volcanic Group is the oldest. They now form the arc of rugged relief that extends south of Gopane to Lobaste and northwards towards Mogobane before striking northeast to the north of Ramotswa station. The next, younger rocks belong to the Transvaal Supergroup, one of the major early Proterozoic sedimentary successions, some 5000m thick. The youngest of the supercrustal rock sequences is the Waterberg SuperGroup, the orthoquartite member (Mannyelanong Formation, 17,000–22,000 million years) being responsible for the flat cappings on the interfluves of the project area. Much of the catchment, including the Gaborone Dam, is underlain by the Gaborone Granite (1600–4000m.y). There are also other intrusive rocks in the area, which may be unrelated to the Gaborone granite. These are the Kgoro Complex and Mmathethe Granite (1600– 4000m.y), also dolerite sills and sheets (e.g intruding the granite at Kgale Hill) of late Proterozoic age. Younger Palaezoic and Mesozoic rocks are completely absent from this area. Minor calcretes occur in the soils above the volcanics of the Transvaal Supergroup and minor pedogenic laterites occur in flat locations over the granites. The tectonic history of these ancient rocks is complex and many of the streams of the catchment are strongly controlled by the structures. Unlike in other areas of Transvaal Supergroup rocks, the economic geology of this area has proved very disappointing, despite substantial exploration. The only metal, which has been mined in area, is manganese at Ramotswa, Otse and Kgwakgwa at Kanye. Occurrences are, however limited in the area and thickness. All mining has now ceased, that at Ramotswa in 1958 and at Otse in 1996. There are no foreseeable prospects of pollution of the Gaborone Dam water by run off from metalworking. As concerns

The effect of socio-economic activities on watershed management

729

industrial rocks and minerals, clays, river sands and gravels have been and still are worked for brick making and aggregate. 2.4 Geomorphology 2.4.1 Weathering The catchment comprises surfaces, plains and bottomlands, of predominantly PostAfrican Surface age. The African erosion Surface was the first to form when Gondwanaland split up into the continents of Africa, South America, India, Australia and Antarctica. It is known to have been deeply weathered and weathering was advanced in the sense that primary minerals were converted first to 2:1 clays and in the upper parts of the profile where weathering was more aggressive the 2:1 clays were converted to 1:1 clays, predominantly kaolinite. This weathering involved severe leaching, so that virtually all of the bases, essential nutrients, were removed. Very small areas of African erosion surface survive on the interfluves as small mesas or plateaux, particularly where Waterberg quartzites form cappings. The deep weathering profiles have been stripped off, leaving an interestingly etched rock surface. A good example is the plateau on which Kanye sits. The plateau is deeply penetrated by fossil gorges, terminating in huge dead waterfalls, which clearly attest to formerly much wetter conditions. The fasts of such plateaux, only penetrated with difficulty, provide sufficient isolation for the survival of kudu and leopard. The sharp drop to the Post-African Surface plains, provides the perimeters of such plateaux with many excellent, but rarely visited, view points offering wide vistas over the inselberg-dotted lowlands. Continental uplift and erosion resulted in the formation of the Post-African Surface, which dominates the project area. Weathering is much less advanced, although deep in places. Clay formation is poor. Typically, weathering has only succeeded in separating the individual mineral components of the rock, leaving them in situ as a disaggregated mass. Such poor weathering yields saprolite (weathered rock) with very little cohesion. It is very susceptible to erosion. The poor development of clays results in poor, thin sandy soils, particularly where the rocks are acidic, e.g. in granitic areas. Here the soils are sandy, dominated by quartz and orthoclase feldspar particles, the most resistant to weathering. The small, isolated Hills, which dot the plains are the exposed parts of the basal surface of weathering (the boundary between weathered and fresh rock). The relative relief on the basal surface of weathering is high in granitic areas; the boundary plunges up and down very abruptly, controlled by the frequency of fracturing. Small outcrops form koppies, larger outcrops inselbergs. These shed rainwater so that weathering is facilitated around their bases. 2.5 Soils Soils in the catchment area are generally of the sandy loams, clays and sandy clay loam types. Good fertile soils are found at depositional areas such flood plains and depressions. In the Gaborone area, soils on alluvial deposits are developed in Notwane and

Water resources of arid areas

730

Segoditshane flood plains where ephemeral rivers bring sand from upstream. These soils are suitable for trees. 2.6 Hydrology and water resources The area is drained by quite a number of ephemeral streams. At the western part of the catchment area are river tributaries such as Gamoralo, Matlhapise, Mokape, and Masinyetse. To the southern part are Gamoswana, Molapowabojang, and Lobatse (from South Africa). The central catchment area consists of Makwena, Molopye, Chawe, and Potsane Rivers. The northern part is the Kgamagadi River. The main rivers in the area that contributes to the Dam reservoir are the Notwane, Nywane, Maratadibe, Fikeng, Metsemotlhaba, Taung, Nywane and Peleng. The hydrological characteristics of the rivers are said to be changing. The river system has become depositional rather than being erosive thereby causing floods. This is noticeable particularly of the Metsemotlhaba, Notwane and Taung rivers, which are prone to flash flooding during rainy season thereby causing Damage to properties and severing of communication linkages between Ramotswa and Tlokweng and the rest of the country. 2.6.1 Wetlands There is a large wetland in the valley of the Taung river, near the village of Mogobane. This wetland is at least 750ha in extent and it currently supports dense wetland vegetation including fragmitis. As will be discussed later, this wetland presents an opportunity for pollution control for Gaborone Dam. All water from the western part of the catchment, i.e. the areas of Khanye, Ranaka, Magolhwane, Ntlhantlhe, and Lotlhakane; will pass through this wetland before entering Gaborone Dam. If properly maintained, the Mogobane wetland will filter and improve the chemical quality all the water from the Taung river and its tributaries before it enters Gaborone Dam. 2.6.2 Aquifers/wellfields The main aquifers in southeastern Botswana are well represented in the project area in the form of two major aquifers, both of which have been targets for exploration. These are the dolomitic aquifers in Ramotswa-Lobatse and Kanye. 2.7 Ecology 2.7.1 Vegetation The Gaborone Dam catchment area comprises a diverse array of vegetation types. Extensive areas of the catchment comprise typically flat sandveld savanna which has been heavily cultivated and is covered by a high proportion of cleared fields, actively growing crops and abandoned fields. The dominant vegetation types in this area are

The effect of socio-economic activities on watershed management

731

acacia dominated bush savannas and tree savannas which range in density from low density areas close to settlements to high density areas in less settled and cultivated areas. Rocky Hills and kopjes occur throughout the catchment area and are generally covered by medium density woodland and bush which shows a high degree of variation in species composition depending on local soil conditions. The rocky outcrops and surrounding Hilly areas are not cultivated and have few settlements, although the runoff from their slopes, following intense rainstorms, contributes significantly to soil erosion on the heavily grazed and cultivated lowlands. Apart from the rocky Hillslopes, considerable concern also surrounds the integrity of the riparian fringes, which experience heavy grazing by domestic stock. Vegetation along the river channels such as Notwane in Gaborone and Peleng in Lobatse exhibit a different shrub and tree cover, which is different from the accustomed savanna type. In view of its uniqueness, the Department of Museum and National Archives has planned to preserve a 2km length and 100m-wide corridor along the Notwane River. 2.7.2 Fauna (Wildlife) Wildlife in the catchment area is low when compared with the northern part of the country. Despite this, quite a number is found in the area. Some little wildlife is found at the Kgale Hill in Gaborone, 300 bird species and a vulture colony called Manyelanong near Otse Hills are also in existence. The area also contains 2 private owned natural reserves; St. Clairs Lion Park and Mokolodi Nature Reserve. The Gaborone Dam and the Sebele sewage ponds are home to over 40 species of water birds. The Gaborone Game Reserve has a large number of antelopes, warthogs ostriches and rhinos. 2.7.3 Crocodiles The occurrence of crocodiles in the Gaborone and Notwane Dams also deserves explicit attention. The Nile Crocodile has in general shown a drastic and well documented decline worldwide, and while not yet endangered, has slipped to a “vulnerable status” (IUCN, 1982). Botswana’s population is concentrated in the Delta, with considerable debate surrounding the size of the pre commercial harvesting (i.e. pre- 1957) crocodile population in the Delta, with estimates varying from a population of 21,000 up to 66,000. 2.7.4 Fisheries Although the fisheries potential of the majority of Dams in eastern Botswana is largely unexploited, Gaborone Dam makes a significant contribution to both the diet and livelihoods of people. Barbus sp. tend to be an abundant and unexploited resource in most Dams throughout the country, probably because such time consuming activities as cleaning and post-harvest processing are essential if the fish is not to be spoilt (Nermark and Mmopelwa, 1994). Good-sized bream appear to be in demand from the larger towns, although local people often prefer catfish (C. gariepinus), a species which is not only drought resistant, but can greatly increase the fish biomass of reservoirs (Nermark and Mmopelwa, 1994).

Water resources of arid areas

732

3 THE PROBLEM The City of Gaborone has been expanding at a phenomenal rate throughout the past decade and it is still continuing to grow. In a period of 30 years, the population of Gaborone has grown by more than 10 times (Table 1). This growth is the single most important factor underlying the phenomenal growth of population in the catchment area as a whole. The population of the catchment area has been growing over the years. In 1971, the catchment area’s population was 133, 211; by 1981 it had increased to 237, 247 at a rate of 5.8 per cent. The 1991 population census indicates that the area’s

Table 1. Population statistics for Gaborone. 1971a 1981a Growth 1991b rate (1971– 1981)

Growth 2001c rate 1981– 1991)

17,718 59,657 12.1% 133,468 8.1%

Growth rate (1991– 2001)

186,007 3.3%

population increased further to 338,068 at a rate of 3.5 per cent. The recent population census held in 2001 indicates a growth of about 3.1 percent in the catchment area. The present population is therefore estimated at 458,370 (see Table 1). The growing urban areas such as Gaborone, Lobatse, Ramotswa, and Kanye, have been the main nuclei for the population pull and hence their growth. These centers have now become dormitory towns for Gaborone, with substantial numbers of people commuting from them to work in Gaborone daily. The highest growth rates between 1991 and 2001 are recorded for Otse, Gaborone, and Gabane ranging between 3.3 per cent and 5.5 per cent. This is as a result of some pull factors (migration) such as economic opportunities, social, educational and recreational activities, which have attracted people to these centres. The growth described above is driving a scramble for land in the catchment area. This in turn has resulted in land use and land development problems. The problem is exacerbated by the fact that land administration within the catchment area falls under the jurisdiction of many authorities, with very little coordination. As a result, the catchment area is experiencing water quality problems. A case in point is in Ramotswa, in the South East District, where although it has abundant ground water, it has been polluted by human waste. More importantly, the water in Gaborone Dam on which the City depends is also affected. The present situation regarding the chemical aspects of the water quality is that it is excellent, falling well within BOBS standards for Class 1 water (Ideal), although aesthetic parameters (odour and taste) have, on occasions been unacceptable. There is a long term trend towards increase in solutes, conductivity and pH, superimposed on seasonal patterns. The cause appears to be regional rather than local, but further research into this is needed. Of some concern is the tentative conclusion that normal rains do not result in very substantial washing of surface materials down into the Dam. These appear to accumulate until above average rainfall conditions are experienced. In effect the catchment stores surface materials, possibly for many years before they are delivered to the Dam. This enhances the already pronounced seasonal patterns of delivery. Data on

The effect of socio-economic activities on watershed management

733

the biochemisty and biology of the Dam water are, at this point, inadequately monitored and it is largely this aspect of water quality which could be of concern (Aqualogic, 2002) 4 THE CAUSES OF THE PROBLEM The causes of the problem should be discernible from the foregoing. Without doubt, the problem is the direct consequence of development in the catchment area without adequate controls. This is the sort of problem that watershed management approach to resource development and utilization could have helped to avoid. In the case of the Gaborone Dam catchment area, the problem can be attributed to the land use and land administration problems. Land ownership in the country is either freehold or tribal (customary law). A large percentage of the study area falls under tribal land, (ie, more than 70%) and is administered under the jurisdiction of several Land Boards (Balete, Bakwena, Ngwaketse, Malete). Usage rights can either be granted communally or to individuals. The majority of residential and agricultural sites are therefore held under customary law land rights, while those for commercial, industrial and institutional are held under common law (leasehold). Customary grants are not registered at the Deeds registry but only at the Land Board. It thus cannot be used as collateral security for the raising of mortgage finance unless surveyed and converted into common law. Freehold land on the other hand, permits management of the land to rest with the owners in accordance with legislation such as the “Agricultural Resources Conservation Act” (Cap 35:06). This legislation puts restrictions on the owner in terms of actual tenure, sales and leases. The study area reflects a range of allocation patterns in various stages of transformation from a traditional pattern to the more recent contemporary (modern) layouts. The study area falls within two countries—Botswana (93%) and South Africa (7%). The land uses in Botswana’s Section can be defined in two broad categories. The first category relates to those predominantly rural uses situated outside the built-up areas of the villages and towns, whilst the second are those uses of an urban nature situated within the boundaries of the villages. The catchment/study area is made up of the built-up areas (settlements), areas of agriculture (both cultivation/cropping and grazing), waterbodies/Dams as well as other tourism-related areas and key public facilities. Agriculture is the second largest sector of Botswana’s economy and is the backbone of the rural economy, which supports two-thirds of the national population. Spatially the cropping and grazing (agriculture) component together, makes up the largest land use within the catchment/study area, with grazing making up the majority of the agricultural land, eg Rankoromane Farm near Otse. The second largest land use is the built-up area—which is all the major villages and towns, as well as all the smaller villages like Ranaka and Ntlhantlhe. The waterbodies/Dams land use category, occupy temporarily or permanently some parts of the study area. A very serious problem is caused by the occasional flooding of

Water resources of arid areas

734

the settlements within the study area. Most of the flooded areas are located in settlements eg Ramotswa. In addition to the above land uses, is the game/nature reserve/parks component, most of which are privately-owned, and related to tourism (eg Mmokolodi, St Clairs Lion Park, Manyelanong Game Reserve, etc). This land use is concentrated to the north and east of the study area, and is one of the smallest land uses within the catchment area. All land within and around the Dam (Forest Hill 9-KO) itself is freehold land. The land use within this area can be categorised into the following uses: industrial, quarrying, residential, commercial, recreational, social and public uses, agricultural and infrastructure, and fall into an area defined as a “greenbelt” and form part of a Regional Park EHES (2002). This portion of the catchment area falls within the jurisdiction of Lehurutshe Council, of the Central District Municipality of the North West Province of South Africa. The land use in this area is a mix of rural/tribal settlements, characterised by low density, scattered homesteads, and privately-owned game farms and/or nature reserves. (A portion of the Madikwe Game Reserve falls within this catchment area.) There are no major public or civic institutions in this portion of the study area, and visually looks like the central and south of the catchment area falling within Botswana. Added to the above is the industrial development taking place in the catchment area. The major economic activity in the catchment area is services, which employs on the average about 45% of the labour force. This is followed by industry, which accounts for about 35% of the labour force with construction employing most people in the industry. Agriculture accounts for about 9% in the catchment area, but however employs more people in the rural areas. Commerce, which entails wholesaling and retailing employs about 11%. Economic activities are concentrated in the Ramotswa area. Over two-thirds of the economic activities (68.1%) are located in and around Ramotswa, and reflects Ramotswa’s proximity to Gaborone; Ramotswa has developed a nucleus of activities in the sectors of grain milling, clothing and furniture industries along with “common” activities such as brick moulding and metal works; ● The most common economic activities are brick moulding, metal works, clothing and furniture production. ● Six tanneries are located in Ramotswa and Kanye. ● Nine scrap yards exist evenly spread over the project area. Livestock production, especially range cattle and goat production, is the main form of agricultural activity in the Gaborone Dam catchment. 4.1.1 Crop production It is observed that the gross maize, sorghum, and beans (pulses) production in the districts is very low. For example, in the 1996/1997 cropping season, which was the best in the records, the whole of Ngwaketse South district produced only 734 tons of sorghum and 916 tons of maize. On the other hand, the same district only produced 15 and 76 tons of sorghum and maize respectively in the 1999/2000 season. These data confirm the point that the catchment is located in an area of very low rainfed crop production potential.

The effect of socio-economic activities on watershed management

735

4.1.2 Range livestock production It is estimated that there are about 125,000 cattle in the catchment, after allowing for the fact that the catchment only takes 10% of Ngwaketse South extension district. The goat population in 1996 was in excess of 175,000. 4.1.3 Dairying There are dairy operations in some of the commercial farms in the eastern part of the catchment area. In the rural sector of the catchment, there are two herds, one at Magobane and the other at Kanye. The former dairy herd had 160 cows in 2001, 77 of them in milk and producing 1463 litres per day. The Kanye dairy comprised 103 cows in 2001, 50 of them in milk and producing 640 litres of milk per day. From an environmental point of view, dairy operations can be a source of water pollution when animal wastes wash into public streams and lakes. The Magobane dairy is the largest dairy close to Gaborone Dam and it is situated on the banks of the Taung River. Fortunately, however, the dairy is adjacent to and upstream of the Magobane wetland. Thus, any wastes that are released from the dairy into the Taung River are likely to be trapped by the wetland, and are therefore unlikely to have a significant direct pollution impact on Gaborone Dam. 4.1.4 Poultry Lately there has been an upsurge in small-scale commercial poultry production in all the villages. Small groups of farmers in most of the villages of the catchment have constructed chicken houses that accommodate between 500 and 6000 chickens at any one time. Chicken manure is rich in nitrogen and therefore has a high potential for polluting water bodies if directly released therein. It was further observed that the manure or waste from these chicken operations is not carefully stored or disposed of. There is a risk of some of this manure being washed directly into the Notwane River and into the Dam. Thus in summary, the problem of pollution in the Gaborone catchment area may be attributed to: ● Land administration problems: different authorities administering land in the same catchment. ● Lack of coordination among the various authorities. ● Land use under different sovereignties—RSA and Botswana. ● Industrial development within the catchment area.

5 ACTION In order to address the problem, a number of actions have been initiated within the paradigm of watershed management. These actions are: (i) Revision and coordination of land use plans. This study is currently underway.

Water resources of arid areas

736

(ii) Some industries have embarked upon pre-treatment of wastewater. (iii) Better monitoring schemes have been designed and are being executed by the DWA Pollution Control Unit. (iv) Most urbanized villages in the catchment area are being provided with waterborne sanitation facilities.

6 CONCLUSION Clearly, environmental protection of the Gaborone Dam goes far beyond the protection of the immediate environment. Whole catchment management is increasingly recognized as essential to the protection of any one component within the system. To achieve this, the functioning of the system must be clearly understood. This can be best done by adopting the paradigm of watershed management. REFERENCES DWA (2002), Environmental impact assessment consultancy for the feasibility study of small to medium Dams in eastern Botswana, Aqualogic Pty Ltd. Final report. Department of Town and Regional Planning (1995). Gaborone Landscape Masterplan 100 pp. EHES (PTY) LTD (2002). A strategic environmental assessment for the Gaborone Dam catchment area. IUCN (1982) Amphibia-Reptilia Red Data Book. Part I. Testudines, Crocodilia, Rhinchocephalia. 426 pp. Nermark, U.P. and Mmopelwa, T.G. (1994). Utilization of small water bodies, Botswana: report of activities towards fisheries exploitation, 1992–1993. Harare (Zimbabwe). 36 pp.

This page intentionally left blank. Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

Author index Ahmed, A.M. 547 Alemaw, B.F. 551, 559 Andam, K.A. 175 Anyemedu, F.O.K. 169 Aslan, Ş. 477 Asmellash, M. 59 Awuor, V.O. 35 Balachandran, K.K. 67 Bauer, P. 03 Bean, J. 73 Bennie, A.T.P. 397 Böttcher, J. 379 Brunner, P. 03 Carter, R.C. 467 Chaoka, T.R. 551, 559 Chavarro, D.C 239, 341, 541 Chengeta, Z. 451 Clarke, S. 143 Cobbing, A.J.E. 87 Davies, J. 87 Dennis, I. 73, 79, 115, 429 Dennis, S.R. 297 Diiwu, J.Y. 347 Drury, S.A. 59 du Preez, M. 297 Duijnisveld, W.H.M. 379 Ellington, R.G. 47 Engelbrecht, P. 143 Farah, H.O. 331 Farr, J.L. 315 Folwell, S.S. 507 Fregoso, A. 541 Frengstad, B. 97 Fry, M.J. 507 Gumiremhete, R. 315

Author index

Hassan, M. 467 Himmelsbach, T. 379 Hötzl, H. 379 Hodgson, F.D.I. 363 Houghton-Carr, H.A. 507 Hussin, Y.A. 239, 341 Ibrahim, A.E. 247, 259 Idowu, E.O. 227 Ilemobade, A.A. 419 Iliya, A. 279 Ilunga, A.M. 23 Keeletsang, M. 541 Kellner, T. 315 Kemiso, D.B. 565 Kheiralla, K.M. 247, 259 Kinzelbach, W. 03 Kistamah, N. 529 Kooke, S.O. 35 Kumar, A. 59 Kumar, H. 323 Lubczynski, M. 239 Lubczynski, M.W. 97, 271, 285, 341, 541 Machingambi, M. 181 Machiridza, R. 181 Mafa, B. 133 Magombedze, L.M. 97 Magowe, M. 29 Makobo, P. 29 Makuya, M. 519 Manzungu, E. 181 Mapanda, W. 541 McCartney, M. 493 Mishra, G.C. 105 Mkwizu, Y.B. 157 Molwalefhe, L. 459 Molwalefhe, L.N. 409 Mpala, T. 201 Msangi, J.P. 207 Mudzingwa, B. 315 Mugabe, F.T. 501 Mulwa, J.K. 303 Muyima, N.Y.O. 535 Mvungi, A. 519 Ndiritu, J.G. 445

738

Author index

Nkotagu, H.H. 157 Ntsatsi, J. 323 Nyarko, K.B. 169, 217 Nzaba, A.S. 331 Obakeng, O. 239, 271, 285, 341, 541 Obakeng, T. 29 Odai, S.N. 169, 175 Oduro-Kwarteng, S. 169 Olubode-Awosola, O.O. 227 Omoto, W.O. 35 Opere, A. 357 Opere, A.O. 35 Paimpillil, J.S. 67 Pironcheva, G. 535 Pretorius, J. 79 Rooke, E.R. 485 Roseunee, S. 529 Rushton, K.R. 467 Sally, H. 493 Sbeih, M.Y. 191 Scheuerlein, H. 435 Schwiede, M. 379 Selemani, M. 515 Senzanje, A. 493, 501 Sharma, T.C. 331 Shemang, E.M. 279, 323, 551, 559 Siegfried, T. 03 Sitters, C.W.M. 331 Stadler, S. 379 Staudt, M. 371 Stephenson, D. 23, 419, 451 Tesfaslasie, F. 59 Thabeng, G.S. 565 Totolo, O. 551 Tredoux, G. 143 Trifunovic, N. 175 Türkman, A. 477 Uka, Z.B. 507 Umoh, U.T. 389 Usher, B. 79 Usher, B.H. 47, 123, 363 van Tonder, G. 73, 429

739

Author index

van Tonder, G.J. 17, 47, 115, 297 Veltman, S. 123 Vermeulen, D. 429 Vermeulen, P.D. 17 Viswanatham, K.S. 59 Vogel, H. 133, 371 von Hoyer, M. 379 Vriend, S. 459 Woyessa, Y.E. 397 Yilmaz, L. 151 Zamxaka, M. 535 Ziwa, C. 541

740