BIOSPHERE IMPLICATIONS OF DEEP DISPOSAL OF NUCLEAR WASTE The Upwards Migration of Radionuclides in Vegetated Soils
This page intentionally left blank
BIOSPHERE IMPLICATIONS OF DEEP DISPOSAL OF NUCLEAR WASTE The Upwards Migration of Radionuclides in Vegetated Soils
H S Wheater J N B Bell A P Butler B M Jackson L Ciciani Imperial College London, UK
D J Ashworth Imperial College London, UK & US Salinity Laboratory, Riverside, California, USA
G G Shaw Imperial College London, UK & University of Nottingham, UK
ICP
Imperial College Press
Published by Imperial College Press 57 Shelton Street Covent Garden London WC2H 9HE Distributed by World Scientific Publishing Co. Pte. Ltd. 5 Toh Tuck Link, Singapore 596224 USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601 UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE
British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library.
Series on Environmental Science and Management — Vol. 5 BIOSPHERE IMPLICATIONS OF DEEP DISPOSAL OF NUCLEAR WASTE The Upwards Migration of Radionuclides in Vegetated Soils Copyright © 2007 by Imperial College Press All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means, electronic or mechanical, including photocopying, recording or any information storage and retrieval system now known or to be invented, without written permission from the Publisher.
For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is not required from the publisher.
ISBN-13 978-1-86094-743-8 ISBN-10 1-86094-743-3
Typeset by Stallion Press Email:
[email protected]
Printed in Singapore.
KwangWei - Biosphere Implications.pmd
1
8/29/2007, 4:06 PM
May 3, 2007
15:1
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Acknowledgements
In addition to the authors, a number of other individuals have played a major role in our 17-year programme. In particular we would like to thank: Steve Burne, who played such a key role in the operation of the lysimeters, supported by the excellent technical assistance of the late John Ions; Penny Wadey who carried out the experiments over the greater part of the programme; Margaret Minski for her contribution to the earlier stages of the work, including the radiological protection aspects; Qing Chen, Jo Marchant and Alison Bostock as PhD students carrying out ancillary research; Jake Tomkins for several years’ contribution to the modelling; numerous staff at Nirex, in particular Paul Degnan in recent years; and Mike Thorne for his quality assurance and technical input.
v
fm
May 3, 2007
15:1
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
This page intentionally left blank
fm
May 3, 2007
15:1
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Preface
Radioactive wastes contain a wide spectrum of radionuclides arising from commercial power production, industry, medical applications and defence-related activities. In the UK today, much of the lowlevel radioactive waste that is generated as a result of these activities is conditioned and packaged as it arises, prior to transport to a near-surface disposal facility at Drigg, West Cumbria. However, the storage and disposal options for other longer-lived solid low-level, intermediate-level and high-level radioactive wastes, as well as for other materials, such as separated plutonium, that could eventually be declared as wastes are still to be decided. A leading option for the long-term management of various types of radioactive wastes is deep geological disposal. In the UK the radioactive waste management agency United Kingdom Nirex Limited (Nirex) has been investigating issues concerned with the deep geological disposal for more than 20 years and, as part of its research and development programme, Nirex has developed a Phased Geological Repository Concept. In this concept the wastes are to be emplaced in a repository constructed at several hundred metres depth in an appropriate host geological formation, but importantly the phased repository concept includes the potential for retrieving the wastes and, by relatively straightforward means, having a capability to monitor them for extended periods of time (up to several hundred years). Over the very long timescales that need to be considered in post-closure radiological safety assessments much of the activity initially associated with the waste will reduce to negligible levels over relatively short periods of time (hundreds to several thousand years). However, due to the long half-lives associated with certain vii
fm
May 3, 2007
15:1
viii
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
radionuclides some would remain hazardous for very much longer periods of time. Although any releases of radionuclides that might eventually be released from a deep geological repository would be in very low concentrations they would largely be precluded from migrating to the inhabited environment (the biosphere) by a number of physical and chemical barriers. These barriers relate to the engineered facility and to the characteristics of the surrounding host and overlying rock. For example, the engineered facility can provide both physical and chemical containment, e.g. by limiting degradation of the waste form and by inducing a chemical environment that suppresses solubility and enhances sorption of many of the radionuclides in the waste. Similarly, the host and overlying rock can inhibit radionuclide transport by slow or non-existent groundwater flow, by sorption of radionuclides onto mineral surfaces, and by diffusion of radionuclides into the rock matrix. Any low concentrations of radionuclides that do migrate away from the immediate repository and host rock environment would be further reduced over time by dispersion and mixing with overlying groundwaters. Radiological safety assessments are typically undertaken by both radioactive waste management agencies and by the regulatory authorities to evaluate the potential and consequences for any such discharges. As a consequence of the effects of solubility limitation and sorption, many of the radionuclides that are of environmental interest in the context of effluent discharges from nuclear power plants, e.g. 137 Cs, 239 Pu and 241 Am, do not generally feature as a concern in safety assessments of deep geological repositories for radioactive wastes. Rather, the emphasis is on the very long-lived radionuclides that are highly soluble in repository conditions and that are adsorbed to only a very limited degree on repository materials or the host and overlying rock. In particular, radionuclides that are released and transported in anionic form tend to dominate the radiological impacts determined in post-closure safety assessments. As a result of the assessment activities undertaken as part of the Nirex research programme and elsewhere, 36 Cl, 79 Se, 99 Tc and 129 I in particular have been identified as being of interest as over the very long timescales that typically need to be investigated (up to and
fm
May 3, 2007
15:1
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Preface
ix
beyond one million years) they have the potential to migrate to the biosphere. In assessment scenarios where radionuclides do migrate to the biosphere, a further consideration relates to the way in which the accessible environment could become contaminated. In the past, the emphasis has been on releases of radioactive materials to the atmosphere, from weapons’ testing in the 1950s and 1960s, the routine operations of the commercial nuclear fuel cycle, and nuclear accidents. Therefore, in considering contamination of soils and subsequent transfers through the food-chain, the emphasis has been on radionuclide deposition from the atmosphere. However, in the context of the geological disposal of radioactive wastes, the routes by which the accessible environment may become contaminated include abstraction of groundwater, for use in irrigation and for other purposes, and the natural discharge of groundwater from depth to soils and the aquatic environment. Contamination by irrigation may be treated somewhat analogously to deposition from the atmosphere. However, contamination by natural discharge presents substantially different considerations and it is those considerations, as they apply to the radionuclides of particular interest in solid radioactive waste disposal, that Imperial College has explored in an extensive experimental programme over the last eighteen years. When the Imperial College experimental programme began, the approaches used in radiological assessments to relate radionuclide concentrations in soil to uptake by plants were generally highly simplified and empirical. With the emphasis being placed on uptake following surface contamination, the standard approach was to average the concentration of a radionuclide in soil over the depth of the rooting zone and then to estimate the concentration in plants by applying a concentration ratio. This concentration ratio was estimated from experimental studies (often conducted in pots in laboratory conditions) or from observations of stable element concentrations in soils and plants, where it was felt that stable element observations could legitimately be applied to radioisotopes of those elements. However, with contamination of the soil from below and with the need to utilise results from a moderately extensive experimental
fm
May 3, 2007
15:1
x
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
programme in a wide range of potential environmental contexts, the existing empirical approach was felt to be inadequate. For this reason and from the outset, the experimental programme at Imperial College was complemented by an extensive, process-based mathematical modelling programme directed first to understanding the processes underlying the experimental results obtained, but subsequently also to determining the implication of those results in an assessment context. Thus, hydrological and contaminant transport models were created that could take into account changes in the hydrological and hydrochemical environment of the soil, both with depth and over the course of time, and the development of the rooting systems of plants over time in that changing hydrological and hydrochemical environment. In turn, this emphasis on process-based modelling strongly influenced the nature of the experiments that were undertaken. Previous experimental studies had typically paid little regard to either hydrological conditions in the soil or variations in the oxidation-reduction regime with depth. In contrast, in the Imperial College experiments, comprehensive hydrological monitoring was included from the outset, with provision made to measure both soil water content and potential. In addition, techniques were developed to measure the development of root profiles. In later studies, the capability was developed to extract small volumes of soil solution during the course of the experiments, so permitting a full suite of chemical and radiochemical analyses to be conducted. Initial studies carried out at Imperial College were conducted in ambient conditions in lysimeters with a cross-sectional area of about 1.6 m2 and a water table that was controlled such that it was maintained at a fixed depth. Subsequently, other experiments were undertaken on substantially smaller soil columns in a laboratory with a controlled environment. Through process-based modelling, it was shown that parameter values determined in the lysimeters could be transferred to the column experiments and vice versa. This justified the later stages of the programme exclusively utilising column experiments. The adoption of soil columns as an experimental basis allowed a wider range of conditions to be explored than would have
fm
May 3, 2007
15:1
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Preface
xi
been possible in the lysimeters. Thus, the column studies included comparisons of radionuclide behaviour in different soil types, in intact versus repacked soil, with fixed and sinusoidally varying water table heights, and in the presence of different concentrations of stable chlorine isotopes. In the most recent experiments, the system has been further augmented by the inclusion of supplementary experiments in mini-columns, which provide additional insights into the influence of different degrees of water saturation and redox conditions on radionuclide behaviour. The Imperial College studies should not be seen in isolation. When the Nirex biosphere research programme was initiated in 1987, a structured review was undertaken to identify those areas in which it would be appropriate for Nirex to undertake research to complement that being undertaken for a variety of purposes by other organisations. As described above, one of those areas was determined to be a detailed study of the transport of key radionuclides in soil and uptake by plants when contamination occurred from below. Other areas included the influence of climate change and landform development on the accessible environments into which radionuclides might emerge far in the future, the development and application of 3D catchment models to represent the transport of radionuclides in such environments, and evaluation of the adequacy of existing models of the development of ice-sheets for use in assessment studies. As with the Imperial College experimental and modelling studies, research in these other areas has been consistently funded by Nirex and research in many of these areas continues, both at a national level and through internationally funded programmes. In a programme extending over eighteen years, it was inevitable that results of the Imperial College work would initially appear in a variety of Imperial College and Nirex reports, as well as in journal articles and conference papers. Fortunately, the team involved in the research has been extremely stable throughout that period and several of the founder members are still involved. Furthermore, from the outset, both the experimental and modelling work was undertaken within a rigorous quality assurance framework, which has meant that all the relevant experimental and modelling results remain readily
fm
May 3, 2007
15:1
xii
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
available for scrutiny. Building on these secure foundations, it has been possible to pull together all the work that has been undertaken into this authoritative monograph. Nirex is proud to have sponsored such an innovative programme of work and hopes that this publication will provide an ongoing useful resource to members of the radioactive waste management and radioecological communities. In addition, it should be of considerable relevance to both research workers and commercial organisations involved in related areas, e.g. to those studying the behaviour of heavy metals in the environment.
Dr. Paul Degnan (United Kingdom Nirex Limited) and Dr. Mike Thorne (Mike Thorne & Associates Limited, scientific advisor to Nirex on biosphere science) Harwell, UK, November 2006
fm
May 3, 2007
15:1
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Contents
Acknowledgements
v
Preface
vii
Section 1
Background
1
Chapter 1.
Introduction
3
Section 2
Methods
11
Chapter 2.
Experimental Protocols
13
2.1. 2.2. 2.3. 2.4. Chapter 3.
Background . . . . . . . . . . . . . . . Phase I and II Lysimeter Experiments Soil Column Experiments . . . . . . . Mini-column Experiments . . . . . . .
. . . .
Modelling Radionuclide Transport and Uptake in Vegetated Soils 3.1. 3.2. 3.3. 3.4. 3.5.
Introduction . . . . . . . . . . . Lysimeter System Model . . . . Physically-Based Modelling . . . Hydrological Model Applications Discussion . . . . . . . . . . . .
xiii
13 14 29 35
41 . . . . .
. . . . .
. . . . .
. . . . .
41 42 57 64 84
fm
May 3, 2007
15:1
xiv
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Section 3
Results
87
Chapter 4.
Radiochlorine
89
4.1. 4.2. 4.3. 4.4. Chapter 5.
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
Background . . . . . . . Experimental Overview Results . . . . . . . . . General Discussion . . .
Background . . . . . . . Experimental Overview Results . . . . . . . . . General Discussion . . .
Background . . . . . . . Experimental Overview Results . . . . . . . . . General Discussion . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
Background . . . . . . . Experimental Overview Results . . . . . . . . . General Discussion . . .
188 192 194 226 230
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
230 237 238 263 266
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
Radiocations 8.1. 8.2. 8.3. 8.4.
89 90 93 184 188
Radioselenium 7.1. 7.2. 7.3. 7.4.
Chapter 8.
. . . .
Technetium 6.1. 6.2. 6.3. 6.4.
Chapter 7.
. . . .
Radioiodine 5.1. 5.2. 5.3. 5.4.
Chapter 6.
Background . . . . . . . Experimental Overview Results . . . . . . . . . General Discussion . . .
266 269 269 294 300
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
300 304 305 338
fm
May 3, 2007
15:1
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Contents
xv
Section 4
Conclusions and Recommendations
343
Chapter 9.
Conclusions
345
9.1. 9.2. 9.3. 9.4. 9.5.
Overview . . . . . . . . . . . . . . Summary of Results . . . . . . . . Experimental Design . . . . . . . . Modelling . . . . . . . . . . . . . . Conclusions and Future Directions
. . . . .
. . . . .
. . . . .
345 346 361 363 365
Colour Section
367
Appendix
379
Data Archive
381
Reference
383
Index
397
fm
This page intentionally left blank
May 3, 2007
14:14
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
SECTION 1
BACKGROUND
1
ch01
May 3, 2007
14:14
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
This page intentionally left blank
2
ch01
May 3, 2007
14:14
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
CHAPTER 1
Introduction
Radioecology is still a relatively young science, having originated after the development of nuclear weapons at the end of World War II, and is concerned with both effects and pathways of radionuclides in the environment, particularly the latter. Initially, pathways research was concerned with routes through food chains by which human exposures could arise from radionuclides deposited as fallout from nuclear weapons testing in the atmosphere (Scott Russell, 1966). Following the international moratorium on such tests in 1963, attention was switched to pathways to humans after operational releases or accidental discharges from nuclear installations, particularly power stations. In the case of accidents, an enormous volume of research was carried out following the widespread deposition of radionuclides over Europe as a result of the Chernobyl disaster in 1986 (Savchenko, 1995; Warner & Harrison, 1993; Desmet et al., 1990). Much of this was concerned with pathways in semi-natural ecosystems, including potential routes to humans via food chains (Shaw & Bell, 2001; Desmet et al., 1990; Bell & Shaw, 2005). Most of this research was concerned with radioactivity deposited from the atmosphere onto both native vegetation and crops. The vast amount of data generated permitted a large number of calculations of soilplant transfer factors, which are defined as radioactivity per dry weight of various plant organs divided by the radioactivity per dry weight of soil within the rooting zone (Shaw & Bell, 2001). Transfer factors have been collated in a data base set up by the International Union of Radioecologists (IUR) (Frissel, 1992) which thus 3
ch01
May 3, 2007
14:14
4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
provides a compilation of the efficiencies of different plant species in taking up individual radionuclides from a range of soil types; in all cases there is a range of values, depending on soil, climate and biological factors. In addition, a large number of models have been developed which predict the distribution and transport of radionuclides in air-soil-plant systems (Thiessen et al., 1999). With the passing of the years, a large inventory of nuclear waste and contaminated areas and artefacts has built up in all developed and some developing countries. This ranges from low level waste, consisting of items such as contaminated gloves, laboratory coats and laboratory equipment, to intermediate level waste which largely consists of components of fuel elements from nuclear reactors, to high level waste which comprises both spent fuel and some reprocessing wastes. Currently in the UK, low level waste is disposed of by placing it in sealed containers in engineered vaults at Drigg on the Cumbrian coast, in a facility operated by the British Nuclear Group (BNG) (formerly British Nuclear Fuels). However, the capacity of this facility is limited and it will be filled in the foreseeable future. In the case of high level waste, this is stored in a concentrated vitrified form in special buildings at the BNG site at Sellafield, Cumbria pending a national policy decision on disposal. Currently there is also no disposal route for intermediate level waste in the UK. Radioactive waste management has been a major government policy issue for over 20 years. In the 1970s considerable use was made of deep sea disposal, but this ceased when an international voluntary moratorium on such action was agreed in 1983, under the London Dumping Convention. In the 1980s the government initially investigated additional shallow disposal methods, but this provoked widespread opposition, which resulted in plans for the development of such a sites for low level waste being abandoned in 1987. This was accompanied by a policy change in favour of a deep repository for both intermediate and low level waste. Initially, some 500 possible sites for this were considered, but this was reduced to a shortlist of 12 and subsequently to just two viz. the nuclear sites at Sellafield and at Dounreay on the north coast of Scotland.
ch01
May 3, 2007
14:14
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Introduction
5
In 1982, UK Nirex was set up by the government to implement a strategy for disposal of low and intermediate level wastes. Since 1988, Nirex has funded extensive research into the deep disposal of such wastes under its Geosphere and Biosphere Research Programmes. The latter concerns the study of processes in the nearsurface and surface environment that might affect radiation doses to humans over very long time periods as a result of radionuclides migrating upwards by various routes to reach the surface and entering into food chains. The aim of this research has been to develop appropriate tools and data-bases for the safety assessment of a deep repository. Currently (2005) in the UK the issue of radioactive waste management is being investigated by the government’s recently established Committee on Radioactive Waste Management (CoRWM). This has reviewed all possible options, including some as bizarre as shooting packages of waste into space and transmutating certain radionuclides to other elements. At the time of writing, these options have been whittled down to thirteen, three of which involve deep geological disposal, which is the route favoured by most experts around the world, but not necessarily by the environmental movement and the general public. Since 1988 Imperial College has had a multi-disciplinary team contracted to Nirex and working on the Biosphere Programme. This team has brought together experts with a range of skills and experience, which has facilitated the essential multi-disciplinary approach to this complex problem. It has involved a number of College departments, these currently being the Department of Civil and Environmental Engineering, the Centre for Environmental Policy and the Division of Biology. In Civil Engineering, the research has been conducted in the Environmental and Water Resource Engineering Section (Professor H. Wheater and Dr. A. P. Butler), which has worked for many years on a whole range of hydrological studies, with strong emphasis on modelling groundwater flow. In both Biology and Environmental Policy (Professor J. N. B. Bell, Professor G. Shaw and Dr. D. Ashworth) experimental research has been
ch01
May 3, 2007
14:14
6
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
conducted since 1980 on the migration of radionuclides into soil from surface deposition and subsequent uptake into crops, representing situations following operational or accidental releases from nuclear installations. This ongoing work consists of both experimental and numerical modelling work on the upward movement in the unsaturated zone of a suite of radionuclides that are of radiological concern in terms of what might reach the surface from a deep repository. The context for this work is the assumption that on the timescales of relevance to radionuclide safety assessment, physical containment and chemical containment in an engineered repository will break down. Groundwater then becomes a potential transport pathway of radionuclides to the near-surface environment. Risk to humans is affected by direct ingestion of contaminated plant material, or indirectly, through ingestion by animals, hence an important element of the safety assessment of subsurface disposal of radioactive wastes is the movement of radionuclides through soils and their uptake by vegetation. At the outset of the Biosphere research programme it was recognised that this was an area of considerable process complexity. Upwards migration of contaminants in soils is determined by the interactions of the soil-plant-atmosphere system that determine the flux of water in response to plant root uptake. Solute movement can occur by advection, or by root uptake and translocation, and is strongly influenced by geochemical interactions with the soil. In turn, these geochemical interactions are strongly influenced for certain radionuclides by the redox status of the soil, in a zone where strong redox gradients occur across the interface between the saturated zone of groundwater, and the unsaturated zone of the soil water system. Given this context, it was felt that the representation in safety assessment procedures, largely based on soil-plant transfer factors derived from atmospheric deposition, was simplistic, and that appropriate data to support the safety assessment (in particular concerning upward migration of radionuclides) did not exist. A programme of research was therefore developed at Imperial College with the
ch01
May 3, 2007
14:14
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Introduction
7
following strategic goals: (a) To develop an appropriate level of scientific understanding of soil migration and soil-plant transfer processes to support a credible safety assessment case for subsurface disposal of radioactive wastes. (b) Development of models to transfer results from the research to safety assessment. (c) Provision of guidance on appropriate parameter ranges and parameter uncertainty for safety assessment. The experimental work has involved outdoor lysimeters and columns in a controlled environment in which the radionuclides are injected into a simulated water-table towards the base of these containers and upward movement measured over a time course. A generic crop is grown in the soils which fill the containers, this being either winter wheat (Triticum aestivum L.) or perennial ryegrass (Lolium perenne L.). A key element of the experimental programme is the measurement of uptake of the radionuclides into the roots and aerial parts of these crops, this being the most likely route for human detriment. The second part of the programme is a series of numerical modelling exercises designed to simulate the movement upwards of radionuclides through the soil/plant system, on a time course basis, with a root uptake model being central to this part of the exercise. Such modelling is applied to both the lysimeter and column experiments, as an essential tool in the analysis of complex data sets, and as a means of generalisation of the results for safety assessment. The experimental programme has been carried out continuously since 1990 in eight phases. The first two phases were conducted in lysimeters from 1990 to 1993, followed by a single phase II lysimeter experiment in 1995/6 (the data from the Phase 1 experiments were subsequently used as the basis of an IAEA model intercomparison project (Butler et al., 1998)). After the end of the first year of the latter phase there was a policy change at Nirex, which resulted in support for the complex and expensive lysimeter studies being withdrawn. Instead future work was conducted in columns in a controlled environment cabinet, which had the same basic experimental set-up
ch01
May 3, 2007
14:14
8
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
as the lysimeters, but on a much reduced scale. The column studies covered Phases III to VIII, with some ancillary research, as well as pot and mini-column studies. Phase II involved the comparison of radionuclide migration in two soil types. However, the columns provided a greater degree of flexibility. This permitted examination of a wider range of soil types, including French soils, as well as use of intact cores on occasion. The French soils were included because Phase VIII, which studied the behaviour of 75 Se, was supported by both Nirex and Andra, the equivalent radioactive waste management authority in France. The radionuclides studied at different stages throughout the programme were 137 Cs, 134 Cs, 75 Se, 36 Cl, 99m Tc, 125 I, 22 Na, and 60 Co. A summary of the eight phases is given in Table 1. This book gives an overview of the entire 17 years’ period for which Nirex and more recently Andra have funded the programme. To the best of the authors’ knowledge, this is a uniquely comprehensive programme and remains the definitive research elucidating the behaviour of the upward migration of radionuclides in the unsaturated zone of soils and their associated vegetation. While it is specifically concerned with radionuclides, the techniques employed and the lessons learned have direct relevance for upward movement of other contaminants from groundwater and are thus of interest to research on contaminated land. Furthermore, a vast amount of information has been generated on plant uptake, which may be applied to the development of phytoremediation as a tool for the management of contaminated sites. While the experiments and modelling described in this book are aimed primarily at the safety case for British, and to a lesser extent French, waste disposal, the results are applicable to other countries where in all cases some form of burial will be the ultimate radioactive waste management solution. Of particular importance is the decision by the USA Federal Government to propose Yucca Mountain in Nevada for a massive subterranean vault for waste disposal in the unsaturated zone (Umstadter & Baciak, 2002). This has generated its own research programmes, including on transport of fission products and actinides (e.g. Patera et al., 1990). Other research into sub-surface disposal has been carried out in Sweden, (S.K.B., 1998,
ch01
May 3, 2007
14:14
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Introduction Table 1.
9
Summary of experimental programme, 1990–2004.
Phase
Year
Container
Soil
I
1990/1 1991/2 1992/3
Lysimeters
Silwood
II
1995/6
Lysimeters
III
1997/8
Columns
IV
1999/2000
Columns
Silwood; Longlands Farm Silwood; Wellesbourne; Robertgate (undisturbed cores) Silwood
V
2000/01
Columns
Silwood
VI VII
2001 2002/3
Columns Columns
Silwood Silwood
VIII
2003/4
Columns, Silwood; mini-columns, Meuse/Hautepots Marne clay loam and sandy loam
Radionuclide 99
Tc, Cl, 22 Na, 137 Cs, 60 Co 22 Na, 137 Cs, 36 Cl 22 Na, 134 Cs, 36 Cl 36
Crop
Watertable
Winter wheat
Fixed — 2 levels
Ryegrass
Fixed
Ryegrass
Fixed
36
Cl + Ryegrass Fixed Stable Cl. 125 I Ryegrass Fixed and fluctuating 99m Tc Ryegrass Fluctuating 36 Cl, 125 I Ryegrass Fixed and fluctuating 75 Se Ryegrass Fixed
1999a, b), Switzerland (NAGRA, 2002; Nuclear Energy Agency, 2004) and Finland (Posiva, 1999). Some more general references on this topic are Chapman & McCombie (2003), Nuclear Energy Agency (1999) and Savage (1995). There is no centralised information on studies of this type, with the results of the studies being scattered in reports and published papers. In fact the most credible programmes are largely focussed on the geosphere. A limited number of studies have examined to some extent upward migration. Examples of these are the work of Torok et al. (1990) at Chalk River, Ontario, who carried out laboratory scale lysimeter experiments with simulated waste forms placed in potential materials for
ch01
May 3, 2007
14:14
10
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
a low-level radioactive waste repository, with development of associated transport models. At least 2 studies have examined the role of burrowing animals (Arthur & Markham, 1983) and earthworms (Muller-Lemans & Van Dorp, 1996) in the upward movement of radionuclides. However, very little research has been conducted on upward movement into vegetation, a notable exception being that reported by Murphy & Johnson (1993), examining the uptake of 99 Tc and 137 Cs into grass, crops and trees from a planned low level waste repository at Savannah River, South Carolina. An important development in recent years has been the BIOMASS project, which has developed a set of “Reference Biospheres” for solid radioactive waste disposal, for use in safety assessments of repositories (Crossland et al., 2005). This was established under the auspices of the International Atomic Energy Agency, and involved regulators, waste disposers and independent experts from around the world. It will be seen that the information on soil-plant transfer of radionuclides due to upwards migration is limited and diverse, and hopefully the work presented here, despite its UK origins, will be seen as a unified and significant contribution to the international literature. Following this introductory chapter, the book consists of two chapters in which the methods used are discussed and defined, namely Chapter 2 — Experimental design and protocols, including the physical and chemical characteristics of the soils; Chapter 3 — Modelling. Experimental and modelling results are integrated by radionuclide in Chapters 4 to 9; Chapter 4 — Radiochlorine; Chapter 5 — Radioiodine; Chapter 6 — Technetium; Chapter 7 — Radioselenium; Chapter 8 — Radiocaesium, radiocobalt and radiosodium. Finally Chapter 9 presents the conclusions and recommendations. An Appendix provides information on web-based access to data files from key parts of the work.
ch01
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
SECTION 2
METHODS
11
ch02
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
This page intentionally left blank
12
ch02
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
CHAPTER 2
Experimental Protocols
2.1.
Background
Experimental studies on sub-surface contamination of soils with radionuclides began in 1989 with the refurbishment of a series of field lysimeters at Silwood Park, situated in Berkshire, southern England (N51◦ 24 29 , W0◦ 38 54 ). These were originally constructed in 1959 for use in a study of soil-plant water relations (Etherington, 1962; Etherington & Rutter, 1964). The whole structure allowed soil monoliths of approximately 1.6 m3 to be established with their surfaces level with the surrounding ground surface, a feature which is essential to preserve realistic micrometeorology when studying surfaceatmosphere exchanges of water vapour and radiant energy. Lysimeters are, by design, hydrologically isolated from the bulk soil, hence their use in studies on the behaviour of soil contaminants of all types is commonplace. They provide an ideal field facility for the study of radionuclides in soil/plant/groundwater systems. Studies on surface radioactive contamination of soils using the Silwood Park lysimeter system began in 1983. As part of a CEC1 funded study, eighteen lysimeters were refurbished and contaminated with 137 Cs, 144 Ce, 106 Ru and 125 Sb as described in detail by Mitchell (1992). This study ran from 1983 to 1985. After lying fallow from 1985 to 1986, a further CEC funded study used ten of the lysimeters originally contaminated in 1983 plus one additional clean lysimeter to 1
Commission of the European Communities. 13
ch02
May 3, 2007
14:19
14
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
study the effects of ageing of radiocaesium in the soil. This involved the fresh application of 134 Cs as described by Munro (1991). Ageing is a process which results in progressively firmer attachment of radionuclides, especially radiocaesium, to the soil matrix, reducing biological uptake but prolonging the residence time of radionuclides in the rooting zone. It is thus an important process which must be taken into account in medium- to long-term radiological assessments. The second CEC study ran from 1986 to 1989. In 1987, a CEGB2 funded study was initiated which involved the addition of 134 Cs, 137 Cs, 60 Co, 54 Mn, 106 Ru and 90 Sr to eight of the lysimeters originally contaminated in 1983, as described by Shaw et al. (1990). Contamination involved excavation of the top 20 cm of the soil from a lysimeter, mixing in of the radionuclide cocktail and replacement of the contaminated soil in the same lysimeter. The sub-soil, below approximately 25 cm, remained essentially uncontaminated at the end of the experiment since each of the radionuclides used has limited mobility in the soil profile. This experiment ran from 1987 to 1990.
2.2. 2.2.1.
Phase I and II Lysimeter Experiments Lysimeter refurbishment and reinstatement
Up until the commissioning of the Nirex Phase I lysimeter experiment, all the studies carried out within the Silwood Park lysimeters involved contamination of the soil surface with radionuclides. In 1988, a major programme of work was initiated to study the migration of radionuclides from the bottom of selected lysimeters to the soil surface. This necessitated a major re-design and refurbishment of nine of the lysimeters in 1989 (Phase I) and a further nine in 1995 (Phase II). In 1989, nine of the uncontaminated lysimeters on the west side of the drainage trench (A, B, C, D, E, F, G, and H, plus an additional lysimeter adjacent to E to house a common reservoir; see Fig. 2.1) were cleaned out and refurbished. This involved excavation of all 2
Central Electricity Generating Board, UK.
ch02
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Experimental Protocols
15
Fig. 2.1. Plan view of the Silwood Park lysimeter system showing the Phase I lysimeters, A–H.
the original soil filling the eight lysimeters followed by repair of the exposed internal concrete surfaces and, finally, painting with a chlorinated rubber paint to minimise sorption of radionuclides to the lysimeter walls. In addition, four of the lysimeters were fitted with raised bottoms to reduce their internal depth. Four of the lysimeters were subsequently re-constructed as ‘deep’ lysimeters with total soil depths of 70 cm (Fig. 2.2), and the remaining four, with raised bottoms, as ‘shallow’ lysimeters which had 40 cm deep soil profiles. All lysimeters had a surface area of 90 × 180 cm. Each of the Phase I lysimeters was filled with soil overlying a 20 cm deep, inert (polythene) bead substrate, the soil and the beads being separated by a geotextile layer. The soil type chosen was a sandy loam (Wicks series, Eutric Cambisol: Dudal & Tavernier, 1985). This soil was sourced in an adjacent orchard at Silwood Park which records indicated had not been subject to any agrochemical inputs (neither pesticides nor fertilisers) since 1947 when the Silwood Park estate was acquired by Imperial College. The general physical and chemical properties of this soil are shown in Table 2.1. Having
ch02
May 3, 2007
14:19
16
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 2.2.
Longitudinal section of a ‘deep’ Phase I lysimeter.
Table 2.1. General physical and chemical properties of the soils packed into the Phase I and Phase II lysimeters (two values separated by a solidus are the results of two separate analyses). Property
Silwood (0–25 cm)
Silwood (25–50 cm)
Longlands (Phase II only)
Unit
pH Sand (0.063–0.002 mm) Silt (0.063–0.002 mm) Clay (<0.002 mm) Loss on ignition Available P NO− 3 Ca (exchangeable) K (exchangeable) Mg (exchangeable) Na (exchangeable) Cation exchange capacity Water soluble Cl− Available SO2− 4
4.3/4.9 75/86 20/12 5/2 6.3 25 44 2.8 0.22 0.22 0.11 13.5
5.2 88 10 2 1.5 — — — — — — —
4.8 54 35 11 — 22 351 4.3 0.33 0.41 0.14 20.6
— % % % % mg L−1 mg kg−1 meq 100 g−1 meq 100 g−1 meq 100 g−1 meq 100 g−1 meq 100 g−1
20 134
— —
24 145
mg kg−1 mg L−1
ch02
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Experimental Protocols
17
placed the polythene beads and the geotextile into the lysimeter bases the soil was screened through a 1 cm mesh before repacking above the geotextile. Care was taken to differentiate between surface soil and sub-soil when excavating and re-packing the soil and there was a slight textural difference between the two, as indicated in Table 2.1. Prior to introduction of any radionuclides the soil was allowed to settle, encouraged by hydraulic wetting tests which involved slow flooding and draining of the lysimeters via the water table control system described in Sec. 2.2.2. In 1994/95, the Phase II experiment was initiated using nine lysimeters, on the east side of the drainage trench, which had originally been contaminated in 1983. This involved removal and disposal of contaminated soil followed by lysimeter refurbishment as described for the Phase I lysimeters above. The Phase II lysimeters were all ‘deep’ (total soil depth was 70 cm) and were filled with either the same soil used in the Phase I experiment, or with a soil sourced from Longlands Farm in Cumbria, UK (N54◦ 25 19 , W3◦ 27 14 ). Table 2.2 gives an overview of the major attributes of lysimeters used in the Phase I and II experiments.
2.2.2.
Water table control
During the Phase I and II experiments the lysimeters were subject to uncontrolled (ambient) inputs of water due to precipitation and losses of water due to evapotranspiration. A principal requirement of both experiments was the maintenance of a fixed water table height at all times, thus it was necessary to design and implement a control system which would export and import water, as necessary, to control the water table height. The water table control system is shown in Fig. 2.3. It consisted of a series of eight buffer tanks within the drainage trench, each one connected to an individual lysimeter via an injection manifold within the lysimeter’s bead substrate. Each buffer tank was also connected via a system of tubing and peristaltic pumps to a large plastic common reservoir fitted into a ninth lysimeter from which the soil
ch02
May 3, 2007
14:19
18
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste Table 2.2.
Summary of Phase I and Phase II lysimeter attributes.
Phase
Identifier
Soil Depth
Type
Crop
Soil
I I I I I I I I II II II II II II II II
A B C D E F G H J K L M N O P Q
70 cm 70 cm 40 cm 40 cm 70 cm 70 cm 40 cm 40 cm 70 cm 70 cm 70 cm 70 cm 70 cm 70 cm 70 cm 70 cm
Monitored Monitored Monitored Monitored Replicate† Replicate† Replicate† Replicate† Monitored Monitored Monitored Monitored Replicate† Replicate† Replicate† Replicate†
Wheat Wheat Wheat Wheat Wheat Wheat Wheat Wheat Ryegrass Ryegrass Ryegrass Ryegrass Ryegrass Ryegrass Ryegrass Ryegrass
Silwood Silwood Silwood Silwood Silwood Silwood Silwood Silwood Silwood Silwood Longlands Longlands Silwood Silwood Longlands Longlands
†
Replicate lysimeters were non-instrumented but provided additional soil and vegetation samples for statistical purposes.
Fig. 2.3.
Design of the Phase I lysimeter water control system.
ch02
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Experimental Protocols
19
had been removed (Burne et al., 1994). The principle of operation of the control system was to maintain the water level within each buffer tank at a pre-determined level, thus maintaining the water level within each lysimeter at the same target level. Control of the water level within each buffer tank was achieved using two optical switches set 0.5 cm above and below, respectively, the target water level. If precipitation exceeded evapotranspiration a water level rise in the buffer tank triggered the upper switch, thus causing a peristaltic pump to export water from the tank (and thus the lysimeter) into the common reservoir. If evapotranspiration exceeded precipitation the reverse occurred and water was imported from the common reservoir, via the buffer tank, into the base of the lysimeter. The water balance of each lysimeter (i.e. import and export volumes) was recorded by counting the number of revolutions, forwards and backwards, of each pump using a shaft encoder plus decoding electronics. Water storage in the buffer tanks was also recorded using a float and potentiometer water level gauge. Due to the positioning of the optical switches the water levels in each buffer tank were maintained to within ±5 mm and this control system proved robust and reliable over a 5 year operational period. Among other regular maintenance checks, the peristaltic pumps were calibrated routinely to determine the volume of water pumped per revolution. The accuracy of this calibration was typically ±5%.
2.2.3.
Lysimeter instrumentation
Figure 2.4 shows the array of in situ instrumentation which was deployed in selected deep and shallow lysimeters to permit the monitoring of soil conditions and crop development. Four of the deep lysimeters had six tensiometers inserted at 2◦ above the horizontal (to facilitate purging of air) through the concrete wall from the trench. These were positioned in a staggered array at 10 cm intervals from a depth of 10 cm to 60 cm below the soil surface. Four of the shallow lysimeters had three tensiometers positioned in the same way at depths of 10 cm to 30 cm below the soil surface. Each tensiometer was connected to a pressure transducer which was, in turn, coupled
ch02
May 3, 2007
14:19
20
Fig. 2.4.
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Plan view of a Phase I lysimeter showing in situ instrumentation.
to a data logger which recorded internal pressure (equivalent to soil matric potential) within each instrument. Parallel pairs of stainless steel rods were also inserted horizontally at the same heights as tensiometers. These were connected to a Tektronix 1502C Metallic TDR3 Cable Tester which allowed the dielectric constant, K, of the soil between electrode pairs to be determined (Topp & Davis, 1980) (see Sec. 2.2.4). K is proportional to the volumetric water content within the soil and the TDR response was calibrated to provide convenient in situ measurement of the soil moisture content which, in combination with synchronous measurements of the soil matric potential, were used to construct a moisture characteristic curve for each lysimeter soil. Other instrumentation deployed in each lysimeter included Perspex access tubes, inserted at 45◦ to the horizontal, into which a rigid endoscope (‘Borescope’, Olympus Industrial, Lake Success, NY) could be inserted to determine root growth, as well as vertical arrays of thermistors and platinum electrodes for the determination of soil 3
Time Domain Reflectometry.
ch02
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Experimental Protocols
21
temperature and oxidation-reduction potential, respectively. Vertical aluminium access tubes (6.5 and 4 cm diameter) were installed in deep lysimeters for the insertion of a neutron probe to determine soil water content. Finally, a vertical narrow (2.5 cm) bore PVC tube was inserted into each lysimeter, which acted as a stilling well to allow sampling of interstitial water from the saturated region of the soil profile. In addition to the instrumentation installed within the lysimeters themselves, an automatic weather station (AWS) was installed immediately adjacent to the lysimeter site (Fig. 2.1). The AWS provided measurements of solar radiation, net radiation, air temperature, relative humidity, wind speed and direction, and rainfall.
2.2.4.
Routine measurements
Routine measurements of a number of variables within the lysimeter system were made either automatically, using data loggers for data storage, or manually. The frequency of measurements depended on the means used to obtain them, as follows. Automatic data loggers were used to record soil matric potential (tensiometers), soil temperature (thermistors), buffer tank water level (float and potentiometer water level gauge), peristaltic pump revolutions (shaft encoder-decoder) and all climatic data (AWS) at a frequency of 100 d−1 (i.e. one measurement every 14.4 minutes). Manual measurements of soil moisture content (TDR, neutron probe), crop root distribution and density (‘Borescope’), crop leaf area index (inclined point quadrat) and crop height (metre rule) were made at weekly or fortnightly intervals although, during Phase II, continuouslymonitored TDR systems were installed. Samples of solutions were taken manually from the polythene bead substrate, the buffer tanks and the common reservoir for radiochemical analysis (Sec. 2.2.7) at approximately fortnightly intervals. Finally, periodic measurements of soil oxidation-reduction potential were taken as the sampling and maintenance schedule allowed. In total, approximately 50 000 data were recorded each week during the entire period of operation of each lysimeter experiment.
ch02
May 3, 2007
14:19
22
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
2.2.5.
Lysimeter dosing
2.2.5.1.
Phase I lysimeters
After testing of the lysimeters for leaks and correct operation of instrumentation and control systems the Phase I lysimeters were deemed operational on 23rd April 1990. The entire system of lysimeter substrates (i.e. polythene beads beneath the geotextile layer), buffer tanks and common reservoir was dosed with radionuclides the following day. A cocktail of 22 Na, 137 Cs, 60 Co, 36 Cl and 99 Tc was injected directly into the buffer tanks and common reservoir. Dosing of the lysimeter substrates was achieved by pumping and recirculating the required volume of radionuclide cocktail through the manifold shown in Fig. 2.3. No additions of radionuclides were made to the surface of the lysimeter soils. The system was re-dosed annually to replace radionuclides which had been lost from solution either by biological uptake or by sorption onto physical surfaces within the lysimeter system. Re-dosing was carried out in April or May when a balance had been achieved between rainwater inputs and evapotranspiration losses. The timetable of dosing during the Phase I experiments is shown in Table 2.3 and the total activities of radionuclides added are shown in Table 2.4. The final year of operation of the Phase I lysimeters (1994) involved the use of only one of the shallow lysimeters for an experiment in which the time course of radionuclide uptake into a winter wheat crop was determined from January ’94 to July ’94. The results from this experiment are not described in this book but, for completeness, Table 2.3 includes the time course experiment within the record of operations of the Phase I lysimeters (the results from the time course experiment can be found in Qing (1998)).
2.2.5.2.
Phase II lysimeters
The timetable of dosing during the Phase II experiments is shown in Table 2.5. Dosing of the system was preceded by sowing the lysimeters with ryegrass in May 1995. The cocktail of radionuclides used to dose the Phase II lysimeters consisted of 22 Na, 137 Cs, and 36 Cl and was added in late June 1995, in the same manner as for the Phase I
ch02
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Experimental Protocols Table 2.3.
23
Timetable of key events during the Phase I experiment.
Date (Calendar)
Date (Julian)
09/11/89 23/04/90 24/04/90 26/07/90 21/11/90 24/01/91 16/05/91 14/08/91 28/11/91 12/05/92 27/07/92 30/10/92 04/05/93 03/08/93 27/10/93 03/01/94 04/05/94 19/07/94 03/08/94
89313 90113 90114 90207 90325 91024 91136 91226 91301 92133 92209 92304 93124 93215 93300 94003 94124 94200 94215
Activity Winter wheat sown Lysimeters deemed operational Lysimeters dosed with radionuclides Winter wheat harvested Winter wheat sown Winter wheat re-sown (rodent damage) Lysimeters re-dosed with radionuclides Winter wheat harvested Winter wheat sown Lysimeters re-dosed with radionuclides Winter wheat harvested Winter wheat sown Lysimeters re-dosed with radionuclides Winter wheat harvested Analysis of earthworm population Winter wheat sown in one shallow lysimeter Dosing of shallow lysimeter with radionuclides Final scheduled time course of winter wheat Final harvest of wheat tillers, end of Phase I
Table 2.4. Total activities (MBq) of radionuclides added to the lysimeter system from 1990 to 1993. Radionuclide (MBq) Year
22
Na
137
Cs
60
Co
36
Cl
99
Tc
1990 1991 1992 1993
152.0 44.6 26.9 30.1
133.4 0.0 0.0 0.0
128.0 37.5 29.3 31.7
247 32 29 32
237 32 35 32
Totals
253.6
133.4
226.5
340
336
lysimeters. Since the Phase II experiment was carried out for only one year re-dosing was not necessary. 2.2.6.
Sampling of soil and crop materials
In both Phase I and Phase II experiments, vertical core samples of soil were taken immediately after harvest of the crop each year
ch02
May 3, 2007
14:19
24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste Table 2.5.
Timetable of key events during the Phase II experiment.
Date (Calendar)
Date (Julian)
15/05/95 31/05/95 01/06/95 26/06/95 27/06/95 27–30/06/95 24/07/95 25–27/07/95 21/08/95 24–25/08/95 25/09/95 26–28/09/95 30/10/95 12/04/96 15–17/04/96 18–19/04/96 26/04/96
95135 95151 95152 95177 95178 95178–95181 95205 95206–95208 95233 95236–95237 95268 95269–95271 95303 96103 96106–96108 96109–96110 96117
Activity Ryegrass sown Lysimeters re-seeded and irrigated Hydrological control system operational Ryegrass sward cut Start of Phase II Lysimeters dosed with radionuclides Ryegrass sward cut Soil cores extracted Ryegrass sward cut Soil cores extracted Ryegrass sward cut Soil cores extracted Ryegrass sward cut Ryegrass sward cut Ryegrass sward cut to stubble Soil cores extracted End of Phase II
in late summer. Core samples, of 3.8 cm diameter, were taken with a commercial tube sampler (Eijkelkamp, Agrisearch Equipment, The Netherlands) down to the geotextile layer which marked the base of the soil within the lysimeters. Having removed these samples from the lysimeter, the remaining holes were packed with uncontaminated soil identical to that within the lysimeters. Cores were taken to the laboratory where they were quickly extruded and cut into depthwise segments. The lowest 10 cm sections of cores from both deep and shallow lysimeters were divided into 2 cm segments whereas the remainder of each core was divided into 10 cm segments: this gave 8 depth samples for shallow lysimeters and 11 depth samples for deep lysimeters. Each depth sample was stored in a refrigerator at 4◦ C while awaiting radiochemical analysis. In Phase I, individual wheat plants were harvested by excision of tillers at a point approximately 2 cm above the soil surface and taken to the laboratory where they were air dried and divided into leaves, stems, chaff and grain. From the 1992 harvest onwards, the wheat ear was further divided to give rachis and chaff samples. In Phase II,
ch02
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Experimental Protocols
25
rye grass samples were taken as series of regular ‘cuts’ during the growing seasons of 1995 and 1996 (see Table 2.5). After sampling, all plant materials were weighed and stored in a dry condition awaiting radiochemical analysis.
2.2.7.
Radiochemical analysis
2.2.7.1.
Analysis of gamma-emitting radionuclides (22 Na, 137 Cs, 60 Co)
Gamma-emitting radionuclide activities of the soil, plant and solution samples were determined by high-resolution gamma-ray spectrometry using a semiconductor (GeLi) detector coupled to a multi-channel analyser with commercial data collection and spectral analysis software (Nuclear Data Systems; GammaVision, EG&G Instruments). Count times were adjusted to give 2σ errors of <10% although count errors were usually considerably less than this. Detectors were calibrated with standards made up to the same geometry as samples. Standards comprised 22 Na, 137 Cs and 60 Co in deionised water prepared from non-certified sources which were cross-calibrated against certified sources of 241 Am and 152 Eu to give a final precision of ±10% (radionuclides were obtained from either Amersham (UK) or DuPont (UK)). Since self absorption of gamma rays in standards and samples was small no attempt was made to correct for it within samples or for differential self-absorption between samples. All radionuclide activities within soil and plant samples were calculated as Bq kg−1 dry weight (all corrected for radioactive decay to the date of harvest) and are referred to as ‘activity concentrations’.
2.2.7.2.
Analysis of beta-emitting radionuclides (36 Cl, 99 Tc)
The 36 Cl and 99 Tc activities in soil, plant and solution samples were determined by liquid scintillation counting (LSC) using a 1219 Rackbeta Spectra Master liquid scintillation counter (Wallac/PerkinElmer). The LSC was calibrated to allow dual
ch02
May 3, 2007
14:19
26
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
counting of 36 Cl and 99 Tc. Radionuclides were obtained from either Amersham (UK) or DuPont (USA). Soil Analysis A simple method was adopted to extract both radionuclides from the soil assuming that 36 Cl and 99 Tc were likely to be present predominantly in the Cl− and TcO− 4 forms. A 10 g (fresh weight) sub sample of soil was removed from each depth sample and placed in a 100 ml polythene centrifuge tube. Deionised water (70 ml) was added and the centrifuge tube was shaken for 2 hours using an end-overend shaker. The tubes were then centrifuged for 15 min in a bench centrifuge at 4000 rpm (2.0 × 103 N kg−1 ). Fifty ml subsamples of the supernatant were decanted from each tube and placed in 75 ml polystyrene containers for concentration by freeze-drying. One hundred ml of 1 M NaOH was added to increase the pH of the sample to >9, to prevent any loss of 36 Cl during freeze drying. On reaching dryness, the extracts were made up to a volume of 2 ml with deionised water and transferred to 6 ml centrifuge tubes for a further 15 minute centrifugation at 4000 rpm, to remove any remaining sediment. One ml of the supernatant was then passed down a cation exchange column (1 cm internal diameter, 10 cm length, packed with Dowex 50WX8[H] 100–200 mesh resin) with deionised water as the eluent. This step removed the (cationic) gamma-emitting radionuclides that were also present in the lysimeter soil and would otherwise have interfered with the analysis of 36 Cl and 99 Tc by LSC. Ten ml of eluent were collected from the cation exchange column in a scintillation vial. Twenty ml of 1M NaOH were added to each vial to increase the pH of the sample to a pH of >9. The samples were taken to dryness by freeze drying and then made up to 1 ml with deionised water and prepared for liquid scintillation counting (LSC) by the addition of 10 ml of OptiPhase Hisafe II scintillation fluid (Wallac/PerkinElmer). To correct for any loss of 36 Cl or 99 Tc during the sample preparation procedure, one standard for each soil core was produced from 50 ml of soil water extract and 70 Bq of each radionuclide. This standard was treated in the same way as water extracts taken directly
ch02
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Experimental Protocols
27
from the soil core samples (i.e. subject to the freeze-drying and anion exchange procedures described above). Recovery from freshly spiked soil was 98% (36 Cl) and 95% (99 Tc) using the deionised water extraction method described here. Subsequent analyses have shown that some reduction in recovery of 36 Cl occurs in soils as a result of binding to organic materials (Lee et al., 2001). Plant Analysis To release 36 Cl and 99 Tc from wheat, a 1 g sample was placed in a pear-shaped glass flask with 10 ml of 16 M HNO3 and heated for 3 h from 90 to 120◦ C. The 36 Cl evolved was collected in a series of three absorption traps containing 5 M NaOH. The majority of the 36 Cl was collected in Trap 1, which contained 25 ml of 5 M NaOH at the start of the extraction. Traps 2 and 3 contained 4 ml of 5 M NaOH each. After 3 hours, approximately 75% of the 36 Cl initially present in the wheat had been released from the sample and trapped. Heating was stopped and the splash head and absorption traps were dismantled. The final weight of solution in the traps was recorded before a 500 mg sub-sample was taken and placed in a scintillation vial with 10 ml of OptiPhase Hisafe III scintillation fluid (Wallac/PerkinElmer) for analysis by LSC. The residue from the first step of the extraction, described above, contained 99 Tc and a mixture of the gamma-emitting radionuclides. In the second step of the extraction procedure, complete digestion of the wheat tissue occurred, resulting in a colourless solution that was passed through a cation exchange column to remove gammaemitting radionuclides before beta analysis by LSC. The residue from the first stage was evaporated in the pear-shaped flask at approximately 120◦ C until approximately 5 ml remained. Several additions of nitric acid, deionised water, and hydrogen peroxide were made as follows: 5 ml of 16 M HNO3 , 5 ml of H2 O, 5 ml of H2 O, 5 ml of 16 M HNO3 , 5 ml of H2 O2 , and finally 5 ml of H2 O with the residue heated until only 2 ml of solution remained in the flask. Between each addition, the residue was evaporated until approximately 5 ml remained. The volume was reduced further to approximately 2 ml before the addition of hydrogen peroxide.
ch02
May 3, 2007
14:19
28
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
The residue, in a final volume of 2 ml, was passed through a cation exchange column (1 × 5 cm, packed with Dowex 50W-X8[H-form] 100–200 mesh resin). Fresh resin was used for each residue sample and each was eluted with 20 ml of deionised water. The sample was collected from the column in 2×10 ml aliquots in pre-weighed scintillation vials. The weight of each residue aliquot was determined before a 500 mg sub-sample was taken from each and placed in a separate scintillation vial with 10 ml of OptiPhase Hisafe III scintillation fluid for analysis by LSC. The 99 Tc activities in each of the aliquots were determined separately and then summed to give the total activity in the original sample. The yields of 36 Cl or 99 Tc (70–78%) were determined by the extraction of standards produced from non-contaminated wheat material spiked with known activities of 36 Cl and 99 Tc, extracted under the same conditions as the samples. Radioactivity associated with the radionuclides was calculated as Bq kg−1 dry weight and is referred to as ‘activity concentration’. Solution Analysis Analyses of 36 Cl and 99 Tc in solutions sampled from the buffer tanks, common reservoir and bead substrates were carried out using the same separation and counting techniques as described above for soil samples. However, it was not necessary to apply the initial extraction step using deionised water. 2.2.8.
Earthworm analysis
During the initial filling of the Phase I lysimeters with Silwood soil few earthworms were observed and no quantitative estimates were made of their numbers. One of the criticisms of experiments using re-packed soil is that they lack the structural features such as earthworm burrows, which make the structure of undisturbed soils susceptible to macropore transport. It was therefore decided to determine the extent to which the earthworm population in the lysimeters was representative of a ‘normal’ population expected for this type of soil. Approximately two months after the 1993 harvest the soil from lysimeter G was dug out to a depth of 20 cm. The soil was passed
ch02
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Experimental Protocols
29
through a coarse screen and all earthworms thus separated were collected. A total undisturbed soil volume of 0.324 m3 was sampled and approximately 230 earthworms were recovered (117 earthworms/m2 ), equal to a fresh biomass of approximately 90 g. This compares with mean annual earthworm densities determined in arable soils of 346–471 individuals m−2 with mean biomass of 56.9–61.2 g m−2 (Curry et al., 1995). 2.3. 2.3.1.
Soil Column Experiments Rationale underlying the use of soil columns
In 1996, a soil column system was designed to replace the lysimeters as a cheaper and more flexible alternative. Apart from the obvious benefits of reduced cost, the advantages of using smaller volumes of soil enclosed in columns include the following: • Greater ease of construction. • Increased replication of experimental units (i.e. columns versus lysimeters). • Ability to add radioisotopes singly to individual columns. • Greater control over the environment to which experimental columns are subjected. • Greater ease of in situ sampling, especially pore waters. • Greater ease of destructive sampling, especially in a planned time course. • Much smaller volumes of radioactive waste generated. The soil columns had to meet the same general specifications of the lysimeters in that water table control had to be achieved and maintained and radionuclides had to be supplied to the subsurface, while supporting a crop throughout the duration of an experiment. The timescales envisaged for soil column experiments were shorter than those of the Phase I lysimeters, but a duration of 12 months was expected. In terms of column size, the key dimension was the column length which equates to the depth of a lysimeter. In the lysimeter Phase I experiment, water tables had been maintained at 65 cm or
ch02
May 3, 2007
14:19
30
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste Soil column
Marriotte bottle
Instrumentation: solution samplers Eh electrodes TDR probes 45 cm
Buffer reservoir Adjustable platform
Fig. 2.5.
Water table height can be held constant or adjusted
Design of the experimental soil column system.
35 cm from the soil surface. For soil columns it was decided to adopt a design in which a water table depth in between these two values could be maintained. A water table depth of 45 cm was selected and the column design, described below and shown in Fig. 2.5, was based on this requirement. 2.3.2.
Soil column design and operation
(a) Undisturbed columns The first soil column experiment (Phase III) was conducted using undisturbed soil columns taken from three separate sites in England. Site locations were chosen to give a range of agricultural soils from areas where detailed information on soil structure and site hydrology was readily available. The sites chosen were Silwood Park, Robertgate (Cumbria: N54◦ 26 42 , W3◦ 28 27 ) and Wellesbourne (Warwickshire: N52◦ 11 40 , W1◦ 33 25 ). The final depth of soil used in the column experiment was 50 cm and contained different soil horizons, of variable thickness, depending on location. The Silwood Park soil properties are given in Table 2.1. The Robertgate soil (a brown earth from the Fordham Series) had a more complex vertical profile, with three distinct layers evident in
ch02
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Experimental Protocols
31
some columns. The dark brown topsoil extended to a depth of 30 cm and a sandy clay loam subsoil was present to a depth of 50 cm, on top of a grey stony, sandy clay layer. Large boulders were also evident during extraction of the Robertgate columns. The Wellesbourne soil (a sandy loam from the Wick series) had a very dense appearance in which it was more difficult to establish defined layers. The top 10 cm layer consisted of a brown sandy loam on top of a reddish brown stony sandy loam that extended to a depth of approximately 45 cm, above a light reddish brown stony sandy loam. A summary of the soil properties for the Robertgate and Wellesbourne soils is provided in Table 2.6. To collect an undisturbed soil column from a given location, a square metre was first marked out on the ground. Surface vegetation was cut to the soil surface and a PVC tube (15 cm diameter, 55 cm in length) placed at the centre of the square. The tube was slowly driven vertically into the soil, while the soil from around the outside of the tube was excavated (each extraction was conducted over a period of about 2 hours). When the tube had been driven to a depth of 55 cm, so that the top of the soil column was level with the outside of the tube, a disc of dense polyurethane foam was placed on top of the soil before an aluminium disk was used to seal the top of the column. A second aluminium disc was driven horizontally at the base of the column before the column was removed from the resulting hole in the ground. The column was then inverted to allow a polyurethane foam disk to be inserted under the aluminium disk before it was sealed
Table 2.6. A summary of the soil properties for the Robertgate and Wellesbourne soils.
Soil Type
Organic Soil particle size analysis content by (Bouyoucos hydrometer method) Depth pH loss on (cm) (H2 O) ignition (%) Sand (%) Silt (%) Clay (%)
Robertgate Robertgate Wellesbourne Wellesbourne
0–30 30–50 0–10 10–45
5.0 5.2 5.9 5.9
6.8 2.8 5.3 4.1
78.6 87.7 84.3 79.4
18.6 7.9 12.5 14.8
2.8 4.4 3.2 5.8
ch02
May 3, 2007
14:19
32
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
in position. This process was repeated with columns taken as close as possible to each other. The presence of large obstructions in the soil meant that, in a few cases, a column was abandoned and the tube moved to a new area close to the previous column extractions. Columns were taken from within a small area to reduce variations in soil structure and in the depth of the different soil horizons. The columns were transported to the laboratory in a vertical position within a padded box and care was taken to disturb them as little as possible during transportation. Plastic bead substrate, water table control and instrumentation (as well as sampling and analytical regimes) for the undisturbed Phase III columns were identical to those described below for repacked columns with the exception of the use of hollow fibre solution samplers, which were unavailable at the time Phase III was conducted. (b) Re-packed columns The main experiments comprising Phases IV to VIII were conducted using soil columns re-packed with soil obtained from Silwood Park. These columns were constructed from PVC cylinders 55 cm in length and 15 cm in diameter, to the bases of which 1 cm thick plastic blanking plates were glued with a marine epoxy resin. A 5 cm layer of polythene beads (4 mm diameter) was placed into the base of each column, followed by several layers of nylon mesh (250 µm aperture). For most soil column experiments the same sandy loam topsoil (Wicks series) which was used in the Phase I lysimeters was collected from the Imperial College field station (Silwood Park, Berkshire, UK): the general physical and chemical properties of this soil are shown in Table 2.1. The soil was sieved through a 1 cm2 screen and then packed incrementally into the columns to give a constant dry bulk density of 1.1 g cm−3 . After packing, selected soil columns were instrumented with 1 cm platinum redox electrodes inserted horizontally through holes drilled in the column walls at regular depth increments (the lowermost probe was inserted into the polythene bead substrate). Pairs of TDR probes (see Sec. 2.2.3), for in situ water content measurements, and hollow fibre solution samplers (‘Rhizon’ samplers, Eijelkamp, the Netherlands) were also inserted horizontally
ch02
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Experimental Protocols
33
through pairs of holes at similar depth intervals as the redox electrodes. The columns were placed in a controlled environment room under conditions of 18 hours light (at 20◦ C) and 6 hours darkness (at 15◦ C) per day. The relative humidity within the room was maintained at 70%. Each column was connected via silicone tubing to a polythene reservoir bottle in which the height of water was controlled by a simple siphon apparatus, with additional solution supplied from a Marriotte bottle (Fig. 2.5). The water levels in the column reservoirs controlled the height of the water table within the soil columns. In most experiments the water tables were controlled at a fixed depth of 45 cm from the soil surface. However, a key feature of the soil column design was that it facilitated control over the water table depth by varying the height of the buffer reservoir using an adjustable platform. In selected experiments water tables were subject to a sinusoidal rise in water table height from 45 cm to a maximum of 30 cm below the soil surface at 3 months. This was followed by a fall in water table back to 45 cm depth at 6 months. In all experiments, no additions of irrigation water were made to the surface of the soils. Regular checks were made to determine the levels of radionuclide solution in the Marriotte bottles. These were calibrated and the drop in solution level within a bottle indicated the loss of water through the soil column to the atmosphere via evapotranspiration. This ‘water use’ by each individual column was logged every two or three days to provide a record of the cumulative water flux through each column. After the establishment of the water table, each column was dosed from the base with the required radionuclide. Dosing was carried out by preparing a large volume of single isotope stock solution, at a target concentration of 20 kBq l−1 , which was used to replace the deionised water in the reservoir bottles. The radionuclides were then free to migrate into the soil columns through the plastic bead substrate as evapotranspiration from the column surfaces created a gradient in water potential from the top to the bottom of the columns. This gradient drove the flux of water described in the previous paragraph. After the establishment of the water table, the surfaces of the columns were sown with perennial ryegrass (Lolium perenne L. var.
ch02
May 3, 2007
14:19
34
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Profit) seed at a rate equivalent to 32 kg ha−1 . Routine (twice weekly) measurements were subsequently made of redox potential, soil moisture content (by TDR) and soil solution composition (after sampling with hollow fibre samplers). Cuts of ryegrass shoots were normally made each month. 2.3.3.
Soil column sampling and analysis
After the required period of operation (maximum 12 months) the columns were split vertically and the soil sampled in discrete layers (2 cm layers in the bottom 10 cm of the columns and 10 cm layers in the upper parts of the columns). Single isotopes were used in all experiments using soil columns in order to facilitate analysis. Gamma-emitting radionuclides used were 22 Na, 134 Cs, 125 I, 95m Tc and 75 Se. After sampling, either in solution samples or soil solids, each of these radionuclides was analysed by direct counting using an automated solid scintillation (NaI(Tl)) gamma-ray spectrometry system (Compugamma 1242, LKB Wallac). Appropriate standards were prepared from stock solutions of each radionuclide and these were used to convert counts per minute obtained in different sample matrices into radioactivities in Bq. 36 Cl was the only beta-emitter used in soil column experiments. This was extracted from the soil with either deionised water or 1 M NaOH. Since 36 Cl was the only beta-emitter present (disregarding naturally occurring radionuclides which were present at much lower activities) the radiochemical separation steps which were necessary when analysing lysimeter soils were not required. Thus, 1 ml aliquots of liquid extracts were injected directly into liquid scintillation vials containing 9 ml of Optiphase HiSafe II scintillation cocktail (Wallac/PerkinElmer) and counted for beta activity using a liquid scintillation counter (Rackbeta, LKB Wallac). Care was taken to apply appropriate quench corrections for different extracts using quench curves individually calibrated for specific extract types. Ryegrass samples were dried and either counted directly for gamma activities, or extracted using hot 1 M HNO3 before evaporation, redilution with deionised water and counting for 36 Cl activity by liquid scintillation counting.
ch02
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Experimental Protocols
2.4. 2.4.1.
35
Mini-column Experiments Rationale underlying the use of mini-columns
One of the key parameters obtained from laboratory scale experiments such as the soil column experiments described above is the solid-liquid distribution coefficient (Kd ). In the case of elements such as Cl, with single valence states and limited tendency to interact with the soil solid phase or colloidal components in the solution phase, the Kd is expected to be low and relatively constant. For many of the key elements in radioactive waste safety studies, however, the chemical species and hence the tendency for sorption and other chemical interactions are strongly controlled by factors such as redox potential which, in turn, is primarily controlled by soil water content. The Kd for most elements is notoriously uncertain because it is a ‘lumped’ parameter, resulting from the simultaneous operation of several ill-defined physical and chemical processes. For waste safety calculations it is often treated stochastically and, thus, it is useful to have databases of Kd values which can be interpreted using statistical methods. Even more useful is to have supporting information within those databases which can be used to interpret the response of Kd to key physico-chemical parameters such as redox potential. Such databases are quite rare, since simultaneous measurements of Kd and supporting data are not straightforward in most experimental or natural systems. The mini-column system described below was designed to obtain such measurements using small volumes of soil which can be maintained at a range of realistic water contents. The choice of a small-scale system was intended to facilitate flexibility both in the numbers of soil types which can be investigated and also in the degree to which individual combinations of soil and water content can be replicated in an individual study. 2.4.2.
Mini-column design and operation
The mini-column design is shown in Fig. 2.6. Individual mini-columns were constructed from easily obtainable PVC plumbing components and, essentially, consisted of a 15 × 5 cm tube enclosed at both ends
ch02
May 3, 2007
14:19
36
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Screw cap
50 mm
Adaptor
34 mm
55 mm
Collar 68 mm
50 mm
Pipe 15 cm 5 cm
82 mm
Collar
Adaptor
Screw cap
Fig. 2.6.
Design of a mini-column.
ch02
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Experimental Protocols
37
with tight-fitting caps. The PVC components were assembled using marine epoxy resin, which ensured that all joints and seals were water tight. A 3 mm diameter hole was drilled through the side wall of the column at half the column height. A platinum electrode (redox probe) was made up using a 10 cm length of insulated copper wire attached to a 1.5 cm length of platinum wire by an insulated aluminium crimpconnector. Epoxy resin was used around the exposed ends of the crimp-connector to prevent the inner metal components becoming wet and corroded when placed into the soil. The probe was then placed against the inner wall of the column with the copper wire running through the hole in the column side wall, with the hole around the wire being sealed with epoxy resin. A 3 mm hole was also drilled into the centre of the base of the column (through the screwon end cap). Through this hole a hollow fibre soil moisture sampler (‘Rhizon’, Eijkelkamp, Netherlands) was inserted vertically so that the entire porous section was within the column. This was also glued into place with epoxy resin. A 1 cm hole was drilled into the centre of the top of the column (through the screw-on end cap). Into this hole was fitted and glued a 1 cm diameter tube to allow access for the addition of water and for a reference electrode. Six such columns were constructed. Several soil types were used to pack the mini-columns. These included the same sandy loam topsoil as used in the foregoing lysimeter and soil column experiments (Table 2.1) plus five additional soils collected from south eastern England and north eastern France. The properties of these soils are shown in Table 2.7. For each column, 325 g of sieved (2 mm), but not ground, dry soil was weighed out into plastic bags. The gravimetric moisture contents of three of these soil samples were adjusted to 25% by the addition of deionised water. The remaining three samples were saturated with deionised water; this equated to a gravimetric moisture content of 40%. To each sample, a 1 ml addition of radionuclide-containing solution (1.5 kBq) was made and mixed into the soil by hand. Triplicate samples of soil were taken from each bag for the determination of initial mean soil activity concentration: these analyses indicated that the coefficient of variation between the three samples was less than 10% in each
ch02
May 3, 2007
French#2
French#3
Soil Type
Rendzina with limestone fragments
Colluvium/ alluvium with limestone fragments
Vegetation
Fallow
pH (H2 O) Sand (%)† Silt (%)
†
Clay (%)
†
Organic matter (%) Location †
UK#1
UK#2
Sandy soil with ironstone nodules
Peaty sand
AoH layer from podzol profile
Fallow
Short pasture
Calluna
Pinus
7.7
7.5
5.1
3.7
3.3
41
42
68
95
91
36
38
22
5
6
23
20
10
0
3
2.6
5.9
2.2
17.2
50.1
N48◦ 28 27 E05◦ 21 21
N48◦ 28 15 E05◦ 20 51
N48◦ 35 33 E05◦ 12 29
N51◦ 22 34 W0◦ 38 11
N51◦ 23 25 W0◦ 41 11
Sand, silt and clay values are percentages based on the total mineral content of the soils only.
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
French#1
Biosphere Implications of Deep Disposal of Nuclear Waste
Properties of soils used in mini-column experiments.
14:19
38
Table 2.7.
ch02
May 3, 2007
14:19
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Experimental Protocols
39
case. Therefore, it was assumed that the soil was homogenous with respect to radionuclide activity concentration. The soils were then packed into the columns to a depth of 12.5 cm, producing an equivalent dry bulk density of 1.3 g cm−3 , before the upper end cap was replaced. In the 40% moisture content columns, standing water was present at the soil surface indicating that the soil was saturated. All columns were then weighed and the masses recorded. A rubber bung was placed into the 1 cm diameter tube on the top of the saturated columns in order to limit the entry of atmospheric oxygen to the column since the aim was to produce anoxic conditions in these columns. The columns were then placed within a controlled environment room with a day-time (18 h) temperature of 20◦ C and a night-time (6 h) temperature of 15◦ C. Relative humidity was controlled at 70%. Mini-column experiments were conducted for approximately 50 days. On days 1, 4, 7, 10, 14, 21, 28, 35, 42 and 49 soil solution samples were removed from each column via the hollow fibre samplers by applying a vacuum (−50 to −60 kPa) to the free end of the sampler. This resulted in the slow extraction of a 3 ml sample of soil solution through the sampler, which was retained for analysis. Following the removal of solution samples, columns were re-weighed and deionised water added to the soil to bring the columns back up to their initial mass (i.e. the moisture contents were maintained gravimetrically). In general, only the volume of water removed by the soil solution sampling was required to reinstate the desired soil moisture content, indicating that evaporative losses from the surfaces of the mini-columns were small. On the days on which solution samples were taken, a measurement of the soil redox potential was also taken by inserting a calomel (mercurous chloride) reference electrode into the soil through the 1 cm diameter tube in the top of the columns and connecting it, via a high impedance voltmeter, to the platinum electrode. The mV reading was corrected to account for the electrical potential of the reference electrode (250 mV). Mini-column experiments were carried out using 125 I and 75 Se, both gamma-emitting radionuclides. Soil solutions obtained from the hollow fibre samplers were analysed by direct counting using an automated solid scintillation (NaI(Tl)) gamma-ray spectrometry system
ch02
May 3, 2007
14:19
40
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
(Compugamma 1242, LKB Wallac). The Kd is defined as the ratio of the radionuclide’s activity concentrations in the solid and liquid phases of the soil, as follows: Kd =
Sd Sn
where Kd = solid-liquid distribution coefficient (ml g−1 ) Sd = solid phase concentration (Bq g−1 soil) Sn = solution phase concentration (Bq ml−1 solution) In practice, the total activity of each radionuclide added per gram of soil was known, in addition to the total volume of the soil, its bulk density and its volumetric moisture content. In addition, the activity concentration of the radionuclide within the soil solution was known at the time of sampling of the solutions. The Kd at this time could then be calculated as follows. 1 (V ρT ) − (V θSn ) Kd = Sn Vρ where V = volume of soil (ml) ρ = soil bulk density (g cm−3 ) T = total soil concentration (solid + solution) (Bq g−1 soil) θ = soil volumetric moisture content (cm3 cm−3 )
ch02
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
CHAPTER 3
Modelling Radionuclide Transport and Uptake in Vegetated Soils
3.1.
Introduction
The previous chapter has described the various field and laboratory experiments used to examine the uptake of radionuclides from a contaminated water table by vegetation, namely winter wheat and perennial ryegrass. These experiments varied markedly in scale (spatial and temporal) and complexity. At one end, there were the shortterm laboratory bench mini-column and pot experiments, at the other the long-term field lysimeters. A key requirement, therefore, is a systematic methodology that enables the experimental data to be interpreted in a consistent manner and also allows their implications for any future repository site characterisation and performance assessment programme to be evaluated. This was achieved though the use of physically-based numerical models of water flow and radionuclide partitioning and transport. This chapter, therefore, describes the modelling tools that have been developed as part of the research programme at Imperial. Section 3.2 shows how a systems approach was used to represent the highly complex, interconnected Phase I field lysimeter experiment and in Sec. 3.3 the use of physically-based soil water flow and contaminant transport models for both lysimeter and laboratory column studies is described. Applications of the hydrological models are discussed in Sec 3.4 as a precursor to the detailed studies of radionuclide migration and uptake in following chapters.
41
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
42
3.2.
Biosphere Implications of Deep Disposal of Nuclear Waste
Lysimeter System Model
As discussed in Chapter 2, the lysimeter experiments involved a highly complex system of eight lysimeters interconnected via a common reservoir allowing the addition or removal of radionuclides. The field location of the lysimeters involved exposure to both rainfall and evapotranspiration. An automated control system kept the water pressure at the base of the lysimeter at a prescribed level and provided a means of determining the fluxes of water across this boundary (Fig. 3.1). Whilst the ultimate aim of the experimental work is to achieve an improved understanding of the soil to plant transfer of radionuclides originating from a contaminated water table, these studies are bound up with the radiochemical behaviour in the entire experimental system. The amount of each radionuclide in the lysimeter soil depends on fluxes across the geotextile layer at the base of the lysimeter and, hence, on concentrations in the underlying substrate. These, in turn, depend on the hydraulic and radiochemical behaviour of the water table control system. Therefore, in order to assist in the interpretation of the experimental data, a model of the entire experimental system was developed (Butler & Wheater, 1999a).
Fig. 3.1.
Diagram of lysimeter facility showing water table control system.
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
43
The model was designed to reproduce the essential features of the data using simplified representations of the key physical processes. This approach allowed simulations to be undertaken without undue expense of computational time, whilst providing a powerful tool for the assessment of data consistency and radiochemical inventories. The experimental facility was conceptualised as a collection of compartments representing each of the different system components (i.e. common reservoir, buffer tanks, substrates, and lysimeter soils). These are linked hydraulically by connections conceived as being of zero volume. The model conceptualisation is shown in Fig. 3.2. The following convention is used to define the various model variables: Mx is the total radioactivity (Bq) in component x (where subscript R is the common reservoir, T the buffer tank, S the lysimeter substrate and L the lysimeter soil). An additional subscript i is used to denote individual lysimeters (where i = A, . . . H). Vx and cx represent the volume of (cm3 ) and concentration in (Bq cm−3 ) component x, respectively.
Lysimeter
Contaminated zone Water table
A
E
ML MS
Substrate
MT
Pump
Buffer Tank
QA
QE
B
F
Vd
AR
QB
QF
C
G MR
hR
VR
QC
Common Reservoir
QG
D
H
QD
Fig. 3.2.
QH
Lysimeter system model conceptualisation.
ch03
May 3, 2007
14:24
44
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
The hydraulic characteristics of the lysimeter facility are given in terms of component volumes and fluxes. The common reservoir volume VR is a variable quantity given by the product of the area of the rectangular container AR and the water depth hR . The design of the water table control system is such that the volume of the buffer tanks VT and substrates VS for each lysimeter can be assumed constant. The transfer of water between these components is controlled by peristaltic pumps located between the buffer tanks and the common reservoir. These fluxes (QA , QB , etc.) are obtained from the number of pump revolutions per time step (N ) multiplied by a pumping parameter (p), the volume of water passed per revolution. A more complex issue is the representation of the lysimeter water balance. The total volume of water in the lysimeter soil is given by the integration of the areal water content over the soil depth. Changes in this value occur as a result of a net difference between flow across the geotextile at the base of the lysimeter soil (due to the operation of the water table control system), rainfall recharge at the soil surface and evapotranspiration losses through the surface and plant root system. In order to simplify the model conceptualisation, recognizing that the lysimeter has a fixed near-surface water table, which constrains variation in the total water content, the volume of the lysimeter soil water is also treated as a constant VL . Thus, it is assumed that the water flux across the base of the lysimeter Qi (obtained from the peristaltic pump data) is equal to evapotranspiration minus rainfall. Therefore, as the buffer tank, substrate and lysimeter volumes are constant, the rate of change in the water depth of the common reservoir is given as: H −1 dhR = Qi dt AR
(3.1)
i=A
Thus, the time-varying behaviour of the common reservoir water level in response to the operation of the peristaltic pumps and the occasional dosing/decanting in order to replenish water lost through
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
45
evapotranspiration or gained by rainfall infiltration is: H t −1 · Qi .dt hR (t) = hR0 + AR t =0 i=A D t
+ δ(t − td )Vd .dt d=1
t =0
(3.2)
where hR0 is the initial common reservoir water level, Vd is the dth (of a total D) dosing/decanting volume change at time td and δ is the Dirac delta function. Having formalised the hydraulic behaviour of the experimental system, the associated radiochemical behaviour can be derived. The governing equations are presented in terms of the fundamental property of total activity M , although model outputs are generally expressed as concentrations. The main assumption is that the radiochemical concentration is uniformly distributed throughout the entire water volume of each of the components that comprise the water table control system, allowing each such component to be represented by a single concentration value. It is further assumed that no radiochemical sorption occurs in the common reservoir, buffer tanks and substrates, based on the material screening procedure undertaken during the experimental design to minimise sorption within the control system (Burne et al., 1994). This assumption was found to hold for 22 Na and 36 Cl, however, subsequent studies for other radionuclides (e.g. 125 I, 95m Tc, 75 Se) required the use of partitioning coefficients in order to represent interactions with material components of the water table control system. The balance for each of these components is the storage change due to radionuclides pumped into or out of the component volume along with losses due to radioactive decay or gains from redosing. For example, the balance equation in terms of the rate of change of the total amount of radionuclide stored in the common reservoir is: H dMR =− (Qi .cRT i ) − λMR (3.3) dt i=A
ch03
May 3, 2007
14:24
46
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
where cRT i represents the concentration of the water flowing between the common reservoir and buffer tank. Its specific value depends on the direction of the flow, hence: cRT i = cR cRT i = cT i
if Qi ≥ 0 if Qi < 0
The parameter λ is a decay constant (s−1 ) which accounts for the loss of radionuclide due to radioactive decay. The common reservoir concentration is defined as: cR =
MR MR = VR AR hR
(3.4)
Equation (3.3) can be integrated to give the time-varying common reservoir total activity: MR (t) = MR0 − −λ
t
t =0
t
t =0
H
Qi .cRT i
i=A
MR (t ).dt +
D d=1
.dt
t t =0
δ t − tRd .MRd .dt
(3.5)
where MR0 is the total activity in the common reservoir at time t = 0 and MRd is the activity introduced/removed during the dosing/decanting of the common reservoir at time tRd . The radiochemical balance of the lysimeter soil requires a more complex formulation than the other system components. The approach adopted was to use as simplified a representation as possible, whilst employing assumptions and conceptualisations that are physically reasonable. This presents a challenge, as models of contaminant transport in unsaturated soils have become increasingly complex (Feyen et al., 1998; Lipiec et al., 2003). The approach that was undertaken was to treat the lysimeter in terms of a contaminated soil ‘box’. However, it is unreasonable to assume that radionuclides entering the soil are uniformly distributed over the entire soil volume, as this represents an unrealistic amount of dispersion. Instead, it is
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
47
necessary to define an active volume within the soil where radionuclides are distributed. At the start of the experiment, the soil in the lysimeters is void of radioactivity. Once the system becomes operational, radionuclides are transported into the lysimeters by the water table control system, as a result of evapotranspiration losses. Later in the year there is a release of radionuclides from the lysimeter soils as a result of infiltrating rainwater exceeding soil water storage capacity. Thus, there is a complex set of physical processes whereby radionuclides enter the base of the lysimeter soil from the underlying substrate through the geotextile layer. They subsequently migrate through the soil due to bulk water movement supplemented by the effects of dispersion and diffusion and possibly retarded by sorption reactions with the soil matrix. However, in addition, the upward movement of the radionuclides is opposed by the infiltration of uncontaminated rainwater moving down the soil profile. These processes give rise to a region in the soil where the two sources of water undergo mixing. In order to represent these effects in as simple a manner as possible, the ‘active’ volume is conceptualised as a box. Its base is fixed at the geotextile layer. The top boundary, however, is free to move and represents the height (hLi ) of the active volume above the soil base. It is assumed that the contaminated water does not undergo ‘piston’ displacement by the infiltrating rainwater, but instead, undergoes a mixing process. In order to represent this effect, the active volume is divided into two, upper and lower, zones. The lower zone represents unmixed water which has entered the soil from the lysimeter substrate, whereas the upper represents the active/rainwater mixing zone. The radionuclide concentration in the lower zone is assumed to be a uniform value c∗Li , whereas the concentration profile in the upper zone varies linearly from c∗Li at the lower boundary of the mixing zone (hmi ) to zero at the top boundary (hLi ) (Fig. 3.3). The activity balance for the lysimeter soil can now be formulated in the same way as the other system components. dMLi = Qi .cSLi − λMLi − ULi dt
(3.6)
ch03
14:24
48
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste Soil surface height
May 3, 2007
HL
Upper (mixing) zone
hL
Lower (uniform) zone
hm Lysimeter base 0
Fig. 3.3.
Concentration
c*L
Lysimeter system model conceptualisation.
where cSLi = cSi
if Qi ≥ 0
c∗Li
if Qi < 0
cSLi =
and ULi represents loss of radionuclide from the soil due to crop uptake. In the initial development of the model, it was assumed that there is no interaction of the radionuclide with the various surfaces of the system components. In general, this is not unreasonable, as the materials utilised in the construction of the experimental system were specifically chosen in order to minimise this effect. However, once the radionuclide enters the lysimeter soil, it is almost certain that sorption effects, such as ion exchange with clay minerals present, will occur to a greater or lesser extent. It is assumed that this is a linear equilibrium process and can, therefore, be represented by a single partition coefficient Kd (m3 kg−1 ). Hence, if cLi (h) represents the soil water activity concentration (Bq cm−3 ) at a height h above the lysimeter base, SLi represents the sorbed concentration (Bq kg−1 ), ρs the dry soil bulk density (km3 ) and θ¯ is the mean water content
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
49
of the soil (mm−3 ) then: SLi = Kd cLi
(3.7)
and the total amount of radionuclide in the lysimeter is therefore: hLi ¯ Li + ρs SLi .dh θc (3.8) MLi = AL h=0
where AL is the surface area of the lysimeter (m2 ). Inserting Eq. (3.7) into Eq. (3.8) and assuming that ρs and Kd are constant, gives: hLi cLi .dh (3.9) MLi = AL θ¯ + ρs Kd h=0
Finally, incorporating the definition of the soil concentration profile, produces: (hmi + hLi ) ∗ .cLi (3.10) MLi = AL θ¯ + ρs Kd . 2 Hence, the relationship between the soil water concentration at the base of the lysimeter and the total amount of a radionuclide stored in the lysimeter soil is given by: c∗Li =
MLi Li ) AL θ¯ + ρs Kd . (hmi +h 2
(3.11)
It should be noted that this relationship means that the radionuclide concentration at the base of the lysimeter depends not only on the total amount of activity but also on the heights hmi and hLi . The movements of the two boundaries, which define the two soil contamination zones, reflect their distinctive characteristics. The upper boundary hLi defines the maximum extent to which radioactivity originating from the lysimeter substrate has risen up the soil profile. It is assumed that this boundary can only move vertically up the soil column in response to evapotranspiration. The uni-directional nature of the movement of the boundary reflects the fact that rainfall leaching is not able to remove all of the contamination present, due to activity residing in small and dead end pores (van Genuchten and Wierenga, 1976). The intermediate boundary hmi is allowed to
ch03
May 3, 2007
14:24
50
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
move in either direction, depending on the net water movement in the soil. This boundary, therefore, defines the lower extent of the mixing zone. In both cases the upper and lower physical limits of the boundaries are the soil surface and the geotextile layer, respectively. The boundary movements are governed by the lysimeter’s mean soil water velocity. This is defined as the specific flux divided by the mean water content. The introduction of soil sorption, however, acts to retard radiochemical movement, therefore, the rates of change of the boundaries are given by the retarded velocity vri , which is defined as: vLi (3.12) vri = d 1 + ρsθK ¯ QLi ALi θ¯ (3.13) Therefore, the rates of change of the two boundary positions are: where vLi is the mean water velocity in the lysimeter =
dhLi = vri dt
for vri > 0 and
0 ≤ hLi < HLi ,
otherwise = 0 (3.14)
dhmi = vri for vri > 0 and hmi ≤ HLi , dt and hmi > 0, otherwise = 0
or vri < 0 (3.15)
In order to complete the model a method for representing the uptake of radionuclides by plants is required. It has been found (e.g. Epstein, 1966; Nye & Tinker, 1977; Barber, 1984) from experimental studies of excised roots that the active uptake of various chemicals across the root boundary can be represented by Michaelis-Menten rate kinetics, as follows: Fmax (3.16) .ca F = km + ca where the specific flux across the root surface (F ) is a function of the concentration at the soil/root boundary (ca ) governed by two
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
51
parameters: Fmax the maximum contaminant flux that can enter the root, and km the Michaelis rate constant. Furthermore, work undertaken at Imperial prior to these investigations had shown that the same relationship could be used to describe the uptake of caesium by winter wheat roots (Shaw & Bell, 1989). If ca is km then Eq. (3.16) can be written as a linear relationship in the following form: F = αca
(3.17)
where α (ms−1 ) is referred to as the root sorption coefficient or ‘root absorbing power’ (Nye, 1966). In the model described herein, it is assumed that Eq. (3.17) can be applied to the entire root system using a single root sorption coefficient. In addition, it is assumed that ca is equal to the lysimeter soil concentration cL . Therefore, if the plant root density, in terms of total root length per unit volume of soil (m m−3 ) is given as ρr (z) (where z is the depth (cm) from the soil surface) and taking a as the mean root radius (cm), then the rate of radionuclide uptake per unit volume of soil f (Bq m−3 s−1 ) by the plant root system is given by: f = 2πaρr αcL
(3.18)
If UL is the cumulative uptake of radionuclide by the complete crop root system for an individual lysimeter (Bq), then the rate of uptake at time t (ignoring radioactive decay) is given as: zRmax dUL = AL f.dz (3.19) dt z=0 where zRmax is the maximum rooting depth. In order to utilise Eq. (3.19), an expression for the time development of the crop’s root density profile is required. The root density profile at time t for many monocotyledon crops can be represented by an exponential distribution (Gerwitz & Page, 1974), where ρr = ρr0 (t) e−z/zr (t)
(3.20)
ch03
May 3, 2007
14:24
52
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
where ρr0 (t) is root density at the soil surface (m m−3 ) and zr (t) is the root density attenuation coefficient at time t. As discussed in Chapter 2, borescope measurements of root densities were part of a weekly set of crop observations undertaken at the experimental site (Burne et al., 1994). In order to simplify the analysis of the data, results for each set of replicate lysimeters for the two water table treatments were combined to form averaged values over 10 cm depth intervals. The exponential density profile, described above, was fitted to these data using a simplex algorithm employing a least squares objective function. This resulted in a time-dependent evaluation of the model parameters ρr0 and zr . A visual inspection of these results indicated that the following functions could be used to represent the time-dependent behaviour of these parameters F 1 + Ge−tc /H zr = D(1 − e−tc /E ) A ρr0 = 1 + Be−tc /C
zRmax =
where tc is the crop growth time in days from sowing. Table 3.1 shows sowing and harvest dates for the five sets of lysimeter wheat crops. The fitted parameter values (A-H) are shown in Table 3.2. It can be observed that there is significant inter-annual variability for some of these parameters. This reflects the variations in crop growth that occurred during the experiment. Figure 3.4 shows a typical example Table 3.1. Sowing and harvest dates for winter wheat during the Lysimeter Experiment. Crop
Sowing Date Calendar
1990 1991 1992 1993 1994
12 24 28 30 16
Dec. Jan. Oct. Oct. Dec.
89 91 91 92 93
Harvest Date
Julian
Calendar
Julian
89346 91024 91301 92304 93350
26 Jul. 90 14 Aug. 91 27 Jul. 92 03 Aug. 93 03 Aug. 94
90207 91226 92209 93215 94215
ch03
May 3, 2007 14:24
1990 1991 1992 1993 1994
Depth
Deep Shallow Deep Shallow Deep Shallow Deep Shallow Deep Shallow
Root Growth Model Parameters A
B
C
D
E
F
G
H
3.331 3.841 4.549 5.926 3.681 6.266 6.630 3.209 1.939 2.263
9911 594.5 2260 119.5 149.0 8079 43.18 253.8 2666 9999
13.24 22.13 10.41 20.63 32.01 21.43 42.85 28.13 18.83 5.458
56.26 18.23 27.78 19.02 41.11 12.91 10.79 8.305 402.3 9.177
269.9 49.00 109.3 101.6 262.8 31.35 127.1 2.802 7255 4.431
63.09 32.66 55.39 32.60 56.72 35.97 55.20 26.55 44.66 22.04
141.3 13.00 1673 23840 14.61 0.722 29.79 5.986 72860 431.4
17.02 19.69 7.675 4.861 25.64 124.6 37.44 45.44 11.52 18.89
53
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Year
Root density model parameter values.
Modelling Radionuclide Transport and Uptake in Vegetated Soils
Table 3.2.
ch03
14:24
54
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste 4.5 4.0 0 - 10 cm observed
3.5 Root density (cm cm-3)
May 3, 2007
10 - 20 cm observed
3.0
20 - 30 cm observed 30 - 40 cm observed
2.5
40 - 50 cm observed 50 - 60 cm observed
2.0
0 - 10 cm si mulated
1.5
10 - 20 cm simulated 20 - 30 cm simulated
1.0
30 - 40 cm simulated 40 - 50 cm simulated
0.5
50 - 60 cm simulated
0.0 91050
91100
91150
91200
Harvest
91250
Julian date
Fig. 3.4. season.
Observed and modelled deep lysimeter root densities, 1991 crop
of the performance of the root density model against the observed values. Although the above relationships represent a rather highly parameterised model for general usage, the importance of correctly representing the root density distribution down the soil profile is illustrated in Fig. 3.5. This shows contaminated water rising up the soil, under capillary suction, and progressively encountering the developing root system. The shaded area represents the zone where contaminated water is in contact with the root system. It is root uptake in this region of the soil, integrated over time, which is used to provide the modelled crop uptake at harvest. As the radionuclide concentration profile and root density distribution are analytical functions and assuming (in the first instance) that α and a are constant over the entire depth of the root system, the rate of radionuclide uptake by the entire lysimeter crop, allowing for decay of stored activity, is given as: dUL = 2πAL aαρr0 c∗L {P + Q} − λUL dt
(3.21)
ch03
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
Root density
55
ρr0
Soil surface
Root density profile
height
May 3, 2007
Co nc en t ra tio
Root uptake zone
np
ro f ile
zRmax
Lysimeter base Concentration Fig. 3.5.
Root uptake model conceptualisation.
where
P = zR e−(H−hm )/zR − e−(H−hRmax )/zR
if hm > hRmax ,
otherwise zero, and
Q=
2 zR hL − hm
if hL > hRmax
−(H−hL )/zR
e
and
hm ,
−
zR + hL − hm zR
−(H−hm )/zR
e
otherwise zero.
Equation (3.21) represents the final component of the lysimeter system model and provides a system of equations representing the time-dependent behaviour of every component of the experiment. The final form of the model incorporates a set of 50 state variables, which include contaminated water levels, component activities, and plant uptake. These represent a complete description of the experimental system, and comprise a set of coupled ordinary differential
ch03
May 3, 2007
14:24
56
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
equations. The time-dependent solution of the model was obtained using the Livermore ordinary differential solver for systems with general sparse Jacobian matrices (Hindmarsh, 1983). Using an output time-step of 12 hours this provided a comprehensive set of results that could be assessed against the detailed experimental data collected at the site (see Burne et al., 1994; Wadey et al., 1994). Prior to undertaking model simulations of radionuclide transport and uptake, the hydrological component of the lysimeter system model was assessed. This is important as this component of the model provides the common reservoir’s water volume (the product of the water level height hR [m] and the area AR [m2 ]), as well as the water fluxes between the common reservoir and the lysimeters. These, in turn, affect radionuclide concentrations in the reservoir and hence in the entire system through the periodic addition of specified amounts of radionuclide, and the removal of known volumes of contaminated water, as well as the movement of radionuclides between common reservoir and lysimeters via the buffer tanks. Changes in water level in the common reservoir arise when either known volumes of water are added or removed, in order to prevent the reservoir from drying out or overflowing, or due to the operation of the water table control system. Measurements of the common reservoir water level were taken manually every 2–3 days and provide a set of observations for comparison with the lysimeter system model. The determination of the reservoir water level by the system model involves calculating the peristaltic pump flux Qi [cm3 s−1 ] using the pumping parameter pi for each lysimeter. Although these were periodically measured, effects of experimental error and tube deterioration, coupled with occasional blockages (caused by the production of a flocculent related to the growth of an iron bacterium Gallionella promoted by redox changes within the various components of the water table control system) meant that it was found preferable to obtain them from model calibration. Therefore, representative values for all eight pumps were calculated using observed common reservoir water level changes over 20 day intervals. The time-series plot of simulated and observed water levels is shown in Fig. 3.6. This shows the frequent changes in water level due to the effects of pumping along with the periodic dosing and decanting of the common
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
57
Fig. 3.6. Simulated and observed common reservoir water levels (Lysimeter system model).
reservoir during the summer and winter seasons, respectively. The good agreement between the simulated and observed water levels is demonstrated by a normalised root mean squared error (i.e. RMSE divided by the mean of the observed values (Loague & Green, 1991) of 0.043). Following calibration the hydrological component, the lysimeter system model was then used to simulate the time-dependent behaviour of selected radionuclides in the Phase I experiment. This is reported in subsequent chapters. 3.3. 3.3.1.
Physically-Based Modelling Soil-plant-water model (SPW1)
Typically, assessment models for complex problems such as quantifying the risks associated with radioactive waste disposal or the
ch03
May 3, 2007
14:24
58
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
impacts of licensed and unlicensed discharges of radionuclides to the environment use a compartmental modelling approach. Boxes within the model represent the source, intermediate storage and final sink for radionuclides. Connecting the boxes are the fluxes representing the transport of radionuclides between compartments. Such an approach is generally necessary because of the complexity of the problem, the long simulation timescales required and the need for simulations to be undertaken within a probabilistic framework. However, it is vitally important that such models are underpinned by physically-based mechanistic models, which are able to represent detailed processes and their associated parameter values. These models acts as vehicles for interpreting experimental data and using such information to inform assessment models. The physically-based models described in this chapter are used to interpret the various sets of experimental results described in succeeding chapters. Their physical and mathematical formulations are, in essence, the same as those that have been used to represent subsurface water flow and solute transport in distributed, catchment-scale models, such as SHETRAN (Parkin et al., 2000). Thus, they provide a formal link from the experiment programme to the larger scale catchment scenarios that are, in turn, used to inform the assessment modelling. Assessment work has indicated that the main risk drivers associated with deep geological radioactive waste disposal of low- and intermediate-level wastes are mobile radionuclides with long halflives, e.g. 36 Cl and 129 I (Nirex, 2001). As the transport and uptake of such radionuclides are primarily driven by the soil water flux, two linked models have been employed. A hydrological model, based on the one-dimensional Richards’ equation, has been used to simulate the time-dependent behaviour of soil water potential (equivalent to water pressure). This, in turn, provides vertically distributed soil water contents and fluxes, which are then passed to a contaminant transport model. The fate and transport of the various target radionuclides are then simulated using the advection-dispersion equation, supplemented with additional process representations of sorption, radioactive decay and root uptake. These models are described in more detail below.
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
59
The transport of water through soil, in response to the forces of gravity and water pressure, has been widely represented by the Richards’ equation (Sposito, 1986; Feddes et al., 1988). This relates the change in soil moisture content to the hydraulic flux divergence. It assumes that water flow can be represented by Darcy’s law, whereby the flux is linearly related to the local hydraulic head gradient. The constant of proportionality is termed the hydraulic conductivity K [ms−1 ], which, in turn, has been found to be related to the soil water pressure (or matric) potential ψ [m]. Furthermore, the moisture storage or moisture content θ [m3 m−3 ] is also a function of the matric potential. A mathematical form of the one-dimensional form of Richards’ equation for vertical soil water flows with losses due to water uptake uw [s−1 ] by plant roots is:
D(ψ, z)
∂ψ = ∂t
∂ K(ψ, z) ∂ψ + 1 ∂z ∂z
− uw
(3.22)
where D(ψ, z) =
dθ dψ
(3.23)
The above model requires formal descriptions of the dependence of soil moisture content and hydraulic conductivity on matric potential, and it is these relationships that characterise the hydraulic properties of the modelled soil. There are various methods for obtaining numerical solutions of the above equations. The approach adopted in the model described here is to discretise Eq. (3.22) onto a vertical mesh and thereby reduce the problem to a system of ordinary differential equations that describe the time differentials of the matric potential at each grid node. An automatic variable-order variable-timestep integration routine is then used to obtain the timedependent behaviour of the discretised matric potentials subject to the imposed boundary conditions. The above model has been described in Karavokyris et al. (1990) and the associated numerical code is referred to by the name SPW1 (Tompkins & Butler, 1996a).
ch03
May 3, 2007
14:24
60
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
The root water uptake term uw depends on the plant transpiration rate. In the formulation of the SPW1 model, plant transpiration can be imposed externally through off-line calculations of potential evapotranspiration using energy balance models, such as those of Penman and Penman-Monteith (Shuttleworth, 1979). Alternatively, the model can calculate transpiration fluxes directly from site climate data, such as net radiation Rn , wind speed v, air temperature T and atmospheric humidity e, measured by an Automatic Weather Station (AWS). These are used by the program to calculate water fluxes from the canopy to the free atmosphere via calculation of the canopy temperature and water potential. The canopy water potential (which is related to canopy temperature and stomatal vapour pressure) is also used as a driving force for moving water from the soil into the plant through the root system. Figure 3.7 shows a graphical representation of the model conceptualisation. The basic resistance network used to drive water through the plant is shown to the right. Vegetative, soil and aerodynamic resistance networks mimic the pressure potentials driving water fluxes from the plant roots through the canopy to the
Fig. 3.7.
Soil-plant-water model conceptualisation.
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
61
atmosphere. The resistance networks allow the plant to respond to soil and atmospheric conditions through a variable resistance term, which is analogous to leaf stomatal control. In the centre, the basic structure of the rainfall, evaporation, transpiration and canopy interception processes is shown. The two equations to the left of the figure describe actual evapotranspiration and water uptake by the root system. Obviously, each of these terms has its own associated sub-model. For detailed descriptions of each of these plant model components, refer to Karavokyris et al. (1990), Jackson (2007).
3.3.2.
Solute transport model (SLT1)
Commensurate with the above representation of soil water flow, a similar general representation has been adopted for solute movement through vegetated soils. The general modelling approach for solute migration through porous media is based on the well established theory of the advection-dispersion equation. Although more complex models have been developed, notably those which include the ˇ unek effects of macropore flow (Sim ˙ et al., 2003), such an approach was not considered (or found to be) necessary owing to the focus on upward migration. Furthermore, the approach adopted is well recognised with the international community for modelling radionuclide transport in the biosphere (Elert et al., 1998; Butler et al., 1998). The rate of change of contaminant stored within a unit volume is related to the divergence of an advective flux due to bulk water motion (Jq ) and a dispersive flux (Jd ) resulting from the combination of molecular diffusion and mechanical dispersion. Mechanical dispersion arises from the movement of the contaminant through the tortuous pathways within the soil pore space. The storage of contaminant in the soil is assumed to be partitioned between the dissolved and sorbed phases. The activity of contaminant stored in solution per unit volume is given as θc (where θ is the soil moisture content [m3 m−3 ] and c is the radionuclide concentration in soil water [Bq m−3 ]), whilst the concentration of sorbed contaminant per unit soil mass is given as S [Bq kg−1 ]. The basic model equation, taking into account the losses due to root uptake us [Bq m−3 s−1 ]
ch03
May 3, 2007
14:24
62
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
and radioactive decay Γ [Bq m−3 s−1 ], takes the form: ∂[Jd + Jq ] ∂[θc + S] =− − us − Γ ∂t ∂z
(3.24)
Sorption of contaminant onto the surface of the soil matrix is represented by a linear partition coefficient Kd [m3 kg−1 ]. This effectively lumps all the retention mechanisms into a single value (Ames & Rai, 1978), hence: S = ρb Kd c
(3.25)
where ρb is the dry bulk density [kg m−3 ]. The advective flux Jq is given by the product of the Darcy water flux and solute concentration: Jq = qc
(3.26)
The dispersive flux Jd is assumed to be linearly related to the vertical concentration gradient through the parameter Dh , representing the effects of hydrodynamic dispersion. In addition, allowance is made for the fact that movement does not occur through the entire soil volume but only in the water phase; hence: Jd = −θDh
∂c ∂z
(3.27)
The hydrodynamic dispersion is given by the sum of the soil ∗ [m2 s−1 ], which takes into account molecular diffusion coefficient Dm the tortuous nature of the solute pathways, and the mechanical dispersion coefficient (Dd ); where Dd = dL q/θ, and dL is referred to as the soil dispersivity [m]. Thus: ∗ ∗ + Dd = Dm + dL Dh = Dm
q θ
(3.28)
A methodology for representing the uptake of radionuclides by plant roots has already been described in the context of the lysimeter system model. The same approach is also used here. Hence, the specific
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
63
flux F across the root surface [Bq m−2 s−1 ] is given as: F =
Fmax km + ca
ca
(3.29)
where Fmax [Bq m−2 s−1 ], the maximum solute flux that can enter the root and the Michaelis rate constant, km [Bq m−3 ]. It is then assumed that ca is km , as the molar concentration of radionuclides are extremely long so that the above equation can be simplified to the following form: F = αca
(3.30)
where α is the root sorption coefficient or ‘root sorbing power’ (Nye, 1966), with the units m s−1 . The main assumptions underlying the representation of root uptake of radionuclides made in the SLT1 model are that the root uptake rate per unit volume can be represented using MichaelisMenten rate kinetics and that the root boundary concentrations are sufficiently small for the above linearised equation to be applicable. It was also assumed that the uptake by the entire root system could be represented by the parameter α, i.e. that the uptake efficiencies are identical for any section of the plant root system. Therefore, if the plant root density, in terms of the total root length per unit volume [m m−3 ], is ρr (z), where z is the depth [m] from the soil surface, and the mean root radius [m] is a, then the rate of radionuclide uptake per unit volume us [Bq m−3 s−1 ] by the plant root system is given by: us = 2πaρr F = 2πaρr αca
(3.31)
Because of the difficulty of measuring the root radius, it is combined with the root sorbing parameter to give the root uptake coefficient α = aα. It is further assumed that the radionuclide concentration at the root/soil boundary ca can be represented by the average concentration in the soil c. Hence: us = 2πα ρr c
(3.32)
ch03
May 3, 2007
14:24
64
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Finally, radioactive decay is represented by the decay constant λ [s−1 ] and, as this is relevant to both sorbed and aqueous phases: Γ = λ (θc + S) = λ (θ + ρb Kd ) c
(3.33)
Therefore, the form of the SLT1 solute transport equation used to simulate radionuclide behaviour in the experimental lysimeters is: ∂c ∂ θDh ∂z − qc ∂ [θ + ρb Kd ] c = − 2πα ρr c − λ (θ + ρb Kd ) c (3.34) ∂t ∂z Time-dependent solutions of the above equation are obtained using the same solution procedure adopted for the Soil Plant Water model (SPW1). The solute transport model (SLT1) based on Eq. (3.34), is then driven by the hydrological outputs from the SPW1 model. The current model version is SLT1v3 (Tompkins & Butler, 1996b). 3.4.
Hydrological Model Applications
The following chapters present insights, derived through model simulations, into the processes governing the migration and uptake of various key radionuclides associated with deep underground disposal of radioactive waste. However, before these are discussed, it is important to note that these simulations are based on the input of timedependent vertical profiles of soil water flow and volumetric moisture content, which, in most cases, were provided from simulations using the SPW1 hydrological model. It is, therefore, helpful to consider applications of this model to both field lysimeters and laboratory soil columns. 3.4.1.
Hydrological inputs
Hydrological simulations of the lysimeters require their formal representation by the model. There are four key aspects in this process, which are: discretisation of the soil column; soil hydraulic properties; initial and time-dependent boundary conditions; and the representation of vegetation in terms of crop and root properties.
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
3.4.1.1.
65
Model discretisation
As previously described, the lysimeters and columns had a vertical extent of between 40 and 70 cm. Therefore, in order to provide a relatively detailed simulation of the soil profile without imposing an undue computational burden when undertaking transient simulations, a uniform grid mesh of either 1 or 2 cm spacing was selected. In the case of the Phase II lysimeters, which were all constructed with a soil depth of 70 cm, model simulations were made using a uniform grid mesh of 2 cm. 3.4.1.2.
Soil hydraulic properties
Simulations of unsaturated flow using Richards’ equation (Eq. 3.22) require two constitutive relationships. The soil moisture characteristic, which relates soil water content to matric potential and the unsaturated hydraulic conductivity curve, which expresses the water conductance of the soil as a function of matric potential or soil water content. In many instances (Haverkamp & Vauclin, 1981), these relationships are general functions, which are used to represent a particular soil through a process of parameter identification. For the soils that have been used in the experimental work described here, the moisture characteristic curve suggested by van Genuchten (1980) has been applied. This is a widely used function, whose continuous form makes it well suited for numerical modelling. It is conventionally written as: Se =
−(1−1/n) θ − θr 1 = θs − θr 1 + |αψ|n
for ψ ≤ 0
(3.35)
where Se represent the degree of saturation, θs the saturated moisture content (i.e. the moisture content at ψ = 0), θr is the residual moisture content and represents the soil water which cannot be removed by capillary suction (i.e. as ψ → −∞), α and n are, respectively, scaling and curvature parameters. Fitting the function to a particular soil involves identification of these four parameters from measured soil water contents and potentials.
ch03
May 3, 2007
14:24
66
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
In the case of the Phase II experiments, there were large amounts of both sets of measurements available for characterisation of both the Silwood and Longlands soils. Following a programme of cleaning the data (i.e. removing erroneous values, such as “spikes” associated with tensiometer bleeding) sets of soil moisture contents and matric potentials obtained during the operational period of the lysimeter experiment were combined to form a comprehensive data set. This provided measurements covering a wide portion of the soil moisture characteristic curve, from saturated conditions to relatively dry soil with matric potentials in the range −400 – −800 cm. It was initially assumed that the soil profile in each set of lysimeters for the two soil types could be treated as a homogeneous medium and that the lysimeters in each set would behave in a similar fashion. Equation (3.35) was manually fitted to the observed unsaturated soil moisture contents and matric potentials in order to obtain the soil moisture release curve, which was assumed to be non-hysteretic. While the model is capable of handling hysteresis (with a minor modification), errors within the experimental data and the problem’s inherent nonlinearity make the optimal fitting of even the non-hysteretic case problematic. It was deemed, therefore, that the addition of hysteresis would add unnecessary complexity to the model formulation. Figure 3.8 shows a plot of the cleaned data for one of the Silwood lysimeters (J). There did not appear to be any significant evidence of soil layering in either lysimeter, so one moisture characteristic relationship was fitted to describe hydraulic properties in both lysimeters at all elevations. The cleaned data showed reasonably identifiable limits for a fitted curve, although both lysimeters show only a limited number of points in the –240 to –70 cm matric potential range once the spikes in the tensiometer data are removed. A range of curves following these data was fitted using the van Genuchten relationship, with the uncertainty resolved through observation of the variance in the simulation results (Jackson, 2007). The chosen best fit for the simulations is included against the experimental data in both figures, with parameter values as given in Table 3.3.
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
Fig. 3.8. Table 3.3.
67
Silwood moisture characteristic curve against Lysimeter J data.
Soil moisture characteristic parameter values for lysimeter soils.
Parameter
Saturated moisture content Residual moisture content Scaling factor Curvature factor Saturated slope Sat. hydraulic conductivity Tortuosity factor
Symbol Units
θs θr α n ∆ Ks η
Silwood soil
Longlands Longlands Top Layer Lower Layer
m3 m−3 0.498 0.351 0.541 m3 m−3 0.084 0.025 0.171 2.92 2.88 5.80 m−1 − 2.70 1.78 1.74 10−5 10−5 10−5 m−1 m s−1 4.83 × 10−5 1.01 × 10−4 5.00 × 10−5 − 1.31 −0.42 0.34
Analysis of the data from the lysimeters containing Longlands soil was more difficult than for those with Silwood soil. This was primarily because of the soil becoming much drier during the initial 100-day period. As a consequence, over this period, the observed matric potentials became very unreliable. Following a comprehensive
ch03
May 3, 2007
14:24
68
Fig. 3.9.
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Longlands moisture characteristic curve against Lysimeter M data.
cleaning of the data the results appeared to show two distinguishable curves, indicating a division in the soil between the upper 20 cm and the bottom half metre. Therefore, two fitted curves were employed for the Longlands lysimeters, as demonstrated in Fig. 3.9 and with parameter values as given in Table 3.3. Under saturated conditions the relationship between θ and ψ for both soil types was defined in terms of the constant slope parameter ∆, where: θ = θs + ∆ψ
(3.36)
The parameter values used to define the Silwood and Longlands soils are given in Table 3.3. Having specified parameters for the soil moisture characteristic used by the SPW1 model, the next step was to define a similar relationship for hydraulic conductivity. This is more complex, as direct measurements of unsaturated hydraulic conductivity at different soil water potentials are not available. A generic method for obtaining
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
69
unsaturated hydraulic conductivities from the degree of saturation (Se ) has been developed by Mualem (1976). When applied to the van Genuchten relationship, this gives:
m 2 K = Ks Seη 1 − 1 − Se1/m (3.37) where η is a “tortuosity” factor, which Mualem suggested should be set as 0.5. However, a study by Schaap & Leij (2000) on a wide range of soil samples has shown that there is no benefit in setting η = 0.5 and better results can be obtained by treating it as an arbitrary parameter to be obtained through calibration to observed data. This was the approach adopted for simulating the Phase II lysimeter data. Estimates of saturated hydraulic conductivities for the Silwood Park soil derived from analyses of hydraulic test data gave values in the range of 2.7 × 10−6 − 5.5 × 10−5 m s−1 . Monte Carlo simulations of the Silwood soil data indicated that parameter values near the upper end of this range were more appropriate. The final results were a saturated hydraulic conductivity (Ks ) of 4.8 × 10−5 m s−1 with a tortuosity factor (η) of 1.31. In the case of the Longlands simulations, Ks and η for both layers were found through a similar calibration method. These parameters are also presented in Table 3.3 and the unsaturated hydraulic conductivity curves are compared graphically in Fig. 3.10. 3.4.1.3.
Boundary conditions and sinks
The previous chapter has described how the experimental lysimeter facility was designed to employ a fixed water table at about 5 cm above the soil base. This design criterion thereby defines the lower boundary condition used in the SPW1 model simulation to be a constant pressure head. From an inspection of the matric potential data for the lowest tensiometer, it was observed that the water pressure potential at the base of the soil column was generally about +4.5 cm. This value was used as the matric potential at the bottom boundary of the model. The application of the constant pressure boundary condition also meant that the observed lower boundary fluxes could be used for
ch03
May 3, 2007
14:24
70
Fig. 3.10. lands soils.
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Unsaturated hydraulic conductivity curves for Silwood and Long-
model performance assessment. With regard to these fluxes, the importance of the peristaltic pumping factors (p) has already been noted, as they enable the amount of water entering and leaving the lysimeters to be calculated (Butler & Wheater, 1999a). In the Phase II experiment, these factors were measured on a regular basis, in order to account for changes in the performance of the silicone tubing associated with wear. The upper boundary condition of the hydrological model involves a specified flux condition that represents the net difference between rainfall infiltration and soil evaporation fluxes. These were derived from data collected from an automatic weather station (AWS) situated adjacent to the lysimeters. Rainfall rates were obtained from two tipping bucket raingauges at the experimental site. Transpiration fluxes were calculated internally through the use of the coupled soil and plant representation (Fig. 3.7). Calculations of aerodynamic and stomatal heat and vapour conductances are used to model the movement of heat and water from the crop. Water
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
71
uptake by the root system occurs as a consequence of a potential difference between the plant canopy and soil rhizosphere. Feedback mechanisms, regulated by the soil moisture content and the plant canopy water pressure (hence potential), reduce transpiration rates during dry conditions, thereby helping to maintain the crop’s viability. Soil evaporation rate at the surface was represented using a scheme based on that of Budyko (1974). This allows the actual evaporation to fall below potential once the soil water storage in the surface layer falls below a critical value. For the Silwood soil simulations this was 0.2 m3 m−3 , which was consistent with the value used to simulate the Phase I experiment. In the case of the Longlands simulations, this value was set to 0.3 m3 m−3 .
3.4.1.4.
Crop parameters
The model simulations require inputs of crop height and root density profile for representation of the root uptake fluxes and total transpiration amounts. Crop heights were directly measured and the model used averaged values. Root density data were derived from borescope measurements using the conversion method given in Wadey et al. (1994). The plant sub-model also requires leaf area indices and stomatal conductance values. Owing to the general extent of crop coverage, leaf area indices were not considered to be a sensitive model parameter and were set at a constant value of 5 [m2 m−2 ] for all simulations. The stomatal conductance, conversely, is a very sensitive model parameter, particularly when it comes to adequately representing boundary fluxes. Guidance on this was provided from previous experimental studies of perennial ryegrass grown under controlled conditions which had yielded measurements of leaf conductances ranging between 2.5 × 10−3 (under minimal light and/or limiting soil moisture conditions) to 6.5 × 10−3 m s−1 , which is representative of wellwatered plants with plentiful artificial light (Lawlor and Lake, 1975). SPW1 requires two further parameters to describe leaf conductance, the minimum achievable value (taken directly from the results just given), and a stomatal aperture scaling factor (for details, refer
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
72
Biosphere Implications of Deep Disposal of Nuclear Waste Table 3.4.
Ryegrass crop parameter values.
Parameter Leaf area index Minimum leaf conductance Leaf conductance stomatal aperture scaling factor
Symbol
Units
Value
L
m2 m−2
5.0
glmin
m s−1
2.5 × 10−3
gsf
m3 s−1 W−1
3.74 × 10−4
to Karavokyris et al., 1990 and Jackson, 2007). This is adjusted during the model run by a function accounting for effects caused by soil water deficits (taking a value between 0 and 1), and through an exact proportionality relationship with net radiation. The stomatal aperture scaling factor was therefore calculated by taking the maximum conductance above the minimum achieved by well-watered plants and dividing this by the net energy flux received by the plant. The result is shown in Table 3.4 along with the other ryegrass crop parameter values. 3.4.2.
Results of Phase II lysimeter simulations
In considering the results of model simulations particular consideration is given to the overall performance of the model in reproducing moisture contents and the lower boundary fluxes, as these are subsequently used by the SLT1 solute transport model to simulate radionuclide transport and plant uptake in the lysimeters. In this section, the results of applying the SPW1 model to the Phase II lysimeters will be described as the more comprehensive datasets, for two different soils, provide a greater test for assessing the performance of the hydrological model than in the case of the Phase I lysimeters. 3.4.2.1.
Soil water status: Silwood lysimeters
Simulations of soil matric potential and moisture content for a generic ‘Silwood’ lysimeter are compared with observed values from lysimeters J and K. Plots of these at soil depths of 10, 20 and 30 cm are given in Figs. 3.11 to 3.13. The plots of moisture content are in the
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
73
Fig. 3.11. Simulated and observed hydrological data at 10 cm depth (Silwood lysimeters).
upper half of each figure. These show water content at 10 cm falling to around 10% during the dry period and then rising to between 20– 30% (Fig. 3.11). Although it performs well during the dry period, the model tends to over-estimate the wetter moisture contents. This is primarily due to errors in the moisture characteristic curve and their interaction with the selected hydraulic conductivity relationship. The results at 20 cm depth (Fig. 3.12) show somewhat better agreement, although they do show a slight delay, of a week or so, in the soil wetting up at the end of the prolonged dry period (around 95250). At 30 cm the model simulation shows a more dynamic response that the rather static moisture content values shown by the TDR data (Fig. 3.13). The corresponding plots of matric potential at each depth are in the lower half of each figure. It can be seen that there is quite close agreement between simulated and observed data (particularly when
ch03
May 3, 2007
14:24
74
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 3.12. Simulated and observed hydrological data at 20 cm depth (Silwood lysimeters).
the variability between the two sets of lysimeter data are taken into consideration). They show the development of dry conditions during the first 100 days of the simulation period, and are able to generate the large suctions that occurred during the period 95180–95250, and which are outside the functioning range of the tensiometers. Overall, it is considered that the matric potentials at all levels were reproduced adequately. 3.4.2.2.
Soil water status: Longlands lysimeters
Figures 3.14 to 3.16 show results at the same soil depths for the Longlands lysimeters. Considering first the matric potential plots, owing to cavitation of the water column at tensions of greater 800 cm, the actual soil water potentials at depths of 10 cm and 20 cm during the first three months of the experiment would be expected to be rather more negative than those recorded. By contrast, the values at 30 cm
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
75
Fig. 3.13. Simulated and observed hydrological data at 30 cm depth (Silwood lysimeters).
are at or around the lower limit for tensiometer measurements of −800 cm. Overall, the model handles this period of the experiment well and provides results which are consistent with these effects. During the subsequent wetter period the model reproduces the response of the wetting events reasonably well, but is not able to reproduce the observed fall in matric potential during the intervening drier periods. The corresponding results for soil water content are shown in the upper plots of Figs. 3.14 to 3.16. In general, the model is able to reproduce well the observed soil moisture contents at 10 cm. However at 20 cm depth, although the model is able to capture the dynamics of the observed response well there is a systematic error of +5% in the model simulation compared to the observed during the wetter period (from 95260 onwards). At the 30 cm soil depth there is generally good agreement between experimental and simulated data apart
ch03
May 3, 2007
14:24
76
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 3.14. Simulated and observed hydrological data at 10cm depth (Longlands lysimeters).
from the model’s somewhat exaggerated response during particularly wet events. 3.4.2.3.
Lower boundary fluxes
Figure 3.17 shows a comparison of the simulated lower boundary fluxes against those measured from the four replicate Silwood lysimeters (J, K, N, O). The plot shows the model simulation drawing somewhat less water into the lysimeters than was observed by the four replicates (overall about 25 mm). In addition, the timing of the reversal from predominantly inflow to outflow in the model simulation occurs about 10 days later than is shown by the observed data. The reduced inflow appears to be related to the model underestimating the potential evaporation rate. It is possible this is due to a systematic error in the peristaltic pump calibration factors, since the simulation also underestimates the total outflow. The timing of the
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
77
Fig. 3.15. Simulated and observed hydrological data at 20cm depth (Longlands lysimeters).
reversal in flow is more difficult to reconcile as the rainfall data do not appear to show amounts around the time of 95250 sufficient to cause this change. Consequently, one explanation for this effect is through preferential flow, possibly via macropores that had developed during the preceding dry period. Although there is also much greater variation between individual lysimeters (L, M, O, P), similar effects to those described above can be seen in the Longlands results (Fig. 3.18). Overall, the model performance against the data is less successful than that for the Silwood lysimeters. In particular, the low soil hydraulic conductivities required to reproduce the very dry soil conditions during the first hundred days of the experiment (and hence the reduced actual evapotranspiration rates) also result in the gap in the timing of the inflow reversal between the model and data being even larger than before.
ch03
May 3, 2007
14:24
78
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 3.16. Simulated and observed hydrological data at 30cm depth (Longlands lysimeters).
These results (both fluxes and soil water status), therefore, highlight the challenges posed when modelling unsaturated flow, particularly when constrained by such a highly detailed and comprehensive dataset. 3.4.3.
Laboratory column simulations
The previous section has provided a detailed description of the hydrological modelling of the Phase II lysimeters and has demonstrated the model’s ability to represent the response of a vegetated soil under semi-natural conditions. In contrast, modelling the response of the laboratory soil columns was far less demanding, as two components of the experimental design markedly constrained hydrologic behaviour. First, the columns were situated in an environmentally controlled growth room, in which temperature, humidity and “day light” length
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
79
Fig. 3.17. Simulated and observed cumulative lower boundary fluxes, Silwood soil.
and intensity were fixed. Second, no additions of water, apart from minor amounts to ensure crop survival, were made to the surface of the columns. This meant that, apart from the first two to three weeks when the columns “wetted up”, the columns were at quasi-steadystate conditions. The only perturbations to this status were either deliberate manipulations of the lower boundary condition (i.e. a controlled change in the water table elevation) or as a result of changes in the condition of the ryegrass sward. During the course of the various experimental phases, slightly different approaches were developed to handle the hydrological conditions of the columns and their input into the solute transport modelling. These are summarised below. The Phase III experiment involved undisturbed columns of three different soil types. Given the heterogeneities in the soil structure and the highly limited information on the soil moisture characteristic due to the imposed steady-state conditions, the required soil
ch03
May 3, 2007
14:24
80
Fig. 3.18. lands soil.
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Simulated and observed cumulative lower boundary fluxes, Long-
moisture contents and water fluxes were derived directly from the observed data (Wadey et al., 2001). Figure 3.19 shows the cumulative water influxes for the six month columns and highlights the variability between replicate columns and the three soil types. A similar approach was also adopted in the Phase IV experiment, which investigated the effect of stable chloride soil concentrations on 36 Cl uptake in re-constituted Silwood soil (Butler et al., 2001). However, in this case rather than simulating each individual column, the hydrological behaviour of the columns was divided into two groups according to the magnitude of the water fluxes. Generic representations of the hydraulic components were then derived for these ‘low’ and ‘high’ flux states. The Phase V experiment focused on the behaviour of radioiodine. Unlike chlorine, iodine is affected by redox conditions. In addition, it was found to interact with various components of the experimental system. These effects meant that the model domain
ch03
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
Silwood soil columns
Water influx( ml)
10000 8000 6M01 6000
6M02
4000
6M03 6M04
2000 0 0
50
100
150
200
Day of Experiment
Water influx( ml)
Robertgate soil columns 4500 4000 3500 3000 2500 2000 1500 1000 500 0
6M05 6M06 6M07 6M08
0
50
100
150
200
Day of Experiment
Wellesbourne soil columns 3500 3000 Water influx( ml)
May 3, 2007
2500
6M09
2000
6M10 `
1500
6M11 6M12
1000 500 0 0
50
100
150
200
Day of Experiment
Fig. 3.19.
Cumulative water influx into 6 month soil columns.
81
ch03
May 3, 2007
14:24
82
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Ryegrass
Soil
Water table elevation
Substrate 1
Substrate 2
Boundary contaminant concentration
Fig. 3.20.
Representation of experimental design used in column simulations.
had to be extended to include both the 5 cm thick bead substrate, which underlay the soil column, and the external water level control system (Figure 3.20). Given this more complex representation, the SPW1 model was used to simulate the hydrological response of the system (Ashworth et al., 2002). The soil column was discretised using 1 cm grid blocks and additional 5 cm grid blocks were used to represent the underlying bead substrate and control reservoir. Soil
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
83
Fig. 3.21. Simulated and observed final moisture contents and cumulative water influx — 12 month columns.
hydraulic parameters were then derived for each group of soil columns (3, 6, 9 and 12 months). Figure 3.21 provides an example of model performance for the 12 month columns. The moisture content time series shows the initial wetting up of the soil column following establishment of the fixed water table and the model’s agreement with the final moisture content distribution obtained from gravimetric analysis. In addition, the plot shows how water influx to the columns can vary over time, in spite of constant climatic conditions. In this case, the progressive decline was related to the deterioration in the ryegrass sward due to prolonged dry conditions in the top soil layer. Simulations of Phases VI, VII and VIII used the same representation of the column experimental set up as just described for Phase V and with similar results. The only difference is that some of the columns in the Phase VI and VIII experiments included a sinusoidal variation
ch03
May 3, 2007
14:24
84
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 3.22. Simulated and observed moisture contents for Phase VI 6 month column with varying water table.
in the water table elevation, which rose and fell by 15 cm over a sixmonth period. In these cases, a clear hysteresis was observed in the moisture content data. However, as the SPW1 model does not include hysteresis and a formal representation of this effect would have been extremely difficult to quantify given the experimental data, these columns were simulated by allowing the lower boundary condition to follow the rise in water table until the maximum elevation was reached, at which point it remained fixed. The increased wetting of the column as a result of the variation in water table elevation can be seen in the results for one of the Phase VI columns given in Fig. 3.22. 3.5.
Discussion
Soil water flow in unsaturated soils is a highly non-linear problem. The challenge in modelling these conditions is compounded when
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Modelling Radionuclide Transport and Uptake in Vegetated Soils
85
there are large amounts of high quality data, as is the case with the lysimeter experiments. The imposed specified flux boundary condition at the upper surface coupled with the detailed measurement of fluxes across the lower boundary present a high degree of constraint on the model. In addition, the existence of high resolution matric potential and moisture content data means that the moisture characteristic curve is also highly constrained. Thus, the main degrees of freedom are the residual hydraulic conductivity values and the stomatal conductances. Overall the SPW1 model performed well in reproducing the various components of the soil water data, particularly when using the coupled soil/plant version. Systematic errors in reproduced soil moisture contents demonstrated the non-linearity of dependence between non-predetermined hydrological parameters, and show the need for careful optimisation procedures. The lower boundary flux response for the Longlands lysimeters showed a lag compared to the observed data. Nevertheless, it was still able to reproduce the overall observed influxes and effluxes, which is the most important aspect for the radionuclide simulations. The simulations of the laboratory soil columns were, by contrast, more straight-forward. However, the highly controlled experimental conditions and their quasi-steady-state meant that characterising the hydraulic properties was more problematic. In the Phase III and IV experiments, hydrological modelling was dispensed with and moisture contents and soil water fluxes were derived directly from the experimental data in order to drive the solute transport simulations. However, in subsequent column experimental phases where radionuclides interacted with components of the experimental system and where, in some cases, time-varying lower boundary conditions were introduced, simulations using the SPW1 model proved necessary for providing the required inputs for the associated solute transport simulations.
ch03
May 3, 2007
14:24
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
This page intentionally left blank
ch03
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
SECTION 3
RESULTS
87
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
This page intentionally left blank
88
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
CHAPTER 4
Radiochlorine
4.1. 4.1.1.
Background 36 Cl
in radioactive waste
Chlorine has two stable isotopes, 35 Cl and 37 Cl. Of the many radioactive isotopes, 36 Cl is the most significant in relation to the disposal of radioactive waste. 36 Cl is an important component of intermediate level radioactive waste, with its primary source in such waste being the operation of nuclear power plants. It is formed within the reactor core by neutron capture by stable 35 Cl which may be present at trace levels (i.e. as an impurity) in various irradiated materials (Sheppard et al., 1996). Radionuclides of primary interest with regard to radioactive waste disposal are those that have a long physical half life and a potentially high degree of environmental mobility. 36 Cl, which is predominantly anionic and has a physical (radioactive) half-life of 3.01 × 105 years fulfils these criteria.
4.1.2.
Chlorine behaviour in soils and plants
Chlorine behaviour in soils is dominated by its presence as the inorganic chloride form (Brady & Weil, 1996). This conservative (nonsorbed) form is often assumed to follow hydrological flows almost quantitatively and is, therefore, routinely used as a tracer for the movement of water through, for example, soil columns, natural soils in the field, and aquifers. Furthermore, 36 Cl was shown by 89
ch04
May 3, 2007
14:33
90
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Ogard et al. (1988) to break through to the surface environment following underground nuclear testing slightly before tritiated water, suggesting that the migration of radiochlorine may be enhanced due to anion exclusion (i.e. inability to enter the less mobile parts of the pore space due to electrical repulsion, Thomas & Swoboda, 1970). This high affinity for the water phase of soils results in the solidliquid distribution coefficient (Kd value) of chlorine generally being assumed to be zero. In addition, it also renders chlorine biologically available. 36 Cl is expected to be one of the most mobile and biologically available of the radionuclides that are important in the context of radioactive waste disposal. Chemical species of chlorine in soils other than chloride are not often reported. However, a small but significant body of work, initiated by Apslund & Grimvall (1991), suggests a role for natural organic matter in the formation of organo-chlorine species within ¨ soils. Oberg (1998) specifies two processes by which the formation of chlorinated soil organic matter may come about: microbial formation of specific chloro-organics which become incorporated into soil organic matter, and microbially induced formation of reactive chlorine which then reacts with functional groups on molecules of soil organic matter. Specific associations of 36 Cl with humic materials extracted from soil with water and sodium hydroxide were identified by Lee et al. (2001), using gel filtration chromatography. Clearly, therefore, processes exist whereby soil chlorine may not be present solely as the chloride form. Since chlorine associated with soil organic matter is likely to exist primarily in the soil solid phase, such conditions would seem likely to result in a Kd value of chlorine that is greater than zero, although the concept of Kd may not be appropriate, as it presumes reversible sorption, whereas incorporation in organic matter may not be reversible.
4.2.
Experimental Overview
Experiments with 36 Cl began with two outdoor lysimeter experiments (Phases I and II) followed by three column experiments
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
91
(Phases III, IV and VII). The experiments looked at the effects of differing soil types, disturbed and undisturbed soil columns, differing levels of stable chlorine, and differing water table levels, on the soil-plant behaviour of 36 Cl. The details of the various experiments are summarised in Table 4.1. The general set up of the lysimeter and soil column experiments is given in Chapter 2. In the first lysimeter experiment (Phase I), 36 Cl was added to the base of Silwood soil (characteristics shown in Table 2.1) lysimeters in a cocktail with a number of other radionuclides. Both ‘deep’ (70 cm) and ‘shallow’ (40 cm) lysimeters were used, and all were cropped with winter wheat. The full experiment lasted from 1990 to 1993, although some further data were collected in 1994. A new radionuclide dosing was carried out in April or May each year. In the first year (1990), this addition of 36 Cl was much greater than in subsequent years in order to initially establish relatively high activity concentrations within the system (Table 4.2). Soil and wheat samples were collected at the end of each growing season. The second lysimeter experiment Table 4.1. Summary of the various programme. Phase Phase Phase Phase Phase Phase
Duration I II III IV VII
4 1 6 9 6
years year months months months
Table 4.2.
36
Cl experiments undertaken in the Description
Deep and shallow lysimeters, Silwood soil Deep lysimeters, Silwood and Longlands soil Soil columns, Silwood, Robertgate, Wellesbourne soils Soil columns, Silwood soil, Stable Cl treatments Soil columns, Silwood soil, Fixed and fluctuating water tables
Activities of
36
Cl (MBq) added to experimental system.
Year Common Reservoir Buffer Tanks Lysimeter Substrates Total 1990 1991 1992 1993 1994
121.73 13.19 12.40 9.76 14.66
9.03 1.63 1.44 1.92 1.44
116.40 17.00 15.00 20.00 15.00
247.16 31.82 28.84 31.68 31.10
ch04
May 3, 2007
14:33
92
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
(Phase II) was similar to the first, although only deep lysimeters were used and the lysimeters were cropped with perennial ryegrass. Silwood soil and Longlands soil (characteristics shown in Table 2.1) were investigated. The lysimeters were dosed once with 36 Cl (together with 137 Cs and 22 Na) at the end of June 1995. Soil sampling was carried out at the end of July, August and September 1995, and mid-April 1996. Ryegrass sampling was carried out at the end of July, August, September and October 1995, and mid-April 1996. Further details of the lysimeter experiments are given in Chapter 2. The first column experiment (Phase III) used undisturbed soil columns of three soil types, Silwood, Wellesbourne and Robertgate (characteristics of latter two shown in Table 2.6). All columns were vegetated with perennial ryegrass. A 36 Cl-contaminated water table, fixed at 45 cm below the soil surface, was imposed on the columns. Sampling of the soil was carried out after 1, 3 and 6 months. Ryegrass was sampled monthly. The second column experiment (Phase IV) used re-packed soil columns of Silwood soil, again with perennial ryegrass cover and a 36 Cl-contaminated water table fixed at 45 cm from the soil surface. However, in this experiment, the effect of stable chloride on the behaviour of 36 Cl was investigated. Therefore, three stable chloride treatments were established: unamended (around 1 mmol Cl kg−1 ), 5 mmol kg−1 and 25 mmol kg−1 . Sampling of soil was carried out at 3 and 9 months and sampling of the ryegrass at monthly intervals. The final column experiment (Phase VII) again used repacked columns of Silwood soil. Fixed (at 45 cm from the soil surface) and fluctuating water tables containing 36 Cl were compared. The height of the fluctuating water tables was increased in increments of 0.5 cm from an initial soil depth of 45 cm up to 30 cm depth over the first 3 months of the experiment, and then decreased to 45 cm depth over the latter 3 months. The days on which 0.5 cm adjustments were required in order for the water table height to follow a sinusoidal curve were pre-determined. Sampling of the soil was carried out at 2, 4 and 6 months and ryegrass sampling monthly. With a significant inventory in radioactive waste and a long physical half life, investigation of the behaviour of 36 Cl is important in order to assess, as noted above, the potential risk associated with its
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
93
disposal. Its very low Kd , compared with other radionuclides, suggests a high potential mobility and makes a quantification of 36 Cl soil-plant behaviour especially important. In addition, little previous work has studied the behaviour of this radionuclide. Some discussion on its behaviour is given by Coughtrey et al. (1983), however, the IAEA (1994) compendium data for a wide range of radionuclides does not provide either Kd values or transfer factors for 36 Cl. 4.3.
Results
4.3.1.
Four year lysimeter experiment (Phase 1)
4.3.1.1.
Lysimeter soil profiles of
36Cl
The mean soil activity profiles of 36 Cl for the 1990 to 1993 destructive samplings of both deep and shallow lysimeters are shown in Figs. 4.1 and 4.2 respectively. In each year, 36 Cl was present throughout the soil profile. The nature of 36 Cl migration through the soil can be best seen by considering the data for 1990 when the behaviour of the first spike of 36 Cl can be considered without the complication of 36 Cl from previous years. In this year, the lysimeters were dosed 40 Height above geotex (cm)
May 3, 2007
1990
35
1991
30
1992 1993
25 20 15 10 5
Shallow
0 0
2
4
6
8
Soil activity concentration (kBq 36
10
12
kg−1)
Fig. 4.1. Mean soil Cl activity concentrations in the Phase I shallow lysimeters (1990–1993). From Shaw et al. (2004).
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
94
Biosphere Implications of Deep Disposal of Nuclear Waste 70 65
Height above geotex (cm)
May 3, 2007
1990
60 55
1991
50 45
1993
1992
40 35 30 25 20 15 10 Deep 5 0 1.00E-03 1.00E-02
1.00E-01 1.00E+00 1.00E+01 1.00E+02
Soil activity concentration (kBq kg−1)
Fig. 4.2. Mean soil 36 Cl activity concentrations in the Phase I deep lysimeters (1990–1993). Note log scale. From Shaw et al. (2004).
with 36 Cl at the end of April and sampled at the end of July. Thus, over the three months of the 1990 study, it was evident that 36 Cl had migrated through the full depth of both types of lysimeter, i.e. up to 70 cm. This equates to a migration rate of around 0.77 cm day−1 , although the rate may have actually been greater because migration of the radionuclide was limited by the extent of the soil. Activity concentrations of the soil tended to decline over time. In the shallow lysimeters, the values ranged from 4 to 10 kBq kg−1 in 1990, and from 1 to 3 kBq kg−1 in 1993. In 1990, the bulk of the activity was found in the 0–25 cm (above geotex) layer of soil. However, in subsequent years a fairly uniform activity profile was observed throughout the full depth of the shallow lysimeters. In the deep lysimeters, declines in activity concentrations were again generally observed from year to year. The activity profiles were non-uniform, with activity concentrations tending to decline with decreasing soil depth. This was particularly noticeable in the 1993 sampling, where, at the bottom of the lysimeter, the activity concentration was around
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
95
3 kBq kg−1 but declined to around 0.01 kBq kg−1 in the 40–60 cm (above geotex) zone. Mean water inflow to the lysimeters over the course of the experiment is shown in Fig. 4.3. These data indicate the net flux of water movement through the lysimeter soil, i.e. whether the predominant water movement was in an upwards or downwards direction. Clearly, this is driven by environmental variables; essentially rainfall (downwards flux) and evapotranspiration (upwards flux). Measures of these values are shown in Fig. 4.4. During the spring and summer of 1990 (the time when the lysimeters were first dosed and sampled) the net flux was positive (upwards), due to low rainfall and a high evapotranspiration flux brought about by relatively high air temperatures (temperature data not shown). This explains the observed upward migration of 36 Cl over this period, since the added 36 Cl is expected
'Deep'
'Shallow'
200
100
0
-100
-200
-300
1990
1991
1992
1993
1994
-400
Sp Su ring m m Au er tu m W n in te Sp r rin Su g m m Au er tu m W n in te Sp r r Su ing m m Au er tu m W n in te Sp r r Su ing m m Au er tu m W n in te Sp r rin g
Water Inflow Through Lysimeter Base (mm)
May 3, 2007
Fig. 4.3. Net water fluxes across the bases of the Phase I deep and shallow lysimeters in each season from spring 1990 to spring 1994. From Wadey et al. (2001).
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
96
Biosphere Implications of Deep Disposal of Nuclear Waste
Evapn. (deep)
Evapotranspiration or Rainfall (mm)
Evapn. (shallow)
Rainfall
350
1990
1992
1991
1993
300 250 200 150 100 50 0
Spring
Winter
Autumn
Spring
Summer
Winter
Autumn
Spring
Summer
Winter
Autumn
Summer
Spring
Winter
Autumn
Spring
-50
Summer
May 3, 2007
Fig. 4.4. Rainfall and evapotranspiration from the Phase I deep and shallow lysimeters in each season from spring 1990 to spring 1994 (absolute amounts per quarter). From Wadey et al. (2001).
to remain as the non-sorbed chloride ion and hence behave conservatively in the soil. During the period from the 1991 dosing to the 1993 sampling, mean water inflow to the lysimeters was negative. Although the presented data do not show shorter time periods over which an upward flux would have occurred, this indicates that the net flux in water movement within the lysimeters was downward during the 1991, 1992 and 1993 migration studies. Given the apparent dependency of 36 Cl behaviour on moisture flux, it would be expected that net upward migration of the 36 Cl in these years would be low. Because of the presence of activity from the 1990 study, however, determining the behaviour of 36 Cl added in the three subsequent years is difficult. Nevertheless, it seems clear that a significant degree of upward migration did not take place in these years since, in all instances, the activity concentrations of the soil decreased from the 1990 levels.
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
97
The apparent dependency of 36 Cl behaviour on the flux of water through the lysimeters indicates a possible reason for the observed annual decreases in soil 36 Cl activity concentrations. These may have been due to a leaching of 36 Cl out through the bottom of the lysimeters as a result of the net downward fluxes for each year after 1990. Nevertheless, the fact that substantial quantities of 36 Cl were found throughout the soil profile at all sampling times, suggests that this leaching occurred only for a certain fraction of the radionuclide. This, in turn, may suggest that some fraction of the 36 Cl present in the soil might have been bound in some way by the soil to prevent it from leaching. A further explanation which may account for the decreases in soil activity concentrations over time, particularly in the upper region of the lysimeters, may have been uptake of the 36 Cl into the wheat crop, i.e. depletion of the soil activity, rather than leaching. 4.3.1.2.
Wheat biomass and activity concentrations
Mean biomass production for the various parts of the wheat crop grown on the shallow and deep lysimeters is shown in Figs. 4.5 and 4.6, respectively. Data for the 1992 harvest often gave the greatest biomass yields, whereas the 1993 harvest gave the lowest yields. Overall, the annual biomass yields generally seemed to decrease in 350 1990
300
1991
Wheat biomass (g)
May 3, 2007
250
1992 1993
200 150 100 50 0 Grain
Chaff
Leaf
Stem
Rachis
Fig. 4.5. Mean wheat biomass production from the Phase I shallow lysimeters (1990–1993).
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
98
Biosphere Implications of Deep Disposal of Nuclear Waste 600 1990 Ryegrass biomass (g)
500
1991 1992
400
1993 300 200 100 0 Grain
Chaff
Leaf
Stem
Rachis
Fig. 4.6. Mean wheat biomass production from the Phase I deep lysimeters (1990–1993).
1600 Wheat activity concentration (Bq g-1)
May 3, 2007
1990 1400
1991
1200
1992
1000
1993
800 600 400 200 0 Grain
Chaff
Leaf
Stem
Rachis
Fig. 4.7. Mean 36 Cl activity concentrations of wheat from the Phase I shallow lysimeters (1990–1993).
the order: 1992 > 1990 > 1991 > 1993. In relation to the biomass of the various plant parts, the grain and stems generally gave greater biomass yields than the chaff and leaf portions. The mean activity concentrations of 36 Cl in the harvested biomass portions are shown in Figs. 4.7 and 4.8. In the shallow lysimeters, the activity concentrations of all parts were by far greatest in the first year of the experiment (1990). Values in 1991, 1992 and 1993
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
Wheat activity concentration (Bq g-1)
May 3, 2007
99
10000
1990 1991
1000
1992 100
1993
10 1 0.1 0.01 Grain
Chaff
Leaf
Stem
Rachis
36
Fig. 4.8. Mean Cl activity concentrations of wheat from the Phase I deep lysimeters (1990–1993). Note log scale.
were broadly similar, although 1992 consistently gave the lowest values. In most cases, activity concentrations of the various plant parts within each year increased in the order grain < chaff < leaf < stem. In the deep lysimeters, highest activity concentrations were found in 1993 and these were higher than those found in any year for the shallow lysimeters. Values in 1990 and 1992 were comparable to those obtained for the shallow lysimeters in those years, whereas the values for 1991 were markedly lower than those for the shallow lysimeters. In 1990 and 1993, activity concentrations of the various wheat parts increased in the same order as for the shallow lysimeters. However, in 1991 and 1992, such a clear trend was not observed. In a simple, concentration ratio, interpretation, a direct relationship between the annual soil and plant activity concentrations would be expected. Essentially, higher soil activity concentrations, particularly in the upper soil zone where roots predominate, should produce higher plant activity concentrations, particularly for a conservative radionuclide such as 36 Cl. Indeed, in the shallow lysimeters, highest activity concentrations in both the soil and wheat were observed in the first year of the experiment. However, a simple relationship between soil and plant activity concentrations may not be expected, due to the influence of other variables, most notably the development of the plant. Although no relationship between wheat biomass and wheat activity concentration was observed, the distribution of roots
ch04
May 3, 2007
14:33
100
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
in relation to the soil contamination is likely to have had a significant impact on the wheat uptake of 36 Cl from the soil. To take this into account, weighted transfer factors were calculated. These calculations used a mean soil activity concentration weighted according to the fractional abundance of roots and measured soil activity concentrations in discrete layers of the soil profile, i.e. a Weighted Mean Activity Concentration (WMAC). The calculation of the WMAC and its application in the transfer factor is shown in Eqs. (4.1) and (4.2). TFw = [R]plant /WMAC
(4.1)
where TFw is the weighted soil-plant transfer factor (dimensionless) [R]plant is the activity concentration of the radionuclide within the plant tissue at harvest (Bq kg−1 ) WMAC is the weighted mean activity concentration of the soil (Bq kg−1 ) WMAC =
n
([R]i fi )
(4.2)
i=1
where WMAC is the weighted mean activity concentration of the soil (Bq kg−1 ), [R]i is the radionuclide activity concentration in the ith soil layer (Bq kg−1 ), fi is the fractional abundance of plant roots in the ith soil layer (dimensionless). The transfer factor values are shown in Figs. 4.9 and 4.10 and again illustrate the enhancement in 1990 (for deep and, particularly, shallow lysimeters), and 1993 (for deep lysimeters) in the transfer of activity from the soil to the plant. The potential for transfer of activity from the soil to the plant is driven by transpiration. If evapotranspiration (Fig. 4.4) is used as an indicator of this, it is clear that the relatively high evapotranspiration in the shallow and deep lysimeters in 1990 provides a good explanation of the observed enhanced degree of uptake in both types of lysimeter in this year. In other
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
101 1990
10000
1991 1992 1993
Transfer factor
1000
100
10
1 Grain
Chaff
Leaf
Stem
Rachis
Fig. 4.9. Weighted soil-wheat transfer factors (dry weight basis) for the Phase I shallow lysimeters (1990–1993). Note log scale.
10000
1990 1991
1000 Transfer factor
May 3, 2007
1992 1993
100 10 1 0.1 0.01 Grain
Chaff
Leaf
Stem
Rachis
Fig. 4.10. Weighted soil-wheat transfer factors (dry weight basis) for the Phase I deep lysimeters (1990–1993). Note log scale.
cases, the lower evapotranspiration fluxes are consistent with the lower soil-plant activity transfers in 1991, 1992 (for shallow and deep lysimeters) and 1993 (shallow lysimeters). However, this still leaves the relatively high transfer factors in the deep lysimeters in 1993 requiring an explanation that is not dependent upon a high evapotranspiration flux. A possible explanation is that the wheat roots were more effectively exploiting the contaminated soil at this time.
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
102
Biosphere Implications of Deep Disposal of Nuclear Waste 40 35 30 Height above geotex (cm)
May 3, 2007
25 20 15 1990 10
1991 1992
5
1993
0 0
1
2
3
4
5
Root density (cm3 cm-3)
Fig. 4.11. Root density distributions in Phase I shallow lysimeters averaged over a period of 6 weeks prior to harvest in 1990–1993. From Wadey et al. (2001).
Since the dominant zone of contamination was the bottom 20–30 cm of the lysimeters, it is possible that the roots penetrated more effectively to these depths in 1993. Although this is not supported by the observed root profile distributions (Figs. 4.11 and 4.12), such an effect would require only a small number of roots, which were perhaps not easily detected, to penetrate to this depth. Interestingly, the observed biomass yields were generally lowest in this year also. The relatively poorly growing plants may have developed some deeper roots in order to exploit water and/or nutrients and, as a consequence, taken up relatively large quantities of 36 Cl from depth. Given the difficulties interpreting the observed data, this highlights the need for a systematic approach to understanding the data through modelling. 4.3.1.3.
Soil transport simulations
A modelling methodology to simulate water flow and radionuclide partitioning and migration within the integrated Phase I lysimeter
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
103
70 60 Height above geotex (cm)
May 3, 2007
50 40 30 1990
20
1991 1992
10
1993
0 0
1
2
3 3
4
5
-3
Root density (cm cm )
Fig. 4.12. Root density distributions in Phase I deep lysimeters averaged over a period of 6 weeks prior to harvest in 1990–1993. From Wadey et al. (2001).
facility has been previously described (Chapter 3, Butler & Wheater, 1999a). This model was used to investigate the migration of 36 Cl within the Phase I lysimeters and its subsequent uptake by the various annual crops of winter wheat. Table 4.2 gives the activities of 36 Cl that were used to dose the experimental system during the spring and summer months over the five cropping seasons (1990–1994) during which the experiment was operational. As already noted, the amount of 36 Cl added during the first year of the experiment was substantially greater than that added in successive years (by approximately a factor of eight). This was due to a desire to maintain a target concentration of 80–120 kBq l−1 in the supplied water over the 1990 crop growth season, which was particularly warm and dry, and during which large volumes of water were exported from the common reservoir to meet the high evapotranspiration demands of the crop. This objective was relaxed in subsequent years owing to health and safety restrictions on the total activity of 36 Cl in the experimental system.
ch04
May 3, 2007
14:33
104
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste Table 4.3. Parameters used in lysimeter system model for 36 Cl simulations. Parameter
Symbol
Value
Units
Height Shallow lysimeter Deep lysimeter
HL HL
40.0 70.0
cm cm
Surface area Common reservoir Lysimeter
AR AL
12210 16562
cm2 cm2
Volume Buffer tank Substrate
VB VS
96 000 960 000
cm3 cm3
Soil parameters Mean moisture content Bulk density
θ¯ ρs
0.3 1.5
cm3 cm−3 g cm−3
Table 4.3 gives a list of general model parameter values previously obtained by Butler & Wheater (1999b). Two further parameters, specific to 36 Cl, which are required to run the model are the linear equilibrium sorption (Kd [cm3 g−1 ]) and the root uptake α [cm2 s−1 ]) coefficients. In the case of soil sorption, it was considered reasonable (a priori) to treat radiochlorine as being conservative, with a Kd value equal to zero, thus leaving only the root uptake coefficient to be obtained through calibration. However, an initial simulation of 36 Cl concentrations in the water table control system and lysimeter soil showed poor agreement between observed and predicted results during the first (1990/1991) winter flushing period. The source of this discrepancy appeared to be the simulated soil concentrations, which were greatly in excess of those observed from extracted soil cores taken shortly after the first crop harvest in 1990 (Figs. 4.13 and 4.14). This resulted in a reinterpretation of 36 Cl behaviour in soil. An inventory balance of the experimental system (Table 4.4) at the time of the first crop harvest on 26 July 1990 (Julian date (JD) 90207) indicated that 49 MBq (i.e. 20%) of the 36 Cl added to the experimental system during the first season of operation was unaccounted for. Therefore, the implication from the model simulation was that a greater activity of 36 Cl had been transferred to the soil
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
105
Fig. 4.13. Comparison of simulated and observed mobile 36 Cl shallow lysimeter soil concentrations at 1990 crop harvest.
than could be accounted for from the soil core analyses. It is usual to assume that chlorine, added in the form of the chloride ion, acts as a conservative tracer in soils (e.g. Appelo & Postma, 1993; Brady & Weil, 1996). However, there are a few studies that have investigated the cycling of chlorine in soil and have shown that there is a signif¨ icant amount of organically bound chlorine present in soil (Oberg, 1998) and that organo-halogens can act as a sink for chlorine. The soil core radiochemical analyses presented above were based on soil water extractions, which indicated the amount of 36 Cl readily available for root uptake or soil leaching. Given the discrepancy between the modelling and experimental results, there was a concern that additional, unextracted 36 Cl may have been present in the soil in a less mobile and hence less readily extractable form. Laboratory studies have demonstrated that additional amounts of 36 Cl could be extracted from contaminated soils using 1M NaOH,
ch04
May 3, 2007
14:33
106
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 4.14. Comparison of simulated and observed mobile soil concentrations at 1990 crop harvest.
36
Cl deep lysimeter
compared with the amounts obtained using de-ionised water (Lee, 1997; Lee et al., 2001). Therefore, 10 g samples of lysimeter soil were taken from a layer 20–30 cm height above the Geotextile layer to ascertain whether a similar situation applied in the lysimeter experiments. To each of these soil samples, 30 ml of 1M NaOH was added and shaken for 2 hours. Following this treatment, the suspension was centrifuged (4000 rpm for 15 minutes) and 1 ml of the sedimentfree supernatant removed by pipette. The colour of the supernatant was dark brown, indicating substantial extraction of native humic substances from the bulk soil. The 1 ml sample of supernatant was injected onto a 40 cm gel filtration column (Fractogel HSW-40, with a nominal upper molecular weight cut-off of 10 000 Daltons) and eluted in a flowing phase of 10% methanol at 1 ml per minute. Fractions of the column eluent were collected at 3 minute intervals and analysed sequentially for both UV absorbance at 280 nm and 36 Cl activity.
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine Table 4.4.
36
107
Cl inventory at 1990 crop harvest.
Component Dosing Common reservoir Buffer tanks Lysimeter substrates Lysimeter soil Wheat crop Total in system Unaccounted
36
Cl (MBq) 247.16 2.69 5.52 108.59 59.95 20.79 197.54 49.62
A plot of the elution profiles for both UV absorbance (indicating humic substances) and 36 Cl activity is shown in Fig. 4.15. Once the above analysis had been carried out for all fractions, the sum of extracted 36 Cl was calculated and compared to the total extracted using de-ionised water from a similar sample. Extraction with de-ionised water gave a 36 Cl activity concentration of 3.3 kBq kg−1 , whereas the amount obtained using sodium hydroxide was 11.5 kBq kg−1 , a difference of around 8 kBq kg−1 and an increase by a factor of 3.5. As NaOH is used as a non-specific extractant for naturally-occurring humic substances in soils, it can be hypothesised that the additional 36 Cl extracted by this method was in some way associated with humic substances. This is supported by the gel filtration chromatogram (Fig. 4.15), which indicates that the largest sub-fraction of the total 36 Cl activity extracted by NaOH was associated with a low molecular weight peak of humic material. There are also two subsidiary peaks, one of which is considered to be the free chloride form of 36 Cl. The other is more difficult to interpret. It is thought that it might be associated with a range of unidentified organic compounds, although there is no clear peak in the UV absorbance spectrum to confirm this. An inventory analysis of the system at the 1991 re-dosing on 16 May (JD 91136) indicated that 73 MBq of 36 Cl (30% of the total input) was unaccounted for. If this apparent loss is assumed to be organically bound 36 Cl and distributed equally between the eight lysimeters over a contaminated soil depth of 40 cm this gives,
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
108
Biosphere Implications of Deep Disposal of Nuclear Waste 25
3.5
36
Cl associated with low molecular weight humic material
Major humic acid peak
3
20
2.5
36
Cl in free chloride form
15
Total beta activity (dpm)
UV absorbance @ 280 nm (-)
May 3, 2007
2
1.5 10
36
Cl associated with unidentified organic material
1
Low molecular weight humic
5
0.5
0
0 1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22
Fraction Number (3 minutes per fraction)
Fig. 4.15. Gel filtration chromatogram showing elution of 36 Cl in relation to humic substances extracted from lysimeter soil using 1M NaOH.
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
109
for a mean bulk density of 1.5 g cm−3 , a stored concentration of 9 kBq kg−1 . This result is remarkably similar to the value of 8 kBq kg−1 obtained above and appears to confirm the presence of organically bound 36 Cl components in the soil. Further implications of this effect were considered using the lysimeter system model. Laboratory studies of free and organically-bound 36 Cl partitioning have indicated that transfer from the free ionic form to association with low molecular weight humic substances takes place over a time period of approximately 50 days (Lee, 1997). To incorporate the effects of 36 Cl partitioning between the mobile ionic form and an immobile organically bound state, the lysimeter system model was revised to include the following reaction, which also acted as a sink in the lysimeter soil water concentration balance equation: ¯ θc dSo = βo (Som − So ) , dt ρb Som where So is the concentration of organically bound 36 Cl per unit dry mass of soil [kBq kg−1 ], Som is the maximum organic concentration the soil can hold [kBq kg−1 ], βo is the binding rate constant [s−1 ], c is the concentration of 36 Cl in solution [kBq l−1 ], θ¯ is the mean soil moisture content [cm3 cm−3 ] and ρb is the dry bulk density [kg l−1 ]. ¯ b converts volumetric soil water concentrations to unit The ratio θ/ρ soil mass. In the above example, the reaction proceeds until a maximum concentration Som is reached. For the purposes of this simulation, it was assumed that the soil organic matter acts as a sink for 36 Cl. However, it would also be expected that 36 Cl could be subsequently released through decomposition of the organic material. Here, it has been assumed that this process is significantly slower (i.e. by at least an order of magnitude) than the rate of uptake and hence, over the time-scale of the experiment, has been ignored. An initial value for Som of 8 kBq kg−1 was adopted (the difference between the NaOH and H2 O extractions). Although this value enabled the system model to reproduce well the vertical distribution of water-extractable 36 Cl in the deep lysimeters at the time of the 1990 crop harvest, the simulated vertical distribution of waterextractable 36 Cl for the shallow lysimeters still exceeded the observed
ch04
May 3, 2007
14:33
110
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
concentrations. This, in turn, led to excess simulated 36 Cl concentrations in the water table control system components. Through calibration, a revised set of organic 36 Cl storage concentrations was derived. For three of the shallow lysimeters (C, G, H), Som was set at 25 kBq kg−1 whereas in lysimeter D (which had shown rather different hydraulic behaviour to the others) a value of 15 kBq kg−1 was employed. Comparisons of simulated water-extractable vertical profiles with core samples taken around 50 and 100 days after the start of the experiment gave a value of 1.2 × 10−6 s−1 for the binding rate constant βo . Combining this with values for Som , ρb and θ¯ along with a mean soil water concentration c over the growth period of 40 kBq l−1 gives an estimated reaction half-time of 20 days. This is about 2–3 times more rapid than estimates derived from laboratory studies of small-scale non-vegetated pot experiments containing 160 g of soil (Lee, 1997). A comparison of the observed water extractable soil concentration profiles of 36 Cl, taken at the time of the 1990 crop harvest, with those for model simulations with and without organic partitioning is shown in Fig. 4.13 (shallow lysimeters) and Fig. 4.14 (deep lysimeters). The effect of 36 Cl partitioning between free and organically bound states in bringing the simulated water extractable soil concentrations into agreement with those obtained from two sampled cores is readily observed. It can also be seen that the model is able to reproduce the general shape of the observed concentration depth profiles, particularly when the effects of spatial variability, demonstrated by the difference between the two core samples for each lysimeter, are taken into account. The reason for the discrepancy in simulated and observed 36 Cl concentrations in the top 10 cm layer for lysimeters G and H is unclear. Distinctions in water flow are not an adequate explanation as Fig. 4.13 shows that 36 Cl had migrated to the surface in lysimeter D, even though the total amount of water which had entered this lysimeter from the start of the experiment to crop harvest was approximately half that of the other three shallow lysimeters (Burne et al., 1994). The 36 Cl concentration profiles in the deep lysimeters are well reproduced apart from the under prediction of 36 Cl concentrations in the top 10 cm
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
111
of soil. One possible cause for the higher than predicted concentrations might be enhanced upward migration velocities arising from reduced soil water contents in the upper half of the lysimeter due to large evapotranspiration losses under the unusually warm and dry conditions leading up to crop harvest. This is not accounted for in the model formulation, which assumes a constant moisture content throughout the lysimeter soil profile. Another possible explanation might be the operation of a rapid biological pathway such as root translocation and exudation, or foliar leaching or litter fall onto the soil surface. Simulated 36 Cl extractable soil concentration profiles for subsequent years (1991–1994) show reasonable agreement with observed data, particularly when the effects of spatial variability are taken into account. These concentrations were generally much lower than in the 1990 profiles, due to projected losses of 36 Cl via leaching and plant uptake, and the lower amounts of 36 Cl added to the system in the years following 1990 (Table 4.1). The effects of the reduced levels of mobile 36 Cl due to organic partitioning in the soil can also be seen in the simulated behaviour of 36 Cl concentrations in the water table control system (Figs. 4.16 and 4.17). Figure 4.16 shows simulated and observed 36 Cl concentrations in the common reservoir over the five crop growing seasons. The effect of organic partitioning on the amount of 36 Cl flushed out of the lysimeters during the 1990–91 winter period can be clearly seen and shows how its introduction has helped to bring the model into agreement with observed data for all components of the integrated experimental system. Differences between the two simulations (i.e. with and without partitioning) are carried over into the following (1991) crop season. During the period from the redosing of the lysimeter with 36 Cl in spring (JD 91136, 16 May 1991) to the next re-dosing a year later (JD 92133, 12 May 1992) there was a steady decline in common reservoir 36 Cl concentration. Both model simulations reproduce this decline with an almost constant 5 kBq l−1 difference between them. Although the original model simulation tends to give a better fit to the data, it is instructive to consider what gives rise to this difference in concentration between the two simulations. Towards the end
ch04
May 3, 2007
14:33
112
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 4.16. Comparison of simulated common reservoir (with/without organic partitioning) with observations.
36
Cl concentrations
of April 1991, water levels in the common reservoir had, in response to evapotranspiration losses from the lysimeters, dropped sufficiently to require the addition of 300 litres of uncontaminated water. This resulted in a 70% fall in common reservoir 36 Cl concentrations, with the result that the original and revised model simulations were 2.8 and 1.2 kBq l−1 , respectively. However, shortly after 29 April (JD 91119), an intense 23 mm rainfall event resulted in substantial fluxes of water being pumped from the lysimeters to the common reservoir over the following days. The higher concentrations of 36 Cl present in the substrate and buffer tank components for the original model formulation result in additional 36 Cl activity entering the common reservoir, compared with the situation for the organic partitioning model, and gave rise to the 5kBq l−1 concentration difference between the two model predictions. By the time of the 1992 re-dosing the difference between the two simulations is extremely small and for the
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
113
Fig. 4.17. Simulated and observed 36 Cl concentrations in buffer tanks and bead substrate for lysimeter A control system.
remaining time period, the two models are almost identical and show good agreement with the observed data. Figure 4.17 shows, as an example, the simulated and observed 36 Cl concentrations in one of the deep lysimeters (A) and demonstrates the dynamic nature of the concentrations in the system components (particularly the buffer tanks), as well as the model’s ability to reproduce this. However, the plot also shows additional variability in the observed radiochlorine concentrations, some of which is due to analytical errors. An indication of this is those occasions when measured concentrations of the buffer tank and the two substrate samples are markedly different from the model, e.g. JD 91225 (13 August 1991) and JD 92227 (14 August 1992). In these instances, there does not appear to be any physical or chemical explanation why such high values should have been obtained, particularly given the good agreement between model and data both prior to and after these dates.
ch04
May 3, 2007
14:33
114
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
In a similar manner, the extremely low concentrations (< 1 kBq l−1 ) recorded in both common reservoir samples (Fig. 4.16) and buffer tank (Fig. 4.17) on 90142 (22 May 1990) along with the substrate drain sample for JD 91044 (13 February 1991) also appear to be outliers. The same argument can be extended to other “outliers” such as the buffer tank concentration for JD 93049 (18 February 1993). One area where there is a consistent discrepancy between simulated and observed concentrations is in the lysimeter substrate during the first 100 days of the experiment. During this period, the simulated concentrations are between 10 and 40 kBq l−1 lower than the sampled values. The most probable explanation for this, apart from data error (which seems unlikely given the consistency between model and data in the other components of the experimental system over the same time period), is that the assumption of complete mixing within the lysimeter substrate is not valid. As it is an enclosed volume, dosing of the lysimeter substrate was achieved by hydraulically isolating it from the buffer tank and then circulating water between the manifold inlet and the drain (see Fig. 3.2). A small amount of 36 Cl contaminated water (approximately 3 litres) was then added to the circulating water at a concentration which, when diluted over the 120 litre substrate pore volume, would produce the required target concentration. Substrate samples were taken from the drain (at its base) and from a tube located immediately beneath the geotextile layer (designated the top tube). Incomplete mixing of the dosing water in the substrate could result in regions of elevated concentrations within the lower part of the system, which would gradually dissipate through dispersion and diffusion. The drain location would therefore tend to sample these higher concentrations. In contrast, the model simulation should approximate to the mean 36 Cl concentration of water entering the lysimeter soil. The effect of mobile and stagnant regions within the substrate and their incomplete mixing, particularly in relation to the location of the inlet manifold within the substrate, became progressively apparent during the first year of the experiment from the large concentration differences observed between drain and top tube samples. In order to address this problem, a regular practice of re-circulating the substrate water was
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
115
introduced. The effectiveness of this can be seen in the improved agreement between the two sets of substrate sample concentrations over subsequent years of the experiment.
4.3.1.4.
Wheat uptake simulations
The representation of soil-to-plant transfer is an important component in radiological assessment models. This has traditionally been achieved through the use of concentration ratios or transfer factors (IUR, 1989). However, such an empirical approach does not incorporate any explicit physical process representation. The modelling approach that has been adopted makes the assumption that uptake per unit length of plant root is proportional to the soil water concentration. Thus, the total crop uptake at harvest is the integration of the soil water concentration and plant root density over both depth and time (Butler & Wheater, 1999a). This method was successfully used to describe the uptake of 22 Na by the lysimeter wheat crop (Butler & Wheater, 1999b). Figure 4.18 shows a plot of the simulated and observed total crop uptake of 36 Cl in the above-surface components of the winter wheat crop in kBq per lysimeter over five crop harvests. A root uptake coefficient of α = 2 × 10−10 cm2 s−1 results in the model under-estimating the observed crop uptakes during the first crop harvest in 1990, but reproduces well total crop uptake of 36 Cl over the subsequent four harvests. In order for the model to reproduce the crop uptake values for 1990, the root sorption coefficient needs to be increased by a factor of 20. Clearly this indicates that there is a degree of simplification that has been introduced by the modelling approach which does not incorporate effects during the first crop season. Nevertheless, overall the results are encouraging as they reproduce the variability observed between lysimeter types and over different crop seasons. Further refinements require the use of more sophisticated representations of water flow, contaminant transport and plant uptake, storage and release, which are described in the following section. One important contrast should be noted between the results discussed here and those described earlier in Sec. 4.3.1.2. The model simulations reproduce the total uptake by
ch04
May 3, 2007
14:33
116
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 4.18. Comparison of simulated and observed total lysimeter in harvested wheat crop for 1990–1994.
36
Cl uptake (kBq) per
the crop in response to 36 Cl concentrations in soil water interacting with the plant root system. Therefore, the model simulation does not directly incorporate shoot biomass, although this is to some extent indirectly incorporated through root density profile and water fluxes in response to evapotranspiration. This explains why the total uptake by the 1993 winter wheat crop harvest, which is generally lower than those for the previous years, produces the largest plant activity concentrations when biomass and radiochlorine distribution in the soil are taken into account. In summary, an important outcome from the Phase I investigation 36 of Cl migration and uptake in a vegetated soil was to call into question the assumption that chlorine is always a conservative element in environmental systems. Although some studies have indicated an ¨ organic phase to chlorine cycling in soil (e.g. Oberg, 1998), there
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
117
has been little direct and quantitative evidence to support this. This analysis has indicated that there is a partitioning of radiochlorine between that in the mobile chloride form (36 Cl− ) and that which becomes bound to organic compounds with a molecular weight of less than 10 000 Daltons. However, the magnitude of the partitioning appears to be surprisingly large (with free to bound ratios of 1:3 to 1:6). Such a result clearly has important implications both for the ultimate fate and transport of 36 Cl in the context of radioactive waste disposal and also for an overall understanding of chlorine behaviour in environmental systems. The radiochlorine that entered the lysimeter at the base of the soil profile during the summer of 1990 was in the free chloride form at a concentration of 60 µg l−1 . However, it then entered a system where the stable chloride levels in the soil water solution were around 60 mg l−1 , i.e. three orders of magnitude greater than 36 Cl concentrations. It is reasonable to assume that the stable Cl existed at some quasi-equilibrium in which Cl fluxes between the soil water solution and the organic soil components were roughly equal and opposite. It is thought that the first six months of the lysimeter experiment established equilibrium between the free and organically bound states of 36 Cl. This resulted in a progressive reduction in soil water concentrations to levels comparable to those added to the experimental system in subsequent years and meant that there was little further net transfer between the two states. This explains why 36 Cl in subsequent years appears to behave as a conservative tracer, but with lower availability to plants than in the first season. It is possible, therefore, that in the event of a release of 36 Cl into the near-surface environment, organic partitioning may initially tend to reduce concentrations of 36 Cl in soil water and thereby reduce plant uptake. The experimental and modelling results have shown that 36 Cl is readily taken up by winter wheat, with 16% of the 36 Cl which entered the lysimeter during the first crop season appearing in the above-surface components of the crop. The calibrated root uptake coefficient (α ) of 2×10−10 cm2 s−1 (1991–1994) and 4×10−9 cm2 s−1 in 1990 is between 25 and 500 times greater than that obtained for 22 Na (8×10−12 cm2 s−1 ) and demonstrates the importance of chloride
ch04
May 3, 2007
14:33
118
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
uptake by the wheat plant in order to maintain ion balance within the crop tissue. It also highlights the role of arable crops as an important vector for 36 Cl transfer from the geosphere to the biosphere and the importance of properly representing its behaviour in the nearsurface environment when developing models for dose assessment in connection with sub-surface radioactive waste disposal. Given these significant outcomes, it was important to assess whether such results were reproducible and could be also observed with a different soil type and plant crop. Consequently, a second lysimeter experiment was undertaken with an additional soil and a different crop to that used in the Phase 1 experiment.
4.3.2.
One year lysimeter experiment (Phase 2)
4.3.2.1.
Lysimeter soil profiles of
36Cl
The Phase II experiment comprised eight lysimeters containing Silwood and Longlands soils, which were dosed with 36 Cl in late June 1995. Soil cores were removed from the lysimeters in July, August and September of 1995 and April 1996. Mean activity profiles for each soil are shown in Figs. 4.19 and 4.20. At the July sampling, upwards migration of the 36 Cl had clearly taken place up to the 30–40 cm (above geotex) layer in both soils. However, in both soils there were relatively small amounts of activity above this, extending to the soil surface, particularly in the Longlands soil. Therefore, within one month of the lysimeters being dosed with 36 Cl, the radionuclide had migrated upwards through the full 70 cm of soil. This equates to a migration rate in excess of 2 cm day−1 . In general, the two soils seemed to show broadly similar patterns in terms of the migration of 36 Cl. However, a comparison of the soil types in terms of activity concentrations at the July sampling illustrates some differences, particularly at the base of the lysimeters. The main fraction of the activity in the Silwood soil was present between 6 and 20 cm (above geotex), whereas in the Longlands soil, it was slightly lower, between 2 and 10 cm (above geotex). This illustrates a greater migration in the Silwood soil.
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
119
70 Jul-95 60
Aug-95 Sep-95
50 Soil depth (cm)
May 3, 2007
Apr-96
40
30
20
10
0 0
1
2
3
4
5
6
Soil activity concentration (Bq g-1)
Fig. 4.19. Mean soil 36 Cl activity concentrations in the Phase II Silwood soil lysimeters (1995–1996).
At the August sampling, 36 Cl in the Silwood soil had continued to migrate up through the soil relative to the July sampling. The activity concentration in the upper 10 cm of soil had increased markedly and the main fraction of the 36 Cl towards the base of the column had moved upwards to 20–30 cm (above geotex) soil layer. Nevertheless, activity concentrations in the 40–60 cm (above geotex) zone remained very close to zero. Activity concentrations were comparable to those seen for the July sampling. However, in the Longlands soil, activity concentrations appeared to be generally lower for the August sampling than were seen in the previous month. Most significantly, the large peak of activity in the 2–10 cm (above geotex) layer present at the July sampling, had all but disappeared in the August sampling. However, some increase in activity concentration was observed in the 20–30 cm and 50–60 cm layers (above geotex). The September sampling showed that, at most depths, activity concentrations decreased relative to the August sampling. However, in the intermediate depths, where activity concentrations had been
ch04
14:33
120
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste 70 Jul-95 Aug-95
60
Sep-95 Apr-96
50 Soil depth (cm)
May 3, 2007
40 30
20
10
0 0
2
4
6
8
10
Soil activity concentration (Bq g-1)
Fig. 4.20. Mean soil 36 Cl activity concentrations in the Phase II Longlands soil lysimeters (1995–1996).
very close to zero in the two previous months, the activity concentrations had increased. For both soil types, a slight accumulation of 36 Cl at the soil surface was observed. Activity concentrations in the Silwood soil peaked in the 10–20 cm (above geotex) layer, i.e. lower down the profile than was observed in August. In the Longlands soil, the greatest activity concentrations were observed at the very base of the soil profile (for the first time). As was generally observed in both July and August, activity concentrations in September were lower for the Silwood soil than the Longlands soil in the bottom 10 cm of soil, but higher throughout the remainder of the soil profile. Bearing in mind that the input (water table solution) activity concentration was identical for both soils, differences in the observed activity concentrations between the soils perhaps suggest that positioning of the 36 Cl was taking place. Since the water extraction used to remove 36 Cl from the soils prior to analysis is likely to have removed only activity remaining in the chloride (free) form, soil-positioned activity would not have been accounted for using this technique. As discussed for
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
121
the Phase I lysimeters, this process is potentially significant. Further attempts to address this issue were made in subsequent experiments (described below). At the final soil sampling (April 1996) the 36 Cl activity concentrations in both soils were low compared with previous measurements. Particularly at the bases of the lysimeters, there was a marked reduction in activity concentrations, to a maximum of around 0.5 Bq g−1 in the lower-most layer. This compares to peak values in July of around 6 and 9 Bq g−1 in the Silwood and Longlands soils, respectively. Mean monthly net water influxes to the base of the lysimeters, rainfall and potential evapotranspiration are shown in Figs. 4.21 and 4.22. These suggest that the rapid migration of 36 Cl to the soil surface during July was due to a net positive water influx to the base of the lysimeters, driven by low rainfall and high evapotranspiration. Similarly, the continued upward migration during August was also likely to be due these processes since similar hydrological conditions were observed. However, during September, rainfall increased and evapotranspiration decreased, leading to a net negative water flux to the base of the lysimeters. This indicates that the dominant moisture flux within the lysimeters was downward and may account for the apparent loss of activity from the lysimeter soils between August and Waterinflux through lysimeter base (mm)
May 3, 2007
150
Silw ood Longlands
100 50 0 Jul-95
A ug-95
Sep-95 Oct-95
No v-95 Dec-95
Jan-96
Feb-96 M ar-96
A pr-96
-50 -100 -150
Fig. 4.21. Net water fluxes across the bases of the Phase II lysimeters during 1995–1996. Data for April relate only to period up to the 19th when the experiment ended.
ch04
14:33
122
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste Rainfall (mm)
140 Evapotranspiration or rainfall (mm)
May 3, 2007
Evapotranspiration (mm) 120 100 80 60 40 20 0 Jul-95
A ug-95
Sep-95
Oct-95
No v-95 Dec-95
Jan-96
Feb-96
M ar-96
A pr-96
Fig. 4.22. Rainfall and potential evapotranspiration from the Phase II lysimeters during 1995–1996 (absolute amounts per month). Data for April relate only to period up to the 19th when the experiment ended.
September, i.e. it was leached out through the base of the lysimeter. Similarly, the April 1996 sampling followed a period of relatively high rainfall and low evapotranspiration (winter months) which appeared to have leached much of the activity from the soil profile. Nevertheless, losses of activity into the ryegrass crop, particularly in September, should also be considered (see below). Overall, however, the soil behaviour of 36 Cl seemed to be well explained by the fluxes of water through the soil. This is consistent with the predominance of the poorly sorbed chloride ion in soil solutions. Nevertheless, its migration appears to be tempered somewhat by a degree of positioning onto the soil solid phase, probably organic matter, as evidenced by the differences in activity concentrations between soil types. 4.3.2.2.
Ryegrass biomass and activity concentrations
The mean monthly production of ryegrass biomass in the lysimeters is shown in Fig. 4.23. For both soil types, yields were greatest in July 1995 and tended to decrease over time. Comparing the effect of soil type, the Silwood soil gave the greater yield in July 1995 but then lower yields until April 1996. The yields of ryegrass were much greater than those observed in the subsequent soil column
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
123
experiments (see below), due to the much larger surface area of the lysimeters, compared to the soil columns, and the beneficial effect of natural light. The Silwood soil gave greater ryegrass activity concentrations than the Longlands soil in all months (Fig. 4.24). For both soils, July 1995 and April 1996 gave the lowest ryegrass activity concentrations and August the highest. September and October gave
Ryegrass dry biomass (g)
700 Silw ood
600
Longlands
500 400 300 200 100 0 July
August
September
October
April
Fig. 4.23. Mean ryegrass biomass production from the Phase II Silwood and Longlands lysimeters (July 1995–April 1996).
Ryegrass activity concentration (Bq g-1)
May 3, 2007
8000 Silw ood
7000
Longlands
6000 5000 4000 3000 2000 1000 0 July
August 36
September
October
April
Fig. 4.24. Mean ryegrass Cl activity concentrations for the Phase II Silwood and Longlands lysimeters (July 1995–April 1996).
ch04
May 3, 2007
14:33
124
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
similar results to one another. Since ryegrass would not be expected to root deeply, the activity concentration of the upper soil is likely to be of some significance in determining the uptake of 36 Cl into the plant. Therefore, in July when activity concentrations in the 0–10 cm soil layers were lowest, the uptake into the crop was relatively low. In contrast, highest activity concentrations in these soil layers were observed in August and this was mirrored by the greatest uptake in this month. Activity concentrations of the soil in this region were generally greater in the Silwood soil than in the Longlands soil, explaining why the activity concentrations of the ryegrass were greater in the Silwood soil. The 36 Cl activity concentration of the upper soil zone (0–10 cm) therefore appeared to a good indicator of uptake by the crop. Nevertheless, the hydrological behaviour of the soil-plant system is also likely to exert a strong control on uptake (e.g. via evapotranspiration flux). For example, the high uptake observed in August coincided with the highest net water influx (low rainfall and high evapotranspiration; Figs. 4.21 and 4.22). Furthermore, the low uptake in September can possibly be attributed to the relatively high rainfall during this month, which appeared to leach 36 Cl out of the rooting zone of the soil profile, and to a low rate of evapotranspiration. The same also appeared to be true of the April 1996 sampling. Thus, plant uptake of the 36 Cl appeared to be conditioned by both the distribution of soil contamination within the rooting zone and hydrological conditions (e.g. favoured by high transpiration). Mean weighted transfer factors (calculated according to Eqs. 4.1 and 4.2) are shown in Fig. 4.25 and were very high, ranging from 12 000 to almost 50 000 across both soil types. 4.3.2.3.
Soil transport simulations
Although some inferences for the processes affecting 36 Cl behaviour in the Phase II lysimeters have been described in the previous section, further insights into the effects of soil water fluxes and the magnitude of soil-to-plant transfers can be obtained through the application of physically-based modelling techniques. Hydrological simulations of soil moisture content and water fluxes (described in the previous chapter) provided the driving inputs for simulations,
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
125
April
Longlands Silwood
September
August
July
0
10000
20000
30000
40000
50000
60000
Soil-plant transfer factor
Fig. 4.25. Mean weighted soil-plant transfer factors for the Phase II Silwood and Longlands lysimeters (July 1995–April 1996).
using the solute transport model SLT1 (also described in the previous chapter), of 36 Cl in the Phase II lysimeters. In order to conduct detailed generic model simulations of the behaviour of radionuclide migration and uptake in the experimental lysimeters using the SLT1 model, several aspects of the lysimeter representation have to be considered. These are the discretisation of the soil profile; the initial and boundary conditions; the transport parameter values; and the representation of the plant root system. A model grid discretisation identical to that employed for the SPW1 hydrological model simulations (i.e. 2 cm grid mesh) was utilised. This ensured that the soil moisture contents and mean internodal, boundary, and plant root water fluxes derived by the SPW1 model could be used to directly drive the SLT1 solute transport model. In order to control the size of the SPW/SLT interface file, which supplies the simulated soil moisture contents and fluxes, but still provide a detailed representation of the dynamic behaviour of the system, a time interval of 0.1 day was selected. In addition to the hydrological inputs, the model also requires definitions of the initial and time-dependent boundary conditions and parameter values. At the start of the experiment all of the lysimeters were uncontaminated and thus the initial solute and sorbed concentrations of the radionuclide were zero.
ch04
May 3, 2007
14:33
126
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
The upper boundary of the model represents the soil surface. Any upward fluxes of water through the upper boundary are due to evaporation and were considered to be pure. In addition, it was assumed that any radionuclides apportioned to the crop remain until harvest when they are completely removed from the lysimeter system (i.e. the contributions of loss of plant parts, leaching and wash-off were generally considered to be insignificant (Coughtrey et al., 1983). Thus, there were no radiochemical fluxes across the upper boundary (i.e. soil surface) in either direction. At the lower boundary of the model, which represents the geotextile interface between the soil and the bead substrate, the controlling factors are the water flux across this interface (which is provided by the SPW1 model) and the radiochemical concentration in the substrate (since this defines the concentration of radiochemical in the water entering the lysimeter). The variation in 36 Cl concentration over time was obtained from the periodic observations of concentrations from a tube immediately below the geotex layer at the base of the lysimeter soil at roughly fortnightly intervals. The time-dependant 36 Cl lower boundary concentrations, derived in this manner, are shown in Figs. 4.26 and 4.27. There is an initial period of nearly 30 days when the concentrations are zero (when the water table control system became operational) followed by a rise in concentrations as the lysimeters were dosed. The interpretation of the double peak in the 36 Cl concentrations is unclear and is possibly an artefact of the sampling and analysis procedures on the second sampling date. This is followed by a decline in substrate concentrations as additional clean water added to the common reservoir, in order to replenish water pumped to the lysimeters in response to evapotranspiration losses, moves through the experimental system. During September 1995, there is a short rise in the lower boundary concentrations, which is due to flushing of the lysimeters by excess rainfall. However, it is noticeable that this effect is relatively small and is followed by a slow decline in the lower boundary (i.e. substrate) concentrations until the end of the experiment. Assuming a homogeneous soil composition within the lysimeter and excluding plant root densities, the SLT1 model is characterised
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
Fig. 4.26.
Lower boundary
36
127
Cl concentrations (Silwood lysimeters).
∗ , longitudinal dispersivby 4 soil parameters (molecular diffusion Dm ity dL , a linear equilibrium sorption coefficient, Kd and the dry bulk density ρb ), a root uptake coefficient α (discussed in Chapter 3) and a radioactive decay coefficient λ. The values of these parameters used in the Phase II 36 Cl simulations are given in Table 4.5. It has already been remarked that substantial uptake of 36 Cl by ryegrass took place in the Phase II experiment. Such an effect has a major impact on soil concentrations. Figures 4.28–4.31 show a comparison of observed 36 Cl concentrations for the lysimeters containing Silwood soil. Setting the root uptake coefficient (α ) to zero (i.e. no plant uptake) results in simulated 36 Cl soil activity concentrations greatly exceeding those observed over the bulk of the soil profile. The accompanying simulation shows the effect of including a “loss” of 36 Cl from the soil through root uptake. An inspection of the soil concentrations shows that, although there is little difference between the simulations for the first coring
ch04
May 3, 2007
14:33
128
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 4.27.
Lower boundary
Table 4.5.
36
Cl concentrations (Longlands lysimeters).
SLT1 model parameter values for Phase II modelling.
Parameter
∗ Soil diffusion coefficient (Dm ) Soil dispersivity dL Sorption coefficient Kd Bulk density ρs Root uptake coefficient α Decay constant λ Decay half-life T1/2
Units
cm2 s−1 cm cm3 g−1 g cm−3 cm2 s−1 s−1 year
Soil type Silwood
Longlands
4.0 × 10−6 1.9 0.0 1.5 8.6 × 10−9 7.3 × 10−14 3.0 × 105
4.0 × 10−6 2.1 0.0 1.5 7.6 × 10−8 7.3 × 10−14 3.0 × 105
date (Fig. 4.28), there are substantial differences by the time of the next two dates (Fig. 4.29 and 4.30). These clearly show the depletion of 36 Cl from the soil into the ryegrass crop, and the effect of this upon the soil profile concentrations.
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
Fig. 4.28.
Simulated and observed
Fig. 4.29.
Simulated and observed
36
36
129
Cl Silwood soil concentrations at 95207.
Cl Silwood soil concentrations at 95236.
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
130
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 4.30.
Simulated and observed
36
Cl Silwood soil concentrations at 95270.
A point where large differences still occur is at the base of the lysimeter soil column. Here the model over-predicts soil concentrations (a similar effect also occurred when simulating 22 Na behaviour (Jackson, 2007)). Given the high concentrations of 36 Cl entering the lysimeter during the first 40 days following dosing, it does not seem possible that the ‘observed’ concentrations of around 1–2 kBq kg−1 can be genuine. In addition, during the original modelling of the soil concentration profiles, the other region where there was a clear difference between the simulated and observed values was in the top 10 cm of soil. One explanation for this effect is a surface influx arising from a combination of leaf wash off and litter fall. This was therefore tested by modifying the model in order to allow a first order release of 36 Cl back to the soil from the leaf storage component (Jackson, 2007) and did lead to an improvement in the model’s performance. Figures 4.32 to 4.35 show the equivalent simulated and observed 36 Cl soil concentrations for the lysimeters filled with Longlands soil. In general, the results are similar to those observed for the Silwood
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
36
131
Fig. 4.31.
Simulated and observed
Fig. 4.32.
Simulated and observed 36 Cl Longlands soil concentrations at 95207.
Cl Silwood soil concentrations at 96109.
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
132
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 4.33.
Simulated and observed 36 Cl Longlands soil concentrations at 95236.
Fig. 4.34.
Simulated and observed 36 Cl Longlands soil concentrations at 95270.
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
Fig. 4.35.
133
Simulated and observed 36 Cl Longlands soil concentrations at 96109.
soil lysimeters. The results for the coring date 95236, however, are of concern, partly due to the low activity concentration in the bottom 10 cm soil layer (similar to that observed with the Silwood lysimeters), but also from the relatively high levels of 36 Cl observed in the top 20 cm of soil. These are also seen to a lesser extent in the following coring date, 95270. Although the model incorporated a leaf release mechanism, this did not result in any substantial increase in 36 Cl activity in this top layer. This is because any release in of 36 Cl in to the top soil layer was rapidly taken back up due to the large number of roots present and the high root uptake parameter. The results for the final coring date (96109) show virtually all of the simulated 36 Cl flushed back out of the lysimeter. The observed data (admittedly only one sample) indicates a greater level of radiochlorine remaining in the lysimeter. It is possible that this is due to some recycling of the radionuclide through biomass release during the winter months.
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
134
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 4.36.
Simulated and observed
4.3.2.4.
36
Cl ryegrass uptake (Silwood lysimeters).
Ryegrass uptake simulations
Figure 4.36 shows the observed and simulated ryegrass uptakes for the Silwood lysimeters over the five crop harvests. Overall, the model simulation shows good agreement with the observed data. The poorest performance is for the first crop harvest and could indicate that the assumption of a constant root uptake coefficient during the period of crop establishment may not be correct. Whilst the simulated uptake for harvest date 95233 is at the low end of the range of observed values, it is somewhat in excess of those for 95268. Nevertheless, given the high degree of variability and the challenges posed when modelling biological systems, the performance of the model is highly encouraging, particularly when its ability in representing the soil concentrations profiles is also taken into account. Thus, the model simulation shows a high degree of consistency between the various internal components. It also indicates that the effects of organic binding of 36 Cl were less significant in the Phase II lysimeters than
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
Fig. 4.37.
135
Simulated and observed 36 Cl ryegrass uptake (Longlands lysimeters).
those that were encountered in the Phase I experiment (Lee et al., 2001). The results for the Longlands lysimeters (Fig. 4.37) are similar to those for Silwood soil, although as before, the simulated uptake is at the low end of the range of observed values for 95233 and exceeds all observed values in 95268. It is also gratifying to note that the reduced water fluxes in the Longlands lysimeters give rise to reduced 36 Cl uptake by the ryegrass in comparison with that observed in the Silwood lysimeters. 4.3.3.
Six month intact soil column experiment (Phase III)
4.3.3.1.
Soil column profiles of
36Cl
Mean water extractable activity concentrations of 36 Cl in the soil column profiles are given in Figs. 4.38 to 4.40 for the 1, 3 and 6 month
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
136
Biosphere Implications of Deep Disposal of Nuclear Waste Soil activity concentration (Bq g-1) 0
1
2
3
4
5
6
7
0
Soil depth (cm)
5
Silw ood
10
Robertgate
15
Wellesbourne
20 25 30 35 40 45 50
Fig. 4.38. Mean soil 36 Cl activity concentrations (water extractable) for the Phase III intact soil column experiment at 1 month.
Soil activity concentration (Bq g-1) 0
1
2
3
4
5
6
7
0 5 Soil depth (cm)
May 3, 2007
10 15
Silw ood Robertgate Wellesbourne
20 25 30 35 40 45 50
Fig. 4.39. Mean soil 36 Cl activity concentrations (water extractable) for the Phase III intact soil column experiment at 3 months.
columns, respectively. At 1 month, 36 Cl had migrated upwards to a depth of around 30 or 35 cm in all soils. This equates to a migration rate of around 0.67 cm day−1 . Within the contaminated zone, activity concentrations generally increased with increasing depth. Whilst broadly similar activity concentrations were observed for the
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
137
Soil activity concentration (Bq g-1) 0
1
2
3
4
5
6
7
0 5 10 Soil depth (cm)
May 3, 2007
15 20 25 30
Silw ood Robertgate Wellesbourne
35 40 45 50
Fig. 4.40. Mean soil 36 Cl activity concentrations (water extractable) for the Phase III intact soil column experiment at 6 months.
Silwood and Robertgate soils, the Wellesbourne soil displayed much lower activity concentrations. Because the input (water table solution) activity concentrations were the same for each soil type, this suggests that the 36 Cl may have become attached to the solid phase in the Wellesbourne soil and so was not as easily removed using the water extraction technique. At 3 months, the 36 Cl had reached the soil surface in the Robertgate soil, had reached the 5–10 cm layer in the Silwood soil, and had reached the 20–25 cm layer in the Wellesbourne soil. In addition to a lower degree of migration, the Wellesbourne soil again showed activity concentrations that were often low compared to the other two soils. At 6 months, in both the Robertgate and Silwood Park columns, 36 Cl had reached similar or higher activity concentrations at the soil surface than in the saturated region of the water table, particularly in the Silwood soil, where activity concentrations were relatively high at the soil surface. This indicates that the 36 Cl was transported to, and accumulated at, the soil surface. In the Wellesbourne columns, the 36 Cl had only just reached the soil surface, and increasing activity concentrations were observed as depth increased. This indicates that the 36 Cl migration in this soil was much slower than in the other two soils.
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
138
Biosphere Implications of Deep Disposal of Nuclear Waste 10000
6 months
9000 Total solution influx (ml)
May 3, 2007
3 months
8000
1 month
7000 6000 5000 4000 3000 2000 1000 0 Silw ood
Fig. 4.41.
Robertgate
Wellesbourne
Mean solution influx to the Phase III intact soil columns.
The soil sorption of chlorine is generally accepted to be very low, and, as was seen in the lysimeter experiments, the migration of 36 Cl is likely to be very sensitive to water movement through the soil. Indeed, the degree of 36 Cl migration upwards through the column seemed to be closely related to the water influx into the columns. The mean water influxes into the columns are shown in Fig. 4.41. For example, at 3 months the greatest water influx was observed to the Robertgate column which also showed the greatest degree of 36 Cl migration. Similarly, at 6 months, the Silwood soil columns had the highest water influx and showed the greatest accumulation of 36 Cl at the soil surface. At each time period, the lowest water influx rates were observed in the Wellesbourne columns and this was apparently reflected in the longer time taken for the 36 Cl to reach the soil surface in these columns. In addition, and as noted above, the apparent sorption of 36 Cl onto the Wellesbourne soil may also have been important in slowing upwards migration in this soil. Despite an expectation for low soil sorption of 36 Cl, the results do appear to suggest that some sorption may have taken place. Most notable were the relatively low activity concentrations found for the Wellesbourne soil, particularly at the first month’s sampling. As with the previous experiments, this suggests that the water extraction was not extracting all of the 36 Cl present, i.e. that it was not
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
139
Soil activity concentration (Bq g-1) 0
1
2
3
4
5
6
7
0 5
Soil depth (cm)
May 3, 2007
Silw ood
10
Robertgate
15
Wellesbourne
20 25 30 35 40 45 50
Fig. 4.42. Mean soil 36 Cl activity concentrations (sodium hydroxide extractable) for the Phase III intact soil column experiment at 1 month.
all present as the chloride form. Since sodium hydroxide is a (nonspecific) extractant of soil organic matter, it was expected that the soil activity concentrations derived using this extractant (Figs. 4.42– 4.44) would be greater than those derived from the water extractions. However, the difference in 36 Cl extractability between the water and sodium hydroxide was inconsistent. In the Wellesbourne soil, very little difference between the water and sodium hydroxide extractions was observed. Nevertheless, it is interesting to note that the organic matter content of the Wellesbourne sub-soil was much greater (4.1% OM content) than the other two soils (1.5 and 2.8% OM content) (see Chapter 2). Largest differences between the two extractants were found for the Robertgate soil, where higher concentrations of 36 Cl were generally observed when sodium hydroxide was used. Although this was not the case at all depths in the profile, it may indicate that some association of 36 Cl with the organic components of the soil occurred. However, based on the measured soil organic matter contents (shown in Chapter 2) it is not clear why this should have occurred for this soil unless the quality, rather than quantity, of the organic matter was a significant factor.
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
140
Biosphere Implications of Deep Disposal of Nuclear Waste Soil activity concentration (Bq g-1) 0
1
2
3
4
5
6
7
0 5
Silw ood
Soil depth (cm)
10
Robertgate
15
Wellesbourne
20 25 30 35 40 45 50
Fig. 4.43. Mean soil 36 Cl activity concentrations (sodium hydroxide extractable) for the Phase III intact soil column experiment at 3 months.
Soil activity concentration (Bq g-1) 0
1
2
3
4
5
6
7
0 5 10 Soil depth (cm)
May 3, 2007
15 20 25 30
Silw ood Robertgate Wellesbourne
35 40 45 50
Fig. 4.44. Mean soil 36 Cl activity concentrations (sodium hydroxide extractable) for the Phase III intact soil column experiment at 6 months.
4.3.3.2.
Ryegrass biomass and activity concentrations
Mean values for the yields of ryegrass dry biomass over the 6 months of the intact soil column experiment are shown in Fig. 4.45. Generally, these mean values were between 0.5 and 2.0 g of dry weight for all three soils, with the lowest values observed in Months 3 and 4.
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
141
Ryegrass dry biomass (g)
2.5
2
1.5
Silw ood Robertgate Wellesbourne
1
0.5
0 Month 1
Fig. 4.45. columns.
Ryegrass activity concentration (Bq g-1)
May 3, 2007
Month 2
Month 3
Month 4
Month 6
Mean dry ryegrass yields at each month for the Phase III intact soil
14000 Silw ood 12000 10000
Robertgate Wellesbourne
8000 6000 4000 2000 0 Month 1
Month 2
Month 3
Month 4
Month 6
Fig. 4.46. Mean 36 Cl activity concentrations for ryegrass harvested from the Phase III intact soil columns.
Activity concentrations for the ryegrass shoots are shown in Fig. 4.46. Activity concentrations tended to increase over time for each of the soils. In Month 1, only the Silwood soil showed a measurable activity concentration for the ryegrass. This soil continued to show the greatest ryegrass activity concentrations over the course of the experiment, up to a maximum of almost 12 000 Bq g−1 in month 6. The
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
142
Total 36Cl uptake into ryegrass (Bq)
May 3, 2007
Biosphere Implications of Deep Disposal of Nuclear Waste 100000 R2 = 0.80
90000 80000 70000 60000 50000 40000 30000 20000 10000 0 0
2000
4000
6000
8000
10000
Total solution influx (ml)
Fig. 4.47. Relationship between total solution influx and total uptake of by ryegrass in the Phase III intact soil column experiment. (p < 0.002)
36
Cl
Robertgate and Wellesbourne soils showed similar activity concentrations to one another, peaking at around 4000–6000 Bq g−1 . Because 36 Cl is generally expected to remain in the soil solution, the transpiration of water through the plant is likely to be strongly related to 36 Cl uptake. If the total water use by the columns is taken as a proxy for transpiration (although this includes both evaporation and transpiration) it can be related to the total uptake of 36 Cl by the ryegrass from a particular column. This relationship is shown in Fig. 4.47. Similarly, total ryegrass biomass for a column seemed to be related to the total uptake of 36 Cl (Fig. 4.48). These relationships suggest that higher yielding plants led to greater water use (transpiration) and hence greater 36 Cl uptake. It also seems apparent that plants below a certain yield (around 2 g dry biomass) were limited in their ability to take up 36 Cl. As seen previously, however, the pattern of contamination within the soil is likely to have a dominant bearing on the uptake of 36 Cl into the plant since uptake can only occur if the plant roots are exploiting contaminated soil. Over time, the mobile nature of the 36 Cl meant that it was transported to, and often accumulated at, the soil surface. Therefore, uptake into the ryegrass could also be expected to increase over time, since the upper 10–20 cm of soil is where the ryegrass roots would be most likely to take up solutes (due to their relative abundance in this zone,
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Total 36Cl uptake into ryegrass (Bq)
Radiochlorine
143
100000 90000
R2 = 0.70
80000 70000 60000 50000 40000 30000 20000 10000 0 0
2
4
6
8
10
12
Total ryegrass dry biomass (g)
Fig. 4.48. Relationship between total ryegrass biomass and uptake of the ryegrass in the Phase III intact soil column experiment. (p < 0.02)
36
Cl into
Root density (cm cm-3) 0
10
20
30
40
50
60
0 5 10 Soil depth (cm)
May 3, 2007
15 20 25 30 35 40 45
Silw ood Robertgate Wellesbourne
50
Fig. 4.49. Mean root density profiles for the 6 month intact soil columns (Phase III).
Fig. 4.49). This is seemingly evidenced in the Silwood soil columns where both the accumulation of 36 Cl at the soil surface and uptake into the ryegrass were greater than for the other two soils. However, mean weighted transfer factors (see Sec. 4.3.1.2) for the 6 month columns are shown in Fig. 4.50, and show that transfer was greatest in the Wellesbourne soil.
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
144
Biosphere Implications of Deep Disposal of Nuclear Waste
Wellesbourne
Robertgate
Silw ood
0
10000
20000
30000
40000
50000
60000
70000
Transfer factor
Fig. 4.50. Mean weighted soil-plant transfer factors (dry weight) for the 6 month intact soil columns (Phase III).
An overall activity balance comparing the total input of 36 Cl to a column, to that recovered at the end of the experimental period for that column showed a range of 56 to 217% recovery. The mean recovery was 90%. Of the recovered activity in the 6 month columns, between 32 and 76% was found in the soil, between 7 and 66% in the ryegrass and between 1 and 12% on the substrate components (polythene beads, geotextile material and silicone sealant). 4.3.3.3.
Soil transport simulations
Modelling of the column data was carried out using the approach outlined in Chapter 3, whereby steady-state soil water fluxes and moisture contents were obtained directly from the measured data and used to drive the 36 Cl transport simulations. Only results from the transport simulations are described here. A summary of the parameter values used for the transport simulations of 36 Cl in these soil columns is provided in Table 4.6. Although average data for replicate columns have thus far been reported, model simulations of individual columns were carried out in order to relate specific hydrological fluxes to 36 Cl transport. Each of the 6 month columns was modelled
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
145
Table 4.6. SLT1 model parameter values used for Phase III intact column simulations. Parameter ∗ Soil diffusion coefficient (Dm ) Soil dispersivity (dL ) Decay constant Sorption coefficient (Kd ) calibration parameter Root uptake coefficient (α ) calibration parameter
Units
Value
cm2 s−1 cm s−1 cm3 g−1
1.0 × 10−6 1.0 7.3 × 10−14 0.01
cm s−1
2.0 × 10−10 , 1.0 × 10−9 , 2.0 × 10−9
Table 4.7. Identifying codes for the 6 month intact soil columns (Phase III). Identifiers were used in the individual model simulations. Column Code 6M01 6M03 6M05 6M07 6M09 6M11
Column Description Silwood soil, Replicate A Silwood soil, Replicate B Robertgate soil, Replicate A Robertgate soil, Replicate B Wellesbourne soil, Replicate A Wellesbourne Soil, Replicate B
and each column was referred to by a unique identity code. These codes are shown in Table 4.7. Following previous simulations of the lysimeter experiments, a constant Kd value of 0.01 cm3 g−1 was employed. This gives a retardation value of around 1.01 and means that, in effect, there is no dependency on sorption due to anion exchange. The importance of plant uptake in determining the fate of 36 Cl was also observed in experimental results where it was found that up to 66% of the recovered activity was present in the ryegrass crop. An important consequence of this high uptake is that uptake by the plant root system perturbs the soil concentrations. Therefore, in seeking to simulate the behaviour of 36 Cl in the soil columns, three root uptake parameter values (α ) were selected (as shown in Table 4.6) and their effect on simulated soil activity concentrations evaluated.
ch04
14:33
146
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
The plots of observed and simulated soil activity concentrations at 6 months (Figs. 4.51–4.53) show that the observed concentrations are reasonably well reproduced over most of the column’s depth. However, the spread in the simulated results envelope due to variations in the root uptake parameter is generally quite low, meaning that for many of the columns the envelope lies outside the observed values.
Column 6M01 0 5
Soil depth (cm)
10 15 20 25 30
Water extraction
35
1M NaOH extraction
40
Simulated Max/Min Simulated
45 50 0
2
4
6
8
10
12
14
16
Radionuclide concentration (kBq/kg)
Column 6M03 0 5 10
Soil depth (cm)
May 3, 2007
15 20 25 30
Water extraction
35
1M NaOH extraction
40
Simulated Max/Min
45
Simulated
50
0
2
4
6
8
10
12
14
16
Radionuclide concentration (kBq/kg)
Fig. 4.51. Simulated and observed soil 36 Cl activity concentrations for the 6 month, Phase III, intact, Silwood soil columns.
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
147
Soil depth (cm)
Column 6M05
0 5 10 15 20 25 30 35 40 45 50
water extraction 1M NaOH extraction Simulated Max/Min Simulated
0
2
4
6
8
10
12
14
16
Radionuclide concentration (kBq/kg) Column 6M07
Soil depth (cm)
May 3, 2007
0 5 10 15 20 25 30 35 40 45 50
water extraction 1M NaOH extraction Simulated Max/Min Simulated
0
2
4
6
8
10
12
14
16
Radionuclide concentration (kBq/kg)
Fig. 4.52. Simulated and observed soil 36 Cl activity concentrations for the 6 month, Phase III, intact, Robertgate soil columns.
Nevertheless, it should be emphasised that the observed concentrations given do not include measurement errors which are likely to further widen the spread of the observed values. The shape of the observed and simulated profiles is roughly uniform from the base to around 10 cm below the surface. This is not unduly surprising as it represents the migration of a uniform lower boundary concentration up the entire soil column. However, marked differences can be seen in the top 10 cm of, even duplicate, soil columns. For example, in column
ch04
14:33
148
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste Column 6M09 0 5
Soil depth (cm)
10 15 20 25 30
Water extraction
35
1M NaOH extraction
40
Simulated Max/Min
45
Simulated
50 0
1
2
3
4
5
6
7
8
9
10
Radionuclide concentration (kBq/kg) Column 6M11
Soil depth (cm)
May 3, 2007
0 5 10 15 20 25 30 35 40 45 50
Water extraction 1M NaOH extraction Simulated Max/Min Simulated
0
1
2
3
4
5
6
7
8
9
10
Radionuclide concentration (kBq/kg)
Fig. 4.53. Simulated and observed soil 36 Cl activity concentrations for the 6 month, Phase III, intact, Wellesbourne soil columns.
6M03 (Silwood soil replicate B) the profile is approximately uniform up to the surface, whereas in column 6M01 (Silwood soil Replicate A) there is a rise in the 36 Cl soil concentrations, which in the top 5 cm are almost 3 times greater than the values in the lower 40 cm. By contrast, the model simulation shows very large soil concentrations, which at the soil surface are about an order of magnitude greater
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
149
than the values lower down. These enhanced values result from a concentrating effect, as water is lost at the surface due to evapotranspiration, leaving 36 Cl to accumulate. The discrepancy between the simulated and observed values may indicate a volatilisation process taking place near the soil surface, possibly biologically induced, which provides a mechanism for 36 Cl to be lost from the columns. An alternative explanation is that root uptake of 36 Cl is being underestimated by the model in the top 10 cm of the soil. Some support for this can be drawn from the lower concentrations obtained using the higher root uptake coefficient. One possible contribution to this is that the root densities near the surface were perhaps under-estimated. The problem with this explanation is that it would produce even higher simulated crop uptakes and, in the case of 6M03, lead to an even greater model over estimation of total uptake, which was already almost 2.5 times greater than that observed (see below). Also, this would be at odds with the good performance of the model in the other cases, where simulated total crop uptake of 36 Cl was generally within 15% of that observed. In the case of the Robertgate soil columns, the lower observed water fluxes did not result in such marked accumulations of 36 Cl at the surface in the simulated results, although, there is a suggestion that this effect may have occurred in the observed data. In the case of column 6M05 (Robertgate soil Replicate A), the model tended to over-estimate the soil concentrations and this may have been due to an under-prediction of crop uptake. The model simulation for column 6M07 (Robertgate soil Replicate B) generally shows good agreement with the observed data. It is worth noting that this column had the best individual inventory balance (closure to within 4%) and that the simulated total uptake of 36 Cl was within 1% of the observed. The Wellesbourne soil columns had the lowest water flux rates, and this was reflected in the observed 36 Cl soil concentration profiles in both columns, where there was a steady decline from the base of the columns to very low values near the surface. In the case of column 6M09 (Wellesbourne soil Replicate A), the model
ch04
May 3, 2007
14:33
150
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
reproduced the observed profile extremely well. However, for column 6M11 (Wellesbourne soil Replicate B) the simulated soil concentrations are larger than the observed and also show an enhancement effect in the top 5 cm. The reason for this appeared to be the very low observed root densities (data for individual columns not shown). These were between 1 and 2 orders of magnitude lower than those observed in column 6M09 and produced a simulated total crop uptake which was only 45% of that observed.
4.3.3.4.
Ryegrass uptake simulations
During the six month column experiments, five harvest cuts to a level of 5 cm were taken of the grass sward. The final cut also included the remaining 5 cm of stubble. The harvest times were at 1 month, 2 months, 3 months, 41/2 months and 6 months. The total 36 Cl contents of the harvested crop (in Bq) were compared with the simulated crop uptakes and are shown in Figs. 4.54–4.55. For the two Silwood Park soil columns, column 6M01 was well reproduced by the model, whereas, as was alluded to above, the model had a tendency to over-predict the uptake in column 6M03 throughout the five cuts. In the case of the Robertgate columns, the model reproduced the low uptakes for 6M07, but tended to underpredict the values for column 6M05. As suggested earlier, this may have also led to the over-estimation in soil concentrations shown in this column and could have been due to low root densities. The ryegrass uptake observed in the Wellesbourne soil column, 6M09, was very well represented by the model, with the model over-predicting by just 11%. It is perhaps significant that this column, like 6M07, had a reasonable closure in the inventory balance (total recovery 79%). However, for the second Wellesbourne column, 6M11, the model tended to under-estimate the amount of 36 Cl uptake. As stated above, this seemed to be related to extremely low observed root densities. Overall, however, the results for a root sorption coefficient of 1.0 × 10−9 cm2 s−1 were encouraging, with simulated crop uptakes generally within 15% of the observed totals, and indicated a robust model performance, both in terms of the representation of
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
151
First Harvest
Crop Uptake (Bq)
2500 2000 1500
Observed
1000
Simulated
500 0 6M01
6M03
6M05
6M07
6M09
6M11
Second Harvest
Crop Uptake (Bq)
14000 12000 10000 8000
Observed
6000
Simulated
4000 2000 0 6M01 6M03 6M05 6M07 6M09 6M11
Third Harvest
Crop Uptake (Bq)
May 3, 2007
18000 16000 14000 12000 10000 8000 6000 4000 2000 0
Observed Simulated
6M01 6M03 6M05 6M07 6M09 6M11
Fig. 4.54. Simulated and observed total 36 Cl uptake by ryegrass in the Phase III intact soil columns over the first three harvests.
uptake over time and between different columns. The fact that a constant parameter value was employed throughout for three different soil types gives a degree of confidence in the transferability of these results to assessment applications.
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
152
Biosphere Implications of Deep Disposal of Nuclear Waste
Fourth Harvest
Crop Uptake (Bq)
25000 20000 15000
Observed
10000
Simulated
5000 0 6M01 6M03 6M05 6M07 6M09 6M11
Fifth Harvest 60000 Crop Uptake (Bq)
May 3, 2007
50000 40000 Observed
30000
Simulated
20000 10000 0 6M01 6M03 6M05 6M07 6M09 6M11
Fig. 4.55. Simulated and observed total 36 Cl uptake by ryegrass in the Phase III intact soil columns over the final two harvests.
4.3.4.
Nine month soil column experiment with differing levels of stable chlorine (Phase IV)
4.3.4.1.
Soil column profiles of
36 Cl
The mean activity concentration of 36 Cl (extracted with water) in the soil profiles at 3 and 9 months are given in Figs. 4.56 and 4.57, respectively. At 3 months there was little discernible difference in the vertical profiles of 36 Cl in relation to stable chlorine treatment. Each profile exhibited a gradual decline from a maximum activity concentration of approximately 4–5 kBq kg−1 at the base of the column to zero at a soil depth between 0 and 20 cm from the soil surface. Based on these data, a rate of migration can be calculated as 0.56 cm day−1 . At 9 months, instead of a gradual decline in 36 Cl activity concentration from the column base to the soil surface, more uniform activity
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
153
Soil activity concentration (Bq g-1) 0
1
2
3
4
5
6
Soil depth (cm)
0 5
1 mmol/kg
10 15
5 mmol/kg 25 mmol/kg
20 25 30 35 40 45 50
Fig. 4.56. Mean soil 36 Cl activity concentrations at each stable chloride treatment of the 3 month soil columns (Phase IV).
Soil activity concentration (Bq g-1) 0
2
4
6
8
10
0 5 10 Soil depth (cm)
May 3, 2007
15 20 25 30 35 40 45
1 mmol/kg 5 mmol/kg 25 mmol/kg
50
Fig. 4.57. Mean soil 36 Cl activity concentrations at each stable chloride treatment of the 9 month soil columns (Phase IV).
profiles had generally developed. However, as in the previous experiment, an accumulation of 36 Cl at the soil surface was clearly evident in some cases. This general trend in 36 Cl migration did not seem to be greatly affected by the addition of stable chloride. As before, with chlorine it is useful to consider the water flux through the columns in relation to upwards migration. Mean water fluxes are shown in Fig. 4.58. Again, there is evidence of higher water
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
154
Biosphere Implications of Deep Disposal of Nuclear Waste 8000 9 month 7000 Total solution influx (ml)
May 3, 2007
3 month
6000 5000 4000 3000 2000 1000 0 1 mmol/kg
5 mmol/kg
25 mmol/kg
Fig. 4.58. Mean water influx at each stable Cl treatment of the 3 and 9 month soil columns (Phase IV).
fluxes in columns in which relatively large amounts of 36 Cl were transported to the soil surface (i.e. in the 5 mmol kg−1 stable Cl treatment), suggesting that the 36 Cl had a strong affinity for the soil solution phase. However, it was also observed that, in some cases, sodium hydroxide yielded greater 36 Cl soil activity concentrations than the water extraction, perhaps again suggesting a role of organic matter in the sorption of 36 Cl (also suggested in the intact column studies above). Of the 12 soil columns in this experiment, 6 showed a greater recovery of 36 Cl from the soil when sodium hydroxide was used as the extractant. Recovery was increased by between 3 and 35% (mean 12%). The inconsistency in the increased recovery using sodium hydroxide makes it difficult to assess the importance of 36 Cl sorption onto soil organic matter. However, the apparently strong relationship between water flux and 36 Cl migration suggest that the process may not be of great importance in governing the overall flux of 36 Cl through the soil, though it may affect the degree of plant availability. 4.3.4.2.
Ryegrass biomass and activity concentrations
The mean biomass production of the columns at each month of the experiment is shown in Fig. 4.59. Over the first four months
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
155
2.0
Ryegrass dry biomass (g)
May 3, 2007
1.8
1 mmolkg-1
1.6
5 mmolkg-1
1.4
25 mmolkg-1
1.2 1.0 0.8 0.6 0.4 0.2 0.0 Month 1
Month 2
Month 3
Month 4
Month 5
Month 7
Month 8
Month 9
Fig. 4.59. Mean dry ryegrass yields at each month and for each stable chloride treatment of the 9 month soil columns (Phase IV).
of the experiment, the lowest stable Cl treatment gave the greatest biomass yields. Beyond this time, highest biomass yields were, in all but one case, found for the 5 mmol kg−1 stable Cl treatment. In all but two cases, the highest stable Cl treatment gave the lowest biomass yields, suggesting that the stable Cl hindered the growth of the ryegrass. In the two lower stable Cl treatments, biomass yields were comparable to those found in the intact column study (described above). No clearly identifiable trend in biomass production over time was seen. The mean activity concentrations for the ryegrass material over the course of the experiment are shown in Fig. 4.60. Detectable 36 Cl activity was found in ryegrass from all cuts at the lowest stable Cl treatment. Detectable 36 Cl was not found in the first cut at the 5 mmol kg−1 Cl treatment, nor in the first 3 cuts at the 25 mmol kg−1 Cl treatment. Activity concentrations in the shoots were, in all cases, greatest in the 1 mmol kg−1 Cl treatment, and lowest in the 25 mmol kg−1 Cl treatment. However, the differences between concentrations for the 1 mmol kg−1 and 5 mmol kg−1 treatments were small (though note the logarithmic scale in Fig. 4.60). In general, the activity concentrations in the higher stable Cl treatment were at least two orders
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
156
Ryegrass activity concentration (Bq g-1)
May 3, 2007
Biosphere Implications of Deep Disposal of Nuclear Waste 10000 1 mmol kg-1 1000
5 mmol kg-1 25 mmol kg-1
100
10
1
0.1 Month 1 Month 2 Month 3 Month 4 Month 5 Month 7 Month 8 Month 9
Fig. 4.60. Mean activity concentrations for ryegrass harvested each month and for each stable chloride treatment of the 9 month soil columns (Phase IV). Note log scale.
of magnitude lower than for the lowest treatment. This indicates that the presence of stable Cl severely limited the uptake of 36 Cl into the ryegrass. There was an apparently strong trend for activity concentrations to increase over time, as was seen in the intact column study. Again, this can be explained by the upward soil migration of the 36 Cl into the rooting zone of the ryegrass over the course of the experiment. The accumulation of 36 Cl towards the soil surface is likely to enhance further the uptake into the plant, due to the dominance of roots in this zone (Fig. 4.61). Once again, in this experiment, strong, positive relationships between total water use, ryegrass biomass, and 36 Cl uptake were observed (Figs. 4.62 and 4.63), indicating the importance of transpiration in controlling the uptake of 36 Cl into the ryegrass. Mean weighted transfer factors for the 9 month columns were calculated using the approach described in Chapter 2 and are shown in Fig. 4.64. Overall activity balances in this experiment ranged from 34 to 74% recovery, with a mean of 59%. These values represent an underrecovery of 36 Cl, however, the reason for this was not clear. Attempts were made to find the ‘lost’ activity, by carrying out extractions of column materials, but these were largely unsuccessful. The most
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
157
Root density (cm3 cm-3) 0
20
40
60
80
100
120
0 5
Soil depth (cm)
10 15 20 25 30 35
1 mmol/kg
40
5 mmol/kg
45
25 mmol/kg
50
Fig. 4.61. Mean root density profiles at each stable chloride treatment of the 9 month soil columns (Phase IV).
Total 36Cl uptake into ryegrass (Bq)
May 3, 2007
10000
1000
100
10
1 0
1000
2000
3000
4000
5000
6000
Total solution influx (ml)
Fig. 4.62. Relationship between total solution influx and total uptake of 36 Cl by ryegrass in Phase IV soil columns (note log scale). R2 value = 0.77 (p < 0.01).
likely sink of the activity was considered to be the bead substrate as replicate extractions of this material yielded very variable, and sometimes large, quantities of 36 Cl. Of the total activity recovered from the 9 month columns, the vast majority (85 to 98%) was found in the soil. Between 0 and 10% was found in the ryegrass (much lower than in the previous experiment) and between 2 and 5% in the substrate components.
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
158
Total 36Cl uptake into ryegrass (Bq)
May 3, 2007
Biosphere Implications of Deep Disposal of Nuclear Waste 10000
1000
100
10
1 0
1
2
3
4
5
6
7
8
Total ryegrass biomass (g)
Fig. 4.63. Relationship between total ryegrass biomass and total uptake of 36 Cl by ryegrass in Phase IV soil columns (note log scale). R2 value = 0.74 (p < 0.01).
1 mmol kg−1
5 mmol kg−1
25 mmol kg−1
0
5000
10000
15000
20000
25000
Transfer factor
Fig. 4.64. Soil-plant transfer factors (dry weight) for each stable chloride treatment of the 9 month soil columns (Phase IV).
4.3.4.3.
Soil transport simulations
Hydrological and transport model simulations of the soil columns were carried out as described in Chapter 3. The columns were treated, in the first instance, as individual realisations of a generic column experiment. Thus a ‘representative’ model was constructed
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
159
Table 4.8. Identifiers for the Phase IV varying-stable-chlorine soil columns (all Silwood soil). Identifiers were used in the individual model simulations. Column Code
Stable Cl Soil Concentration (mmol kg−1 soil dw)
Time to Destructive Analysis (months)
N3MA01 N3MA02 N3MB03 N3MB04 N3MC05 N3MC06
1.0 1.0 5.0 5.0 25.0 25.0
3 3 3 3 3 3
N9MA07 I9MA08 N9MB09 I9MB10 N9MC11 I9MC12
1.0 1.0 5.0 5.0 25.0 25.0
9 9 9 9 9 9
and then compared against the observed data. Again here, the individual columns were given identifying codes for use in the modelling simulations (Table 4.8). In seeking to develop such a generic representation of the column experiments, it was recognised that the total water influx (and therefore 36 Cl influx) into the column has a substantial effect on the values and distributions of contaminant concentrations. However, an inspection of the influx rates revealed that the columns could be divided into two categories designated low and high. The ‘low’ influx group (columns N3MA01, N3MC06, N9MA07, I9MB10, N9MC11, I9MC12) covered a range of values between 12.1 and 17.1 ml day−1 and had a mean value of 14.3 ml day−1 . The ‘high’ rate (columns N3MA02, N3MB03, N3MB04, N3MC05, I9MA08, N9MB09) had a range from 22.1 to 26.2 ml day−1 and averaged 22.7 ml day−1 . These mean values are equivalent to evapotranspiration rates of 0.8 and 1.2 mm day−1 , respectively. It should be noted that the labels ‘low’ and ‘high’ are relative, and the larger daily value of 1.2 mm day−1 was only about a third of the highest rate observed in the previous column experiment. Clearly, there was no effect of experimental treatment (i.e. stable chloride treatment) on whether high or low flux rates were observed.
ch04
May 3, 2007
14:33
160
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
A comparison of soil concentrations obtained by water and sodium hydroxide extraction indicated that there was no clear evidence of radiochlorine binding onto organic compounds present in the soil. This result is at variance with results published by Lee et al. (2001). In the absence of evidence of significant binding in this experiment, it was assumed that 36 Cl behaved in essentially a conservative manner. However, for continuity with previous studies in this programme, a small degree of interaction was included (Kd = 0.01), and the soil diffusion and dispersivity coefficients were set at 1.0 × 10−6 cm2 s−1 and 1 cm, respectively. An inspection of the measured ryegrass root density distribution showed a relatively large degree of variability between columns. However, in the model simulations that follow these were simplified to a uniform distribution of 50 cm cm−3 as well as representing approximately the mean ryegrass distribution it, in effect, makes the root uptake representation (which is the product of the root density ρr and the root sorption parameter α ) solely dependent on the vertically distributed soil water contaminant concentration. A summary of the parameter values used in the model simulations is provided in Table 4.9. Owing to the large inventory imbalances highlighted above, model simulations of the columns were developed in a progressive manner. Initial ‘benchmark’ simulations were undertaken for the two flux rates with a uniform water flux over the entire soil profile and with no uptake of 36 Cl by the crop (i.e. α = 0). The results for the 3 and 9 month columns are shown in Fig. 4.65. Apart from the 3 month low influx simulation, the results show enhanced concentrations of 36 Cl at the soil surface as water was evaporated and the radiochlorine accumulated in Table 4.9. SLT1 model parameter values used in the Phase IV varying-stable-chloride column simulations. Parameter ∗ Soil diffusion coefficient (Dm ) Soil dispersivity (dL ) Decay constant (λ) Sorption coefficient (Kd ) Root uptake coefficient (α )
Units cm2 s−1 cm s−1 cm3 g−1 cm2 s−1
Value 1 × 10−6 1.0 7.3 × 10−14 0.01 1.0 × 10−10 − 1.0 × 10−9
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
161
(a)
0 5 10
A01 A02 B03 B04 C05 C06 Low High
Depth (cm)
15 20 25 30 35 40 45 50 0
1
2
3 4 5 6 Activity conc. (kBq/kg)
7
8
(b)
0 5 A07 A08 B09 B10 C11 C12 Low High
10 15
Depth (cm)
May 3, 2007
20 25 30 35 40 45 50 0
2
4
6 8 10 Activity conc. (kBq/kg)
12
14
16
Fig. 4.65. Simulated and observed 36 Cl concentrations — benchmark simulation (a) 3 months, (b) 9 months (Phase IV).
the soil water. In the case of the 9-month simulations, these soil concentrations (not shown for purposes of clarity) were 66 and 140 Bq g−1 for low and high fluxes, respectively. It is clear that these values were greatly in excess of those observed in the columns. The largest value for the 0–5 cm depth layer was 14.6 Bq g−1 (Column B09) and values for the other five 9-month columns were all below 3.5 Bq g−1 . The results for the 3-month columns were not quite so dramatic, nevertheless apart
ch04
14:33
162
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
from column B04 (which had the highest daily influx rate of 26.2 ml day−1 ) all the other columns showed essentially no accumulation of 36 Cl at the surface. This is in marked contrast to the simulated values, which showed generally higher concentrations and significant amounts of surface accumulation. Although the simulated concentrations could be reduced through increasing uptake by the crop (Figs. 4.66 and 4.67),
(a)
0 5
A01 A02 B03 B04 C05 C06 0.0 2E-10 4E-10 1E-9
10
Depth (cm)
15 20 25 30 35 40 45 50 0
1
2
3 4 5 Activity conc. (Bq/g)
6
7
8
(b)
0 5 A01 A02 B03 B04 C05 C06 0.0 2E-10 4E-10 1E-9
10 15 Depth (cm)
May 3, 2007
20 25 30 35 40 45 50 0
1
2
3 4 5 Activity conc. (Bq/g)
6
7
8
Fig. 4.66. Simulated and observed 3 Month 36 Cl soil concentrations for a range of plant root uptake coefficients (cm2 s−1 ). (a) low flux rate; (b) high flux rate (Phase IV).
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
163
(a)
0 5 10
A07 A08 B09 B10 C11 C12 0.0 1E-10 2E-10 1E-9
Depth (cm)
15 20 25 30 35 40 45 50 0
2
4
6 8 10 Activity conc. (Bq/g)
12
14
16
(b)
0 5 10
Depth (cm)
May 3, 2007
15 20
A07 A08 B09 B10 C11 C12 0.0 1E-10 1E-10 1E-9
25 30 35 40 45 50 0
2
4
6 8 10 12 Activity conc. (Bq/g)
14
16
Fig. 4.67. Simulated and observed 9 Month 36 Cl soil concentrations for a range of plant root uptake coefficients (cm2 s−1 ). (a) low flux rate; (b) high flux rate (Phase IV).
accumulations of 36 Cl still occurred and the simulated profiles are at variance with observed data, particularly in the case of the 9 month columns. The general trend shown by these results was that increasing root uptake in order to reduce concentrations at the surface led to an under-prediction of soil concentrations within the middle portion of the profile. This led to the conclusion that the assumption of uniform flow throughout the soil column was incorrect. This was also supported
ch04
May 3, 2007
14:33
164
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
by low soil moisture contents in the upper 5 and 10 cm layer of the column soils which indicated limited water reaching the soil surface (data not shown). In order to represent the effect of water uptake by the ryegrass root system in response to losses of water through transpiration, the water flux field along the column was modified. Following an approach previously adopted with the Silwood lysimeters, it was assumed that the partitioning between evaporation at the soil surface and plant transpiration was 1:9 (i.e. 10% of the water influx to the column left by evaporation at the surface and the remaining 90% was transpired). However, it should be noted that simulations were found to be relatively insensitive to this ratio and a split of 1:3 (25% soil evaporation) gave very similar results. The effect on soil concentrations of water uptake by the root system where there is no commensurate uptake by the roots of solute (i.e. α = 0.0) is shown in Figs. 4.68 and 4.69. Water uptake by the root system introduces two effects. The first is that the uptake of water by the roots resulted in an increase in solute concentrations in the middle portion of the profile, which exceed those at the base of the column. The second is that the progressively reducing upward fluxes were unable to transport large quantities of 36 Cl to the surface. The result is much reduced accumulations at the surface. Even so, there is still a noticeable amount of accumulation shown in the 9 month high-flux simulation. Although the inclusion of distributed root water uptake over the entire depth of the column was able to reduce surface accumulations of radiochlorine, the resulting concentrations within the bulk of the soil profile were generally much higher than the observed values for both the 3 and 9 month columns. However, this is without the transfer of 36 Cl into the crop. Figures 4.68 and 4.69 also show the effect on simulated soil concentrations per unit soil mass for a range of root uptake coefficients between 1.0 × 10−10 and 1.0 × 10−9 cm2 s−1 . They demonstrate how the uptake of radiochlorine by the ryegrass crop modified the soil concentrations and brought them more into agreement with observations. It was not easy to define a single value that best reproduces the observed data. However, an inspection of
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
165
Depth (cm)
(a)
0 5 10 15 20 25 30 35 40 45 50
A01 A02 B03 B04 C05 C06 0.0 2E-10 4E-10 1E-9
0
1
2
3 4 5 6 Activity conc. (Bq/g)
7
8
(b)
0 A01 A02 B03 B04 C05 C06
5 10 15 Depth (cm)
May 3, 2007
20
0.0 1.6E-10 4E-10 1E-9
25 30 35 40 45 50 0
1
2
3
4
5
6
7
8
Activity conc. (Bq/g)
Fig. 4.68. Simulated and observed 3 month 36 Cl soil concentrations including water uptake by plant roots for a range of plant root uptake coefficients (cm2 s−1 ). (a) low flux rate; (b) high flux rate (Phase IV).
Fig. 4.68 indicates that a value of 2.0 × 10−10 cm2 s−1 gave a reasonable representation of the results for the 3 month columns, whereas a value between 1.0 × 10−10 and 1.6 × 10−10 cm2 s−1 was more appropriate for the 9 month columns (Fig. 4.69). It is also clear that there was a large amount of noise in the data and that there were various columns that showed large deviations
ch04
14:33
166
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste (a)
0 5 10
A07 A08 B09 B10 C11 C12 0.0 1E-10 2E-10 1E-9
Depth (cm)
15 20 25 30 35 40 45 50 0
2
4
6
8
10
12
14
16
Activity conc. (Bq/g) (b)
0 5 10 15
Depth (cm)
May 3, 2007
A07 A08 B09 B10 C11 C12 0.0 1E-10 2E-10 1E-9
20 25 30 35 40 45 50 0
2
4
6 8 10 Activity conc. (Bq/g)
12
14
16
Fig. 4.69. Simulated and observed 9 month 36 Cl soil concentrations including water uptake by plant roots for a range of plant root uptake coefficients (cm2 s−1 ). (a) low flux rate; (b) high flux rate (Phase IV).
from the simulated profiles. Perhaps the most striking examples of this were columns A02 and A08. The concentration profile for ‘high flux’ column A02 appeared to match very closely that of its counterpart column A01 (which is in the low flux category), in spite of having an average flow that is 50% greater. Tailoring the model to reproduce both of these distributions for the appropriate
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
167
water flux rates involved root uptake coefficients of 4.0 × 10−10 and 6.0 × 10−10 cm2 s−1 , respectively. Overall, it appeared that the generic model representation was able to reproduce the general characteristics of the observed data in spite of the noise that appeared to exist. However, this success relied on a substantial amount of the 36 Cl that enters the column being taken up by the ryegrass crop. This was not consistent with the observed data as discussed further below. 4.3.4.4.
Ryegrass uptake simulations
Figures 4.68 and 4.69 showed that, in order for the generic model to be able to reproduce the observed soil concentration data, a substantial amount of 36 Cl must have been removed from the soil into the crop. In Table 4.10 observed total plant uptakes for the twelve Table 4.10. Observed and simulated total Phase IV experiment.
36
Cl uptake by ryegrass in the
Time
Treatment
Column
Flux rate
Total uptake (Bq)
3 months
1 mmol kg−1
N3MA01 N3MA02 N3MB03 N3MB04 N3MC05 N3MC06 1.0 × 10−10 cm2 s−1 2.0 × 10−10 cm2 s−1 1.0 × 10−10 cm2 s−1 2.0 × 10−10 cm2 s−1
Low High High High High Low Low Low High High
168 84 47 7 147 0 5050 8960 7780 13700
N9MA07 I9MA09 N9MB09 I9MB10 N9MC11 I9MC12 1.0 × 10−10 cm2 s−1 2.0 × 10−10 cm2 s−1 1.0 × 10−10 cm2 s−1 2.0 × 10−10 cm2 s−1
Low High High Low Low Low Low Low High High
1693 3529 6897 2525 3 57 37 000 53 000 64 000 85 000
5 mmol kg−1 25 mmol kg−1 Simulated
9 months
1 mmol kg−1 5 mmol kg−1 25 mmol kg−1 Simulated
ch04
May 3, 2007
14:33
168
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
columns are compared with the simulated total crop uptake for the two flux rates for selected root uptake parameter values. Although there were differences in total observed ryegrass uptake between the various stable chloride treatments (particularly with the high value of 25 mmol kg−1 ), the most striking feature of the data was that the simulated ryegrass uptakes were between 1 and 4 orders of magnitude greater than observed in the experimental columns. It is this discrepancy which may help to explain the very poor 36 Cl inventory balances for the experimental columns. Tuning the model to fit the observed soil concentration data could only be achieved by transferring large quantities of radiochlorine into the plant. Before considering the observed uptakes in more detail, it is useful to make a comparison of the above result with those for the intact Silwood columns described previously. The best performing model simulations of the intact Silwood columns were undertaken using a root uptake parameter of α = 1.0 × 10−9 cm2 s−1 , whereas the best performing simulations here used values 5 or 10 times smaller. However, if a comparison of the root density profiles for the two experiments is undertaken, there is a marked contrast, with the root densities here being at least 5 times greater, particularly over the bottom 40cm of the column, than previously. Thus, there was a possible hint that the root uptake model required some revision as it appeared that the key uptake control parameter was the product of the root density and root uptake parameter (i.e. ρr α ), which seemed to be comparable in both sets of column experiments. It is also instructive to point out that levels of observed 36 Cl uptake in the 6 month intact columns were generally in the range of 5000–10 000 Bq, and in the case of column 6M01 up to 50 000 Bq. These values were comparable with those produced by the model simulations. However, the values observed in the current experiment were less than 200 Bq in the case of the 3 month columns and less than 7000 Bq (and generally less than 3500) for the 9 month columns. Unfortunately, it was not possible to resolve this discrepancy between the experiments. Specifically, radiochemical analyses of the ryegrass samples were quality assured with spiked test samples in order to determine the yield coefficient. The results of 70%
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
169
in the current experiment were comparable with those of between 60 and 70% found previously. In addition, the calculation method for the final form of the results was checked and found to be in order. Thus, there were no obvious grounds for rejecting the observed values. Nevertheless, these clearly indicate that, if correct, there was a further mechanism that was either transporting or storing 36 Cl in substantial quantities. However, it is difficult to speculate on what this mechanism might have been. Losses due to volatilisation were considered. However, it is difficult to see how such large quantities of radiochlorine (and presumably stable chlorine as well) could undergo such a transformation. 4.3.5.
Six month soil column experiment with fixed and fluctuating water tables (Phase VII)
4.3.5.1.
Soil column profiles of
36Cl
Mean 36 Cl activity concentrations of the soil (extracted with water) at 2, 4 and 6 months are shown in Figs. 4.70 and 4.71. After 2 months (fluctuating water table) 36 Cl had migrated up to around
Soil activity concentration (Bq g-1) 0
1
2
3
4
5
6
7
8
0
Soil depth (cm)
May 3, 2007
5
2 month
10
4 month
15 20 25 30 35 40 45 50
Fig. 4.70. Mean soil 36 Cl activity concentrations (extracted with water) at 2 and 4 months in the Phase VII experiment (fluctuating water table).
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
170
Biosphere Implications of Deep Disposal of Nuclear Waste Soil activity concentration (Bq g-1) 0
2
4
6
8
10
12
14
0 5 10 Soil depth (cm)
May 3, 2007
15 20 25 30 35 40
Fluctuating
45
Fixed
50
Fig. 4.71. Mean soil 36 Cl activity concentrations (extracted with water) at 6 months for the fluctuating and fixed water tables in the Phase VII experiment.
15 cm depth, giving a migration rate of around 0.58 cm day−1 . Activity concentrations generally increased with increasing depth. After 4 months (fluctuating water table), the 36 Cl concentrations again generally increased with increasing depth but had reached the soil surface and started to accumulate. Therefore, greater activity concentrations were found in the 0–5 cm soil layer than in the soil directly below this. At 6 months and with a fluctuating water table, a uniform 36 Cl soil profile (4–5 Bq g−1 ) had become established between 5 and 50 cm depth. In the uppermost 5 cm, the accumulation of 36 Cl had become more significant. This accumulation was due to migration of 36 Cl solution to the soil surface followed by evaporation of water only, leaving the 36 Cl behind. This process was observed in the previous soil column experiments using 36 Cl. The marked effect of the manipulation of the water table height can be seen by comparing the 36 Cl profiles of the fixed and fluctuating water table columns (Fig. 4.71). At 6 months, the fixed water table led to the 36 Cl only just reaching the soil surface, giving a migration rate of 0.28 cm day−1 . Generally, increased activity concentrations were observed with increasing depth. Clearly the forcing of solution into the base of the columns by increasing the water table height
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
171
6 month fluctuating
6 month fixed
4 month fixed
2 month fixed
0
1000
2000
3000
4000
5000
6000
Total solution influx (ml)
Fig. 4.72. Mean total solution influxes to the fixed and fluctuating water table soil columns at each time interval (Phase VII).
led to a much faster movement of 36 Cl to the soil surface. Over the 6 months, mean total water influx to the fixed water table columns was 3450 ml and to the fluctuating water table columns was 5360 ml (Fig. 4.72). This is consistent with the previously observed effect of water flux into the columns on 36 Cl migration (described above) and implies that the 36 Cl had a strong affinity for the soil solution phase. In contrast to previous experiments, this experiment had soil solution samplers at various heights within the soil columns, thereby allowing for a consideration of 36 Cl levels measured directly in the soil solution. 4.3.5.2.
Soil solution activity concentrations
The 36 Cl activity concentrations in the soil solution of the 6-month columns over the course of the experiment are shown in Figs. 4.73 and 4.74. For the fixed water table columns, the 36 Cl concentrations decreased with decreasing depth as the incoming solution became diluted by the non-active water already in the column system prior to dosing the columns with 36 Cl. However, at each sampling depth, the activity concentration tended to increase over time as the 36 Cl input accumulated. The fluctuating water table also led to increases in 36 Cl activity concentrations at each depth over time, although
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
172
Biosphere Implications of Deep Disposal of Nuclear Waste 36
0
Cl activity concentration in soil solution (Bq ml-1) 5 10 15
20
0
10
Soil depth (cm)
0.5 months 20
1.5 months 2.5 months 3.5 months
30
4.5 months 6 months
40
50
Fig. 4.73. Mean soil solution 36 Cl activity concentrations for the fixed water table columns over the course of the Phase VII experiment.
36
0
Cl activity concentration in soil solution (Bq ml-1) 5
10
15
20
0
10 0.5 months
Soil depth (cm)
May 3, 2007
20
1.5 months 2.5 months
30
3.5 months 4.5 months 6 months
40
50
Fig. 4.74. Mean soil solution 36 Cl activity concentrations for the fluctuating water table columns over the course of the Phase VII experiment.
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
173
these increases were generally much greater than those observed for the fixed water table. This was presumably due to the greater input of 36 Cl solution to these columns over the course of the experiment (a higher water influx due to the manipulation of the water table). Thus, the dilution effect of the non-active solution initially present in the column was overcome much more rapidly. This led the slope of the activity depth profile to become steep relatively quickly and thus approach a uniform vertical distribution throughout the soil column. In general, 36 Cl activity concentrations in the soil solutions were relatively high in relation to the input solution activity concentration (20 Bq ml−1 ) but never actually reached this level. This provides further evidence that some removal of 36 Cl from soil solution took place. 4.3.5.3.
Ryegrass biomass and activity concentrations
Mean ryegrass biomass yields for the columns over the course of the experiment (Fig. 4.75) were generally lower than were found for the previous experiments (described above). The reason for this is not clear but may have been due to a lower light intensity in the experimental facility. No clear effects of water table treatment or time, on the biomass yields, were observed. 0.6 Fixed
Ryegrass dry biomass (g)
May 3, 2007
Fluctuating
0.5 0.4 0.3 0.2 0.1 0.0 Month 1
Month 2
Month 3
Month 4
Month 5
Month 6
Fig. 4.75. Mean dry biomass yields from the fixed and fluctuating 6 month columns (Phase VII).
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
174
Ryegrass activity concentration (Bq g-1)
May 3, 2007
Biosphere Implications of Deep Disposal of Nuclear Waste 1200 Fixed 1000
Fluctuating
800 600 400 200 0 Month 1
Month 2
Month 3
Month 4
Month 5
Month 6
Fig. 4.76. Mean 36 Cl activity concentrations of ryegrass from the fixed and fluctuating 6 month columns (Phase VII).
The mean activity concentrations of the ryegrass over the course of the experiment are shown in Fig. 4.76. In all but the first month, the fixed water table columns gave lower shoot activity concentrations than the fluctuating water table columns. In the fluctuating water table columns, the shoot activity concentrations showed a general trend to increase over time, although they did ‘tail off’ in Month 6. The increased movement of water through the fluctuating water table columns therefore seems to have transported 36 Cl into the rooting zone, and hence the ryegrass, much more effectively. In general, a relatively strong positive relationship between water flux through a column and 36 Cl uptake by the ryegrass was observed (Fig. 4.77). The uptake observed in the fixed water table columns was somewhat lower than that observed in the previous soil column experiment (differing stable chlorine levels) and, particularly, that in the intact soil column experiment. This may have been due to the relatively low ryegrass biomass yields found for these columns compared with previous experiments. However, in this experiment, a clear relationship between biomass and total uptake of the ryegrass was not seen. The distribution of ryegrass roots in the soil columns is shown in Table 4.11. Roots were only found in the top 10–15 cm. However,
ch04
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
Total 36Cl uptake into ryegrass (Bq)
May 3, 2007
175
1200 1000 800 600 400 200 0 0
1000
2000
3000
4000
5000
6000
7000
Total solution influx (ml)
Fig. 4.77. Relationship between total solution influx and total uptake of by ryegrass in the Phase VII experiment. R2 value = 0.93 (p < 0.001).
36
Cl
Table 4.11. Mean root density (g dry root kg dry soil−1 ) at each depth interval for the fixed and fluctuating water table columns (Phase VII).
Fixed Fluctuating
0–5 cm
5–10 cm
10–15 cm
15–50 cm
0.85 0.66
0.46 0.22
0.11 None detected
None detected None detected
this does not preclude a small number of roots having penetrated to greater depths. Such roots may not have been detected by the approach used to extract the roots from the soil. Based on the observed root distribution, mean weighted transfer factors for the 6 month columns were calculated using the approach described in Chapter 2. These were 324 for the fixed water table columns and 1081 for the fluctuating water table columns. In terms of overall activity balances, these column experiments gave acceptable recoveries ranging from 63 to 116%, with a mean of 84%. Of the total amount of activity recovered from within the 6 month columns, 90 to 97% was found within the soil, less than 1% in the ryegrass, and the remainder in the substrate components.
ch04
May 3, 2007
14:33
176
4.3.5.4.
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Soil transport simulations
The modelling of the soil columns was carried out as described in Chapter 3. The model domain included the soil column along with two additional components, representing the beads substrate (called Sub1) and the moveable reservoir, with its associated tubing and connectors (called Sub2). In the case of the beads substrate, partitioning between sorbed and solution phases are denoted as ASub1 and SSub1, respectively. It was decided to incorporate these into the model in order to represent any losses due to sorption effects. This also enabled sources of any observed inventory imbalances (shown in the following figures by the term Miss) of the total amount of radionuclide activity entering the system from the Marriot bottle (denoted as TotIN) to be investigated. The previous modelling studies of 36 Cl migration and uptake in experimental soil columns had only been able to use final soil concentrations for model calibration and performance assessment. However, the addition of the hollow fibre solution samplers (HFSS) provided an important alternative dataset to the destructive soil analyses. In particular, the ability to sample solution concentration measurements in both space and time meant that there was the opportunity to assess the temporal dynamics of the SLT model simulations of the experimental columns. The importance (and power) of this additional source of data for simulating the transport and uptake of 36 Cl is illustrated below using data from one of the six month columns with a fixed water table. 4.3.5.4.1.
Simulation 1
The initial simulation utilised the soil parameter values shown in Table 4.12. However, in order to account for the observed sorption of 36 Cl by the sealant in the underlying beads substrate, a Kd of 0.2 l kg−1 was introduced for this component. Furthermore, for this simulation it was assumed that there was no plant uptake. Hence the root uptake coefficient, α , was set equal to zero. This is reasonable as the amount of 36 Cl detected in the ryegrass was comparatively low (i.e. < 1% of the total inventory).
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine Table 4.12.
177
Initial parameter values used in the Phase VII simulations.
Parameter Molecular diffusion Dispersivity coefficient Decay constant Soil sorption coefficient Plant root uptake
Symbol
Units
Value
∗ Dm dL λ Kd α
cm2 s−1 cm s−1 cm3 g−1 cm2 s−1
4.0 × 10−6 1.0 0.0 0.01 Variable, see text
Fig. 4.78. Simulated and observed 36 Cl inventory for the example 6 month column (fixed water table) (Simulation1) (Phase VII).
An inspection of the system inventory (Fig. 4.78) shows the model simulation slightly underestimating the bead substrate concentration and markedly overestimating the amount of 36 Cl stored in the soil. However, it should also be observed that this excess was comparable to the amount of unaccounted 36 Cl for this particular column (which was just over 30% of the total input). An excess of radiochlorine in the soil column at the end of the experiment is also apparent from Fig. 4.79, where the simulated soil water activity concentrations were greater than those observed
ch04
May 3, 2007
14:33
178
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 4.79. Simulated and observed water extractable 36 Cl soil concentrations for the example 6 month column (fixed water table) (Simulation1) (Phase VII).
throughout the entire soil profile. However, what is also of interest was the model’s ability to reproduce the time-dependent concentrations obtained from the HFSS. Figure 4.80 shows a plot of solution concentrations at the six HFSS locations down the soil column and in the beads substrate (Sub1) combined with sample concentrations of the adjustable reservoir (Sub2). The results show that the concentrations in the reservoir reflect those of the Marriot bottle, which were at, or just slightly below, the target concentration of 20 cm3 g−1 . Furthermore, there is good agreement between simulated and observed substrate concentrations. This was achieved through the introduction of the sorption coefficient, Kd = 0.2 l cm3 g−1 . It is important to note that whilst this value produced a good fit to the sampled solution concentrations, this was achieved at the expense of reproducing the inventory analysis for this component. This discrepancy also demonstrates that there is a need to assess the worth and weight of the various forms of available data.
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
179
Fig. 4.80. Simulated and observed 36 Cl soil solution concentrations for the example 6 month column (fixed water table) (Simulation1) (Phase VII).
An inspection of Fig. 4.78 shows that the disagreement in the amount of 36 Cl stored in the beads substrate lay in the sorbed concentration. However, there was high variability, and hence uncertainty, in the analysis of the replicate samples used to determine this value. Therefore, it seemed reasonable to weight the model in favour of reproducing the HFSS data. Figure 4.80, therefore, indicates that the model correctly reproduces the input concentrations of radiochlorine into the soil column. This appears to be borne out by the soil water concentrations obtained from the lowest soil solution sampler in the soil column (at a depth of 47.5 cm), which shows good agreement between simulated and observed data during the first 150 days. However, after this time the observed soil water concentrations appear to level out, whereas the simulated values continue to increase towards the target concentration of 20 cm3 g−1 . Furthermore, a similar effect can be seen in the results for locations higher up the soil profile. As the
ch04
May 3, 2007
14:33
180
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
example column had a fixed water table, this result cannot be due to changes in water flux arising from the manipulation of the external reservoir. 4.3.5.4.2.
Simulation 2
In order to investigate the data further, it was decided to rerun the initial simulation with plant uptake added. Analyses of plant root densities (Table 4.6) indicated that the root biomass was predominantly located in the top 15 cm. However, in the following simulation it is assumed that there existed a uniform root density of 20 cm cm−3 down to a depth of 44 cm (i.e. just above the water table) with a root uptake coefficient set at 1.0 × 10−8 cm2 s−1 . The primary objective here was to provide a sink term which removes 36 Cl from the soil water. It is important to add that no inference was being drawn that the 36 Cl removed via this method was actually taken up by the ryegrass root system. Rather, the root uptake sub-model allows a first-order loss mechanism to be introduced and its effects explored. Figure 4.81 shows the inventory balance for the revised simulation. The main outcome is that the addition of root uptake transferred radionuclides from the soil into the “plant sink”. Consequently, there is better agreement between simulated and observed storage of 36 Cl in the soil column. The effects of the radionuclide transfer can also been seen in the plot of the final soil activity concentrations (Fig. 4.82). There is now improved agreement in the upper part of the soil profile, although the soil concentrations at the base of the column still exceed the observed values. Figure 4.83 shows the corresponding time series concentrations for the soil solution and substrate samples. It is interesting to observe that the loss of radionuclide from the soil solution results in improved agreement in concentrations at soil depths of 12.5, 22.4 and 32.5 cm. However, the final concentrations at the lower soil depths still exceed those obtained from the soil solutions. 4.3.5.4.3.
Simulation 3
A further simulation was therefore undertaken in which the rate of 36 Cl uptake was artificially enhanced at depth though manipulating
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
181
Fig. 4.81. Simulated and observed 36 Cl inventory for the example 6 month column (fixed water table) (Simulation2) (Phase VII).
Fig. 4.82. Simulated and observed water extractable 36 Cl soil concentrations for the example 6 month column (fixed water table) (Simulation2) (Phase VII).
ch04
May 3, 2007
14:33
182
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 4.83. Simulated and observed 36 Cl soil solution concentrations for the 6 month example soil column (fixed water table) (Simulation2) (Phase VII).
the root density profile. In this case, root densities were set as follows: 0–25 cm:10 cm cm−3 ; 25–35 cm: 20 cm cm−3 ; 35–48 cm: 40 cm cm−3 ; 48–50: 0 cm cm−3 and the root uptake coefficient (α ) was set at 5.0 × 10−9 cm2 s−1 . Figure 4.84 shows the effect on the simulated inventory balance. There was a slight increase in the amount of 36 Cl transferred from the soil into the plant and therefore a slight improvement in the agreement between simulated and observed storage in the soil profile. A plot of the final soil activity concentrations (Fig. 4.85) shows that the soil profile was reasonably well reproduced, although the simulated values tend to slightly exceed observed values. It is perhaps appropriate also to mention that there was clearly a peculiarity in the observed values over the depth interval 44–48 cm. These values appear to be anomalously low, but there was no clear reason for this.
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
183
Fig. 4.84. Simulated and observed 36 Cl inventory for the example 6 month column (fixed water table) (Simulation3) (Phase VII).
Fig. 4.85. Simulated and observed water extractable 36 Cl soil concentrations for the example 6 month soil column (fixed water table) (Simulation3) (Phase VII).
ch04
May 3, 2007
14:33
184
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 4.86. Simulated and observed 36 Cl soil solution concentrations for the example 6 month column (fixed water table) (Simulation3) (Phase VII).
Figure 4.86 shows experimental and modelled time-series concentrations for simulation 3. It reveals that, through careful selection of the plant uptake coefficient and root density profile, excellent agreement between the two sets of data was achieved. It also presents a dilemma in terms of a physical explanation for this inferred loss mechanism. 4.4.
General Discussion
Across all the lysimeter and column experiments, upwards migration of 36 Cl from the water tables through the soil was relatively rapid compared with other radionuclides studied within this programme. Over time, particularly in the absence of a downward water flux (i.e. in the soil columns), the 36 Cl profile in the soil tended to approach a uniform distribution, but with an accumulation of 36 Cl at the soil
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
185
surface. The rates of upward migration for 36 Cl observed across the experiments ranged from 0.28 to > 2 cm per day, but were most commonly around 0.6 cm per day. Whilst this rate provides a way of comparing the datasets, it is clearly dependent upon the hydrological conditions within the experiment and these conditions varied, particularly between the lysimeters and the columns. In many cases, high water use by the columns, driven by evapotranspiration, or an increase in water table height, could be related to a greater extent of 36 Cl migration. Similarly, in the lysimeters, the periods of greatest upwards migration were related to periods of low rainfall and high evapotranspiration. Periods of rainfall were often associated with losses of activity from the soil, presumably via leaching through the base of the lysimeter. In concurrence with this, chlorine is often assumed to have a Kd value of zero, i.e. it remains in the solution phase and follows the water flux. However, in the experiments described in this chapter where different soil types were used, differences in activity concentrations in the different soil types were observed. In addition, the periods of rainfall in the lysimeter experiments, although associated with leaching of 36 Cl, did not lead to complete removal of the radionuclide from the soil profile, suggesting that some of the 36 Cl was retained by the soil. This was thought to be due to adsorption of the 36 Cl onto the soil, probably onto organic matter, and suggests that a Kd value of zero is not correct. Because it was not possible to determine Kd values from the experiment directly due to the problem of quantifying the ‘total’ soil activity concentration, the best indication of a Kd value from these experiments can be obtained from the modelling parameters. The Kd value used throughout the modelling was 0.01 l kg−1 , indicating that a small amount of interaction with the solid phase was indeed required to simulate the experimental data. Uptake of 36 Cl into vegetation seemed to be dependent upon the distribution of activity within the soil in relation to root distribution, and evapotranspiration. Often a relationship between the activity concentrations of the upper 10 or 20 cm of soil and the ryegrass was seen. However in the lysimeters, this was seemingly tempered by the
ch04
May 3, 2007
14:33
186
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
net water flux. When the net water flux was positive, relatively high uptake was seen. At times of net negative water influx, relatively low uptake was seen, probably due to a combination of low evapotranspiration and a leaching of 36 Cl from the rooting zone of the crop. In the columns, good positive correlations between uptake, biomass and evapotranspiration were often seen. This suggests that a healthy, high yielding crop led to high transpiration and, hence, a high uptake of 36 Cl. As with the soil data, this indicates the strong influence of the evapotranspiration stream through the columns on the fate of 36 Cl. In the soil columns, once the 36 Cl had migrated upwards through the column into the rooting zone, uptake by the ryegrass appeared to be strongly related to evapotranspiration. The principal exception to this was where the soil was amended with increasing quantities of stable chlorine. This markedly reduced uptake of 36 Cl into the ryegrass. Soil-plant transfer factors for 36 Cl ranged from zero to around 50 000, with large differences in transfer factors observed between the experiments. Lowest values were generally observed in the soil column experiment comparing fixed and fluctuating water tables. Here, even with a fluctuating water table seeming to transport 36 Cl into the rooting zone very effectively, lower transfer factors than were previously observed for fixed water tables were found. The relatively low biomass yields of the ryegrass appeared to have been important in limiting uptake. The model simulations suggested that, in some cases, low plant uptake of 36 Cl in the latter two soil column experiments (those studying the effects of stable chlorine content and water table height) may have been associated with poor 36 Cl activity balances for the experiment. Only by simulating much greater uptake by the ryegrass could the observed soil activity concentrations be simulated by the model. However, the analytical procedures used for digestion of the plant samples were checked and no potential problems could be identified. This may suggest that the ‘lost’ activity was elsewhere within the column system but was not efficiently extracted. A further possible explanation is loss via volatilisation, however, the potential for loss via this pathway is considered to be very low.
ch04
May 3, 2007
14:33
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiochlorine
187
Overall, the clear tendency for 36 Cl to behave conservatively and follow the flux of moisture through the columns suggests that it is likely to of significant importance in relation to the risk associated with its potential migration from a radioactive waste repository to the biosphere, especially under conditions of a high water table or a high evapotranspirative (upwards) flux.
ch04
May 3, 2007
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
CHAPTER 5
Radioiodine
5.1. 5.1.1.
Background 129I
in radioactive waste
Radioactive isotopes of iodine in the environment occur as a result of nuclear weapons tests and the operation of nuclear facilities. Iodine has over 30 known isotopes, nine of which are produced in the fission cycle (Bulman & Cooper, 1986; Szabo et al., 1993). The most commonly studied isotopes include 127 I (the only stable isotope), 125 I, 129 I and 131 I. The tendency of iodine to accumulate in the human thyroid gland means that all forms of radioactive iodine have the potential to induce tumours of the thyroid (Coughtrey et al., 1983; Szabo et al., 1993; Muramatsu & Yoshida, 1995). The radionuclides 131 I and 129 I are of particular importance in direct release scenarios and the air-grass-cow-milk pathway, which is a dominant dose-impact pathway for 131 I following releases to the environment, has been studied intensively. Studies on the fate of 129 I, which has a 15.7 million year physical half life, are less abundant. In the terrestrial environment, the principal sink for stable iodine (127 I) is the soil and it is reasonable to assume that this will also be the case for 129 I. When considering long-term disposal of low and intermediate level wastes, the 15.7 million year half life of 129 I makes releases to the biosphere almost inevitable, regardless of containment measures (Bors & Martens, 1992; Szabo et al., 1993). This, combined with the inventory of 129 I within both low and intermediate level waste, makes 129 I an
188
ch05
May 3, 2007
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
189
important radionuclide in safety assessments of proposed radioactive waste repositories (Sheppard et al., 1993; Nirex, 1997).
5.1.2.
Iodine behaviour in soils and plants
Iodine belongs to Group VII of the periodic table, the halogens. This is the most reactive group of elements and includes fluorine, chlorine, bromine and iodine. The halogens are prominent anions in the environment, forming largely ionic molecules. Of particular importance, when considering the soil-sorption behaviour of iodine, is its chemical speciation. Muramatsu & Yoshida (1999) reported that the valence states of the major iodine species are: iodide, I− (−1); iodate, IO− 3 (+5); elemental iodine, I2 (0); and methyl iodide, CH3 I (+1). Geometric mean Kd values (cm3 g−1 ) for iodine in four generic soil types, obtained from Sheppard & Thibault (1990), are: Sand 1.0, Loam 5.0, Clay 1.0, Organic 25.0. Muramatsu et al. (1996) found sorption of iodine onto soils to be almost complete after only 3 days, although work by Sheppard et al. (1995) suggests a much slower sorption process, taking approximately 20 days. Such variations in time scales hint at the complexity of iodine sorption processes and highlight the necessity of examining soil characteristics in detail. The work of Raja & Babcock (1961) suggests that anion exchange plays no major role in the association of iodine with soil solids and that sorption processes are primarily mediated by organic matter. This has been questioned in view of results of more recent anion competition experiments (Sheppard et al., 1995), which suggest a more important role for anion exchange. The role of anion exchange in soil iodine behaviour, although recognised (Sheppard et al., 1995), is thought to be of secondary importance to organic matter mediated sorption behaviour (Whitehead, 1973, 1974; Tikhomirov et al., 1980; Bors et al., 1988). Bors, et al. (1991) found that soil clay minerals (kaolinite, montmorillonite and chlorite-illite) showed little propensity for iodine sorption. Whitehead (1974) found that freshly precipitated hydrated ferric oxide substantially sorbed iodine at pH < 5.5, but that sorption decreased to zero at neutral pH. The effects of soil
ch05
May 3, 2007
14:35
190
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
pH have also been studied by Schmitz & Aumann (1995), who found that both stable and radioactive iodine were more readily extracted from soil at pH 8.2 than at pH 4.8. The reduction-oxidation status of soil potentially plays an important role in determining iodine speciation. In turn, speciation appears to play an important role in determining the sorption behaviour of iodine in soil. For example, Muramatsu et al. (1996) studied the effects of waterlogging on the desorption of 125 I from soil and found that, following waterlogging, the redox potential decreased and led to the increased desorption of 125 I. Similarly, Sheppard & Hawkins (1995) found lower soil-sorption of added 125 I under anoxic conditions than under oxic conditions. Sheppard & Motycka (1997) found that such redox-induced decreases in iodine sorption led to higher concentrations of iodine in plants grown on flooded soil. However, there appears to be a paucity of information relating redox effects to the soil-transport of iodine. This may be particularly significant in the vicinity of a soil water table, where spatio-temporal variations in redox potential would be expected. Although information on natural ‘in situ’ iodinated organohalogen production is scarce, much work has been undertaken using reconstituted soil cores and small soil samples. Reported volatilisation from these studies has varied a great deal from 0.07 percent of added iodine over 60 days (Sheppard et al., 1995) to 57 percent over 30 days (Whitehead, 1981). Murumatsu & Yoshida (1995) suggest iodine is primarily volatilised in organic forms. This is supported by the gas chromatography work of Reineger (1977), which found the major volatilised iodine product to be methyl iodide; ethyl iodide and iodate were also detected, but in much smaller quantities. Formation of volatile forms of iodine is thought to be mediated by enzymes produced by lignocellulose degrading microorganisms, especially basiodiomycete fungi. The rate and total amount of volatilisation have been shown to decline with increasing soil organic matter content (Reineger, 1977; Whitehead, 1981; Sheppard & Hawkins, 1995). Volatilisation has been demonstrated to increase in planted soils. This has been attributed to both an accessible transport route to the atmosphere (via intercellular gas
ch05
May 3, 2007
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
191
spaces and the aerenchyma system) and reducing conditions present in the rhizosphere (Muramatsu et al., 1989). Iodine transfer factors (TFs) for a range of plant species have been determined. For example: Chinese white cabbage 0.063 (Yu et al., 2000), wild rice 0.17–0.25 (Sheppard & Motycka, 1997), generic species 0.11 (Sheppard & Evenden, 1997), grass 0.0034 (IAEA, 1994), beet 0.069–0.15 (Sheppard et al., 1993), cabbage 0.027–0.10 (Sheppard et al., 1993), maize 0.024–0.19 (Sheppard et al., 1993), grass 0.0018 (IUR, 1989), and various species 0.4 (Coughtrey et al., 1983). In a study of iodine uptake by wild rice (Zizania aquatica), Sheppard & Motycka (1997) observed that 125 I in flooded soil was an order of magnitude more soluble than in drained soil. The soil solid/liquid partition coefficient Kd , in flooded soil was about 0.6 cm3 g−1 whereas in drained soil it was about 6 cm3 g−1 . This tenfold difference in iodine solubility in the soil only led to a small difference in the values of transfer factors for iodine in the rice seed in flooded and drained soils, which were 0.25 and 0.17, respectively. However, Sheppard & Motycka (1997) measured transfer factors for iodine in the lower leaves of flooded rice plants of up to 69. The most recent study on soil-plant transfer of radioactive iodine has been that of Yu et al. (2000) who studied uptake from soil of 131 I by Chinese white cabbage (Brassica chinensis L.). These authors found that the soil-plant transfer factor for iodine depended on the mass of vegetable present (Chinese white cabbage can be grown at biomass densities up to 10 kg m−2 ), the growth period of the vegetable and season in which it was grown. The mean transfer factor determined by Yu et al. (2000) was 0.063, which is intermediate between the geometric mean values obtained from the IUR data base (IUR, 1989) and by Sheppard & Evenden (1997). In summary, therefore, 129 I requires investigation in the context of upwards soil migration because (a) it has a very long physical half life and so will persist in the environment, (b) it is potentially poorly sorbed by soils and so may exhibit a high degree of mobility, and (c) its behaviour is thought to be strongly affected by redox conditions, variations in which are likely to be significant throughout the geosphere and biosphere. Although plant uptake of iodine appears to
ch05
May 3, 2007
14:35
192
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
be low, it is still important to quantify this under the conditions of groundwater contamination, since this has rarely been done. 5.2.
Experimental Overview
The study of 129 I was carried out using a mini-column experiment and two separate large-scale soil column experiments. 125 I was used as a convenient, relatively short-lived (physical half-life ∼ 60 days), gamma emitting surrogate for 129 I. The general nature of each of these experiments is described in Chapter 2. All of these experiments were carried out using the Silwood sandy loam soil and 125 I added as sodium iodide in a carrier-free form. In the mini-column experiments, two moisture contents were used, 25% (gravimetric) and saturation, with each set up in triplicate. No fumigation treatment was imposed. Thus, a total of six mini-columns were established. The principal aim of this experiment was to determine radioiodine Kd values under conditions that were more realistic than those of the more ‘traditional’ batch sorption studies. Therefore, in order to compare the Kd data derived from the mini-columns with those from a more conventional approach, a batch sorption experiment was also set up, which also used the Silwood soil. Since the batch sorption experiment was not a general method, it is not described in Chapter 2. Details are, therefore, given here. 20 ml of 125 I-containing solution was added to triplicate 4 g samples of air-dried soil. Several 125 I activity treatments were established viz: 8.7, 17.4, 43.5, 87.0, 174.0, 348.0 Bq. The samples were then placed on a slow end-over-end shaker for 2 hours. At the start and end of this time, a platinum redox electrode and reference electrode (as described for the other experiments in Chapter 2) were inserted into several randomly selected samples to record the redox potentials. Following the 2 h shaking period, the samples were centrifuged at 2686 g for 15 mins. The supernantant solution was then filtered through a Whatman cellulose nitrate 0.22 µm filter before a 5 ml subsample was taken for radiochemical analysis. Soil adsorption was determined by the difference between the 125 I activity added initially and that remaining in the solution at the end of the experiment.
ch05
May 3, 2007
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
193
An adsorption isotherm was prepared by plotting adsorbed activity (Bq kg−1 ) against equilibrium solution concentration (Bq l−1 ). The slope of this linear relationship was taken as the Kd value. The data from both the mini-column and batch studies were decay corrected. The first large-scale soil column experiment (Phase V) was carried out over 12 months and using a fixed (at 45 cm depth) water table. Twelve columns were established with three columns destructively sampled every 3 months. All columns were vegetated with perennial ryegrass. The data from this experiment were not decay corrected. This was because several stock solutions (each with an initial activity concentration of 20 Bq ml−1 ) were prepared for use (addition to the Marriott bottles) over the course of the column experiment. At the times when these newly prepared solutions began to be added to the Marriott bottles, there was clearly a marked increase in the activity flux to the columns. Decay correcting the data back to, for example, Day 0 of the experiment (when the first stock solution had an activity concentration of 20 Bq ml−1 ) would therefore give a false impression of data from columns which had received significant contributions of activity from subsequent stock solutions. Essentially, there is no single date to which decay correction can be carried out. Therefore, 125 I activities for this experiment are given as those present on the day of sampling. The second column experiment (Phase VII) was run over 6 months and featured both fixed and fluctuating water tables. The fixed water tables were, again, maintained at 45 cm depth. The height of the fluctuating water tables was increased in 0.5 cm intervals from an initial depth of 45 cm to 30 cm depth at 3 months, and then decreased in 0.5 cm intervals to 45 cm depth at 6 months. The days upon which 0.5 cm adjustments were made were pre-determined in order to follow a sinusoidal curve. Eight columns were set up. Six had fluctuating water tables with two columns destructively sampled every two months. The further two columns were those with fixed water tables and these were destructively sampled at 6 months. All columns were vegetated with perennial ryegrass. In contrast to the first column experiment, all data from this experiment were decay corrected to Day 0 of the experiment. This could be done here because 125 I stock
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
194
Biosphere Implications of Deep Disposal of Nuclear Waste
solutions prepared subsequent to the first stock solution were made up with an activity concentration equal to that which the first stock solution would have had at that time. Essentially, this approach is the same as having a single stock solution used throughout the experiment.
5.3.
Results
5.3.1.
Mini-column and batch experiments
5.3.1.1.
Redox potentials
Over the course of the batch experiment, the soil samples remained oxic (with redox potentials between 300 and 400 mV). The adsorption isotherm resulting from the batch experiment is shown in Fig. 5.1. The slope of this relationship gives the Kd value for iodine in this system as 2.5 l kg−1 (cm3 g−1 ). In the mini-column experiment, the general trend in redox potential over time is shown in Fig. 5.2. The drier (25% gravimetric moisture content) treatment led to oxic conditions (around 400 mV) being
Soil activity concentration 125I (Bq kg−1)
May 3, 2007
30000 y = 2.5x - 694.5 R2 = 0.99 20000
10000
0 0
2000
4000
Equilibrium solution
Fig. 5.1.
125
6000 125
8000
10000
12000
I activity concentration (Bq l−1)
I adsorption isotherm onto the Silwood sandy loam soil.
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
195 25 % Moisture content
600
40 % Moisture content
500 Redox potential (mV)
May 3, 2007
400 300 200 100 0 -100
0
10
20
30
40
50
-200 Experiment day
Fig. 5.2. Mean redox potential at dry (25%) and saturated (40%) moisture contents in the mini-columns over the course of the experiment.
maintained over the course of the experiment. Such conditions were also observed in the saturated columns (40% gravimetric moisture content) over the first 30 days of the experiment. However, after this time the redox potential began to fall, reaching −100 mV by the end of the experiment. The relatively low redox potentials recorded in the saturated soils are typical of anoxic soils (Sposito, 1989). In addition, the results from the nitrate and sulphate analyses of the soil solutions at the start and end of the experiment offer a further insight into the redox chemistry of the soils. For example, from Fig. 5.3 it is evident that, in the drier columns, neither nitrate nor sulphate concentrations decreased significantly between the start and end of the experiment. However, in the saturated columns, nitrate concentrations decreased by around 60% between the start and end of the experiment. This suggests that, due to saturation, the redox status of the soil was such that nitrate was reduced, possibly to nitrite. However, the lack of a decrease in sulphate concentrations over the course of the experiment suggests that the redox potential was not sufficiently depressed (reaching −100 mV) to bring about sulphate reduction to sulphide. From these data, it is evident that the higher, saturated, moisture content only brought about the partial reduction of the soil.
ch05
14:35
196
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste 700
Nitrate Sulphate
600 Solute concentration (mg l−1)
May 3, 2007
500 400 300 200 100 0 25 % Start
40 % Start
25 % End
40 % End
Fig. 5.3. Mean nitrate and sulphate concentrations in the soil solutions extracted from the mini-columns at both moisture contents at the start and end of the experiment.
5.3.1.2.
Kd values
Kd values (Fig. 5.4) were determined for all columns on the various sampling occasions using Eq. (2.1). Total soil activity concentration was taken as the initial activity concentration measured in the soils before packing into the columns. Over the first two weeks of the experiment, Kd values were below 1 cm3 g−1 in both treatments, indicating a low degree of iodine sorption onto the soil. More specifically, over this period mean Kd values tended to be slightly lower in the drier soils than in the saturated soils (although at day 14 almost identical mean values were observed). Whilst in both treatments, the Kd value tended to increase with time, this increase was much greater in the drier treatment, particularly after three weeks. By the end of the experimental period, Kd values in the drier columns had increased to, and stabilised at, around 7 cm3 g−1 . The increase in Kd value in the saturated columns was much smaller, to a maximum of approximately 2 cm3 g−1 . Because the effects of time-dependent parameters (e.g. redox potential) are likely to be limited over the initial period of the experiment, the fact that Kd values were slightly lower in the drier soils
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
197
10 25 % Moisture content 40 % Moisture content Kd value(l cm3g-1)
May 3, 2007
1
0.1 0
10
20
30
40
50
Experiment day
Fig. 5.4. Mean (± 1 standard deviation) Kd values over the course of the minicolumn experiment. Note log scale.
than in the saturated soils over the first two weeks of the experiment (Fig. 5.4) suggests that the prevailing moisture content rapidly affected Kd values. The data indicate that at the lower moisture content, higher activity concentrations of 125 I were present in the soil solution (i.e. a lower Kd value was observed). This is likely to be due, at least to some extent, to dilution of soluble 125 I at higher moisture contents. Such a phenomenon has significant implications for the usefulness of batch techniques which use low soil:moisture ratios and which may result in lower solute concentrations and overestimates of Kd value. Indeed, this appears to be the case, as the Kd value determined here by the batch technique using a soil:solution ratio of 1:5 (2.5 cm3 g−1 ) was greater than the initial Kd values observed in the mini-columns (1 cm3 g−1 ). Interestingly, this effect of moisture content on soluble ion concentrations can also be seen in the nitrate and sulphate data (Fig. 5.3). For example, at the start of the experiment, when no redox-induced effect on ion concentrations would be expected, the saturated soil exhibited soil solution concentrations of both nitrate and sulphate that were lower than in the drier treatment. The fact that Kd values generally increased over the course of the experiment, illustrates the importance of contact time between iodine
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
198
Biosphere Implications of Deep Disposal of Nuclear Waste
and the soil. Again, this raises questions over the reliability of Kd values determined using batch techniques over short time periods. For example, in an oxic system, the current mini-column work suggests that a Kd value determined after one day may be around fifty times lower than a Kd determined after around 40–50 days. In assessing environmental fate, the behaviour of a contaminant in equilibrium with its environment is often of primary interest. This result suggests that using a Kd value derived from short-term experiments can result in an overestimation of iodine solubility. Again, this appears to be borne out by the results from the batch experiment in which the short contact time (2 hours) led to a much lower Kd value than was observed in the mini-columns at the end of the experiment (although this difference is offset by the effect of moisture content, as described above). Differences in the magnitude of the time-dependent increases in Kd value between the two moisture treatments are thought to be a result of the observed trends in redox potential. The relationship between these two variables is shown in Fig. 5.5. The data suggest that the falling redox potential in the saturated treatment maintained iodine solubility much more effectively than in the drier treatment. This should be considered in the context of iodine speciation.
9
25 % Moisture content
8
40 % Moisture content
7 Kd value (l cm3g-1)
May 3, 2007
6 5 4 3 2 1 0 -200
-100
0
100
200
300
400
500
600
Redox potential (mV)
Fig. 5.5. Relationship between redox potential and Kd value in the minicolumns. Direction of the arrows represents increasing time.
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
199
Of the two main inorganic species of iodine likely to exist in soil solution, the predominantly oxic form (iodate) is known to be relatively more strongly adsorbed onto the soil than the predominantly reduced form (iodide) (Murumatsu et al., 1996). Thus, in the saturated treatment, as redox potential fell, the Kd value was apparently prevented from increasing over time as significantly as it did in the drier columns, due to the chemical reduction of the iodine. 5.3.2.
Twelve month, fixed water table, soil column experiment (Phase V)
5.3.2.1.
Soil column profiles of
125 I
Mean (± 1 s.d.) migration of 125 I from the base up through the soil columns from the 12 month experiment is shown in Figs. 5.6–5.9 for the 3, 6, 9 and 12 month columns, respectively. At 3 months, highest activity concentrations were found in the bottom 2 cm of the soil columns. The concentrations generally decreased with decreasing depth up to 25–30 cm; above that point no 125 I activity was detected in the soil. By 6 months, radioactive decay of the 125 I within the
Activity concentration (Bq g-1) 0
1
2
3
4
5
6
0 5 10 Soil depth (cm)
May 3, 2007
15 20 25 30 35 40 45 50
Fig. 5.6. Mean (solid line) and one standard deviation (dashed line) soil profile of 125 I activity at 3 months in the 12 month soil column experiment (Phase V).
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
200
Biosphere Implications of Deep Disposal of Nuclear Waste Activity concentration (Bq g-1) 0
0.5
1
1.5
2
2.5
3
0 5
Soil depth (cm)
10 15 20 25 30 35 40 45 50
Fig. 5.7. Mean (solid line) and one standard deviation (dashed line) soil profile of 125 I activity at 6 months in the 12 month soil column experiment (Phase V).
Activity concentration (Bq g-1) 0
0.5
1
1.5
2
2.5
3
0 5 10 Soil depth (cm)
May 3, 2007
15 20 25 30 35 40 45 50
Fig. 5.8. Mean (solid line) and one standard deviation (dashed line) soil profile of 125 I activity at 9 months in the 12 month soil column experiment (Phase V).
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
201
Activity concentration (Bq g-1) 0.0
0.2
0.4
0.6
0.8
1.0
0 5 10 15
Soil depth (cm)
May 3, 2007
20 25 30 35 40 45 50
Fig. 5.9. Mean (solid line) and one standard deviation (dashed line) soil profile of 125 I activity at 12 months in the 12 month soil column experiment (Phase V).
column system led to lower activity concentrations being measured in the soil (it was not possible to decay correct the data in this experiment). However, of more interest than the actual activity concentrations is the pattern in 125 I distribution observed at this time. A zone of 125 I accumulation was clearly apparent at 20–30 cm depth. Activity concentrations within this zone were generally slightly greater than those found at the base of the soil columns. This accumulation indicates that the migration of 125 I was being arrested at this point, although small concentrations (< 0.1 Bq g−1 ) were observed above this region. At 9 months, the zone of 125 I accumulation between 20 and 30 cm was still evident. Interestingly, at 9 months, similar activity concentrations were observed throughout the profile as were observed at 6 months, i.e. radioactive decay had not led to a reduction in measured activity concentrations. This suggests that activity influx at the base of the columns compensated for radioactive decay of 125 I within the soil. In the 9 month columns, slightly greater 125 I activity concentrations were found in the 0–20 cm region (above the zone of accumulation) than were found in this region at 6 months. Therefore, some further migration towards the soil surface had occurred.
ch05
May 3, 2007
14:35
202
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Although the general trend of 125 I migration below 20 cm depth was observed at 12 months, the zone of accumulation previously found was not as distinct. This may have been a result of radioactive decay of the 125 I present in the soil, compounded by the fact that during this period, the controlled environment room in which the columns were housed malfunctioned and, as a consequence, poor ryegrass growth over the latter stages of the experiment was seen (discussed below in relation to ryegrass growth and 125 I uptake). This resulted in lower influxes of activity to these columns than would have otherwise been expected, due to a lower evapotranspirative flux. However, it is important to note that overall, the pattern in 125 I soil migration over the course of the experiment is clear and was not thought to have been greatly affected by the malfunction of the controlled environment room. The most important feature in the soil profile trends would seem to be the accumulation of 125 I at around the column mid-point. The extent of migration of 125 I seemed to show a relationship with the presence of anaerobic soil conditions in and above the saturated region at the base of the soil columns (Fig. 5.10). The experiment suggests that 125 I was significantly mobile only within the reduced zone of the soil column, resulting in an 125 I accumulation at the boundary between the anaerobic and aerobic regions. A few studies exist on the effects of redox potential on iodine solubility and mobility. For example, Muramatsu et al. (1996) found that reduced soil conditions brought about the desorption of added 125 I from soil into solution. They suggested that relatively high desorption occurred when the redox potential dropped to around −100 mV. This decrease in iodine sorption at low redox potential is thought to have been primarily due to the chemical reduction of iodine species in the soil. In aerobic soil conditions, the dominant iodine species is thought to be iodate which is relatively strongly adsorbed onto the soil. However, a fall in redox potential leads to the transition of iodate towards iodide; the latter is less strongly adsorbed by comparison. This redox-dependent effect on the sorption behaviour of iodine was clearly seen in the mini-column experiments described above. Therefore, the anaerobic conditions at the base of the experimental soil columns probably led to the formation of iodide which was readily transported upwards
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
203
Reduction-Oxidation potential (mV) -400
-200
0
200
400
600
800
1000
0
10 Soil depth (cm)
May 3, 2007
20
30
40
50
Fig. 5.10. Steady-state soil redox potential profile (mean ± 1 standard deviation) (Phase V).
through the anaerobic zone of soil by the evapotranspirative flux. At the boundary with the aerobic zone, however, the redox gradient in the soil appears to have given rise to redox conditions under which the transition of iodide towards iodate occurred. Thus, 125 I sorption was increased and this led to the observed accumulation on the soil solid phase. In concurrence with this, Thomson et al. (1995) found a large accumulation of stable iodine at the boundary between the aerobic and anaerobic regions of marine sediment. 5.3.2.2.
Soil extractability of
125 I
In order to further understand the soil sorption behaviour of 125 I, extractions of selected samples of contaminated soil from the 12 month columns were carried out with deionised water and 1M sodium hydroxide. These extractants were chosen to remove the soluble and organic-bound soil pools, respectively. The mean percentage of the total 125 I activity that was extracted with water was 6.1% (standard error = 1.2%; n = 26). It may be expected that 125 I extractability with water would be high since the in situ redox potential of the soil within the contaminated zone was low. Solubility
ch05
May 3, 2007
14:35
204
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
values of up to around 30% have been found for 125 I using in situ porous cup samplers at redox potentials of around −200 mV (Muramatsu & Yoshida, 1999). However, in the work described here, no attempt was made to retain low redox conditions during the extraction and so the soil-water mix is likely to have become re-oxidised during shaking. Since manipulation of the oxygen status of soil has shown that 125 I sorption is reversible (Bird & Schwartz, 1996), it is perhaps unsurprising that relatively low percentage extractability values were observed. Sodium hydroxide extracted a mean of 60.1% of the total 125 I (standard error 2.5%; n = 26). Since sodium hydroxide is considered a non-specific extractant of soil humic material, the results suggest that the 125 I resided principally in the humic soil pool, thereby agreeing with the work of Whitehead (1973); Whitehead (1974); Tikhomirov et al. (1980); and Bors et al. (1988). This association may be due to the formation of covalent bonds between the iodine and humic macromolecules (Mercier et al., 2000). However, Sheppard & Thibault (1992) suggested that iodine retention may be primarily through physical association with surfaces and entrapment in micropores and structural cavities of the intricate fabric of organic matter.
5.3.2.3.
Ryegrass biomass and activity concentrations
It is important to note that during the 12 month experiment, ryegrass health/growth was adversely affected by mechanical breakdown of the controlled environment room on two separate occasions. The first of these arose from a failure of the temperature control system. Between days 160 and 163 of the experiment, the temperature within the room reached in excess of 40◦ C. This caused the death of the ryegrass. Therefore, the columns were re-seeded with ryegrass at day 199 and a temperature control switch was installed in order to cut all power to the room should the temperature rise above 25◦ C. At day 304, the temperature did rise above 25◦ C and all power to the room was cut, thereby preventing further temperature increase. However, a major part of the room’s cooling system was
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
205
found to be irreparably damaged and required replacement, meaning that the room would potentially be without power for some time. In an attempt to circumvent this problem, power was returned to the room for around three hours each day. Longer periods of time were considered unsuitable as, in the absence of an effective cooling system, the temperature in the room was found to increase rapidly once power, and in particular lighting, was restored. The three hour period of power supply was found to increase the room temperature from around 21◦ C to 29◦ C. The room was not appropriately fixed until day 354 of the experiment. Over this period, it became clear that the ryegrass was growing very poorly and, after being cut at 5 cm from the soil surface on day 335, failed to re-grow for the remainder of the experiment (overall the experiment lasted 384 days including the initial period where a non-contaminated water table was imposed on the columns). Mean monthly biomass production for the > 5 cm portion of ryegrass from the 12 month columns is shown in Fig. 5.11. Yields of ryegrass from all columns were relatively low over the first 6 months of the experiment. No ‘> 5 cm’ sample was taken at Month 6 because the first of the controlled temperature room problems discussed above meant that no growth occurred following the cut at Month 5. Thus,
3.5 Ryergrass fresh biomass (g)
May 3, 2007
3 2.5 2 1.5 1 0.5 0 1
2
3
4
5
6
7
8
9
10.5
12
Month
Fig. 5.11. Mean monthly ryegrass shoot fresh biomass production in the 12 month columns over the course of the 12 month experiment (Phase V).
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
206
Biosphere Implications of Deep Disposal of Nuclear Waste
I activity concentration (Bq g-1)
the Month 6 cut constituted only the 0–5 cm sample of dead grass from each column. The timing of this cut coincided with the destructive sampling of the 6 month columns. Immediately following the removal of this dead ryegrass, the 9 and 12 month columns were re-seeded with ryegrass. Following an initial low yield of ryegrass at the first cut following re-seeding (Month 7), yields at Months 8 and 9 were generally greater than those previously recorded. The yield at 10.5 months was affected by the second of the controlled environment room malfunctions described above and was low. As with the Month 6 cut, the Month 12 cut constituted only the 0–5 cm grass sample from each of the remaining columns, since no re-growth of the grass occurred following the cut at 10.5 months. Mean 125 I activity concentrations of ryegrass shoot material harvested during successive cuts from each of the columns are shown in Fig. 5.12. In many cases, no activity was detected in the ryegrass shoots over the course of the experiment. This suggests that the ryegrass roots did not penetrate down to the dominant zone of 125 I soil contamination. However, towards the end of the experiment, in particular during the growth of the second ryegrass sward, a small amount of activity was transferred to the above-ground biomass, i.e. in the 9 and 12 month columns. However, to put this amount of transfer into context, the
125
May 3, 2007
1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 1
2
3
4
5
6
7
8
9
10.5
12
Month
Fig. 5.12. Mean ryegrass shoot activity concentrations (dry weight basis) in the 12 month columns over the course of the 12 month experiment (Phase V).
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
207
Root biomass (g dry root kg-1 dry soil) 0
2
4
6
8
10
12
14
0 5 10 Soil depth (cm)
May 3, 2007
15 20 25 30 35 40 45 50
Fig. 5.13. Depth profile of root biomass for the 12 month soil columns (mean ± 1 standard deviation) (Phase V).
total amount of activity present in the ryegrass from any particular column represented a maximum of 0.25% of the total activity present in the soil of that column. The relationship between root biomass and soil depth is shown in Fig. 5.13. This should be considered in light of the knowledge that the roots were removed from the soils using a sieving/washing technique which did not discriminate between roots from the experimental ryegrass and roots present in the soil when it was collected. This ‘background’ root density was determined as 2.23 ± 0.21 g dry root per kg dry soil. Thus, it would appear that the experimental ryegrass roots only contributed significantly to the uppermost 5 cm of soil. This indicates that the living ryegrass roots may not have been exploiting the 125 I contaminated soil to any great extent, since the contaminated soil was generally below 20 cm depth. Nevertheless, it is possible that a small, indeterminable, number of roots may have penetrated to sufficient depth to allow for uptake. Alternatively, the small amount of observed uptake may have been due to root removal from the small reservoirs of 125 I present in the upper soil layers.
ch05
May 3, 2007
14:35
208
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste Table 5.1. Percentage recovery of 125 I activity from the 12 month soil column experiment (Phase V).
Soil Ryegrass Substrate Res. Solutions Total
5.3.2.4.
3 month
6 month
9 month
12 month
21 0 4 57 82
16 < 0.1 23 54 93
24 < 0.1 5 63 92
10 < 0.1 3 69 82
Activity inventory
Percentage recovery of 125 I activity added to the soil columns was determined and is shown in Table 5.1. Average overall recovery ranged from 82 to 93%, suggesting a relatively good recovery with little unaccounted loss. Discounting the residual solutions (those present within the Marriot bottle, reservoir and polythene bead substrate), most of the activity was generally recovered from the soil. The exception was at 6 months, when a relatively large amount of activity was associated with the substrate materials, particularly the silicone sealant, at the base of the columns. In all cases, the substrate materials had appreciable quantities of activity associated with them. Amounts associated with the ryegrass were very low. 5.3.2.5.
Soil transport simulations
The modelling approach described in Chapter 3 was applied to data from the 12 month experiment. This initially involved hydrological (solution influxes and soil moisture contents) modelling (see Chapter 3) before simulation of the transport of 125 I through the soil profile. To do this, data for the three columns at each destructive sampling period (3, 6, 9 and 12 months) were averaged and simulation of each ‘average’ column was then undertaken. Generic parameter values for the transport simulations are given in Chapter 3 and only those parameters specific to the simulation of 125 I are given significant consideration here. Most notably, these are the sorption ∗ ), the dispercoefficient (Kd ), the molecular diffusion parameter (Dm sivity coefficient (dL ), and the decay constant (λ). Due to the very
ch05
May 3, 2007
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
209
low degree of observed plant uptake, the activity uptake by the plant was assumed to not affect the activity concentration of the soil and so the root uptake parameter (α ) was tuned separately in order to reproduce the measured activity in the perennial ryegrass (described below). The decay constant for 125 I was set at 1.348 × 10−7 s−1 . Since no decay correction was applied to the experimental data, no decay correction was employed in the modelling of this experiment. The dispersivity coefficient for the simulations was set to 1.0 cm. In relation to its soil sorption behaviour, 125 I was shown in the experimental results to exhibit a strong redox dependency. Given the changes in redox potential throughout the soil profile, the effects on Kd value had to be reflected in the modelling of 125 I migration in the soil columns, i.e. a constant Kd value could not be used throughout the entire soil profile. In addition, based on the experimental data, sorption of 125 I onto the substrate layer (polythene beads, silicone sealant and nylon mesh used at the base of the columns) clearly took place. Therefore, a sorption coefficient for these interactions was also required, both in terms of its magnitude and its spatial distribution, for the modelling simulations. The molecular diffusion coefficient required the same treatment. The value and spatial distribution of the Kd ∗ parameters required to produce the simulations shown in and Dm Figs. 5.14–5.17 are summarised in Table 5.2. In order to simulate the accumulations of 125 I at around the column mid-point and in the lowermost 2 cm of soil, higher Kd values were required at these points than for the rest of the column. The reason for differing Kd values being required within and between columns was thought to be differences in redox potential throughout the soil profiles. As noted earlier in this chapter, redox potential appeared to have a strong effect on the sorption behaviour of the 125 I. Thus, towards the base of the soil column (32–49 cm depth), lower Kd values (0.01–0.1 cm3 g−1 ) were required to simulate the experimental data due to the predominance of the reduced, and relatively poorly sorbed, iodide species. The higher Kd values required at around 31–32 cm depth in order to simulate the accumulation of
ch05
May 3, 2007
14:35
210
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 5.14. Simulated and observed soil activity concentrations for the 3 month columns in the 12 month soil column experiment (Phase V).
Fig. 5.15. Simulated and observed soil activity concentrations for the 6 month columns in the 12 month soil column experiment (Phase V).
ch05
May 3, 2007
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
211
Fig. 5.16. Simulated and observed soil activity concentrations for the 9 month columns in the 12 month soil column experiment (Phase V).
Fig. 5.17. Simulated and observed soil activity concentrations for the 12 month columns in the 12 month soil column experiment (Phase V).
ch05
May 3, 2007
14:35
212
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Table 5.2. Distribution of Kd and D∗m values in the 12 month soil column simulations (Phase V). Simulation
Parameter Values
3 month
Depth (cm) 0–31 31–32 32–49 49–50 50–55 Kd (cm3 g−1 ) 0.60 0.40 0.05 0.55 0.33 ∗ (cm3 g−1 ) 4.0 × 10−6 4.0 × 10−6 1.0 × 10−6 1.0 × 10−6 1.0 × 10−6 Dm
6 month
Depth (cm) 0–29 29–30 30–49 49–50 50–55 Kd (cm3 g−1 ) 1.50 1.00 0.10 0.30 0.55 ∗ (cm3 g−1 ) 4.0 × 10−6 4.0 × 10−6 1.0 × 10−6 1.0 × 10−6 1.0 × 10−6 Dm
9 month
Depth (cm) 0–30 30–31 31–49 49–50 50–55 Kd (cm3 g−1 ) 1.80 0.90 0.05 0.60 0.55 ∗ (cm3 g−1 ) 4.0 × 10−6 4.0 × 10−6 1.0 × 10−6 1.0 × 10−6 1.0 × 10−6 Dm
12 month
Depth (cm) 0–25 25–26 26–49 49–50 50–55 Kd (cm3 g−1 ) 3.00 2.00 0.01 0.55 0.40 ∗ Dm (cm3 g−1 ) 1.0 × 10−6 1.0 × 10−6 1.0 × 10−6 1.0 × 10−6 1.0 × 10−6
125 I,
were thought to be associated with the shift to oxic soil conditions where the more strongly adsorbed iodate species would be expected to dominate iodine speciation. The reason that the bottommost layer of soil (49–50 cm) also required a relatively high Kd value was thought to arise from dissolved oxygen present in the dosing water, which meant that iodine was initially present as the iodate ion when this water came into contact with the reduced soil at the base of the column. Some distance in, probably a few mm to 1 cm, biological activity probably led to low redox conditions and reduction to iodide. Once in the reduced form, more rapid migration occurred. 5.3.2.6.
Ryegrass uptake simulations
Based on the results of previous soil column experiments, where the density of living ryegrass roots was determined by hand-picking such roots from the soil, an assumption of a uniform and constant root density distribution of 50 cm cm−3 from the surface to a depth of 40 cm for the 125 I columns, was made. However, in order to assess the sensitivity of the results to the root density distribution, a further set of simulations was conducted using a root density profile that decreased linearly from 50 cm cm−3 at the surface to 0 cm cm−3 at 40 cm depth.
ch05
May 3, 2007
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
213
Modelling of two scenarios was carried out, as described in Chapter 3, for the ‘average’ 9 and 12 month columns. For these simulations, calibration of the root uptake parameter (α ) and the shoot/root partitioning ratio were carried out in order to achieve concurrence with the experimental data for the total uptake in the above-ground biomass. The values of these parameters required to effectively simulate the observed data are shown, together with the simulated results for total activity in the perennial ryegrass shoots, in Table 5.3. The simulated results of both root distribution scenarios show reasonably good agreement with the, low, uptake observed. The predicted uptake was not highly sensitive to the root distribution scenario although the linearly decreasing 50 cm cm−3 scenario gave slightly higher values in both the 9 and 12 month columns. 5.3.3.
Six month, fixed and fluctuating water tables, soil column experiment (Phase VII)
5.3.3.1.
Soil column profiles of
125 I
Mean (± 1 s.d.) soil activity concentrations of 125 I in the 6 month experiment are shown in Figs. 5.18–5.20 for the 2 month, 4 month and 6 month columns (all with fluctuating water tables), respectively. Figure 5.21 shows the data for the 6 month columns with fixed water tables. Under conditions of a fluctuating water table, the data show that at 2 months, 125 I had migrated up to 30 cm depth, at which point an abrupt decrease in concentrations was observed. At 4 months, 125 I had accumulated significantly at around the mid-point of the column. Some, limited migration above this accumulation was determined, although, as with the 2 month data, a sharp decrease in activity concentrations was observed. The soil profile of 125 I then appeared to change little between 4 and 6 months. Fixed and fluctuating water table columns gave generally similar soil 125 I profiles after 6 months. However, with a fixed water table, the activity accumulation seemed to occur at slightly greater depth. Reassuringly, there are marked similarities in the trends observed between the 12 month soil column experiment (described above) and the 6 month experiments. As such, similar explanations as to the
ch05
May 3, 2007
Experimental
Root density (cm cm−3 )∗ α (cm2 s−1 ) Shoot/root ratio Shoots (Bq) ∗
Case 1
Case 2
9 month
12 month
9 month
12 month
9 month
12 month
unknown
unknown
0.62
1.90
50 (uniform) 2.6 × 10−14 0.006 0.69
50 (uniform) 6.0 × 10−15 0.09 2.00
50–0 (linear) 4.4 × 10−13 0.006 0.73
50–0 (linear) 2.0 × 10−13 0.09 2.80
See text for explanation of the two root profile distribution scenarios.
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
I in the 9 and 12 month soil columns in the 12 month soil
Biosphere Implications of Deep Disposal of Nuclear Waste
125
14:35
214
Table 5.3. Experimental and simulated ryegrass uptake of column experiment (Phase V).
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
125
0
1
215
I activity concentration (Bq g-1) 2
3
4
5
0
Soil depth (cm)
5 10 15 20 25 30 35 40 45 50
Fig. 5.18. Mean (solid line) and one standard deviation (dashed line) soil profile of 125 I activity at 2 months in the 6 month soil column experiment (fluctuating water table) (Phase VII).
125
0
1
I activity concentration (Bq g-1) 2
3
4
5
0 5 Soil depth (cm)
May 3, 2007
10 15 20 25 30 35 40 45 50
Fig. 5.19. Mean (solid line) and one standard deviation (dashed line) soil profile of 125 I activity at 4 months in the 6 month soil column experiment (fluctuating water table) (Phase VII).
migration behaviour of the 125 I, i.e. in relation to redox potential throughout the soil profile, can be given. The redox potential profiles for the fixed and fluctuating water table columns are shown in Fig. 5.22 and explain why the accumulation of activity occurred at a greater depth in the fixed water table columns. To explain, the
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
216
Biosphere Implications of Deep Disposal of Nuclear Waste
125
0
1
I activity concentration (Bq g-1) 2
3
4
5
0
Soil depth (cm)
5 10 15 20 25 30 35 40 45 50
Fig. 5.20. Mean (solid line) and one standard deviation (dashed line) soil profile of 125 I activity at 6 months in the 6 month soil column experiment (fluctuating water table) (Phase VII).
125
0
1
I activity concentration (Bq g-1) 2
3
4
5
0 5 Soil depth (cm)
May 3, 2007
10 15 20 25 30 35 40 45 50
Fig. 5.21. Mean (solid line) and one standard deviation (dashed line) soil profile of 125 I activity at 6 months in the 6 month soil column experiment (fixed water table) (Phase VII).
increase in water table height in the fluctuating water table columns led to the boundary between anoxic and oxic soil conditions being pushed further up the column (due to saturation of the soil) and this boundary did not move back down the column (due to hysteresis in the soil moisture characteristic curve) in response to the water table
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
217
Redox potential (mV) -400
-200
0
200
400
600
800
0 10 Soil depth (cm)
May 3, 2007
20 30 40 Fixed 50
Fluctuating
Fig. 5.22. Final soil redox potential profiles for the fixed and fluctuating 6 month soil columns (Phase VII).
being lowered over the latter half of the experiment. Thus, over much of the experiment, the redox conditions in the fluctuating water table columns would have favoured the formation of the reduced, more soluble, iodine species (iodide) further up the column than in the fixed water table columns. Thus, a greater extent of migration was observed. The relationship between 125 I migration and redox potential is further explored in Figs. 5.23 and 5.24 where the soil activity concentration is plotted against the measured redox potential of the soil at the point (both spatially and temporally) where the activity concentration was recorded. With a fixed water table (Fig. 5.23), oxic conditions were associated with relatively low soil activity concentrations of 125 I. However, under anoxic conditions, higher activity concentrations in the soil were generally observed. The relationship suggests that the presence of oxic and anoxic conditions gave rise to two ‘clusters’ of points rather than a linear regression. This is further borne out by the data for the fluctuating water table columns (Fig. 5.24) where the same two groupings are even more apparent. However, in this case, a third cluster of points is seen at relatively high activity concentrations under oxic conditions. These relate to instances in the fluctuating water table columns where the 125 I
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
218
Biosphere Implications of Deep Disposal of Nuclear Waste 800
Redox potential (mV)
600 400 200 0 0
0.5
1
1.5
2
3
2.5
-200 -400 Soil 125I activity concentration (Bq g-1)
Fig. 5.23. Relationship between soil activity concentration and redox potential in the 6 month soil column experiment — 6 month, fixed water table columns (Phase VII).
800 600
Redox potential (mV)
May 3, 2007
400 200 0 0
0.5
1
1.5
2
2.5
3
-200 -400 Soil 125I activity concentration (Bq g-1)
Fig. 5.24. Relationship between soil activity concentration and redox potential in the 6 month soil column experiment — 6 month, fluctuating water table columns (Phase VII).
migrated through the redox boundary into the oxic soil towards the surface of the column. This was probably due to the increased flux of water through these columns, associated with increasing the height of the water table. In addition, It is likely that redox changes in the
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
219
20–30 depth layer resulting from the raising and lowering of the water table led to these increased levels of 125 I becoming ‘trapped’ in this region of the soil column. 5.3.3.2.
Soil solution
125 I
The solution activity concentrations of 125 I over the course of the experiment are shown in Figs. 5.25 and 5.26 for the fixed and fluctuating water tables, respectively. Although the input solution had an activity concentration of 20 Bq ml−1 , it was observed that significant adsorption of the radionuclide onto plastic surfaces within the column system occurred during the experiment (discussed below). Consequently, even within the bead substrate, the activity concentration of the solution was significantly below 20 Bq ml−1 . It is most useful, therefore, to think of the activity concentration within the bead substrate as the input activity concentration to the soil. As such, it becomes apparent that, especially at the start of the experiment, significant adsorption of the 125 I onto the soil took place. Manipulation of the water table height led to 125 I being present in soil solution further up the column than for the fixed water table
125
0
1
I activity concentration in soil solution (Bq ml-1) 2
3
4
5
0 0.5 months 10
Soil depth (cm)
May 3, 2007
20
1.5 months 2.5 months 3.5 months
30 40
4.5 months 6 months
50
Fig. 5.25. Mean soil solution activity concentrations throughout the soil profile of the 6 month fixed water table columns over the course of the experiment (Phase VII).
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
220
Biosphere Implications of Deep Disposal of Nuclear Waste
125
0
1
I activity concentration in soil solution (Bq ml-1) 2
3
4
5
0 0.5 months 10
Soil depth (cm)
May 3, 2007
20
1.5 months 2.5 months 3.5 months
30 40
4.5 months 6 months
50
Fig. 5.26. Mean soil solution activity concentrations throughout the soil profile of the 6 month fluctuating water table columns over the course of the experiment (Phase VII).
columns. This would seem consistent with the solid phase soil distribution of 125 I. For both fixed and fluctuating water tables, activity concentration tended to increase at each soil depth over time. However, towards the end of the experiment (particularly between the last two sampling times), it can be seen that activity concentrations decreased at most depths. Since all of these data are decay corrected, this is not a consequence of radioactive 125 I decay. The most likely explanation would seem to be a further, time-dependent, sorption mechanism for the 125 I. Since both solid and solution phase 125 I activity concentrations were determined at the end of the experiment (6 month columns), it was possible to calculate in-situ solid-solution partition coefficients (Kd values) for the lower (contaminated) column depths (32.5 cm and below) at this time. These are given in Table 5.4. In most cases, Kd values were less than 2.5 cm3 g−1 . The Kd values appear to strengthen the explanation for the migration behaviour of the 125 I. For example, towards the base of the column, low Kd values were observed, presumably due to the predominance of the poorly sorbed iodide species. Towards the boundary between the anoxic and oxic soil regions, Kd values tended to increase, presumably as a result
ch05
May 3, 2007
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
221
Table 5.4. In-situ Kd values for 125 I for fixed and fluctuating water tables at 6 months in the Phase VII soil column experiment. Soil Depth
32.5 37.5 42.5 47.5
cm cm cm cm
Fixed (mean)
Fixed (s. dev)
2.17 2.11 1.06 0.65
0.51 2.24 0.84 0.33
Fluctuating Fluctuating (mean) (s. dev) 10.75 1.72 0.84 0.99
13.24 1.69 0.03 0.17
of the transition of the iodide towards iodate These results seem to provide further evidence that redox potential affected significantly the speciation, and hence sorption and migration potential, of 125 I. 5.3.3.3.
Ryegrass biomass and activity concentrations
The > 5 cm portion of ryegrass from each column was harvested on a monthly basis for the duration of the life of each column. Mean monthly ryegrass biomass production in the 6 month columns is shown in Fig. 5.27. In general, monthly biomass production increased over the first 3 months as the grass was becoming established. However, at 4 months and beyond, the production of biomass declined in all columns. There was no apparent reason for this observation, but it should be taken into account when considering the uptake of 125 I by the ryegrass (discussed below). A possible explanation for the loss of ryegrass vitality is the artificial conditions within the controlled environment room, e.g. lighting, humidity and temperature. Alternatively, an exhaustion of the supply of plant nutrients (e.g. nitrate, potassium) from the soil may have been experienced over time. Determination of 125 I activity concentrations of the ryegrass showed that in only one column and in one monthly cut was an activity concentration above background detected. This column gave an activity concentration of 7.5 Bq g−1 (dry weight basis) at month 4 of the experiment. In addition to the data observed for the 12 month experiment, this indicates a very low transfer of 125 I from the soil into the plant. This is not unsurprising when the soil 125 I profiles
ch05
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
222
Biosphere Implications of Deep Disposal of Nuclear Waste 2 Fluctuating
Ryegrass fresh biomass (g)
May 3, 2007
Fixed 1.5
1
0.5
0 1
2
3
4
5
6
Month
Fig. 5.27. Mean ryegrass shoot fresh biomass over the course of the 6 month experiment (Phase VII).
are considered, since the 125 I generally did not migrate into the rooting zone of the plant. Mean ryegrass root biomass for the columns is shown in Table 5.5. In contrast to the 12 month experiment described above, the root densities in this 6 month experiment were determined by hand-picking living ryegrass roots from the soil. The data clearly illustrate that the roots generally penetrated measurably only into the top 10 cm of soil; although in a few cases penetration to 15 cm depth was observed. However, it should also be noted that even this method of root extraction may not determine the presence of a very small number deeper-penetrating roots. The possibility of this cannot be ruled out. Clearly, however, even if this did occur, it did not lead to a high degree of soil-plant transfer of the 125 I. Table 5.5. Mean root biomass (g) in the 6 month columns of the Phase VII soil column experiment. Soil Depth
Fixed Water Table
Fluctuating Water Table
0–5 cm 5–10 cm 10–15 cm 15–50 cm
0.45 0.29 0.13 None detected
0.73 0.09 None detected None detected
ch05
May 3, 2007
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
5.3.3.4.
223
Activity inventory
Recovery of activity from the 6 month column experiment was generally lower than for the previous 12 month experiment, ranging from 62 to 78% (Table 5.6). Mean recovery was 69%. Greatest recovery was from the substrate materials (e.g. polythene beads, nylon mesh, silicone sealant, tubing, connectors), although appreciable quantities were found in the soil and also the residual solutions (those within the Marriot bottle, reservoir and bead substrate). Because activity was recovered from the ryegrass in only one column, on one occasion, ryegrass has not been included in the inventory. 5.3.3.5.
Soil transport simulations
In the simulations for the 12 month soil column experiment, average columns for the various destructive sampling periods were modelled. However, the modelling approach for the 6 month experiment was to simulate individual columns. In contrast to the 12 month columns, many of these columns had fluctuating water tables meaning that the capillary fringe was raised by around 15 cm during the first three months and then remained in this new position during the latter half of the experiment. This hysteretic effect is discussed in relation to the modelling approach in Chapter 3. Associated with this increase in the saturated region of the columns was a rise in the low redox zone, thereby allowing additional migration of 125 I up the soil column in comparison with that observed in the fixed water table columns (Figs. 5.20 and 5.21), due to the strong redox dependency of iodine sorption and hence migration. Therefore, model simulations used to interpret the data divided the soil into two zones: an oxic region, Table 5.6. Percentage recovery of 125 I activity from the 6 month soil column experiment (Phase VII).
Soil Substrate Res. Solutions Total
2 month
4 month
6 month
6 month (fixed)
16 41 21 78
25 32 13 70
19 35 8 62
20 31 16 67
ch05
May 3, 2007
14:35
224
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
which extended from the soil surface down to a depth of between 30 and 44 cm (depending on observed differences between columns), and an anoxic zone, which was from the bottom of the high redox zone to the base of the soil column. Differing sorption coefficients (Kd values) were required for these zones. Furthermore, the model was also used to simulate losses of radioiodine to the polythene beads (given the designation Substrate 1), and moveable reservoir and connecting tubing (given the designation Substrate 2). Additional sorption parameters were required in order to represent the losses that occurred to these components. The ranges of parameter values obtained from manual model calibration are presented in Table 5.7. The decay constant for 125 I was again set at 1.348 × 10−7 s−1 . In this experiment, both the experimental and model data were decay corrected. In the following discussion, the results from a 6 month column with a fluctuating water table are presented as a typical example of the model results as a whole. The parameter values specific to this simulation are also shown in Table 5.7. Figure 5.28 shows simulated 125 I activity concentrations at one month time intervals combined with means and variances of observed soil activities obtained from triplicate samples for various depth intervals down the column at the end of the experiment. It is worth remembering that, in contrast to the 12 month experiment, these data were decay corrected. As shown in Table 5.7, the model simulation includes a sorption parameter specific to the bottom 2 cm of soil. A similar approach was used in the simulation of the 12 month soil columns (described above) and the need for this is probably associated with the presence of the 125 I in an iodate form due the oxic nature of the dosing solution. In the example described here, this sorption was represented by a Kd of 0.3 cm3 g−1 for the bottom 2 cm of the soil column. Beyond this 2 cm, the soil was determined to be very anoxic due to saturation, meaning that the 125 I was most probably reduced to the iodide form and leading to more rapid migration. This is represented in the model simulation by a low Kd value of 0.01 cm3 g−1 . At the capillary fringe, atmospheric oxygen partitioned into the soil water and raised the redox potential. In the case being described here, this occurred at a depth of around 30 cm. This effect
ch05
May 3, 2007 14:35
Kd and dispersion parameter values in the 6 month soil column simulations (Phase VII).
Range for all columns Example simulation
Oxic Depth
Dispersion parameters
Sub2
Sub1
48–50 cm
Anoxic
Oxic
(cm)
Dm (cm2 s−1 )
dL (cm)
1.00–2.50
0.45–1.60
0.01–0.60
0.01
1.0–6.0
0–31 to 0–44
4.0 × 10−6 –2.0 × 10−5
1.0–5.0
2.45
1.60
0.30
0.01
2.00
0–31
2.0 × 10−5
5.0
Radioiodine
Kd (cm3 g−1 )
225
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Table 5.7.
ch05
May 3, 2007
14:35
226
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 5.28. Simulated and observed soil activity concentrations of a 6 month (fluctuating water table) column in the Phase VII experiment.
is reflected in the increase in Kd , to 2.0 cm3 g−1 , at 31 cm (Table 5.7). This change can be clearly observed in the model simulation, where a marked accumulation of 125 I can be seen in both the simulated and observed profiles over the 25–30 cm depth interval. Similarly accurate simulations of the experimental soil data were produced for the other columns. No simulation of plant uptake was undertaken as, apart from one of the columns in month 4 of the experiment, no detectable amounts of radioiodine were measured in any of the ryegrass cuts taken during the 6 month experiment (discussed earlier). 5.4.
General Discussion
In relation to the behaviour of radioiodine in the soil environment, these experiments and associated simulations have allowed the time-dependent soil-plant behaviour of radioiodine migrating from a
ch05
May 3, 2007
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
227
near-surface water table to be elucidated. This has clearly shown the importance of the redox-dependent chemical speciation of radioiodine. Most importantly, this speciation affected the mobility of the radionuclide within the soil columns. In the lower part of the soil column, microbial consumption of dissolved oxygen took place, thereby creating anoxic conditions. A fluctuating water table was able to increase the extent of this low redox zone. Under such conditions, the preferred state for 125 I was, seemingly, the poorly sorbed iodide ion, which was able to migrate relatively freely. However, above the top of the capillary fringe, oxygen was able to diffuse downwards from the soil surface and thereby maintain a high redox state in the upper part of the column. Under these conditions, iodine was apparently oxidised to the iodate ion and, hence, its further migration was severely limited, leading to an accumulation of radioiodine at the boundary between anoxic and oxic soil conditions. Nevertheless, small amounts of radioiodine were observed to migrate above this accumulation, towards the soil surface, suggesting that the radionuclide does have some potential to migrate in the oxic, iodate form. Such trends in the behaviour of 125 I were consistently observed in both soil column experiments. The mini-column experiments provided clear, additional evidence for the effect of high redox potential on the increased soil adsorption of 125 I (and vice versa). Extractions of contaminated soil from the columns suggested that organic matter may have been very important in relation to the sorption of radioiodine. This implies that any zone of radioiodine accumulation may ultimately be only temporary since organic matter degradation over time may alter the ability of the organic material to retain iodine. In turn, this could result in iodine remobilisation and further upward migration. It is useful to compare Kd values determined for 125 I in the work presented here with those previously reported by other workers. The values found in the present work through model calibration are summarised in Table 5.7. Overall, the values obtained compare reasonably well with one another. However, the values from the modelling studies, particularly in the anoxic soil zone, tended to be slightly lower than those found experimentally. In summarising
ch05
May 3, 2007
14:35
228
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
iodine Kd values for sandy soils (as were used here) from a number of sources, both Sheppard and Thibault (1990) and IAEA (1994) reported geometric mean values of 1 cm3 g−1 . More recently, in reviewing biosphere model values for iodine, Sheppard et al. (2002) recommended a geometric mean Kd value for sandy soils of 8 cm3 g−1 , based on a large number of studies. The former value of 1 cm3 g−1 is consistent with the range of values shown in Table 5.7. Although the latter is outside the range in Table 5.7, it is comparable with the results obtained from the mini-column analyses (Fig. 5.4), thus giving a degree of confidence in the reliability of the values found in this work. However, the range of Kd values available in the literature for sandy soils is relatively large (0.23–695 cm3 g−1 , in the review by Sheppard et al., 2002) and offers little assistance in elucidating the importance of redox potential on iodine Kd values. Since the literature values were generally obtained using batch experiments, it is likely that the large degree of variability is a result of differences in basic soil characteristics (e.g. pH, organic matter content) rather than time-dependent variables such as redox potential. Therefore, it is difficult to know which Kd values are appropriate for oxic and, particularly, anoxic environments, based on current data. This highlights a major advantage of the mini-column approach which, for a given soil (e.g. from the site of a proposed nuclear waste repository), allows Kd values to be readily determined for both oxic and anoxic conditions over a longer time scale than is convenient using the batch method. The fact that these Kd values were similar to those obtained in the, albeit slightly, more realistic scenario of the larger soil columns is encouraging. Nevertheless, the fact that lower values were required to successfully simulate the experimental data may indicate that the values obtained experimentally tended to be overestimates. 125 I was found to be poorly transferred from the soil to the crop. In general, it appeared that the ryegrass roots did not penetrate into the dominant zone of 125 I contaminated soil in the lower half of the soil columns, thereby limiting uptake. On the few occasions where uptake was seen, the amounts transferred were very small in relation to the 125 I inventory of the soil. In the 12 month experiment, this
ch05
May 3, 2007
14:35
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioiodine
229
may have been due, at least in part, to the initially low vitality of the ryegrass, and the malfunctioning of the controlled environment room in which the columns were housed. However, the same was observed in the 6 month experiment, where these problems were not present. Thus, it would appear that in these experiments the poor uptake was largely a result of the lack of root penetration into the contaminated soil. The question of whether uptake from iodine contaminated, rootexploited soil would occur cannot be answered unequivocally based on these data. Literature data would suggest that whereas uptake from, presumably oxic, soil can occur, transfer factors tend to be very low. This is likely to be a product of the relatively high Kd values for iodine in oxic soils, and a low translocation of iodine within the plant (Sheppard & Motycka, 1997). Overall, it is clear that two properties, redox dependent Kd and low plant uptake, make the transfer of radioiodine (such as 129 I) from a contaminated water table into ryegrass (at least) very low. Such a result has important implications for safety assessment calculations undertaken for this radionuclide where transport and vegetation bioaccumulation in the near-surface environment are taken into account. It is important to also consider the implications of an accumulation of 129 I at the anoxic/oxic soil boundary. Such an accumulation could potentially become a significant biosphere reservoir. The influence of the height of the water table in determining the closeness of this accumulation to the soil surface is a further important consideration.
ch05
May 3, 2007
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
CHAPTER 6
Technetium
6.1. 6.1.1.
Background 99 Tc
in radioactive waste
Sixteen isotopes of technetium (Tc) are known, with mass numbers ranging from 92 to 107, none of which are stable. Despite many early claims that Tc had been isolated from natural sources (Noddack et al., 1925) it is now accepted that the physical half lives of all the known isotopes are too short to allow any primordial Tc to remain in the present-day Earth’s crust. Of the sixteen Tc isotopes, it is 99 Tc (half life 2.1×105 years) and 99m Tc (half life 6 hours) which have been most studied, the former because of its potential radiological impact following release to the environment, the latter because of its wide variety of uses in medical diagnostics. Due to its relatively long physical half life, the specific activity of 99 Tc is very low (6.37×108 Bq g−1 ) compared with that of 99m Tc (2.0 × 1017 Bq g−1 ). 99 Tc can be produced by direct fission of 235 U in nuclear reactors with a maximum yield of 6.06%, thus giving it a relatively high abundance among fission products (Till, 1986). In addition to direct fission, 99 Tc and 99m Tc can be produced by nuclear interactions with molybdenum, specifically by neutron activation of 98 Mo to 99 Mo, followed by beta decay to 99m Tc and 99 Tc, which eventually decay to 99 Ru. Its long physical half life and inventory in both low and intermediate level radioactive waste makes 99 Tc a key radionuclide in the assessment of the risks associated with radioactive waste disposal. In studying
230
ch06
May 3, 2007
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
231
the behaviour of 99 Tc, the relatively short-lived, gamma ray-emitting isotope 95m Tc has often been used as a surrogate (half life ∼ 60 days). 6.1.2.
Technetium behaviour in soils and plants
The mobility and bioavailability of Tc within soils and sediments is controlled predominantly by the chemical form in which it exists, which in turn is largely controlled by the oxidation-reduction potential of the soil or sediment. Permanently saturated aquatic sediments can experience continuous highly anoxic conditions, in which the existence of Tc(IV) would be expected. However, most fertile soils can experience a range of redox potentials spanning oxic, sub-oxic and anoxic domains, depending on seasonal changes in water content. According to Beasley & Lorz (1986), the conversion of Tc(IV) to Tc(VII) occurs at approximately 150 mV at a range of pH from 5 to 7, which is typical of many agricultural soils. Thus, it is likely that under normal conditions of fluctuating soil redox potentials the inter-conversion of Tc between its two predominant chemical forms will occur. Sorption of Tc to the solid phase within soils and sediments is generally low. Table 6.1 summarises solid-liquid Kd values for Tc in generic soil types reported in two major compilations of data (Sheppard & Thibault, 1990; IAEA, 1994). Both data sources identify clay and organic soils as having the greatest potential for sorption of Tc, though the geometric mean Kd values (1–1.5 cm3 g−1 ) even for these soil types indicate a relatively low sorption tendency in soils. The highest maximum Kd value is 340 cm3 g−1 , reported by Sheppard & Thibault (1990) for an organic soil. The generally low sorption of Tc in soils probably reflects the fact that the pertechnetate ion predominates in most soils. Van Loon (1986) and Echevarria et al. (1998) have determined that the pertechnetate anion is very similar in its behaviour to the nitrate anion in soil-plant systems. Sorption of nitrate is extremely low in agricultural soils, giving rise to problems of water pollution after fertiliser applications to land and, similarly, the mobility of Tc in the pertechnetate form in soils is likely to be high. Wildung et al. (1986) showed that sorption of Tc
ch06
May 3, 2007
14:37
232
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste Table 6.1.
Reported Kd values (cm3 g−1 ) for Tc in generic soil types. Minimum
Geometric Mean
Maximum
Source: Sheppard & Thibault (1990) Sand Loam Clay∗ Organic
0.01 0.01 1.16 0.02
0.10 0.10 1.00 1.00
16 0.4 1.32 340
Source: IAEA (1994) Sand Loam Clay Organic
0.037 0.011 1.100 0.041
0.140 0.100 1.200 1.500
5.0 90.0 1.40 55.0
∗
The geometric mean quoted by Sheppard & Thibault (1990) lies below the range quoted by these authors.
in soils is low under aerobic conditions even over extended periods. These authors examined a large number of readily measurable soil variables and determined that Fe oxides were correlated strongly with early, short-term sorption of Tc, whereas organic carbon appeared to be the most significant soil component associated with Tc sorption over extended periods. Organic matter in soils and sediments has been identified by several authors as being of primary significance in the sorption of Tc. Stalmans et al. (1986) considered that organic matter was most likely to provide a geochemical sink for Tc, though with generally low Kd values being reported by most authors it is debatable whether this is likely to be a sink of much quantitative significance. Rice paddy soils are flooded and dried on a seasonal basis and experience an annual cycle of oxidation and reduction. Tagami & Uchida (1999) used these soils as a model to study the chemical transformations of Tc during flooding and drying. These authors expected that chemical reduction of Tc(VII) to Tc(IV) would decrease Tc extractability from the soils, but they principally set out to determine whether Tc sorbed to soil under reducing conditions would be remobilised following re-aeration of the soils. During 52 days of waterlogging the calcium chloride extractable fraction of Tc reduced
ch06
May 3, 2007
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
233
from 0.85 to 0.25. Even after drying this soil for 43 days the extractable fraction remained almost unchanged at 0.25, reinforcing the conclusion that irreversible reduction of Tc in rice paddy soils can and does occur. This is somewhat contradictory to results obtained by Sheppard & Evenden (1991) who determined that reoxidation of soil resulted in solubilisation of Tc, presumably as a result of the reconversion of Tc(IV) to Tc(VII). To reinforce the importance of soil redox potential in controlling Tc behaviour in soils, Tagami & Uchida (1998) found that amounts of both calcium chloride and acetic acid extractable Tc were dramatically reduced under conditions of waterlogging and that this sorption at low redox potential was also time dependent. Under aerobic soil conditions, the extractable Tc fraction was more or less constant at 0.7 to 0.8, whereas under low redox waterlogged conditions extractablity was reduced rapidly to approximately 0.2 over 50 to 60 days, and thereafter remained more or less constant at this value up to 180 days. The reaction kinetics involved in Tc sorption to soil organic matter have been investigated by Van Loon (1986), who found that long-term sorption is rather slow. He determined that 90% immobilisation of 99 Tc in a contaminated soil may take as long as 30 to 40 years, which compares with an immobilisation time for radiocaesium in most agricultural soils of less than one year. A question of major importance is the role, if any, the soil microbiota play in the reduction of Tc(VII) to Tc (IV). Tagami & Uchida (1996) autoclaved seven selected soils from upland areas or rice paddies in Japan and incubated these, together with unsterilised soils, to determine the sorption of 95m Tc onto each. They determined that adsorption of Tc was higher in non-sterile soils, which they attributed to the onset of reducing conditions in these soils. The conclusion concerning the role of microbial activity in controlling Tc behaviour in soils was that it played an important though indirect role, acting to consume oxygen in waterlogged soil and thereby to reduce the soil’s redox potential. Consequent on this reduction in redox potential is the conversion of Tc(VII) to Tc(IV). Tagami & Uchida (1996) concluded that Tc(IV) could occur in the soil as insoluble TcO2 ,
ch06
May 3, 2007
14:37
234
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
TcO(OH)2 or TcS2 under reducing conditions. The last species is likely to be of particular significance, since microbial reduction of sulphur to sulphide occurs strongly under reducing conditions, leading to the easily detectable formation of H2 S. Pignolet et al. (1989) studied the role of various microorganisms in Tc sorption in sediments and concluded that sulphate-reducing bacteria played a particularly important role (a) through concentration of Tc in bacterial cells but also (b) through the formation of insoluble forms of Tc by reaction with H2 S. The importance of sulphate-reducing bacteria in the behaviour of Tc was also inferred by Abdelouas et al. (2002) who found that their presence led to the reduction of Tc(VII) and the precipitation from solution of TcO2 and/or TcS2 . Bunzl & Schimmack (1988) also investigated the role of microbial activity in the sorption of Tc by a range of soils and peats by sterilising them using gamma irradiation, employing absorbed doses of up to 80 kGy. They found that Kd values for 95m Tc increased for irradiated peat samples, which is contradictory to the results obtained by Tagami & Uchida. Bunzl & Schimmack suggested that irradiation increased the number of sorption sites on the peat on which Tc could be chemically reduced. Any sterilisation technique has the potential to introduce physical, chemical or biological artefacts into experimental soil-plant systems and the results obtained from such studies need to be interpreted with caution. Tagami & Uchida (1996) determined that short-term sorption of Tc to their soils was most strongly correlated with organic matter content, thus confirming the observation of Sheppard et al. (1990) that increased soil organic matter tends to increase Tc sorption. However, Tagami & Uchida also found that Tc sorption was not significantly correlated with the soil’s anion exchange capacity, suggesting that simple anion exchange of pertechnetate is unimportant as a sorption mechanism. This suggests that complexation reactions may govern the adsorption of Tc to organic materials within soils. Tables 6.2 and 6.3 summarise the generally high soil-plant transfer factors (TFs) for Tc in a variety of common crop types. The IUR (1989) database provides a more comprehensive source of transfer factors than the IAEA (1994) publication (the former is the
ch06
May 3, 2007
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
235
Table 6.2. Reported soil-plant transfer factors for Tc in common crop types calculated as the ratio of plant:soil activity concentrations expressed on a dry weight basis.
Cereals Fodder Grass Pea, Bean∗ Turnip Potato∗ Cabbage Lettuce Spinach
Crop Lower 95th Percentile
Expected Value
Upper 95th Percentile
0.07 0.81 10.00 10.00
0.73 8.10 76.00 4.30 79.00 0.24 12.00 200.00 2600.00
3.70 81.00 760.00 43.00
2.40 10.00 10.00 260.00
2.40 120.00 2000.00 7800.00
Source: IAEA, 1994: ∗ values reproduced here as reported in the original document Table 6.3. Reported soil-plant transfer factors for Tc in common crop types calculated as the ratio of plant:soil activity concentrations expressed on a dry weight basis.
Barley Grain Maize Grain Maize Leaf Maize Stem Grass Bean Pod Pea Turnip Flesh Potato Tuber Cabbage Leaf Lettuce Spinach
Crop Minimum
Geometric Mean
Maximum
0.18 0.50 12.50 0.84 7.90 1.12 22.90 13.80 0.01 4.53 108.80 1700.00
0.65 0.93 24.27 2.72 75.52 1.30 26.04 32.98 0.23 12.12 182.70 2617.41
2.37 1.40 52.00 7.60 471.70 1.41 29.60 78.80 0.65 33.00 306.80 3400.00
Source: IUR, 1989
largest collection of soil-plant transfer factors available). It can be seen from the IUR compilation that transfer factors for leafy tissues can be particularly high, with maximum TF values for spinach leaves reported to be up to 7800. This confirms the early observations of Wildung et al. (1977) who surmised that active accumulation
ch06
May 3, 2007
14:37
236
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
of Tc was indicative that the element was acting as a nutrient analogue. Cataldo et al. (1989) found close parallels between the uptake and metabolism of Tc and Echevarria et al. (1998) found that pertechnetate is absorbed proportionally to nitrate by rye grass and Van Loon (1986) determined that nitrogen in the form of nitrate was an effective inhibitor of pertechnetate uptake in spinach plants. In well-aerated soils, in which most crop plants are grown, the pertechnetate anion is likely to be the most important chemical form of Tc, as described in the previous section. Van Loon (1986) found that reduced and complexed Tc appeared not to be available for uptake by plants, but could be made available for uptake by reoxidation to Tc(VII). An interesting hypothesis concerns the ability of plants to make reduced Tc (Tc(IV)) available for uptake by oxidation of their rhizosphere, a phenomenon known to occur in wetland plants in particular. Sheppard & Evenden (1991) experimented with rice and an artificial macrophyte root and determined that reoxidation of Tc by this means was likely to be slow and relatively unimportant. It has been shown that, due to high soil to plant transfer, even over one cropping season, the majority of 99 Tc within a soil will be incorporated by wheat and other crops (Grogan et al., 1987). Echevarria et al. (1997) showed that rye grass shoots accumulated 62 to 78% of pertechnetate from soils and that the fraction of 99 Tc absorbed from soils depended primarily on the biomass of plant material present. No reduction in 99 Tc bioavailability was observed over a period of 6 months, suggesting that, with continued cropping, the principal means of loss of this radionuclide from soils in agricultural ecosystems is likely to be the removal of harvested crop tissues. The presence of 99 Tc in radioactive waste, together with its long physical half life and anionic nature suggest that it may be capable of migrating from the geosphere into the biosphere and hence be of significance in the assessment of radioactive disposal. On this basis it requires investigation, especially since little is known about its upwards migration from groundwater and its migration behaviour in relation to changes in redox potential.
ch06
May 3, 2007
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
6.2.
237
Experimental Overview
Two types of study were carried out to determine the upward migration and plant uptake of Tc. Both used the Silwood soil (characteristics shown in Table 2.1). An outdoor lysimeter experiment (Phase I) was dosed annually in April or May with 99 Tc (amongst other radionuclides) and the upwards migration and uptake into a winter wheat crop determined at the end of each growing season. A soil column experiment (Phase VI) used 95m Tc as a short-lived (physical half life ∼ 60 days) surrogate for 99 Tc. In both cases, the pertechnetate form of Tc was used. Whilst the specific design of the lysimeter experiment is described in Chapter 2, only the general nature of the soil column experiments is given. The more specific details of the column experiment are, therefore, given here. The experiment was 6 months in duration and was carried out using 9 columns of the Silwood soil. Three columns were destructively sampled every 2 months. Perennial ryegrass was grown on all the columns and a fluctuating water table was imposed. As described in Chapter 2, the fluctuating water table was manipulated from an initial setting of 45 cm depth, up to 30 cm depth by 3 months (mid experiment), and back down to 45 cm depth by 6 months. Adjustments of 0.5 cm in the height of the water table were made and the days on which these adjustments were required, in order to follow a sinusoidal curve, were pre-determined. In neither the column nor the lysimeter experiments were the data decay corrected. Whereas for 99 Tc (used in the lysimeter experiment) decay correction is clearly not required due to the long half life, ideally data for a short-lived isotope (e.g. 95m Tc, as used in the column experiment) should be decay corrected. However, as was the case in the 12 month 125 I soil column experiment, decay correction of the data was not appropriate. This was because several stock solutions (each with an initial activity concentration of 20 Bq ml−1 ) were prepared and used (for addition to the Marriott bottles) over the course of the column experiment. At the times when these newly prepared solutions began to be added to the Marriott bottles, the activity flux to the columns would clearly have increased. Decay correcting the data back to, for example, Day 0 of the experiment (when the first
ch06
May 3, 2007
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
238
Biosphere Implications of Deep Disposal of Nuclear Waste
stock solution had an activity concentration of 20 Bq ml−1 ) would therefore give a false impression of data from columns which had received significant contributions of activity from subsequent stock solutions. Essentially, there is no single date to which decay correction can be carried out. Therefore, in the soil column experiment, where 95m Tc was used, the activities are given as those present on the day of sampling. The transport model for the column experiment was, therefore, also set to include radioactive decay. 6.3.
Results
6.3.1.
Four year lysimeter experiment (Phase I)
6.3.1.1.
Soil profiles of
99 Tc
The mean annual soil 99 Tc activity profiles for shallow and deep lysimeters at the end of each growing season are shown in Figs. 6.1 40
1990
35
1991 30
1992 1993
25 20 15 10 5 99Tc
'shallow'
0 1.00E-03
1.00E-02
1.00E-01 1.00E+00 1.00E+01 1.00E+02
Soil Activity Concentration (kBq kg-1) Fig. 6.1. Mean soil profile of 99 Tc activity at each annual sampling of the Phase I shallow (40 cm) lysimeter experiment.
ch06
May 3, 2007
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
239
70 65
1990
60
1991
55 50
1992
45
1993
40 35 30 25 20 15
99Tc
'deep'
10 5 0 1.00E-03
1.00E-02
1.00E-01
1.00E+00 1.00E+01 1.00E+02
Soil Activity Concentration (kBq kg-1) Fig. 6.2. Mean soil profile of 99 Tc activity at each annual sampling of the Phase I deep (70 cm) lysimeter experiment.
and 6.2, respectively. In each year, 99 Tc was present throughout the soil profile indicating that it moved readily up through the lysimeter soil. This is particularly evidenced by considering the data for 1990, when the behaviour of the first ever spike of 99 Tc can be considered without the complication of 99 Tc from previous years. Dosing in this year was carried out in late April and the soil sampling in late July. Thus, over the three months of the initial growing season, the 99 Tc had migrated through the full 70 cm of the deep lysimeter soil. Differences in the activity concentrations of 99 Tc throughout the soil profiles were observed from year to year. However, the general patterns observed were very similar between years. Greatest 99 Tc activity concentrations were observed at the base of the lysimeters (the bottom 5–10 cm). Above this, fairly uniform profiles of activity concentration up to the soil surface were found. This trend suggests an accumulation, via soil adsorption, of 99 Tc at the base of the lysimeters. Since the water table was fixed at 5 cm above the base of the lysimeters, this apparent accumulation would lie within
ch06
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
240
Biosphere Implications of Deep Disposal of Nuclear Waste
the reduced, anoxic soil. Tc is known to exhibit redox-dependent behaviour. For example, at high redox potential (oxic conditions), the poorly sorbed pertechnetate ion (Tc(VII), the form in which the Tc was added in this experiment) tends to persist in soils. However, under anoxic conditions Tc can be chemically reduced to much more strongly sorbed (Tc(IV)) forms (Bachuber et al., 1986; Bird & Schwartz, 1997). Redox conditions within the lysimeters are shown in Fig. 6.3 and illustrate the reduced conditions towards the base of the lysimeters. Such conditions would be likely to lead to the conversion of Tc to more strongly sorbed forms. Nevertheless, this apparent accumulation of reduced 99 Tc clearly did not completely prevent the upward migration of the 99 Tc throughout the soil profile, suggesting that some 99 Tc was mobile through the anoxic soil zone. This mobile fraction was then perhaps converted back to the poorly sorbed, pertechnetate, ion once within oxidised soil, leading to the significant and rapid degree of upwards migration that was observed. Indeed, the behaviour of pertechnetate has been likened to that of nitrate (Echevarria et al., 1998) which is both highly mobile and bioavailable in soils. This also helps to explain the differences in
70 60
Height above geotex (cm)
May 3, 2007
50 40 30 20
Deep Shallow
10 0 -400
-200
0
200
400
600
800
Redox potential (mV)
Fig. 6.3. Steady-state (fixed water table) soil redox potential profiles in the Phase I lysimeter experiment.
ch06
May 3, 2007
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
241
99 Tc
migration between the shallow and deep lysimeters. For example, considering 1990 again, it is apparent that greater activity concentrations above the main zone of accumulation were observed in the shallow lysimeters. Redox potentials for the shallow lysimeters (Fig. 6.3) indicate that the soil at the base of these lysimeters was not as strongly reduced, perhaps limiting the conversion of Tc(VII) to the more strongly sorbed Tc(IV). As the adsorption of 99 Tc generally appeared to be low in the lysimeters, the movement of water through the lysimeter should explain the observed pattern of radioactivity distribution. The water fluxes, rainfall and evapotranspiration measured throughout this lysimeter experiment were presented in Chapter 4 (Figs. 4.3 and 4.4). It is clear from these data that 1990 was the only year with a positive (upward) net water influx to the base of the lysimeters due, primarily, to relatively low rainfall. Thus 99 Tc-containing water was ‘pulled’ into, and up through, the lysimeter soil by the evapotranspiration flux in this year and gave rise to the observed soil contamination throughout the soil profile. Conversely, in subsequent years, the net negative water influxes to the lysimeters generally led to reductions in soil activity concentrations from the 1990 levels. This is likely to have been due to a low degree of upward migration of 99 Tc added in these subsequent years, leaching of residual (1990) activity from the soil, and plant uptake.
6.3.1.2.
Wheat biomass and activity concentrations
The mean biomass production from the shallow and deep lysimeters is shown in Figs. 6.4 and 6.5. The 1992 harvest often gave the greatest biomass yields, whereas the 1993 harvest consistently gave the lowest yields. Overall, the annual biomass yields generally decreased in the order: 1992 > 1991 > 1990 > 1993. In relation to the biomass of the various plant parts, the grain and stems generally gave greater biomass yields than the chaff and leaf portions. In the vast majority of cases for both the deep and shallow lysimeters, the activity concentration of the wheat portions was greatest for the 1990 harvest (Figs. 6.6 and 6.7). Activity concentrations in
ch06
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
242
Biosphere Implications of Deep Disposal of Nuclear Waste 350 1990
Wheat dry biomass (g)
300
1991 250
1992 1993
200 150 100 50 0 Grain
Chaff
Leaf
Stem
Rachis
Fig. 6.4. Mean biomass yield of wheat for each year of the Phase I shallow (40 cm) lysimeter experiment.
600 1990
500 Wheat dry biomass (g)
May 3, 2007
1991 400
1992 1993
300 200 100 0 Grain
Chaff
Leaf
Stem
Rachis
Fig. 6.5. Mean biomass yield of wheat for each year of the Phase I deep (70 cm) lysimeter experiment.
the subsequent years were broadly similar to one another. The reason for the greater uptake during the first year is probably as it was for the 36 Cl in these experiments (described in Chapter 4) and is tied in with the observed pattern of 99 Tc soil migration over time. As described above, in 1990 upwards migration of 99 Tc was observed
ch06
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
243
Wheat activity concentration (Bq g-1)
1000 1990 1991 100
1992 1993
10
1
0.1 Grain
Chaff
Leaf
Stem
Rachis
99
Fig. 6.6. Mean Tc activity concentrations of wheat for each year of the Phase I shallow (40 cm) lysimeter experiment.
100 Wheat activity concentration (Bq g-1)
May 3, 2007
1990 1991 1992 1993
10
1
0.1 Grain
Chaff
Leaf
Stem
Rachis
99
Fig. 6.7. Mean Tc activity concentrations of wheat for each year of the Phase I deep (70 cm) lysimeter experiment.
throughout the soil profile and the greatest soil activity concentrations were found in this year. In addition, the root distributions (Figs. 4.9 and 4.10 for the shallow and deep lysimeters respectively) show that wheat roots were present throughout most of the profile, but particularly towards the soil surface. The relatively high soil activity concentrations within this region in 1990 would therefore
ch06
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
244
Biosphere Implications of Deep Disposal of Nuclear Waste
explain the greater uptake into the wheat at that time. From 1991 to 1993, the soil activity concentrations were depleted, leading to lower uptake by the wheat. The mean transfer factors, weighted according to the distribution of roots and soil activity (see Eqs. (4.1) and (4.2)), are shown in Figs. 6.8 and 6.9 for the shallow and deep lysimeters,
Transfer factor
200 180
1990
160
1991 1992
140
1993
120 100 80 60 40 20 0 Grain
Chaff
Leaf
Stem
Rachis
Fig. 6.8. Soil-plant transfer factors for wheat components throughout the Phase I shallow lysimeter experiment.
1200 1990 1000
1991 1992
Transfer factor
May 3, 2007
800
1993
600 400 200 0 Grain
Chaff
Leaf
Stem
Rachis
Fig. 6.9. Soil-plant transfer factors for wheat components throughout the Phase I deep lysimeter experiment.
ch06
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
245
respectively. Generally, these were less than 200, although two higher values (up to around 1000) were also recorded. 6.3.2.
Six month, fluctuating water table, soil columns (Phase VI)
6.3.2.1.
Soil column profiles of
95m Tc
The mean (+ 1 standard deviation) log 95m Tc soil activity concentrations for the 2, 4 and 6 month columns are shown in Figs. 6.10 to 6.12, respectively. It should be noted that the limit of detection for 95m Tc was around 0.03 Bq g−1 although values below this are shown for interest since they were thought to perhaps hint at some migration in excess of the otherwise apparently very limited migration observed. The 95m Tc was found primarily to be present at the base of the columns. In all of the nine columns, over 80% of the total soil activity was present in the 48–50 cm soil layer and, in all but one column, more than 90% was present in this layer. The dominance of these lower soil layers in terms of the overall budget of soil 95m Tc illustrate that significant retardation of 95m Tc migration through the soil occurred. However, if the standard deviations are 95m
0.001
0.01
Tc activity concentration (Bq g-1) 0.1
1
10
0 5 10 Soil depth (cm)
May 3, 2007
15 20 25 30 35 40 45 50
Fig. 6.10. Mean (solid line) plus one standard deviation (dashed line) soil profile of 95m Tc activity at 2 months in the Phase VI soil column experiment. Vertical dashed line shows the analytical detection limit. Note log scale.
ch06
14:37
246
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste 95m
0.001
Tc activity concentration (Bq g-1)
0.01
0.1
1
10
0
Soil depth (cm)
5 10 15 20 25 30 35 40 45 50
Fig. 6.11. Mean (solid line) plus one standard deviation (dashed line) soil profile of 95m Tc activity at 4 months in the Phase VI soil column experiment. Vertical dashed line shows the analytical detection limit. Note log scale.
95m
0.001
0.01
Tc activity concentration (Bq g-1) 0.1
1
10
0 5 Soil depth (cm)
May 3, 2007
10 15 20 25 30 35 40 45 50
Fig. 6.12. Mean (solid line) plus one standard deviation (dashed line) soil profile of 95m Tc activity at 6 months in the Phase VI soil column experiment. Vertical dashed line shows the analytical detection limit. Note log scale.
taken into account, the data indicate that some migration did appear to occur over time. For example, at 2 months, activity was detected only within the 46–48 cm layer and below. At 4 months, activity was also detected in the 44–46 cm layer. At 6 months, activity was detected as far up as the 35–40 cm layer.
ch06
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
247
5000 4500 Mean solution influx (ml)
May 3, 2007
4000 3500 3000 2500 2000 1500 1000 500 0 2 month
Fig. 6.13. sampling.
4 month
6 month
Mean solution influx to the Phase VI soil columns at each destructive
The soil columns were designed to allow for the free movement of water from the water table to the atmosphere above the column (see Chapter 2). This was clearly achieved, as shown in Fig. 6.13, since it can be seen that the cumulative water influxes to the columns increased throughout the experiment. This suggests that chemical, rather than physical, factors are likely to provide an explanation for the general lack of 95m Tc migration. The presence of the water table led to strongly anoxic conditions in the soil columns and these conditions were ‘pushed’ further up the column as the height of the fluctuating water table was increased (Fig. 6.14). In addition, it is clear that the lowering of the water table over the latter half of the experiment did not lead to a re-oxidation of the soil (also observed in the radioiodine experiments (Chapter 5)). Such conditions are likely to have had a strong effect on the solubility of 95m Tc. For example, the soluble pertechnetate ion (Tc(VII)) (the form in which the Tc was added in this experiment) can be chemically reduced to much less soluble (Tc(IV)) forms. Abdelouas et al. (2002) found that the presence of soil-indigenous sulphate-reducing bacteria led to the reduction of Tc(VII) and the precipitation from solution of TcO2 and/or TcS2 . These workers surmised that co-precipitation with newly formed iron sulphide contributed to Tc removal. Reduction of sulphate occurs at around −200 mV (Sposito, 1989). In the soil columns described here,
ch06
14:37
248
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste Oxidation-Reduction potential (mV) -400
-200
0
200
400
600
800
0 10
Soil depth (cm)
May 3, 2007
45 cm (start) 40 cm
20
35 cm 30 cm (max)
30
35 cm 40 cm
40
45 cm (end) 50
Fig. 6.14. Soil redox potential profile in relation to water table depth in the Phase VI soil column experiment.
redox potential values commonly reached as low as −250 mV at the base of the columns, indicating that such a sulphate-reduction mechanism could have taken place and led to sorption of 95m Tc onto the soil. Furthermore, Sheppard & Sheppard (1986) found that complexation by soil organic matter is a further mechanism by which reduced Tc species may become soil-adsorbed. Nevertheless, it is clear, as was seen for the lysimeter experiments, that a small amount of Tc migration above the main zone of accumulation occurred. Taking into account the redox potential profiles, this migration occurred within the reduced soil zones, suggesting that some migration of reduced Tc species did take place. 6.3.2.2.
Soil
95m Tc
extractability
The sorption behaviour of 95m Tc was further assessed by extracting the contaminated soil (primarily the 48–50 cm sample in each column) with 1M sodium hydroxide and deionised water. Sodium hydroxide was used as a ‘non-specific’ extractant of soil organic matter. Of the total 95m Tc present in the soils (determined by direct gamma counting), between 34 and 70% was removed by sodium hydroxide (mean 55%) suggesting, in concurrence with
ch06
May 3, 2007
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
249
Sheppard et al. (1990), that the organic matter acted as an important sink of 95m Tc in the soil. The solubility of 95m Tc was calculated as the percentage of total 95m Tc (determined by direct gamma counting) that was extracted with deionised water. This ranged from 2 to 15% (mean 8%). These values are somewhat less than those found by Tagami & Uchida (1998) who periodically extracted Tc (using 0.05M calcium chloride) from soil subjected to a 52 day water-logging. They found that solubility fractions decreased from 0.85 to 0.25 over the course of the water-logging. Given the low degree of upward migration of 95m Tc in the soil columns, solubility of up to 15% seems inconsistently high. Therefore, it is considered likely that the column soil may have become re-oxidised during the extraction process, resulting in a reoxidation, and hence release, of adsorbed 95m Tc. The potential for re-oxidation of soil leading to re-mobilisation of adsorbed Tc has important implications for the interpretation of the soil-column experiment. Although not observed in this experiment, a reduction in water table height may, in practice, result in re-oxidation of soil over longer time periods. If Tc were to become re-oxidised, its potential for upward migration, and hence its biospheric significance, could be greater than the current column migration rates imply. Despite this, several workers have suggested that Tc re-oxidation does not occur or, at least, occurs very slowly (e.g. Sheppard & Evenden, 1991; Tagami & Uchida, 1999). However, there is also evidence that re-mobilisation of 99 Tc may take place, e.g. in marine sediments (Leonard et al., 1997; Morris et al., 2000). In order to quantify this potential for reoxidation, anaerobic 95m Tc-contaminated soil samples from the soil columns were allowed to air dry (at 20◦ C by day and 15◦ C by night, 70% relative humidity) over a period of 17 days and the water extractability of the 95m Tc was measured periodically. It was assumed that re-oxidation of the soil occurred as drying proceeded. Results of this experiment are shown in Fig. 6.15 and indicate that solubilisation of 95m Tc did occur as the soil dried and, presumably, became re-oxidised. Over the 17 days, the gravimetric moisture content of the soils decreased from around 45 to 2%, whereas 95m Tc solubility increased from around 4 to 20%.
ch06
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
250
Biosphere Implications of Deep Disposal of Nuclear Waste
Tc extractability in water (% of total)
30 y = 0.9467x + 5.0281 R2 = 0.9224
25 20 15 10 5
95m
May 3, 2007
0 0
2
4
6
8
10
12
14
16
18
Time (days)
Fig. 6.15. Mean (± 1 standard deviation) percentage water extractability of total soil 95m Tc over time. Increasing time also represents a drying, and re-oxidation, of the soil.
In the context of the column study, this suggests that a potential may exist for Tc accumulated over a long period in a reduced zone to be released as a pulse with a transition from reducing to oxidising conditions (for instance under conditions of a falling water table, or following the installation of soil drains). Thus, the Tc could become mobilised and bioavailable over a short period giving rise to increased radiological impacts in consequence. Indeed, if converted to the pertechnetate form, the literature suggests that the mobility and plant uptake of Tc would be substantial. 6.3.2.3.
Ryegrass biomass and activity concentrations
Mean fresh weight biomass production for the 6 month columns is shown in Fig. 6.16. Similar trends were observed for the 2 and 4 month columns over the relevant time periods. Biomass production tended to increase each month over the first 3 months of the experiment before declining in each of the final 3 months. The reason for this decrease is unclear but may have been due to such factors as a time-dependent decrease in soil nutrient availability, or the
ch06
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
251
3.5 Ryegrass fresh biomass (g)
May 3, 2007
3 2.5 2 1.5 1 0.5 0 1
2
3
4
5
6
Month
Fig. 6.16. Mean monthly ryegrass biomass production (fresh weight) for the 6 month soil columns (Phase VI).
movement of anoxic soil conditions up the column over time (i.e. into the zone occupied by some roots). Roots were removed from the soil using the washing and sieving technique. Therefore, both experimental ryegrass roots and roots naturally present in the soil were removed. When the background density of residual roots within the soil was taken into account (determined as 2 g kg dry soil−1 ), it was evident that the experimental ryegrass roots were generally found to be present only in the top 10 cm of soil for all columns (Fig. 6.17). However, this method would not have detected a small number of roots extending to greater depths, and this cannot be discounted. Nevertheless, with the bulk of the root mass in the upper 10 cm, and the bulk of the 95m Tc in the bottom few centimetres of soil, it is unsurprising that 95m Tc concentrations in ryegrass were very low, and often non-detectable, over the course of the experiment (Table 6.4). Whilst it is worth noting that the relatively low ryegrass biomass production observed over the latter half of the experiment may have been influential, at least to some extent, in limiting 95m Tc uptake over this time, the principal cause for low uptake is thought to be the observed lack of overlap between the roots and the contaminated soil.
ch06
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
252
Biosphere Implications of Deep Disposal of Nuclear Waste Dry root biomass (g kg dry soil-1 ) 0
2
4
6
8
10
0 5 10 Soil depth (cm)
May 3, 2007
15 20 25 30 35 40 45 50
Fig. 6.17. Mean (± 1 standard deviation) root distribution profile in the 6 month soil columns (Phase VI). Native root density was determined as around 2 g kg−1 . Table 6.4. Mean (and 1 standard deviation) 95m Tc activity concentrations (Bq g−1 dry material) of the harvested perennial ryegrass shoots over the course of the Phase VI soil column experiment.
Month Month Month Month Month Month
1 2 3 4 5 6
2 Month Columns
4 Month Columns
6 Month Columns
Mean
S. Dev
Mean
S. Dev
Mean
S. Dev
n. det. n. det. — — — —
— — — — — —
n. det. n. det. 0.11 n. det — —
— — 0.19 — — —
n. det n. det 0.35 0.39 0.23 0.11
— — 0.40 0.68 0.40 0.20
N.det.: not detected.
Given the apparent lack of detectable 95m Tc in the rooting zone, it is difficult to accurately determine transfer factors. However, if the soil activity concentrations in the rooting zone are used despite the fact that they were below analytical detection limits, transfer factors can be determined for one column at 4 months and one column at 6 months (only these two columns showed activity in the ryegrass crop at the time of destructive sampling of the column). Despite the fact
ch06
May 3, 2007
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
253
that the rooting zone appeared to be the upper 10 cm of soil, the activity concentration of the uppermost 5 cm of soil was used in the transfer factor calculation, since in one of the two cases no activity was apparent in the 5–10 cm layer. Transfer factors of around 100 and 240 were found using this approach, compared with an expected value for grass of 76 (IAEA, 1994). Clearly however, this approach does not allow for the possibility of a small number of roots penetrating into the more contaminated soil at depth, notwithstanding the prevailing anoxic conditions within this region. It was not possible to account for this given that the presence of such roots was not ascertained. However, in this scenario, lower transfer factors would be expected. 6.3.2.4.
System activity balances
The input of 95m Tc activity to each column was compared with the recovery of activity from the various column components (e.g. soil, ryegrass, residual solutions, polythene beads, nylon mesh, silicone sealant, silicone tubing, nylon tubing connectors and the various plastic surfaces). Activity balances ranged from 67 to 100%, with all but one column in the range 81 to 100%. Average recovery from all columns was 85%. Due to a large amount of sorption onto the substrate components within the column system, particularly the nylon tubing connectors, only around 2–16% of the total activity added to the columns was found in the soil. Nevertheless, these low quantities are not thought to have compromised the experiment, since sufficient quantities (kBq) of activity reached the soil to allow for an interpretation of the behaviour of the 95m Tc. 6.3.2.5.
Soil transport simulations
No modelling of 99 Tc migration in the lysimeter experiment was carried out. Discussion here is therefore limited to the soil column experiment. The general modelling approach described in Chapter 3 was adopted. This initially entailed detailed hydrological modelling of the soil columns, including a simplified representation of hysteretic behaviour due to the fluctuating water table. A review of the detail of the experimental data revealed that each of the 2 month columns
ch06
May 3, 2007
14:37
254
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
and each of the 4 month columns demonstrated similar hydrological behaviour, in terms of water influxes to the columns and final soil moisture contents. However, of the 6 month columns, one column (named as Column A) showed atypically dry conditions in the upper portion of the soil profile and a much lower water inflow than the other two columns (named as Columns B and C) (data not shown). It was thought that this was related to the poorer ryegrass growth that was also observed in Column A. Since the subsequent modelling of the transport behaviour of 95m Tc was driven by the hydrological model (see Chapter 3), it was decided to model the ‘average’ 2 and 4 month columns but to model the 6 month columns individually, so as to be able to take into account their differing hydrological characteristics. The experimental data showed several features pertinent to the modelling of 95m Tc transport in the column system. Due to the effect of redox potential on Tc sorption behaviour, 95m Tc accumulated in the lower soil zones. Therefore, this redox dependency needed to be taken into account, via differing Kd values, in the model simulations. In addition, the various plastic column components were seen to sorb significant quantities of 95m Tc. These interactions also needed to be taken into account in the modelling. Therefore, the polythene bead compartment was given the designation Substrate 1, whereas the external reservoir, associated tubing and tubing connectors were collectively given the designation Substrate 2. Generic model parameter values are described in Chapter 3. Here, the molecular diffusion (Dm ) was set to 4.0 × 10−6 cm2 s−1 , the dispersivity coefficient (dL ) to 1.0 cm, and the decay constant (λ) to 1.32E × 10−7 s−1 . In order to comply with the experimental data, the model data were reported as activities on the day of destructive sampling of the soil columns, i.e. activities in the model simulations were also not decay corrected. Plant uptake of 95m Tc in these experiments was very low and inconsistent (in most cases, no uptake was detected). Therefore, no plant uptake modelling was carried out and, furthermore, no plant uptake coefficient was used in the soil transport simulations. The primary parameter requiring calibration in the model simulations was therefore the Kd value. As described
ch06
May 3, 2007 14:37
Layer (cm) 0–40 40–48 48–50 Sub. 1 Sub. 2
4 month (average)
6 month (Column A)
Kd Layer Kd (cm3 g−1 ) (cm) (cm3 g−1 ) 0.01 1.50 25.0 1.7 5.0
0–40 40–48 48–50 Sub. 1 Sub. 2
0.01 6.00 40.0 1.3 5.3
6 month (Column B)
Tc Phase VI
6 month (Column C)
Layer Kd (cm) (cm3 g−1 )
Layer (cm)
Kd (cm3 g−1 )
Layer (cm)
Kd (cm3 g−1 )
0–30 30–48 48–50 Sub. 1 Sub. 2
0–20 20–48 48–50 Sub. 1 Sub. 2
0.01 1.00 60.0 2.5 6.0
0–30 30–48 48–50 Sub. 1 Sub. 2
0.01 1.30 22.0 0.6 4.4
0.01 2.00 25.0 1.0 4.0
Technetium
2 month (average)
95m
255
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Table 6.5. Magnitude and distribution of sorption parameter (Kd ) values used in the soil column simulations.
ch06
May 3, 2007
14:37
256
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
above, calibration of this parameter for the soil and the substrates was necessary. Given the redox dependency of the sorption of 95m Tc, initial simulations using a simple two-state anaerobic (high Kd )/aerobic (low Kd ) representation were carried out but were unable to reproduce the observed 95m Tc soil distribution. Therefore, the soil column was divided into three zones, an oxic zone (with depths ranging between 0–20 and 0–40 cm), an intermediate zone (ranging from 20– 48 and 40–48 cm) and an anoxic zone (48–50 cm). The existence of an intermediate state is thought to be due to the reduction of Tc by Fe(II) in solution (Cui & Eriksen, 1996). Model calibration was then performed through the adjustment of Kd values for each of these soil zones and for the two substrates. The Kd values required to reproduce the experimental data are shown in Table 6.5. Inventory balances and soil activity profile simulations from the modelling exercise are given in Figs. 6.18 to 6.27.
Fig. 6.18. imulated and observed inventory balance for 2 month columns (Phase VI).
ch06
May 3, 2007
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
257
Fig. 6.19. Simulated and observed final soil activity profile for 2 month columns (Phase VI) (dotted line is analytical detection limit).
Fig. 6.20. Simulated and observed inventory balance for 4 month columns (Phase VI).
ch06
May 3, 2007
14:37
258
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 6.21. Simulated and observed final soil activity profile for 4 month columns (Phase VI) (dotted line is analytical detection limit).
Fig. 6.22. Simulated and observed inventory balance for 6 month Column A (Phase VI).
ch06
May 3, 2007
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
259
Fig. 6.23. Simulated and observed final soil activity profile for 6 month Column A (Phase VI) (dotted line is analytical detection limit).
Fig. 6.24. Simulated and observed inventory balance for 6 month Column B (Phase VI).
ch06
May 3, 2007
14:37
260
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 6.25. Simulated and observed final soil activity profile for 6 month Column B (Phase VI) (dotted line is analytical detection limit).
Fig. 6.26. Simulated and observed inventory balance for 6 month Column C (Phase VI).
ch06
May 3, 2007
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
261
Fig. 6.27. Simulated and observed final soil activity profile for 6 month Column C (Phase VI) (dotted line is analytical detection limit).
All the inventory balances showed that a substantial proportion of the 95m Tc added to the column experiments was retained by Substrate 2, i.e. the reservoir, associated tubing and tubing connectors. The remaining amounts were almost equally distributed between the beads substrate (Substrate 1) and the soil. However, it should also be noted that there is frequently a significant amount of technetium unaccounted for. This may have been due to an underestimation of activity adsorbed on the tubing connectors, which showed the highest percentage of total sorbed solute, and might have also been due to the presence of solute sorbed to the reservoir walls, which was not measured. Therefore, as the transport model is mass conservative, it was decided that all the remaining unrecovered activity present in the system should be assigned to Substrate 2 (which is why the histogram bar exceeds the error bar in some instances). The observed soil activity profiles for the average two and four month column plots (Figs. 6.19 and 6.21 respectively) showed
ch06
May 3, 2007
14:37
262
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
relatively high activity concentrations in the lowest 2 cm soil layer, which then fell rapidly to levels at or below the detection limit. The model simulations reflect this with Kd values of 25 and 40 cm3 g−1 in the 48–50 cm layer (for the 2 and 4 month columns respectively). In order to reproduce the rest of the bottom 10 cm of soil, somewhat lower Kd values of 1.5–6 cm3 g−1 were employed. Similar results were also obtained in the individual six month columns (Figs. 6.23, 6.25 and 6.27). Sorption coefficients for the bottom 2 cm of soil showed a similar range (22–60 cm3 g−1 ). The second layer, of varying thickness (20–48 or 30–48 cm), required Kd values of 1.0–2.0 cm3 g−1 . In general, activity concentrations for the soils above this zone of 95m Tc accumulation were well below the analytical detection limit and so no attempt to model these values was made. However, in 6 month Column C, (which had the highest solution influx) activity concentrations throughout the soil profile showed a degree of continuity and were generally not so far below the detection limit. In this case, the model was able to predict the apparent activity seen throughout the soil profile (Fig. 6.27). It is reassuring that the model parameters were again comparable with those obtained for other columns (e.g. 6 month Column A, and the ‘average’ 2 month column). The low levels of 95m Tc activity transported above the bottom 2 cm of soil also had an impact on the observed plant uptakes. These were very low due to the probable lack of overlap between the contaminated soil and the perennial ryegrass roots. Again, analytically, many of the activity concentrations for the plant material were around the analytical detection limit. Owing to these uncertainties in the root density profile and the amount of technetium uptake, simulations of the plant uptake were not undertaken and consequently no root uptake parameter values are reported. Rather, it was considered that pot experiments may be more beneficial in determining this parameter since the 95m Tc could be mixed homogenously into the soil to ensure that the plant roots were exploiting contaminated soil.
ch06
May 3, 2007
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
6.4.
263
General Discussion
The migration behaviour of technetium appeared to be strongly influenced by soil redox conditions. This was particularly noticeable in the soil column experiment where a zone of strongly reduced soil at the base of the columns (established prior to the introduction of the 95m Tc) effectively limited the upward migration of the radionuclide. In most cases, over 90% of the total soil activity was present in the lowermost layer of soil (48–50 cm depth). This was thought to be due to the chemical reduction of the added pertechnetate, Tc(VII), to Tc(IV) forms. A similar, although less pronounced, accumulation of 99 Tc was observed in the reduced zone of the lysimeters. However, rapid migration of relatively small amounts of 99 Tc, upwards through the lysimeter soil, was also observed. Within around 3 months, added 99 Tc had reached the soil surface, even in the 70 cm lysimeters. Although this was not obvious in the soil columns, the observation of 95m Tc throughout the soil column profile, albeit at or below the analytical detection limit, may suggest that in the soil columns too, some migration through the soil profile took place. Greater confidence can be placed in the low-level activity concentrations for the lysimeter soil, since these samples were analysed using a GeLi detector with a greater efficiency than the NaI detector used in the soil column study (see Chapter 2). An alternative explanation for the apparently greater extent of migration in the lysimeter soil is the presence of wheat cover and field conditions in the lysimeter experiments, which may have resulted in a greater upwards soil migration of the technetium, relative to the soil columns, due to a stronger upwards water flux. The importance of such conditions in facilitating upwards transport was especially noticeable in the first year of the lysimeter experiment, 1990. Model simulations of 95m Tc transport within the soil columns showed good agreement with experimental data. These simulations allowed for further analysis of the data, particularly in relation to the determination of the Kd values required to successfully simulate the experimental data. In order to reproduce the accumulation of 95m Tc seen in the bottom 2 cm of the soil activity profiles, high
ch06
May 3, 2007
14:37
264
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
sorption coefficients (between 22 and 60 cm3 g−1 ) were required, associated with the strongly anoxic conditions in this region. In the oxic zone, Kd values of 0.01 cm3 g−1 were used. However, an intermediate zone (of varying depth depending on column) required Kd values of between 1.0 and 6.0 cm3 g−1 in order to reproduce the observed data. Although the presence of an intermediate redox zone is possible, based on redox chemistry, this was not borne out by measurements of redox potential made in this region of the profile. This may suggest that the effect of the redox chemistry on technetium sorption behaviour is not simply dependent on the actual value (mV reading) of redox potential. For example, pH, in conjunction with redox potential, is known to have significant bearing on the severity of reduction in soils (Rowell, 1994). The model Kd values can be compared to the minimum (0.01 cm3 g−1 ) and maximum (16 cm3 g−1 ) values reported by Sheppard & Thibault (1990) for sandy soils. Clearly, the oxic and ‘intermediate’ values fall within this range giving some confidence in their suitability for such modelling purposes. The relatively high values required for the 48–50 cm layer are likely to be a result of the strongly anoxic nature of the soil in this region, compared to the presumably, oxic, environments generally reported on in the literature. The extent of soil migration in the experiments also defined the degree of plant uptake observed. In the soil columns, activity concentrations of perennial ryegrass were very low and, in many cases, monthly cuts yielded non-detectable 95m Tc levels. This was most likely due the lack of overlap between the ryegrass roots and the contaminated soil. Furthermore, the extremely low amounts of technetium measured in the cut ryegrass meant that it was not possible to assess a root uptake parameter value, using the modelling approach, for this radionuclide under the conditions employed in these experiments. Because the 99 Tc in the lysimeter experiments readily migrated to the soil surface, a relatively high degree of uptake into the wheat crop was found. Transfer factors were, in most cases, less than 200, although a couple of higher values were observed. Although, in general, it was not possible to determine transfer factors for the soil columns due to the absence of 95m Tc uptake by the ryegrass in
ch06
May 3, 2007
14:37
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Technetium
265
many cases, values of 100 and 240 were calculated. Due to the oxic conditions found within the predominant zone of rooting, it can be assumed that the form of technetium taken up by the plants was pertechnetate. The calculated transfer factors are broadly consistent with those previously determined for oxic environments (e.g. IUR, 1989; IAEA, 1994; Echevarria et al., 1997; Denys et al., 2003; Echevarria et al., 2003). Despite a potential for roots to penetrate into the anoxic, more highly contaminated, soil, it seems unlikely that this contributed significantly to the observed plant uptake because the plant availability of reduced Tc species has been shown to be very low (Yanagisawa & Muramatsu, 1995; Tagami & Uchida, 1996). Redox-dependent chemical speciation was thought to be of fundamental importance in determining the potential for upward migration and plant uptake of Tc. As a consequence of this redox-dependent behaviour, reduced Tc forms have been found to be less mobile through soils than the pertechnetate form. Bird & Schwartz (1997) found that, in Canadian Shield lake sediments, Kd values were generally much greater for Tc in anoxic conditions than in oxic conditions. Based on their results, these workers suggested that anoxic conditions at depth should form a barrier to upward migration of Tc. Such a scenario appeared to develop in the soil column experiment described here, as the 95m Tc did not effectively migrate through the low redox soil. In the context of radioactive waste disposal, this has significant implications in relation to the potential for 99 Tc to reach the biosphere. In addition to likely variations in redox potential within the near-field environment, the influence of the water table is likely to lead to wide variations in soil redox potential. In order for 99 Tc from a radioactive waste repository to reach the biosphere, it would need to migrate through both anoxic and oxic soil. These experiments suggest that significant accumulation of 99 Tc may occur where anoxic soil is encountered. However, the importance of considering 99 Tc in the risk assessment of radioactive waste disposal is highlighted by the fact that, (a) small quantities of Tc did migrate within and through the anoxic zone of the soils and (b) reduced Tc (TcIV) species appear to have the potential to become re-oxidised (to TcVII), hence potentially becoming more mobile and bioavailable.
ch06
May 3, 2007
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
CHAPTER 7
Radioselenium
7.1. 7.1.1.
Background 79 Se
in radioactive waste
The physical half life of 79 Se is uncertain. Originally thought to be around 6.5 × 104 years (ICRP, 1983), more recent estimates are around 6.5×105 years (e.g. Chu et al., 1999). This re-evaluation of its half life has led to a reconsideration of its importance in the context of nuclear waste disposal. 79 Se is mostly formed as a product of 235 U fission. However, small quantities can also arise from the activation of trace Se present within reactor materials (Nirex, 2004). 79 Se is therefore a component of spent nuclear fuel, radioactive wastes resulting directly from the reprocessing of spent fuel, and radioactive wastes associated with the operation of nuclear reactors and reprocessing facilities (Argonne National Laboratory, 2004; Nirex, 2004). According to the UK Radioactive Waste Inventory, 79 Se is present primarily in high-level radioactive waste but is also found in intermediate and low-level waste. The total UK activity of this waste is around 350 TBq (Nirex, 2004).
7.1.2.
Se behaviour in soils and plants
As a member of group VI of the Periodic Table (atomic number 34), Se shares an extensive oxy/hydroxy chemistry with sulphur, tellurium and polonium. Within soils, Se can most conveniently be 266
ch07
May 3, 2007
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
267
thought of as resembling sulphur in its chemical behaviour. Like sulphur, Se exists as an anion and in four possible oxidation states, viz: selenate (+6), selenite (+4), elemental Se (0), and selenide (−2) (Fio et al., 1991). The speciation and distribution of Se in soils is dependent on a number of (often interacting) factors, including the pH and chemical and mineralogical composition of the soil, microbial interactions and the nature of adsorbing surfaces (Neal, 1995). For example, the clay sized fraction of soils, hydrous oxides of Fe and Al, and soil organic matter have been shown to be important Se sorbing substrates within soils (Singh et al., 1981; Dhillon & Dhillon, 1999). However, perhaps the most important influence on Se behaviour is the soil redox potential (Eh). Under oxic (high redox) conditions Se(VI) in the form of selenate (SeO2− 4 ) dominates, while at lower redox potentials Se(IV) is 3− found in the form of both selenite (SeO2− 3 ) and biselenite (HSeO ) ions. The reduction of selenate to selenite occurs at a redox potential of 440 mV (at pH 7) and selenite is reduced to elemental selenium (Se0 ) at a redox potential of 270 mV (at pH 7). Elemental selenium probably exists as a precipitated solid. At much lower redox potentials (strongly anoxic conditions), such as may be found in saturated soils, elemental Se can be further reduced to highly insoluble selenide (Se−2 ) species (e.g. H2 S). Therefore, over the range of pH and redox potentials most likely to be encountered in soils a range of inorganic species is possible. Se speciation has important implications with respect to its bioavailability. Generally, the more oxidising the soil conditions, the more soluble and mobile Se becomes. Less is known about the behaviour of organic Se compounds in soils as they are often unidentifiable, forming as products and by-products of microbial processes, or through direct reactions with humic and fulvic acids (Neal, 1995). However, it has been demonstrated that the microbial methylation of inorganic Se can result in the generation of volatile organic species. As a measure of overall sorption of Se to soils it is useful to obtain estimates of the solid-liquid distribution coefficient (Kd ) for different soil types. Kd data for Se have been published by the IAEA (1994) and these are based on estimates by Sheppard & Thibault
ch07
May 3, 2007
15:4
268
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
(1990): sandy soil 1.5 × 102 cm3 g−1 , loam soil 4.9 × 102 cm3 g−1 , clay soil 7.4 × 102 cm3 g−1 , and organic soil 1.8 × 103 cm3 g−1 . However, Thorne et al. (2001) recommended Se Kd values of: brown acidic soil 18 cm3 g−1 , colluvial soil 35 cm3 g−1 , brown rendzina soil 70 cm3 g−1 and organic soil 170 cm3 g−1 . These authors noted that the uncertainty range for each group was one order of magnitude above or below these recommended values. Se concentrations in plants are generally very low, of the order of parts per billion (ppb) but occasionally of the order of parts per million (ppm) in plants such as cereals. Some plant species hyperaccumulate Se, notably Astragalus, which may contain up to 0.5% w/w Se in its tissues. The biochemistry of Se in plants is complex but well known, with different groups of seleno-amino acids being produced by accumulator and non-accumulator plants. Methyl selenide, a volatile organic form of Se, is synthesised by higher plants and several bacteria and fungi can also methylate Se from inorganic forms, resulting in the potential for biogenic volatilisation of Se from soils and plants. Coughtrey et al. (1983) gave concentration ratios (dry mass plant basis) of 5 for selenite-amended soil and 30 for selenate-amended soil. Data from Saas et al. (1982) indicate that higher concentration ratios than this can occur, whereas Johnsson (1991) found concentration ratios of the order of 1 (on a dry mass plant basis). Kloke et al. (1984) provided an order of magnitude range of soil-plant transfer coefficients for Se from 0.1 to 10. Thorne et al. (2001) recommended a value of 1.0 (fresh weight plant/dry weight soil) with an uncertainty range of 0.1 to 30. The increase in the estimate of the physical half life of 79 Se has led to increased interest in its environmental behaviour should it be released from a radioactive waste repository. Being anionic, it is potentially poorly sorbed onto soils. Nevertheless, it apparently displays relatively high Kd values and low soil-plant transfer factors. Understanding of its behaviour is further complicated by its redox dependency. Since its behaviour in the soil-plant system has received only very limited research effort, it was included in this program of work.
ch07
May 3, 2007
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
7.2.
269
Experimental Overview
The study of 79 Se (Phase VIII) was carried out using mini-column and large scale soil column experiments. 75 Se was used as a convenient, relatively short-lived (physical half-life 120 days), gammaemitting surrogate for 79 Se. The general nature of each of these experiments is described in Chapter 2. For each experiment, the data were decay corrected to a fixed date at the start of the experiment. The mini-column experiments were carried out with the Silwood soil (characteristics shown in Table 2.1) and five other soils: two organic soils from the UK (denoted UK#1 and UK#2), and three mineral soils from the Meuse/Haute Marne region of Eastern France (denoted French#1, French#2 and French#3). Characteristics of these five soils are shown in Table 2.7. All soils were studied at two moisture contents, field capacity and saturation. Three of the soils (Silwood, French#2 and French#3) also had fumigation (with methyl bromide) treatments at both moisture contents. Volatilisation measurements were made on the UK#1, UK#2, French#1 and French#2 soils. The larger-scale soil column experiments were undertaken using the Silwood sandy loam soil only. Five columns were run over 3 months (3 with ryegrass cover and 2 with no ryegrass cover (1 of which also had a volatilisation trap fitted over its surface)). Three columns were run over 6 months (all with ryegrass cover) and five columns were run over 9 months (all 5 with ryegrass cover, and 2 of which had volatilisation traps fitted over their surfaces). No MeBr fumigation was carried out in this experiment.
7.3.
Results
7.3.1.
Mini-column experiments
7.3.1.1.
Redox potentials
Time course trends in redox potential for each of the non-fumigated soils at field capacity are shown in Fig. 7.1. Here, all soils remained oxic but large differences between soils were observed. Nevertheless, the redox potential of each soil remained relatively constant
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
270
Biosphere Implications of Deep Disposal of Nuclear Waste Silw ood
900
UK#1
Redox potential (mV)
800
UK#2
700
French#1
600
French#2
500
French#3
400 300 200 100 0 0
10
20
30
40
50
Day
Fig. 7.1. columns.
Mean redox potentials in the non-fumigated field capacity mini-
over time. In contrast, the saturated (non-fumigated) treatment (Fig. 7.2) led to marked reductions in redox potential over time. This was particularly noticeable in the Silwood and French#3 soils where anoxic (negative mV) conditions developed after around 10 days. Despite an initial decline, the UK#1 and UK#2 soils remained very oxic at almost +300 mV for the remainder of the experiment. Since these two soils were the most organic, being a Calluna peat and a Silw ood
500
UK#1 400 Redox potential (mV)
May 3, 2007
UK#2
300
French#1
200
French#2 French#3
100 0 -100
0
10
20
30
40
50
-200 -300 Day
Fig. 7.2.
Mean redox potentials in the non-fumigated saturated mini-columns.
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
271
forest topsoil, respectively, this result suggests that the presence of large quantities of organic matter did not bring about reduction of the soils despite the potential for this organic material to serve as a substrate for increased microbial activity. For the three soils which also had a methyl bromide fumigation treatment, the redox potentials in both moisture contents showed similar trends to the non-fumigated treatment. The field capacity soils stayed oxic over time (Fig. 7.3). However, in all three saturated soils, redox potentials became negative (Fig. 7.4). This happened more quickly in the Silwood and French#2 soils (after around 10 days) than in the French#3 soil (after around 30 days). Clearly, saturation of the soils led to reductions in redox potential. In many cases, anoxic conditions developed. This suggests that as oxygen within the soil was consumed it was not replenished by the diffusion of atmospheric oxygen, due to the waterlogged nature of the soil pores. Fumigation was intended to disrupt and reduce the soil microbial community. However, this did not appear to have a large effect on soil redox potential. Nevertheless, it did seem to bring about anoxic conditions in French#2 soil and delay the onset of anoxic conditions in French#3 soil, compared with the non-fumigated treatments.
800 Silw ood
700 Redox potential (mV)
May 3, 2007
French#2
600
French#3 500 400 300 200 100 0 0
10
20
30
40
50
Day
Fig. 7.3.
Mean redox potentials in the fumigated field capacity mini-columns.
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
272
Biosphere Implications of Deep Disposal of Nuclear Waste 500 Silw ood
Redox potential (mV)
400
French#2
300
French#3 200 100 0 -100
0
10
20
30
40
50
-200 -300 Day
Fig. 7.4.
7.3.1.2.
Mean redox potentials in the fumigated saturated mini-columns.
Kd values
Time-course trends in Kd values for the non-fumigated field capacity and saturated soils are shown in Figs. 7.5 and 7.6, respectively. For the fumigated field capacity and saturated soils, these trends are shown in Figs. 7.7 and 7.8 respectively. Missing values are due to an inability to extract a soil solution from the mini-column, or non-detectable levels of 75 Se being recorded in the soil solution. Silw ood
450
UK#1
400
UK#2
350 Kd (cm3g-1)
May 3, 2007
French#1
300
French#2
250
French#3
200 150 100 50 0 0
10
20
30
40
50
Day
Fig. 7.5.
Mean Kd values in the non-fumigated field capacity mini-columns.
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
273 Silw ood
Kd (cm3g-1)
700
UK#1
600
UK#2
500
French#1 French#2
400
French#3
300 200 100 0 0
10
20
30
40
50
Day
Fig. 7.6.
Mean Kd values in the non-fumigated saturated mini-columns.
450
Kd (cm3g-1)
May 3, 2007
400
Silw ood
350
French#2
300
French#3
250 200 150 100 50 0 0
10
20
30
40
50
Day
Fig. 7.7.
Mean Kd values in the fumigated field capacity mini-columns.
A similar range of Kd values (generally between 50 and 400 cm3 g−1 ) was observed for each soil type and, overall, the differing soil characteristics did not seem to exert a consistent effect on the degree of 75 Se sorption. The observed Kd values for the mineral soils (Silwood, and the three French soils) generally compare well to the IAEA (1994) values which range from 150 to 740 cm3 g−1 for mineral soils. However,
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
274
Biosphere Implications of Deep Disposal of Nuclear Waste 700 Silw ood
600
French#2 500 Kd (cm3g-1)
May 3, 2007
French#3
400 300 200 100 0 0
10
20
30
40
50
Day
Fig. 7.8.
Mean Kd values in the fumigated saturated mini-columns.
the observed values for the UK#1 and UK#2 soils were somewhat below the expected IAEA values for organic soils (1800 cm3 g−1 ). Overall, no strong effects of time or moisture content on 75 Se sorption were noted. Consequently, a single, average Kd value for each soil/treatment over the entire experiment was calculated and these values are shown in Table 7.1. These values show that, in all but one case (the saturated Silwood soil), fumigation led to a reduction in Kd value, when compared to the non-fumigated treatment. It is also apparent from Fig. 7.9, that for the Silwood soil, a clear negative relationship between redox potential and Kd value was observed (Fig. 7.9). However, it should be noted that such a strong relationship was not observed for the other soils. Table 7.1. Average Kd values over the course of the experiment in each of the soils and treatments. Non-Fumigated
Silwood UK#1 UK#2 French#1 French#2 French#3
Fumigated
Field capacity
Saturated
Field capacity
Saturated
188 158 302 147 143 190
265 100 237 172 189 190
96 — — — 118 164
313 — — — 74 114
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
275
1400 1200 1000 Kd (cm3g-1)
May 3, 2007
800 600 400 200 0 -400
-200
0
200
400
600
800
1000
Redox potential (mV)
Fig. 7.9. Relationship between redox potential and Kd value measured in the Silwood soil mini-columns.
7.3.1.3.
Volatilisation
Volatilisation rates of 75 Se from the surfaces of the UK#1, UK#2, French #1, French#2 soil mini-columns are shown in Table 7.2. These indicate that readily measurable volatilisation of 75 Se did take place. No literature values exist against which to compare these rates. In general, these rates correspond to half times (the time for half of the total soil activity to be lost via volatilisation) of 8 to 70 years. No clear effect of moisture content on 75 Se volatility was observed. Similarly, no consistent difference between the organic (UK) soils and the mineral (French) soils was seen. However, in the one soil in which a fumigated treatment was also used, this fumigation led to an increase Table 7.2. Average volatilisation rates (×10−5 Bq day−1 ) for each of the soils and treatments. Non-Fumigated
UK#1 UK#2 French#1 French#2
Fumigated
Field capacity
Saturated
Field capacity
Saturated
2.7 8.0 11.2 5.9
2.7 7.1 15.5 3.8
— — — 15.8
— — — 23.9
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
276
Biosphere Implications of Deep Disposal of Nuclear Waste
in the volatilisation rate. The reason for this was not clear since the volatilisation of 75 Se is thought to be microbially driven (via methylation) and so the assumed reduction of the microbial population should have perhaps led to a reduction in volatilisation. However, if the fumigation treatment merely disrupted the microbial community, then it is possible that a sub-population of microbes, with a greater ability to induce the volatilisation of 75 Se, may have proliferated.
7.3.2.
Soil columns
7.3.2.1.
Soil column profiles of
75 Se
Mean (± 1 s.d.) soil 75 Se activity concentration profiles for the vegetated 3, 6 and 9 month columns without volatilisation traps are shown in Figs. 7.10–7.12. At all three time periods, the 75 Se profile looked very similar, displaying upward migration to the 35–40 cm depth layer. Given that the influxes of 75 Se solution to the columns increased over time, and that the moisture contents in the soils reached a steady state, this lack of continued upward migration over time suggests that the 75 Se was adsorbed onto the soil very effectively.
Soil activity concentration (Bq g-1) 0.01
0.1
1
10
100
0 5 Soil depth (cm)
May 3, 2007
10 15 20 25 30 35 40 45 50
Fig. 7.10. Mean (solid line) ± 1 standard deviation (dashed lines) soil activity concentrations throughout the soil column profile at 3 months. Note log scale. Activity concentrations on the y-axis should be considered non-detectable.
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
277
Soil activity concentration (Bq g-1) 0.01
0.1
1
10
100
0
Soil depth (cm)
5 10 15 20 25 30 35 40 45 50
Fig. 7.11. Mean (solid line) ± 1 standard deviation (dashed lines) soil activity concentrations throughout the soil column profile at 6 months. Note log scale. Activity concentrations on the y-axis should be considered non-detectable.
Soil activity concentration (Bq g-1) 0.01
0.1
1
10
100
0 5 Soil depth (cm)
May 3, 2007
10 15 20 25 30 35 40 45 50
Fig. 7.12. Mean (solid line) ± 1 standard deviation (dashed lines) soil activity concentrations throughout the soil column profile at 9 months. Note log scale. Activity concentrations on the y-axis should be considered non-detectable.
Despite the apparent lack of significant upward migration through the soil, an interesting feature observed in one of the 6 month, and one of the 9 month columns is that a low activity concentration was found for the uppermost soil layer (0–5 cm). However, it should be pointed out that the levels observed were only slightly above analytical detection limits.
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
278
Biosphere Implications of Deep Disposal of Nuclear Waste
The two non-vegetated columns provide an interesting comparison to the vegetated columns in terms of upward 75 Se migration over 3 months (Fig. 7.13). It should be noted that one of these columns had a volatilisation trap fitted and the other did not. The soil activity profiles for these columns showed a very similar trend to that observed in the 3-month vegetated columns (Fig. 7.10), with the majority of activity in the 48–50 cm layer and small amounts migrating higher, especially where no volatilisation trap was present. In both non-vegetated columns however, the activity concentrations observed in all layers were lower than in the corresponding layers of the vegetated 3 month columns. This was thought to be due to the lower water fluxes observed for these columns, meaning that less activity was entering the columns. The absence of vegetation reduced the influx of 75 Se solution into the base of the columns by around 36%. This effect is further borne out by the fact that the soil activity concentrations for the column which had a volatilisation trap were lower than for the column without a trap fitted. The inclusion of a volatilisation trap on a non-vegetated column reduced the water influx by around 69%. In addition, the soil activity profiles for the 9-month columns which were vegetated but had volatilisation traps Soil activity concentration (Bq g-1) 0.001
0.01
0.1
1
10
100
0 5 10
Soil depth (cm)
May 3, 2007
15 20 25 30 35 40 45 50
Fig. 7.13. Soil activity concentrations for non-vegetated soil columns at 3 months. The dashed line represents a column with a volatilisation trap fitted and the solid line a column without a volatilisation trap. Note log scale. Activity concentrations on the y-axis should be considered non-detectable.
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
279
Soil activity concentration (Bq g-1) 0.01
0.1
1
10
0 5
Soil depth (cm)
May 3, 2007
10 15 20 25 30 35 40 45 50
Fig. 7.14. Mean (solid line) ± 1 standard deviation (dashed lines) soil activity concentrations throughout the soil column profile at 9 months with volatilisation traps fitted. Note log scale. Activity concentrations on the y-axis should be considered non-detectable.
fitted (Fig. 7.14), showed a lesser degree of migration and lower activity concentrations than the 9-month columns without volatilisation traps (Fig. 7.12). The reduction in 75 Se solution influx due to the presence of these volatilisation traps was around 79%. 7.3.2.2.
Soil solution
75 Se
Time-course trends in the mean activity concentrations of 75 Se in the extracted soil solutions are shown in Fig. 7.15 for the instrumented 9-month columns. These should be considered in relation to the input 75 Se solution which had an activity concentration of 20 Bq ml−1 . The low values (less than 1 Bq ml−1 ) observed even within the polythene bead substrate indicate that significant sorption of 75 Se occurred onto the various non-biological column components. The activity concentration of the solution reaching the soil above the bead substrate was therefore much reduced relative to the input solution. Generally, the lowermost hollow fibre soil solution sampler (HFSS) inserted into the soil (47.5 cm depth) yielded a soil solution with an activity concentration of around 0.2 Bq ml−1 . The activity concentration decreased further at 42.5 cm depth (to around 0.1 Bq ml−1 ) and again at 37.5 cm depth (to around 0.05 Bq ml−1 ). Occasionally,
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
280
Biosphere Implications of Deep Disposal of Nuclear Waste Soil solution activity concentration (Bq ml-1) 0.00
0.20
0.40
0.60
0.80
0 Month 1 10
Soil depth (cm)
May 3, 2007
Month 2 Month 3
20 30
Month 4 Month 5 Month 6
40
Month 7 Month 8
50
Month 9
Fig. 7.15. Mean soil solution activity concentrations throughout the soil profile over the course of the experiment.
low activity concentrations were observed at less than 37.5 cm depth but these were inconsistent and close to analytical detection limits. Therefore, the activity concentrations within the soil further declined relative to those within the bead substrate indicating that significant adsorption of 75 Se onto the soil solid phase took place. The soil solution data therefore appear to show concurrence with the solid phase activity concentrations and provide further evidence for a high degree of Se sorption and, hence, limited migration. In addition, no strong effects of time on 75 Se activity concentrations throughout the soil profile were noted, again illustrating that the observed upwards migration of 75 Se occurred relatively quickly following the introduction of the radionuclide into the water table. The solid and solution phase data from the 9-month data can be used to estimate solid-solution distribution coefficients (Kd values) within the lower region of the columns. Average values were calculated to be 4 cm3 g−1 at 42.5 cm depth and 13 cm3 g−1 at 47.5 cm depth. Interestingly, these Kd values are much lower than those generally found in the mini-columns. Indeed, the Kd values produced from the soil column experiment seem to be more in line with the recommended values given by Thorne et al. (2001). These authors suggested a Kd value for a brown acidic soil (comparable to the Silwood sandy loam soil used here) of 18 cm3 g−1 .
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
281
Sorption of the 75 Se onto both the column components and the soil was likely to have been caused by its chemical speciation in relation to the prevailing redox conditions. As described earlier, at low redox potential, the selenide form of Se is likely to dominate. Insolubility of this form is potentially caused by precipitation from the solution, which may explain the low activity concentrations observed even within the bead substrate. Within the soil, precipitation, together with soil adsorption phenomena (e.g. complexation by organic matter and hydrous oxides) are likely to have served to further remove selenide from the soil solution. 7.3.2.3.
Ryegrass biomass and activity concentrations
Mean monthly ryegrass biomass production in the 9 month columns is shown in Fig. 7.16. Results are expressed on a fresh weight basis because of concerns over the loss of volatile 75 Se during the drying process. The results allow for a comparison of those columns with volatilisation traps and those without. In the columns without volatilisation traps, monthly biomass production peaked at 2 months and thereafter tailed off to a fairly uniform level with time. However, the columns with traps fitted showed a generally increasing
18 16
Ryegrass fresh biomass (g)
May 3, 2007
With volatilisation trap Without volatilisation trap
14 12 10 8 6 4 2 0 Month 1 Month 2 Month 3 Month 4 Month 5 Month 6 Month 7 Month 8 Month 9
Fig. 7.16. Mean ryegrass fresh biomass in the 9 month columns over the course of the experiment.
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
282
Biosphere Implications of Deep Disposal of Nuclear Waste
trend in biomass up to month 4 before dropping suddenly to a relatively low level from month 5 on. This suggests that the volatilisation trap itself initially aided vigorous plant growth but ultimately hindered it. Visual observation of the plants during the experiment supports this as the plants beneath the traps became unhealthy looking (wilted and brown) over time. These trends should be taken into account when considering the uptake of 75 Se by the ryegrass (discussed below). The relationship between biomass yield (for columns without volatilisation traps only) and solution influx to the columns was investigated as shown in Fig. 7.17. The strong positive relationship does indeed suggest that plant growth drove the flux of solution, and hence radioactivity, into the columns. Ryegrass roots were found to be predominantly present in the top 5 or 10 cm of the soil columns, where a concentration of around 2–3 g dry root kg−1 dry soil was found. However, relatively small quantities of roots were found below this, throughout the soil profile (generally around 0.05–0.5 g dry root kg−1 dry soil). This suggests that the roots were penetrating to depth in search of water and hence were present within the contaminated zone of soil (generally the bottom 10–15 cm). This further indicates a potential for uptake of 75 Se into the ryegrass.
16000
Total column solution influx (ml)
May 3, 2007
R2 = 0.82
14000 12000 10000 8000 6000 4000 2000 0 0
10
20
30
40
50
60
70
80
Total column ryegrass biomass (g)
Fig. 7.17. Relationship between total column biomass and total solution influx for all vegetated columns without volatilisation traps.
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
Ryegrass activity concentration (Bq g-1)
May 3, 2007
0.9
With volatilisation trap
0.8
Without volatilisation trap
283
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Month 1 Month 2 Month 3 Month 4 Month 5 Month 6 Month 7 Month 8 Month 9
Fig. 7.18. Mean ryegrass activity concentrations in the 9 month columns over the course of the experiment.
Mean 75 Se activity concentrations of the above-ground ryegrass harvested monthly from the 9 month columns are shown in Fig. 7.18. Results are again expressed on a fresh weight basis. Comparisons can be made between those columns with volatilisation traps and those without. Generally, in columns without traps, activity concentrations tended to increase up to around 0.6 Bg g−1 at four months before tailing off over the remainder of the experiment. With the traps fitted, 75 Se uptake in ryegrass showed a similar trend but with somewhat lower activity concentrations than in treatments without traps. From month 5 onwards, no 75 Se was detected in the ryegrass samples from the columns with volatilisation traps. In the vast majority of cases, activity concentrations for the above-ground ryegrass were below 1 Bq g−1 . It is interesting to note that activity was present in the ryegrass 1 month after dosing indicating that the ryegrass roots had reached the contaminated soil at the base of the soil columns very quickly; presumably in search of moisture. Linear regression correlations of total ryegrass uptake for a column (Bq) with total water use of a column (ml), and total ryegrass uptake for a column (Bq) with total ryegrass biomass of a column (g) and are shown in Figs. 7.19 and 7.20 respectively.
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste 40 35 30 R2 = 0.87
25
75Se
20 15
Total column
uptake by ryegrass (Bq)
284
10 5 0 0
2000
4000
6000
8000
10000
12000
14000
16000
Total column solution influx(ml)
75Se
uptake by ryegrass (Bq)
Fig. 7.19. Relationship between total solution influx and total ryegrass uptake of 75 Se for all vegetated columns without volatilisation traps.
Total column
May 3, 2007
40 35 30 25
R2 = 0.92
20 15 10 5 0 0
10
20
30
40
50
60
70
80
Total column ryegrass biomass (g)
Fig. 7.20. Relationship between total ryegrass biomass and total ryegrass uptake of 75 Se for all vegetated columns without volatilisation traps.
Determination of soil-shoot transfer was calculated using the weighted soil-plant transfer factor (see Eqs. (4.1) and (4.2)). Since only fresh weight activity concentrations were recorded for the ryegrass, these were adjusted to dry weight basis assuming a dry:wet ratio of 1:10. This ratio was found to be approximately correct by drying of selected ryegrass samples after analysis of 75 Se. For the
ch07
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
285
3 month column, transfer factor values ranged from 141-336, for the 6 month columns from 75 to 167, and for the 9 month columns from 78 to 173. These indicate a relatively high degree of soil-plant transfer. Thorne et al. (2001) report fresh weight biomass : dry weight soil transfer factors for selenium of 0.1–30. These appear consistent with the observed data here since the observed values should be divided by ten to convert to these units, i.e. an overall range of 7.5 to 33.6. 7.3.2.4.
Volatilisation
Volatilisation rates were measured on two 9-month columns with vegetation and one 3 month column without vegetation. Volatilisation of 75 Se was readily observed. By comparing the two sets of data (Fig. 7.21) an understanding of the relative contributions of the plant and the bare soil surface in terms of 75 Se volatilisation can be gained. In months one and two, this comparison shows that around two-thirds of the volatilisation observed in the vegetated columns was due to volatilisation directly from the bare soil surface. In month 3, volatilisation from the bare soil surface seemed to decrease. Whether this decrease would have continued over further time is not clear. However, in the vegetated columns, volatilisation increased markedly at month 3, reaching its highest value of the
12 Volatilisation rate (Bq/m2/day)
May 3, 2007
Vegetated Non-vegetated
10 8 6 4 2 0
Month 1 Month 2 Month 3 Month 4 Month 5 Month 6 Month 7 Month 8 Month 9
Fig. 7.21. columns.
Volatilisation rates for
75
Se from vegetated and non-vegetated soil
ch07
May 3, 2007
15:4
286
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
whole 9 month period. At this time, the data indicate that the bare soil appeared to account for around one-quarter of the volatilisation from the vegetated columns, suggesting an increased significance of the vegetation in determining volatilisation.
7.3.2.5.
Activity inventory
75 Se
activity recovered from the various components of each soil column at the end of the experiment was compared with the input of activity to that column. Recoveries ranged from 60 to 108%, with a mean of 86%. Of the recovered activity, between 5 and 20% was recovered from the soil, 0.01 to 0.1% in the ryegrass and 79 to 95% in the substrate components of the column structure (e.g. polythene beads, tubing and connectors). Recoveries were, on the whole, considered acceptable. The high affinity of 75 Se for the substrate components was not considered to compromise the experiment, since significant quantities of 75 Se reached the soil and clear, consistent, patterns of activity distribution were observed within the soil.
7.3.2.6.
Soil transport simulations
Details of the modelling approach are given in Chapter 3. The transport simulations for 75 Se described here were preceded by detailed hydrological modelling of the soil columns (described in Chapter 3). An important feature of the 75 Se column experiment was the inclusion of two non-vegetated soil columns, which ran for 3 months. These columns provided an opportunity to investigate 75 Se transport in soil without any additional impact arising from the presence of the ryegrass. Having done so, there was then an opportunity to compare the outcome with the results from the 3-month vegetated soil columns. Finally, model simulations of the longer-term (in particular the 9 month) columns provided an opportunity to assess the understanding gained from the 3 month data. As the emphasis here was on developing an improved understanding, 75 Se transport and uptake simulations of selected columns were carried out and are presented here.
ch07
May 3, 2007
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium Table 7.3.
Basic parameter values used in the transport simulations of 75 Se.
Parameter Molecular diffusion Dispersivity coefficient Decay constant Soil sorption coefficient Plant root uptake
7.3.2.6.1.
287
Symbol
Units
D∗m dL λ Kd α
cm2 s−1 cm s−1 cm3 g−1 cm2 s−1
Value 4.0 1.0 0.0 see see
× 10−6 (decay corrected) discussion discussion
Non-vegetated columns
Values of parameters used in the transport simulations of the columns are shown in Table 7.3. Initially, a number of simulations of the nonvegetated 3 month column were carried out. Studies conducted using the mini-columns had demonstrated that, for Silwood soil under saturated (and therefore anoxic) conditions, selenium sorption was high, with calculated Kd values up to around 300 cm3 g−1 . Therefore, an initial transport simulation was conducted using this Kd value. In addition, a substantial amount of sorption within the various components of the experimental system (e.g. external reservoir, tubing and connectors) was observed, these effects also needed to be incorporated into the model simulation. This was achieved through calibration of sorption coefficients for the two “substrates” that comprise the bottom layers of the model domain (described in Chapter 3). Due to the lack of plants in this column, the plant uptake coefficient was not included in the soil transport simulations. The simulation was calibrated in order to ensure that the amount of selenium stored in the soil agreed with the experimental total. As a consequence, in terms of the overall distribution of activity within the model system, it can be observed that the there was some discrepancy between the prediction of stored amounts in the beads, reservoir and tubing, and those actually observed (Fig. 7.22). Figure 7.23 shows the simulated and observed soil activity profile for 75 Se for this simulation. The most obvious aspect of this is that simulated 75 Se activity was solely located in the bottom 2 cm. However, given that a sorption coefficient (Kd ) of 300 cm3 g−1 was employed in the model simulation this is not surprising. Indeed such
ch07
May 3, 2007
15:4
288
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 7.22. Simulated and observed 75 Se inventories for a non-vegetated, 3 month column (without a volatilisation trap fitted).
Fig. 7.23. Simulated and observed soil 75 Se activity profiles for a non-vegetated, 3 month column (without a volatilisation trap fitted).
ch07
May 3, 2007
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
289
a value gives an extremely high (approximately 800) retardation coefficient. By contrast, the experimental data show detectable migration of 75 Se to between 35 and 40 cm depth. In order to try and improve agreement between model and experiment additional simulations were undertaken with lower sorption coefficients. Figure 7.24 shows the results where the effect of sorption has been reduced by an order of magnitude (i.e soil Kd = 30 cm3 g−1 ). In this case, there was now improved agreement between the lowest two points, but still the model was not able to reproduce the next three data points up the soil profile. Further reductions in Kd did not lead to any further improvement. As these could not reproduce the 75 Se activities at around 10−1 kBq l−1 over the 40–46 depth interval, it was therefore concluded that there was an additional mechanism operating, which could enable reasonably rapid transport of small amounts of 75 Se. An alternative explanation, therefore, is that, as some of the 75 Se entered the soil, it was modified into a mobile form, which may have been colloidal in nature (Zawislanski et al., 2003), and that the key
Fig. 7.24. Simulated and observed soil 75 Se activity profiles for a non-vegetated, 3 month column (without a volatilisation trap fitted) — using a reduced Kd .
ch07
May 3, 2007
15:4
290
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
effect of this change in speciation was a marked reduction in Kd . In order to simulate this potential process, a modified model simulation was undertaken in which a fraction of the 75 Se entering the column was assumed to partition into a relatively mobile “colloidal” state and the remainder remained strongly attached to the (immobile) soil matrix. Figure 7.25 shows the result when 95% of the 75 Se was considered to be strongly sorbed (Kd = 45 cm3 g−1 ) and the other 5% was more mobile (with a Kd set at 2 l cm3 g−1 ). Although this represents a further improvement, it is also clear that the increase in selenium activity seen in the 40–42 cm layer, could not be reproduced through this process. It was therefore further postulated that the “colloidal” transport mechanism might have been redox dependent and that once the mobile 75 Se moved into the oxic zone above the capillary fringe then a further change in speciation took place which caused increased sorption. Based on moisture content measurements, the oxic zone of the column being simulated
Fig. 7.25. Simulated and observed soil 75 Se activity profiles for a non-vegetated, 3 month column (without a volatilisation trap fitted) — with additional mobile transport mechanism.
ch07
May 3, 2007
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
291
here was assumed to be to a depth of around 40 cm (this column was not one of those with in situ redox potential probes). Therefore, a further simulation was undertaken, which used the same 95:5 fraction, with a reduced Kd for the mobile fraction (in this case 0.3 cm3 g−1 ) in the anoxic zone but with an increased Kd of 7.0 cm3 g−1 for the oxic and sub-oxic zones, which were assumed to extend from the surface to 42 cm depth (Fig. 7.26). The result showed a marked improvement and appeared to be able to reproduce the general distribution of detected 75 Se over the bottom 10 cm of the column. Although there are clearly some differences between the two sets of data, given the complexity of the problem and the low activity levels for the layers above the bottom 2 cm, this was regarded as a very encouraging result. Above 40 cm depth there was no detectable 75 Se and this was matched by the model simulation, which produced concentrations well below the detection limit. Therefore, having developed a theory for 75 Se transport in soil using data
Fig. 7.26. Simulated and observed soil 75 Se activity profiles for a non-vegetated, 3 month column (without a volatilisation trap fitted) — with additional mobile transport mechanism in anoxic zone.
ch07
May 3, 2007
15:4
292
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
from a non-vegetated soil column, the next stage was to consider its applicability to vegetated columns. 7.3.2.6.2.
Vegetated columns
The model parameters shown in Table 7.3 were also employed in the simulation of a 3 month vegetated column. Despite this column being vegetated, the plant uptake coefficient was not used in the simulation because of the very low plant uptake observed in the columns, i.e. the plant uptake was assumed not to have significantly affected the soil activity concentrations. The Kd values used for the two (mobile/immobile) 75 Se fractions were kept as they were for the previous simulation described above. The only difference in parameterisation from the previous simulation was that the percentage of 75 Se assumed to be in a mobile form was increased from 5 to 15%. Given the small amount of adjustment used in this simulation, Fig. 7.27 shows an extremely encouraging result. In particular it appears to support the assumption that the mobile transport component was
Fig. 7.27. Simulated and observed soil 75 Se activity profiles for a vegetated, 3 month column — with additional mobile transport mechanism in anoxic zone.
ch07
May 3, 2007
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
293
redox dependent, as the simulated results show a marked accumulation of 75 Se in the 40–42 cm layer, which is consistent with the experimental data. Two 9 month vegetated columns were then simulated. Again, the model parameters shown in Table 7.3 were employed, and the plant uptake coefficient was not used. Figures 7.28 and 7.29 show the simulated and observed soil activity profiles for the two vegetated 9 month columns with the greatest variation in total solution influx. In the case of the column with lowest inflow (Fig. 7.28) the mobile fraction was set at 10%, whereas for the highest flux column (Fig. 7.29) it was 15%. In both cases, it was found that the Kd for the mobile anoxic phase needed to be increased from 0.3 cm3 g−1 to 2.0 cm3 g−1 . Although it is not exactly clear why these relatively small adjustments were required, there did seem to be some degree of similarity between the various simulations. In regard to the mobile/immobile fraction, the required adjustments seemed to be related to water
Fig. 7.28. Simulated and observed soil 75 Se activity profiles for a vegetated, low solution influx, 9 month column — with additional mobile transport mechanism in anoxic zone.
ch07
May 3, 2007
15:4
294
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Fig. 7.29. Simulated and observed soil 75 Se activity profiles for a vegetated, high solution influx, 9 month column — with additional mobile transport mechanism in anoxic zone.
influx and the higher the influx rate the greater the size of the mobile fraction. It is possible that the value of the mobile fraction Kd might have been time dependent and this is why the value used in the 9 month columns is larger than that used in the 3 month simulations. In any case, once again the results were very encouraging. A further feature of the data is that both sets show a greater degree of 75 Se migration up the soil column, with detectable amounts measured in the 35–40 cm layer and this result is also borne out by the model simulations.
7.4.
General Discussion
The mini-column experiments provided a very convenient way of determining Kd values for 75 Se under somewhat more realistic conditions than a batch sorption experiment. A comparison of these two approaches is given in Chapter 5 relating to radioiodine. In
ch07
May 3, 2007
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
295
addition, the Kd values could be calculated without the complication of radionuclide transport, as would have to be taken into account in the larger soil columns. The mini-columns gave Kd values which generally ranged from 50 to 400 cm3 g−1 , suggesting a relatively high degree of sorption onto the soil solid phase. Organic matter is often reported to be particularly important in the sorption of 75 Se (Neal, 1995). However, the observed data suggest otherwise since the two, highly organic, UK soils did not show greater Kd values than the other, mineral, soils. These two soils were a calluna peat and a forest top soil and so their constituent organic matter may have been only poorly humified. This may explain their lack of increased sorption potential, i.e. less-well degraded material may have a lower affinity for sorption. The reason for the apparent decrease in 75 Se sorption under fumigated conditions is not clear. Since Se sorption is known to be strongly affected by redox potential, an effect of fumigation on redox potential, and hence 75 Se sorption could be hypothesised. Indeed, it is interesting to note that, for the Silwood soil, a clear negative relationship between redox potential and Kd value was observed (Fig. 7.9). Nevertheless, overall, the effect of fumigation on redox potential was inconsistent, suggesting that this hypothesis does not stand up to scrutiny. It seems probable, therefore, that soil microbes may play an important role in the immobilisation of Se. Thus, in the fumigated soils where soil microbial populations were presumably reduced, lower sorption was observed. This fumigation/microbial effect appeared to be more significant in influencing 75 Se solubility than the other variables in the experiments, e.g. soil texture, pH, organic matter, moisture content, redox potential and time. Overall, the relatively high Kd values observed in the mini-column study suggest that the sorption of 75 Se to soils is high. Its low affinity for the soluble phase of a soil indicates that its migration through soil may be somewhat retarded. Indeed, in the soil column experiments, the 75 Se added to the base of the column was effectively retained on the soil and showed very limited upwards migration. Se adsorption in soils is thought to be driven by its redox-dependent speciation. The order of decreasing solubility is given by Neal (1995)
ch07
May 3, 2007
15:4
296
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
as selenate (SeO4 2− ) > selenite (SeO3 2− ) > elemental Se (Se0 ) > selenide (Se2− ). Therefore under reduced soil conditions, the more strongly sorbed forms of Se would be expected to dominate. As described in previous chapters, the redox potential at the base of the soil columns quickly became negative following the imposition of the water table. It is therefore highly probable that the reduced, highly sorbed, forms of Se were dominant within this soil region. It is perhaps unsurprising therefore that upward migration of the 75 Se was limited. Adsorption onto soil surfaces such as organic matter and clay sized particles (e.g. iron and aluminium hydrous oxides) are deemed probable under such conditions (Neal, 1995). However, because the 75 Se consistently reached the 35–40 cm depth it is clear that some upward migration did take place within the first three months of the experiment (i.e. it was not all adsorbed in the lowermost (48–50 cm) layer). It is thought that this is most likely due to the 75 Se being introduced to the columns after fifteen days when the soil at the base of the columns was not strongly reduced. Therefore, a less readily adsorbed form of Se (e.g. selenite; the form in which the 75 Se was added to the columns) may have predominated at this time, allowing an initial pulse of upward migration. Once the strongly anoxic conditions at the base of the columns had developed it would appear that this migration was halted as the speciation favoured the reduced, readily adsorbed Se forms. Indeed, it was shown in the Silwood soil mini column experiments (Fig. 7.9) that a negative linear correlation existed between redox potential and 75 Se Kd value (i.e. increased sorption at low redox potential). Nevertheless, small amounts of activity did, in some cases, reach the soil surface. Given the apparently high degree of sorption onto the soil, it seems unlikely that this activity reached the surface by an advective-diffusive mechanism, since it would then be expected that activity should also be found in the intermediate soil layers. This result could, therefore, indicate a role of the ryegrass in the transport of the 75 Se from the deep soil layers to the soil surface. In concurrence, neither of the non-vegetated columns had any detectable activity present at the soil surface (0–5 cm layer).
ch07
May 3, 2007
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
297
In spite of a high degree of sorption, the flux of water through the columns appeared to exert some control on the behaviour of 75 Se. In the absence of vegetation (or when a volatilisation trap was fitted to a column), fluxes of water through the columns were reduced. In such columns, lower soil activity concentrations were observed and the extent of migration upwards through the columns was also reduced when compared with columns with higher water fluxes. This clearly illustrates that even for a highly sorbed contaminant, the flux of water through soil is still important in determining its environmental fate. In both the mini-column and larger-scale soil column experiment, volatilisation of 75 Se was observed. This was particularly significant in the larger experiment, since the 75 Se was present only towards the base of the column and so it suggests that the 75 Se was transported upwards through a considerable depth of soil, either by diffusion through soil pores or through the ryegrass roots and shoots. The data from the soil column experiment suggest that between onethird and three-quarters of the total 75 Se volatilised occurred through the plant. However, no direct linear correlation between volatilisation and plant biomass, or water use, for the vegetated columns was found. Indeed, high volatilisation rates were observed at both high and low ryegrass biomass yields. Overall, the percentage of the total soil activity lost via volatilisation from the vegetated columns was just below 4%. The factors controlling the volatilisation from the soil/plant are not clear. It is likely that microbial factors, not quantified in this experiment, may have been significant. Certainly, the literature would seem to support the importance of microbes in the formation of volatile methylated selenium, particularly at low redox potential (Neal, 1995). This is consistent with the presence of the 75 Se, almost exclusively, within the low redox zone at the base of the soil columns (i.e. within the water table). However, the volatilisation of 75 Se from the mini-columns, even those which were not anoxic, suggests that low redox potential is not a pre-requisite for Se volatility. Activity concentrations of the above ground biomass were generally low. This was presumably due to the majority of the soil 75 Se
ch07
May 3, 2007
15:4
298
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
remaining at some depth, i.e. a strong overlap between the roots and the contamination was not observed. However, clearly, some roots did penetrate into the contaminated soil. This is taken into account by calculation of weighted transfer factors which gave relatively high values with an average of 156. Uptake into the ryegrass of a particular column was strongly positively correlated with both total biomass and total water use for the column. Total biomass and total water use were themselves strongly positively correlated with one another. Therefore, it seems likely that uptake of 75 Se by ryegrass in the column experiment was dependent upon water use (evapotranspiration), which was in turn dependent upon the growth of the ryegrass. Again, this illustrates the importance of understanding the water flux when considering the fate of 75 Se. Modelling studies of the soil transport of 75 Se could not reproduce the, albeit limited, migration of 75 Se in the soil profiles when a Kd of 300 cm3 g−1 was used. Much lower, and variable, Kd values were required in order to simulate the experimental data. Variation in Kd was required because the model simulations pointed to the existence of Se fractions of differing mobility. The studies indicated that the mobile fraction is between 5 and 15% of the total amount of 75 Se entering the column Although the exact form could not be determined from the experimental measurements, it is known that organic forms of selenium can occur in soils (Zawislanski et al., 2003). Whatever its form, model investigations indicated that its mobility was restricted to the saturated, low redox zone at the base of the column. In order to successfully simulate the experimental data, this fraction was given relatively low Kd values, ranging from 0.3 to 2.0 cm3 g−1 in the anoxic zone to 7.0 cm3 g−1 in the oxic/sub-oxic zone. The remaining, immobile, fraction required a Kd value of 45 cm3 kg−1 . Radioselenium, therefore, displays complex behaviour in the soilplant system. Clearly, redox potential plays an important role in limiting its migration, however, the formation of a more mobile fraction, as suggested by the modelling, indicates that its migration may be more significant than that of radionuclides such as 99 Tc, the migration of which was shown to be severely limited in the anoxic soil at the base of the columns. Although measured levels of plant uptake
ch07
May 3, 2007
15:4
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radioselenium
299
and volatilisation of 75 Se were low, the fact that both occurred readily, after just 1 month of the soil column experiment, indicates their potential importance in the transfer of activity into the biosphere. Indeed, relatively high weighted transfer factors were observed in the soil column experiment.
ch07
May 3, 2007
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
CHAPTER 8
Radiocations
8.1. 8.1.1.
Background Radiocations in radioactive waste
Radioactive isotopes of cobalt, caesium and sodium from a number of sources are present in radioactive waste. Their primary source is the waste from the operation of nuclear reactors. Cobalt has only one stable isotope (59 Co) and a number of radioactive isotopes, with physical half lives ranging from minutes or days to years. The longest lived is 60 Co which has a physical half life of around 5.3 years. Due to its relatively long half life, 60 Co is the cobalt isotope of primary importance in relation to radioactive waste disposal. The only stable caesium isotope is 133 Cs. More than thirty radioactive caesium isotopes exist, including 134 Cs and 137 Cs. Both of these isotopes are important in relation to radioactive waste disposal due to their production in nuclear reactors as fission products of uranium and plutonium. In addition, they have relatively long physical half-lives of around 2 and 30 years, respectively. Since the Chernobyl reactor accident in 1986, much research effort has focussed on the environmental behaviour of released 134 Cs and, particularly, 137 Cs. 23 Na is the only stable isotope of sodium. Of the 18 radioactive isotopes, 22 Na has the longest physical half life (2.6 years). Although 22 Na is a minor component of radioactive waste, its chemical properties complement and extend those described above and led to its inclusion in the range of radionuclides studied as part of this investigation into the medium to long-term implications of radioactive waste disposal in the biosphere. 300
ch08
May 3, 2007
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
8.1.2.
301
Co, Cs and Na behaviour in soils and plants
Cobalt is a metallic, transition element. It reacts with most inorganic (or organic) acids to form simple Co(II) salts, e.g. CoCl2 , Co(NO3 )2 , and can also form sulphides, carbonates and a vast range of organic complexes. It exhibits six valence states with the Co(II) and Co(III) states being the most common. The background concentration of cobalt in soils is reported as being between 1.6 and 21.5 mg kg−1 (Perez-Espinosa et al., 2005). Cobalt in soil has been shown to have a low extractability with water and weak salt solution compared with its extractability with DTPA (Perez-Espinosa et al., 2005). This suggests that its environmental significance may be low, since it indicates that cobalt has a strong affinity for the solid soil phase, i.e. a high Kd value. Furthermore, Tagami & Uchida (1998) found that the Ca-exchangeable fraction of 60 Co added to two Japanese agricultural soils decreased to almost zero after only a few days following addition. Nevertheless, Albrecht et al., (2003) determined a Kd value of 6 l kg−1 for 58 Co in a neutral, silty loam soil using a batch sorption technique. Despite finding this relatively low Kd , these workers also observed that the migration of 58 Co through soil monoliths was much lower than would be suggested by this Kd . They observed that 84% and 91% of 58 Co added to the surfaces of intact and homogenised soil monoliths (respectively) was retained in the top 4 cm of soil. No 58 Co was detected in the water leaching from the base of the monoliths. Indeed, much greater Kd values have been reported by IAEA (1994) with ‘expected’ Kd values of 60, 1300, 540 and 990 l kg−1 for sand, loam, clay and organic soils, respectively. The importance of clay and organic matter in the sorption of cobalt can be seen by comparing the values for the soils dominated by these components to that of sandy soils. Clearly, adsorption onto cation exchange sites is likely to be an important mechanism limiting the solubility of the positively charged cobalt ion. However, soil adsorption mechanisms of cobalt were studied by Fujikawa & Fukui (1991), and they attributed the tendency for cobalt to become strongly soil-adsorbed to the diffusion and fixation of cobalt into the crystal lattice of iron-bearing minerals. Furthermore, Perez-Espinosa
ch08
May 3, 2007
14:42
302
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
et al. (2005) suggested that complexation by organic matter and precipitation as carbonates and hydroxides of iron were also important mechanisms controlling cobalt sorption behaviour in soil. Smith & Carson (1981) reported that, as for most metals, the adsorption of cobalt increased at high pH. Given an apparent tendency for strong soil adsorption, the transfer of cobalt from soils to plants would be expected to be low. Indeed, the IAEA (1994) gave soil-plant transfer factors of 0.054 for grass and from 0.03 to 0.29 for a range of vegetables. On the basis of the literature Kd values, greater uptake from sandy soils than from other soil types would be expected. Despite a low transfer from the soil, the uptake of cobalt has been observed to increase in the presence of soluble organic matter (Smith & Carson, 1981; Perez-Espinosa et al., 2005). Thus, the exudation of organic acids from plant roots into the rhizosphere may enhance cobalt uptake. Caesium is an alkali metal in Group I of the periodic table and displays similar chemical behaviour to potassium. It exists in solution predominantly as the monovalent cation, Cs+ , due to its weak interaction with ligands (Avery, 1995). Due to the cationic nature of caesium, its sorption onto soil solids is partially controlled by non-specific cation exchange reactions. Therefore, negatively charged surfaces such as clay minerals and organic matter serve as sinks for caesium. However, the most important process dictating caesium sorption in soils is its specific, and almost irreversible attachment, to clay minerals, most notably the 2:1 clay mineral illite (Shaw & Bell, 2001). This attachment can occur onto either the planar or frayed edge sites of the clay structure, or via exchange of the interlayer potassium ions with caesium ions. This latter process leads to the long-term fixation, and hence low mobility and bioavailability, of caesium. The importance of clay minerals in the sorption of caesium is demonstrated by considering its reported Kd values for a range of soil types. Sheppard & Thibault (1990) give ‘expected’ values of 2.7 × 102 , 4.4 × 103 , 1.8 × 103 , and 2.7 × 102 l kg−1 for sand, loam, clay and organic soils, respectively. These values reflect the tendency for clay minerals in clay and loam soils to sorb caesium by the processes described above.
ch08
May 3, 2007
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
303
Clearly, organic matter is less significant but, even so, the Kd values for organic and sandy soils are high in comparison with those of many other radionuclides. As is the case for cobalt, the high Kd values suggest a low potential for the soil-plant transfer of caesium. In concurrence, the IAEA (1994) reported ‘expected’ transfer factors ranging from 5 × 10−3 (for rice) to 5.3 × 10−1 (for grass). Clearly, lower uptake would be expected in clay soils due to stronger sorption on the solid soil phase. In addition, soil-plant transfer is known to depend on factors such as pH and the presence of competing ions (e.g. potassium, ammonium), increasing concentrations of which can reduce the uptake of caesium (White & Broadley, 2000). Sodium is also a Group 1 alkali metal, with its solution chemistry dominated by the monovalent, Na+ , ion. There are very few studies on the soil-plant behaviour of sodium. However, it is thought to be poorly sorbed in soils due to its low potential for interaction with solid phase surfaces. Cation exchange reactions are likely to be the dominant process by which sodium can become sorbed. Therefore, greater sorption in organic and clay soils, than in sandy soils, would be expected. Sheppard & Thibault (1990) reported 22 Na Kd values ranging from approximately 1–10 cm3 g−1 , which seem indicative of its relatively low sorption potential. However, even with an apparent affinity for the solution phase of soils, the IAEA (1994) reported, perhaps surprisingly, a low ‘expected’ sodium transfer factor of 3.0 × 10−1 (crop not specified). As for caesium, the plant uptake of sodium is likely to be reduced by the presence of other monovalent cations, e.g. potassium and ammonium. For the radionuclides discussed in previous chapters, their satisfying of three general criteria led to their inclusion in this programme of work, viz. their presence in radioactive waste, their significant physical half-life, and their potential for a high degree of mobility. Whilst the radiocations clearly satisfy the first of these criteria and, to some degree, the second, their ability to satisfy the last criterion is much less clear, since, compared with the radioanions studied, they are much more likely to become adsorbed onto soil solid phases due to their charge. Nevertheless, on the basis that their upwards migration
ch08
May 3, 2007
14:42
304
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
behaviour had not been hitherto quantified, they were included in this programme of work. 8.2.
Experimental Overview
Experiments with the radiocations 60 Co, 134 Cs, 137 Cs and 22 Na began with two outdoor lysimeter experiments (Phases I and II) followed by a soil column experiment in a controlled environment room (Phase III). Details of the experimental approaches are given in Chapter 2. However, an overview of each experiment is given below. In the first lysimeter experiment (Phase I), 60 C, 134 Cs and 22 Na were added to the base of Silwood soil (characteristics shown in Table 2.1) lysimeters in a cocktail with a number of other radionuclides. Both ‘deep’ (70 cm) and ‘shallow’ (40 cm) lysimeters were used. The experiment was undertaken over four years (1990–1993) with a new radionuclide dosing carried out in April/May of each year. Radionuclide uptake in a winter wheat crop, and distribution within the soil profile, were determined at the end of each growing season. The second lysimeter experiment (Phase II) was similar to the first although only deep (70 cm) lysimeters were used and the lysimeters were cropped with perennial ryegrass. The Silwood soil and Longlands soil (characteristics shown in Table 2.1) were investigated. The lysimeters were dosed with 137 Cs and 22 Na (as well as 36 Cl; described in Chapter 4) at the end of June 1995. In order to determine the fate of the radionuclides, soil sampling was carried out at the end of July, August and September 1995, and mid-April 1996, and ryegrass sampling was carried out at the end of July, August, September and October 1995, and mid-April 1996. The soil column experiment (Phase III) used 134 Cs and 22 Na in undisturbed soil columns of three soil types: Silwood, Wellesbourne and Robertgate (characteristics of latter two shown in Table 2.6). A fixed water table (at 45 cm from the soil surface) was imposed on the columns and sampling of the soil was carried out after 1, 3 and 6 months. Ryegrass was sampled monthly. These samples were analysed for 134 Cs and 22 Na. In all experiments, decay correction of the radiochemical data was carried out to the date of the start of the relevant experiment.
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
8.3.
305
Results
8.3.1.
Four year lysimeter experiment (Phase I)
8.3.1.1.
Soil profiles of radiocations
The shallow and deep lysimeter profiles of 60 Co, 137 Cs and 22 Na activity concentrations over the course of the experiment are shown in Figs. 8.1 to 8.6 respectively. Overall, the greatest activity concentrations for each radionuclide were found within the bottom 5 or 10 cm of the lysimeters, i.e. within the region of the water table. Although migration of each radionuclide to the soil surface was observed, the activity concentrations near the soil surface were generally low. Overall, activity concentrations tended to be lower in the deep lysimeters. Nevertheless, in the case of 60 Co and 137 Cs in the deep lysimeters, distinct accumulations of these radionuclides were seen at the soil surface. Such accumulations were not as recognisable in the shallow lysimeters. Migration of 60 Co, 137 Cs and 22 Na to the soil surface occurred during the first year of the experiment (1990) in both the shallow 40 Height above geotex (cm)
May 3, 2007
35
1990 1991
30
1992
25
1993
20 15 10
Shallow
5 0 1.00E-03 1.00E-02 1.00E-01 1.00E+00 1.00E+01 1.00E+02 -1
Soil activity concentration (kBq kg ) Fig. 8.1. Mean soil 60 Co activity concentrations in the Phase I shallow lysimeters (1990–1993). Note log scale.
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Height above geotex (cm)
306
70 65 60 55 50 45 40 35 30 25 20 15 10 5 0
1990 1991 1992 1993
Deep
1.00E-03
1.00E-02
1.00E-01
1.00E+00
1.00E+01
1.00E+02
-1
Soil activity concentration (kBq kg ) Fig. 8.2. Mean soil 60 Co activity concentrations in the Phase I deep lysimeters (1990–1993). Note log scale.
40
Height above geotex (cm)
May 3, 2007
35
1990
30
1991 1992
25
1993
20 15 10 Shallow
5 0 1.E-03
1.E-02
1.E-01
1.E+00
1.E+01
1.E+02
1.E+03
-1
Soil activity concentration (kBq kg ) Fig. 8.3. Mean soil 137 Cs activity concentrations in the Phase I shallow lysimeters (1990–1993). Note log scale.
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Height above geotex (cm)
Radiocations
70 65 60 55 50 45 40 35 30 25 20 15 10 5 0 1.00E-03
307
1990 1991 1992 1993
Deep 1.00E-02
1.00E-01
1.00E+00 1.00E+01 1.00E+02 -1
Soil activity concentration (kBq kg ) Fig. 8.4. Mean soil 137 Cs activity concentrations in the Phase I deep lysimeters (1990–1993). Note log scale.
40 35 Height above geotex (cm)
May 3, 2007
1990 1991
30 25
1992 1993
20 15 10 5 Shallow 0 1.00E-03
1.00E-02
1.00E-01
1.00E+00
1.00E+01
1.00E+02
-1
Soil activity concentration (kBq kg ) Fig. 8.5. Mean soil 22 Na activity concentrations in the Phase I shallow lysimeters (1990–1993). Note log scale.
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
308
Biosphere Implications of Deep Disposal of Nuclear Waste
Height above geotex (cm)
May 3, 2007
70 65 1990 60 1991 55 1992 50 1993 45 40 35 30 25 20 15 10 Deep 5 0 1.00E-03 1.00E-02 1.00E-01 1.00E+00 1.00E+01 1.00E+02 -1
Soil activity concentration (kBq kg ) Fig. 8.6. Mean soil 22 Na activity concentrations in the Phase I deep lysimeters (1990–1993). Note log scale.
and deep lysimeters. This migration, through even the deep lysimeters, occurred between dosing of the lysimeters in April 1990 and soil sampling in June 1990, indicating rapid migration (i.e. within 3 months). On the basis of activity concentrations, 22 Na appeared to be the most mobile of the radionuclides, reaching much higher activity concentrations, above the bottom few centimetres of soil, than either 60 Co or 137 Cs. Yearly variations were noted for each of the radionuclide activity profiles, particularly 22 Na. However, the profiles showed no discernible time-wise trend, i.e. activity concentrations were not seen to increase consistently or markedly over time. Therefore, it is unclear from the data whether the activity observed in the profiles from 1991 to 1993 was, in fact, residual activity from 1990, or due to further migration of the radionuclides in the subsequent years. For 36 Cl in these experiments (described in Chapter 4), it was considered that the rapid migration observed in 1990 was due to the positive water influx to the base of the lysimeters in this year and that upwards
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
309
migration after that time was likely to have been very limited, due to the negative water influxes to the lysimeters in the years 1991 to 1993, i.e. a net downward flux of moisture due to relatively high rainfall. Although this is likely to be tempered somewhat by the contrasting sorption behaviour of the radiocations, the importance of the water flux in determining solute transport suggests that upwards migration of radiocations during 1991 to 1993 was also likely to be limited. However, no evidence of leaching (marked reductions in soil activity concentrations) was observed over this time either. 8.3.1.2.
Wheat biomass and activity concentrations
Mean biomass production for the various parts of the wheat crop grown on the shallow and deep lysimeters is shown in Figs. 8.7 and 8.8, respectively. The 1992 harvest often gave the greatest biomass yields, whereas the 1993 harvest gave the lowest yields. Overall, the annual biomass yields generally seemed to decrease in the order: 1992 > 1991 > 1990 > 1993. In relation to the biomass of the various plant parts, the grain and stems generally gave greater biomass yields than the chaff and leaf portions. Mean 60 Co, 137 Cs and 22 Na activity concentrations of the wheat portions for each year of the experiment in the shallow and deep 350 1990
300
1991
Wheat biomass (g)
May 3, 2007
250
1992 1993
200 150 100 50 0 Grain
Chaff
Leaf
Stem
Rachis
Fig. 8.7. Mean wheat biomass production from the Phase I shallow lysimeters (1990–1993).
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
310
Biosphere Implications of Deep Disposal of Nuclear Waste 600 1990
Ryegrass biomass (g)
500
1991 1992
400
1993 300 200 100 0 Grain
Chaff
Leaf
Stem
Rachis
Fig. 8.8. Mean wheat biomass production from the Phase I deep lysimeters (1990–1993).
lysimeters are shown in Figs. 8.9 to 8.14, respectively. Overall, these were very low in comparison with the values observed for 36 Cl and 99 Tc in the same experiment (described in Chapters 4 and 6, respectively). Highest values were observed for 137 Cs in 1990, reaching a maximum mean level of 13.8 Bq g−1 . Maximum mean levels for 60 Co and 22 Na were 1.3 Bq g−1 (1990) and 6.3 Bq g−1 (1991), respectively. The years in which these maximum values were obtained suggest 10 Wheat activity concentration (Bq g-1)
May 3, 2007
1 1990 1991
0.1
1992 1993
0.01
0.001 Grain 60
Chaff
Leaf
Stem
Rachis
Fig. 8.9. Mean Co activity concentrations of wheat from the Phase I shallow lysimeters (1990–1993). Note log scale.
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
311
Wheat activity concentration (Bq g-1)
1
0.1
1990 1991 1992 1993
0.01
0.001 Grain
Chaff
Leaf
Stem
Rachis
Fig. 8.10. Mean 60 Co activity concentrations of wheat from the Phase I deep lysimeters (1990–1993). Note log scale.
100 Wheat activity concentration (Bq g-1)
May 3, 2007
10 1990
1
1991 1992
0.1
1993
0.01
0.001 Grain
Chaff
Leaf
Stem
Rachis
Fig. 8.11. Mean 137 Cs activity concentrations of wheat from the Phase I shallow lysimeters (1990–1993). Note log scale.
that the plant biomass (greatest in 1992) was not directly related to plant uptake. Generally, the shallow lysimeters seemed to give greater activity concentrations in the wheat than the deep lysimeters, which is consistent with the greater activity concentrations observed in the shallow lysimeter soils, and the closer proximity of the wheat roots to the dominant zone of soil activity at depth. In terms of the activity
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
312
Biosphere Implications of Deep Disposal of Nuclear Waste
Wheat activity concentration (Bq g-1)
10
1 1990 1991
0.1
1992 1993
0.01
0.001 Grain
Chaff
Leaf
Stem
Rachis
Fig. 8.12. Mean 137 Cs activity concentrations of wheat from the Phase I deep lysimeters (1990–1993). Note log scale.
10 Wheat activity concentration (Bq g-1)
May 3, 2007
1990 1991
1
1992 1993
0.1 Grain
Chaff
Leaf
Stem
Rachis
Fig. 8.13. Mean 22 Na activity concentrations of wheat from the Phase I shallow lysimeters (1990–1993). Note log scale.
concentrations of the various wheat portions, the three radionuclides showed differing behaviour. Generally, 60 Co showed greatest activity concentrations in the grain, 137 Cs in the leaf and 22 Na in the stem. As discussed in previous chapters for the other radionuclides, the uptake into the crop is dependent upon the soil profile of activity and the distribution of roots within the soil. Distribution of the roots in
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
313
Wheat activity concentration (Bq g-1)
10
1
1990 1991 1992 1993
0.1
0.01 Grain
Chaff
Leaf
Stem
Rachis
Fig. 8.14. Mean 22 Na activity concentrations of wheat from the Phase I deep lysimeters (1990–1993). Note log scale.
these lysimeters was given previously (Figs. 4.9 and 4.10). This can be taken into account by calculating weighted transfer factors (see Eqs. (4.1) and (4.2)). For the three radionuclides considered here, mean weighted transfer factors for the shallow and deep lysimeters are shown in Figs. 8.15 to 8.20. As was the case for the activity concentrations of the plant material, transfer factors were low, with the majority of values in the range 0.1 to 10 for all three radionuclides 10
1 Transfer factor
May 3, 2007
1990 1991
0.1
1992 1993
0.01
0.001 Grain
Chaff
Leaf
Fig. 8.15. Mean soil-plant transfer factors for lysimeters (1990–1993). Note log scale.
Stem 60
Rachis
Co in the Phase I shallow
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
314
Biosphere Implications of Deep Disposal of Nuclear Waste
Transfer factor
100
10
1990 1991 1992 1993
1
0.1 Grain
Chaff
Leaf
Fig. 8.16. Mean soil-plant transfer factors for ters (1990–1993). Note log scale.
Stem 60
Rachis
Co in the Phase I deep lysime-
100
10 Transfer factor
May 3, 2007
1990
1
1991 1992
0.1
1993
0.01
0.001 Grain
Chaff
Leaf
Fig. 8.17. Mean soil-plant transfer factors for lysimeters (1990–1993). Note log scale.
Stem 137
Rachis
Cs in the Phase I shallow
(total range from 0 to 70 for all radionuclides). The transfer factors also reveal that soil-plant transfer for the deep lysimeters was greater than for the shallow lysimeters, which is the opposite of the trend for the activity concentration data. This suggests that roots in the deep lysimeters were more efficient in absorbing radionuclides than the roots in the shallow lysimeters.
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
315
Transfer factor
100
10
1990 1991 1992 1993
1
0.1 Grain
Chaff
Leaf
Stem
Rachis
Fig. 8.18. Mean soil-plant transfer factors for 137 Cs in the Phase I deep lysimeters (1990–1993). Note log scale. 10
Transfer factor
May 3, 2007
1
1990 1991 1992 1993
0.1
0.01 Grain
Chaff
Leaf
Fig. 8.19. Mean soil-plant transfer factors for lysimeters (1990–1993). Note log scale.
Stem 22
Rachis
Na in the Phase I shallow
Due to its greater mobility through the soil, and hence presence within the plant rooting zone, transfer factors for 22 Na were consistently greater than for the other two radionuclides considered here. Thus, as would be expected, the most mobile radiocation in the soil also exhibited the greatest degree of bioavailability. For 60 Co and 137 Cs, transfer factors were greatest in 1990, declining in subsequent years. If the root-profile distributions (shown in Figs. 4.9 and 4.10)
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
316
Biosphere Implications of Deep Disposal of Nuclear Waste 100
Transfer factor
May 3, 2007
10
1990 1991 1992 1993
1
0.1 Grain
Chaff
Leaf
Fig. 8.20. Mean soil-plant transfer factors for ters (1990–1993). Note log scale.
Stem 22
Rachis
Na in the Phase I deep lysime-
are considered, it is apparent that, in 1990, the root density in both deep and shallow lysimeters was greater in the 10 cm layer of soil above the geotex than in subsequent years. Thus, in 1990, a greater number of roots were exploiting the dominant region of soil contamination, producing the higher transfer factors. The subsequent decline of root presence in this region is likely to have led to the reduction in transfer factors. However, in addition, the tendency for the 60 Co and 137 Cs transfer factors to decline over the course of the four years of the experiment may also have been due to time-dependent sorption of these radionuclides onto the soil surfaces. Caesium, in particular, is known to undergo strongly time-dependent fixation in soils. Clearly, such a reduction in soluble levels could have led to reductions in transfer from the soil to the plant. 8.3.2.
One year lysimeter experiment (Phase II)
8.3.2.1.
Soil profiles of radiocations
Following dosing in April 1995, soil cores were removed from the Silwood and Longlands soil lysimeters in July, August and September of 1995, and in April 1996. Results of the 137 Cs and 22 Na radiochemical analysis of these soils are shown in Figs. 8.21 to 8.24. At the July 1995
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
317
70 Jul-95 60
Aug-95
Height above geotex (cm)
Sep-95 Apr-96
50
40
30
20
10
0 0.00
0.01
0.10
1.00
10.00
100.00
Soil activity concentration (Bq g-1)
Fig. 8.21. Mean 137 Cs soil activity concentrations throughout the Silwood lysimeter profile (Phase II). Note log scale. 70 Jul-95
60
Aug-95 Sep-95
Height above geotex (cm)
May 3, 2007
50
Apr-96
40
30
20
10
0 0.0
0.1
1.0
10.0
100.0
1000.0
Soil activity concentration (Bq g-1)
Fig. 8.22. Mean 137 Cs soil activity concentrations throughout the Longlands lysimeter profile (Phase II). Note log scale.
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
318
Biosphere Implications of Deep Disposal of Nuclear Waste 70
60
Height above geotex (cm)
Jul-95 Aug-95
50
Sep-95 Apr-96
40
30
20
10
0 0.00
0.01
0.10
1.00
10.00
-1
Soil activity concentration (Bq g ) 22
Fig. 8.23. Mean Na soil activity concentrations throughout the Silwood lysimeter profile (Phase II). Note log scale. 70 Jul-95 60
Aug-95 Sep-95
Height above geotex (cm)
May 3, 2007
Apr-96
50
40
30
20
10
0 0.00
0.01
0.10
1.00
10.00
100.00
-1
Soil activity concentration (Bq g ) 22
Fig. 8.24. Mean Na soil activity concentrations throughout the Longlands lysimeter profile (Phase II). Note log scale.
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
319
sampling, the profiles show that the bulk of activity was retained at the base of both sets of lysimeters. For 22 Na, this generally occurred in the bottom 10 cm, and for 137 Cs in the bottom 2 cm, again indicating the greater mobility of 22 Na. As in the first lysimeter experiment, fairly uniform distributions of both radionuclides, albeit at very low activity concentrations, were evident above these accumulations up to the soil surface, indicating rapid (within 1 month) migration through the lysimeter. Whereas the pattern in the measured 137 Cs profiles remained relatively unchanged over the remainder of the experiment, the dominant fraction of 22 Na, initially present in the bottom 10 cm of soil, continued to migrate up through the lysimeters of both soil types over the summer months, i.e. up to and including September 1995. At the April 1996 sampling, the net negative water fluxes to the lysimeters over the winter and spring months (see Figs. 4.15 and 4.16), probably led to the markedly lower activity concentrations of 22 Na observed in the soil at this time, i.e. 22 Na was leached from the soil over this relatively long period. Overall, the two soil types exhibited similar behaviour in relation to the migration of 22 Na. 8.3.2.2.
Ryegrass biomass and activity concentrations
The mean monthly production of ryegrass biomass in the lysimeters is shown in Fig. 8.25. For both soil types, yields were greatest in July 700
Ryegrass dry biomass (g)
May 3, 2007
Silwood
600
Longlands
500 400 300 200 100 0 July
August
September
October
April
Fig. 8.25. Mean monthly ryegrass biomass production from the Phase II Silwood and Longlands lysimeters (July 1995–April 1996).
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
320
Biosphere Implications of Deep Disposal of Nuclear Waste
1995 and tended to decrease over time. Comparing the effect of soil type, the Silwood soil gave the greater yield in July 1995 but then lower yields until April 1996. The yields were much greater than those observed in the subsequent soil column experiments (see below), due to the much larger surface area of the lysimeters, compared with the soil columns, and the beneficial effect of natural light. Mean 22 Na activity concentrations of the ryegrass from the Silwood and Longlands lysimeters ranged from around 5 to 100 Bq g−1 (Fig. 8.26), i.e. generally greater than those found in the previous lysimeter experiment. The values were generally greater in the Longlands soil than in the Silwood soil, but, in both cases, increased from initially relatively low levels in July up to a peak in September. The fact that uptake was recorded at all in the first month of the experiment indicates that an overlap between the plant roots and the soil contamination occurred rapidly. Indeed, root density measurements taken from these lysimeters in July (Fig. 8.27) showed that roots were present towards the base of the lysimeters, where the greatest amounts of activity were also found. Increasing uptake through August and September would be expected due to the relatively high evapotranspiration during these months (Figs. 4.14 and 4.15). Thus, not only would the plants be removing greater quantities of (contaminated) water from the soil, but also upwards soil Ryegrass activity concentration (Bq g-1)
May 3, 2007
120 Silwood 100
Longlands
80 60 40 20 0 July 22
August
September
October
April
Fig. 8.26. Mean Na activity concentrations of ryegrass from the Phase II Silwood and Longlands lysimeters (July 1995–April 1996).
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Height above geotex (cm)
Radiocations
321
70
July
60
Aug Sept
50
Apr 40
July
30
Aug Sept
20
Apr
10 0 0
2
4 Root density (cm3 cm -3)
6
8
Fig. 8.27. Mean root density for the Phase II Silwood (solid lines) and Longlands (dashed lines) lysimeters (July 1995–April 1996). 4000 3500 3000 Transfer factor
May 3, 2007
2500 Silwood
2000
Longlands
1500 1000 500 0 July
August
September
April
Fig. 8.28. Mean 22 Na soil-plant transfer factors for the Phase II Silwood and Longlands lysimeters (July 1995–April 1996).
migration of activity into zones of increasing root density occurred. The reductions in plant uptake in October 1995 and then April 1996 would seem likely to be due to the reverse of these processes, i.e. reduced transpiration and increased leaching of activity away from the dominant rooting zone. The calculated weighted transfer factors for 22 Na are shown in Fig. 8.28. No transfer factor was calculated for October because of the lack of soil sampling in this month. The transfer factors ranged from 400–3700 and suggest that greatest transfer occurred in the Longlands soil, especially in August and
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
322
Biosphere Implications of Deep Disposal of Nuclear Waste
Ryegrass activity concentration (Bq g-1)
September. The higher transfer factors observed here, compared with the previous lysimeter experiment, may suggest that higher transfer of 22 Na from soil to grass occurs than from soil to wheat. Mean 137 Cs activity concentrations were much lower than those of 22 Na, but similar to those observed for 137 Cs in the first lysimeter experiment, ranging from 0.1 to 2.5 Bq g−1 (Fig. 8.29). Mean weighted transfer factors for 137 Cs are shown in Fig. 8.30. These were low compared with those for 22 Na (10–185) but again showed some degree of concurrence with those obtained in the first lysimeter 3 Silwood 2.5
Longlands
2 1.5 1 0.5 0 July
August
September
October
April
Fig. 8.29. Mean 137 Cs activity concentrations of ryegrass from the Phase II Silwood and Longlands lysimeters (July 1995–April 1996). 200 180 160 Transfer factor
May 3, 2007
140 120
Silwood
100
Longlands
80 60 40 20 0 July 137
August
September
April
Fig. 8.30. Mean Cs soil-plant transfer factors for the Phase II Silwood and Longlands lysimeters (July 1995–April 1996).
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
323
experiment. The Silwood soil generally showed lower activity concentrations and transfer factors than the Longlands soil, particularly in July and August. The relatively high values for the Longlands soil in July and August again indicate the rapidity with which the ryegrass roots penetrated to the contaminated soil at depth. However, no marked differences, which would explain greater uptake in the Longlands soil, between the root profiles of the Silwood and Longlands soils were observed (Fig. 8.27). Again, the importance of the direction of the net water flux in determining the degree of plant uptake is apparent from the observation that the summer months gave greater activity concentrations of 137 Cs than October and April 1996. 8.3.3.
Six month intact soil column experiment (Phase III)
8.3.3.1.
Soil activity concentrations
The mean activity concentration profiles of 22 Na detected in the soil are shown in Figs. 8.31 to 8.33 for the 1, 3 and 6 month columns, Soil activity concentration (Bq g-1) 0
5
10
15
20
25
0 5 10
Silwood Robertgate Wellesbourne
15 Soil depth (cm)
May 3, 2007
20 25 30 35 40 45 50
Fig. 8.31. Mean 22 Na soil activity concentrations in the Phase III intact Silwood soil columns at 1 month.
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
324
Biosphere Implications of Deep Disposal of Nuclear Waste Soil activty concentration (Bq g-1) 0
5
10
15
20
25
30
35
40
0 Silwood
5
Robertgate
10
Wellesbourne
Soil depth (cm)
15 20 25 30 35 40 45 50
Fig. 8.32. Mean 22 Na soil activity concentrations in the Phase III intact Silwood soil columns at 3 months. Soil activity concentration (Bq g-1) 0
10
20
30
40
50
60
70
0 5
Silwood Robertgate
10
Wellesbourne
15 Soil depth (cm)
May 3, 2007
20 25 30 35 40 45 50
Fig. 8.33. Mean 22 Na soil activity concentrations in the Phase III intact Silwood soil columns at 6 months.
ch08
May 3, 2007
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
325
respectively. At 1 month the dominant zone of activity was in the bottom 10 cm of soil. In each soil type, this accumulation was seen to progress up the soil profile over the course of the experiment. For example, by the end of 3 months, it had moved up to the 35–40 cm layers of the Silwood and Wellesbourne soils. In addition, activity concentrations between 40 and 48 cm of all the soils had increased compared with the 1 month data, suggesting further accumulation in this region. These trends were continued at 6 months, with relatively high activity concentrations observed up to the Silwood soil surface. In the Robertgate soil, the dominant 22 Na zone had moved up to the 30–35 cm layer, and in the Wellesbourne soil, up to the 25–30 cm layer. Furthermore, the highest concentrations of 22 Na were observed at the base of the Robertgate columns and these were much greater than those at 3 months. Indeed, increases in activity concentrations were noted for all the soil types relative to the 3 month data. In the same experiment, 36 Cl was observed to migrate to, and accumulate at, the soil surface by 6 months (described in Chapter 4). This indicates that retardation, due to soil adsorption, of the 22 Na was taking place within the columns of all three soil types. In relation to the effect of soil type, differences in the migration of 22 Na were observed. Migration was somewhat retarded in the Robertgate and Wellsbourne soils compared with the Silwood soil. This may be explained by the lower clay and organic matter content of the Silwood soil, since these are potentially important sorbing surfaces for cations. The mean activity concentrations of 134 Cs in the soil profile are shown in Figs. 8.34 to 8.36 for the 1, 3 and 6 month columns, respectively. Low activity concentrations for this radionuclide, compared with 22 Na, were observed due to its adsorption onto plastic surfaces within the column system, i.e. before it reached the soil. Nevertheless, the trend in 134 Cs distribution can still be seen. At all sampling times, the vast majority of the 134 Cs was within the lowermost layer (48–50 cm depth) for each soil type. The activity concentration of this layer tended to increase over time, suggesting that accumulation, due to adsorption onto the soil solids, was taking place. Given the importance of clay minerals (quantity and quality) in
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
326
Biosphere Implications of Deep Disposal of Nuclear Waste Soil activity concentration (Bq g-1) 0.0
0.1
1.0
10.0
0 5
Silwood Robertgate
10
Wellesbourne Soil depth (cm)
15 20 25 30 35 40 45 50
Fig. 8.34. Mean 134 Cs soil activity concentrations in the Phase III intact Silwood soil columns at 1 month.
Soil activity concentration (Bq g-1) 0.0
0.1
1.0
10.0
0 5 10
Silwood Robertgate Wellesbourne
15 Soil depth (cm)
May 3, 2007
20 25 30 35 40 45 50
Fig. 8.35. Mean 134 Cs soil activity concentrations in the Phase III intact Silwood soil columns at 3 months.
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
327
Soil activity concentration (Bq g-1) 0.0
0.1
1.0
10.0
100.0
0 5
Silwood Robertgate
10
Wellesbourne
15 Soil depth (cm)
May 3, 2007
20 25 30 35 40 45 50
Fig. 8.36. Mean 134 Cs soil activity concentrations in the Phase III intact Silwood soil columns at 6 months.
the sorption of caesium, it is likely that these surfaces were largely responsible for the observed sorption. Although differences in clay content between the three soils was observed, the lack of marked differences in the behaviour of 134 Cs suggests that in all soils, the amount and type of clay present was sufficient to result in significant adsorption. However, as was noted for 137 Cs in the lysimeter experiments, migration above the dominant region of accumulation was observed, particularly in the 40–48 cm depth region. Moreover, very small, but detectable levels of 134 Cs were found up to the soil surface. Even at the 1 month sampling, 134 Cs was detected at the surface of the Silwood and Robertgate soils. This was also found at 3 and, to a lesser extent, 6 months with the Robertgate soil consistently showing the highest activity concentrations in the upper regions of the soil columns.
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
328
Biosphere Implications of Deep Disposal of Nuclear Waste
8.3.3.2.
Ryegrass biomass and activity concentrations
Mean values for the yields of ryegrass dry biomass over the 6 months of the intact soil column experiment are shown in Fig. 8.37. These mean values were predominantly between 0.5 and 2.0 g of dry weight for all three soils, with the lowest values observed in Months 3 and 4. Activity concentrations of 22 Na and 134 Cs for the ryegrass shoots are shown in Figs. 8.38 and 8.39, respectively. 22 Na activity
Ryegrass dry biomass (g)
3.00 2.50 2.00 Silwood 1.50
Robertgate Wellesbourne
1.00 0.50 0.00 Month 1
Month 2
Month 3
Month 4
Month 6
Fig. 8.37. Mean ryegrass dry biomass at each harvest of the Phase III intact soil columns.
Ryegrass activity concentration (Bq g-1)
May 3, 2007
400 350 300 250
Silwood Robertgate
200
Wellesbourne
150 100 50 0 Month 1
Month 2 22
Month 3
Month 4
Month 6
Fig. 8.38. Mean ryegrass Na activity concentrations at each harvest of the Phase III intact soil columns.
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
Ryegrass activity concentration (Bq g-1)
May 3, 2007
329
0.6 0.5 0.4 Silwood Robertgate
0.3
Wellesbourne 0.2 0.1 0 Month 1
Month 2
Month 3
Month 4
Month 6
Fig. 8.39. Mean ryegrass 134 Cs activity concentrations at each harvest of the Phase III intact soil columns.
concentrations tended to increase over time for each of the soils. Much greater 22 Na activity concentrations were observed in the Silwood soil (up to around 350 Bq g−1 ) than in the other two soils, which is consistent with the greater upwards soil migration of 22 Na into the dominant rooting zone of the Silwood soil. The Robertgate and Wellesbourne soils gave similar ryegrass activity concentrations to one another, which were consistently below 30 Bq g−1 , probably due to the greater depth of soil contamination. In contrast, the 134 Cs data show that uptake by perennial ryegrass from the Silwood soil was the lowest. Indeed, in many months no detectable uptake by the ryegrass in the Silwood soil columns (and, to a lesser extent, the Robertgate and Wellesbourne columns) was observed. Ryegrass from the Robertgate columns reached the greatest activity concentrations, which is consistent with the greater upwards soil migration observed in this soil. However, these activity concentrations were still low (below 0.6 Bq g−1 ) compared with, for example, the 22 Na data. No clear trend in the uptake of 134 Cs over time was observed. Again, the significant accumulation of much of the added 134 Cs at the base of the soil columns, generally within the water table, is likely to have been a major factor limiting the uptake by the ryegrass. Soil-plant transfer factors for 134 Cs and 22 Na in the 6 month soil columns were calculated using the weighted approach (Eqs. (4.1) and (4.2)). For
ch08
May 3, 2007
14:42
330
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
22 Na
values of 874, 164 and 48 were found for the Silwood, Robertgate and Wellsbourne soils, repectively. For 134 Cs, values of 0, 14 and 0 were found for these soils, respectively. The values confirm the relatively high transfer of 22 Na, particularly in the Silwood soil. 8.3.3.3.
Soil transport simulations
Modelling of the soil-plant behaviour of 22 Na in the intact soil columns was carried out using the approach given in Chapter 3. Hydrological modelling was followed by transport simulations. Whilst average data are reported in the description of the experimental results above, the modelling focused on simulating individual (duplicate) 6 month columns. The column identifying codes used in these simulations are given in Table 8.1. Whereas the details of the parameterisation of the hydrological modelling are given in Chapter 3, the parameters, and their values, required for the Table 8.1. Intact soil column identifying codes used in the Phase III model simulations. Identifying Code
Soil Column Description
6M02 6M04 6M06 6M08 6M10 6M12
Silwood soil, replicate A Silwood soil, replicate B Robertgate soil, replicate A Robertgate soil, replicate B Wellsbourne soil, replicate A Wellsbourne soil, replicate B
Table 8.2. Parameter values used in the 22 Na model simulations of the Phase III intact soil columns. Parameter
Units
∗ Soil Diffusion Coefficient (Dm ) Soil Dispersivity (dL ) Half-life Decay Constant Sorption Coefficient (Kd ) calibration parameter Root Sorption Coefficient (α ) calibration parameter
cm2 s−1 cm yr s−1 cm3 g−1 cm s−1
Value 1.0 × 10−6 1.0 2.6 8.5 × 10−9 0.0, 0.5, 1.0, 1.5, 2.0 4.0 × 10−11
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
331
transport simulations are given in Table 8.2. It should be noted that the soil and root sorption coefficients were calibration parameters, i.e. their values were adjusted in order to fit the model to the data. In the case of the soil sorption coefficient, a range of values was used (0.0 to 2.0 cm3 g−1 ) in order to assess the sensitivity of the model to this parameter.
Column 6M02 0 5
Soil depth (cm)
10
Observed
15
Simulated (Kd=1 l/kg)
20
Simulated Max/Min
25 30 35 40 45 50
0
5
10
15
20
25
Radionuclide concentration (kBq/kg) Column 6M04
0
Soil depth (cm)
May 3, 2007
5 10 15 20 25 30
Observed Simulated (Kd=1 l/kg) Simulated Max/Min
35 40 45 50
0
5
10
15
20
25
Radionuclide concentration (kBq/kg) Fig. 8.40. Simulated and observed Phase III intact Silwood soil columns.
22
Na soil activity concentrations in the
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
332
Biosphere Implications of Deep Disposal of Nuclear Waste
Figure 8.40 shows the simulated and observed 22 Na concentration profiles for the two, duplicate, 6 month Silwood Park soil columns, identified as 6M02 and 6M04. It can be seen that the observed concentrations were contained within the envelope of simulations using sorption values ranging from 0.0 to 2.0 cm3 g−1 . However, of more interest is that the two model simulations employing a soil sorption coefficient of 1.0 cm3 g−1 were able to reproduce the markedly different profiles for both columns reasonably well. This demonstrates that the disparity between the two columns arose directly from their differing water flux rates, as can be seen from Fig. 8.41. This result again reinforces previous observations that, when seeking to simulate the environmental behaviour of relatively mobile contaminants, it is critical that the soil water fluxes are represented accurately (Butler & Wheater, 1999b). The main weakness in the result for column 6M02 was that the observed concentration profile was almost linear between the base and 10 cm from the soil surface. This contrasts with the more curved shape of the model simulation, which arose from the solution of the advection-dispersion equation. The reason for this discrepancy is unclear, although it may have been due to a degree of heterogeneity in the soil structure of the intact soil cores that was not accounted for in the model parameterisation.
8000
Silwood
Robertgate
Wellsbourne
7000 Total water influx (ml)
May 3, 2007
6000 5000 4000 3000 2000 1000 0 6M02
6M04
6M06
6M08
6M10
6M12
Soil column indentifier
Fig. 8.41. Total water influx to the duplicate, 6 month, intact soil columns (Phase III).
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Soil depth (cm)
Radiocations
333
Column 6M06
0 5 10 15 20 25 30 35 40 45 50
Observed Simulated (Kd=1 l/kg) Simulated Max/Min
0
5
10 15 20 25 30 35 40 45 50 55 60 Radionuclide concentration (kBq/kg)
Soil depth (cm)
May 3, 2007
Column 6M08
0 5 10 15 20 25 30 35 40 45 50
Observed Simulated (Kd=1 l/kg) Simulated Max/Min
0
5
10 15 20 25 30 35 40 45 50 55 60 Radionuclide concentration (kBq/kg)
Fig. 8.42. Simulated and observed 22 Na soil activity concentrations in the Phase III intact Robertgate soil columns.
By contrast, the Robertgate & Wellesbourne columns showed far less variability in the water influx (Fig. 8.41). This is also reflected in the model simulations for the replicate 6 month columns. Figure 8.42 shows the results for the Robertgate soil columns (identified as 6M06, and 6M08). Interestingly, in spite of the slightly higher clay content of this soil compared with the Silwood soil, the model simulation utilising a soil sorption coefficient of 1.0 cm3 g−1 still gave a reasonable representation of the observed data. However, there is a large discrepancy at the base of column 6M08 where the observed soil activities in the bottom 4 cm of soil are between 2 and 4 times
ch08
14:42
334
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
higher than the model simulation. This may reflect the presence of a more highly sorbing soil layer. Indeed, the Kd value of 2.0 cm3 g−1 , whilst an improvement, still underestimated this accumulation by around a factor of 2. The model performed well in reproducing the soil concentration profile for the two Wellesbourne columns (identified as 6M10 and 6M12), again using a soil sorption coefficient of 1.0 cm3 g−1 (Fig. 8.43). Overall, these results indicate a certain degree of parameter consistency between soil types, which gives some degree of confidence in a priori parameter estimation.
Soil depth (cm)
Column 6M10
0 5 10 15 20 25 30 35 40 45 50
Observed Simualted (Kd=1 l/kg) Simulated Max/Min
0
5
10
15
20
25
Radionuclide concentration (kBq/kg) Column 6M12 0 5 10
Soil depth (cm)
May 3, 2007
Observed
15
Simulated (Kd=1 l/kg)
20
Simulated Max/Min
25 30 35 40 45 50 0
5
10
15
20
25
Radionuclide concentration (kBq/kg)
Fig. 8.43. Simulated and observed 22 Na soil activity concentrations in the Phase III intact Wellsbourne soil columns.
ch08
May 3, 2007
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
8.3.3.4.
335
Ryegrass uptake simulations
Simulation of the total ryegrass uptake of 22 Na over the 6 month period was carried out as described in Chapter 3. The value of the root sorption coefficient is shown in Table 8.2. In the 6 month columns, ryegrass shoots (> 5 cm) were harvested at 1 month, 2 months, 3 months, 41/2 months and 6 months. At 6 months, the 0– 5 cm portion of ryegrass was also harvested and analysed. The total 22 Na uptake of the harvested crop (in Bq) at each of these periods and for each of the soils was compared with the simulated crop uptake. The simulated and observed uptake of 22 Na by the ryegrass is shown in Figs. 8.44 and 8.45. As mentioned above, there was a general increase in uptake of 22 Na over time as the contaminated water moved up the column and came into contact with an increasing volume of roots. During the first four harvests there was a tendency for the model to over-predict. This may have been due to the sensitivity of the small overlap of contaminated soil with the low root densities at the base of the soil column. During the final six weeks of the column experiments, there was a marked increase in observed 22 Na uptake (Fig. 8.38). In particular, the Silwood Park soil column, 6M02, had an uptake of 22 Na that was vastly greater than that for the other five columns. This was presumably a result of the much greater migration of 22 Na up the soil profile in column 6M02 compared with the others, which was, in turn, probably due to the higher water influx observed in this column. However, the model simulation for 6M02 does not reflect this effect. It is not entirely clear why the simulated uptake was not substantially greater than that for the other Silwood column (6M04), particularly when the marked difference in migration penetration between the two columns is considered. Owing to the influence of the observed uptake in column 6M02 on the plotted results for the final harvest, a second plot using a logarithmic scale is also given. This shows that the simulated uptake in column 6M04 was comparable to 6M02, whereas the observed value is about two orders of magnitude lower. The simulated and observed results for the Robertgate columns (6M06 and 6M08) and one of the
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
336
Biosphere Implications of Deep Disposal of Nuclear Waste First Harvest
Crop Uptake (Bq)
30 25 20 15 Observed
10
Simulated
5 0
6M02
6M04
Silwood
6M06
6M08
Robertgate
6M10
6M12
Wellesbourne
Second Harvest
Crop Uptake (Bq)
60 50 40 30 Observed
20
Simulated
10 0
6M02
6M04
Silwood
6M06
6M08
Robertgate
6M10
6M12
Wellesbourne
Third Harvest
Crop Uptake (Bq)
May 3, 2007
80 70 60 50 40 30 20 10 0
Observed Simulated
6M02
6M04
Silwood
6M06
6M08
Robertgate
6M10
6M12
Wellesbourne
Fig. 8.44. Simulated and observed 22 Na uptake by ryegrass over the first three harvests of the intact soil columns (Phase III).
ch08
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
337
Fourth Harvest
Crop Uptake (Bq)
160 140 120 100 80 60 40 20 0
Observed Simulated
6M02 6M04 6M06 6M08 6M10 6M12 Silwood
Robertgate
Wellesbourne
Fifth Harvest
Crop Uptake (Bq)
3000 2500 2000
Observed
1500
Simulated
1000 500 0 6M02 6M04 6M06 6M08 6M10 6M12 Silwood
Robertgate
Wellesbourne
Fifth Harvest
1.E+04 Crop Uptake (Bq)
May 3, 2007
1.E+02
Observed Simulated
1.E+00 1.E-02 6M02 6M04 6M06 6M08 6M10 6M12 Silwood
Robertgate
Wellesbourne
Fig. 8.45. Simulated and observed 22 Na uptake by ryegrass over the final two harvests of the Phase III intact soil columns. The two lower charts show the same data but the lowest uses a log scale.
ch08
May 3, 2007
14:42
338
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Wellesbourne columns (6M10) compare reasonably well. However, the model greatly under-predicted the uptake in the Wellesbourne soil column, 6M12. Nevertheless, given the highly variable behaviour of biological systems, the model provided a useful insight into soilto-plant transfer of 22 Na. Overall, its performance was regarded as reasonably encouraging, particularly when the limited amount of calibration is considered. Even so, the results also indicate that other factors not included in the model formulation (e.g. soil nutrient status) may have been affecting the uptake of 22 Na by ryegrass.
8.4.
General Discussion
Given the electro-positive nature of Co, Cs and Na, a high degree of soil adsorption for each radionuclide would be expected due to sorption reactions with negatively charged soil surfaces (e.g. organic matter and clay minerals). Therefore, in comparison with electronegative radionuclides (e.g. 36 Cl, discussed in Chapter 4), a lower extent of upwards soil migration would also be expected. However, in the lysimeter experiments, despite a dominant zone of activity accumulation at the base of the soil profile, the potential for radioactive isotopes of these cations to reach the soil surface was exhibited. Nevertheless, the activity concentrations above the zone of accumulation were generally very low, for example compared with those from the 36 Cl studies. This highlights the importance of drawing a distinction between the extent of migration, and the amounts of activity transported, when assessing the mobility of the radionuclides. On this basis, 22 Na can be considered the most mobile of the radiocations studied here, since it generally showed the highest activity concentrations throughout most of the soil profile. Overall, the extent of migration (i.e. to the soil surface) of the cationic radionuclides in the lysimeter experiment was the same as that of the anions 36 Cl and 99 Tc also used in the lysimeter experiments (described in Chapters 4 and 6 respectively). This was surprising given that, in contrast to anionic elements, cationic elements would be expected to sorb to the soil due to their positive charge.
ch08
May 3, 2007
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
339
Consequently, the migration for cations would be expected to be much lower. However, it should be pointed out that the activity concentrations of the radioanions towards the soil surface were much greater than for the radiocations and, in this sense, can be considered more mobile. For radionuclides of low Kd , water flux is likely to play a dominant role in determining their migration behaviour. This probably explains the greater migration of 22 Na, compared with 60 Co and radiocaesium, since this radionuclide has a relatively low Kd of approximately 1–10 cm3 g−1 (Sheppard and Thibault, 1990) and is, therefore, likely to be relatively poorly sorbed. Thus, under conditions of a net upwards moisture flux (e.g. 1990 in the first lysimeter experiment, the summer months in the second lysimeter experiment and the entire soil column experiment), 22 Na migrated upwards through the soil. Similarly, when net negative water fluxes were observed during the winter and spring months of the second lysimeter experiment, leaching (downwards movement) of the 22 Na was apparent. Nevertheless, because it shows a stronger interaction with soil surfaces than radionuclides such as 36 Cl, its migration in both directions was somewhat retarded compared with the behaviour of 36 Cl in these lysimeters (described in Chapter 4). This explains why activity concentrations of 22 Na towards the soil surface of both lysimeter experiments were much lower than those of 36 Cl in the same experiments, i.e. substantially more 22 Na than 36 Cl was adsorbed on its passage through the soil profile. This is borne out in the modelling studies where a Kd value of 1.0 cm3 g−1 was required to simulate the behaviour of 22 Na, compared with 0.01 cm3 g−1 for 36 Cl. In the soil columns, the upwards flux was likely to be low, particularly when compared with the lysimeters over the summer months, resulting in the slower migration up through the soil. However, the absence of a downwards flux in these experiments resulted in the continued upward migration of 22 Na throughout the experiment. Significant migration towards the soil surface was therefore seen. Despite elevated activity concentrations of 60 Co and 137 Cs, throughout the lysimeter soil profile, the dominant feature of the activity distribution for these radionuclides was their accumulation
ch08
May 3, 2007
14:42
340
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
at the base of the lysimeters. This was further evidenced in the soil column experiment for 134 Cs. This indicates that upon reaching the soil, the dominant process controlling the behaviour of these radionuclides was strong adsorption. This is a reflection of their relatively high Kd values, reported by Sheppard & Thibault (1994) as 1.3× 103 and 4.6 × 103 cm3 g−1 for 137 Cs and 60 Co, respectively (substantially greater than that of 22 Na). However, due to these high Kd values, it is difficult to explain their rapid migration to the soil surface of the lysimeters, and the occasional presence of 134 Cs at the surface of the soil columns. For example, the values suggest retardation factors of around 2.5 to 8 × 103 , meaning that an advective water velocity of around 1 m yr−1 should result in upwards migration of 60 Co and 137 Cs over just a fraction of a millimetre in a year. Nevertheless, if chemical conditions within the soil were such as to reduce the sorption potential of these radionuclides, movement with the water flux may seem more reasonable. One possible explanation for this effect could be the attachment of these cationic radionuclides to colloids present in the soil. Indeed, such an explanation has already been proposed for the transport of selenium observed in the Phase VIII soil columns (Sec. 7.3.2.6). Although a minor process, if a few percent of the total amount of radiocations that entered through the base of the soil were in this form, this could readily explain the low activities observed up the soil profile. Another, and possibly related, aspect is that the presence of anoxic conditions within the region of the water table in all these experiments was recorded (e.g. see Figs. 5.10 and 6.3). Although the onset of such conditions would not have a direct effect on the speciation of 60 Co and radiocaesium, their effect on the general soil chemistry may have been influential in determining the behaviour of these radionuclides. For example, low redox potential is known to bring about the formation of ammonium ions, as a reduced nitrogen form. The ammonium ion has a strong ability to displace other ions from exchange sites on the soil solid phase and is often used as a standard ion in analytical extractions of metals and nutrient ions from soils. Therefore, an increase in ammonium ions in soil would be expected to lead to decreased sorption (i.e. decreased Kd ) of other cations. The migration potential of such
ch08
May 3, 2007
14:42
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Radiocations
341
cations would thus be increased. This process is suspected to be particularly important for Cs since fixation of this element has been shown to be markedly decreased due to an increase in ammonium ions brought about by low redox conditions (Pardue et al., 1991). Therefore, in the case of 60 Co and radiocaesium, this might suggest that a combination of hydrological and altered chemical conditions accounted for the rapid upwards migration observed. An alternative explanation for the migration of these strongly sorbed radionuclides above the zone of accumulation may involve transport through plant roots. Indeed, this may seem more reasonable since, even for the relatively poorly sorbed 22 Na, migration to the soil surface within one month (in the second lysimeter experiment), suggests movement over and above that which would be expected purely on the basis of an advective flux. Even towards the start of the experiments, the presence of roots in the zones of dominant soil contamination (i.e. at the base) was observed. Moreover, the potential for root uptake in this region is potentially increased through the reduction in sorption of these radionuclides due to the redox-dependent processes described above. Following uptake, two scenarios by which the activity is returned to the soil higher up the profile are possible. First is that the radionuclide is taken up by the plant roots, transferred upwards through the roots, and then exuded by the roots into the soil further up the profile. Second is that radionuclide is taken up by the roots, transferred to the above ground biomass (e.g. leaves), and then returned to the soil as those leaves fall from the plant and become incorporated into the soil through biological action. The rapid (within one month) migration to the soil surface of 137 Cs and 22 Na in the second lysimeter experiment suggests that the first of these processes may be the most likely, since it would seem that this is insufficient time to allow for the translocation, senescence, and degradation required by the second scenario. As might be expected, the most mobile of the radiocations in the soil (22 Na) also showed the greatest transfer to the above-ground biomass, due to its relatively low sorption and, hence, presence in the soil solution. In general, conditions conducive to upwards migration of the radionuclides through the soil also appeared to be conducive
ch08
May 3, 2007
14:42
342
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
to plant uptake. Consequently, greatest transfer of activity occurred during the periods of greatest water influx (e.g. the summer months of the second lysimeter experiment). During wetter periods of the lysimeter experiments, evapotranspiration was reduced and 22 Na was seen to be leached down the soil profile and, hence, away from the dominant rooting zone. Uptake was thus reduced. The generally low degree of soil transport of radiocaesium and 60 Co led to low transfer of these radionuclides from the soil to plants. Transfer factors for both of these radionuclides were consequently lower than those of 22 Na. Nevertheless, transfer of activity into the plant again appeared to be related to the flux of water through the system. For example, transfer was again greatest during 1990 of the first lysimeter experiment and the summer months of the second lysimeter experiment, presumably due to the high evapotranspiration flux at these times.
ch08
May 3, 2007
14:44
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
SECTION 4
CONCLUSIONS AND RECOMMENDATIONS
343
ch09
May 3, 2007
14:44
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
This page intentionally left blank
344
ch09
May 3, 2007
14:44
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
CHAPTER 9
Conclusions
9.1.
Overview
The migration of solutes in vegetated soils and their associated uptake by plants is an important and complex area of science that is fundamental to understanding the functioning of vegetation and the major biogeochemical cycles and that underlies a wide range of issues of environmental management. Applications include the assessment of risks of pollution from surface deposition and subsurface sources and the potential for contamination remediation, for example by phytoremediation. This monograph presents an extensive programme of experimental research that considers radionuclides of contrasting chemical properties. Although the context for the work has been risk assessment for the subsurface disposal of radioactive waste, the radionuclides are direct analogues for a much wider set of potential contaminants, and thus the work is of wide generic applicability. The upwards migration of contaminants is controlled by complex interactions within the soil-plant-atmosphere system, and the field and laboratory experiments have been carefully designed to represent these at a realistic scale and to allow comprehensive monitoring of the hydrological, geochemical and biological controls on solute migration and uptake. The work has benefited greatly from the creation of a truly multidisciplinary team, and we are not aware of any other concerted research programme, at least in the radiochemical context, that has addressed hydrological, geochemical and biological aspects in comparable detail. 345
ch09
May 3, 2007
14:44
346
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
These experiments have been successful in providing new insights, as well as new data, that have important implications for the safety assessment of nuclear waste disposal. The data obtained have additional value in providing a basis for model testing and intercomparison — the use of the Phase I experiments for the IAEA BIOMOVS intercomparison has already been noted in Chapter 3. An important aspect of the experiments is the complexity of response, which has meant that simple interpretation of results has been of limited value. Our approach has been to use both simple and detailed, physically-based, numerical models to support the design of the experimental programme and to analyse the data. At a basic level, it has been necessary to use models to establish experimental boundary conditions and experimental budgets. More fundamentally, the models allow us to explore hypotheses of response and provide a means of parameter estimation and of generalisation of the results. Such direct integration of experimental and modelling research is often advocated, but rarely achieved. We believe that it has been fundamental to the success of the programme.
9.2. 9.2.1.
Summary of Results Chlorine
In the one-dimensional soil-plant system the upwards migration of radionuclides above a near surface water table occurs primarily due to water flux, which itself is largely the result of the uptake of water by the plant root system. In the first two experimental phases, based on outdoor lysimeters, the natural variability of climate led to an annual cycle of upwards water flux in Summer, which could be interrupted by rainfall events, downwards drainage in Winter, and transitional behaviour in Spring and Autumn. In the subsequent column experiments, following the establishment of the crop, only upwards flux occurred. The water flux is thus determined by the specified environmental conditions, the stage of crop development, and the health and vitality of the crop.
ch09
May 3, 2007
14:44
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Conclusions
347
In the case of chlorine, it is generally assumed that geochemical interactions with the soil solid phase are negligible (i.e. Kd = 0), and hence it is expected that radiochlorine transport will be determined by the advective water flux with no soil interactions. And indeed, across all the lysimeter and column experiments, upwards migration of 36 Cl from the water tables through the soil was relatively rapid compared to other radionuclides studied within this program. In the experimental columns, the 36 Cl profile in the soil tended to approach a uniform distribution over time, but with an accumulation of 36 Cl at the soil surface. The rates of migration were clearly dependent upon crop transpiration, and hence on crop vitality, with marked differences arising between experimental replicates as a result. In the lysimeters, the periods of greatest upwards migration were related to periods of low rainfall and high evapotranspiration, with exceptionally high migration in the dry summer of 1990 (Phase I). Periods of rainfall were often associated with losses of activity from the soil, due to drainage through the base of the lysimeter. However, in the experiments in which different soil types were used, differences in activity concentrations in differing soil types were observed. In addition, the periods of rainfall in the lysimeter experiments, although associated with leaching of 36 Cl, did not lead to complete removal of the radionuclide from the soil profile, suggesting that some of the 36 Cl was retained by the soil. The experiments therefore provided clear evidence that some degree of interaction was occurring with the soil solid phase. Known mechanisms of chloride retardation within soil include physical entrapment in small dead-end pores although evidence shows that adsorption of 36 Cl onto organic matter is also possible. The best indication of a Kd value from these experiments can be obtained from the modelling; it was not possible to determine Kd values from the experiment directly due to the problem of quantifying the ‘total’ soil activity concentration. The Kd value used throughout the modelling was 0.01 cm3 g−1 , indicating that a small amount of interaction with the solid phase was indeed required to simulate the experimental data. Data from field studies ¨ (Oberg, 1998) and lysimeter experiments (Rodstedth et al., 2003) on stable (non-radioactive) chloride in Sweden have indicated that
ch09
May 3, 2007
14:44
348
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
significant interaction between chlorine and humic substances occurs under natural conditions. One major question is the rate at which trace activities of 36 Cl will join the pool of stable chlorine in natural soil-vegetation systems over the long periods (hundreds of years) of relevance to radioactive waste disposal. Uptake of radionuclides into vegetation is expected to be dependent on the distribution of activity within the soil in relation to root distribution, and on the upwards water flux. In the lysimeters, this was seemingly tempered by the net water flux. When the net water flux was upwards (positive), relatively high uptake was seen. At times of net negative water influx, relatively low uptake was seen, probably due to a combination of low evapotranspiration and a leaching of 36 Cl from the root zone of the crop. In the columns, significant positive correlations between uptake, biomass and evapotranspiration were often seen. This suggests that a healthy, high yielding crop led to high transpiration and, hence, a high uptake of 36 Cl. As with the soil data, this indicates the strong influence of the evapotranspiration stream through the columns on the fate of 36 Cl. Soil-plant transfer factors for 36 Cl ranged from zero to around 50 000, with large differences in transfer factors observed between the experiments. Lowest values were generally observed in the soil column experiment comparing fixed and fluctuating water tables. Here, even with a fluctuating water table seeming to transport 36 Cl into the rooting zone very effectively, lower transfer factors than were previously observed for fixed water tables were found. The relatively low biomass yields of the ryegrass may therefore have been important in limiting uptake. The model simulations suggested that, in some cases, low plant uptake of 36 Cl may have been associated with poor activity balances for the experiment. Only by simulating much greater uptake by the ryegrass could the observed soil activity concentrations be simulated by the model. However, the analytical procedures used for digestion of the plant samples were checked and no potential problems could be identified. This suggests that the ‘lost’ activity was elsewhere within the column system but was not efficiently extracted. A further possible explanation is loss via volatilisation, however, the potential for loss via this pathway is considered to be very low. In
ch09
May 3, 2007
14:44
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Conclusions
349
attempting to account for an imbalance in stable chlorine inputs and outputs in a small scale lysimeter experiment Rodstedth et al. (2003) postulated three mechanisms which may have been responsible; formation of non-volatile chlorinated organic compounds, formation of volatile chlorinated organic compounds and anion exchange. Anion exchange in most soils is known to be weak so the most likely explanation lies with the first two mechanisms. The generally high soil to plant transfer factors for 36 Cl obtained from the various experimental phases of the research programme are explained by the high root uptake coefficients (α ) (Table 9.1)that have been derived from the accompanying model simulations. Although these cover an overall range of two orders of magnitude (1.0 × 10−10 – 1.0 × 10−8 cm2 s−1 ), typical values were around 1.0 × 10−9 cm2 s−1 . In addition, it is important to note that the above results are also dependent on the root density distributions used in the model simulations and, as already noted, these were often difficult to accurately identify, particularly in the case of perennial ryegrass. However, when used in association with the densities derived from these experiments, the above table provides an important outcome from the Table 9.1. studies.
36
Experiment
Description
Crop
Soil+
Root uptake coefficient (α ) [cm2 s−1 ]
Phase I
Lysimeter
Winter wheat
Silwood
4.0 × 10−9 (1990) 2.0 × 10−10 (1991+)
Phase II
Lysimeter
Ryegrass
Silwood Longlands
Phase III
Column
3.6 × 10−9 6.0 × 10−9 )
Cl root uptake coefficients obtained from lysimeter and column
Silwood Robertgate Weelsbourne
1.0 × 10−9
Phase IV
Column
Silwood
1.0 × 10−10 6.0 × 10−10
Phase VII
Column
Silwood
5.0 × 10−9 1.0 × 10−8
+
All soils were re-constituted, apart from undisturbed columns used in Phase III.
ch09
May 3, 2007
14:44
350
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
experimental programme in regard to modelling the fate and transport of radiochlorine in vegetated soils and how soil-to-plant transfer of 36 Cl can be represented for assessment modelling.
9.2.2.
Iodine
Moving from chlorine to iodine, the interactions between solutes and the soil solid phase become more important in influencing radionuclide mobility, and in this case the results clearly show the importance of the redox-dependent chemical speciation of radioiodine, and the associated effects on mobility. This means that biogeochemical effects can be important in influencing the chemical environment, which in turn determines radionuclide mobility. In the mini-column experiments, strong time-dependence of Kd was observed and variable effects of soil moisture content. Initially, lower Kd values were obtained for the drier soil conditions, but once the results had stabilised (after 1 month), this effect was reversed. These effects are associated with redox potential, and hence speciation. It is suggested that short duration batch experiments with high moisture contents may overestimate Kd . In the lower part of the soil columns, microbial consumption of dissolved oxygen took place, thereby creating anoxic conditions. A fluctuating water table was able to increase the extent of this low redox zone. Under such conditions, the preferred state for 125 I was, we infer, the poorly sorbed iodide ion, which was able to migrate relatively freely. However, above the top of the capillary fringe, oxygen was able to diffuse downwards from the soil surface and thereby maintain a high redox state in the upper part of the column. Under these conditions, iodine is expected to be oxidised to the iodate ion which is known to be much more strongly sorbed by the soil solid phase, in particular organic matter. This sorption process meant that the radionuclide was retarded at the redox front with the result that an accumulation of 125 I took place in the oxidation transition zone. Such effects were consistently observed in both soil column experiments. The position of the oxidation transition zone, and hence the
ch09
May 3, 2007
14:44
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Conclusions
351
radioiodine accumulation, was dictated by the hydrological conditions within the column, specifically the movement of the water table. The mini-column experiments provided clear, additional evidence of the effect of low redox potential on the decreased soil adsorption of 125 I. However, it should also be noted that small amounts of 125 I were found to migrate above this redox boundary over time. This suggests that the radionuclide has the ability to move as the oxic, iodate form. Clearly, the long term implications of this should be considered in the risk assessment of radioactive waste disposal. 125 I was found to be poorly transferred to the crop. In general, it appeared that the ryegrass roots did not penetrate into the 125 I contaminated soil towards the base of the soil columns. In the 12 month experiment, this may have been due, at least in part, to the apparently low vitality of the ryegrass seed initially used, and the malfunctioning of the constant environment room in which the columns were housed. However, in the 6 month experiment, where these problems were not apparent, essentially no uptake of 125 I was again observed. Despite the reseeded 9 and 12 month columns in the 12 month experiment showing some uptake into the ryegrass, these amounts were very small in relation to the overall 125 I inventory of the column systems. In conclusion, it is clear that that the redox dependence of sorption, and hence Kd , means that modelling for safety assessment must take into account of the water and associated chemical environment, with strong differences occurring across the interface between the saturated and unsaturated zones (i.e. the water table and capillary fringe). The modelling here required a spatially-variable distribution of Kd to represent the observed profiles (Table 9.2); clearly this empirical adjustment is crude and ideally a geochemical submodel would be developed to represent explicitly the dependence of sorption on redox and hence speciation. Similarly the current modelling capability does not allow for time varying Kd . The reduction in iodine mobility in the unsaturated zone with, in addition, an extremely small plant uptake coefficient (Table 9.2), makes the transfer of radioiodine (such as 129 I) from a contaminated water table into ryegrass (at least) very low. Such a result has important implications for safety assessment calculations undertaken for
ch09
May 3, 2007
14:44
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
352
Biosphere Implications of Deep Disposal of Nuclear Waste Table 9.2.
125
I parameter values from column experiments.
Experiment+
Root uptake coefficient (α ) [cm2 s −1 ]
Soil oxidation state
Sorption coefficient (Kd ) [cm3 g−1 ]
Phase V
6.0 × 10−15 − 4.4 × 10−13
Oxic Anoxic
0.01–0.1 0.6–3.0
Phase VII
N/A
Oxic Anoxic
0.01 1.0–6.0
+
Soils were re-constituted Silwood soil sown with perennial ryegrass.
this radionuclide where transport and vegetation bioaccumulation in the near-surface environment are taken into account. It is important to also consider the implications of an accumulation of 129 I at the anoxic/oxic soil boundary. Such an accumulation could potentially become a significant biosphere reservoir. The influence of the height of the water table in determining the closeness of this accumulation to the soil surface is a further important consideration. More work is needed into the effects of water table fluctuation on the accumulation and release of iodine.
9.2.3.
Technetium
As for iodine, the mobility of technetium is dependent on redoxmediated geochemical interactions. The two experiments to study the upward migration of Tc showed marked differences in the behaviour of the radionuclide. In the soil columns, a zone of strongly reduced soil occurred at the base of the columns (established prior to the introduction of the 95m Tc) and the upward migration of the radionuclide was extremely limited. In most cases, over 90% of the total soil activity was present in the lower-most layer of soil (48–50 cm depth). This was thought to be due to the chemical reduction of the added pertechnetate (TcVII) to TcIV forms. These reduced forms are thought to precipitate from solution in the presence of sulphate reducing bacteria, and to become adsorbed onto soil organic matter.
ch09
May 3, 2007
14:44
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Conclusions
353
In contrast, migration upwards through the lysimeter soil was rapid and did not appear to be strongly limited, despite an apparent zone of 99 Tc accumulation towards the base of the lysimeter. Within around 5 months, the Tc had reached the soil surface in even the 70 cm lysimeters. Although it this difference in migration behaviour could be due to differences in the redox status of the soil between the two experiments, an alternative explanation might be the difference in crop types with the deeper rooting winter wheat roots able to access 99 Tc from the base of the lysimeter compared with the ryegrass used in the soil column experiment. Model simulations of the 95m Tc transport within the soil columns had to take account of the varying water table, which gave rise to strongly hysteretic behaviour. This was reproduced in a simplified way, which nevertheless showed good agreement with experimental data. These simulations allowed for further analysis of the data, particularly in relation to the determination of the Kd values required to successfully simulate the experimental data. In order to reproduce the observed soil activity profiles in the soil columns, high sorption coefficients (between 22 and 60 cm3 g−1 ) were required. These are much higher than the values of around 0.1 cm3 g−1 reported elsewhere for, presumably, oxic soil environments. The accuracy of the activity data for the upper soil layers was questionable due to detection limits. Consequently, there was some concern over the reliability of the sorption coefficients in the anoxic/oxic transition and fully oxic layers. The extent of solute migration in the experiments also defined the degree of plant uptake observed. In the column experiments, activity concentrations of perennial ryegrass were very low and, in many cases, non-detectable. This was most likely due to the apparent lack of overlap between the ryegrass roots and the contaminated soil. Furthermore, the extremely low amounts of technetium measured in the cut ryegrass meant that it was not possible to assess a root uptake parameter value. Because the 99 Tc in the lysimeter experiments readily migrated to the soil surface, a relatively high degree of uptake into the wheat crop was found. Transfer factors were, in most cases, less than 200. Due
ch09
May 3, 2007
14:44
354
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
to the oxic conditions found within the predominant zone of rooting, it can be assumed that the form of technetium taken up by the plants was pertechnetate. The calculated transfer factors are broadly consistent with those previously determined for oxic environments. Despite a potential for roots to penetrate into the anoxic, more highly contaminated, soil, it seems unlikely that this contributed significantly to the observed plant uptake because the plant availability of reduced Tc species has been shown to be very low. The results clearly indicate that redox-dependent chemical speciation is of fundamental importance in determining the potential for upward migration and plant uptake of Tc. In the context of radioactive waste disposal, this has significant implications in relation to the potential for 99 Tc to reach the biosphere. In addition to likely variations in redox potential within the near-field environment, the influence of the water table is likely to lead to wide variations in soil redox potential. Thus, in order for 99 Tc from a radioactive waste repository to reach the biosphere, it would need to migrate through both anoxic and oxic regions of the soil profile. These experiments suggested that significant accumulation of 99 Tc may occur where anoxic soil is encountered. However, the importance of considering 99 Tc in the risk assessment of radioactive waste disposal is highlighted by the facts that, (a) an, albeit small, amount of migration within the anoxic zone of the soil columns did occur, and (b) reduced Tc (TcIV) species appeared to have the potential to become re-oxidised (to TcVII) and, hence, become potentially more mobile. The potential for relatively rapid solute migration and plant uptake of 99 Tc when the soil was not as strongly or consistently reduced was demonstrated by the lysimeter experiment.
9.2.4.
Selenium
Along with the other key radionuclides discussed above, there are major gaps in our knowledge of the fate and transport of selenium within the biosphere. A particular issue for selenium is the role of volatilisation. The experimental programme was designed to investigate transport and uptake of 79 Se in vegetated soils, using 75 Se as a
ch09
May 3, 2007
14:44
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Conclusions
355
surrogate. As above, column experiments were used, in this case for both vegetated and non-vegetated soils. These were supplemented by laboratory bench-top studies of 75 Se sorption (mini-columns) and plant uptake (pots). Complex, and variable solute-solid interactions were observed, and the column results suggested that an additional migration process was occurring. In the mini-columns, six differing soil types were studied (the Silwood soil, three from the region of the Meuse/Haute-Marne underground laboratory in France, and two organic soils from the UK). The mini-columns were incubated at field capacity and saturated moisture contents. Fumigating some of the mini-columns with methyl bromide enabled the effects of microbial activity to be investigated. The addition of a vapour trap to some of the columns also allowed the effect of Se volatilisation to be assessed. Overall, the results showed that Kd values ranged from around 50 to 400 cm3 g−1 across all soil types, suggesting a relatively high degree of sorption. In general, no strong effects of time or moisture content on the Kd values were observed. However, fumigation of the soils often led to a reduction in Kd values suggesting that sorption of 75 Se was a microbially-mediated process. The Silwood sandy loam also showed a strong negative relationship between Kd value and redox potential, indicating that under anoxic conditions, 75 Se sorption was increased. Volatilisation of 75 Se was readily measurable from all of the four soils with vapour traps fitted. No differences between the two organic and two mineral soils were apparent. However, again, fumigation seemed to have an effect, in increasing the observed volatilisation rate. In the pot experiments, Kd values and soil-ryegrass transfer factors for the Silwood soil and two of the French soils (also used in the mini-columns) were determined. Three moisture contents were used, from dry to field capacity. Again, a fumigation (with methyl bromide) treatment was included. Across the soils and treatments, high Kd values, consistent with those from the mini-columns, were observed. Therefore the potential for 75 Se mobility and bioavailability in soil can be considered low. This was further confirmed by the low transfer factors observed in the pot experiment. Whilst some differences
ch09
May 3, 2007
14:44
356
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
in transfer factors were observed between treatments, these could not be consistently explained on the basis of Kd values. Consistently lower Kd values were observed for fumigated soils, than for non-fumigated soils, in both the mini-column and pot experiments and this is thought to explain the greater transfer factors observed in Silwood sandy loam and, particularly, French clay loam, under fumigated conditions. The reason for the apparent increase in 75 Se solubility under fumigated conditions is not clear. However, it seems probable that soil microbes may play an important role in the immobilisation of Se. Thus, in the fumigated soils, where soil microbial populations were presumably reduced, lower sorption was observed. This effect appeared to be more significant in influencing 75 Se solubility and plant uptake than, for example, soil texture, pH, organic matter, moisture content and redox potential. The soil column experiment investigated the upward soil migration, plant uptake and volatilisation of 75 Se from a fixed water table over a 9 month period. As noted above, the experiment included two non-vegetated columns, and selected columns were fitted with volatilisation traps to determine volatile losses of 75 Se from the soil and plant surfaces. Detectable amounts of volatile 75 Se were measured for the three columns that were fitted with volatilisation traps, although the amounts involved were extremely small. It should be noted, however, that the presence of the trapping apparatus had the effect of markedly reducing the transpiration rate and hence the influx of 75 Se. Consequently, the amount lost under normal conditions might be somewhat higher, although still relatively minor in terms of the overall inventory. Soil solution measurements of 75 Se taken using HFSS combined with analyses from destructive sampling at the end of the experiment enabled values of Kd within the column to be calculated. These showed values that were substantially lower than those obtained from either the mini-column or pot experiments. This was further confirmed by supporting model simulations. The upward migration of 75 Se in the columns was mainly limited to the bottom 10–15 cm, where there were low (i.e. negative) redox potentials due to saturated soil moisture conditions. Nevertheless, in spite of this
ch09
May 3, 2007
14:44
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Conclusions
357
limited zone of migration, uptake of 75 Se by the ryegrass was observed to take place within one month of the start of the experiment, indicating active and rapid root penetration down the bulk, if not all, of the column. This was subsequently confirmed during destructive soil analyses. Soil-to-plant transfer factors calculated from these analyses were generally about an order of magnitude larger than those derived from the pot experiments. The increased values were possibly due to a combination of higher water fluxes and lower Kd values. Model simulations supported the observation that the Kd s at the base of the 50 cm soil columns were about an order of magnitude lower than those derived from the mini-columns. However, the model simulations also pointed to the existence of a mobile Se fraction. Although the exact form could not be determined from the experimental measurements, it is known that organic forms of selenium can occur in soils. Modelling studies indicated that this mobile fraction is between 5 and 15% of the total amount of 75 Se (in the form of selenite) entering the column. Whatever its form, model investigations indicated that its mobility was restricted to the saturated, low redox zone at the base of the column. That the bulk of the 75 Se is sorbed is not unduly surprising, as selenite tends to sorb strongly onto oxide surfaces and clay minerals. A method was developed to quantify the root uptake coefficient used in the reactive transport model from soil-to-plant transfer factors obtained from the pot experiments. However, when implemented this led to values substantially less than those observed, which was also consistent with the fact that transfer factors derived from the column experiments were several orders of magnitude higher that those from the pot studies. Alternatively, based on these results, it can be argued that the actual root uptake coefficient α for 75 Se is approximately 1.0 10−10 cm2 s−1 . The discrepancies in Kd and soil-to-plant transfer values derived from the mini-columns and pot experiments in comparison with those obtained from the larger scale column experiments, coupled with a minor (but still important) mobile fraction within the anoxic zone indicate that there are still major challenges for small scale
ch09
May 3, 2007
14:44
358
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
measurements to provide reliable parameter estimates for assessment studies, particularly when the problems of upscaling to field and catchment scale are considered.
9.2.5.
Radiocations
Given the electro-positive nature of Co, Cs and Na, a high degree of soil adsorption for each radionuclide would be expected due to sorption reactions with negatively charged soil surfaces (e.g. organic matter and clay minerals). Therefore, in comparison with electronegative radionuclides (e.g. 36 Cl), a lower extent of upwards soil migration would also be expected. However, in the lysimeter experiments, despite a dominant zone of activity accumulation at the base of the soil profile, the potential for radioactive isotopes of these cations to reach the soil surface was demonstrated. Nevertheless, the activity concentrations above the zone of accumulation were generally very low, for example compared to those from the 36 Cl studies. 22 Na can be considered the most mobile of the radiocations studied here since it generally showed the highest activity concentrations throughout most of the soil profile. Overall, the vertical extent of migration (i.e. to the soil surface) of the cationic radionuclides in the lysimeter experiment was the same as that of the anions 36 Cl and 99 Tc also used in the lysimeter experiments. This was surprising given that, in contrast to anionic elements, cationic elements would be expected to sorb to the soil due to their positive charge. For radionuclides of low Kd , water flux is likely to play a dominant role in determining their migration behaviour. This probably explains the greater migration of 22 Na, compared to 60 Co and radiocaesium. Thus under conditions of a net upwards moisture flux (e.g. 1990 in the Phase I lysimeter experiment, the summer months in the Phase II lysimeter experiment and during all soil column experiments), 22 Na migrated upwards through the soil. Similarly, when net negative water fluxes were observed during the winter and spring months of the second lysimeter experiment, leaching (downwards movement) of the 22 Na was apparent. Nevertheless, its migration in
ch09
May 3, 2007
14:44
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Conclusions
359
both directions was somewhat retarded compared to the behaviour of 36 Cl, as expected due to the greater extent of soil interaction. Hence activity concentrations of 22 Na towards the soil surface of both lysimeter experiments were much lower than those of 36 Cl, i.e. substantially more 22 Na than 36 Cl was adsorbed on its passage through the soil profile. This is borne out in the modelling studies where a Kd value of 1.0 cm3 g−1 was required to simulate the behaviour of 22 Na, compared to 0.01 cm3 g−1 for 36 Cl. In the soil columns, the upwards flux was low, particularly when compared to the lysimeters over the summer months, resulting in the slower migration up through the soil. However, the absence of a downwards flux in these experiments resulted in the continued upward migration of 22 Na throughout the experiment. Significant migration towards the soil surface was therefore seen. Despite elevated activity concentrations of 60 Co and 137 Cs, throughout the lysimeter soil profiles, the dominant feature of the activity distribution for these radionuclides was their accumulation at the base of the lysimeters. This was further evidenced in the Phase III soil column experiment using 134 Cs. This indicates that upon reaching the soil, the dominant process controlling the behaviour of these radionuclides was strong adsorption. This is a reflection of their relatively high Kd values (substantially greater than that of 22 Na). However, due to these high Kd values, it is difficult to explain their rapid migration to the soil surface of the lysimeters, and the occasional presence of 134 Cs at the surface of the soil columns. For example, the values suggest retardation factors of around 2.5 to 8 × 103 meaning that an advective water velocity of around 1 m y−1 should result in upwards migration of 60 Co and 137 Cs over just a fraction of a millimetre in a year. Nevertheless, if chemical conditions within the soil were such to reduce the sorption potential of these radionuclides, movement with the water flux may seem more reasonable. The presence of anoxic conditions within the region of the water table in all these experiments was recorded (e.g. see Figs. 5.10 and 6.3). Although the onset of such conditions would not have a direct effect on the speciation of 60 Co and radiocaesium, their effect on the general soil chemistry may have been influential in determining the behaviour of these
ch09
May 3, 2007
14:44
360
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
radionuclides. For example, low redox potential is known to bring about the formation of ammonium ions, as a reduced nitrogen form. The ammonium ion has a strong ability to displace other ions from exchange sites on the soil solid phase and is often used as a standard ion in analytical extractions of metals and nutrient ions from soils. Therefore, an increase in ammonium ions in soil would be expected to lead to decreased sorption (i.e. decreased Kd ) of other cations. The migration potential of such cations would thus be increased. This process is suspected to be particularly important for Cs since fixation of this element has been shown to be markedly decreased due to an increase in ammonium ions brought about by low redox conditions (Pardue et al., 1991). Therefore, in the case of 60 Co and radiocaesium, this might suggest that a combination of hydrological and chemical conditions accounted for the rapid upwards migration observed. An alternative explanation for the migration of these strongly sorbed radionuclides above the zone of accumulation may involve transport through plant roots. Indeed, this may seem more reasonable since, even for the relatively poorly sorbed 22 Na, migration to the soil surface within one month (in the second lysimeter experiment), suggests movement over and above that which would be expected purely on the basis of an advective flux. Even towards the start of the experiments, the presence of roots in the zones of dominant soil contamination (i.e. at the base) was observed. Moreover, the potential for root uptake in this region is potentially increased through the reduction in sorption of these radionuclides due to the redox-dependent processes described above. Following uptake, two scenarios by which the activity is returned to the soil higher up the profile are possible. First is that the radionuclide is taken up by the plant roots, transferred upwards through the roots, and then exuded by the roots into the soil further up the profile. Second is that radionuclide is taken up by the roots, transferred to the above ground biomass (e.g. leaves), and then returned to the soil as those leaves fall from the plant and become incorporated into the soil through biological action. The rapid (within one month) migration to the soil surface of 137 Cs and 22 Na in the second lysimeter experiment
ch09
May 3, 2007
14:44
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Conclusions
361
suggests that the first of these processes may be the most likely since it would seem that this is insufficient time to allow for the translocation, senescence, and degradation required by the second scenario. As might be expected, the most mobile of the radiocations in the soil (22 Na) also showed the greatest transfer to the above ground biomass, due to its relatively low sorption and, hence, presence in the soil solution. In general, conditions conducive to upwards migration of the radionuclides through the soil also appeared to be conducive to plant uptake. Consequently, greatest transfer of activity occurred during the periods of greatest water influx (e.g. the summer months of the second lysimeter experiment). During wetter periods of the lysimeter experiments, evapotranspiration was reduced and 22 Na was seen to be leached down the soil profile and, hence, away from the predominant rooting zone. Uptake was thus reduced. The generally low degree of soil transport of radiocaesium and 60 Co led to low transfer of these radionuclides from the soil to plants. Transfer factors for both of these radionuclides were consequently lower than those of 22 Na. Nevertheless, transfer of activity into the plant again appeared to be related to the flux of water through the system. For example, transfer was again greatest during 1990 of the first lysimeter experiment and the summer months of the second lysimeter experiment, presumably due to the high evapotranspiration flux at these times. 9.3.
Experimental Design
Each of the experimental systems used during the course of this study has its advantages and limitations. From the initial Phase I and Phase II studies using field- scale lysimeters the move to the use of undisturbed soil columns, re-packed soil columns and finally to ‘mini-columns’ appears to mark a shift from a ‘more’ to a ‘less’ realistic approach in terms of both scale and complexity of experiments. Aside from the obviously lower running costs of soil columns versus lysimeters, however, the use of smaller scale experimental systems has several advantages, as described in Chapter 2.
ch09
May 3, 2007
14:44
362
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Our experience from Phase III onwards has shown that the hugely increased ease of access for the purposes of sampling has improved our understanding of experimental results, particularly when considering the chemical and radiochemical compositions of soil solutions. The hollow fibre solution samplers used for the first time in Phase VI substantially improved the detail at which experimental results could be interpreted and allowed, for the first time in our experimental programme, in situ estimates of Kd to be obtained under realistic conditions of moisture content and redox potential. This is an important advance since the data bases of Kd values commonly available in the wider literature are for the most part obtained using batch techniques with a long history of development but with a dubious record of reliability. Discrepancies between our in situ measured Kd values and those estimated by modelling have been identified, however, and this leads to the more fundamental consideration of just how representative and reliable this parameter is. Even more important for the application of experimental results to safety calculations for radioactive waste disposal is the degree to which parameters measured over periods of a few months to a few years can be extrapolated to much longer time scales under evolving climatic and pedological conditions. Experiments conducted at relatively small scale (soil columns and mini-columns) are useful in identifying key mechanisms such as accumulation of iodine at redox boundaries and the apparent role of soil microbial activity in the sorption of selenium. The wider spatial and temporal relevance of these processes, however, can only be judged by observing the environmental distributions of these elements in biomes and over timescales which are of interest in the context of waste disposal assessments. A logical extension of the experiments reported here, therefore, might be the study of natural ecosystems across landscapes which provide sequences of vegetation and soil types which represent the evolution of biomes over suitable time scales. Work ¨ already carried out in Sweden (Oberg, 1998) and the UK (Bostock, 2004) has already demonstrated the value of understanding the natural biogeochemical cycles of chlorine and iodine. Much more could be learned about these elements and others which occur naturally,
ch09
May 3, 2007
14:44
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Conclusions
363
such as selenium, in ecosystems with relevance to radioactive waste safety studies. This approach would not be possible for Tc although a potentially suitable analogue element exists in the form of rhenium. This and other ‘rare’ elements can be measured with increasing reliability using modern mass spectrometry techniques which can also be used to make elegant measurements of chemical speciation of such elements. To consolidate our ability to make credible waste safety assessments it is evident that improved understanding of radionuclide behaviour at the scale of both the landscape and the molecule is desirable.
9.4.
Modelling
As discussed above, without the close integration of modelling and experimental research, the above analyses would not have been possible. The use of a simplified model of the lysimeter system was required to establish boundary conditions for the individual lysimeters, and provided an important means of quality assurance of the lysimeter flux data. More generally, the use of detailed physically-based models (SPW and SLT) has been required to represent the spatial detail of the experiments, to test hypotheses of response and to derive parameter values that have generic applicability. The BIOMOVS analysis of the lysimeter data, referred to in Chapter 3, provides a detailed assessment of the different performance of different model types, but it is evident that given the complexity of hydrological and geochemical response as one considers the upwards migration of solutes from a chemically-reduced groundwater system, across a water table and associated capillary fringe, and into the oxic unsaturated zone, a physically-based model with high spatial resolution is required to capture the process detail — at least for the present purposes of experimental analysis. In general, the model applications have been extremely successful in reproducing the experimental data, and deriving associated
ch09
May 3, 2007
14:44
364
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
parameter values for generic application. One important feature has been the relative consistency of parameter values obtained at the very different scales of the field lysimeters and the laboratory columns. The modelling has also successfully identified some unresolved issues in the experimental data — for example, in the case of Selenium, apparent fast flow pathways, possibly associated with colloidal transport. The research has pointed to some limitations in modelling capability: (a) Clearly there is a need to represent the dependence of geochemical solute/solid phase interactions on the ambient chemical environment. At present this has been achieved by empirical adjustment of Kd values; it would be much more powerful to be able to represent these dependencies explicitly within the model. (b) Similarly, the results of varying the water table elevation within the soil column pointed to the importance of hysteresis in the soil water characteristic relationships, and this has been represented by empirical adjustment. It would be helpful to represent hysteresis directly within the model (c) There is some evidence from related experiments of a dependence of sorption processes on soil moisture status. This is an area that requires fundamental research, but is of particular importance in the migration of solutes across a water table. We also acknowledge that the modelling analysis presented here has been explicitly deterministic — clearly there is considerable scope to extend the modelling here to formally recognise the uncertainty in both flow and transport properties, and further work is currently underway at Imperial to achieve this. We conclude that the models used here demonstrably provide a powerful capability for use in risk assessment and waste management, not only in the context of nuclear waste. We also conclude that the results of this programme illustrate the gross assumptions of some current assessment methods, e.g. soilplant transfer factors, and help to explain the large range of published values of transfer factors.
ch09
May 3, 2007
14:44
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Conclusions
9.5.
365
Conclusions and Future Directions
The research reported here has wide relevance to the scientific understanding of solute transport and plant uptake processes, and to a range of associated environmental management issues. The experimental results are of course imperfect, and we have taken pains to record the difficulties as well as the successes. Certain aspects of the experimental results remain unexplained, but we believe that the results reported here provide both important scientific insights for risk assessment of nuclear waste management, the modelling tools and numerical parameter values required to guide that assessment, and the potential for wider application. Over the course of the 17 year long programme of Nirexfunded research at Imperial, the experimental and modelling methods reported here have been developed and, taken together, provide a powerful research tool to understand and characterise the complex interactions involved. This methodology clearly has potential for wide application. The results here focus on a limited range of radionuclides, prioritised according to their importance for current waste management issues. There is obvious potential for direct extension to other radionuclides, soils, vegetation and environmental conditions. More fundamentally, the research has identified a number of open research questions concerning transport pathways and processes, and illustrated the strong dependence that can arise between hydrological conditions, microbial and chemical processes and solute-solid interactions. For example, there is a clear need for research on sorption processes under unsaturated conditions, and for better understanding of the redox-dependence of radionuclide speciation and mobility. The experimental procedures reported here can readily be applied to address these research questions. An additional dimension of experimental complexity is encountered when trying to consider the effects of microorganisms on radionuclide sorption and transport in both saturated and unsaturated soils. The Phase VIII experiment has scratched the surface of this issue using a crude but apparently effective means of demonstrating the role of gross microbial activity on
ch09
May 3, 2007
14:44
366
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
the sorption of 75 Se. This remains, however, a clear priority for future research. The research reported here has been funded mainly by UK Nirex to support safety assessment of the deep disposal of nuclear waste. A continuing programme of work has been identified to strengthen the linkage between this research and assessmentprocedures. We intend, for example: (a) to demonstrate the credibility of the models and parameter values by long term simulation of the effects of atmospheric deposition within a framework of uncertainty analysis, (b) to consider the necessary linkage between the coarse temporal resolution climate scenarios commonly available to support assessment and the fine-time-scale modelling reported here (through downscaling), (c) to consider the issues of upscaling experimental results to the scale required for current safety assessment modelling tools, and (d) to use the models to define future experimental priorities. As noted in 9.4, there is also a need to develop modelling capability with respect to dynamic geochemical controls on mobility. However, several other extensions of modelling capability are desirable. These include a dynamic plant capability, given the strong dependence of root uptake proceeses on plant vitality and stage of development. soil bioturbation, which has not as yet been represented in our models, and for the long time scales of nuclear waste assessment, there are issues of soil physical and chemical development to consider. In conclusion, the results reported here represent the outcomes from a long road of research, with successes, some failures and hopefully overall a significant contribution to knowledge and understanding of solute transport in vegetated soils. The authors hope that these results will provide the reader with insight and stimulation to pick up the challenge of research into these difficult and important scientific questions.
ch09
May 3, 2007
15:7
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
COLOUR SECTION
367
color
May 3, 2007
15:7
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
This page intentionally left blank
368
color
May 3, 2007 15:7
369
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Colour Section Plate 1. View of Phase I lysimeter experiment with winter wheat crop. Automatic weather station and water level control system also shown.
color
May 3, 2007
15:7
370
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Phase I lysimeter experiment with mature winter wheat crop. Plate 2.
color
May 3, 2007 15:7
371
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Colour Section Plate 3. Phase I lysimeter experiment. Dr Penny Wadey and (late) John Ions taking leaf area index measurements of young winter wheat crop.
color
May 3, 2007
15:7
372
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Plate 4. Phase I lysimeter experiment. Detail of buffer tank water level control system and side entry tensiometers.
color
May 3, 2007
15:7
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Colour Section
373
Plate 5. Phase I lysimeter experiment showing common reservoir and peristaltic pumps.
color
May 3, 2007
15:7
374
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Plate 6. Detail of boroscope and access tube for root density measurements (Phase I experiment).
color
May 3, 2007
15:7
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Colour Section
Plate 7. crop.
375
Phase II experiment showing early development of perennial ryegrass
color
May 3, 2007
15:7
376
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Marriotte bottles
Buffer reservoir
Adjustable platform
Soil columns
Plate 8. Environmental growth room with soil columns and water table control systems (Phases VI & VII).
color
May 3, 2007
15:7
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Colour Section
Plate 9.
Phase VIII (75 Se) experiment, showing volatilisation trap.
377
color
May 3, 2007
15:7
378
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Plate 10. Destructive sampling of Phase III soil column, showing colour change due to anoxic (lower) and oxic (upper) condictions (interval between white lines in 5 cm).
color
May 3, 2007
15:0
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
APPENDIX
379
Appendix
May 3, 2007
15:0
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
This page intentionally left blank
380
Appendix
May 3, 2007
15:0
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Appendix: Data Archive
The experimental data reported here are available from the following URL: www3.imperial.ac.uk/ewre/research/nirex/data archive The archive is categorised according to experimental phase (see Chapter 2). Each phase has a main folder, which includes an overview file describing the structure of the associated data archive. Phases I and II involved field lysimeters and included highly detailed logged climate and soil water hydrology data (at a resolution of 0.01 day), along with periodic plant (approx 2 weeks) and radionuclide (variable, but typically related to dosing/draining the control system and crop harvests) measurements. Phases III to VIII took place in an environmental growth room under pseudo-steady-state conditions. Consequently, there are no climate data. Measurements of water influx, soil water status, redox status and radionuclide concentrations (associated with the water table control system and, for phases VII and VIII, soil water samples) were made manually at intervals of approximately two weeks. In addition, plant growth data were collected at a similar time interval. Detailed soil profile measurements of plant root distribution, gravimetric soil water content and radionuclide concentration were then made at the end of each column experiment through destructive sampling. Included in the data archive are associated reports and papers that have been prepared for each experimental phase.
381
Appendix
This page intentionally left blank
May 3, 2007
14:58
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
References
Abdelouas, A., Fattahi, M., Grambow, B., Vichot, L. and Gautier, E., 2002. Precipitation of technetium by subsurface sulphate-reducing bacteria. Radiochimica Acta, 90, 773–777. Albrecht, A., Schultze, U., Bello Bugallo, P., Wydler, H., Frossard, E. and Fl¨ uhler, H., 2003. Behavior of a surface applied radionuclide and a dye tracer in structured and repacked soil monoliths. Journal of Environmental Radioactivity, 68(1), 47–64. Ames, L. L. and Rai, D., 1978. Radionuclide interactions with soil and rock media, Vol. 1 and 2. Las Vegas, NV: US EPA, Office of Radiation Programs; EPA 520/6-78-007-b. Appelo, C. A. J. and Postma, D., 1993. Geochemistry, Groundwater and Pollution, Balkema, Rotterdam, 536pp. Apslund, G. and Grimvall, A., 1991. Organohalogens in nature, more widespread than previously assumed. Environmental Science and Technology, 25, 1346–1350. Argonne National Laboratory, Illinois, USA website: http://www.ead. anl.gov/pub/doc/selenium.pdf (Accessed April 2004). Arthur, W. J. III and Markham, O. D., 1983. Small mammal soil burrowing as a radionuclide transport vector at a radioactive waste disposal area in southeastern Idaho. Journal of Environmental Quality, 12(1), 117–122. Ashworth, D., Ciciani, L., Butler, A. P. and Shaw, G., 2002. Radiochemical Analysis and Modelling of the Phase V Soil Column Experiment. Imperial College London Report, Imperial/NRP 009. Avery, S. V., 1995. Caesium accumulation by microorganisms: uptake mechanisms, cation competition, compartmentalization and toxicity. Journal of Industrial Microbiology, 14, 76–84. Bachhuber, H., Bunzl, K. and Schimmack, W., 1986. Spatial variability of the distribution coefficients of Cs-137, Zn-65, Sr-85, Co-57, Cd-109, Ce141, Ru-103, Tc-95m and I-131 in a cultivated soil. Nuclear Technology, 72, 359–371. 383
Reference
May 3, 2007
14:58
384
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Barber, S. A., 1984. Soil Nutrient Bioavailability: A Mechanistic Approach. John Wiley & Sons, New York, 398pp. Beasley, T. M. and Lorz, H. V., 1986. A review of the biological and geochemical behaviour of technetium in the marine environment. Journal of Environmental Radioactivity, 3(1), 1–22. Bell, J. N. B. and Shaw, G., 2005. Ecological lessons from the Chernobyl accident. Environment International, 31, 771–777. Bird, G. A. and Schwartz, W., 1996. Distribution coefficients, Kd s, for iodine in Canadian shield lake sediments under oxic and anoxic conditions. Journal of Environmental Radioactivity, 35, 261–279. Bird, G. A. and Schwartz, W. J., 1997. Partition coefficients (Kd values) for Tc in shield lake sediments under oxidizing and reducing conditions. Water Research, 31, 1673–1678. Bors, J. and Martens, R., 1992. Microbial effects on radioiodine behaviour in soils: I. Studies on the role of microbial biomass on the sorption of radiodine in soils. Microbial Aspects in Radioecological Research, Workshop, Proceedings of the International Union of Radioecologists, 6.–10.10.1993, Udine, Italy. Bors, J., Ersten, H. and Martens, R., 1991. Sorption studies of radioiodine on soils with special references to soil microbial biomass. Radiochemica Acta, 52–3. Bors, J., Martens, R. and Kuhn, W., 1988. Studies on the role of natural and anthropogenic organic substances in the mobility of radio-iodine in soils. Radiochimica Acta, 44/45, 201–206. Bostock, A. C., 2004. Chemical Speciation, Volatilisation and Cycling of Cl-36, I-129 and Tc-99 in Coniferous Systems. PhD Thesis, University of London (Imperial College). Brady, N. C. and Weil, R. R., 1996. The Nature and Properties of Soils, 11th edition Prentice Hall, USA. Budyko, M. I., 1974. Climate and Life, Academic Press, New York, 508pp. Bulman, R. A. and Cooper, J. R., 1986. Speciation of Fission and Activation Products in the Environment. Elsevier Applied Science Publishers, London and New York. Bunzl, K. and Schimmack, W., 1988. Effect of microbial biomass reduction by gamma irradiation on the sorption of Cs-137, Sr-85, Ce-139, Co-57, Cd-109, Zn-65, Ru-103, Tc-95m and I-131 by soils. Radiation and Environmental Biophysics, 27(2), 165–176. Burne, S., Wheater, H. S., Butler, A. P., Johnston, P. M., Wadey, P., Shaw, G. and Bell, J. N. B., 1994. Radionuclide transport above a nearsurface water table. I An automated lysimeter facility for near-surface contaminant transport studies. Journal of Environmental Quality, 23, 1318–1329.
Reference
May 3, 2007
14:58
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
References
385
Butler, A. P. and Wheater, H. S., 1999a. Modelling radionuclide transport in an integrated lysimeter experiment (I): Model development, Journal of Environmental Quality, 28, 1938–1946. Butler, A. P. and Wheater, H. S. 1999b. Modelling radionuclide transport in an integrated lysimeter experiment (I): Application to 22 Na. Journal of Environmental Quality, 28, 946–1956. Butler, A. P., Chen, J., Aguero, A., Edlund, O., Elert, M., Kirchner, G., Raskob, W. and Sheppard, M., 1998. Performance assessment studies of models for water flow and radionuclide transport in vegetated soils using lysimeter data. Journal of Environmental Radioactivity, 42(2–3), 271–288. Butler, A. P., Wadey, P., Shaw, G. and Ciciani, L., 2001. Radiochemical Analysis and Modelling of the Phase IV Soil Column. Imperial College London Report, ICSTM/NRP 003. Cataldo, D. A., Garland, T. R., Wildung, R. E. and Fellows, R. J., 1989. Comparative metabolic behaviour and interrelationships of Tc and S in soybean plants. Health Physics, 57, 281–288. Chapman, N. A. and McCombie, C., 2003. Principles and Standards for the Disposal of Long-lived Radioactive Wastes. Waste Management Series, Pergamon. Chu, S. Y. F., Ekstr¨ om, L. P. and Firestone, R. B., 1999. The Lund/LBNL Nuclear Data Search, Version 2.0. http://nucleardata.nuclear.lu.se/ NuclearData/toi/index.asp. Coughtrey, P. J., Jackson, D. and Thorne, M. C., 1983. Radionuclide Distribution and Transport in Terrestrial and Aquatic Ecosystems: A critical Review of Data, Vol. 3, A. A. Balkema, Rotterdam. Crossland, I. G., Pinedo, P., Kessler, J. H., Torres-Vidal, C. and Walters, B., 2005. “Reference Biospheres” for solid radioactive waste disposal: the BIOMASS Methodology. Journal of Environmental Radioactivity (in press). Cui, D. and Eriksen, T. E., 1996. Reduction of pertechnetate by ferrous iron in solution: influence of sorbed and precipitated Fe(II). Environmental Science and Technology, 30(7), 2259–2262. Curry, J. P., Byrne, D. and Boyle, K. E., 1995. The earthworm population of a winter cereal field and its effects on soil and nitrogen turnover. Biology and Fertility of Soils, 19, 166–172. Denys, S., Echevarria, G., Florentin, L., Leclerc-Cessac, E. and Morel, J. L., 2003. Availability of Tc-99 in undisturbed soil cores. Journal of Environmental Radioactivity, 70, 115–126. Desmet, G., Nassimbeni, P. and Belli, M. (eds), 1990. Transfer of Radionuclides in Natural and Semi-Natural Environments. Elsevier Applied Science, London.
Reference
May 3, 2007
14:58
386
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Dhillon, K. S. and Dhillon, S. K., 1999. Adsorption-desorption reactions of selenium in some soils of India. Geoderma, 93, 19–31. Dudal, R. and Tavernier, R., 1985. Classification and definitions of the soil and map units. Chapter 2 in Soil Map of the European Communities (1:1,000,000), CEC, Office for Official Publications of the European Community, Luxembourg. Echevarria, G., Morel, J. L., Florentin, L. and Leclerc-Cessac, E., 2003. Influence of climatic conditions and soil type on 99 TcO4− uptake by ryegrass. Journal of Environmental Radioactivity, 70, 85–97. Echevarria, G., Vong, P. C., Leclerc-Cessac, E. and Morel, J. L., 1997. Bioavailability of technetium-99 as affected by plant species and growth, application form, and soil incubation Journal of Environmental Quality, 26(4), 947–956. Echevarria, G., Vong, P. C. and Morel, J. L., 1998. Effect of NO3− on the fate of 99 TcO4− in the soil-plant system. Journal of Environmental Radioactivity, 38, 163–171. Etherington, J. R., 1962. The Growth of Alopecurus pratensis L and Agrostis tenuis Sibth in Relation to Soil Moisture Conditions. PhD thesis, University of London. Etherington, J. R. and Rutter, A. J., 1964. Soil water and the growth of grasses — the interaction of water table depth and irrigation amount on the growth of Agrostis tenuis and Alopecurus pratensis. Journal of Ecology, 55, 677–689. Elert, M., Butler, A. P., Chen, J., Dovlete, C., Konoplev, A., Golubenkov, A., Sheppard, M., Togawa, O. and Zeevaert, T., 1998. Effects of model complexity on uncertainty estimates. Journal of Environmental Radioactivity, 42(2–3), 255–270. Epstein, E., 1966. Dual patterns of ion adsorption by plant cells and by plants. Nature. 212, 1324–1327. Feddes, R. A., Kabat, P., Van Bakel, P. J. T., Bronswijk, J. J. B. and Halbertsma, J., 1988. Modelling soil water dynamics in the unsaturated zone — State of the art, Journal of Hydrology, 100, 69–111. Feyen, J., Jacques, D., Timmerman, A. and Vanderborght, J., 1998, Modelling water flow and solute transport in heterogeneous soils: A review of recent approaches. Journal of Agricultural Engineering Resources, 70, 231–256. Fio, J. L., Fujii, R. and Deverel, S. J., 1991. Selenium mobility and distribution in irrigated and nonirrigated alluvial soils. Soil Science Society of America Journal, 55, 1313–1320. Frissel, M. J., 1992. An update of the recommended soil-to-plant transfer factors of Sr-90, Cs-137 and transuranics. VIIIth Report of the Working Group Soil-to-Plant Transfer Factors, International Union of Radioecologists.
Reference
May 3, 2007
14:58
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
References
387
Fujikawa, Y. and Fukui, M., 1991. Analysis of radioactive caesium and cobalt adsorption to rocks using the two-site kinetic model. Journal of Contaminant Hydrology, 8, 43–69. Gerwitz, A. and Page E. R., 1974. An empirical mathematical model to describe plant root systems. Journal of Applied Ecology 11, 773–781. Grogan, H. A., Mitchell, N. G., Minski, M. J. and Bell, J. N. B., 1987. Pathways of radionuclides from soils to wheat, in: Coughtrey, P.J., Martin, N.H., Unsworth, M.H. (eds.), Pollutant Transport and Fate in Ecosystems. Special publication N6 of the British Ecological Society. Haverkamp, R. and Vauclin, M., 1981. A comparative study of three forms of the Richard’s equation used for predicting one-dimensional infiltration in unsaturated soils. Soil Science Society of America Journal, 45, 13–20. Hindmarsh, A. C., 1983. ODEPACK A systemized collection of ODE solvers, Stepleman, R. S., et al. (eds.), Scientific Computing. North Holland, Amsterdam, pp. 55–64. I.A.E.A., 1994. Handbook of Parameter Values for the Prediction of Radionuclide Transfer in Temperate Environments. Technical Report Series No. 364. Vienna: International Atomic Energy Agency. ICRP Publication 38, 1983. Radionuclide transformations: Energy and intensity of emissions, Annals of the ICRP, 11–13. International Union of Radioecologists (IUR), 1989 VIth Report of the Working Group on Soil-to Plant Transfer Factors. Grimselpass, Switzerland. IUR, Wageningen, The Netherlands. Jackson, B. M., 2007. Modelling water and solute transport within vegetated soils using a stochastic framework. PhD Thesis, University of London, 385pp. Johnsson, L., 1991. Selenium uptake by plants as a function of soil type, organic matter content and pH. Plant and Soil, 133, 57–64. Karavokyris, I., Butler, A. P. and Wheater, H. S., 1990. The Development and Validation of a Coupled Soil-Plant-Water Model (SPW1), Nirex Safety Series Report, NSS/R225, available from UK Nirex Ltd. Kloke, A., Sauerbeck, D. R. and Vetter, H., 1984. The contamination of plants and soils with heavy metals and the transport of metals in terrestrial food chains, Nriagu, J.O. (ed.), Changing Metal Cycles and Human Health. Dahlem Konferenzen, pp. 113–141. Lawlor, D. W. and Lake, J. V., 1975. Evaporation rate, leaf water potential and stomatal conductance in Lolium, Trifolium and Lysimachia in drying soil. Journal of Ecology, 13, 639–646. Leonard, K. S., McCubbin, D., Brown, J., Bonfiel, R. and Brooks, T., 1997. Distribution of technetium-99 in UK coastal waters. Marine Pollution Bulletin, 34, 628–636.
Reference
May 3, 2007
14:58
388
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Lee, R. T., 1997. 36 Cl Transport and Fate in Soils. PhD thesis, University of London, 117pp. Lee, R. T., Shaw, G., Wadey, P. and Wang, X., 2001. Specific association of 36 Cl with low molecular weight fractions of humic acids. Chemosphere, 43(8), 1063–1070. Lipiec, J., Arvidsson, J. and Murer, E., 2003. Review of modelling crop growth, movement of water and chemicals in relation to topsoil and subsoil compaction. Soil & Tillage Research 73, 15–29. Loague, K. and Green, R. E., 1991. Statistical and graphical methods for evaluating solute transport models: Overview and application. Journal of Contaminant Hydrology, 7(1–2), 51–73. Mualem, Y. (1976), A new model of predicting hydraulic conductivity of unsaturated porous media. Water Resources Research, 12, 513–522. Mercier, F., Moulin, V., Guittet, M. J., Barre, N., Toulhoat, N., GautierSoyer, M. and Toulhoat, P., 2000. Applications of NAA, PIXE and XPS for the quantification and characterization of the humic substances/iodine association. Radiochimica Acta, 88, 779–785. Mitchell, N. G., 1992. The Influence of Soil Type and Climate on the Uptake of Radionuclides into Winter Wheat. PhD thesis, University of London. Morris, K., Butterworth, J. C. and Livens, F. R., 2000. Evidence for the remobilisation of Sellafield waste radionuclides in an intertidal salt marsh, West Cumbria, UK. Estuarine, Coastal and Marine Science, 51, 613–625. Muller-Lemans, H. and van Dorp, F., 1996. Bioturbation as a mechanism for radionuclide transport in soil: relevance of earthworms. Journal of Environmental Radioactivity, 31, 7–20. Munro, W. R. C., 1991. The Time-Dependent Transfer of Radiocaesium Through Aerial and Soil Pathways to Crops. PhD thesis, University of London. Muramatsu, Y., Uchida, S., Sumiya, M., Ohmono, Y. and Obata, H., 1989. Tracer experiments on transfer of radioiodine in the soil-rice plant system. Water Air and Soil Pollution, 45. Muramatsu, Y. and Yoshida, S., 1995. Volatilization of methyl iodide from the soil-plant system. Atmospheric Environment, 29, 21–25. Muramatsu, Y. and Yoshida, S., 1999. Effects of microorganisms on the fate of iodine in the soil environment. Geomicrobiology Journal, 16(1). Muramatsu, Y., Yoshida, S., Uchida, S. and Hasebe, A., 1996. Iodine desorption from rice paddy soil. Water, Air and Soil Pollution, 86, 359–371. Murphy, C. E. and Johnson, T. L., 1993. Vegetative uptake of Tc-99 from buried, solidified, low-level radioactive-waste. Journal of Environmental Quality, 22, 793–799.
Reference
May 3, 2007
14:58
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
References
389
NAGRA, 2002. Project Opalinus Clay: Safety Report. Demonstration of disposal feasibility for spent fuel, vitrified high-level waste and long-lived intermediate-level waste (Entsorgungnackweis). Technical Report 02-05. NAGRA (National Cooperative for the Disposal of Radioactive Waste), Wettingen, Switzerland. Neal, R. H., 1995. Selenium, Alloway, B. J. (ed.), Heavy Metals in Soils, 2nd Edition. London, Blackie Academic. Nirex, 1997. An Assessment of the Post-Closure Performance of a Deep Waste Repository at Sellafield. Nirex Report No. S/97/012. Available from UK Nirex Ltd. Nirex, 2001. The Scientific Foundations of Deep Geological Disposal. Nirex Report No. N/016, pp. 118. Available from UK Nirex Ltd. Nirex, 2004. Briefing note on Selenium-79. Files TPA 80 612 and TPA 25 040. Available from UK Nirex. Noddack, W., Tacke, I. and Berg, O., 1925. Die Ekamangane. Die Naturwissenschaften, 13, 567. Nuclear Energy Agency, 1999. Progress Towards Geologic Disposal of Radioactive Waste: Where Do We Stand? An International Assessment. NEA, Paris. Nuclear Energy Agency, 2004. Safety of disposal of spent fuel, HLW and long-lived ILW in Switzerland — an international peer review of the postclosure radiological safety assessment for disposal in the Opalinus Clay of the Z¨ urcher Weinland. Radioactive Waste Management publication No. 5568, Nuclear Energy Agency of the OECD, Paris. Nwabuzor, S. S., 1984. Soil Water Dynamics under Field Crops, PhD thesis, University of London. Nye, P. H., 1966. The effect of nutrient intensity and buffering power of a soil, and the absorbing power, size and root hairs of a root, on nutrient absorption by diffusion. Plant Soil, 25, 81–105. Nye, P. H. and Tinker, P. B., 1977. Solute Movement in the Soil-Root System. Blackwell Scientific Publications. Oxford, pp. 324. ¨ Oberg, G., 1998. Chloride and organic chlorine in soil. Acta Hydrochemica et Hydrobiologica, 26, 137–144. Ogard, A. E., Thompson, J. L., Rundberg, R. S., Wolfsberg, K., Kubik, P. W., Elmore, D. and Bentley, H. W., 1988. Migration of chlorine-36 and tritium from an underground nuclear test. Radiochimica Acta, 44/45, 213–217. Pardue, J. H., DeLaune, R. D., Patrick Jr, W. H. and Whitcomb, J. H., 1991. Effect of redox potential on fixation of 137 Cs in lake sediment. Health Physics, 57, 781–789. Parkin, G., Ewen, J. and O’Connell, P. E., 2000. SHETRAN: a coupled surface/subsurface modelling system for 3D water flow and sediment
Reference
May 3, 2007
14:58
390
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
and solute transport in river basins. American Society of Civil Engineers Journal of Hydrologic Engineering, 5, 250–258. Patera, E. S., Hobart, D. E., Meijer, A. and Rundberg, R. S., 1990. Chemical and physical processes of radionuclide migration at Yucca Mountain, Nevada. Journal of Analytical and Nuclear Chemistry, 142, 331–347. Perez-Espinosa, A., Moral, R., Moreno-Caselles, J., Cort´es, A., PerezMurcia, M. D. and G´ omez, I., 2005. Co phytoavailability for tomato in amended calcareous soils. Bioresource Technology, 96(6), 649–655. Pignolet, L., Auvray, F., Fonsny, K., Capot, F. and Moureau, Z., 1989. Role of various microorganisms on Tc behaviour in sediments. Health Physics, 57(5), 791–800. Posiva, 1999. Safety Assessment of Spent Fuel Disposal in H¨astholmen, Kivetty, Olkiluoto and Romuvaara: TILA-99. Report Posiva 99-07. Posiva Oy, Helsinki, Finland. Qing, H., 1998. Migration and Plant Uptake of Radionuclides in Laboratory soil Columns and Lysimeters with Contaminated Water Tables. PhD thesis, University of London (Imperial College). Raja, H. E. and Babcock, K. L., 1961. On the soil chemistry of radioiodine. Soil Science, 91. Reineger, P., 1977. Transfer of Iodine in terrestrial ecosystems. Proc. NEA Specialist Meeting, 1977, Paris. ¨ Rodstedth, M., St˚ ahlberg, C., Sand´en, P. and Oberg, G., 2003. Chloride imbalances in soil lysimeters. Chemosphere, 52, 381–389. Rowell, D. L., 1994. Soil Science: Methods and Applications. Longman, Harrow. Saas A., Grauby A., Colle C., Denardi J. L., Nomura T. and Joyer Ph., 1982. Evolution physico-chimique du technetium-99 et du selenium-79 dans les sols. IAEA-SM-257/28. Savage, D. (Ed), 1995. The Scientific and Regulatory Basis for the Geological Disposal of Radioactive Waste. John Wiley and Sons Ltd. Savchenko, V. K., 1995. The ecology of the Chernobyl disaster. Man and the Biosphere Series. Vol. 16. UNESCO, Paris/The Parthenon Publishing Group, New York. Schaap, M. G. and Leij, F. J., 2000. Improved prediction of unsaturated hydraulic conductivity with the Mualem-van Genuchten model. Soil Science Society America Journal, 64, 843–851. Schmitz, D. and Aumann, D. C., 1995. A study of the association of two iodine isotopes of natural 127 I and of the fission product 129 I with soil components using a sequential extraction procedure. Journal of Radioanalytical and Nuclear Chemistry, Articles, 198(1), 229–236. Scott Russell, R. S., 1966. Radioactivity and Human Diet. Pergamon Press, Oxford.
Reference
May 3, 2007
14:58
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
References
391
Shaw, G. and Bell, J. N. B., 1989. The kinetics of caesium absorption by roots of winter wheat and the possible consequences for the derivation of soil-to-plant transfer factors for radiocaesium. Journal of Environmental Radioactivity, 10, 213–231. Shaw, G. and Bell, J. N. B., 2001. Significance of soil-to-plant transfers of radionuclides, Van der Stricht, E., Kirchmann, R. (eds), Radioecology: Radioactivity and Ecosystems. International Union of Radioecologists, Oupeye, pp. 112–135. Shaw, G., Bell, J. N. B., Minski, M. J. and Nair, S., 1990. Experimental Studies of Radionuclide Uptake and Retention by Crops under UK Field Conditions. Final report of a research contract jointly undertaken by Imperial College Reactor Centre and Berkeley Nuclear Laboratories, CEGB. Shaw, G., Wadey, P. and Bell, J. N. B., 2004. Radionuclide transport above a near-surface water table: IV. Soil migration and crop uptake of chlorine36 and technetium-99, 1990 to 1993. Journal of Environmental Quality, 33, 2272–2280. Sheppard, S. C. and Evenden, W. G., 1991. Can aquatic macrophytes mobilize technetium by oxidizing their rhizosphere? Journal of Environmental Quality, 20, 738–744. Sheppard, S. C. and Evenden, W.G., 1997. Variation in transfer factors for stochastic models: soil to plant transfer. Health Physics, 72(5), 727–733. Sheppard, S. C. Evenden, W. G. and Amiro, B.D., 1993. Investigation of the soil-to-plant pathway for I, Br, Cl and F. Journal of Environmental Radioactivity, 21, 9–32. Sheppard, M. I. and Hawkins, J. L., 1995. Iodine and microbial interactions in an organic soil. Journal of Environmental Radioactivity, 29, 91–109. Sheppard, S. C., Johnson, L. H., Goodwin, B. W., Tait, J. C., Wuschke, D. M. and Davison, C. C., 1996. Chlorine-36 in nuclear waste disposal. 1. Assesment results for used fuel with comparison to I-129 and C-14. Waste Management, 16, 607–614. Sheppard, S. C. and Motycka, M., 1997. Is the Akagare phenomenon important to iodine uptake by wild rice (Zizania aquatica)? Journal of Environmental Radioactivity, 37, 339–353. Sheppard, M. I. and Sheppard, S. C., 1986. Technetium behaviour in soils of the Canadian Precambrian Shield, in: Desmet, G., Myttenaere, C. (eds.). Technetium in the Environment, pp. 131–141. Elsevier Applied Science Publishers, London. Sheppard, S. C., Sheppard, M. I. and Evenden, W. G., 1990. A novel method used to examine variation in Tc sorption among 34 soils. Journal of Environmental Radioactivity, 11, 215–233.
Reference
May 3, 2007
14:58
392
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Sheppard, M. I., Sheppard, S. C. and Sanipelli, B., 2002. Recommended Biosphere Model Values for Iodine. Ontario Power Generation, Nuclear Waste Management Division Report 06819-REP-01200-10090R00, Toronto, Canada. Sheppard, M. I. and Thibault, D. H., 1990. Default solid/liquid partition coefficients, Kds, for four major soil types: a compendium. Health Physics, 59(4), 471–482. Sheppard, M. I. and Thibault, D. H., 1992. Chemical behaviour of iodine in organic and mineral soils. Applied Geochemistry, 7, 265–271. Sheppard, M. I., Thibault, D. H., McMurry, J. and Smith, P. A., 1995. Factors affecting the soil sorption of iodine. Water, Air and Soil Pollution, 83, 51–67. Shuttleworth, W. J., 1979. Evaporation, Institute of Hydrology Report No. 56, Wallingford. Singh, M., Singh, N. and Relan, P. S., 1981. Adsorption and desorption of selenite and selenate selenium on different soils. Soil Science, 132, 134–141. ˇ Simunek, J., Jarvis, N. J., van Genuchten, M. Th. and G¨ arden¨ as, A., 2003. Review and comparison of models for describing non-equilibrium and preferential flow and transport in the vadose zone. Journal of Hydrology, 272, 14–35. SKB, 1998. Project SAFE. Update of the SFR-1 safety assessment — Phase 1. Technical Report R-98-43. Swedish Nuclear Fuel and Waste Management Co., Stockholm. http://www.skb.se/templates/ skbiframe 9413.aspx SKB, 1999a. Deep Repository for Spent Nuclear Fuel: SR 97-Post-closure Safety. Main Report (3 vols — Summary plus Volumes I and II) Technical Report TR-99-06. Swedish Nuclear Fuel and Waste Management Co., Stockholm. http://www.skb.se/templates/skbiframe 9413.aspx SKB, 1999b. Deep repository for long-lived low- and intermediatelevel waste: Preliminary safety assessment. Technical Report TR-9928. Swedish Nuclear Fuel and Waste Management Co., Stockholm. http://www.skb.se/templates/skbiframe 9413.aspx Smith, I. C. and Carson, B. L., 1981. Trace Metals in the Environment, Vol 6, Cobalt. Ann Arbor Science Publication Inc, Ann Arbor, MI. Sposito, G., 1986. The “physics” of soil water physics. Water Resources Research, 22(9), 83S–88S. Sposito, G., 1989. The Chemistry of Soils. Oxford University Press, New York. Stalmans, M., Maes, A. and Cremers, A., 1986. Role of organic matter as a geochemical sink for technetium in soils and sediments, Desmet, G. and
Reference
May 3, 2007
14:58
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
References
393
Myttenaere, C. (eds.). Technetium in the Environment. Elsevier Applied Science Publishers, London, pp. 91–113. Szabo, G., Angelova, A., Guczi, J. and Bulman, R. A., 1993. An examination of the uptake of radioiodide by chemically bound humic acid and by some solid phases of soil. Science of the Total Environment, 130–31, 375–382. Tagami, K. and Uchida, S., 1996. Analysis of technetium-99 in soil and deposition samples by inductively coupled plasma mass spectrometry. Applied Radiation and Isotopes, 47, 1057–1060. Tagami, K. and Uchida, S., 1998. Aging effect on bioavailability of Mn, Co, Zn and Tc in Japanese agricultural soils under waterlogged conditions. Geoderma, 84, 3–13. Tagami, K. and Uchida, S., 1999. Chemical transformation of technetium in soil during the change of soil water conditions. Chemosphere, 38, 963–971. Thiessen, K. M., Thorne, M. C., Maul, P. R., Pr¨ ohl, G. and Wheater, H. S., 1999. Modelling radionuclide distribution as transport in the environment. Environmental Pollution, 100, 151–177. Thomas, G. W. and Swoboda, A. R., 1970. Anion movement in soils. Soil Science 110, 163–166. Thomson, J., Higgs, N. C., Wilson, T. R. S., Croudace, I. W., de Lange, G. J. and van Santvoort, P. J. M., 1995. Redistribution and geochemical behaviour of redox-sensitive elements around S1, the most recent eastern Mediterranean sapropel. Geochimica et Cosmochimica Acta, 59, 3487–3501. Thompson, N., Barrie, I. A. and Ayles, M., 1981. The Meteorological Office Rainfall and Evaporation Calculation System: MORECS (July 1981). Hydrological Memorandum No. 45. Meteorological Office, Bracknell, Berkshire, pp. 69. Thorne, M. C., Kelly, M., Rees, J. H. and Sanchez-Friera, P., 2001. Review of Kd and Transfer Factors for Selenium Relevant to a Temperate Biosphere and the Site-Specificity of the Meuse/Haute Marne Underground Laboratory. Andra Report ID: C RP 0AET 01-04. Till, J. E., 1986. Source terms for technetium-99 from nuclear fuel cycle facilities. pp. 1–20 in ‘Technetium in the Environment’, Eds. G. Desmet and C. Myttenaere, Proceedings of a CEC seminar, Elsevier Applied Science Publishers, London and New York. Tompkins, J. A. and Butler, A. P., 1996a. Documentation for the Imperial College Coupled Soil-Plant-Water Model (SPW1v5). Imperial College Internal Report: IC/ MAN/MDL/SPW/1.5/9607.
Reference
May 3, 2007
14:58
394
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Tompkins, J. A. and Butler, A. P., 1996b. Documentation for the Imperial College 1-D Finite Difference Unsaturated Soil Solute Transport Model (SLT1v3). Imperial College internal report: IC/MAN/MDL/SLT/ 1.3/9608. Topp, G. C. and Davis, J. L., 1980. Measurement of soil water using TDR: a field evaluation. Soil Science Society of America Journal, 49, 19–24. Torok, J., Buckley, L. P. and Woods, B. L., 1990. The separation of radionuclide migration by solution and particle transport in soil. Journal of Contaminant Hydrology, 6, 185–203. Tikhomirov, F. A., Kasparov, S. V., Prister, B. S. and Sal’nikov, V. G., 1980. Role of organic matter in iodine fixation in soils. Soviet Soil Science, 12(1), 64–72. Umstadter, D. and Baciak, J., 2002. Pro Yucca Mountain Final Paper. NERS 211, Winter 2002. pp. 21. van Genuchten, M., 1980. Predicting the hydraulic conductivity of unsaturated soils. Soil Science Society of America Journal, 44, 892–897. van Genuchten, M. Th. and Wierenga, P. J., 1976. Mass transfer studies in sorbing porous media: 1. Analytical solutions. Soil Science Society of America Journal, 40, 473–480. van Loon, L., 1986. Kinetic aspects of the soil-plant transfer of technetium. Doctoraatproefschrift no. 150. Fakulteit der Landbouwetenschappen van de Katholieke Universeit Leuven, Belgium. Wadey, P., Shaw, G. and Bell, J. N. B., 2001. Radionuclide transport above a near-surface water table: III. Soil migration and crop uptake of three gamma-emitting radionuclides, 1990 to 1993. Journal of Environmental Quality, 30, 1341–1353. Wadey, P., Shaw, G., Bell, J. N. B. and Minski, M. J., 1994. Radionuclide transport above a near-surface water table: II. vertical distribution of gamma activities within soil profiles in relation to wheat rooting density and soil-to-plant transfers. Journal of Environmental Quality, 23, 1330–1337. Wadey, P., Butler, A. P., Tompkins, J. A. and Shaw, G., 2001. Radiochemical Analysis and Modelling of the Phase III Soil Column Experiment. Imperial College London Report, ICSTM/NRP 004. Warner, F. and Harrison, R. M. (eds.), 1993. Radioecology after Chernobyl. SCOPE 50. John Wiley & Sons, Chichester. White, P. J. and Broadley, M. R., 2000. Mechanisms of caesium uptake by plants. New Phytologist, 147, 241–256. Whitehead, D. C., 1973. Studies on iodine in British soils. Journal of Soil Science, 24. Whitehead, D. C., 1974. The sorption of iodide by soil components. Journal of Science, Food and Agriculture, 25, 73–79.
Reference
May 3, 2007
14:58
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
References
395
Whitehead, D. C., 1981. The volatilisation, from soils and mixtures of soil components, of iodine as potassium iodide. Journal of Soil Science, 32. Wildung, R. E., Desmet, G. M., Cataldo, D. A. and Weiss, S. G., 1989. The behaviour of technetium in terrestrial and aquatic environs: a symposium. Health Physics, 57, 235–351. Wildung, R. E., Garland, T. R. and Cataldo, D., 1977. Accumulation of technetium by plants. Health Physics, 32, 314–317. Wildung, R. E., Garland, T. R., McFadden, K. M. and Cowan, C. E., 1986. Technetium sorption in surface soils. pp. 115–129, ‘Technetium in the Environment’, G. Desmet and C. Myttenaere, (eds.). Proceedings of a CEC Seminar, Elsevier Applied Science Publishers, London. Yanagisawa, K. and Muramatsu, Y., 1995. Transfer of technetium from soil to paddy and upland rice. Journal of Radiation Research, 36, 171–178. Yu, K. N., Cheung, T., Luo, D. L. and Li, M. F., 2000. Factors affecting uptake of 131 I in Chinese white cabbage (Brassica Chinensis linn). Journal of Environmental Radioactivity, 49, 209–216. Zawislanski, P. T, Benson, S. M., Terberg, R. and Borglin, S. E., 2003. Selenium, speciation, solubility and mobility in land-disposed degraded sediments. Environmental Science and Technology, 37, 2415–2420.
Reference
May 3, 2007
14:58
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
This page intentionally left blank
Reference
May 3, 2007
14:57
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Index
activity balances, 175 activity concentrations, 25, 154 actual evaporation, 71 adsorption, 296 adsorption isotherm, 194 advection, 6 advection-dispersion equation, 58 agrochemical, 15 air temperature, 21, 60 air-soil-plant systems, 4 analogue element, 363 Andra, 8 anion exchange, 27 anoxic conditions, 190 anoxic zone, 256 atmosphere exchanges, 13 atmospheric deposition, 6 atmospheric humidity, 60 automated solid scintillation, 34 automatic variable-order, 59 automatic weather station, 21, 60
Kd , 197 106 Ru, 13, 14 125 I, 8 125 I accumulation, 201 125 Sb, 13 127 I, 188 129 I, 188 131 I, 188 133 Cs, 300 134 Cs, 8, 14, 300 137 Cs, 8, 13, 14, 22, 25, 300 144 Ce, 13 152 Eu, 25 22 Na, 8, 22, 25, 300 23 Na, 300 241 Am, 25 36 Cl, 8, 22, 25 36 Cl migration, 153 54 Mn, 14 59 Co, 300 60 Co, 8, 14, 22, 25, 300, 301 75 Se, 8 90 Sr, 14 95m Tc, 248 99m Tc, 8 99 Tc, 22, 25 Astragalus, 268 Calluna peat, 270, 295 in situ instrumentation, 19
batch sorption experiment, 192 batch sorption studies, 192 bead substrate, 208 bioaccumulation, 352 biological, 345 biological uptake, 22 biomass, 142 BIOMASS project, 10 biomass yields, 297 biosphere, 5
accidental releases, 6 actinides, 8 397
index
May 3, 2007
14:57
398
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Borescope, 20, 52 boundary, 47 boundary condition, 59, 84 boxes, 58 British, 8 british nuclear group, 4 bromine, 189 buffer tank, 19 bulk soil, 13 burrowing animals, 10 Ca-exchangeable fraction, 301 caesium, 300 canopy, 60 canopy temperature, 60 capillary fringe, 223 capillary suction, 54, 65 catchment-scale models, 58 cation exchange, 26 cation exchange sites, 301 centrifuged, 26 chaff, 24 chemical containment, 6 chemical properties, 15 chlorine, 80, 189 climatic data, 21, 60 Co(II), 301 Co(II) salts, 301 Co(III), 301 Co(NO3 )2 , 301 cobalt, 300 CoCl2 , 301 colloidal, 35 column experiments, 7 common reservoir, 14 complexation, 302 contaminant transport model, 58 contaminated areas, 4 controlled environment cabinet, 7 crop leaf area index, 21 crop root distribution and density, 21 crops, 10 Darcy’s law, 59 data logger, 20, 21 dead-end pores, 347
decay constant, 208 decay corrected, 193 deep repository, 4 deionised water, 26 detection limits, 252 dielectric constant, 20 direct ingestion, 6 dispersive flux, 61 dispersivity coefficient, 208 dosing, 22 Dounreay, 4 drainage trench, 14, 17 Drigg, 4 dry bulk density, 39 earthworm, 10, 28 elemental iodine, I2 (0), 189 engineered repository, 6 England, 13 entrapment, 347 environment room, 204 evaporative losses, 39 evapotranspiration, 17, 22, 149, 361 evapotranspiration losses, 47 experimental research, 5 exponential density profile, 52 extractant, 248 extracting, 248 field, 41 Finland, 9 fission products, 8 fixation, 316 fixed water table, 170 fluctuating water table, 170, 193 flux boundary, 85 fluxes, 58 food chains, 3 Fordham Series, 30 freeze-drying, 26 French, 8 French soils, 8 fumigation, 295 Gamma-emitting, 26 Gamma-emitting radionuclide, 25
index
May 3, 2007
14:57
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Index
399
gamma-ray spectrometry, 25, 34 GeLi detector, 263 generic crop, 7 geochemical, 345 geochemical interactions, 6 geological disposal, 5 geosphere, 5, 9 geotextile, 17, 22 government policy, 4 grain, 24 grass, 10 gravimetric moisture contents, 37 groundwater flow, 5
leaching, 358 leaf area indices, 71 least squares objective function, 52 leaves, 24 linear equilibrium process, 48 linear regression correlations, 283 linearised equation, 63 liquid scintillation counting (LSC), 25 London Dumping Convention, 4 Longlands Farm, 17 low level waste, 4 LSC, 25 lysimeters, 7
halogens, 189 harvest, 23 head gradient, 59 high level waste, 4 hollow fibre solution samplers (HFSS), 176 hydraulic conductivity, 59 hydraulic flux divergence, 59 hydraulic properties, 59 hydrological, 345 hydrological models, 41 hydrological studies, 5 hysteresis, 66, 364 hysteretic, 66 hysteretic effect, 223
marine epoxy resin, 37 Marriot bottle, 208 Marriotte bottle, 33 matric potential, 20, 59, 85 mechanical dispersion, 61 methyl bromide, 269 methyl iodide, CH3 I (+1), 189 methylated selenium, 297 Michaelis rate constant, 51 Michaelis-Menten rate kinetics, 63 microbial population, 276 micrometeorology, 13 microorganisms, 365 migration, 6 mini-column, 8, 35, 272 mini-column experiment, 192 mobile fraction, 294 modelling, 5 modelling tools, 41 moisture characteristic curve, 65 moisture content, 20, 85, 356 molecular diffusion, 61 molecular diffusion parameter, 208 monocotyledon, 51
IAEA, 7 immobilisation, 295 intermediate level waste, 4 intermediate zone, 256 International Atomic Energy Agency, 10 interstitial water, 21 inventory, 4 iodate, IO− 3 (+5), 189 iodide, I− (−1), 189 iodine, 80, 188, 189 iodine sorption, 189 iodine transfer factors, 191 iron-bearing minerals, 301 laboratory, 41
NAGRA, 9 near-surface environment, 6 near-surface water table, 227 net radiation, 21, 60 neutron probe, 21 Nirex, 5 nitrate, 195
index
May 3, 2007
14:57
400
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
Nuclear Energy Agency, 9 nuclear installations, 6 nuclear weapons, 3 numerical modelling, 6 one-dimensional, 59 organic matter, 139, 356 oxic conditions, 190 oxic zone, 256 oxidation-reduction potential, 21 parameter identification, 65 parameter values, 364 partition coefficient, 62 pathways, 3 Penman, 60 Penman-Monteith, 60 perennial ryegrass, 7, 71 peristaltic pumps, 17 pH, 356 phase I, 14, 16 phase II, 7 phase III, 8 phase VIII, 8 physical containment, 6 physically-based numerical models, 41 plant canopy, 71 plant root density, 63 plant root system, 63 plant rooting zone, 315 plastic bead substrate, 32 platinum electrode (redox probe), 37 platinum electrodes, 20 platinum redox electrode, 32, 192 polythene bead substrate, 21 polythene beads, 17, 144 polyurethane foam, 31 pot, 8 pot experiments, 355 potential evaporation, 76 potentiometer, 19 precipitation, 19 pressure transducer, 19 rachis, 24 radiant energy, 13
radioactive waste management, 4 radioactive decay, 25 radioactive isotopes, 300 radiocaesium, 10, 14 radiochemical analysis, 21 radiochlorine, 10, 177 radiocobalt, 10 radioecology, 3 radioiodine, 10 radionuclide activity concentration, 39 radionuclide cocktail, 14 radionuclides, 3, 6 radioselenium, 10, 298 radiosodium, 10 rainfall, 21 rainwater, 22 rainwater mixing zone, 47 rate kinetics, 50 redox, 80 redox dependency, 256 redox potential, 35, 190, 356 redox status, 6 reduction-oxidation, 190 reference biospheres, 10 reference electrode, 39, 192 regulators, 10 relative humidity, 21 residual moisture content, 65 resistance network, 60 retard, 50 retardation, 325 rhenium, 363 Rhizon, 32 Richards’ equation, 58, 59 rigid endoscope, 20 Robertgate, 31 Robertgate soil, 30 root boundary, 50 root density, 52 root growth, 20 root mean squared error, 57 root uptake, 6 root uptake parameter, 209 root-profile distributions, 315 rooting zone, 186
index
May 3, 2007
14:57
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Index rye grass, 25 ryegrass biomass, 154 S.K.B., 8 safety assessment, 5 sandy loam, 15 saturated moisture content, 65 saturated region, 21 saturation, 65 savannah river, 10 scintillation, 26 scintillation cocktail, 34 seleno-amino acids, 268 Sellafield, 4 shallow and deep lysimeters, 243 SHETRAN, 58 Silwood Park, 13 simplex algorithm, 52 SLT1, 63 sodium, 300 sodium hydroxide, 139 soil activity concentration, 37 soil column experiments, 37, 192 soil contaminants, 13 soil matric potential, 21 soil microbial activity, 362 soil moisture content, 21, 59 soil moisture status, 364 soil monoliths, 13 soil organic matter, 139 soil temperature, 21 soil texture, 356 soil type, 8, 15 soil water characteristic relationships, 364 soil water deficits, 72 soil water potential, 58 soil-plant transfer, 10, 191 soil-plant transfer factor, 3, 6, 186, 191 soil-plant-atmosphere system, 6 soil-shoot transfer, 284 soil-sorption behaviour of iodine, 189 soil/plant/groundwater systems, 13 solar radiation, 21 solid phase concentration, 40
401 solid radioactive waste disposal, 10 solid-liquid distribution coefficient (Kd ), 35 solute movement, 6 solution phase concentration, 40 sorbed phases, 61 sorption, 22, 138 sorption coefficient, 178, 208 sowing, 22 spatial resolution, 363 SPW1, 59 stems, 24 stomatal heat, 70 stomatal vapour pressure, 60 sulphate, 195 sulphate reducing bacteria, 352 surface deposition, 6 Sweden, 8 Switzerland, 9 TDR, 20 technetium, 10, 230 temperature, 21 tensiometer, 19, 66 time domain reflectometry, 20 tortuosity factor, 69 tortuous, 62 translocation, 6 transmutating, 5 transpiration, 186 transpiration fluxes, 60 transport models, 10 transport pathway, 6 trees, 10 unsaturated soils, 84 unsaturated zone, 6 uptake, 6 upward migration, 9 upward movement, 10 USA, 8 van Genuchten relationship, 66 vapour conductances, 70 variable-timestep integration, 59 vegetation, 10
index
May 3, 2007
14:57
402
SPI-B441 Biosphere Implications of Deep Disposal of Nuclear Waste – 9 x 6in
Biosphere Implications of Deep Disposal of Nuclear Waste
volatilisation, 190, 285 volumetric moisture content, 40 water water water water water water water water
balance, 19 content, 21 influx, 159 level, 19 pressure, 58 resource engineering, 5 table, 170 table control system, 17, 47
water table elevation, 84 water vapour, 13 water-table, 7 waterlogged, 271 Wellesbourne, 138 wheat ear, 24 wind speed, 21 winter wheat, 7 Yucca Mountain, 8
index