Advances in Agronomy, Volume 103

V O LU M E O N E ADVANCES H U N D R E D IN AGRONOMY T H R E E ADVANCES IN AGRONOMY Advisory Board PAUL M. BERT...

Author: Donald L. Sparks

100 downloads 1443 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

V O LU M E

O N E

ADVANCES

H U N D R E D

IN

AGRONOMY

T H R E E

ADVANCES IN AGRONOMY Advisory Board

PAUL M. BERTSCH

RONALD L. PHILLIPS

University of Kentucky

University of Minnesota

KATE M. SCOW

LARRY P. WILDING

University of California, Davis

Texas A&M University

Emeritus Advisory Board Members

JOHN S. BOYER

KENNETH J. FREY

University of Delaware

Iowa State University

EUGENE J. KAMPRATH

MARTIN ALEXANDER

North Carolina State University

Cornell University

Prepared in cooperation with the American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America Book and Multimedia Publishing Committee DAVID D. BALTENSPERGER, CHAIR LISA K. AL-AMOODI

CRAIG A. ROBERTS

WARREN A. DICK

MARY C. SAVIN

HARI B. KRISHNAN

APRIL L. ULERY

SALLY D. LOGSDON

V O LU M E

O N E

ADVANCES

H U N D R E D

T H R E E

IN

AGRONOMY EDITED BY

DONALD L. SPARKS Department of Plant and Soil Sciences University of Delaware Newark, Delaware, USA

AMSTERDAM • BOSTON • HEIDELBERG • LONDON NEW YORK • OXFORD • PARIS • SAN DIEGO SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO Academic Press is an imprint of Elsevier

Academic Press is an imprint of Elsevier 525 B Street, Suite 1900, San Diego, CA 92101-4495, USA 30 Corporate Drive, Suite 400, Burlington, MA 01803, USA 32 Jamestown Road, London, NW1 7BY, UK Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands First edition 2009 Copyright # 2009 Elsevier Inc. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: [email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at http://elsevier.com/locate/permissions, and selecting Obtaining permission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made ISBN: 978-0-12-374819-5 ISSN: 0065-2113 (series) For information on all Academic Press publications visit our website at elsevierdirect.com Printed and bound in USA 09 10 11 12 10 9 8 7 6 5 4 3 2 1

CONTENTS

Contributors Preface

1. Clearing the Air: Livestock’s Contribution to Climate Change

vii ix

1

Maurice E. Pitesky, Kimberly R. Stackhouse, and Frank M. Mitloehner 1. Introduction 2. Life Cycle Assessment 3. Effects of Agriculture on Climate Change 4. Livestock Types and Production Systems 5. Enteric Fermentation 6. Animal Manure 7. Livestock Related Land-Use Changes 8. Livestock Induced Desertification 9. Release from Cultivated Soil 10. Carbon Emissions from Feed Production 11. On-Farm Fossil Fuel Use: Diesel and Electricity 12. Postharvest: CO2 from Livestock Processing 13. Conclusions Acknowledgment References

2. Improvement of Drought Resistance in Rice

3 8 9 11 15 18 20 23 24 26 29 30 33 35 36

41

R. Serraj, A. Kumar, K. L. McNally, I. Slamet-Loedin, R. Bruskiewich, R. Mauleon, J. Cairns, and R. J. Hijmans 1. Introduction 2. Drought Characterization 3. Rice Responses to Drought 4. Concepts and Tools for Phenotyping 5. Conventional Breeding 6. Marker-Assisted Selection 7. Drought-Resistance Genes and GM Technology 8. Conclusions and Future Prospects References

42 44 49 57 64 71 78 86 88

v

vi

Contents

3. Problems, Challenges, and Strategic Options of Grain Security in China

101

Huixiao Wang, Minghua Zhang, and Yan Cai 1. Introduction 2. Connotation of Grain Security in China 3. Development Stages of Grain Production in China 4. Achievements and Experiences of Grain Production in China 5. Problems and Challenges of Grain Security in China 6. Strategy Options and Countermeasures for Grain Security in China 7. Case Studies 8. Concluding Remarks Acknowledgment References

4. Weed Management in Rice-Based Cropping Systems in Africa

103 106 109 114 120 129 139 143 144 144

149

J. Rodenburg and D. E. Johnson 1. 2. 3. 4. 5.

Introduction Weed Species in Rice in Africa Weed Management Practices in African Rice-Based Cropping Systems Emerging Weed Problems and Weed Management Issues A Strategic Vision for Weed Management and Research in African Rice Production Systems 6. Concluding Remarks Acknowledgments References Index See color insert section at the end of Chapter 2

150 155 165 186 189 200 200 201 219

CONTRIBUTORS

Numbers in Parentheses indicate the pages on which the authors’ contributions begin

R. Bruskiewich (41) IRRI-CIMMYT Crop Research Informatics Laboratory, International Rice Research Institute (IRRI), Metro Manila, Philippines Yan Cai (101) Key Laboratory for Water and Sediment Sciences, Ministry of Education, College of Water Sciences, Beijing Normal University, Beijing, People’s Republic of China J. Cairnsk (41) Crop and Environmental Sciences Division, International Rice Research Institute (IRRI), Metro Manila, Philippines R. J. Hijmansk (41) Social Sciences Division, International Rice Research Institute (IRRI), Metro Manila, Philippines D. E. Johnson (149) Crop, Soil and Water Sciences Division, International Rice Research Institute (IRRI), Metro Manila, Philippines A. Kumar (41) Plant Breeding, Genetics, and Biotechnology Division, International Rice Research Institute (IRRI), Metro Manila, Philippines R. Mauleon (41) IRRI-CIMMYT Crop Research Informatics Laboratory, International Rice Research Institute (IRRI), Metro Manila, Philippines K. L. McNally (41) TTChang-Genetic Resources Center, International Rice Research Institute (IRRI), Metro Manila, Philippines Frank M. Mitloehner (1) Department of Animal Science, University of California, California, USA

k

Left IRRI in June, 2009

vii

viii

Contributors

Maurice E. Pitesky (1) School of Veterinary Medicine, University of California, California, USA J. Rodenburg (149) Africa Rice Center (WARDA), Dar es Salaam, Tanzania R. Serraj (41) Crop and Environmental Sciences Division, International Rice Research Institute (IRRI), Metro Manila, Philippines I. Slamet-Loedin (41) Plant Breeding, Genetics, and Biotechnology Division, International Rice Research Institute (IRRI), Metro Manila, Philippines Kimberly R. Stackhouse (1) Department of Animal Science, University of California, California, USA Huixiao Wang (101) Key Laboratory for Water and Sediment Sciences, Ministry of Education, College of Water Sciences, Beijing Normal University, Beijing, People’s Republic of China Minghua Zhang (101) Department of Land, Air and Water Resources, University of California, California, USA

PREFACE

Volume 103 contains four excellent reviews on topics that are of global significance—impacts of animal production on air quality and global climate change, food security, and enhancement of food production via drought resistance and weed management. Chapter 1 is a timely review on the impacts of livestock production on air quality and climate change. Chapter 2 is a comprehensive treatise on advances in improving drought resistance in rice including conventional breeding and molecular approaches. Chapter 3 discusses some of the problems, challenges, and options for protecting grain security in China. Chapter 4 provides a thorough review on weed management in rice-based cropping systems in Africa including details on weed species, management practices, and emerging weed challenges. I thank the authors for their fine contributions. DONALD L. SPARKS Newark, Delaware, USA

ix

C H A P T E R

O N E

Clearing the Air: Livestock’s Contribution to Climate Change Maurice E. Pitesky,* Kimberly R. Stackhouse,† and Frank M. Mitloehner†,1 Contents 1. Introduction 1.1. Overview of global, national, and state (California) reports on livestock’s role in climate change 1.2. Global estimates for livestock’s impact on climate change 1.3. United States estimates for livestock’s impact on climate change 1.4. California estimates for livestock production effects on climate change 2. Life Cycle Assessment 3. Effects of Agriculture on Climate Change 4. Livestock Types and Production Systems 5. Enteric Fermentation 5.1. Carbon dioxide emissions from livestock respiration 6. Animal Manure 7. Livestock Related Land-Use Changes 8. Livestock Induced Desertification 9. Release from Cultivated Soil 10. Carbon Emissions from Feed Production 11. On-Farm Fossil Fuel Use: Diesel and Electricity 12. Postharvest: CO2 from Livestock Processing 12.1. Transportation 12.2. Waste and biomass 13. Conclusions Acknowledgment References

* { 1

3 3 3 4 5 8 9 11 15 17 18 20 23 24 26 29 30 31 32 33 35 36

School of Veterinary Medicine, University of California, California, USA Department of Animal Science, University of California, California, USA Corresponding author: email: [email protected]

Advances in Agronomy, Volume 103 ISSN 0065-2113, DOI: 10.1016/S0065-2113(09)03001-6

#

2009 Elsevier Inc. All rights reserved.

1

2

Maurice E. Pitesky et al.

Abstract The United Nations, Food and Agricultural Organization [FAO, Steinfeld, Gerber, Wassenaar, Castel, Rosales, and de Haan (2006). Livestock’s Long Shadow. Food and Agriculture Organization of the United Nations] report titled Livestock’s Long Shadow (LLS) stated that 18% (approximately 7100 Tg CO2eq yr 1) of anthropogenic greenhouse gases (GHGs) are directly and indirectly related to the world’s livestock. The report’s statement that livestock production is responsible for a greater proportion of anthropogenic emissions than the entire global transportation sector (which emits 4000–5200 Tg CO2-eq yr 1) is frequently quoted in the public press [Fox News and Kroll (2009). A Tearful, Reluctant Farewell to My Favorite Food: Meat; LA Times (2007). A warming world; pollution on the hoof; livestock emissions are a leading source of greenhouse gases. One solution may be to eat less meat, Los Angeles; NY Times, Op-ed. (2009). Meat and the Planet. New York City] and continues to inform public policy. Recent estimates by the United States Environmental Protection Agency [EPA, Hockstad, Weitz (2009). Inventory of U.S. greenhouse gases and sinks: 1990–2007. Environmental Protection Agency] and the California Energy Commission [CEC—California Energy Commission (2005). Inventory of California Greenhouse Gas Emissions and Sinks: 1990 to 2002 Update] on the impacts of livestock on climate change in the United States and California have arrived at much different GHG estimates associated with direct livestock emissions (enteric fermentation and manure), totaling at less than 3% of total anthropogenic GHG and much smaller indirect emissions compared to the global assessment. Part of the difference of the global versus national predictions is due to the significant weight that has been assigned to the category of ‘‘land-use change’’ patterns related to livestock production (mainly deforestation). Furthermore, LLS attempts a life cycle assessment for global livestock production but does not use an equally holistic approach for its transportation prediction numbers. The primary focus of the present paper is to examine the relative contributions of livestock to climate change at different geographical and production scales. [Note:CO2equivalents (CO2-eq.) represent the total impact (radiative forcing) of GHG in the atmosphere, thereby making it possible to determine the climate change impact of one GHG versus another EPA [EPA and Holtkamp, Irvine, John, Munds-Dry, Newland, Snodgrass, and Williams (2006). ‘‘Inventory of U.S. Green House Gases and Sinks: 1996–2006.’’]. The definition of the Global Warming Potential (GWP) for a particular GHG is the ratio of heat trapped by one unit mass of the GHG to that of one unit mass of CO2 (the GWP of CO2 is one) over a specific period of time [IPCC (2001). IPCC Third Assessment Climate Change 2001. A Report of the Intergovernmental Panel on Climate Change]. The 100-year GWP for CH4 and N2O are 23 times and 296 times the GWP of CO2, respectively [IPCC (2001). IPCC Third Assessment Climate Change 2001. A Report of the Intergovernmental Panel on Climate Change]. Therefore, for simplicity sake it is common practice to combine the total effects of CO2, CH4, and N2O into CO2 equivalents (or CO2-eq).]

Clearing the Air: Livestock’s Contribution to Climate Change

3

1. Introduction 1.1. Overview of global, national, and state (California) reports on livestock’s role in climate change Livestock’s Long Shadow (LLS) (FAO et al., 2006) is a life cycle assessment (LCA) of livestock’s global impact on biodiversity, land-use, water depletion, water pollution, air pollution, and anthropogenic GHG emissions. The report attempts to quantify the global direct and indirect GHG emissions associated with livestock. Direct and indirect sources of GHG emissions in animal production systems include physiological processes from the animal (enteric fermentation and respiration), animal housing, manure storage, treatment of manure slurries (compost and anaerobic treatment), land application, and chemical fertilizers (Casey et al., 2006; Monteny et al., 2001). Direct emissions refer to emissions directly produced from the animal including enteric fermentation and manure and urine excretion ( Jungbluth et al., 2001). Specifically, livestock produce CH4 directly as a byproduct of digestion via enteric fermentation (i.e., fermenting organic matter via methanogenic microbes producing CH4 as an end-product) ( Jungbluth et al., 2001). Methane and N2O emissions are produced from enteric fermentation and nitrification/denitrification of manure and urine, respectively (Kaspar and Tiedje, 1981). Previous agricultural estimates have included emissions associated with indirect energy consumption (e.g., electricity requirements, off-site manufacturing, etc.) as five times greater than on-site emissions for cropland production (Wood et al., 2006). Therefore, to accurately estimate the full environmental impact of livestock, indirect emissions need to be quantified. For livestock production, the term indirect emissions refers to emissions not directly derived from livestock but from feed crops used for animal feed, emissions from manure application, CO2 emissions during production of fertilizer for feed production, and CO2 emissions from processing and transportation of refrigerated livestock products (IPCC, 1997; Mosier et al., 1998a). Other indirect emissions include net emissions from land linked to livestock including deforestation (i.e., conversion of forest to pasture and cropland for livestock purposes), desertification (i.e., degradation of above ground vegetation from livestock grazing), and release of C from cultivated soils (i.e., loss of soil organic C (SOC) via tilling, natural processes) associated with livestock (IPCC, 1997).

1.2. Global estimates for livestock’s impact on climate change LLS estimates the global contribution of anthropogenic GHG emissions from the livestock sector at 7100 Tg CO2-eq yr 1, which is approximately 18% of global anthropogenic GHG emissions (FAO et al., 2006). For

4

Maurice E. Pitesky et al.

comparison, global fossil fuel burning accounts for 4000–5200 Tg CO2-eq yr 1 (FAO et al., 2006). According to FAO et al. (2006), the major categories of anthropogenic GHG emissions are: 1. 2. 3. 4. 5. 6. 7. 8.

Enteric fermentation and respiration (1800 Tg CO2-eq yr 1) Animal manure (2160 Tg CO2-eq yr 1) Livestock related land-use changes (2400 Tg CO2-eq yr 1) Desertification linked to livestock (100 Tg CO2-eq yr 1) Livestock related release from cultivated soils (230 Tg CO2-eq yr Feed production (240 Tg CO2-eq yr 1) On-farm fossil fuel use (90 Tg CO2-eq yr 1) Postharvest emissions (10–50 Tg CO2-eq yr 1)

1)

Using the first seven of the eight categories listed above, livestock account for 9, 35–40, and 65% of the total global anthropogenic emitted CO2, CH4, and N2O, respectively (FAO et al., 2006).

1.3. United States estimates for livestock’s impact on climate change A second recent report issued by the United States Environmental Protection Agency (EPA) titled ‘‘Inventory of United States Greenhouse Gases and Sinks: 1990–2007’’ (EPA et al., 2007) uses a similar comprehensive LCA methodology compared to LLS (FAO et al., 2006) to characterize the contribution of livestock (and other industries) within the United States with respect to anthropogenic GHG emissions. The EPA et al. (2007) report provides a United States national inventory of anthropogenic GHG sources categorized by industry and location (i.e., states within the United States). Based on the total gross anthropogenic emissions of 7150 Tg CO2-eq yr 1 produced within the United States, the EPA calculates that 5.8% (or 413 Tg CO2-eq yr 1) is associated to the entire agricultural sector (i.e., enteric fermentation, livestock manure management, rice cultivation, agricultural soil management, and burning of crop residues, etc.). Specifically, agriculture in the United States represents 32% of the anthropogenic CH4 emission and 68% of the N2O emission (EPA et al., 2009). Within the United States, approximately 198 Tg CO2-eq yr 1 or 2.8% is associated with livestock (i.e., enteric fermentation and manure management). However, as a reference point for the United States, the transportation sector accounted for 26% (or 1887 Tg CO2-eq yr 1) of the total (7150 Tg CO2-eq yr 1) United States anthropogenic GHG portfolio, reflecting the significance of fossil fuel combustion (EPA et al., 2009) and the relative significance of transportation versus animal agriculture. Therefore, the global prediction that livestock account for 18% of GHG emissions

5

Clearing the Air: Livestock’s Contribution to Climate Change

and therefore have a ‘‘larger’’ GHG ‘‘footprint’’ than the transportation sector (FAO et al., 2006) is not accurate for the United States. Within the agricultural sector, the EPA et al. (2009) has identified several ‘‘key’’ categories (both direct and indirect sources of GHG emissions). The sources are: 1. 2. 3. 4. 5.

Agricultural soil management (209 Tg CO2-eq yr 1) Enteric fermentation (139 Tg CO2-eq yr 1) Manure management (59 Tg CO2-eq yr 1) Rice cultivation (6.2 Tg CO2-eq yr 1) Field burning of agricultural residues (1.4 Tg CO2-eq yr

1)

1.4. California estimates for livestock production effects on climate change In accordance with EPA and IPCC methods, the state of California compiled its own GHG inventory (CEC, 2005). In 2004, the California inventory estimated that 27 Tg CO2-eq yr 1 or 5.4% of California’s gross anthropogenic GHG profile (492 Tg CO2-eq yr 1) is associated directly and indirectly with agriculture. Within California agriculture, approximately 14 Tg CO2-eq yr 1 or 2.8% is associated with livestock (i.e., enteric fermentation and manure management). Consistent with global (i.e., FAO et al., 2006) and national (i.e., EPA et al., 2009) data, agricultural soil management and enteric fermentation were the greatest emitters of anthropogenic CH4 and N2O in California (California Environmental Protection Agency, 2007). As a reference point for California, in 2004 the transportation sector accounted for 182 Tg CO2-eq yr 1 or 37% of the total (492 Tg CO2-eq yr 1) California anthropogenic GHG portfolio, reflecting the significance of fossil fuel combustion (CEC, 2005) to overall GHG emissions. Again, the global prediction for the relative contribution of livestock versus transportation to climate change (livestock account for 18% of GHG emissions which is more than transportation) is a significantly inaccurate when applied to California, which is the largest dairy and agricultural state within the United States (NASS, 2009). The major categories of anthropogenic GHG emissions investigated by the State of California (California Environmental Protection Agency, 2007) within the agricultural sector include the following (from highest to lowest emissions): 1. 2. 3. 4. 5.

Agricultural soil management (9.1 Tg CO2-eq yr 1) Enteric fermentation (7.2 Tg CO2-eq yr 1) Manure management (6.9 Tg CO2-eq yr 1) Rice cultivation (0.6 Tg CO2-eq yr 1) Field burning of agricultural residues (0.2 Tg CO2-eq yr

1)

6

Maurice E. Pitesky et al.

While all three reports (CEC, 2005; EPA et al., 2009; FAO et al., 2006) have similar goals (to quantify the relative role of agricultural sources relative to overall anthropogenic GHG emissions), the scope of each report coupled with specific assumptions makes comparison, extrapolation, and interpretation of one report to another cumbersome. These differences are due to several factors including geography (i.e., regional vs global), scope, and methodology (i.e., different assumptions, coefficients, and models). For example, with respect to scope, the EPA et al. (2009) and CEC (2005) reports currently do not identify CO2 emissions from fossil fuel burning related to agriculture. However, the CEC (2005), EPA et al. (2009), and FAO et al. (2006) reports are largely similar from a methodology perspective. Figure 1 shows a comparison of predicted relative GHG emissions across all three reports. Globally, FAO et al. (2006) predicts land-use change

A

Global livestock GHG emissions, % Land-use change accounted for

Land-use change unaccounted for 2.0

5.0

2.2

5.2 31

34

47

39

1.3 1.4 26

3.3 3.4

Livestock related land-use change

Animal Manure

Cultivated livestock related soils

B

Enteric Fermentation

Desertification

Feed production

On-farm fossil fuel use

Agricultural GHG emissions, % United States

California 34

36

31

50

0.4 0.4

14 1.5

Agricultural soil management Rice cultivation

3 Enteric fermentation

30 Manure management

Field burning of agricultural residues

Figure 1 GHG emissions associated with global livestock (A), United States emissions, and California agricultural emissions (B). Direct and indirect N2O emissions associated with application and deposition of manure are accounted for in the "agriculture soil management" section in the EPA and CEC reports; while in the FAO report, those emissions are accounted for in the animal manure section. Source: data from CEC (2005), EPA et al. (2006), and FAO (2006).

Clearing the Air: Livestock’s Contribution to Climate Change

7

(35.3%) as the primary source of livestock related anthropogenic GHGs (Fig. 1A). The ranking of GHG sources from highest to lowest emissions is identical between EPA et al. (2009) and CEC (2005) (Fig. 1B). However, agricultural soil management is a larger source of emissions in the United States as a whole versus California (50.0% vs 36.0%, respectively) (CEC 2005; EPA et al., 2009). All three reports (CEC, 2005; EPA et al., 2009; FAO et al., 2006) use a combination of Intergovernmental Panel on Climate Change (IPCC) Tier I (uses population data coupled with global emissions factors) and Tier II (same data as Tier I applies more accurate equations based on diet and digestibility coupled with uncertainty analysis). The EPA uses a sophisticated Tier III process-based model (DAYCENT) model to estimate direct emissions from major crops and grassland. The Tier III model uses detailed predictions incorporating local management and weather conditions (among other variables). The Tier I–III models conform the United Nations Framework Convention on Climate Change (IPCC, 2007). However, some differences in assumptions between the three reports were noted: 1. Some parameters were modified to make them more relevant to national and California livestock systems. For example, the State of California adjusted residue-to-crop mass ratio and the fraction of residue applied to reflect the decreased agricultural burning within California (California Environmental Protection Agency, 2007). The EPA report incorporates the Cattle Enteric Fermentation Model (CEFM), which is a refinement of the Tier II calculation (EPA et al., 2009). Major refinements include linkage of livestock performance data to the growth stage of the animal. Specifically, factors such as weight gain, birth rates, pregnancy, feedlot placements, diet, and animal harvest rates are tracked to characterize the United States cattle population on a monthly basis versus the Tier II model, which is updated annually with respect to those variables. Furthermore from a statistical perspective, the EPA report includes a range (e.g., upper and lower boundaries) of emissions estimates predicted by Monte Carlo simulations for a 95% confidence interval (EPA et al., 2009). 2. Another major difference across the three reports is that FAO et al. (2006) focuses on livestock while the EPA et al. (2009) and California (CEC, 2005) reports include agriculture as a whole (i.e., livestock and plant crops). With respect to the EPA et al. (2009) data, it is important to define the agricultural soil management category, which includes applying fertilizers and manure, growing N-fixing crops, retaining crop residues, liming of soils, depositing waste by domestic and grazing animals, and cultivating histosols (i.e., soils with high organic matter content). For example, in the CEC (2004) and EPA et al. (2009) reports, agricultural soil management (the largest source of GHG emissions in the United States and California), includes GHG emissions associated with growing fruits, vegetables, fiber grain, as well as livestock pasture and rangeland.

8

Maurice E. Pitesky et al.

2. Life Cycle Assessment According to International Standard ISO 14040, an LCA is a ‘‘compilation and evaluation of the inputs, outputs, and the potential environmental impacts of a product or service throughout its life cycle’’ (International Organization for Standardization, 2006). A LCA is a methodology used to assess both the direct and indirect environmental impact of a product from ‘‘cradle to grave.’’ Environmental impacts that can be measured include fossil fuel depletion, water use, GWP, ozone depletion, and pollutant production. Figure 2 shows a partial LCA for livestock production (NRC, 2003). While there are international standards with respect to LCA analysis, uncertainties exist regarding the definitions and ‘‘boundaries’’ of indirect environmental impacts. For example, should the energy required to extract the coal that is used to make the fertilizer, that is applied to the cropland to grow animal feed be included in a ‘‘true’’ LCA of livestock? According to ISO 14040 (International Organization for Standardization, 2006) a comprehensive approach would be ideal but is often not practical. Hence further refinement of the scope and methodology is necessary to increase comparability between LCAs. Lal (2004) described primary (i.e., tilling, sowing, harvesting, pumping water, grain drying), secondary (i.e., manufacturing, packaging, and storing fertilizers and pesticides), and tertiary (i.e., acquisition of raw materials and fabrication of equipment and buildings) emission sources (Lal, 2004). Therefore, based on Lal (2004), one possible method would include LCAs with a numerical suffix indicating the ‘‘degree of separation’’ between the product (e.g., animal protein) and the indirect emissions source input (i.e., the greater the number the more complete and complex the LCA). Emissions

Export

Feed

Import/ export

Crop Herd

Product

Emissions Emissions

Fertilizer

Soil Emissions

Manure

Import/ export

Emissions

Figure 2 Example of an LCA model for livestock. The model reflects on-site and off-site inputs associated with livestock production. This would not be considered a complete LCA since emissions are only estimated for feed, herd, manure, soil, and crop. Source: NRC (2003).

Clearing the Air: Livestock’s Contribution to Climate Change

9

For example, the LCA in Fig. 2 would be an LCA-1 because only feed, herd, manure, soil, and crop emissions are being accounted for. Regardless, the goal of the LCA is to understand all (or the major) environmental impacts of a product or service to identify the main pollution sources. Aside from LCA analysis there are several other types of assessment tools for determining the environmental impact of various products and services at a local or global scale. Halberg et al. (2005) reviewed multiple assessment tools and concluded that LCAs are ideal for global analysis of products (including livestock production systems (LPSs)) while ecological footprint analysis (EFA) are better suited for studying specific local geographical target areas such as nutrient surplus per hectare (Halberg et al., 2005).

3. Effects of Agriculture on Climate Change Biogenic emissions of CO2, CH4, and N2O are emitted as part of the natural biogeochemical cycling of C and N (e.g., decomposition or burning of plant material). Anthropogenic emissions of CO2, CH4, and N2O are emitted due to human decisions, activity, and influence of our abiotic and biotic environment (Bruinsma, 2003). Since the industrial revolution in 1750, CO2 concentrations have increased from 280 to 379 ppm, CH4 concentrations have increased from 715 to 1732 ppb, and N2O concentrations have increased from 270 to 319 ppb (IPCC, 1997). Since 1970, atmospheric concentration of CO2, CH4, and N2O has increased by approximately 31, 151, and 17%, respectively, in the United States (USDA, 2004). Figure 3 shows global CH4 and N2O emissions (magnitude and source) within the agricultural sector for 10 different global regions (Smith et al., 2007a). While the gross emissions are not normalized to population (e.g., approximately 20% of the world’s population live in developed countries), it is important to recognize that the developing world emits approximately two thirds of all anthropogenic agricultural GHG. In addition, Fig. 3 predicts an increased rate of agricultural emissions through 2020. In six of the 10 world regions, N2O from soils was the primary agricultural source of GHGs. These N2O emissions are primarily due to fertilizer and animal manure applied to agricultural soils. In the other four regions (Latin America and the Caribbean, Central and Eastern Europe, the Caucasus and Central Asia, and OECD Pacific), CH4 from enteric fermentation was the primary source of agricultural emissions (Smith et al., 2007a). Currently, over half of the total global CH4 emissions and one third of N2O emissions are from anthropogenic sources including agriculture, landfills, biomass burning, industrial activities, and natural gas (IPCC, 1997). The IPCC (1997) estimated that the agricultural sector contributes between 10 and 12% of global anthropogenic CO2 emissions (i.e., fossil fuel

2000

Developing countries of South Asia Mt CO2-eq

Developing countries of East Asia

Sub-Saharan Africa N2O manure N2O soils N2O burning

1500 1000

CH4 rice CH4 manure

500

CH4 enteric CH4 burning

0 Latin America and the Carribean

Middle East and North Africa

Caucasus and Central Asia

Western Europe (EU15, Norway and Switzerland)

Central and Eastern Europe

OECD Pacific (Australia, New Zealand, Japan, Korea)

Developing regions

Developed regions

2000 1500 1000 500 0

2000 1500 1000 500 0 OECD North America (Canada, USA, Mexico) 6000

2000 1500

4000

1000 2000

500

20

15

20

10

20

05

20

00

20

95

20

90

19

19

20

15

20

10

20

05

20

00

20

95

20

19

19

20

15

20

10

20

05

20

00

20

95

20

90

19

19

90

0

0

Figure 3 Estimated agricultural N2O and CH4 emissions on 10 world regions between 1990 and 2020. Source: Adapted from Fourth Assessment Report of the IPCC (2007) and Smith et al. (2007a).

Clearing the Air: Livestock’s Contribution to Climate Change

11

burning), 40% of global anthropogenic CH4 emissions (i.e., enteric fermentation, wetland rice cultivation, decomposition of animal waste), and 65% of global anthropogenic N2O emissions (i.e., agricultural soils, use of synthetic and manure fertilizers, manure deposition, biomass burning) (De Gryze et al., 2008; IPCC, 1997). Therefore, agriculture is considered the largest source of anthropogenic CH4 and N2O at the global, national, and state level (CEC, 2005; De Gryze et al., 2008; EPA et al., 2009), while transportation is considered the largest anthropogenic source of CO2 production (EPA et al., 2009). C and N are part of dynamic cycles that are dependent on multiple environmental conditions. Specifically, oxidation state, pH, water activity, nitrification, denitrification, fermentation, ammonia volatilization, and the microbial ecology of the environment quantitatively and qualitatively affect GHG emissions (CAST, 2004). In addition, emission sources are dispersed and largely driven by biological activity with significant variability over time, space, and management practices (CAST, 2004). Emissions are further affected by local and regional meteorological and soil conditions. Several examples of qualitative variability of GHG production due to environmental conditions have been cited in the literature. For example, under aerobic conditions CO2 is preferentially produced relative to CH4 production (De Gryze et al., 2008). However, under anaerobic conditions via methanogenesis (i.e., in rice fields or in a bovine’s rumen), CH4 is preferentially produced relative to CO2 production. The CH4 produced can then be converted to CO2 by microorganisms via CH4 oxidation (De Gryze et al., 2008). Because CH4 has 21–23 times the GWP of CO2, understanding the environmental conditions of CH4 and CO2 formation is integral toward both the development of an accurate model and mitigation.

4. Livestock Types and Production Systems Greenhouse gas emissions from livestock are inherently tied to livestock population size (USDA, 2004). However, due to their greater biomass and unique metabolic function, ruminants are the most significant livestock producer of GHGs (USDA, 2004). Figure 4 shows the estimated global distribution of pigs, poultry, cattle, and small ruminants. There are currently 1.5 billion cattle and domestic buffalo, and 1.7 billion domestic sheep and goats in the world, which account for over two thirds of the total biomass of livestock (FAO et al., 2006). Within the United States, there are over 94 million beef cattle and 9.3 million dairy cows (NASS, 2009). Cattle are the largest contributing species to enteric fermentation in the United States (EPA et al., 2009). In all three reports discussed in the present chapter (CEC, 2005; EPA et al., 2009; FAO et al., 2006), CH4 from enteric

12

Maurice E. Pitesky et al.

Livestock units per square km 0 0.1–0.5 0–0.1 0.5–1

1–2.5 >2.5

National boundaries

Figure 4 Global estimates of aggregate distribution of pigs, poultry, cattle, and small ruminants (FAO, 2006).

fermentation is the second leading source of GHG from livestock. Therefore, when evaluating LLS (FAO et al., 2006) with respect to GHGs, domesticated ruminants are the primary species studied. However, it is important to recognize the significance of other nonruminant livestock. For example, in the United States swine are the second greatest source of CH4 and N2O emissions from manure management and have had a CH4 and N2O emissions increase of 34% between 1990 and 2006 (EPA et al., 2006). In addition, pork and poultry production currently consume over 75% of cereal and oil-seed based on concentrate that is grown for livestock (Galloway et al., 2007). Therefore, while ruminants consume 69% of animal feed overall, nonruminates consume 72% of all animal feed that is grown on arable land (Galloway et al., 2007). Consequently, while enteric fermentation from nonruminants is not a significant source of GHG, indirect emissions associated with cropland dedicated to nonruminant livestock might be significant. The types of LPSs utilized are typically based on socioeconomics, tradition, and available resources. LLS states that extensive (i.e., grazing animals) and intensive (i.e., animals are contained and feed is brought to them) LPSs emit 5000 and 2100 Tg CO2-eq yr 1, respectively (FAO et al., 2006). While these emissions numbers are not normalized to a per animal unit scale, the type of production system utilized (i.e., landless vs grassland) affects direct (i.e., from the animal) and indirect (i.e., emissions associated with livestock) emissions quantitatively and qualitatively. For example, the low animal density coupled with high land area utilized by extensive systems

13

Clearing the Air: Livestock’s Contribution to Climate Change

(e.g., grazing animals occupy 26% of the earth’s terrestrial surface) can affect land degradation, deforestation, soil erosion, biodiversity loss, and water contamination (Bruinsma, 2003; FAO et al., 2006). Likewise, because of their high animal density, intensive farming systems can lead to N and P saturation, salinization, and water contamination in addition to reliance on external feed-crop production (Bruinsma, 2003; Mosier et al., 1998a). Therefore, to characterize the GHG ‘‘footprint’’ of livestock, the type of LPS needs to be identified and characterized. On the basis of the system parameters (e.g., feed type, animal density, manure storage, and use etc.), FAO et al. (2006) divides the LPS into two major types (solely LPSs (L) and mixed farming systems (M)). Figure 5 shows the global distribution of production systems (FAO et al., 2006). The solely LPSs are further divided into landless LPS (LL) and grasslandbased LPS (LG): 1. Landless LPS: Intensive/feedlot type system (defined as systems in which less than 10% of the dry matter fed to animals is farm-produced and where the annual stocking rates are above 10 livestock units per km2). Developed countries are the primary users of this system with 54.6% of total LL meat production produced in LL systems (FAO et al., 2006). Globally LL-systems account for 75% of the world’s broiler poultry supply, 40% of its pork, and over 65% of all poultry eggs (Bruinsma, 2003).

Livestock production systems Mixed, irrigated Grazing Mixed, rainfed

Other type

Areas dominated by landless production

National boundaries

Boreal and arctic climates

Figure 5 Estimated distribution of livestock production systems. Landless production systems refer exclusively to monogastric production (FAO, 2006).

14

Maurice E. Pitesky et al.

2. Grassland-based LPSs are defined as areas where more than 10% of dry matter fed to animals is produced at the farm and where annual stocking rates are less than 10 livestock units per hectare of agricultural land (FAO et al., 2006). Grassland-based LPSs are usually present on land that is considered unfit for cropping (primarily semiarid or arid areas). These systems cover the largest global land area and are currently estimated to occupy some 26% of the earth’s ice-free land surface (FAO et al., 2006). In South and Central America and part of South East Asia, grazing is often pursued on land cleared from rainforests, where it fuels soil degradation and further deforestation. In semiarid environments, overstocking during dry periods frequently brings risks of desertification (e.g., in sub-Saharan Africa), although it has been shown that marginal pastures do recover quickly if livestock are taken off and rainfall occurs (Bruinsma, 2003). In general, the LG system is characterized by a lower feed quality and a higher feed intake, which leads to higher methane emissions per animal relative to LL production system (Kebreab et al., 2008). Mixed farming, in which livestock provide manure and power in addition to milk and meat, still predominates for cattle. Mixed farming systems can be divided into the Rain-fed LPS (MR) and the Irrigated LPS (MI). 1. Rain-fed LPS: Mixed systems in which greater than 90% of the value of nonlivestock farm production come from ‘‘rain-fed’’ land use (Ash and Scholes, 2005). In MR, the livestock and cropping components are interwoven. The MR systems are prevalent in temperate, semiarid, and subhumid areas. Approximately two thirds of the total livestock population in India are raised in rain-fed LPS due to the availability of forest grazing and wasteland (Dash and Misra, 2001). These systems typically have large and overstocked livestock populations (Ash and Scholes, 2005). The excess manure is used for cultivation of crops; however, the high animal density can contribute to land-use degradation (Ash and Scholes, 2005). 2. Irrigated mixed farming systems: More than 10% of the value of nonlivestock farm production comes from irrigated land-use. Crop production under irrigated conditions used primarily for rice production with goats as the primary food animal (Ash and Scholes, 2005). Goats typically have low growth and relatively high mortality rates (Ash and Scholes, 2005). Most GHG production is from methane associated with animal manure and irrigated rice cultivation (FAO et al., 2006). Using the eight categories that most LCA uses to divide anthropogenic GHG emissions associated with global and regional livestock, a comprehensive analysis of each category follows with respect to current literature. Based on the comparison the overall relevancy of each category is then assessed for United States livestock.

15

Clearing the Air: Livestock’s Contribution to Climate Change

5. Enteric Fermentation Methane production from enteric fermentation is considered the primary source of global anthropogenic CH4 emissions accounting for approximately 73% of the 80 Tg of CH4 produced globally per year ( Johnson and Johnson, 1995). Globally as well as in the United States and California, CH4 released from enteric fermentation accounts for ~1800, 139, and 7 Tg CO2-eq yr 1, respectively (CEC, 2005; EPA et al., 2009; FAO et al., 2006). LLS (FAO et al., 2006) estimated that 1800 Tg CO2-eq yr 1 is produced globally via CH4 from enteric fermentation following only land-use change as an emission category. Ruminants are unique in their ability to convert plants on nonarable land to protein. This characteristic allows ruminants to utilize land and feed that would otherwise be un-used for human food production. At the same time, ruminant livestock is an important contributor to CH4 in the atmosphere (FAO et al., 2006; IPCC, 2000; USDA, 2004). Methane is produced from the microbial digestive processes of ruminant livestock species such as cattle, sheep, and goats. Nonruminant livestock such as swine, horses, and mules produce less CH4 than ruminants (USDA, 2004) (Fig. 7).

Tonnes of CO2 equivalent Dairy cattle Cattle and buffalo Small ruminant

Poultry Pigs

150 mil. CO2 tonnes eq.

Grazing Mixed Industrial

Figure 6 Total GHG emissions from enteric fermentation and manure per species and main productions system (FAO, 2006).

16

Maurice E. Pitesky et al.

The primary source of CH4 from ruminant livestock is from the process of enteric fermentation during rumination (Casey et al., 2006; Jungbluth et al., 2001; Kaspar and Tiedje, 1981; Sun et al., 2008). Initial microbial breakdown (essential in ruminant digestion) occurs in the rumen, or large fore-stomach, where microbial fermentation converts fibrous feed into products digested and utilized by the animal (Boadi et al., 2004; USDA, 2004). Rumination promotes digestion of cellulose and hemicellulose through hydrolysis of polysaccharides by microbes and protozoa, which is followed by microbial fermentation generating H2 and CO2. Methane is produced as a by-product of enteric fermentation and carbohydrate digestion and is expelled through the mouth via eructation (Monteny et al., 2001). Global CH4 emissions are difficult to predict because specific biochemical components of diets are often overlooked in empirical models. Important differences in feed components of the diets used in extensive and intensive LPSs are often overlooked and these systems are viewed as similar. This can result in over- and underestimates of enteric derived CH4 emissions regionally; especially, where diet components may differ based on the availability of nutrients. Kebreab et al. (2008) suggested that IPCC values overestimate CH4 emissions by 12.5% and underestimate CH4 emissions by 9.8% for dairy and feedlot cattle, respectively. Mechanistic models might be better suited than empirical models for determining CH4 emissions as the models are capable of changing source of carbohydrate or addition of fat to decrease methane (Kebreab et al., 2008). Models that predict methane emissions should depend on the diet being fed and the variables relevant to an animal on a particular diet (Ellis et al., 2009). Predictions of CH4 for an animal on a high grain diet should include some aspect of crude fiber (FC), starch, or forage percentage; while, an animal on a high-fat diet, predictions should include a fat variable (Ellis et al., 2009). In ruminant livestock, enteric fermentation is strongly affected by quantity and quality of their diet ( Johnson and Johnson, 1995). Production of CH4 in ruminants is directly correlated to a loss of metabolizable energy and has been studied in depth during performance studies that aimed at improvements of feed efficiency ( Johnson and Johnson, 1995; Jungbluth et al., 2001; Mosier et al., 1998b). Cattle typically lose 2–12% of their ingested energy as eructated CH4 ( Johnson and Johnson, 1995). Many factors affect CH4 emissions from livestock including feed intake, animal size, diet, growth rate, milk production, and energy consumption ( Johnson and Johnson, 1995; Jungbluth et al., 2001). Diet and level of production directly affect CH4 emission rates (Holter and Young, 1992; Jungbluth et al., 2001; Sun et al., 2008). For example, CH4 outputs are estimated to range from 3.1 to 8.3% of gross energy intake for dry, non-lactating cows and from 1.7 to 14.9% of gross energy intake for lactating cows (Holter and Young, 1992). Enteric CH4 ethane emissions per unit of production are highest when feed quality and level of production are low (Crutzen et al., 1986).

Clearing the Air: Livestock’s Contribution to Climate Change

17

Mitigation through improved feed efficiency could reduce CH4 emissions and result in economic benefits to producers while improving global methane emissions ( Johnson and Johnson, 1995). The use of high energy concentrate feed typically used in landless (LL) LPSs results in relatively higher animal production rates ( Johnson and Johnson, 1995) and thus less CH4 emitted per unit of output. Due to the regional differences in animal species, diets, and production systems, (Figs. 5 and 6) globally it is very difficult to determine accurate CH4 emissions. Most LL LPSs feed high concentrate diets that meet the specific energy requirements of the animal and thus increase production efficiency, using less resources (feed) to obtain a useable product (meat or milk) in less time. In contrast, extensive (grassland) LPSs, where inputs are less controlled and animals roam freely, feed production efficiency decreases. In other words, these animals require more feed and more time to reach an endpoint that yields useable products. Emissions from livestock can be mitigated through animal management techniques including nutrition, housing, and waste management (Clemens and Ahlgrimm, 2001; Johnson and Johnson, 1995; Mosier et al., 1998b; Phetteplace et al., 2001; Saggar et al., 2004). Recent work has focused on manipulating the abundance and/or activity of rumen methanogens, to improve the efficiency of ruminant production in an ecologically sustainable way (Wright et al., 2004). One major mitigation technique for CH4 from livestock is through improvement of production efficiency. For example, in the United States, Capper et al. (2009) suggests that continued improvement of management systems and technologies in commercial operations would reduce resource use and environmental impact without sacrificing production. When comparing 1944 with 2007 dairies in the United States, Capper et al. (2009) found that modern dairies require 21% of animals, 23% of feedstuffs, 35% of the water, and 10% of the land to produce the same one billion kg of milk. Emissions have also been reduced since 1944; dairies today produce 43% of CH4 and 56% of N2O per billion kg of milk (Capper et al., 2009). Management with particular emphasis on improvements of production and reproduction efficiency will likely be among the most viable tools to most significantly reduce environmental impact of livestock systems.

5.1. Carbon dioxide emissions from livestock respiration The CO2 from respiration of livestock amounts to ~3000 Tg CO2-eq yr 1 but this CO2 had previously been absorbed via plants (FAO et al., 2006). According to EPA et al. (2006), FAO et al. (2006), and the Kyoto Protocol (1997), emissions from livestock are part of continuous cycling biological system where plant matter that had once sequestered CO2 is consumed by livestock and then released back into the atmosphere by respiration to be

18

Maurice E. Pitesky et al.

U.S. greenhouse gas emissions from livestock, 2001 Tg CO2 eq. 140

Livestock GHG emissions Manure N2O 27%

120 100

Manure CH4 18%

80 60

Enteric fermentation CH4 55%

40 20 0

Beef cattle

Dairy cattle

Swine Poultry Horses Sheep

Enteric fermentation CH4

Manure CH4

Goats

Manure N2O

Figure 7 United States greenhouse gas emissions by livestock type, 2001 (USDA, 2004). Note, the United States has approximately 10 times more beef than dairy cattle leading to differences in total contributions (FAO, 2006).

reabsorbed by plants (FAO et al., 2006; Kyoto Protocol, 1997). Consequently, the emitted and absorbed quantities are considered equivalent making livestock a net zero source of CO2.

6. Animal Manure The management of animal manure can produce anthropogenic CH4 via anaerobic decomposition of manure and N2O via nitrification and denitrification of organic N in animal manure and urine (Bouwman, 1996). LLS (FAO et al., 2006) estimated that global emissions associated with livestock manure (i.e., manure management, manure land application, and indirect manure emissions) total 2160 Tg CO2-eq yr 1. The EPA et al. (2009) and the state of California (CEC, 2005) have assessed that emissions associated with livestock manure (i.e., manure management) in the United States and California total 59.0 and 6.9 Tg CO2-eq yr 1, respectively. The EPA and CEC place manure land application and indirect manure emissions in the agricultural soil management section. For the EPA and CEC these total 21 and 2.3 Tg respectively.

Clearing the Air: Livestock’s Contribution to Climate Change

19

Typically, when livestock manure is stored or treated in lagoons, ponds, or tanks (i.e., anaerobic conditions), CH4 emissions are produced in higher amounts than when manure is handled as a solid (e.g., stacks or drylot corrals), or deposited on pasture where aerobic decomposition occurs thereby reducing CH4 emissions (EPA et al., 2006). Because a strong relationship exists between manure application on land and N2O emissions (Bouwman, 1996; Jarecki et al., 2008), the emissions associated with fertilization need to be considered a GHG source. However, LLS (FAO et al., 2006) only takes into account emissions from N fertilizer applied to animal feed crops dedicated to food animals, yet including emissions from manure when applied both to animal feed and human crops. The displacement of chemical N fertilizer that is not needed because of N from manure is not considered in LLS. In contrast to chemical fertilizers, the energy input is lower for animal manure (FAO et al., 2006). Therefore, while the direct CO2-eq kg 1 of manure is significantly higher for manure (7–8 kg CO2-eq kg 1 of N) than for fertilizer (between 0.03 and 1.8 CO2-eq kg 1) (Lal, 2004), the indirect emissions from chemical fertilizer that is not produced need to be accounted for to make an appropriate LCA analysis. Investigating LCAs of GHG emissions associated with fertilizer or manure application on cropland are essential toward understanding the significance of animal manure in agriculture. A major factor influencing N2O emissions from agricultural land is N application ( Jarecki et al., 2008). The form of fertilizer applied as well as the placement in the soil influences the flux of N2O emissions (Breitenbeck et al., 1980; Bremner et al., 1981). Both CH4 and N2O can be produced by the decomposition of manure. However, N fertilization reduces soil CH4 oxidation ( Jarecki et al., 2008). Methane is produced via the anaerobic decomposition of manure while N2O is produced via nitrification and denitrification of lad incorporated manure (Chen et al., 2008). Both CH4 and N2O production are influenced by multiple variables including climate, soil conditions, substrate availability, and land management practices (Chen et al., 2008). With respect to management in the developed world, the increased use of liquid versus dry manure waste systems (liquid systems produce significantly more methane) in dairy and pig operations has resulted in a relative increase in methane production (FAO et al., 2006). Specifically, in the United States, CH4 emissions from manure management increased by 34% between 1990 and 2006 primarily due to an increase in liquid manure systems (EPA et al., 2006). One reason for the trend toward liquid-based systems is a response to regulations in the United States including the United States Clean Water Act, which restricts land application rates of manure. The emerging use of CH4 digesters offers a potential mitigation of CH4 emissions from liquid manure systems coupled with electricity, gas, and biofuel generation. Current assumptions predict a 50– 75% reduction (depending on environmental conditions) in digester GHG emissions from manure when compared with the current system where the manure would otherwise be stored as a liquid slurry in a lagoon (AgStar, 2002).

20

Maurice E. Pitesky et al.

Wisconsin, New York, Pennsylvania, and California currently have 20, 16, 16, and 15 operating CH4 digesters, respectively (AgStar, 2002). Nitrogen assimilation efficiencies vary considerably among different livestock with a range between 10% in beef cattle and 38–75% for swine (Castillo et al., 2001; Hoekstra et al., 2007). As a result, a significant amount of N is returned to the environment through animal excretions (Clemens and Huschka, 2001; Hoekstra et al., 2007). This N can reenter the crop-production cycle, or depending on the conditions be emitted as N2O or NH3 (Mosier et al., 1998b). Direct N2O emissions are produced as part of the N cycle through the nitrification and denitrification of organic N in livestock manure and urine (Mosier et al., 1998b). Annual N losses via N2O have been previously calculated between 0 and 5% of N applied for manure ( Jarecki et al., 2008). Indirect N2O emissions are produced from N lost as runoff, and leaching of N during treatment, storage, and transportation (Mosier et al., 1998b). Due to primarily anaerobic conditions of rice production globally, methane production indirectly associated with animal manure application to irrigated rice fields is considered a significant source of emissions. Specifically, due to microbial breakdown of animal manure under anaerobic conditions, global methane emissions account for approximately 60 Tg CO2-eq yr 1 (Verburg and Van der Gon, 2001). In most of the developing world, most rice is grown under these conditions while within the developed world rice is grown with urea as N source. With respect to animal diet, higher energy feed will have increased methane production from manure. For example, feedlot cattle fed a concentrate diet (i.e., high energy) generate manure with up to 50% higher CH4 compared to range cattle eating a forage (i.e., low energy) diet (this trend is reversed for enteric fermentation where feedlot versus range cattle produce much less CH4 per unit of production). Consequently, according to LLS (FAO et al., 2006), the United States (highly intensive production systems) currently has the highest methane emissions factor for manure globally for both dairy and beef cattle (FAO et al., 2006). However, as mentioned earlier (see Section 5.0 on enteric fermentation), high levels of methane emissions from manure management are typically associated with high levels of productivity (FAO et al., 2006). Therefore, per unit of production, more efficient productions systems are superior in the reduction of GHG (Capper et al., 2009).

7. Livestock Related Land-Use Changes Forests cover approximately 4.1 109 ha of the Earth’s land area (Dixon et al., 1994) and are estimated to contain 80% of all above ground C and 40% of all below ground terrestrial C (Dixon et al., 1994) (Fig. 8).

21

Clearing the Air: Livestock’s Contribution to Climate Change

Land degradation in drylands Net loss of forest

Figure 8

Current forest cover Net gain of forest

National boundaries

Forest transition and land degradation in dry lands (FAO, 2006).

Russia and Brazil are home to the largest forested areas accounting for 21 and 10% of the total global forestland, respectively (Dixon et al., 1994). High and low latitude forests contain the largest C pools; hence changes (anthropogenic or nonanthropogenic) to specific forested areas can have a greater effect upon on C storage than other forested areas (Dixon et al., 1994). Land-use change is defined as greenhouse gas emissions from human activities which either change the way land is used (e.g., clearing of forests for agricultural use) or has an effect on the amount of biomass in existing biomass stocks (e.g., forests, village trees, woody savannas, etc.) (IPCC, 2000). From a livestock perspective, land-use changes would include any land adapted for livestock rearing (e.g., animal grazing, production of cropland for livestock feed). Forested areas are particularly sensitive to land-use change. When forest ecosystems undergo relatively abrupt land-use changes, such as deforestation, forest regrowth, biomass burning, wildfires, agriculture abandonment, wetland drainage, plowing, accelerated soil erosion, and so on, a significant loss of SOC and increase in GHG emissions occur (CAST, 2004; Dixon et al., 1994; Houghton et al., 1999). Using the IPCC’s definition of land-use change, livestock uses directly (i.e., pasture, LPS) and indirectly (i.e., production of feed crops) the largest land mass in the world (Bruinsma, 2003; Naylor et al., 2005) and is a primary driver for land-use change. LLS (FAO et al., 2006) estimated that livestock related land-use change produces 2400 Tg CO2-eq yr 1 or 35% of the total GHGs attributed to livestock. LLS (FAO et al., 2006) identifies

22

Maurice E. Pitesky et al.

deforestation in Latin America as the primary source of GHG emissions associated with global livestock. Specifically, land-use changes, including expansion of pasture and arable land for feed crops, primarily occur at the expense of forested land. Forest conversion for permanent crops, cattle ranching, cultivation shifts, and agriculture colonization are considered to contribute equally to the agriculturally driven land-use changes in these countries (Geist and Lambin, 2002). Smith et al. (2007b) estimates that over the last 40 years, an average of 6 and 7 Mha of forestland and nonforestland, respectively, was converted to agricultural land in the developing world. Houghton (2003) estimated that ‘‘Indonesia and Brazil accounted for approximately 50% of the global land-use change C flux in the 1990s.’’ While, LLS (FAO et al., 2006) assigned the largest portion of the GHG livestock portfolio to land-use changes, data from EPA et al. (2009) show that the United States overall actually increase forestland and that the nation’s forests sequester 1078 Tg CO2-eq yr 1 (EPA, 2009). Between 1990 and 2006, the forestland use in the United States increased by 25% from 244 to 304 million hectares (Alig et al., 2003; Smith et al., 2004), resulting in a net uptake in C through trees (EPA, 2009). This gross increase in C sequestration is thought to be related to increased forest area, improved, sustainable timbering (timber growth exceeding harvest), and abandonment of agricultural lands (Alig and Wear, 1992; Alig et al., 1998; Anderson and Magleby, 1997; Flather et al., 1999; Lubowski et al., 2008). LLS’s current LCA methodology (FAO et al., 2006) does not take into account increases in C sinks due to increased management of timberlands in regions like the United States. Forest regeneration, timberland management, and harvesting contribute positively to C sequestration and are highly managed through private landowners. Though harvesting trees as a resource remove much of the aboveground C, there is a positive growth rate of timberlands when it is harvested (Newell and Stavins, 2000). EPA (2009) established through modeling of forest growth that C sequestration is increased if trees are periodically harvested and allowed to regrow rather than maintained as permanently established. For the United States, forest regeneration and expansion is expected to continue and in contrast to some developing countries, deforestation is not a livestock related land-use issue. In conclusion, LLS (FAO et al., 2006) estimated net C losses associated with converting forested land to grasslands and croplands either directly (pasture) or indirectly through livestock feed production on a global scale. These global predictions result in a significant overestimation of GHG emissions from livestock in developed countries that have established land-use patterns since centuries.

Clearing the Air: Livestock’s Contribution to Climate Change

23

8. Livestock Induced Desertification Arid and semiarid ecosystems cover greater than 45% of the global land surface (Asner et al., 2003). The most common human agricultural activities on these lands are cattle and sheep grazing/ranching, wood collection, and cultivation (Asner et al., 2003). Desertification (a form of land degradation) primarily occurs in arid, semiarid, and dry subhumid grazing areas (pasture and rangeland) and causes a net loss of C to the atmosphere, ultimately leading to land with reduced biological productivity (Schlesinger et al., 1990). Desertification is generally caused by excessive grazing by livestock, fire, soil erosion, and salinization (Oba et al., 2008). LLS (FAO et al., 2006) estimated that global emissions associated with livestock induced desertification totals 100 Tg CO2-eq yr 1. These calculations are based upon studies that show a 25–80% decline in SOC in areas with long-term grazing (Asner et al., 2003). Desertification (i.e., land degradation of pasture) is mainly an issue in Africa (2.4 million km2), Asia (2.0 million km2), and Latin America (1.1 million km2) (FAO et al., 2006). The United Nations Environmental Program (UNEP) estimates that 35% of the world’s land surface is currently at risk for desertification and more than 20 million hectares are reduced annually to near or complete uselessness (Hellde´n, 1991). As mentioned above, nonanimal factors, such as soil erosion and geographical location (higher latitudes may have increased rates of decomposition of soil C), account for some of the SOC losses ( Jenkinson, 1991). However, animal factors (i.e., degradation of above ground vegetation) most likely have a more significant contribution to the nonrenewal of decaying organic matter stocks (Asner et al., 2003). Calculating the specific amount that livestock production is responsible for is difficult (FAO et al., 2006). However, livestock do occupy two thirds of the global arable dry land area and the rates of desertification are estimated to be higher in pasture than other land uses (Bruinsma, 2003). In the United States there are roughly 86 million hectares of federal land grazed by domestic livestock in 17 western states (Bock et al., 1993). Currently, the EPA does not have a desertification category in their inventory of United States GHG emissions and sinks. For the last 150 years, desertification and land degradation in the southwestern United States has led to significant land change (e.g., grassland to shrubland) and to some extend land degradation (Mueller et al., 2007). Historic overgrazing of livestock coupled with climate variation and altered fire regimens are considered some of the drivers of desertification in parts of the South West United States (Mueller et al., 2007; Yanoff and Muldavin, 2008). While grazing of grasslands is considered part of a ruminants natural history, not all grasslands have a symbiotic relationship with grazing ruminates (Bock et al., 1993). These grasslands that are ‘‘intolerant’’ of grazing

24

Maurice E. Pitesky et al.

animals are the most sensitive to desertification and hence SOC loss secondary to overgrazing (Bock et al., 1993). However, according to Loeser et al. (2007), areas like the semiarid grasslands of Northern Arizona that have been used at some intermediate level of cattle grazing may be ideal for grazing to maintain native plant diversity. Loeser et al. (2007) did not study C emissions or plant biomass (i.e., indicators of C flux); therefore, further research is required to study potential sequestration. However, the concept of livestock being an integral component of ecosystem health is important to recognize.

9. Release from Cultivated Soil Plowing and tilling coupled with wind, rain, and irrigation exacerbate soil erosion of cropland (Lal, 2004). Approximately 20–30% of SOC is mineralized and released into the atmosphere as CO2 (Lal, 1999). During the past 40 years, almost one third of the world’s cropland has been abandoned due to erosion and degradation (Wood et al., 2006). LLS (FAO et al., 2006) estimated that the loss of C from cultivated soils (i.e., tilling, liming, and emissions related to leguminous feed crops) associated with livestock totals 230 Tg CO2-eq yr 1. These estimates have a high degree of error based on environment, land management, and annual loss rate coefficients used under those conditions. As mentioned in the introduction, agricultural soil emissions include nonlivestock sources such as emissions associated with production of fruit, vegetables, fiber, grain, as well as livestock and grassland-based emissions. Direct and indirect emissions from agricultural soils related to synthetic N and manure utilization on agricultural soils account for 215 Tg CO2-eq yr 1 in the United States (EPA et al., 2009) and 9.1 Tg CO2-eq yr 1 in California (CEC, 2005). Within the agricultural soil management category, the primary sub-category of emissions for both the U.S. (28%) and California (55%) are emissions associated with synthetic fertilizers. The contribution of the livestock to soil emissions has not been determined in EPA et al. (2009). Figure 9 illustrates the sources and pathways of N that result in direct and indirect N2O emissions. Greenhouse gas emissions associated with cultivated soils are higher in the United States based on EPA et al. (2009) versus the global FAO et al. (2006) numbers due to several factors. Before the 2005 ‘‘Inventory of United States Greenhouse Gas Emissions’’ (EPA et al., 2009), GHG estimates within the agricultural sector were based on IPCC emission factors. However, the 2005 inventory includes N2O emissions using a combination of Tier I and Tier III process-based model (DAYCENT) approaches (Del Grosso et al., 2006). Among other differences, the DAYCENT model includes direct and indirect emissions from agricultural soils due to N additions to cropland and grassland and direct and indirect emissions form

25

Clearing the Air: Livestock’s Contribution to Climate Change

N volatilization N deposition to all soils and water bodies

Synthetic N fertilizers

Organic amendments

Urine and dung from grazing animals

Crop residues

N flows: N inputs to managed soils Direct N2O emissions N volatilization and deposition Indirect N2O emissions

Mineralization of soil organic matter

Asymbiotic fixation

Histosol cultivation

ch

ing

Run o ff

Su wa rfac te e r

Lea

This graphic illustrates the sources and pathways of nitrogen that result in direct and indirect N2O emissions from soils using the methodologies described in this inventory. Emission pathways are shown with arrows. On the lower right-hand side is a cut-away view of a representative section of a managed soil; histosol cultivation is represented here.

Ground water

Figure 9 Direct and indirect N2O emissions from agricultural soils. Sources and pathways of N that result in N2O emissions from agricultural soil management modified from (EPA et al., 2006).

soils due to the deposition of manure by livestock. In addition, the model is sensitive to inter-annual changes in temperature and management practices. Consequently, the DAYCENT model is considered a more accurate estimate of agricultural N2O emissions (EPA et al., 2009). In contrast, FAO et al. (2006) use IPCC Tier I calculations, which are primarily based on loss of C due to soil erosion.

26

Maurice E. Pitesky et al.

Cropland versus grassland account for approximately 71 and 29% of total direct anthropogenic GHG emissions from soils, respectively (EPA et al., 2009). Agronomic practices particularly tillage have a significant negative impact on N2O emissions and SOC losses (CAST, 2004). The N2O emissions are produced naturally in soils through the microbial processes of nitrification and denitrification (Khalil et al., 2004). Quantitatively the rate of N2O emissions from soil is highly dependent on several variables including rate of synthetic N-fertilizer application, organic manure application, presence/absence of crop residues, mineralization of soil matter, presence of N-fixing crops, irrigation, and tillage practices (Del Grosso et al., 2006). Consequently, it is important to understand and accurately characterize cropland at high resolution to calculate GHG emissions from cultivated soils. For example, estimates of CO2 emissions for United States corn, soybean, and wheat production vary from 79 kg C ha 1 yr 1 for no till soybean to 268 kg C ha 1 yr 1 for reduced till corn (CAST, 2004). Agricultural soils and vegetation both emit and sequester C. Therefore, mitigation strategies related to cropping practices are an area of interest. In 2000, the IPCC estimated that conservation tillage can sequester 0.1–1.3 tones C ha 1 yr 1 globally and could feasibly be adopted on up to 60% of arable lands. Currently the Kyoto protocols do not include C sinks in the emissions inventory for agriculture (Kyoto Protocol, 1997).

10. Carbon Emissions from Feed Production Historically, most of the resources utilized for livestock nutrition came from the farm itself. While this type of farming is still practiced in some parts of the developing world, most modern livestock operations require a variety of external inputs (i.e., feed production and transport, herbicides, pesticides, etc.) that directly or indirectly utilize fossil fuels and hence produce GHGs (Sainz, 2003). This increased utilization of external inputs allows for increased animal density or intensification of livestock production. In fact, more than half the energy expenditure during livestock production is for feed production (nearly all in the case of intensive beef operations) (FAO et al., 2006). LLS (FAO et al., 2006) estimated that fossil fuel use in manufacturing fertilizer used for animal feed plus emissions associated with application and indirect emissions emits approximately 240 Tg CO2eq yr 1 globally. Total GHG emissions for mineral fertilizer production are based on synthesis of 14 million tones of mineral fertilizer directly used for fertilization of cropland used solely for animal feed (FAO et al., 2006). The energetic cost of synthetic fertilizer synthesis is between 7 and 65 MJ kg 1 of N depending on the fertilizer type and mode of manufacturing (e.g., natural gas versus coal) (FAO et al., 2006). Lal (2004) compiled data

Clearing the Air: Livestock’s Contribution to Climate Change

27

estimating C emissions for production, transportation, storage, and transfer of various fertilizers between 0.03 and 1.8 kg CO2-eq kg 1. The EPA currently does not have a United States domestic value specifically for CO2 emissions from manufacturing of mineral fertilizer for livestock applications. The only United States numbers currently available are for CO2 emissions from ammonium manufacture and urea application (13.8 Tg CO2-eq yr 1) (EPA et al., 2009). Approximately 1% of the world’s net energy is utilized in making synthetic fertilizer (Smith, 2002). Carbon dioxide and N2O emissions are the GHGs associated with the indirect and direct use of fertilizers (Lal, 2004). The primary use of fertilizer in the animal food chain is for the production of corn (FAO et al., 2006). The average corn fertilizer application rate in the United States is 150 kg N ha 1 of corn (CAST, 2004). While N2O emissions occur naturally via nitrification and denitrification, the application of excess N increases the rate of N2O emissions (Bouwman, 1996). Different rotational farming systems that utilize N-fixing plants before planting corn do not seem to mitigate application rates. This may in part be due to the relatively cheap cost of mineral fertilizer coupled with a ‘‘more is better’’ approach. Moiser et al. (1996) estimated that worldwide application of N as synthetic fertilizer (77.4 Tg yr 1) is in the same range as that of N from manure (77.4 Tg yr 1). Synthetic fertilizers have reduced CH4 emissions (ammonium nitrate and ammonium sulfate appear to inhibit CH4 formation) relative to manure, while synthetic fertilizers have relatively higher N2O emissions. Therefore, attempts to reduce CH4 from manure sources may well increase other emissions including N2O (Mosier et al., 1996). In the United States, mineral fertilizers are the dominant form of crop N supplementation and many semi-developed areas of the world are quickly switching to this model. For example, in the United States in 2001, 10,800 Gg yr 1 of N from synthetic fertilizer was used versus 2950 Gg yr 1 from livestock manure applied (USDA, 2004). However, it is important to recognize that the synthetic fertilizer produced and utilized in the developed world in general has lower ammonia losses to the environment (4% compared to up to 30% depending on the type of fertilizer and conditions) than the mineral fertilizers used in the developing world (Bouwman, 1996). Although the use of manure leads to higher direct GHG emissions than mineral fertilizers (Khalil et al., 2008), data comparing net direct and indirect emissions was not incorporated into the LCA of LLS (FAO et al., 2006). In addition, as previously noted in Section 6.0, while LLS addresses the gross GHG emissions produced via production of mineral fertilizer, whereas the potential displacement of synthetic fertilizer production via the ‘‘free’’ production and usage of animal manure is not being discussed. This information will eventually have to be integrated into a more complex (and more accurate) LCA model that would account for the flow of energy from fossil fuels to N fertilizer, from N fertilizer to feed, and from feed to

28

Maurice E. Pitesky et al.

animal protein. Instead, only the emissions associated with N-fertilization of food animal crops (1 Tg CO2-eq yr 1 using a N application rate of ~150 kg ha 1 of corn) were assessed in LLS (FAO et al., 2006). Concentrates are a primary component of livestock feed, fed in the developed world. Concentrates comprise roughly 40% of all animal feed in the developed world versus 12% in the developing world (FAO et al., 2006). Overall, 32% of the world’s cereal production (the primary concentrate) is consumed by livestock (Bruinsma, 2003). The main crops utilized for feed production for livestock are corn (52% of concentrates), barley (19%), wheat (19%), and sorghum (5%) (Bruinsma, 2003; FAO et al., 2006). Within the United States, the state of California is unique from an animal nutrition perspective. The diversity of crops grown in California and their adaptability for both human and animal consumption allows the dairy industry to utilize cropland in a ‘‘dual’’ noncompetitive fashion. Crops, such as rice (rice hulls), almonds (almond hulls), and citrus fruits (citrus pulp) to name a few, have multiple uses for both humans and animals. This dual-utilization decreases the ‘‘footprint’’ of total cropland required for animal feed while integrating these ‘‘waste’’ products for animal feed (plant residues are rarely utilized as soil amendments in the developing world). Feeding crop by-products to livestock reduces decomposition of organic material and releases of GHG to the atmosphere. Instead of these ‘‘waste’’ products being underutilized and hence off-gassing methane as part of landfill or even as municipal solid waste (MSW) (Zhao et al., 2008), dairy cows are able to supplement their diet with these products. In this situation the net benefit of having ruminates needs to be further investigated and included in a California specific model. LLS (FAO et al., 2006) does not address emissions from production of pesticides, herbicides and other amendments commonly added to cropland. However, in intensive systems the combined-energy use for seed and herbicide/pesticide production and fossil fuel for machinery ‘‘generally’’ exceeds that for fertilizer production (Swanton et al., 1996). Lal (2004) conducted a comprehensive review of energy required for production, transportation, and storage of herbicides, insecticides, and fungicides. Means CO2-eq kg 1 for herbicides, insecticides, and pesticides were 6.3, 5.1, and 3.9, respectively, which were higher than all N-based fertilizers investigated (Lal, 2004). Estimates compiling C emissions for production, transportation, storage, and transfer of herbicides, insecticides, and fungicides had average equivalent C emissions higher than fertilizer (Lal, 2004). These numbers are complicated as some research shows that emission factors from production are superseded by net reduction in emissions on the cropland primarily due to no-till farming (Hisatomi et al., 2007). From a technology perspective it should also be noted that superior genetics and technology have made food animal nutrition more efficient from both a production and GHG perspective. For example, a study

Clearing the Air: Livestock’s Contribution to Climate Change

29

regarding bovine somatotropin (BST) hormone calculated that if all the dairy cows in the United States were using BST, the current milk supply could be reduced by 11% fewer cows, who would be fed 9% less feed, that would be produced on 6% less land ( Johnson and Johnson, 1995). These reductions translate to 6% less fossil fuel use and 9% less methane production (Capper et al., 2008). A recent study by Capper et al. (2009) using National Research Council (NRC) nutrient recommendations demonstrated that modern dairy practices in the United States in 2007 versus those in 1944 required 79% fewer animals, 78% less feedstuffs, 90% less land, and 65% less water to produce one billion kg of milk. In addition, the same study showed a 74% reduction in manure, 56% reduction in CH4, and 46% reduction in N2O per billion kg of milk produced in 2007 versus 1944 cows. In 1944, the United States dairy population totaled 25.6 million cows and produced 53 billion kg of milk annually (average milk yield per cow of 2074 kg yr 1) versus 9.2 million cows producing 84.2 billion kg of milk annually (average milk yield of 9193 kg yr 1) in 2007 (Capper et al., 2009). The authors attribute this dramatic increase in production to genetics, nutrition, and management. The average time needed to produce a broiler in the United States has gone from 72 days in 1960 to 48 days in 1995 with a 1.8–2.2 increase in slaughter weight and a 15% decrease in feed conversion ratios (kg feed per kg meat) (Naylor et al., 2005).

11. On-Farm Fossil Fuel Use: Diesel and Electricity On-farm fossil fuel use is highly dependent on the intensity and type of livestock production and the environment of the farm. Once on the farm, fossil fuels are utilized for tilling, irrigation, sowing, the movement of feed, for control of the environment (i.e., cooling, heating, and/or ventilation), for animal waste collection and treatment (i.e., land application, solid separation), and for transportation of products ( Johnson and Johnson, 1995; Lal, 2004; Sainz, 2003). LLS (FAO et al., 2006) estimated that on-farm fossil fuel use emits 90 Tg CO2-eq yr 1. Equivalent estimated for the United States do not exist currently for CO2 emissions from on-farm fossil fuel use. However, in an intensive system, on-farm use of fossil fuel often produces greater GHG emissions than those from chemical N fertilizer (Sainz, 2003). For the assessment of global on-farm fossil fuel use associated with livestock production, LLS (FAO et al., 2006) utilizes a single study by Ryan and Tiffany (1998). FAO et al. (2006) then extrapolates intensive farming globally and adjusts based on latitude (e.g., at lower latitudes less energy would be required for corn drying). Specifically, LLS focuses at on-

30

Maurice E. Pitesky et al.

farm energy use for nine different commodities (corn, soybeans, wheat, dairy, swine, beef, turkeys, sugar beets, and sweet corn/peas). The study identifies diesel or liquefied petroleum gas (LPG) as the primary source of energy for on-farm energy use for eight of the nine commodities. Overall, predictions fossil fuel use associated with livestock production are weak globally and nationally. However, studies in the United States and France have shown a decrease in energy use dedicated to agriculture since 1980 (Bonny, 1993; Cleveland, 1995).

12. Postharvest: CO2 from Livestock Processing The postharvest system includes processing, distribution (transport and storage), and preparation. LLS (FAO et al., 2006) estimated United States emissions between 10–50 Tg CO2-eq yr 1 based on research done in Minnesota (Ryan and Tiffany, 1998). The EPA et al. (2006) report does not address postharvest emissions. Figure 10 shows energy use by energy source on United States farms between 1965 and 2001. While total energy use has leveled since 1990, output per unit of energy input has increased significantly (USDA, 2004). In addition, the adoption of no-till land management has the secondary benefit of decreasing fuel use on farms. While postharvest CO2 relative to the other categories listed is not a major emitter of GHG, the wide range of data available creates some uncertainty. This uncertainty is primarily related to the myriad of valueadded food animal by-products combined with multiple food processing technologies. For example, for a simple product such as processed beef, the energetic cost ranges between 0.84 and 5.02 MJ kg 1 live weight (Ward et al., 1977). In addition, differences in types of energy used for electricity (hydroelectric versus coal) affects on the GHG output. From an energy perspective, depending on the efficiency and the product, agriculture represents between 20 and 50% of the energy consumed within the food supply chain (Wood et al., 2006). For example, the state of California’s energy portfolio will change based on implementation of Assembly Bill 32 (AB-32). Specifically, by 2020, by law the state of California can only produce 1990 levels of anthropogenic GHGs (California Environmental Protection Agency, 2007). In order to achieve this cap, one third of California’s energy portfolio will be renewable compared to roughly 10% currently (California Environmental Protection Agency, 2007). Therefore, while postharvest emissions are a relatively low proportion of total livestock emissions, regional differences in emissions factors are expected.

31

Clearing the Air: Livestock’s Contribution to Climate Change

Quadrillion btu 1.8 1.6 1.4 1.2 1 Natural gas LP gas 0.8

Electricity Diesel

0.6 0.4 0.2

Gasoline

0 1965

1970

1975

1980

1985

1990

1995

2000

Figure 10 Energy use by agriculture by source 1965–2001 (USDA, 2004).

Postharvest emissions associated with animal feed production and processing of non-food related animal products were not included in FAO et al. (2006).

12.1. Transportation The GHG emissions associated with the transport of animal products (‘‘farm-to-fork’’) vary according to mode (truck, rail, water) of transport and type of animal product. Previous studies have shown barge to be over eight times more energy efficient than truck and twice as efficient as rail (Rose, 2006). However, these values do not take into account emissions associated with refrigeration for perishable items. LLS (FAO et al., 2006) estimated CO2 emissions from transport of livestock products to be 0.9 Tg CO2-eq yr 1. The EPA currently does not measure CO2 emissions associated with respect to livestock in the United States. 1. When calculating GHG production on a national or regional level, production areas are ‘‘assessed’’ emissions while the receiver region is not assessed any ‘‘emissions.’’ These ‘‘virtual’’ emissions are ‘‘tallied’’ solely for the producer and not the consumer. There have been estimates that China’s total GHG ‘‘footprint’’ would be reduced by 1/3 if emissions based on usage were calculated instead of emissions based on production (FAO et al., 2006). Likewise, while total CH4 emissions

32

Maurice E. Pitesky et al.

from enteric fermentation for Central and South America are approximately one-quarter (486 Tg CO2-eq yr 1) of the global livestock CH4, only Central and South America is identified (Steinfeld and Wassenaar, 2007). In a study by Galloway et al. (2007), Japan’s pig and chicken consumption resulted in the equivalent usage of 50% of Japan’s total arable land (Galloway et al., 2007), while trade between Brazil and China is responsible for 15% of the virtual N left behind in Brazil and 20% of Brazil’s area to grow soy (Galloway et al., 2007). 2. Reducing ‘‘farm-to-fork’’ transportation emissions does not necessarily reduce GHG emissions from an LCA perspective. If an animal product can be produced internationally in such a way that gross GHG emissions are lower than that same animal product produced under local conditions, then consumption of the product with the shorter ‘‘farm-to-fork’’ distance may in fact have a greater GHG footprint. The point being that the proper integration of international food trade can potentially play an integral role in mitigation of global GHG emissions. While issues related to ‘‘food security’’ encourage local sources of food, the balance between productions for domestic consumption (and food security) needs to be balanced with the ‘‘outsourcing’’ of food production for GHG mitigation. Within the transportation sector, it is important to discuss the energy and environmental impact of ‘‘farm-to-fork’’ costs for food animals. In the developed world, food animals are often concentrated in landless systems, the transportation of feed grains and other feedstuffs often involves a massive transfer of nutrients between regions. Currently, the EU gets the majority of their soybeans for animal feed from Brazil (Smaling et al., 2008). In the United States, pig operations in the Southeast get the majority of their grain from farms in the mid-west (USDA, 2004). LLS does not assess the transportation or potential environmental costs of these food animal farm-to-fork costs.

12.2. Waste and biomass To complete a LCA analysis, the waste/use ratio should be determined. Neither FAO et al. (2006) nor EPA et al. (2006) addresses GHG production due to waste (for animal feed and food animal produced for human consumption). The EPA estimates that 3.6 Tg CO2-eq yr 1 are produced from processing of both meat and poultry from CH4 emissions associated with industrial waste water (typically anaerobic lagoons) (EPA, 2009). When incorporating these numbers into the United States, the EPA estimates total GHG of the United States agricultural sector to increase from 413 to 417 Tg CO2-eq yr 1. The authors were unable to find a specific national or global data on food waste directly related to livestock. However, a study by the USDA Economic Research Service estimated that 2.45 billion kg of edible food at the

Clearing the Air: Livestock’s Contribution to Climate Change

33

retail level and 41.36 billion kg at the consumer and foodservice are lost annually accounting for 26% of the total edible food supply (Kantor et al., 1997). This does not include preharvest, on-the-farm, and farm-to-retail losses. Nearly half of the retail losses came from perishable items such as fluid milk and other dairy products and fresh fruits and vegetables (Heller and Keoleian, 2000). No estimates are given on ‘‘wasted’’ GHGs produced during production of food that was never consumed. In addition, no data was found on waste streams for food products with a livestock component.

13. Conclusions With global meat production projected to more than double the current rate by 2050 (Smith et al., 2007b) and the majority of this livestock production growth occurring in the developing world (Wood et al., 2006), assessment of the holistic impacts of food animals in the context of global and regional environmental policy and food security becomes imperative. Much of the growth in the global livestock sector will occur in areas that are currently forested (i.e., parts of South America and South East Asia). It has been well established that significant reductions of carbon sequestering forests will have large effects on global climate change. LLS (FAO et al., 2006) has been most instrumental in pointing the public attention to the kinds of environmental consequences in which livestock production can potentially result, with special emphasis on climate change. Unfortunately, some of the report’s key conclusions (i.e., livestock produces more GHG than transportation) have been applied regionally and out of their intended context, leading to significant consequences on major public policy affairs. For example, the statement that 18% of anthropogenic global GHGs is caused by livestock production and that livestock produces more GHG than transportation (FAO, 2007) is based on inappropriate or inaccurate scaling of predictions, and thus is open to intensive debate throughout the scientific community. Livestock production in most countries of the developed world (e.g., United States and Europe) has a relatively small GHG contribution within the overall carbon portfolios, dwarfed by large transportation, energy, and other industry sectors. In contrast, livestock production in the developing world can be a dominant contributor to a country’s GHG portfolio, due to the developing world’s significantly smaller transportation and energy sectors. In the United States, transportation accounts for at least 26% of total anthropogenic GHG emissions compared to roughly 5.8% for all of agriculture, which includes less than 3% associated with livestock production. However, in countries like Paraguay, the trend is likely reversed because of

34

Maurice E. Pitesky et al.

Paraguay’s much smaller transportation and energy sectors, and a relatively large livestock sector, which might contribute to more than 50% of that county’s carbon footprint. The fact that land-use changes associated with livestock (i.e., forested land converted to pasture or cropland used for feed production) are a significant source of anthropogenic GHGs in Latin America and other parts of the developing world is apparent. However, it is likely that any kind of land-use change from the original forestland will lead to great increases in global warming. LLS (FAO et al., 2006) attributes almost half of the climate-change impact associated with livestock to the change of land-use patterns. Latin America has the greatest pool of ‘‘unused but suitable’’ land that is currently covered by forests but could be turned into agricultural crop or livestock production (Bruinsma, 2003). In 2000, Latin America had 203 million hectares arable land in use and 863 million hectares of unused land suitable for cropland (19% in use) (Bruinsma, 2003). Over the same time span, developed countries had 387 million hectares arable land in use and 487 million hectares of unused land suitable for cropland and livestock (44% in use) (Bruinsma, 2003). Transformation of land from forest to agriculture has occurred in the developed countries centuries ago to make way for industrialization and general societal wealth. Not surprisingly, numerous developing countries are currently attempting to develop their economies by turning economically marginal land into production. The United States and most other developed countries have not experienced significant land-use change practices around livestock production within the last few decades. Instead, over the last 25 years forestland has increased by approximately 25% in the United States and livestock production has been intensified (concentrated geographically), thus reducing its geographical footprint. Modern livestock production has experienced a marked improvement of efficiencies, leading to significantly decreased numbers of animals to produce a given amount product that satisfies the nutritional demands by society (Capper et al., 2009). According to LLS, intensification of livestock production provides large opportunities for climate change mitigation and can reduce greenhouse gas emissions from deforestation, thus becoming a long-term solution to a more sustainable livestock production. When comparing GHG portfolio sectors such as livestock versus transportation, comparable assessment tools should be used. For example, the transportation figures used in LLS are ‘‘direct emissions’’ associated mainly with combustion during transportation and do not include indirect emissions associated with the transportation or oil industries (i.e., manufacturing of vehicles, resource extraction, etc.). On the other hand, the report assesses livestock holistically from a direct and indirect perspective. A comparison between livestock production versus transportation, with one (livestock) assessment based on a complex LCA and the other (transportation) without LCA, is generally questionable.

Clearing the Air: Livestock’s Contribution to Climate Change

35

Comparing LLS (FAO et al., 2006) with several regional reports (CEC, 2005; EPA et al., 2006) shows large agreement with respect to emission predictions from most livestock related categories. There is general consensus that as a direct GHG category, enteric fermentation in ruminants and manure management are the most important categories within livestock production. Categories like on-farm fuel use or feed production are dwarfed by emissions coming from the animals and their manure. Many investigators use the international standard (ISO 14040) for LCAs that are often rigid, impractical, and not sufficiently transparent. One means of improvement would be the use of a ‘‘numerical suffix system’’ indicating the ‘‘degrees of separation’’ between the product (e.g., animal protein) and the indirect emissions source input (i.e., the greater suffix number, the more complete the LCA). Furthermore, all current and future assessments of GHG impacts should include mass-balance accounting of energy per GHG unit basis to assess the true environmental impact of direct and indirect emissions. Examples include GHGs associated with displaced fertilizer production through use of animal manure. LLS does not currently account for fertilizer that is not produced because animal manure is present. LLS (FAO et al., 2006) does not account for ‘‘default’’ emissions. Specifically, if domesticated livestock were reduced or even eliminated, the question of what ‘‘substitute’’ GHGs world be produced in their place has never been estimated. While never explicitly stated in any publication, the idea that if livestock were simply eliminated, 18% of anthropogenic GHGs would also be eliminated as well, is unrealistic. In fact, many of the resources previously dedicated to domesticated livestock would be utilized by other human activities, many of which produce much greater climate change impacts. It is also important to realize that livestock provides not only meat, dairy products and eggs, but also wool, hides, and many other value-added goods and services. Livestock are often closely integrated into mixed and some landless (e.g., landless dairy) farming systems as consumers of crop by-products and sources of organic fertilizer, while larger animals also provide power for plowing and transport. Therefore, to estimate accurately the ‘‘footprint’’ of all livestock, ‘‘default’’ emissions for nonlivestock substitutes need to be estimated and compared to livestock emissions (e.g., manure versus fertilizer, leather versus vinyl, wool versus microfiber, etc.). The net GHG differences between livestock and other land-use forms can then be used to estimate a more accurate GHG ‘‘footprint’’ of livestock’s impact.

ACKNOWLEDGMENT The authors would like to thank Veronica Arteaga for her assistance in reference compilation and editing.

36

Maurice E. Pitesky et al.

REFERENCES AgStar. (2002). Managing manure with biogas recovery systems improved performance at competitive costs.Environmental Protection Agency. 430-F02-004 http://www.epa. gov/agstar/pdf/manage.pdf. Alig, R., and Wear, D. (1992). Changes in private timberland: Statistics and projections for 1952 to 2040. J. For. 90, 31–37. Alig, R., Adams, D., and McCarl, B. (1998). Impacts of incorporating land exchange between forestry and agriculture in sector models. Environ. Resour. Econ. 9, 259–274. Alig, R. J., Plantinga, A. J., Ahn, S., and Kline, J. D. (2003). Land use changes involving forestry in the United States: 1952 to 1997, with projections to 2050. ‘‘US Forest Service Pacific Northwest Research Station General Technical Report PNW-GTR, I-92’’. Anderson, M., and Magleby, R. (1997). Agricultural Resources and Environmental Indicators, 1996–97. United States Department of Agriculture Economic Research Service, Washington, DC. Ash, N., and Scholes, R. (2005). Ecoysystems and human well-being: Current state and trends: Findings of the condition and trends working group of the millennium ecosystem assessment, Millennium Ecosystem Assessment. World Resources Institute, Washington D.C http:// www.scribd.com/doc/5250332/millennium-ecosystem-assessment-2005. Asner, G. P., Borghi, C. E., and Ojeda, R. A. (2003). Desertification in central Argentina: Changes in ecosystem carbon and nitrogen from imaging spectroscopy. Ecol. Appl. 13, 629–648. Boadi, D., Benchaar, C., Chiquette, J., and Masse, D. (2004). Mitigation strategies to reduce enteric methane emissions from dairy cows: Update review. Can. J. Anim. Sci. 84, 319–335. Bock, C. E., Bock, J. H., and Smith, H. M. (1993). Proposal for a system of federal livestock exclosures on public rangelands in the western United States. Conserv. Biol. 7, 731–733. Bonny, S. (1993). Is agriculture using more and more energy—A French case study. Agric. Syst. 43, 51–66. Bouwman, A. (1996). Direct emissions of nitrous oxide from agricultural soils. Nutr. Cycl. Agroecosyst. 46, 53–70. Breitenbeck, G., Blackmer, A., and Bremner, J. (1980). Effects of different nitrogen fertilizers on nitrous oxide from soil. Geophys. Res. Lett. 7, 85–88. Bremner, J., Breitenbeck, G., and Blackmer, A. (1981). Effects of anhydrous ammonia fertilization on emission of nitrous oxide from soils. J. Environ. Qual. 10, 77–80. Bruinsma, J. (2003). World Agriculture: Towards 2015/2030, an FAO Perspective. Earthscan, London, UK. California Environmental Protection Agency. (2007). California 1990 Greenhouse Gas Emissions Level and 2020 Emissions Limit.Air Resources Board www.ARB.ca.gov/ cc/inventory/pubs/reports/staff_report_1990_level.pdf. Capper, J. L., Castaneda-Gutierrez, E., Cady, R. A., and Bauman, D. E. (2008). The environmental impact of recombinant bovine somatotropin (rbST)use in dairy production. PNAS 105, 9668–9673. Capper, J. L., Cady, R. A., and Bauman, D. E. (2009). The environmental impact of dairy production: 1944 compared with 2007. J. Anim. Sci. 1910 10.2527/jas.2009-1781. Casey, K., Bicudo, J., Schmidt, D., Singh, A., Gay, S., Gates, R., Jacobson, L., and Hoff, S. (2006). Air quality and emissions from livestock and poultry production/waste management systems, 40 pp. Animal Agriculture and the Environment: National Center for Manure and Waste Management White Papers. CAST (2004). Climate change and greenhouse gas mitigation: Challenges and opportunities for agriculture. Council for Agricultural Science and Technology. Task Force Report No 141.

Clearing the Air: Livestock’s Contribution to Climate Change

37

Castillo, A. R., Kebreab, E., Beever, D. E., Barbi, J. H., Sutton, J. D., Kirby, H. C., and France, J. (2001). The effect of energy supplementation on nitrogen utilization in lactating dairy cows fed grass silage diets. J. Anim. Sci. 79, 240–246. CEC. (2005). Inventory of California Greenhouse Gas Emissions and Sinks: 1990 to 2002 Update.Air Resources Board http://www.climatechange.ca.gov/inventory/index.html. Chen, D., Li, Y., Grace, P., and Mosier, A. R. (2008). N2O emissions from agricultural lands: A synthesis of simulation approaches. Plant Soil 309, 169–189. Clemens, J., and Ahlgrimm, H. J. (2001). Greenhouse gases from animal husbandry: Mitigation options. Nutr. Cycl. Agroecosyst. 60, 287–300. Clemens, J., and Huschka, A. (2001). The effect of biological oxygen demand of cattle slurry and soil moisture on nitrous oxide emissions. Nutr. Cycl. Agroecosyst. 59, 193–198. Cleveland, C. J. (1995). The direct and indirect use of fossil fuels and electricity in USA agriculture, 1910–1990. Agric. Ecosyst. Environ. 55, 111–121. Crutzen, P., Asselmann, I., and Seiler, W. (1986). Methane production by domesticated animals, wild ruminants, other herbivorous fauna and humans. Tellus Series B Chem. Phys. Meteorol. 38, 271. Dash, S. S., and Misra, M. K. (2001). Studies on hill agro-ecosystems of three tribal villages on the Eastern Ghats of Orissa, India. Agric. Ecosyst. Environ. 86, 287–302. De Gryze, S., Catala, R.E., Howitt, R.E., and Six, J. (2008) An integrated assessment of the biophysical and economic potential for greenhouse gas mitigation in California agricultural soils. California Energy Commission. PIER Energy-related environmental research. Del Grosso, S. J., Parton, W. J., Mosier, A. R., Walsh, M. K., Ojima, D. S., and Thornton, P. E. (2006). DAYCENT national-scale simulations of nitrous oxide emissions from cropped soils in the United States. J. Environ. Qual. 35, 1451–1460. Dixon, R. K., Brown, S., Houghton, R. A., Solomon, A. M., Trexler, M. C., and Wisniewski, J. (1994). Carbon pools and flux of global forest ecosystems. Science 263, 185–190. Ellis, J. L., Kebreab, E., Odongo, N. E., Beauchemin, K., McGinn, S., Nkrumah, J. D., Moore, S. S., Christopherson, R., Murdoch, G. K., McBride, B. W., Okine, E. K., and France, J. (2009). Modeling methane production from beef cattle using linear and nonlinear approaches. J. Anim. Sci. 87, 1334–1345. EPA. Hockstad, L., and Weitz, M. (2009). Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990–2007. Environmental Protection Agency, Washington DC; 430-R-09-004. EPA.Holtkamp, J., Hayano, D., Irvine, A., John, G., Munds-Dry, O., Newland, T., Snodgrass, S., and Williams, M. (2006). Inventory of U.S. greenhouse gases and sinks: 1996–2006. Environmental Protection Agency http://www.epa.gov/climatechange/ emissions/downloads/08_Annex_1-7.pdf. FAO.Steinfeld, H., Gerber, P., Wassenaar, T., Castel, V., Rosales, M., and de Haan, C. (2006). Livestock’s Long Shadow. Food and Agriculture Organization of the United Nations. http://www.fao.org/docrep/010/a0701e/a0701e00.HTM. Flather, C., Brady, S., and Knowles, M. (1999). Wildlife resource trends in the United States: A technical document supporting the 2000 USDA forest service RPA assessment. United States Department of Agriculture, Forest Service, Fort Collins Co. Gen. Tech. Rep. RMRS-GTR-33. Rocky Mountain Research Station. Fox News., and Kroll, A. (2009). A Tearful, Reluctant Farewell to My Favorite Food: Meat. Fox News http://green.foxnews.com/2009/03/18/a-tearful-reluctant-farewell-to-myfavorite-food-meat/. Galloway, J. N., Burke, M., Bradford, G. E., Naylor, R., Falcon, W., Chapagain, A. K., Gaskell, J. C., McCullough, E., Mooney, H. A., Oleson, K. L. L., Steinfeld, H., Wassenaar, T., et al. (2007). International trade in meat: The tip of the pork chop. Ambio 36, 622–629. Geist, H. J., and Lambin, E. F. (2002). Proximated causes and underlying driving forces of tropical deforestation. Bioscience 52(2), 143–150.

38

Maurice E. Pitesky et al.

Halberg, N., Van Der Werf, H., Basset-Mens, C., Dalgaard, R., and de Boer, I. (2005). Environmental assessment tools for the evaluation and improvement of European livestock production systems. Livest. Prod. Sci. 96, 33–50. Hellde´n, U. (1991). Desertification: Time for an assessment? Forestry and the Environ. 20, 372–382. Heller, M. C., and Keoleian, G. A. (2000). Life Cycle-Based Sustainability Indicators for Assessment of the U.S. Food System. Center for Sustainable Systems, University of Michigan; Ann Arbor. Hisatomi, H., Hitomi, K., and Hayato, S. (2007). Reduction in greenhouse gas emissions by no-tilling rice cultivation in Hachirogata polder, northern Japan: Life-cycle inventory analysis. Soil Sci. Plant Nutr. 53, 668–677. Hoekstra, N. J., Schulte, R. P. O., Struik, P. C., and Lantinga, E. A. (2007). Pathways to improving the N efficiency of grazing bovines. Eur. J. Agron. 26, 363–374. Holter, J. B., and Young, A. J. (1992). Methane prediction in dry and lactating Holstein cows. J. Dairy Sci. 75, 2165–2175. Houghton, R. A. (2003). Revised estimates of the annual net flux of carbon to the atmosphere from changes on land use and land management. Tellus 55B, 378–390. Houghton, R. A., Hackler, J. L., and Lawrence, K. T. (1999). The U.S. carbon budget: Contributions from land-use change. Science 285, 574–578. International Organization for Standardization (2006). ISO 14040:2006. (FAO, Ed.). http://www.iso.org/iso/iso_catalogue/catalogue_tc/catalogue_detail.htm?csnumber=37456 (accessed 2 April 2009). IPCC. (1997). Intergovernmental panel on climate change/organization for economic cooperation and development: Guidelines for national greehhouse gas inventories, Paris. http://www.ipcc-nggip.iges.or.jp/public/mtdocs/pdfiles/rockhamp.pdf. IPCC. (2000). Good Practice Guidance and Uncertainty Management in National Greenhouse Gas Inventories. Intergovernmental Panel on Climate Change (IPCC), IPCC/OECD/IEA/ IGES, Hayama, Japan. IPCC. (2007). Fourth Assessment Report (AR4). Intergovernmental Panel on Climate Change (IPCC), http://ipcc-wg1.ucar.edu/index.html. Jarecki, M., Parkin, T., Chan, A., Hatfield, J., and Jones, R. (2008). Greenhouse gas emissions from two soils receiving nitrogen fertilizer and swine manure slurry. J. Environ. Qual. 37, 1432–1438. Jenkinson, D. (1991). The Rothamsted long-term experiments: Are they still of use? J. Agron. 83, 2–12. Johnson, K., and Johnson, D. (1995). Methane emissions from cattle. J. Anim. Sci. 73, 2483–2492. Jungbluth, T., Hartung, E., and Brose, G. (2001). Greenhouse gas emissions from animal houses and manure stores. Nutr. Cycl. Agroecosyst. 60, 133–145. Kantor, L. S., Lipton, K., Manchester, A., and Oliveira, V. (1997). Estimating and addressing America’s food loses. United States Department of Agriculture. Food Rev. 2–12. Kaspar, H. F., and Tiedje, J. M. (1981). Dissimilatory reduction of nitrate and nitrite in the bovine rumen nitrous oxide production and effect of acetylene. Appl. Environ. Microbiol. 41, 705–709. Kebreab, E., Johnson, K. A., Archibeque, S. L., Pape, D., and Wirth, T. (2008). Model for estimating enteric methane emissions from US dairy and feedlot cattle. J. Anim. Sci. 86, 2738–2748. Khalil, K., Mary, B., and Renault, P. (2004). Nitrous oxide production by nitrification and denitrification in soil aggregates as affected by O-2 concentration. Soil Biol. Biochem. 36, 687–699. Khalil, M. A. K., Shearer, M. J., Rasmussen, R. A., Duan, C. L., and Ren, L. X. (2008). Production, oxidation, and emissions of methane from rice fields in China. J. Geophys. Res. Biogeosci. 113, 12.

Clearing the Air: Livestock’s Contribution to Climate Change

39

Kyoto Protocol. (1997). United Nations Framework Convention on Climate Change. Kyoto Protocol, Kyoto. Lal, R. (1999). Soil management and restoration for C sequestration to mitigate the accelerated greenhouse effect. Progr. Environ. Sci. 1, 307–326. Lal, R. (2004). Carbon emission from farm operations. Environ. Int. 30, 981–990. LA Times, E. (2007). A Warming World; Pollution on the Hoof; Livestock Emissions Are a Leading Source of Greenhouse Gases. One Solution May Be to Eat Less Meat. LA Times, Los Angeles. Loeser, M. R. R., Sisk, T. D., and Crews, T. E. (2007). Impact of grazing intensity during drought in an Arizona grassland. Conserv. Biol. 21, 87–97. Lubowski, R., Plantinga, A., and Stavins, R. (2008). What drives land-use change in the United States? Land Econ. 84, 529–550. Monteny, G. J., Groenestein, C. M., and Hilhorst, M. A. (2001). Interactions and coupling between emissions of methane and nitrous oxide from animal husbandry. Nutr. Cycl. Agroecosyst. 60, 123–132. Mosier, A., Duxbury, J., Freney, J., Heinemeyer, O., and Minami, K. (1996). Nitrous oxide emissions from agricultural fields: Assessment, measurement and mitigation. Dev. Plant Soil Sci. 68, 589–602. Mosier, A., Kroeze, C., Nevison, C., Oenema, O., Seitzinger, S., and Van Cleemput, O. (1998a). Closing the global N2O budget: Nitrous oxide emissions through the agricultural nitrogen cycle: OEDC/IPCC/IEA phase II development of IPCC guideline for national greenhouse gas methodology. Nutr. Cycl. Agroecosyst. 52, 225–248. Mosier, A. R., Duxbury, J. M., Freney, J. R., Heinemeyer, O., Minami, K., and Johnson, D. (1998b). Mitigating agricultural emissions of methane. Clim. Change 40, 39–80. Mueller, E. N., Wainwright, J., and Parsons, A. J. (2007). The stability of vegetation boundaries and the propagation of desertification in the American Southwest: A modeling approach. Ecol. Model. 208, 91–101. NASS. (2009). USDA: National Agricultural Statistics Service: NASS, Washington, DC. Naylor, R., Steinfeld, H., Falcon, W., Galloway, J. N., Smil, V., Bradford, G. E., and Mooney, H. A. (2005). Losing the links between livestock and land. Science 310, 1621–1622. Newell, R., and Stavins, R. (2000). Climate change and forest sinks: Factors affecting the costs of carbon sequestration. J. Environ. Qual. 40, 211–235. NRC. (2003). Air Emissions from Animal Feeding Operations. National Academy Press, Washington DC. NY Times, Op-ed. (2009). Meat and the Planet. NY Times, New York City. Oba, G., Sjaastad, E., and Roba, H. G. (2008). Framework for participatory assessments and implementation of global environmental conventions at the community level. Land Degrad. Dev. 19, 65–76. Phetteplace, H. W., Johnson, D. E., and Seidl, A. F. (2001). Greenhouse gas emissions from simulated beef and dairy livestock systems in the United States. Nutr. Cycl. Agroecosyst. 60, 99–102. Rose, M. H., Seely, B. E., and Barrett, P. F. (2006). The best transportation system in the world: Railroads, trucks, airlines, and American public policy in the twentieth century. Ohio State University Press, Ohio0814210368. Ryan, B., and Tiffany, D. G. (1998). Energy use in Minnesota agriculture. Minn. Agric. Econ. No. 693, Fall 2008. Saggar, S., Bolan, N. S., Bhandral, R., Hedley, C. B., and Luo, J. (2004). A review of emissions of methane, ammonia, and nitrous oxide from animal excreta deposition and farm effluent application in grazed pastures. NZ J. Agric. Res. 47, 513–544. Sainz, R. (2003). Framework for calculating fossil fuel use in livestock systems. http://www. fao.org/wairdocs/lead/x6100e/x6100e00.htm#Contents.

40

Maurice E. Pitesky et al.

Schlesinger, W. H., Reynolds, J. F., Cunningham, G. L., Huenneke, L. F., Jarrell, W. M., Virginia, R. A., and Whitford, W. G. (1990). Biological feedbacks in global desertification. Science 247, 1043–1048. Smaling, E. M. A., Roscoe, R., Lesschen, J. P., Bouwman, A. F., and Comunello, E. (2008). From forest to waste: Assessment of the Brazilian soybean chain, using nitrogen as a marker. Agric. Ecosyst. Environ. 128, 185–197. Smith, B. E. (2002). Nitrogenase reveals its inner secrets. Science 297, 1654–1655. Smith, W., Miles, P., Vissage, J., and Pugh, S. (2004). Forest Resources of the United States, 2002. USDA, Forest Service North Central Research Station, St. Paul, MN. Smith, P., Martino, D., Cai, Z., Gwary, D., Janzen, H., Kumar, P., McCarl, B., Ogle, S., O’Mara, F., Rice, C., Scholes, B., and Sirotenko, O. (2007a). Agriculture in Climate Change 2007: Mitigation, Contribution of Working Group III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, IPCC, Paris. Smith, P., Martino, D., Cai, Z. C., Gwary, D., Janzen, H., Kumar, P., McCarl, B., Ogle, S., O’Mara, F., Rice, C., Scholes, B., Sirotenko, O., et al. (2007b). Policy and technological constraints to implementation of greenhouse gas mitigation options in agriculture. Agric. Ecosyst. Environ. 118, 6–28. Steinfeld, H., and Wassenaar, T. (2007) The role of livestock production in carbon and nitrogen cycles. Annual Rev. Environ. Res. 32, 271–294. Sun, H., Trabue, S. L., Scoggin, K., Jackson, W. A., Pan, Y., Zhao, Y., Malkina, I. L., Koziel, J. A., and Mitioehner, F. M. (2008). Alcohol, volatile fatty acid, phenol, and methane emissions from dairy cows and fresh manure. J. Environ. Qual. 37, 615–622. Swanton, C. J., Murphy, S. D., Hume, D. J., and Clements, D. R. (1996). Recent improvements in the energy efficiency of agriculture: Case studies from Ontario, Canada. Agric. Syst. 52, 399–418. USDA. (2004). U.S. Agriculture and Forestry Greenhouse Gas Inventory: 1990–2001.Global Change Program Office, Office of the Chief Economist, USA. Verburg, P. H., and Van der Gon, H. (2001). Spatial and temporal dynamics of methane emissions from agricultural sources in China. Global Change Biol. 7, 31–47. Ward, G., Knox, P., and Hobson, B. (1977). Beef production options and requirements for fossil fuel. Science 198, 265–271. Wood, R., Lenzen, M., Dey, C., and Lundie, S. (2006). A comparative study of some environmental impacts of conventional and organic farming in Australia. Agric. Syst. 89, 324–348. Wright, A. D. G., Kennedy, P., O’Neill, C. J., Toovey, A. F., Popovski, S., Rea, S. M., Pimm, C. L., and Klein, L. (2004). Reducing methane emissions in sheep by immunization against rumen methanogens. Vaccine 22, 3976–3985. Yanoff, S., and Muldavin, E. (2008). Grassland-shrubland transformation and grazing: A century-scale view of a northern Chihuahuan Desert grassland. J. Arid Environ. 72, 1594–1605. Zhao, X., Musleh, R., Maher, S., Khire, M. V., Vpoce, T. C., and Hashsham, S. A. (2008). Start-up performance of a full-scale bioreactor landfill cell under cold-climate conditions. Waste Manag. 28, 2623–2634.

C H A P T E R

T W O

Improvement of Drought Resistance in Rice R. Serraj,* A. Kumar,† K. L. McNally,‡ I. Slamet-Loedin,† R. Bruskiewich,§ R. Mauleon,§ J. Cairns,*,k and R. J. Hijmans},k Contents 1. Introduction 2. Drought Characterization 2.1. Drought-occurrence parameters 2.2. Distribution of rice production systems and rainfall 2.3. Drought-prone target environments 2.4. Systems analysis and characterization 3. Rice Responses to Drought 3.1. Plant water use 3.2. Spikelet sterility and grain failure 3.3. Dehydration-avoidance mechanisms 4. Concepts and Tools for Phenotyping 4.1. Field-managed drought screening 4.2. The FTSW dry-down approach 4.3. Control and monitoring of soil moisture profiles 4.4. Nondestructive methods for plant growth and water status 4.5. Model-based phenotyping 4.6. Phenotyping for gene expression and profiling 5. Conventional Breeding 5.1. Screening of rice germplasm 5.2. Heritability of yield components 5.3. Direct selection for yield under drought

42 44 45 45 46 48 49 49 50 52 57 57 59 60 61 62 63 64 64 66 67

Crop and Environmental Sciences Division, International Rice Research Institute (IRRI), Metro Manila, Philippines { Plant Breeding, Genetics, and Biotechnology Division, International Rice Research Institute (IRRI), Metro Manila, Philippines { TTChang-Genetic Resources Center, International Rice Research Institute (IRRI), Metro Manila, Philippines } IRRI-CIMMYT Crop Research Informatics Laboratory, International Rice Research Institute (IRRI), Metro Manila, Philippines } Social Sciences Division, International Rice Research Institute (IRRI), Metro Manila, Philippines k Left IRRI in June, 2009 *

Advances in Agronomy, Volume 103 ISSN 0065-2113, DOI: 10.1016/S0065-2113(09)03002-8

#

2009 Elsevier Inc. All rights reserved.

41

42

R. Serraj et al.

6. Marker-Assisted Selection 6.1. QTLs for grain yield under drought 6.2. Selective genotyping and bulk segregant analysis 6.3. Application of association studies in drought research 6.4. Marker-assisted breeding 7. Drought-Resistance Genes and GM Technology 7.1. Gene discovery 7.2. Rice transformation 7.3. Bioinformatics and gene functional analysis 8. Conclusions and Future Prospects References

71 71 72 73 76 78 78 79 84 86 88

Abstract The unpredictability of drought patterns and the inherent complexity of the physiological responses involved have made it difficult to characterize component traits required for improved performance, thus limiting crop improvement efforts to enhance drought resistance in rice. The various stress–response mechanisms and options to enhance plant survival under severe stress do not usually translate into yield stability under water deficit. Increased crop yield and water productivity require the optimization of the physiological processes involved in the critical stages of plant response to soil drying, water-use efficiency, and dehydrationavoidance mechanisms. New high-throughput phenotyping methodologies have been developed to allow fast and detailed evaluation of potential drought-resistant donors and the large number of lines identified by drought-breeding programs. Similarly, large collections of rice germplasm, including minicore sets, wild relatives, and mutant lines have been screened for drought-resistance traits. Genetic sources of drought resistance have now been identified for all major rice agroecosystems and some of the associated traits have been characterized. The identification and genetic mapping of major QTLs for performance under drought stress across environments are currently a major focus. This approach provides a powerful tool to dissect the genetic basis of drought resistance. If validated with accurate phenotyping and properly integrated in marker-assisted breeding programs, these approaches will accelerate the development of drought-resistant genotypes. This chapter reviews the recent progress and achievements in dissecting drought resistance in rice and presents future perspectives for the genetic enhancement of drought adaptation.

1. Introduction Drought is the most important source of climate-related risk for rice production in rainfed areas (Pandey et al., 2007). Water deficit is a major challenge for all agricultural crops, but for rice it is even more so because of its semiaquatic origins and the diversity of rice ecosystems and growing

Improvement of Drought Resistance in Rice

43

conditions (O’Toole, 2004). Expansion of irrigation has been the major approach to increasing yield potential in (formerly) rainfed rice. However, due to increased water scarcity and the projected risks related to climate change (Wassmann et al., 2009), there appears to be limited scope for further expansion of irrigation in areas of rainfed rice production. Rice cultivars combining improved drought resistance with yield potential under favorable conditions are, hence, the most promising and deliverable technology for increasing productivity in these drought-prone areas. In this chapter, we review the progress made in our ability to develop such cultivars. In contrast with some other abiotic stresses such as submergence tolerance (Xu et al., 2006), there has been little progress in trait-based and markerassisted selection (MAS) for improved drought resistance in rice (Bernier et al., 2008; Fukai and Cooper, 1995), whereas direct empirical selection for grain yield under managed drought stress has been more successful recently in rice (Kumar et al., 2008; Venuprasad et al., 2007a) and other crops (Ba¨nziger et al., 2006; Edmeades et al., 1999; Nigam et al., 2005). However, this approach also has to face the challenges of applying large-scale and reliable protocols for managed-stress screening in rice-breeding programs and resolving genotype by environment (G E) interactions in the various drought-prone target environments (Sinclair and Muchow, 2001). One way to overcome the large G E limitation is to understand the basic processes accounting for the drought-resistance trait and how the mechanism reacts under a range of weather scenarios. Simulation models can also help overcome the G E limitations, by combining a mechanistic understanding of a drought trait with a range of weather scenarios (Chapman, 2008). Breeding for specific drought-resistance characters can thus be targeted to those geographical regions that would have the highest probability of frequent yield increases. There has been a significant increase in physiological understanding of the mechanisms of crop growth and development, partly bridging the ‘‘phenotype gap’’ generated by the recent progress in genomics (Boote and Sinclair, 2006; Crouch and Serraj, 2002; Miflin, 2000). However, efforts to dissect drought resistance, by identification and characterization of component traits, which can be transferred through plant breeding into cultivars with high-yielding genetic backgrounds, have had very limited success. Nevertheless, there are a few cases in which trait-based selection for drought resistance has resulted in actual yield improvement (Richards, 2006; Sinclair et al., 2004). A common feature of these success stories is that the timing and intensity of drought are critical for these traits to be effective, in addition to the time scale and crop phenological stage in which they operate. This also indicates that any putative drought-resistance trait is unlikely to be relevant under all water-deficit scenarios due to the high G E interactions generally observed in the phenotypic expression of component traits, and their impacts on crop productivity (Hammer and Jordan, 2008).

44

R. Serraj et al.

Recent progress in molecular genetics and genome-sequencing technologies is now enabling high-throughput whole-genome genotyping platforms, and the cost will decline further (Leung, 2008). By comparison, the phenotyping of large germplasm collections and populations for drought-related traits in field trials is still laborious, imprecise, and costly. This has made phenotyping the current bottleneck of crop improvement and molecular mapping. The dynamics of crop responses to water deficit is a continuous process that involves distinct phases of water use and stress profiles. Although substantial research efforts have been devoted to investigating the mechanisms of ‘‘drought tolerance,’’ focusing on survival stage under severe stress, this has led to little progress in crop improvement. Dehydration avoidance is more relevant as a strategy for relieving agricultural drought and maintaining crop performance, before survival drought develops. In rice, the conclusion emerging from long-term multilocation drought studies was that rainfed lowland rice is mostly a drought avoider, with the genotypes that produce higher grain yield under drought being those that are able to maintain better plant water status around flowering and grain setting (Fukai et al., 2008). The objectives of this chapter are to describe the recent progress in crop improvement research on drought resistance in rice and to review the current knowledge of the key traits and physiological processes involved in dehydration avoidance, growth regulation, and reproductivestage processes under drought stress, toward an integrated strategy for drought resistance improvement in rice.

2. Drought Characterization Drought definitions depend on the context and disciplinary outlook, including meteorological, hydrological, and agricultural perspectives. Generically, drought is the occurrence of a rainfall shortage relative to the expected (average) amount for a target region. However, this may not be a meaningful measure from a crop improvement perspective as drought incidence and its effects on crop productivity can depend as much on rainfall distribution as on the total seasonal rainfall. For example, in a recent experiment at IRRI during the wet season of 2006, seasonal rainfall exceeded 1200 mm, including 320 mm of rainfall in a single day associated with a major typhoon (Xangsane). Yet, a short dry spell during the flowering stage resulted in a dramatic decrease in grain yield and harvest index compared with those of the irrigated control treatment (R. Serraj et al., unpublished data). Beyond the search for global solutions to a generic ‘‘drought,’’ the precise characterization of droughts in the target population of environments (TPE) is a prerequisite for better understanding their consequences for crop production (Heinemann et al., 2008).

Improvement of Drought Resistance in Rice

45

2.1. Drought-occurrence parameters From an agricultural perspective, drought occurs when soil moisture is insufficient to meet crop water requirements, resulting in yield losses. Depending on timing, duration, and severity, this can result in catastrophic, chronic, or inherent drought stresses, which would require different coping mechanisms, adaptation strategies, and breeding objectives. The 2002 drought in India could be described typically as a catastrophic event, as it affected 55% of the country’s area and 300 million people. Rice production declined by 20% from the trend values (Pandey et al., 2007). Similarly, the 2004 drought in Thailand affected more than 8 million people in almost all provinces. Severe droughts generally result in starvation and impoverishment of the affected population, resulting in production losses during years of complete crop failure, with dramatic socioeconomic consequences for human populations (Pandey and Bhandari, 2009). Chronic dry spells of relatively short duration can often result in substantial yield losses, especially if they occur around flowering stage. In addition, drought risk reduces productivity even during favorable years in drought-prone areas because farmers avoid investing in inputs when they fear crop loss (Pandey et al., 2007). This partly explains why the average rice yield in predominantly rainfed eastern India during ‘‘normal’’ years varies between 2.0 and 2.5 t ha 1, far below yield in irrigated areas. Inherent drought is associated either with the low (average) expected rainfall or with an increasing problem of water scarcity in irrigated areas, due to rising demand and competition for water uses. This is, for instance, the case in parts of China, where the increasing shortage of water for rice production is a major concern, although rice production is mostly irrigated (Ding et al., 2005). Improved water management strategies such as ‘‘alternative wetting and drying,’’ together with drought-resistant cultivars may be a viable strategy for these areas (Bouman et al., 2006). Increasing rice productivity in the drought-prone rainfed areas requires adapted solutions and strategies in response to the different types of drought, based on precise characterization of the TPE. With milder chronic droughts being generally more frequent than catastrophic ones, overall crop productivity in rainfed areas would probably benefit more from breeding for enhanced water productivity and resistance to the chronic type of water deficit. Irrigation, where possible, is probably the only technological solution to dealing with catastrophic droughts that are generally best dealt with via insurance and other economic coping mechanisms.

2.2. Distribution of rice production systems and rainfall Worldwide, there are more than 100 million ha of rice, with 89% in Asia. About 45% of the rice area is rainfed, of which 25% is never flooded (upland). Asia has large areas of rainfed rice in eastern India and Bangladesh, northeast

46

R. Serraj et al.

Thailand, Cambodia, and the island of Sumatra in Indonesia (Fig. 1). The majority of rice production in Africa is rainfed (Balasubramanian et al., 2007). It is not a surprise that rainfed rice is not produced much in very dry areas. In Asia, about 11% of the irrigated rice is produced in areas with less than 750 mm of average annual rainfall, versus 0.5% of rainfed rice. About 23% of the irrigated rice is in areas with less than 1000 mm versus 4% of rainfed rice (Fig. 2). Rainfed rice in very dry areas is either a misclassification or it is planted in atypical humid locations on the landscape, such as valley bottoms and marshes. Very humid areas also have a large amount of irrigated rice. This is in part because irrigation in many cases provides only a part of the water required, if and when necessary, for example, during a dry spell. It is also because irrigation allows for the production of a second or third rice crop in the dry season, and in some of these areas, irrigated rice during the rainy season is in most years equivalent to rainfed rice.

2.3. Drought-prone target environments Drought-prone rainfed rice ecosystems can be classified based on toposequence position and defined by the water regime encountered (Garrity et al., 1986). The upland ecosystem occupies more than 10 million ha in Asia and has been subdivided based on production systems and agroecological characterization (Courtois and Lafitte, 1999). Several studies have addressed the biophysical characteristics of the rainfed lowland ecosystem and their implications for breeding (Fukai et al., 2001; Mackill et al., 1996; Wade et al., 1999a). The hydrology of rice fields is highly dynamic and

Figure 1 Global rice area by major rice ecosystems.

47

Improvement of Drought Resistance in Rice

0.30

Fraction of area planted

0.25

Rainfed rice Irrigated rice

0.20

0.15

0.10

0.05

> 4000

3750–4000

3500–3750

3250–3500

3000–3250

2750–3000

2500–2750

2250–2500

2000–2250

1750–2000

1500–1750

1250–1500

1000–1250

750–1000

500–750

250–500

0–250

0.00

Annual precipitation (mm)

Figure 2 Distribution of rainfed and irrigated rice area by annual precipitation.

variable, with frequent shifting between flooded and aerobic conditions depending on the position along a toposequence, which can create strong local spatial variation in crop growth (Cooper et al., 1999a). Recent analysis of rainfall distribution data in eastern India, northeast Thailand, and southern China for the period 1970–2003 (Pandey et al., 2007) indicated that drought is a recurrent phenomenon in all three regions, with probabilities of drought occurrence between 10% and 40% of the years, and highest in eastern India. Late-season drought is more predominant than early-season drought, highlighting the critical importance of rice sensitivity to drought during the reproductive stage (Pandey et al., 2007). In northwest Bangladesh, average annual rainfall varies between 1500 and 2000 mm, with more than 200 mm of rainfall per month during the monsoon period ( June to September), when transplanted rice (in Bangladesh rice grown in the season is referred to as ‘‘T. aman’’) is grown mostly under rainfed conditions. However, the erratic rainfall distribution causes frequent drought in this region, and results in yield losses that are generally higher than the damage caused by flooding and submergence (Islam, 2008). A recent characterization and modeling study showed that the recurrence

48

R. Serraj et al.

interval of terminal drought is around 2–3 years, especially during the latter part of the T. aman crop (Islam, 2008). Short-duration varieties such as BRRI dhan 39 are generally used to escape terminal drought in this region. However, the risk of early drought is also very serious with periods of 10-mm rainfall deficits occurring on average every 1.3 years in some districts, illustrating the need of new drought-adapted T. aman rice varieties for these areas.

2.4. Systems analysis and characterization Targeted breeding strategies for improved drought resistance require detailed and precise characterization of the TPE. Crop simulation models can be instrumental for this exercise. Because of the inherent variability of soil hydrology in the rainfed lowlands, the simulation of soil moisture variation is more complex. A water balance model (RL Rice) was also previously used to characterize the drought patterns in multienvironment trials to examine attainable-yield variation in north and northeast Thailand (Fukai et al., 2001). A new simulation model has been recently reported for soil moisture in rainfed lowland rice (Reshmidevi et al., 2008). It simulates the soil-water balance as well as the overland water flow to estimate soil hydraulic characteristics. Heinemann et al. (2008) have recently used a crop simulation model derived from the generic model SARRA to determine the patterns of drought stress for short- and medium-duration upland rice across 12 locations in Brazil. This study allowed the characterization of drought-prone TPE and confirmed the greater yield impact of drought stress when it occurred around flowering and early grain filling. Simulation models can also provide a tool to better understand G E interactions by combining a mechanistic understanding of a drought-related trait with a range of weather scenarios (Chapman, 2008). Given a historical record of weather for a location, the probability of a yield increase (and maybe a decrease) resulting from the incorporation of any trait into the crop can be simulated. Combining the probabilities for yield change with farmers’ adversity to risk gives a strong indication to a breeder of the desirability of incorporating a particular drought trait for cultivars to be grown in a specific location. System analysis can hence allow breeding for specific drought-adaptive traits to be targeted to those geographical regions that will have the highest probability of yield increases (Sinclair and Muchow, 2001). However, most rice simulation modeling has focused on irrigated conditions (e.g., Bouman et al., 2003) and an integrated rice model needs to be developed or adapted specifically for drought-prone rainfed systems, based on a better physiological understanding of rice interaction with the environment under water deficits.

Improvement of Drought Resistance in Rice

49

3. Rice Responses to Drought 3.1. Plant water use It is well known that plants can sense water availability around the roots and respond by sending hydraulic and/or chemical signals to the shoot to elicit several adaptive responses, including stomatal closure, a decrease in leaf expansion, and gas exchange (Tardieu and Davies, 1993). The typical response curve of a particular physiological process to plant-available soil water can be described by two straight lines that intersect at the threshold value for which the rate of the process in stressed plants starts to diverge from a reference value (Ray and Sinclair, 1997; Sadras and Milroy, 1996). Based on this relationship, plant responses to soil-water deficits can be typically described as a sequence of three successive stages of soil dehydration. Stage I occurs before reaching the threshold value, when water is still freely available from the soil and transpiration is not limited by soil-water availability. Stage II starts when the plant reaches the threshold value of available water and the rate of water uptake cannot match the potential transpiration rate. Stomatal conductance declines, thus limiting the transpiration rate to a level similar to that of soil-water uptake, and results in the maintenance of plant water balance. Stage III is reached when the plants are no longer able to limit transpiration through stomatal conductance; they must then resort to other mechanisms of drought adaptation for survival. Virtually all major processes contributing to crop yield, including leaf expansion, photosynthetic rate, and growth, start to be downregulated late in stage I or early in stage II of soil drying (Serraj et al., 1999). At the end of stage II, these growth-supporting processes have effectively reached zero and no further net growth occurs in the plants. The focus of stage III is mostly on survival, which generally involves mechanisms such as osmotic adjustment (Serraj and Sinclair, 2002) and which can be critical in natural dry-land ecosystems, but have little relevance to increasing/stabilizing crop yield in most agricultural situations. Thus, increased crop yields and water productivity require the optimization of the physiological processes involved in the critical stages (mainly stage II) of plant response to soil drying. Extensive experimental evidence has established a general response of plant gas exchange to soil drying when expressed as a function of the fraction of extractable volumetric soil water content (Sadras and Milroy, 1996). Plant gas exchange is generally constant until about FTSW (fraction of transpirable soil water) 0.5–0.4, and then soil drying results in a linear decrease in leaf gas exchange until available soil water is almost exhausted. This response pattern has been observed over a broad range of environments, species, and soils (review by Sadras and Milroy, 1996). The transpiration rate on drying soil relative to plants on well-watered soil (relative

50

R. Serraj et al.

transpiration rate, RT) was described as a function of the FTSW, using the following equation for soybean (Serraj and Sinclair, 1997): RT ¼

1 1 þ 5:25 expð11:23 FTSWÞ

ð1Þ

This pattern has been confirmed in rice by describing the response using a two-segment model with a linear-plateau segment for the initial phase and a linear-decreasing segment for the second phase of soil drying (Serraj et al., 2008b). The relationships of normalized transpiration rate (NTR) to soil drying, as measured by FTSW, are established according to a dry-down protocol described previously (Serraj et al., 1999; Sinclair and Ludlow, 1986). The NTR response curves were well described by linear-plateau functions that allowed the calculation of the soil-water thresholds at which transpiration of drought-stressed plants began to decrease relative to the well-watered treatment (Ray and Sinclair, 1997). The FTSW thresholds varied significantly among rice genotypes, from 0.35 in upland-adapted cultivar Apo to 0.74 in lowland-adapted cultivar IR72 (Serraj et al., 2008b). Based on this analysis, cultivars with lower threshold values (e.g., Apo) are able to maintain a higher NTR during the drying cycle than those with higher thresholds (e.g., IR72), which would be associated with genetic differences in the control of transpiration and drought-avoidance mechanisms. Those with lower thresholds would allow a greater amount of dry matter accumulation during drought, which might result in higher transpiration efficiencies, as has been shown recently in groundnut cultivars transformed with the DREB1A gene (Bhatnagar-Mathur et al., 2007). Similar data of genotypic variation of NTR response curves were also observed with other species, including maize (Ray and Sinclair, 1997). However, as vapor pressure deficit (VPD) also influences transpiration rates greatly, with possible genotypic variation in the magnitude of the response (Sinclair et al., 2008), this would affect the NTR response to soil dry-down. Given the little scope for improving rice performance under drought based on survival mechanisms of stage III, dehydration-avoidance mechanisms offer more promising avenues, by improving plant characteristics to prolong phases I and II during soil drying. Some of these characteristics include deeper rooting and ability to overcome soil physical barriers and hardpans, and water conservation by controlling transpiration response to soil drying and decreasing stomatal conductance under high VPD conditions.

3.2. Spikelet sterility and grain failure Although drought affects all stages of rice growth and development, stress during the flowering stage depresses grain formation much more than drought at other reproductive stages (Boonjung and Fukai, 1996);

Improvement of Drought Resistance in Rice

51

therefore, screening for resistance near the flowering stage has been considered to be more useful in breeding for improved drought resistance. The strong effects of drought on grain yield are largely due to the reduction in spikelet fertility and panicle exsertion. Several studies have found that reproductive development from meiosis in the spore mother cells to fertilization and early seed establishment was extremely sensitive to various stresses, including drought. These stresses cause various structural and functional disruptions in reproductive organs, leading to failure of fertilization or premature abortion of the seed (Saini and Westgate, 2000). Drought can inhibit the development of reproductive organs, such as the ovary and the pollen at meiosis stage (Saini and Westgate, 2000), and it can also inhibit processes such as anther dehiscence, pollen shedding, pollen germination, and fertilization (Ekanayake et al., 1990). The drought-induced inhibition of panicle exsertion has been identified as a consequence of a decrease in peduncle elongation, which can usually account for 70–75% spikelet sterility under water deficit (O’Toole and Namuco, 1983). Drought stress inhibits peduncle elongation, and rewatering can only partially restore elongation. Recent studies at IRRI found that drought significantly delayed peduncle elongation, trapping a significant fraction of the panicle within the flag leaf sheath due to the repression of the expression of cell-wall invertase genes ( Ji et al., 2005). The spikelets left inside the leaf sheath are usually sterile, resulting in poor yield, which indicates that peduncle elongation may play a major role in panicle exsertion and spikelet fertility under stress. Mutant studies showed that the cause of spikelet sterility can be of two types: inhibition of starch accumulation in pollen grains and failure of anther dehiscence and/or synchronization with anthesis due to the suspension of septum degradation and stomium breakage (Zhu et al., 2004). If drought stress occurs during these processes, the reproductive organs will be abnormal and damaged and spikelets will be sterile. Liu et al. (2005) reported a significant difference in number of pollen grains between IR64 and Moroberekan in the top four rachis under drought conditions. The variation in spikelet fertility between the genotypes was mainly due to the difference in locule-wall structure, and to variation in the number of pollen grains on stigma (Liu et al., 2005). The effect of drought on spikelet sterility and grain failure has long been found to be associated with plant water status, and several reports have previously shown that leaf and panicle water potential were highly associated with panicle exsertion and anther dehiscence (Ekanayake et al., 1990; Garrity and O’Toole, 1995; O’Toole and Namuco, 1983). Garrity and O’Toole (1995) were able to screen rice varieties for reproductive-stage drought-avoidance traits, using canopy temperature as a surrogate trait for plant water status under stress. They showed that grain yield and spikelet fertility were highly correlated with midday canopy temperature on the day

52

R. Serraj et al.

of flowering, and lines with high drought-avoidance scores consistently remained the coolest under stress (Garrity and O’Toole, 1995).

3.3. Dehydration-avoidance mechanisms Given the low potential for improving rice performance under drought based on survival mechanisms of stage III, dehydration-avoidance mechanisms offer more promising avenues, by improving plant characteristics to prolong stages I and II during progressive soil drying. Based on long-term multilocation drought studies in Southeast Asia, Fukai et al. (2008) concluded that rainfed lowland rice is mostly a drought avoider, with the genotypes that produce higher grain yield under drought being those able to maintain better plant water status, especially when stress occurs around flowering and grain formation Dehydration avoidance is defined as the ability to maintain plant water status under soil-water deficits, mainly through increased water uptake and/or reduced transpiration (Levitt, 1980). 3.3.1. Plant water status Leaf relative water content (RWC) and water potential (LWP) have long been associated with rice performance under water deficit (O’Toole and Moya, 1978) and were found to be correlated with yield under drought stress applied around flowering stage ( Jongdee et al., 2006; Lafitte, 2002). Several QTLs have been mapped for RWC (Courtois et al., 2000), LWP (Liu et al., 2006), stomatal conductance (Khowaja and Price, 2008), and several other traits putatively linked to drought resistance. Some inconclusive attempts have been made to exploit such QTLs in MAS schemes, but most QTLs identified have not been repeatable across environments and/or populations, or have not consistently affected either the target trait (Steele et al., 2006) or grain yield under stress when introgressed into a susceptible cultivar (Steele et al., 2007). The physiological basis of a major QTL for yield under drought (qtl12.1), mapped in a Vandana/Way Rarem population (Bernier et al., 2007), was recently investigated by analyzing plant water status traits under both field and controlled environments (Bernier et al., 2009). The large effect of qtl12.1 on grain yield under drought stress in the Vandana/Way Rarem F3-derived population was confirmed in this study. When exposed to severe drought stress at the reproductive stage, lines with the Way Rarem allele of qtl12.1 had improved grain yield, harvest index, and water status, and a higher spikelet number and weight. The lines with the Way Rarem allele of the QTL exhibited a reduced rate of leaf rolling and leaf drying, a higher LWP and significantly increased stomatal conductance, as well as a higher RWC. These results clearly indicated that the major effect of qtl12.1 on grain yield under drought was associated with an improvement in plant water status under stress, most likely through a dehydration-avoidance mechanism. A similar

Improvement of Drought Resistance in Rice

53

approach has also been used for dissecting the dehydration-avoidance mechanisms associated with the major loci for performance under drought in near-isogenic lines derived from IR64/Aday Sel (Venuprasad et al., 2007b). 3.3.2. Root traits Significant progress has been made over the past decade in identifying QTLs associated with several root traits, including rooting depth, length, and thickness, which have long been emphasized as important adaptations to water deficits in rice (Nguyen et al., 1997). Several QTLs have been mapped in rice for various root traits (see Kamoshita et al., 2008 for a review), but it has not yet been possible to translate these results into a successful markerassisted breeding (MAB) program for drought avoidance. One of the main reasons for the lack of success is probably related to logistical and methodological problems of root phenotyping. The heritability of root-related traits is generally low and field screening for these traits is laborious and complicated because of the high plasticity of root growth in response to even small changes in the environment and the complexity of G E interactions. Therefore, most QTLs identified for root traits have been detected in controlled environments, which are very different from the rooting environment plants experience in the field (Serraj et al., 2004). The proof of concept for the ability to introgress root QTLs in rice was recently shown by Steele et al. (2006). If validated with accurate phenotyping protocols and properly integrated into MAB programs, root QTL mapping and MAS strategies could then allow pyramiding of drought-avoidance mechanisms in rice and wheat cultivars and accelerate the development of locally adapted drought-resistant genotypes. To investigate the physiological basis of a major QTL for yield under drought, which was previously mapped in a Vandana/Way Rarem population (Bernier et al., 2007), root traits were analyzed under both field and controlled environments (Bernier et al., 2009). Traits related to deep root growth were positively correlated to water consumption in the droughtstress treatment. Based on regression analysis, the two most important traits were deep root length (below 30 cm) and maximum rooting depth. The lines with the Way Rarem allele of the QTL had 18% higher deep root lengths as compared to lines with the Vandana allele at the locus, leading to the conclusion that rooting depth may be the most important difference in explaining the increased water uptake in the lines with the favorable allele of the QTL. Another candidate root trait to explain the improved water uptake is better/faster root growth at depth. This hypothesis was also supported by the fact that the chromosome region where qtl12.1 is located has previously been identified as affecting rooting depth in a different population (Yue et al., 2006). The overall conclusion of this study is that the Way Rarem allele of qtl12.1 improved root architecture (Bernier et al., 2009) and resulted in increased water uptake, although this was relatively

54

R. Serraj et al.

small (7%). However, the field root sampling methods were not precise enough to clearly capture these differences in root architecture among the recombinant inbred lines (RILs). Better evidence as to the nature of the effect of the QTL on roots can probably be obtained by conducting precise root phenotyping experiments using near-isogenic lines rather than the RIL used in this study. BC3F2 near-isogenic lines with and without the Way Rarem allele of qtl12.1 in the Vandana genetic background are currently being developed at IRRI to clarify the effect of this locus. 3.3.3. Photosynthesis and stomata aperture traits Stomatal regulation in response to soil drying is triggered by root-shoot chemical and/or hydraulic signaling (Tardieu and Davies, 1993) and is a key adaptation strategy to avoid tissue dehydration under drought. Various stomatal characteristics such as density, low conductance, high sensitivity to leaf water status, and ABA accumulation have been suggested as desirable traits in crop improvement for water-limited conditions. Although depending on the environment and occurrence of stress, they may also have tradeoffs in terms of limiting carbon assimilation and crop yield. Significant genetic variation for the sensitivity of stomata to leaf water status was reported in rice (Price et al., 1997), and the process of plant water conservation through stomatal regulation has been associated with reduced spikelet sterility and increased grain yield under reproductive-stage drought (Pantuwan et al., 2002). The sensitivity of stomata to leaf water status is thought to be a highly heritable trait (Ludlow and Muchow, 1990) and has long been shown to have significant genetic variation in rice (Dingkuhn et al., 1991; Price et al., 1997). Dry-down experiments carried out in pot studies under controlled conditions indicated that the FTSW thresholds vary significantly among rice genotypes (Serraj et al., 2008b). Based on this comparison, cultivars with lower threshold values closed their stomata at lower FTSW (drier soil) and hence were able to maintain higher transpiration rates during the drying cycle than those with higher thresholds. These differences in transpiration response to FTSW are most likely associated with genetic differences in the control of transpiration and drought-avoidance mechanisms. However, transpiration response to soil drying is also influenced by VPD (Sinclair et al., 2008), and the interactive effects of FTSW and VPD need to be investigated to better understand stomatal regulation and transpiration response to soil-water deficits. The role of hormonal regulation and ABA is also still debated. Recent studies of stomatal responses to partial soil drying in rainfed lowland rice suggested a possible role of root signals (Siopongco et al., 2008). A drought-induced increase in leaf ABA concentration under field conditions and strong association with soil moisture tension and stomatal conductance (gs) suggested its involvement in mediating stomatal responses during the early steps of drought responses in rice. However,

Improvement of Drought Resistance in Rice

55

stress recovery after severing of drought-stressed roots in a greenhouse could be attributed to increased hydraulic conductance. These responses imply a role for both chemical and hydraulic signals in rice (Siopongco et al., 2008). Instantaneous leaf gas exchange and chlorophyll fluorescence measurements were taken in dry-season upland screening experiments to compare the physiological stress responses of contrasting cultivars (Centritto et al., 2009). Photosynthesis (A) was inhibited by water deficits but to a variable extent in the genotypes tested. In general, the highest inhibition was observed in the genotypes that had inherently lower A, gs, and mesophyll conductance (gm). A very strong correlation was found between low A and low gm, both in genotypes with inherently low A and in water-stressed plants of all varieties, indicating that the main limitation of photosynthesis in rice is the low chloroplastic CO2 concentration (Centritto et al., 2009). This evidence of correlation between photosynthesis and diffusional conductances could provide a mechanistic basis for distinguishing photosynthetic causes of differences in water-use efficiency (WUE) from causes related to stomatal regulation. 3.3.4. Carbon isotope discrimination and WUE As an integrative parameter of stomatal aperture traits, carbon isotope discrimination (CID) has been successfully used as a selection criterion for water-use efficiency and yield under drought (Condon et al., 2004). The negative correlation between CID and WUE, consistent with theoretical predictions, was confirmed in several crop plants, including rice (Dingkuhn et al., 1991; Impa et al., 2005; Peng et al., 1998). However, field studies in wheat have also revealed that the correlation between CID and grain yield was often positive rather than negative (Condon et al., 2004; Monneveux et al., 2006), leading to the conclusion that the anticipated association was hidden because correlations between CID and yield can be influenced by several physiological differences among genotypes (Rebetzke et al., 2002). A successful breeding program began in Australia, aiming at the introduction of enhanced WUE into germplasm within similar genetic backgrounds. The lines selected for increased WUE using CID produced significantly greater yields in eight of the nine field environments, with a relative yield increase reaching 11% in the driest environment (Rebetzke et al., 2002). Two wheat varieties have been released so far from this program, one of the few success stories of trait-based selection for improved crop performance under drought (Sinclair et al., 2004). In rice, grain CID was also found to be strongly associated with grain yield under both irrigated and drought conditions. Two years of dry-season field testing at IRRI have recently confirmed the association between grain yield and CID among rice lines segregating for a major QTL of yield under drought ( J. Bernier et al., unpublished data).

56

R. Serraj et al.

3.3.5. Canopy temperature Canopy temperature is a sensitive indicator of plant stress level, and has long been associated with stomatal conductance at the leaf level. This has led to the development of infrared sensing of canopy temperature as a technique for estimating gs and plant water status ( Jones, 1992). During a period of water shortage, plants conserve water by closing their stomata, raising their internal temperature. Leaf temperature is thus correlated with plant stress level. However, leaf temperature is influenced by stomatal and boundary layer resistances, as well as by meteorological conditions. The rate of leaf transpiration is only one of many components of the canopy energy balance that affect canopy temperature: factors such as radiation, wind speed, air temperature, humidity, and VPD all have major effects ( Jones, 1992). Thermal sensing has been used successfully in field screening for monitoring differences in stomatal responses to drought in crops such as rice and wheat (Amani et al., 1996; Garrity and O’Toole, 1995). Using an infrared thermometer gun on single rice leaves, Garrity and O’Toole (1995) were able to screen rice varieties for reproductive-stage drought-avoidance traits. Their major findings were that mean canopy temperature increased from 28 to 37 C during the drought-stress period, and both grain yield and spikelet fertility were significantly correlated to midday canopy temperature on the day of flowering. They also found highly significant differences in canopy temperature among rice cultivars, reporting that lines with high drought-avoidance scores consistently remained coolest under stress (Garrity and O’Toole, 1995). However, using an infrared gun on single plants is often subject to high field variation and has limited scope as a high-throughput and reproducible screening method. The recent development of infrared thermal imagery has revived interest in using canopy temperature as a screening tool, especially if applied at plot or field levels (Leinonen and Jones, 2004). Infrared cameras can now be used to detect whole-plant, or even plot, temperature within seconds. IR thermal imaging has been used successfully to screen for stomatal and other mutants in the laboratory (Merlot et al., 2002) and for precision phenotyping of plant water status under water deficits, especially when combined with automated image analysis to facilitate analysis (Leinonen and Jones, 2004). Based on the energy balance theory (Monteith, 1973), Horie et al. (2006) developed a methodology based on field measurements of sunlit and suddenly shaded canopy surface temperatures. Simultaneous recording of microclimate data allowed estimation of the evapotranspiration rate, aerodynamic resistance, and canopy diffusive resistance under field conditions. This study further demonstrated the possibility of relating the quantitative estimation of rice canopy physiological characteristics under field conditions to crop growth rate during the period preceding flowering and grain formation stages (Horie et al., 2006). Recent work at IRRI

Improvement of Drought Resistance in Rice

57

has demonstrated that there is potential for large-scale field drought phenotyping with thermal imaging, even in the humid climate of the Philippines (Serraj and Cairns, 2006; Serraj et al., 2008b).

4. Concepts and Tools for Phenotyping Improving the precision and throughput of phenotyping is now often highlighted as the bottleneck and the main priority of droughtresistance studies. Yet identifying a relevant measure of plant sensitivity to water-deficit dynamics is still a challenging question. Time (days after stress application) is often used in drought experiments, but time is obviously an inaccurate covariable as plant physiological responses depend mainly on environmental conditions and stress-occurrence parameters. Similarly, the use of plant water status parameters such as leaf water potential or relative water content as stress covariables is also often biased by variability, G E interaction, and mostly by the impracticality for high-throughput field applications (Lafitte, 2002). Beyond the search for generic drought phenotyping recipes that would fit all situations and environments, our current approach for developing a relevant phenotyping methodology starts by attempting to answer the question: What is the independent variable in the soil–plant–atmosphere continuum system that can be quantitatively and reproducibly related to crop response to water deficits? It has long been found that genetic differences in physiological responses to water deficits are mainly related to differences in soil-water extraction. In rice, Lilley and Fukai (1994a) demonstrated that cultivar differences in the rates of stress development were strongly associated with the variation in extractable soil water and water extraction rate. After accounting for differences in water extraction ability, cultivar differences in sensitivity of physiological processes to water deficit were small (Lilley and Fukai, 1994b). It is, hence, crucial for phenotyping crop responses to drought to characterize precisely the dynamics of soil-water extraction.

4.1. Field-managed drought screening Farmers do not usually want cultivars that are drought resistant but have low yield potential and thus give low yield in favorable years. They prefer cultivars that respond to favorable conditions but that ‘‘protect’’ an economically useful yield under drought conditions. To achieve this combination, breeding lines should be screened under both stress and nonstress situations to select lines that combine high yield potential and good drought resistance. A few studies have reported a low but positive correlation

58

R. Serraj et al.

between yield under stress and yield potential (Kumar et al., 2008; Venuprasad et al., 2007a) and have also demonstrated the feasibility to combine high yield potential and drought resistance (Kumar et al., 2008). 4.1.1. Screening sites and seasons Breeders deploy considerable efforts to disseminate the breeding lines selected at research stations (selection environments, SE) to the TPE. Selection efforts will only be successful if yield gains achieved in SE are expressed in farmers’ fields. Therefore, one of the primary requirements for selection to be effective is representativeness of the selection environment to the TPE. The relationship between cultivar performance in SE and the TPE can be measured as genetic correlation (rG). The higher the rG, the better the relationship between SE and TPE. The rG between SE and TPE can be maximized by ensuring that screening conditions at research stations are similar to those in farmers’ fields. This should not only include soil type and input application but also land topography, rainfall distribution, and irrigation management practices. Carrying out selection directly in the target environment would be the best option. However, it is not realistic for a breeder to grow large populations in farmers’ fields. Therefore, before embarking on a screening program, breeders should confirm that performance in SE is predictive of performance in the TPE. Generally, screening for drought resistance is carried out mostly in the wet season in the target environments in the large rainfed drought-prone areas of eastern India and northeastern Thailand ( Jongdee et al., 2002; Kumar et al., 2008; Pantuwan et al., 2002). However, the success of wetseason screening is hampered by erratic rainfall during the season. Most of the drought screening at the IRRI station in the Philippines is carried out in the dry season. The performance of lines identified as resistant in dry-season screens is generally predictive of the wet-season screens in rainfed areas in India (IRRI, unpublished data). However, this relationship is not always true. At some places, dry-season screening is equivalent to out-of-season screening and climatic factors such as low temperature during seeding, and/ or high radiation, low humidity, and high temperature during flowering interact with the drought screening. The main purpose of dry-season screening under such situations should be to obtain additional information about drought resistance (Lafitte, 2003). Wet-season late-planting drainage screens are largely practiced in rainfed areas of eastern India (Kumar et al., 2008) and northeastern Thailand ( Jongdee et al., 2002; Pantuwan et al., 2002). In the late-planting wetseason screens, stress trials are planted 3–5 weeks late so as to synchronize the reproductive stage of the rice crop with the occurrence of drought due to absence of the rain at the terminal stage of the cropping season. The probability of successful wet-season late-sown screens has been very high (IRRI, unpublished data). Precise screening of a limited number of lines

Improvement of Drought Resistance in Rice

59

under a rainout shelter, especially for detailed characterization of the lines, has also been suggested (Lilley and Fukai, 1994a). 4.1.2. Control of drought initiation and severity Lines with high yield potential perform reasonably well under mild to moderate levels of drought stress (Kumar et al., 2008). To identify a true drought-resistant line, most breeding programs aim to reduce yield of the stress trial by 50% (Lafitte, 2003). However, Kumar et al. (2008) did not observe any response to selection in screenings that had a yield reduction up to 56%. They have suggested that an effective stress screening protocol should reduce mean yield by 65–85% relative to a fully irrigated control in order to allow the identification of drought-resistant lines (Kumar et al., 2008). Most breeding programs screening for drought resistance generally fail to impose a sufficiently severe stress in their trials, and, as a result, are not able to accurately select drought-resistant lines. The use of tensiometers to monitor soil moisture in stress trials, including a number of droughtsusceptible and -resistant checks at repeated intervals, monitoring of leaf rolling and leaf drying in these checks as well as in the experiment as a whole, and proper monitoring of water-table depth in lowland trials are among the tools that can assist breeders in managing the proper level of stress in their screening. Lowland soil is generally saturated before stress begins. Upon initiation of stress, the number of days stress will develop in any field depends on the moisture-holding capacity of the soil, losses from seepage and percolation, and the amount of water transpired by the crop. In most cases, it takes around 18–20 days for stress to develop after a field is drained in clay soil (Kumar et al., 2008). In contrast, in sandy lowland soil, stress can develop within 14 days or so. In upland soil, stress can appear in severe form starting from 7 to 9 days after withholding irrigation. Depending upon the soil type and crop duration, the stress should be initiated in the experiments between 45 and 60 days after sowing so as to target the drought-sensitive stage of rice that lies between approximately 15 days before flowering and 7 days after flowering (O’Toole and Namuco, 1983). Once the drought stress begins, a line that will flower shortly after stress initiation is likely to be less stressed than a line that flowers later. To remove differences in days to flowering affecting the severity of stress/cycles of stress, genotypes can be grouped into early, medium, and late duration and planting dates can be staggered so that all genotypes flower at the same time.

4.2. The FTSW dry-down approach The total amount of soil water available to support plant water uptake was defined by Sinclair and Ludlow (1986) as the ‘‘transpirable soil water’’ and the relative dryness of the soil between upper and lower limits as the FTSW.

60

R. Serraj et al.

By definition, FTSW has a value of 1.0 at the upper limit and 0.0 at the lower limit. An FTSW decrease results in a progressive water deficit that influences many physiological processes such as transpiration, photosynthesis, or leaf expansion (Serraj et al., 1999). These processes are generally inhibited when FTSW decreases to 0.4–0.5, with a consistent trend across a wide range of environments and plant genotypes (Sadras and Milroy, 1996). A simple derivation model was proposed by Sinclair (2005), defining plant water flux in drying soil relative to that in well-watered soil and examining the response in a range of soil volumetric water contents. This derivation resulted in a relatively simple expression that predicted daily transpiration rate response to drying soil, which was consistently independent of the absolute value of transpiration rate, root length density, and soil depth. The expression of relative transpiration as a function of volumetric soil water content available to support transpiration minimized the influence of soil texture on the overall response of water flux to progressive soil drying. The derivation offers a theoretical basis to explain the stability in daily transpiration response to drying soil that has been observed over a wide range of crops and conditions. This also supported the robustness of using FTSW as a stress covariable in drought studies. This technique is now being used for a field-based system (Serraj et al., 2008b), under both upland and lowland conditions, in which soil moisture profiles are automatically monitored in parallel to plant water status and leaf gas exchange measurements to analyze the dynamics of rice response to water deficits. The use of FTSW as a stress covariable makes possible the integration of on-station field experiments with multilocation trials and controlled-environment experiments that analyze specific trait responses to stress and related gene expression (Serraj et al., 2008b).

4.3. Control and monitoring of soil moisture profiles Several irrigation regimes, including sprinkler, surface, furrow, and drip, have been used at IRRI to induce drought during specific periods and to investigate the response of rice varieties to drought stress (Lafitte and Courtois, 2002). For many of the breeding experiments, water is withheld during the reproductive period—the most sensitive stage to drought stress— and flowering, grain filling, and yield are examined. Rice has a large variation in the time at which the reproductive stage is reached; therefore, experiments are generally separated into varieties with similar duration.

In upland fields, sprinkler irrigation is generally used to control soil-water balance and plant water status for managed drought screening. Withholding irrigation allows targeted drought timing (reproductive or vegetative stage) and severity. For instance, a 2-week water-deficit treatment during the dry season resulted in a dramatic decrease in FTSW

Improvement of Drought Resistance in Rice

61

(Serraj et al., 2008b). If stress begins around 10 days before flowering, this can induce a yield reduction of 40–50% compared to the fully irrigated aerobic control treatment. Line-source sprinkler irrigation is used to generate a differential gradient of soil moisture varying from well watered to extremely dry (Cruz and O’Toole, 1984), and thus allows the side-by-side comparison of different levels of drought stress on the same plots. A drip-irrigation system uses drip tapes to apply water to individual rice plants allowing water management to be controlled in individual plots with application of variable drought-stress periods. This is preferred to screen varieties with major differences in flowering time, by initiating the stress period at different times to target a specific phenological stage for each variety (Lafitte and Courtois, 2002). In rainfed lowlands, drainage of flooded paddy fields at specific times (generally 2–4 weeks after transplanting) results in progressive soil drying with a stress intensity that varies with toposequence position and environmental conditions (rainfall and evaporative demand). This system has been used successfully during both the dry and wet seasons to simulate intermittent dry spells, to screen large collections of breeding lines and to identify donor parents for drought-resistance breeding (Kumar et al., 2007). A variety of devices are being tested for the optimization of soil moisture measurement precision, including classical home-made Hg tensiometers, gauge-type tensiometers, equitensiometers, thetaprobes, and capacitance sensors such as Diviner-2000. A 3-year international comparative study conducted by the International Atomic Energy Agency (IAEA) carried out numerous soil moisture comparisons on a wide range of sensors, including the soil moisture neutron probe (SMNP), time domain reflectometry (TDR), capacitance probes (EnviroSCAN and Diviner 2000), the Delta-T thetaprobe, and a profiling probe, tested under a wide range of soil types, vegetations, and experimental sites, under both irrigated and rainfed conditions in field environments and in the laboratory (http://www.iaea.org/ programmes/nafa/d1). The main conclusions indicated that all the devices, except the conventional TDR, required soil-specific calibration systems, and the success in using capacitance sensors such as the Sentek and Delta-T devices depends critically on proper calibration and careful installation of access tubes to avoid soil disturbance.

4.4. Nondestructive methods for plant growth and water status The classical techniques used for measuring plant water use, water status, growth, and gas exchange are tedious and time-consuming, and thus they are not suitable for precise high-throughput phenotyping. Recent efforts

62

R. Serraj et al.

have focused on developing automated and high-throughput phenotyping platforms, such as ‘‘Phenopsis,’’ which is based on automated pot-weighing systems and digital imagery techniques (Granier et al., 2006). Imaging techniques, including chlorophyll fluorescence, NDVI, infrared thermography, reflectance, autoluminescence, and multispectral fluorescence imagery, have all been used for monitoring the effects of drought and other abiotic stresses on plants (see Chaerle et al., 2007 for a review). Some of these techniques are specifically promising for application in highthroughput drought phenotyping. For instance, thermal imaging has proved useful in rapid screening for stomatal responses to stress; it can also be combined with fluorescence imaging to study photosynthesis (Chaerle et al., 2007). Clear advantages of these imaging techniques are that they are nondestructive, as they allow large-scale, continuous, and possibly automated monitoring of dynamic spatial variations at wholeplant or canopy levels. They can also reveal the early signs of stress responses, thus making it possible to visualize the kinetics of stress responses in screening experiments and phenotyping dehydration-avoidance traits. The application of thermography has also been tested for large-scale field screening of rice genotypes for stomatal behavior and water use under drought (Serraj and Cairns, 2006). This technique could also be used as an indirect indicator for screening diverse germplasm accessions for drought-avoidance root traits under lowland conditions, before extracting roots for detailed morphological analyses. Thermography is showing promising results for capturing side-by-side differences in canopy temperature and plant water status, but large-scale extrapolation of the approach for field use is not straightforward given the fluctuations in environmental conditions (Leinonen et al., 2006). Imaging approaches can also be combined with the use of carbon and oxygen isotope discrimination techniques (Condon et al., 2004). A combination of CID with nondestructive visual, thermal, and fluorescence imaging to measure stomatal conductance, and photosynthetic activity, respectively, might allow a better resolution of stomatal and nonstomatal traits in plant responses to water deficits (Chaerle et al., 2007). The integration of the FTSW dry-down methodology with nondestructive imaging and isotopic techniques, for detailed measurements of plant growth and development parameters and yield components, allows high-throughput, field-based, and precise drought phenotyping (Serraj et al., 2008b).

4.5. Model-based phenotyping Simulation models have been used to understand and overcome the complexity of G E M interactions. Models can provide a tool to combine a mechanistic understanding of a drought trait with a range of weather

Improvement of Drought Resistance in Rice

63

scenarios. Models can also be instrumental in environmental characterization to support weighted selection for specific traits and adaptations. Modeling can play a role in enhancing the precision and integration of phenotyping either by linking model coefficients directly to QTLs (Chapman, 2008; Tardieu, 2003; Yin et al., 1999) or more heuristically to guide integrated phenotyping approaches (Hammer and Jordan, 2008). Although there is some general agreement on the potential benefits of model-based phenotyping, there are still pending questions regarding the general scope of the strategy, the link between crop simulation and genetic models, and the prototype of models appropriate for phenotyping. Hammer and Jordan (2008) concluded that, although an integrated systems approach to crop improvement is still in its infancy, rapid technology developments will speed its progress toward more relevance and a greater role in breeding for adaptation to drought.

4.6. Phenotyping for gene expression and profiling Phenotypic traits in gene expression experiments have been generally assumed to be fixed. However, crop growth and development traits are dynamic processes controlled by a complex network of genes. Analyzing phenotypic data measured at a single time in mapping and gene expression studies is not suitable to capture the genetic control of developmental changes in response to stress. The concept of functional mapping has been proposed to characterize the QTLs or genes that underlie complex dynamic traits (Wu and Lin, 2006). The genotypic variation of dynamic responses of transpiration and leaf growth to soil-water deficit was recently investigated under controlled and field conditions. The relationship between relative transpiration and soil drying, as measured by FTSW, was fit by linear-plateau functions that allowed the determination of the soil-water thresholds at which transpiration of drought-stressed plants began to decrease compared to the well-watered treatment. FTSW soil-water thresholds varied significantly among genotypes, suggesting a link between the kinetics of transpiration response to water deficits, leaf gas exchange, and plant growth parameters under drought. The FTSW dry-down phenotyping approach was used for analyzing gene expression and stress–response mechanisms of leaf elongation rate (LER). Transpiration and stomatal conductance both declined with increasing soil moisture deficit (Serraj et al., 2008b). LER was highly sensitive to water deficit, declining at a higher FTSW threshold value than transpiration and stomatal conductance. Expression profiles of expansin and XTH genes were determined in the leaf-elongating zone at specific soil-water deficits. Several genes showed similar changes in expression and consistent associations with changes in leaf elongation (Serraj et al., 2008a).

64

R. Serraj et al.

5. Conventional Breeding Although drought has long been recognized as the primary constraint to rice production in the rainfed ecosystem, most breeding programs in rainfed areas have not made long-term concentrated efforts to improve the drought resistance of presently grown cultivars. Considering the highly variable nature of rainfed environments, breeding for drought-prone rainfed ecosystem requires the development of genotypes that meet the preference of farmers, combine high-yielding ability with good levels of drought resistance (Atlin et al., 2008) and in addition possess tolerance of prevalent biotic stresses. This goal can be achieved only with a large-scale, long-term, and product-oriented breeding program designed for the rainfed ecosystem. As a result of the lack of such concentrated efforts, most improved cultivars grown in drought-prone areas are ones originally bred for irrigated conditions and they were never selected for drought resistance. Although these cultivars have high yield potential, they are highly prone to a yield reduction under drought (Atlin et al., 2008; Kumar et al., 2008). On the other hand, some traditional rice cultivars grown in rainfed areas are highly drought resistant but have low yield potential.

5.1. Screening of rice germplasm Success in breeding for improved drought resistance depends essentially on the choice of parents, selection criteria, and robustness of the managedscreening protocols. The objectives of a screening system are to focus on the TPE and adaptation to major stress-occurrence scenarios, and to minimize field variability for detecting heritable differences in drought resistance. Because of high rainfall levels and the unreliability of weather scenarios during the wet growing season at Los Ban˜os, drought screening at IRRI is mainly carried out during the dry season. Comparing several drought-screening protocols in upland or in drained lowland paddies, Lafitte and Courtois (2002) found that intermittent stress, imposed by withholding irrigation during the period bracketing the entire flowering and grain-filling stages, is generally reliable for ranking cultivar performance under drought, similar to stress targeted precisely at the flowering period of individual cultivars. Recent research findings at IRRI have demonstrated the feasibility of direct selection for yield under drought (Kumar et al., 2008; Venuprasad et al., 2007a). Since yield under stress is a function of yield potential, escape, and drought response, the use of the drought-resistance index (DRI) can help to distinguish drought resistance from escape and yield potential (Bidinger et al., 1987; Ouk et al., 2006), and therefore further enhance the precision and reproducibility of drought screening.

Improvement of Drought Resistance in Rice

65

While breeding for upland and aerobic rice has recently made significant progress in developing new rice cultivars for water-short environments (Bernier et al., 2008), progress in rainfed lowlands has been relatively slow. Most improved cultivars grown in drought-prone rainfed lowlands were originally bred for irrigated conditions, and were never selected for drought resistance (Kumar et al., 2008). Drought escape has been exploited in the drought-prone areas of eastern India and Bangladesh, through shortduration varieties, mainly of the aus germplasm group. But most of these varieties are not necessarily drought-resistant. The slow progress in the genetic improvement of grain yield in rainfed lowlands was explained by two major factors: the complexity of the target genotype environment system and the insufficiency of genetic resources available to breeding programs (Cooper et al., 1999b). Several studies have previously discussed the biophysical characteristics of the rainfed lowland ecosystem and the implications of these for breeding for yield and improved adaptation (Fukai et al., 2001; Mackill et al., 1996; Wade et al., 1999a). Large genetic variation exists within rice and its wild Oryza relatives for performance under drought stress, but progress in developing improved cultivars has been relatively slow. Many parental lines and donors have been identified for drought resistance in upland (Bernier et al., 2007), but only a few have been reported for the more extensive rainfed lowland system (Atlin et al., 2006). The identification of parental materials and development of new populations were a major target for the IRRI rainfed lowland breeding program in the 1990s, focusing on the major target environments in eastern India and northeast Thailand (Sarkarung and Pantuwan, 1999). Breeding populations were developed in the backgrounds of Mahsuri, Safri17, and Sabita for eastern India and KDML105 for northeast Thailand. An extensive G E study in rainfed lowland by Wade et al. (1999b) analyzed the interactions of 37 genotypes across 36 environments in India, Bangladesh, Thailand, Indonesia, and the Philippines from 1994 to 1997. Only a small group of genotypes was stable across environments. Cultivar NSG19 was found to be adapted to environments with rapidonset late drought, whereas Sabita and KDML105 showed adaptation to environments with late maturity or recovery after drought. Stress-sensitive mega varieties are still widespread across South and Southeast Asian rainfed rice production systems, including Swarna, Sambha Mahsuri, IR36, IR64, BR11, and MTU 1010. These varieties generally preferred by farmers for their yield potential and quality traits are not resistant to drought. As they were bred for the irrigated ecosystem, these varieties provide high yield in nondrought years, but they show a high yield reduction in mild to moderate drought years and collapse completely in severe drought-stress years (Kumar et al., 2008). In field experiments conducted at IRRI during the dry seasons of 2006–2008, large-scale field-managed drought screening has recently

66

R. Serraj et al.

focused on the confirmation of drought-resistant breeding lines and identification of new potential donors of drought resistance within genebank germplasm collections and molecular breeding lines developed by backcrossing a series of donors to one of three elite recurrent parents (Lafitte et al., 2006), Oryza glaberrima introgression lines, hybrids, and their parental lines (Atlin et al., 2008; Serraj et al., 2008b), as well as mutants and transgenic lines (Herve´ and Serraj, 2009).

5.2. Heritability of yield components Breeders can make better gains in selection if differences among genotypes are large, selection intensity is kept high through screening large number of genotypes and where the trait under selection has high broad-sense heritability (Atlin et al., 2008). To plan breeding programs and allocate resources efficiently, breeders must have a clear idea of the repeatability or broadsense heritability (H) of estimates of genotypic value. H is the proportion of the variance among line means that is explained by genotypic differences. It is a measure of the reliability or precision with which one can detect differences under a given selection protocol: H¼

s2G s2G þ

s2GL l

þ

s2GY y

þ

s2GLY ly

s2

þ rlyE

;

where s2G , s2GL , s2GY , s2GLY , s2E , l, y, and r are the genotype (G), genotype location (GL), genotype year (GY), genotype location year (GLY), and within-trial error variances and the number of locations, years, and replicates of testing, respectively. s2G , s2GL , s2GY , s2GLY , and s2E are estimated from multienvironment trials (MET) repeated over locations and years within the TPE. It is important for breeding and cultivar testing programs to estimate these parameters (Atlin et al., 2008). Estimates of H can be maximized through careful management of drought screens, high levels of replication within trials, and increasing the number of trial locations and years of testing. Taking these steps will minimize the error variance (s2E ) and allow the detection of genotypic differences among lines (Atlin et al., 2008). Genotype location year (s2GLY ) and error variance (s2E ) have been found to be the largest contributors to random noise in field trials. The contribution of s2E can be reduced by choosing uniform screening sites, increasing the number of replications, adopting improved methods of controlling within-block error (by using lattice design or neighbor analyses), or increasing the number of locations or years of testing. The contribution of s2GLY can be reduced only by increasing the number of tests across locations or years.

67

Improvement of Drought Resistance in Rice

Earlier studies reported low H values for grain yield under stress compared to a control, suggesting low efficiency of selection for grain yield under stress compared to indirect selection for grain yield under a control or indirect selection for secondary traits (Blum, 1988; Edmeades et al., 1989; Rosielle and Hamblin, 1981). From a series of experiments conducted at IRRI under standardized managed screens, H estimates for a selection unit consisting of mean line yield from single-row plots in two replicated trials under upland stress ranged from 0.45 to 0.66 while in the upland control trials, the range was 0.47–0.78. The range was 0.48–0.67 in the lowland stress trials and 0.41–0.68 in the lowland nonstress trials. H estimates were similar for stress and nonstress experiments. In some cases (Table 1), H was higher for nonstress trials (Kumar et al., 2008). Similar H estimates for grain yield under stress and nonstress situations have also been reported in several other studies (Babu et al., 2003; Blum et al., 1999; Lanceras et al., 2004; Venuprasad et al., 2007a). These studies suggested that selection for grain yield under reproductive drought stress in rice can be conducted with the same level of precision as that achieved in nonstress trials.

5.3. Direct selection for yield under drought Recent studies at IRRI that included several populations have shown that secondary traits associated with drought resistance rarely have higher broadsense heritability (H ) than grain yield under drought stress and are often not highly correlated with grain yield (Bernier et al., 2008; Kumar et al., Table 1 Heritability of grain yield under nonstress and reproductive-stage drought stress in standardized managed screens Heritability Population

Situation

Nonstress

Stress

Reference

CT9993-5-10-1/ IR62266-42-6-2 Vandana/Way Rarem

Lowland

0.62

0.54

Upland

0.23

0.70

Kumar et al. (2007) Bernier et al. (2007) Kumar et al. (2008) Kumar et al. (2008) Kumar et al. (2008) Kumar et al. (2008)

IR55419-04/Way Rarem IR55419-04/IR64

Upland

0.61

0.45

Upland

0.48

0.49

IR55419-04/IR64

Lowland

0.58

0.61

Abhaya/Safri 17

Lowland

0.53

0.62

68

R. Serraj et al.

2008; Venuprasad et al., 2008). These studies also reported a significant positive response to direct selection for grain yield under stress.

5.3.1. Developing breeding populations Selection of appropriate parents for developing breeding populations is necessary for extracting good lines from segregating generations. Although breeders have different approaches to the selection of parents and have achieved success in different ways, many successful crosses have some common features (Atlin, 2003). Atlin (2003) suggested the use of at least one locally adapted popular cultivar as a parent to ensure the recovery of a high proportion of progenies with adaptation and quality acceptable to farmers. Choosing parents to complement the weakness of the other parent for disease/insect susceptibility has also been suggested. Kumar et al. (2008) reported crosses involving drought-resistant lines (e.g., IR42253-61-1-12-3, JGL 384, Badshah Bhog) and high-yielding, but drought-susceptible lowland cultivars that significantly outyielded crosses between two lowland high-yielding but drought-susceptible cultivars. Similarly, the best lines from the crosses involving drought-resistant donors also outyielded the best lines from crosses not involving drought-resistant donors. A strong positive effect of the use of drought-resistant donors in obtaining highyielding drought-resistant segregants has been observed (Kumar et al., 2008). Most of the traditional drought-resistant donors carry a linkage drag along with their drought resistance. The use of improved donors in crosses with a locally adapted variety could provide better opportunities.

5.3.2. Population size and selection intensity A drought-breeding program must be large to make progress. A major departure from an irrigated breeding program that is required for a rainfed system is the need for early-generation yield testing in selection environments that represent the TPE (Atlin, 2003; Jongdee et al., 2002). The drought-breeding program then aims to develop fixed lines for early testing at a large number of sites under stress and nonstress situations for identifying lines combining high yield potential and a good level of drought resistance. Several studies have reported positive but low rG between yield under stress and nonstress situations, indicating a low frequency of lines combining high yield and drought resistance. Thus, to identify sufficient desirable progenies, a large number of fixed lines from each breeding population should be grown under stress and nonstress conditions. This can be achieved only through large-scale breeding programs that produce many F2 or F3 progenies. A schematic diagram for one of the possible approaches of breeding for drought resistance is illustrated in Fig. 3.

Improvement of Drought Resistance in Rice

69

F2 (8000–10 000 plants; irrigated) (selection for plant type, yield potential, height, grain type: 400–500 selections)

F2:3 (400–500 plants from each cross; irrigated) (selection for plant type, yield potential, height, grain type: 200–250 selections)

F3:4 (200–250 plants from each cross; irrigated) (selection for yield potential, grain type, blast, BLB, quality traits: 50 selections – 4 plants per family)

F4:5 blocked yield nursery (50 families per population, 4 lines/family; drought and irrigated). Test in at least 2 SEs representative of TEs (selection for high yield potential and drought resistance, insects, diseases in SEs)

F5:6 OYT (1000 lines from several populations) Unreplicated yield testing under irrigated and drought conditions. Test at least in 5 SEs (selection for high yield potential and drought resistance, insects, diseases)

F6:7 (100 advanced lines from several populations) Replicated multi-environment trials under irrigated and drought conditions (selection for high yield potential and drought resistance, insects, diseases in SEs)

Participatory varietal on-farm testing (mother trial – 20 entries) Larger plot size in drought-prone areas Also entered in national system of testing for varietal release process

Participatory varietal on-farm testing (baby trial – 2 entries with check for each farmer) Larger plot size in drought-prone areas Promising entries continue testing in national system for varietal release process Identify lines for release, initiate breeder seed multiplication

Release of line and large-scale seed multiplication Distribution of truthfully labeled seed and later certified seed

Figure 3

Schematic approach in breeding for drought resistance.

Selection intensity must be high to make progress. Although increasing selection intensity is expensive, it is a simple and sure way to increase the selection response (Atlin, 2003). The initial population of lines evaluated must be large enough to permit intensive phenotypic selection for the highly heritable quality traits, plant type, and pest and disease resistance,

70

R. Serraj et al.

while retaining a population with adequate genetic variation for yield and drought resistance. Most breeding programs select for highly heritable characters such as plant type, quality traits, and resistance to insects and diseases in the early segregating generations, and then select for yield potential and drought resistance from the F5 generation. On-farm participatory varietal selection (PVS) trials have been introduced as one of the essential components of a drought-breeding program. Farmers’ preference analysis on varietal adaptability at different stages of the crop as well as postharvest analysis that includes cooking characteristics have been very helpful to breeders in assessing the likelihood of a breeding line becoming popular among farmers. 5.3.3. Combining drought resistance and high yield In rainfed areas, many traditional drought-resistant cultivars have evolved because of continuous farmer selection pressure in drought-prone environments over time. With the spread of high-yielding semidwarf varieties, farmers of rainfed regions have largely abandoned cultivating traditional drought-resistant but low-yielding cultivars. Considering the frequency of occurrence of drought in major drought-prone regions of eastern India being almost two in 5 years and in northeast Thailand being one in 5 years (Pandey et al., 2007), for a drought-resistant cultivar to become popular among farmers, it must combine resistance with high yield potential in favorable years. Transgressive segregants that yielded significantly higher than the parents under both drought and well-watered conditions have been reported in several studies (Bernier et al., 2007; Kumar et al., 2008; Venuprasad et al., 2008), reflecting the high possibility of combining yield potential with drought resistance. A positive response to selection under severe drought stress has been reported under similar stress levels in lowland and upland (Table 2). Kumar et al. (2008) reported that selection carried out under severe stress that reduced the grain yield of stress trials by 65–85% resulted in a positive response to selection both under upland situations in the dry season at IRRI and under lowland situations in the wet season in eastern India. Venuprasad et al. (2008) reported that, across populations, selection for yield under severe upland stress resulted in consistently greater yield gains in severely stressed environments than did selection under nonstress conditions. In the severely stressed dry-season upland trials, stress-selected lines significantly outyielded nonstress-selected lines in three of the four populations. Interestingly, stress-selected lines outyielded nonstress-selected lines in all four populations when evaluation was under natural wet-season stress. The advantage of stress-selected lines over nonstress-selected lines ranged from 14% to 41% and averaged 25% across populations (Venuprasad et al., 2008).

71

Improvement of Drought Resistance in Rice

Table 2 Response to selection on grain yield under severe stress in different populations % advantage over nonstress Situation Stress Nonstress Random selected Reference Mean performance of selected lines under

Population

a

Upland

585a

382

399

53

Kumar et al. (2008)

Upland

548a

417

402

31

Lowland 2076a 1111

848

87

790

690

38

260

220

115

Kumar et al. (2008) Kumar et al. (2008) Venuprasad et al. (2008) Venuprasad et al. (2008)

IR55419-04/ Way Rarem IR55419-04/ IR64 Abhaya/Safri 17 Apo/IR64

Upland

1090a

Apo/IR72

Upland

560a

Significantly superior over nonstress and random selected lines at p0.05.

6. Marker-Assisted Selection Recent developments in molecular marker technology and genomics have provided new approaches for discovering and tagging novel genes and alleles. These tools can enhance the efficiency of breeding programs through their use in MAS. In this way, the selection of target traits can be achieved indirectly using molecular markers that are closely linked to underlying genes or that have been developed from the actual gene sequence (Xu and Crouch, 2008). Drought has been a difficult trait to manage through conventional phenotypic selection and is one of the most ideal traits suitable for improvement through MAS.

6.1. QTLs for grain yield under drought An important reason for the slow progress in breeding for drought resistance has been the failure to identify QTLs with large and consistent effects that can be used for MAB. A recent literature survey shows that, although approximately 150 research papers reporting original QTL data are published yearly for Arabidopsis, soybean, rice, sorghum, maize, barley, and wheat, only a handful of studies have pursued their use in MAB/MAS.

72

R. Serraj et al.

In rice, considerable research effort has been devoted, primarily in the widely used populations CT9993/IR62266 and IR64/Azucena, to the mapping of QTLs for secondary drought-related traits such as root morphology and osmotic adjustment, but few loci with large effects for either of these traits have been identified (see Kamoshita et al., 2008 for a review). Progress in mapping QTLs for secondary traits associated with drought resistance in rice has been repeatedly reviewed (Bernier et al., 2008; Kamoshita et al., 2008; Price et al., 1997, 2002), but MAS for such QTLs has not been successfully used to improve rice yield under drought stress. Two recent reports indicated that QTLs with large effects on yield under drought stress may not be uncommon in rice. Bernier et al. (2007) reported a QTL on chromosome 12 in a Vandana/Way Rarem population explaining about 51% of the genetic variance for yield under severe upland drought stress over 2 years. Kumar et al. (2007) reported a major QTL for grain yield under lowland drought stress in a CT9993/IR62266 population on chromosome 1 explaining 32% of the genetic variance for the trait over 2 years. Most of the QTL identification studies carried out in rice and other crops have used two parents with large differences in the character of interest. A desirable QTL allele discovered in nonelite genetic material may not offer any improvement in the improved genetic background because the allele may already be ubiquitous in current varieties (Collins et al., 2008). The QTL from a particular genetic background usually show smaller effects or may disappear altogether in different genetic backgrounds, even under similar experimental conditions (Collins et al., 2008). With the prevalence of a few prominent rice varieties being cultivated on millions of hectares in major drought-prone areas in eastern India and northeastern Thailand, the two major drought-prone areas in the world, identifying major QTLs for grain yield in the background of improved mega varieties and introgressing the identified QTLs in the same background for improving the drought resistance of presently grown mega varieties have been suggested as an alternative approach. Recently, Venuprasad et al. (2009), using this approach and bulk segregant analysis (BSA), have identified two major QTLs located on chromosomes 2 and 3 for grain yield under lowland drought and one QTL located on chromosome 6 for yield potential and adaptation to aerobic soil conditions. These QTLs have been identified in the background of the drought-susceptible variety Swarna, grown on millions of hectares in India, Nepal, and Bangladesh, and they are being introgressed using MAB to improve the drought resistance of Swarna.

6.2. Selective genotyping and bulk segregant analysis In expanding efforts to screen germplasm for alleles with large effects on yield under stress, the cost of genotyping is a serious impediment, particularly if conventional QTL analysis (i.e., phenotyping and genome-wide scan

Improvement of Drought Resistance in Rice

73

of all individuals in a large RIL population) is used. Hence, alternative strategies based on selective genotyping (Navabi et al., 2008) could be cost effective for testing. Selectively genotyping the phenotypic tails of a population under selection can detect marker–QTL linkage with only a fraction of the genotyping effort required for conventional trait-based QTL analysis, and has been shown to have sufficient power to detect QTLs explaining 15% or more of the phenotypic variance for population sizes and broadsense heritabilities characteristic of cereal-breeding programs (Navabi et al., 2008). BSA is a particularly efficient form of selective genotyping wherein DNA samples of extreme individuals from each tail of a phenotypic distribution for a given trait are pooled and the two resultant bulks are genotyped (Michelmoore et al., 1991). Markers linked to a QTL affecting the trait are expected to be present at different frequencies in the contrasting tails, resulting in polymorphic expression of genotype signals (e.g., bands on an electrophoresis gel) between the two bulks. BSA has been widely used for the genetic analysis of qualitative traits such as disease resistance (e.g., Shen et al., 2003), but not for quantitative traits such as grain yield. However, there is at least one recent report in which BSA was used to identify markers linked to grain yield in rice (Shashidhar et al., 2005). Similarly, Quarrie et al. (1999) have used BSA to identify QTLs for grain yield under drought in maize.

6.3. Application of association studies in drought research Although progress in the development of improved germplasm for drought-prone environments is being made through a combination of empirical breeding and QTL mapping/marker-assisted selection, alternative methodologies can be used simultaneously to complement and sustain varietal development. Even though rice is more highly sensitive to water deficits than other cereals, large genotypic variation in drought resistance still exists within the cultivated rice gene pool (Ali et al., 2006; Jongdee et al., 2006; Monneveux and Ribaut, 2006). Empirical breeding programs exploit only a fraction of the available diversity, leaving many novel ‘‘droughttolerant’’ alleles undiscovered and unused. QTL mapping can identify genomic regions controlling quantitative traits for dissecting complex traits such as drought resistance and/or for use in MAB programs. However, the identification of QTLs requires that functionally relevant allelic variation occur within a mapping population; hence, many loci associated with improved performance under water-limited conditions will not be detected. Association studies for allele mining offer an alternative approach to enhance the use of genetic resource collections for crop improvement and can provide an interface between functional genomics and varietal development by applying the products of functional genomics (candidate

74

R. Serraj et al.

genes) to identify the best donor alleles within a wide range of germplasm and deliver these to breeding programs. Although most association analysis-based studies to date have concentrated on agronomically important traits for favorable environments (Breseghello and Sorrels, 2006; Cockram et al., 2008; Garris et al., 2003; Missaoui et al., 2007; Olsen et al., 2006; Thornsberry et al., 2001; Wilson et al., 2004), the application of association studies in the development of drought-resilient crops has good potential. In general, crop species have experienced strong selective pressure directed at genes controlling traits of agronomic importance under favorable conditions during selective breeding (Vigouroux et al., 2002). Thus, association analysis-based crop improvement strategies for unfavorable environments, which rely on extensive genetic diversity, have the potential to identify unique variants associated with improved drought resistance in a wide range of germplasm adapted to environments where periods of drought often prevail. In the case of rice, the availability of whole-genome sequences of indica and japonica subspecies and their knock-on effects on genetic studies (Kamoshita et al., 2008; Kathiresan et al., 2006; McCouch et al., 2003; McNally et al., 2006; Wanchana et al., 2008) provide a large number of potential droughtresponsive candidate genes and regions for partial validation through association studies. For allele mining, a representative collection needs to be identified that covers the relevant genetic diversity while maintaining the number of accessions at a workable level for genotypic and phenotypic characterization. The genetic variation within a diverse natural collection for allele mining is much higher than within a biparental mapping population. This wider background allows for genotype–phenotype associations to be tested in combinations that are not achievable in mapping populations. Population structure, as a result of domestication and selection, can cause a departure from the standard neutral expectations behind association studies, producing spurious associations and thereby reducing the effectiveness of association studies (Ersoz et al., 2007). Fingerprinting by anonymous markers across the genome can be applied to estimate the genome-wide effects of structure and used in statistical methods to remove the effects of population structure and kinship (Pritchard et al., 2000). In addition, only those traits and haplotype regions that have sufficient diversity within a subpopulation may be tested for associations. This implies the potential need for bridging populations of MAGIC (multiple-parent advanced intercross) lines, RILs, CSSLs, or BILs between representative subpopulations or the use of sufficient lines with admixture or founder effects so that relevant variation can be explored at a statistically significant power. Such a design for rice would be the equivalent of the nested association mapping panels used in maize (Yu et al., 2008). In rice, an association approach is being applied to determine the effect of molecular variation within drought-responsive candidate genes on

Improvement of Drought Resistance in Rice

75

phenotypic response to drought stress. A representative collection was selected from the International Rice Genebank by a semistratified scheme using species and ecocultural types, characterization data, geographical information, and estimates of usage and deployment based on pedigree information in the International Rice Information System (IRIS, http:// iris.irri.org) (Bruskiewich et al., 2003). From this set, a subset of 1536 accessions was selected that maintained the genetic diversity while its smaller size was more suitable for the medium- to low-throughput activities of genotyping and phenotyping. Accessions were genotyped with 48 SSR markers and population structure determined using Instruct (Gao et al., 2007). Five distinct subgroups (aus, aromatic, indica, tropical japonica, and temperate japonica) plus ‘‘admixed’’ types with alleles from two or more subgroups were identified, showing good correspondence with previous isozyme group classifications (Glaszmann, 1987). Many genomic regions and genes have been associated with performance under drought stress in rice (Kathiresan et al., 2006; Lafitte et al., 2004; Salekdeh et al., 2002). Candidate genes were identified by combining evidence from expression analyses, including transcriptomics, proteomics and metabolomics, functional annotation, QTLs, and shifts in allele frequencies under selection during drought stress. Eight candidate genes with strong evidence of their involvement in the regulation of responses to drought stress were selected. Natural molecular variation within candidate genes or single nucleotide polymorphisms (SNPs) was determined using EcoTILLING (Comai et al., 2004). Mismatch patterns were detected on agarose-based gels (Raghavan et al., 2007) in samples contrasted independently against Nipponbare ( japonica type) and IR64 (indica type). The number of SNPs identified per gene ranged from 7 to 17 while the number of haplotypes identified per gene ranged from 7 to 14. Sequencing of selected accessions for each haplotype revealed that most SNPs were transitions and the percentage of variation due to indels (insertion or deletion events) was low within all candidate genes except sucrose synthase. Accessions were phenotyped under field conditions over three consecutive dry seasons. Yield stability under drought stress is largely dependent on the timing of stress, with the irreversible processes during the reproductive stage being the most sensitive. Within the reproductive phase, the largest reduction in grain yield occurs when drought stress coincides with flowering (Cruz and O’Toole, 1984). Large variation in days to flowering (99 days) made it unfeasible to apply drought stress exactly at flowering for all accessions; therefore, stress was imposed at the vegetative stage and secondary traits associated with performance under drought stress (plant water status, biomass accumulation, and flowering delay) were measured. Large phenotypic variation was observed within each subpopulation. Preliminary tests have indicated association between one of the indica haplotypes and biomass accumulation under stress for some of the candidate genes.

76

R. Serraj et al.

One major drawback with this approach is the requirement for prior knowledge of candidate genes. Drought stress is known to affect thousands of genes and new candidate genes and regions are continuously being identified. Falling costs and new advances in genotyping are increasing the possibility of genome-wide association studies in rice. High-density whole-genome scans have the potential to identify new genomic regions associated with interesting phenotypes without assuming their involvement. The OryzaSNP project (http://www.oryzasnp.org) has discovered 160,000 SNPs across the entire rice genome in 20 diverse rice varieties (McNally et al., 2009). These data provide a foundation to create a high-density platform for deep genotyping of diverse rice germplasm that will allow genome-wide association studies with phenotyping data for drought-resistance traits. The advent of whole-genome SNP information has great potential; however, the success of association studies, like other molecular-based crop improvement strategies, still relies on accurate phenotyping of genetic material. Association studies depend on phenotypic variation, and the phenotypic variation within association collections is generally larger than that within mapping populations (Ersoz et al., 2007). Large genetic variation increases precision within association studies but can cause problems in the precision and validity of phenotyping. In rice, the application of association studies for drought-resistance improvement is complicated by flowering time. Successful application of association studies for drought breeding programs will require this key stage of yield failure to be targeted. Large genetic variation in flowering time exists in rice; hence, because of the broad diversity within an association studies population, variation within flowering time will also be increased. Photoperiod-sensitive rice accessions, in which the date of heading is partially determined by the critical daylength, will not flower at all locations. Furthermore, large differences in phenology within a diverse collection complicate the ability to impose drought stress in all accessions at flowering. All of these factors will need to be considered in the design of phenotyping experiments for association studies.

6.4. Marker-assisted breeding In rice, progress has been made with the introgression of major genes for improving tolerance of bacterial leaf blight, brown spot, brown planthopper, and for several other traits. However, there are few reports of the introgression of major QTLs in rice. Recently, Sub1, a major QTL for submergence tolerance (Xu et al., 2006), was introgressed into Swarna, Sambha Mahsuri, and BR11 mega varieties. The newly developed cultivars having the Sub1 QTL have more than 97% of the genome from the recurrent varieties (Swarna, Sambha Mahsuri, BR11), except for the Sub1 region on chromosome 9, and provide enhanced submergence tolerance for

Improvement of Drought Resistance in Rice

77

up to 14 days to these varieties (Septiningsih et al., 2009). In the case of drought, QTLs related to grain yield under drought stress have been reported on numerous occasions in rice (Babu et al., 2003; Bernier et al., 2007; Kumar et al., 2007; Lafitte et al., 2004; Lanceras et al., 2004; Yue et al., 2006), but there have been no reports of a successful use of such QTLs in MAS (Bernier et al., 2008). The lack of repeatability of QTL effects across different populations (QTL genetic background interaction) and across environments (QTL environment interaction) are two factors that limit the use of QTLs for MAS by plant breeders (Bernier et al., 2008; Courtois et al., 2003; Lafitte et al., 2004; Price et al., 2002; Serraj et al., 2005). A good example of the problem of QTL environment can be found from the work published on the rice population CT9993/IR62266, phenotyped for grain yield under drought stress in a wide range of environments. Quantitative trait loci for grain yield under drought stress have been reported under upland conditions in southern India in two trials (Babu et al., 2003) and in Israel in one trial (Blum et al., 1999), under lowland conditions for 2 years in eastern India (Kumar et al., 2007), and under transplanted linesource conditions in one trial in Thailand (Lanceras et al., 2004). The number of loci reported to affect grain yield under drought was five in southern India, two in Israel, four in Thailand, and one in eastern India. Of all these QTLs, none was detected at more than one site, indicating a high level of QTL by environment interaction for all loci and rendering it impractical to use any of them in MAB. A considerable effort has been devoted to the mapping of root-related QTLs in rice. Large chromosomal segments corresponding to QTLs associated with root length in a population derived from a cross between the deep-rooted upland variety Azucena and the shallow-rooted lowland variety IR64 were introgressed into the IR64 background. Most of the lines carrying the desired introgressions failed to have deeper roots than IR64 (Shen et al., 2001). The lack of effect of the QTL-containing segments on root length and yield may be because those QTLs were responsible for a small proportion of the total phenotypic variation (6–18%) and had not been fine-mapped. The QTL regions were very long and the desirable genes may have been lost due to recombination during backcrossing. Azucena rootrelated QTLs have also been introduced into the indica variety Kalinga III, but only one of the five target QTLs had an effect on root length and none had a consistent effect on grain yield under water-limited conditions (Steele et al., 2006). These results indicate that only fine-mapped alleles with large confirmed effects on performance under stress are appropriate targets for MAS. The recently identified major QTLs for grain yield under drought in the background of the improved mega varieties (Venuprasad et al., 2009) have potential for improving the drought resistance of cultivar Swarna through introgression of the identified region after fine mapping. Similarly, grain yield and drought resistance of the upland cultivars can be improved

78

R. Serraj et al.

using major QTLs for grain yield under upland drought stress (Bernier et al., 2007) using marker-assisted recurrent selection (MARS).

7. Drought-Resistance Genes and GM Technology 7.1. Gene discovery The first genome sequences of rice were available in 2002 and the finished genome was published in 2005 (IRGSP, 2005), with the genome annotation available online (Ohyanagi et al., 2006; Yuan et al., 2003). The crop research community is working hard to annotate the function of the resulting crop genomic sequences. For rice, the Rice Annotation Project databases in Japan (http://rapdb.dna.affrc.go.jp/) and the U.S. (http://rice. plantbiology.msu.edu/) provide complementary but mildly competing genome annotation data sets curated by teams of experts. The J. Craig Venter Institute (formerly The Institute for Genome Research or ‘‘TIGR’’) continues to provide several sets of plant annotation resources across many genomes, including rice (http://www.jcvi.org/cms/research/ groups/plant-genomics/). The Gramene comparative grasses resource (http://www.gramene.org) has worked steadily since 2000 to provide the plant community with a curated genetic, genomic, and comparative genomics database for major crop species, including rice. The accurate and complete functional annotation of genomes remains a daunting task ( Jung et al., 2008). Efforts to characterize gene function are largely driven by a paradigm of ‘‘intersecting evidence’’ using experimental results about genes from the following: Position. QTL results that link segregating chromosome markers with plant traits. Function. Rice genome annotation with literature-documented biochemical or sequence homology analysis, combined with genetic dissection of rice mutants, can reveal the role of specified genes in specific biological processes. Expression. Gene expression experiments at the transcript, protein, and metabolic level can associate the expression of specific genes with specific processes, both constitutively and under conditional treatments. Selection. Analysis of genetic resource collections (through association genetics) and bulk population selection experiments can identify genomic regions in linkage disequilibrium with traits or alleles conserved under selection pressure in molecular breeding. This provides additional support for the role of specified loci and alleles in pertinent biological processes and phenotypes.

Improvement of Drought Resistance in Rice

79

Crop models. This is a relatively new source of evidence, but it is expected that linking gene systems information with whole-crop models will help support or refute gene candidacy in specific field-level processes, such as drought-stress responses.

The development of effective information systems for capturing and integrating such intersecting experimental evidence is the focus of research activities in crop bioinformatics. These efforts will, hopefully, allow researchers to take intersection sets of gene candidates and to narrow their focus to a few key candidate gene loci or alleles for further cost-effective laboratory or field validation.

7.2. Rice transformation Genetic enhancement in rice through transgenic approaches complements conventional breeding when the desired gene is not available in the gene pool or when the expression of the gene is not at a preferred level, timing, or tissue, thus needing modification of regulatory elements. The molecular mechanisms regulating responses of plant genes to water stress include the sensing mechanisms of osmotic stress, modulation of the stress signals to cellular signals, transduction of cellular signals to the nucleus, transcriptional control of stress-inducible genes, and the expression of specific stress-related genes and metabolites. Transgenic technology also allows functional validation studies of the genes responsible for these molecular mechanisms. Two major groups of genes have been generally employed to improve stress resistance by gene transfer: genes encoding functional proteins with known enzymatic or structural functions and genes encoding for regulatory proteins of transcription and signal transduction. The first group includes the functional proteins that are involved in dehydration-stress resistance and cellular adaptation such as protection factors of macromolecule chaperones, osmoprotectants, water channel proteins, detoxification enzymes, proteases, and the second group includes transcription factors, protein kinases, and enzymes of phospholipid metabolism, and ABA biosynthesis (Bhatnagar-Mathur et al., 2008; Shinozaki and Yamaguchi-Shinozaki, 2007). Slow progress has been achieved, however, in developing rice varieties adapted to drought stress through transgenic breeding. Initial transgenic experiments were generally based on the use of single genes was usually restricted to the modulation of the plant response to one stress factor. Furthermore, they relied on constitutive expression of the genes that can have a detrimental effect in certain cases (i.e., Dubouzet et al., 2003). The lack of desirable progress is attributable to the fact that resistance to abiotic stress is a very complex multigenic trait influenced by coordinated and differential expression of a network of genes (Garg et al., 2002), and the challenge to translate the drought researches from a different branch of

80

R. Serraj et al.

biology into actual crop yield improvement (Sinclair et al., 2004). Plant performances, under drought stress and under normal conditions, have to be evaluated to ensure that the overall performance of a crop under optimal conditions is not impaired (Herve´ and Serraj, 2009; Xiao et al., 2009). 7.2.1. Perspectives of using functional protein genes A series of transgenic work in rice (Table 3) involves the introduction of genes encoding key enzymes regulating synthesis of compatible solutes that function as osmoprotectants, such as amino acids (e.g., proline), synthesis of osmolytes, quaternary and other amines (e.g., glycinebetaine and polyamines), and a variety of sugars and sugar alcohol (e.g., trehalose). However, several reports (Bhatnagar-Mathur et al., 2008; Serraj and Sinclair, 2002; Turner et al., 2007) indicated that osmotic adjustment generally results in little benefit for yield under water stress. LEA are a major group of proteins that typically accumulate during the late stages of embryogenesis or in response to dehydration, low temperature, salinity, or exogenous ABA treatment that indicate their responsiveness to cellular dehydration (Ramanjulu and Bartels, 2002). LEA proteins may also act as a chaperone-like protective function against cellular damage (Umezawa et al., 2006). Transgenic rice lines with enhanced stress tolerance expressing LEA protein were reported by Babu et al. (2004), Xiao et al. (2007), and Xu et al. (1996). Xiao et al. (2007) reported significant enhancement of drought resistance and no yield penalty observed by overexpressing OsLEA3-1 gene under field conditions. Delayed leaf wilting and senescence caused by water-deficit stress were observed by Xiao et al. (2007) and Xu et al. (1996). As a consequence of dehydration stress, plants accumulate reactive oxygen species (ROS), which damage cellular structures (Smirnoff, 1993). Under optimal conditions, leaves are equipped with sufficient antioxidant enzymes and metabolites to cope with ROS. The accumulation of enzymes such as superoxide dismutases, ascorbate peroxidases, catalases, glutathioneS-transferases (GST), and glutathione peroxidases has been observed during stress conditions (Ramanjulu and Bartels, 2002). Wang et al. (2005) reported improvement in rice abiotic stress tolerance through a detoxification strategy under the control of an inducible promoter. A higher expression of SOD protected the photosynthetic apparatus since transgenic plants showed a higher net photosynthetic rate than the wild-type (WT) controls under drought stress and the transgenic plants had a normal growth habit and appearance when grown under nonstressed conditions. A water channel transporter aquaporin gene has also been reported to enhance drought resistance in rice (Lian et al., 2004).

Table 3 Transgenic rice for drought resistance: gene, gene product, and putative mechanism Mechanism

Osmoprotectant

Compound

Gene

Reference

Proline

D -pyrroline-5-carboxylate synthetase (P5CS) Choline oxidase (cox), betaine-aldehyde dehydrogenase (BADH) Arginine decarboxylase (adc)

Zhu et al. (1998)

Glycinebetaine Polyamines Trehalose

Chaperones (protection factors of macromolecule) Detoxification enzymes

LEA protein

ROS-scavenging enzyme superoxide dismutase Aquaporin

Water channel transporters Regulatory protein Transcription factors (transcription factor) Regulatory protein Protein kinases (signaling mechanism) Regulatory protein ABA biosynthesis key enzyme (signaling mechanism)

1

Trehalose-6-phosphate synthase/ phosphatase (TPSP) fusion; TPS and TPP (trehalose-6-phosphate phosphatase) HVA1 PMA OsLEA3-1 Manganese superoxide dismutase (MnSOD)

Nakamura et al. (1997) and Su et al. (2006) Capell et al. (2004) and Roy and Wu (2001) Garg et al. (2002) and Jang et al. (2003) Babu et al. (2004) and Xu et al. (1996) Cheng et al. (2002) Xiao et al. (2007) Wang et al. (2005)

RWC3

Lian et al. (2004)

CBF/DREB, NAC, HD-Zip, WRKY

Agalou et al. (2008), Hu et al. (2006), Ito et al. (2006), Oh et al. (2005), and Wu et al. (2008) Saijo et al. (2000) and Xiong and Yang (2003)

Calcium-dependent protein kinase (CDPK), MAPK 9-cis-epoxycarotenoid dioxygenase (NCED), LOS5

Xiao et al. (2009)

82

R. Serraj et al.

7.2.2. Transcription factors and signaling genes Significant progress has been made in analyzing the complex cascades of genes and, in particular, those involved in the specificity and crosstalk in stress signaling pathways (Shinozaki and Yamaguchi-Shinozaki, 2007). One of the promising approaches to improve the stress resistance of rice by gene transfer would be through the use of genes encoding protein factors that are involved in regulation of gene expression and signal transduction since they can regulate many of the downstream genes involved in stress resistance when combined with an appropriate promoter. In rice, DRE/CBF transcription factors have been reported to be useful in improving stress resistance in transgenic plants, through influencing the expression of a number of stress-related target genes (Oh et al., 2005; Yamaguchi-Shinozaki and Shinozaki, 2005). ABRE and DRE/CRT are cis-acting elements where the DRE/CBF transcription factors bind and are involved in ABA-dependent and ABA-independent gene expression, respectively, in response to abiotic stress. Oh et al. (2005) reported the Arabidopsis transcription factors, CBF3 and ABF3, activated 12 and 7 target genes in transgenic rice plants, respectively. The target genes together with 13 and 27 additional genes were induced on stress treatments, consequently making the transgenic plants more resistant to stress conditions. A stress-responsive NAC transcription factor, SNAC1, was characterized in rice under field conditions (Hu et al., 2006) and showed 22–34% higher seed setting than a control in the field under severe drought-stress conditions at the reproductive stage. This gene was specifically induced in the guard cells of rice under drought-stress conditions. Overexpression of this gene in rice resulted in significantly increased stomata closure and drought resistance in the drought-stressed field conditions while the photosynthesis rate and yield of transgenic plants were not affected under normal growth conditions. Members of several different classes of transcription factors have been implicated in stress responses, including MYC, MYB, bZIP, AP2, homeodomain (HD-Zip), and zinc-finger proteins. Homeodomain-leucine zipper (HD-Zip) genes encode proteins that have only been identified in plants so far and are thought to regulate development and responses to environmental cues (Ramanjulu and Bartels, 2002). Agalou et al. (2008) constitutively overexpressed one of the HD-Zip genes, Oshox4, in rice and Arabidopsis and showed enhanced resistance to drought in Arabidopsis. They indicated that HD-Zip may function in ethylene signaling pathways. Another approach suggested to contribute to enhancing drought resistance in transgenic plants is the use of constitutively active forms of transcription factors modified by point mutations (Shinozaki and Yamaguchi-Shinozaki, 2007). Transgenic plants expressing a phosphorylated form of AREB1 with multisite mutations induced many ABA-

Improvement of Drought Resistance in Rice

83

responsive genes without exogenous ABA application (Fujita et al., 2005; Furihata et al., 2006). The other regulatory factors have been shown useful for improving drought resistance in rice are genes involve in ABA signaling pathways. NCED (Iuchi et al., 2001) and LOS5 (Xiong et al., 2001) both regulate many stress-related genes in transgenic plants; their overexpression in transgenic plants led to enhanced drought resistance (Xiao et al., 2009). ABA is synthesized de novo primarily in response to drought and high salinity stress. Other regulatory factors, such as MAPK, also involved in ABA biosynthesis, and CDPK, were reported to enhance drought resistance in rice (Saijo et al., 2000; Xiong and Yang, 2003). Recently, Xiao et al. (2009) compared seven genes under constitutive (Actin) and stress-inducible (HVA22P) promoters. Transgenic families of eight constructs (HVA22P:CBF3, HVA22P:NPK1, Actin1:LOS5, HVA22P:LOS5, Actin1:ZAT10, HVA22P:ZAT10, Actin1: NHX1, and HVA22P:NHX1) involving five genes (CBF3, NPK1, LOS5, ZAT10, and NHX1) showed significantly higher relative yield than WT under both drought-stress field and PVC tube conditions. Among the seven stress-responsive genes used in this study, transgenic families of three genes (LOS5, ZAT10, and NHX1) under the control of either the stress-inducible promoter (HVA22P) or the constitutive promoter (Actin1) showed significantly higher relative yields than WT in two drought-stressed conditions. Generally, promoters from drought-inducible genes have advantages in maximizing the effects of transgenes when transcription factors are employed to obtain stress-resistance improvement compared to constitutive promoters since detrimental effects were clearly observed in Arabidopsis (Agalou et al., 2008; Dubouzet et al., 2003). However, the report of Oh et al. (2005) showed that Ubi1:CBF3 and Ubi1:ABF3 rice exhibited neither growth inhibition nor visible phenotypic alterations in rice, despite the constitutive expression of the transgenes. Oh et al. (2005) suggested this may have occurred because fewer target genes are activated by CBF3 or ABF3 in rice than in Arabidopsis, and, hence, the effects on plant growth might be minimized in rice. Another possibility is that lower expression of the transgene occurred due to its integration position and possibly reduced any detrimental effects. A large number of genes have been reported to enhance drought-stress resistance in rice. Nevertheless, although the transformation methodology is somehow becoming more standard, integration is still random; therefore, a large population of transgenic events is needed for evaluation, and only a few studies provide actual data about enhanced drought resistance under field conditions (Hu et al., 2006; Wang et al., 2005; Xiao et al., 2007, 2009). The availability of regulatory-proven and drought-standardized field facilities has limited globally the ability to conduct field trials. The protocols for

84

R. Serraj et al.

plant evaluation and the parameters used to assess plant resistance are diverse and difficult to compare and the low number of independent transgenic lines and poor assessment of a gene effect versus an overall effect due to gene insertion and somaclonal variation could be a drawback (Herve´ and Serraj, 2009). Overall, transgenic technology allows the validation of gene function and current progress shows that this approach can be complementary for other breeding approaches if the phenotypic evaluation is properly conducted (Herve´ and Serraj, 2009).

7.3. Bioinformatics and gene functional analysis Bioinformatics research and development now offer a rich combination of protocols, tools, databases, and computing infrastructure that can be applied to help answer biological research questions, often at a significant savings in time and laboratory resources. Bioinformatics can integrate information across a diverse collection of crop data about germplasm, genotype, phenotype, cellular expression (of transcripts, proteins, and metabolites), growth characteristics, applied treatments, and environmental conditions. In this section, we review a few of the bioinformatics activities and resources that have evolved over the past several years for crop research in general and for rice and drought-resistance research in particular. A few years ago, IRRI commissioned the development and sequencing of an approximately 10,000-clone drought-stressed panicle expressed sequence tag (EST) library. The sequence annotation for this library is publicly available at http://www.iris.irri.org/clone. Sequence, annotation, and trace files are available for each clone. This library was subsequently arrayed on microarray slides at IRRI and various hybridization experiments were undertaken. Experiments using these panicles cDNA microarray slides include a case study involving hybridization of field-grown panicle samples from drought-resistant and -susceptible rice germplasm. The results indicated contrasting drought responses among both upland- and lowlandadapted cultivars and also between traditional and improved upland types (Kathiresan et al., 2006). QTL position data are also being integrated with gene data for managing and visualizing information from various experiments. Open-source software from the Generic Model Organism Database (GMOD, http://www. gmod.org) consortium, in particular, the Genome Browser (GBrowse), is being used for this purpose. The Comparative Mapping Tool (CMAP) is also applied to the storage, comparison, and visualization of genetic and physical maps. The Gramene resource (http://www.gramene.org) uses these GMOD tools to integrate and publish QTL information for many traits, with a few of these studies being drought-response-related. An example of data integration is an anchoring of QTL maps to the rice

Improvement of Drought Resistance in Rice

85

physical map in order to find the candidate genes underlying a droughtresistance QTL flanked by RM212–RM319 on chromosome 1 (Wang et al., 2005). A set of candidate genes of known or inferred function was identified in this region using rice genome annotation. Published literature supports the candidacy of some of these genes in drought-stress response (Rabbani et al., 2003). The Generation Challenge Program (GCP, http://www.generationcp. org) research themes are directed to crop improvement through genomics and comparative biology across species, as well as molecular and phenotypic characterization of genetic resources to discover valuable alleles for crop improvement. One major research and development subprogram of the GCP focuses on crop informatics. This bioinformatics subprogram is striving to develop global public standards for crop information management, including a comprehensive crop scientific domain model (see http:// pantheon.generationcp.org; Bruskiewich et al., 2006) and a platform of tools for accessing and analyzing information available from the globally networked databases of GCP partners and other data sources. Of particular interest to drought research is the Comparative Plant Stress-Responsive Gene Catalog (Wanchana et al., 2008; see http:// dayhoff.generationcp.org) being developed to facilitate the integration of knowledge about stress-responsive genes across crops. This public catalog is a compendium of protein families, phylogenetic trees, multiple sequence alignments, and associated experimental evidence. The central objective of this resource is to elucidate orthologous and paralogous relationships between plant genes that may be involved in response to environmental stress, mainly abiotic stresses such as drought. The Web-based graphical user interface (GUI) of the resource includes query and visualization tools that allow diverse searches and browsing of the underlying project database. So far, 500 genes collected from the literature and GCP partners have been compiled into this stress inventory tool. Based on this gene collection, 180 gene families have been annotated, validated, and loaded into the stress-gene catalog database. The drought-responsive gene families currently available in the database include drought-responsive element binding factors (DREB), aquaporins, cell-wall and vacuolar invertases, sucrose synthase, and sucrose–phosphate synthase. These candidate genes have been well characterized by several experiments and their functions have been proven to be involved in pathways of drought-stress resistance. The International Crop Information System (ICIS, http://www.icis. cgiar.org) is one example of a computerized database system and suite of tools for general integrated management and use of genealogy, nomenclature, evaluation, and characterization data for a wide range of crops. The International Rice Information System (IRIS, http://www.iris.irri.org; Bruskiewich et al., 2003; McLaren et al., 2005) is of specific interest as the largest curated public rice installation of ICIS. In addition to ICIS databases,

86

R. Serraj et al.

Table 4 Partial inventory of publicly accessible plant databases Database

Organism

URL

TAIR NASC MATDB RAPDB Gramene TIGR plant genomes MaizeGDB Barleybase SINGER GRIN

Arabidopsis thaliana Arabidopsis thaliana Arabidopsis thaliana Oryza sativa Comparative grasses Oryza sativa

http://www.arabidopsis.org http://www.nasc.org http://www.mips.gsf.de http://rapdb.lab.nig.ac.jp http://www.gramene.org http://www.tigr.org/ plantProjects.shtml http://www.maizegdb.org http://www.barleybase.org http://www.singer.cgiar.org http://www.grin.usda.gov

Maize Barley (microarray data) Plant genetic resources Plant genetic resources

many additional excellent public online plant and crop databases are available. Table 4 contains a partial inventory of such resources. The establishment of the International Rice Functional Genomics Consortium (IRFGC, http://www.iris.irri.org/IRFGC) aims to coordinate research in the post-sequencing ‘‘functional genomics’’ era by exploring ways to consolidate international rice functional genomics resources and to build common strategies. The consortium is striving to encourage sharing and consolidation of several useful resources for gene characterization of rice in the areas of genomic stocks, expression arrays, proteomics, and bioinformatics.

8. Conclusions and Future Prospects Although the need for precise characterization of the drought-prone TPE has long been emphasized, and this effort is yet to be carried out systematically across the rainfed rice environments in South and Southeast Asia and sub-Saharan Africa. Simulation models can play a role in both characterizing and enhancing the precision and integration of phenotyping either by linking model coefficients directly to or more heuristically guiding integrated phenotyping approaches. Increased crop yield and water productivity require the optimization of the physiological processes involved in the initial critical stages of plant response to soil drying, water-use efficiency, and dehydration-avoidance mechanisms. Overall, it is now well accepted that the complexity of the drought syndrome can be tackled only with a holistic approach integrating plant breeding with physiological dissection of resistance traits and molecular genetic tools together with

87

Improvement of Drought Resistance in Rice

agronomical practices that lead to better conservation and use of soil moisture and matching crop genotypes with the environment. Some of the steps involved in this multidisciplinary approach (Fig. 4) are described below: 1. Define the target drought-prone environment(s) and identify the predominant type(s) of drought stress and the rice varieties preferred by farmers. Define the phenological and morphological traits that contribute substantially toward adaptation to drought stress(es) in the target environment(s). 2. Use simulation modeling and systems analysis to evaluate crop response to the major drought patterns and to assess the value of candidate physiological traits in the target environment. 3. Develop and refine appropriate screening methodologies for characterizing genetic stocks that could serve as donor parents for the traits of interest. 4. Identify the genetic stocks for various putative constitutive and inducible traits in the germplasm and establish genetic correlations between the traits of interest and the degree of adaptation to the targeted drought stress. 5. Harness functional genomics, transgenics, reverse genetics, and bioinformatics tools and resources to understand the genetic control of the relevant traits.

Germplasm and breeding lines

Historical climate series simulation models, GIS

Target environments

Physiological bases of genetic variation Traits: Phenology, root water uptake, water use, biomass accumulation and partitioning

Field selection Screening and evaluation

QTL & LD mapping MAS

Transformation Functional genomics

Yield = TR ⫻ TE ⫻ HI Multi-location trials, PVS, dissemination

Figure 4 Physiological framework of an integrated strategy for genetic enhancement of crop grain yield and its components under drought. TR, total plant water uptake; TE, transpiration efficiency; HI, harvest index.

88

R. Serraj et al.

6. Use mapping populations and/or linkage disequilibrium mapping to identify genetic markers and QTLs for traits that are critical for drought resistance. 7. Incorporate some of the components of relevant physiological traits into various agronomic genetic backgrounds to provide a range of materials with specific traits of interest (i.e., developing NILs, RILs, and BC populations) for improving drought adaptation of locally adapted varieties. 8. Test the MAS products under well-managed field screening and in farmer participatory multilocation trials. This framework is discussed in more detail in relation to a strategy for a Drought Frontier Project on drought-resistant rice (Serraj and Atlin, 2008).

REFERENCES ¨ vernas, E., Johannesson, H., Zhu, X., Estiati, A., Agalou, A., Purwantomo, S., O De Kam, R. J., Engstro¨m, P., Slamet-Loedin, I. H., Zhu, Z., Wang, M., Xiong, L., et al. (2008). A genome-wide survey of HD-Zip genes in rice and analysis of droughtresponsive family members. Plant Mol. Biol. 66, 87–103. Ali, A. J., Xu, J. L., Ismail, A. M., Fu, B. Y., Vijaykumar, C. H. M., Gao, Y. M., Domingo, J., Maghirang, R., Yu, S. B., Gregorio, G., Yanaghihara, S., Cohen, M., et al. (2006). Hidden diversity for abiotic and biotic stress tolerances in the primary gene pool of rice revealed by a large backcross breeding program. Field Crop Res. 97, 66–76. Amani, I, Fischer, R. A., and Reynolds, M. P. (1996). Canopy temperature depression association with yield of irrigated spring wheat cultivars in a hot climate. J. Agron. Crop Sci. 176, 119–129. Atlin, G. (2003). Improving drought tolerance by selecting for yield. In ‘‘Breeding Rice for Drought-Prone Environments’’ (K. S. Fischer, R. Lafitte, S. Fukai, G. Atlin, and B. Hardy, Eds.), pp. 14–22. International Rice Research Institute, Los Ban˜os, Philippines. Atlin, G. N., Lafitte, H. R., Tao, D., Laza, M., Amante, M., and Courtois, B. (2006). Developing rice cultivars for high-fertility upland systems in the Asian tropics. Field Crop Res. 97, 43–52. Atlin, G. N., Venuprasad, R., Bernier, J., Zhao, D., Virk, P., and Kumar, A. (2008). Rice germplasm development for drought-prone environments: Progress made in breeding and genetic analysis at IRRI. In ‘‘Drought Frontiers in Rice—Crop Improvement for Increased Rainfed Production’’ (R. Serraj, J. Bennett, and B. Hardy, Eds.), IRRI World Scientific Publishing, Singapore. Babu, R. C., Nguyen, B. D., Chamarerk, V., Shanmugasundaram, P., Chezhian, P., Jeyaprakash, P., Ganesh, S. K., Palchamy, A., Sadasivam, S., Sarkarung, S., Wade, L. J., and Nguyen, H. T. (2003). Genetic analysis of drought resistance in rice by molecular markers: Association between secondary traits and field performance. Crop Sci. 43, 1457–1469. Babu, R. C., Zhang, J., Blum, A., Hod, T. H. D., Wue, R., and Nguyen, H. T. (2004). HVA1, a LEA gene from barley, confers dehydration tolerance in transgenic rice (Oryza sativa L.) via cell membrane protection. Plant Sci. 166, 855–864. Balasubramanian, V., Sie, M., Hijmans, R. J., and Otsuka, K. (2007). Increasing rice production in sub-Saharan Africa: Challenges and opportunities. Adv. Agron. 94, 55–133.

Improvement of Drought Resistance in Rice

89

Ba¨nziger, M., Setimela, P. S., Hodson, D., and Vivek, B. (2006). Breeding for improved abiotic stress tolerance in maize adapted to southern Africa. Agric. Water Manage. 80, 212–224. Bernier, J., Kumar, A., Ramaiah, V., Spaner, D., and Atlin, G. (2007). A large-effect QTL for grain yield under reproductive-stage drought stress in upland rice. Crop Sci. 47, 507–516. Bernier, J., Atlin, G. N., Serraj, R., Kumar, A., and Spaner, D. (2008). Breeding upland rice for drought resistance (a review). J. Sci. Food Agric. 88, 927–939. Bernier, J., Serraj, R., Kumar, A., Venuprasad, R., Impa, S., Gowda, V., Owane, R., Spaner, D., and Atlin, G. (2009). Increased water uptake explains the effect of qtl12.1, a large-effect drought-resistance QTL in upland rice. Field Crop Res. 110, 139–146. Bhatnagar-Mathur, P., Jyostna-Devi, M., Reddy, D. S., Lavanya, M., Vadez, V., Serraj, R., Yamaguchi-Shinozaki, K., and Sharma, K. K. (2007). Stress-inducible expression of Arabidopsis thaliana DREB1A in transgenic peanut (Arachis hypogaea L.) improves transpiration efficiency under water limiting conditions. Plant Cell Rep. 26, 2071–2082. Bhatnagar-Mathur, P., Vadez, V., and Sharma, K. S. (2008). Transgenic approaches for abiotic stress tolerance in plants: Retrospect and prospects. Plant Cell Rep. 27, 411–424. Bidinger, F. R., Mahalakshmi, V., and Rao, G. D. P. (1987). Assessment of drought resistance in pearl millet [Pennisetum americanum (L.) Leeke]. II. Estimation of genotype response to stress. Aust. J. Agric. Res. 38, 49–59. Blum, A. (1988). Plant Breeding for Stress Environments. CRC Press, Boca Raton, FL. Blum, A., Mayer, J., Golan, G., and Sinmena, B. (1999). Drought tolerance of a doubledhaploid line population of rice in the field. In ‘‘Genetic Improvement of Rice for Water-Limited Environments’’ (O. Ito, J. O’Toole, and B. Hardy, Eds.), pp. 319–330. IRRI, Los Ban˜os, Philippines. Boonjung, H., and Fukai, S. (1996). Effects of soil water deficit at different growth stages on rice growth and yield under upland conditions. 2. Phenology, biomass production and yield. Field Crop Res. 48, 47–55. Boote, K. J., and Sinclair, T. R. (2006). Crop physiology: Significant discoveries and our changing perspective on research. Crop Sci. 46, 2270–2277. Bouman, B. A. M., Kropff, M. J., Tuong, T. P., Wopereis, M. C. S., ten Berge, H. F. M., and van Laar, H. H. (2003). http://www.knowledgebank.irri.org/oryza2000 ORYZA2000: Modeling lowland rice. IRRI. Bouman, B. A. M., Humphreys, E., Tuong, T. P., and Barker, R. (2006). Rice and water. Adv. Agron. 92, 187–237. Breseghello, F., and Sorrels, M. E. (2006). Association mapping of kernel size and milling quality in wheat (Triticum aestivum L.) cultivars. Genetics 172, 1165–1177. Bruskiewich, R. M., Cosico, A. B., Eusebio, W., Portugal, A. M., Ramos, L. M., Reyes, M. T., Sallan, M. A. B., Ulat, V. J. M., Wang, X., McNally, K. L., Sackville Hamilton, R., and McLaren, C. G. (2003). Linking genotype to phenotype: The International Rice Information System (IRIS). Bioinformatics 19, i63–i65. Bruskiewich, R., Davenport, G., Hazekamp, T., Metz, T., Ruiz, M., Simon, R., Takeya, M., Lee, J., Senger, M., McLaren, C. G., and Van Hintum, T. (2006). The Generation Challenge Programme (GCP): Standards for crop data. OMICS 10, 215–219. Capell, T., Bassie, L., and Christou, P. (2004). Modulation of the polyamine biosynthetic pathway in transgenic rice confers tolerance to drought stress. Proc. Natl. Acad. Sci. USA 101, 9909–9914. Centritto, M., Lauteri, M., Monteverdi, C., and Serraj, R. (2009). Response of leaf gas exchange to water deficits during reproductive stage in rice. J. Exp. Bot. 60, 2325–2339. Chaerle, L., Leinonen, I., Jones, H. G., and Van der Straeten, D. (2007). Monitoring and screening plant populations with combined thermal and chlorophyll imaging. J. Exp. Bot. 58, 773–784.

90

R. Serraj et al.

Chapman, S. C. (2008). Use of crop models to understand genotype by environment interactions for drought in real-world and simulated plant breeding trials. Euphytica 161, 195–208. Cheng, S. H., Willmann, M. R., Chen, H., and Sheen, J. (2002). Calcium signaling through protein kinases: the Arabidopsis calcium-dependent protein kinase gene family. Plant Physiol. 129, 469–485. Cockram, J., White, J., Leigh, F. J., Leav, V. J., Chiapparino, E., Laurie, D. A., Mackay, I. J., Powell, W., and O’sullivan, D. M. O. (2008). Association mapping of partitioning loci in barley. BMC Genet. 9, 16. Collins, N. C., Tardieu, F., and Tuberosa, R. (2008). Quantitative trait loci and crop performance under abiotic stress: Where do we stand? Plant Physiol. 47, 469–486. Comai, L., Young, K., Till, B. J., Reynolds, S. H., Greene, E. A., Codomo, C. A., Enns, L. C., Johnson, J. E., Burtner, C., Odden, A. R., and Henikoff, S. (2004). Efficient discovery of DNA polymorphisms in natural populations by EcoTILLING. Plant J. 37, 778–786. Condon, A. G., Richards, R. A., Rebetzke, G. J., and Farquhar, G. D. (2004). Breeding for high water-use efficiency. J. Exp. Bot. 55, 2447–2460. Cooper, M., Rajatasereekul, S., Immark, S., Fukai, S., and Basnayake, J. (1999a). Rainfed lowland rice breeding strategies for Northeast Thailand. I. Genotypic variation and genotype environment interactions for grain yield. Field Crop Res. 64, 131–151. Cooper, M., Fukai, S., and Wade, L. J. (1999b). How can breeding contribute to more productive and sustainable rainfed lowland rice systems? Field Crop Res. 64, 199–209. Courtois, B., and Lafitte, H. R. (1999). Improving rice for drought-prone upland environments. Proceedings of the Workshop on Genetic Improvement of Rice for Water-Limited Environments, Los Ban˜os, Philippines, 1–3 December 1998. Courtois, B., McLaren, G., Sinha, P. K., Prasad, K., Yadav, R., and Shen, L. (2000). Mapping QTL associated with drought avoidance in upland rice. Mol. Breed. 6, 55–66. Courtois, B., Shen, L., Petalcorin, W., Carandang, S., Mauleon, R., and Li, Z. (2003). Locating QTLs controlling constitutive root traits in the rice population IAC 165 Co39. Euphytica 134, 335–345. Crouch, J. H., and Serraj, R. (2002). DNA marker technology as a tool for genetic enhancement of drought tolerance at ICRISAT. In ‘‘International Workshop on Field Screening for Drought Tolerance in Rice’’ (N. P. Saxena, Ed.). International Crops Research Institute for the Semi-Arid Tropics, Patancheru, India. Cruz, R. T., and O’Toole, J. C. (1984). Dryland rice response to an irrigation gradient at flowering stage. Agron. J. 76, 178–183. Ding, S., Pandey, S., Chen, C., and Bhandari, H. (2005). Drought and rice farmers’ coping mechanisms in the rice production systems of southern China. In ‘‘A final research report submitted to the Rockefeller Foundation’’. Department of Agricultural Economics, Huazhong Agricultural University, Hubei, China. Dingkuhn, M., Farquhar, G. D., De Datta, S. K., and O’Toole, J. C. (1991). Discrimination of 13C among upland rices having different water use efficiencies. Aust. J. Agric. Res. 42, 1123–1131. Dubouzet, J. G., Sakuma, Y., Ito, Y., Kasuga, M., Dubouzet, E. G., Miura, S., Seki, M., Shinozaki, K., and Yamaguchi-Shinozaki, K. (2003). OsDREB genes in rice, Oryza sativa L., encode transcription activators that function in drought-, high salt- and coldresponsive gene expression. Plant J. 33, 751–763. Edmeades, G. O., Bolanos, J., Lafitte, H. R., Rajaram, S., Pfeiffer, W., and Fischer, R. A. (1989). Traditional approaches to breeding for drought resistance in cereals. In ‘‘Drought Resistance in Cereals’’ (F. W. G. Baker, Ed.), pp. 27–52. CAB International, Wallingford, Oxon, UK.

Improvement of Drought Resistance in Rice

91

Edmeades, G. O., Bolanos, J., Chapman, S. C., Lafitte, H. R., and Banziger, M. (1999). Selection improves drought tolerance in tropical maize populations. I. Gains in biomass, grain yield, and harvest index. Crop Sci. 39, 1306–1315. Ekanayake, I. J., Steponkus, P. L., and De Datta, S. K. (1990). Sensitivity of pollination to water deficits at anthesis in upland rice. Crop Sci. 30, 310–315. Ersoz, E. S., Yu, J., and Buckler, E. S. (2007). Applications of linkage disequilibrium and association mapping in crop plants. In ‘‘Genomics-Assisted Crop Improvements’’ (R. J. Varshney and R. Tuberosa, Eds.), Vol. 1, pp. 97–119. Springer, Berlin. Fujita, Y., Fujita, M., Satoh, R., Maruyama, K., Parvez, M. M., Seki, M., Hiratsu, K., Ohme-Takagi, M., Shinozaki, K., and Yamaguchi-Shinozaki, K. (2005). AREB1 is a transcription activator of novel ABRE-dependent ABA signaling that enhances drought stress tolerance in Arabidopsis. Plant Cell 17, 3470–3488. Fukai, S., and Cooper, M. (1995). Development of drought-resistant cultivars using physiomorphological traits in rice. Field Crop Res. 40, 67–86. Fukai, S., Basnayake, J., and Cooper, M. (2001). Modeling water availability, crop growth, and yield of rainfed lowland rice genotypes in northeast Thailand. In ‘‘Proceedings of the International Workshop on Characterizing and Understanding Rainfed Environments’’ (T. P. Tuong, S. P. Kam, L. Wade, S. Pandey, B. A. M. Bouman, and B. Hardy, Eds.), pp. 111–130. International Rice Research Institute, Los Ban˜os, Philippines, Bali, Indonesia 5–9 December 1999. Fukai, S., Basnayake, J., and Makara, O. (2008). Drought resistance characters and variety development for rainfed lowland rice in Southeast Asia. In ‘‘Drought Frontiers in Rice— Crop Improvement for Increased Rainfed Production’’ (R. Serraj, J. Bennett, and B. Hardy, Eds.), World Scientific Publishing, Singapore. Furihata, T., Maruyama, K., Fujita, Y., Umezawa, T., Yoshida, R., Shinozaki, K., and Yamaguchi-Shinozaki, K. (2006). Abscisic acid dependent multisite phosphorylation regulates the activity of a transcription activator AREB1. Proc. Natl. Acad. Sci. USA 103, 1988–1993. Gao, H., Williamson, S., and Bustamante, C. D. (2007). A Markov chain Monte Carlo approach for joint inference of population structure and inbreeding rates from multilocus genotype data. Genetics 176, 1635–1651. Garg, A. K., Kim, J. K., Owens, T. G., Ranwala, A. P., Choi, Y. D., Kochian, L. V., and Wu, R. J. (2002). Trehalose accumulation in rice plants confers high tolerance levels to different abiotic stresses. Proc. Natl. Acad. Sci. USA 99, 15898–15903. Garris, A., McCouch, S., and Kresovich, S. (2003). Population structure and its effect on haplotype diversity and linkage disequilibrium surrounding the Xa5 locus of rice (Oryza sativa L.). Genetics 165, 759–769. Garrity, D. P., and O’Toole, J. C. (1995). Selection for reproductive stage drought avoidance in rice, using infrared thermometry. Agron. J. 87, 773–779. Garrity, D. P., Oldeman, L. R., and Morris, R. A. (1986). Rainfed lowland rice ecosystems: Characterization and distribution. In ‘‘A Progress in Rainfed Lowland Rice’’, pp. 3–23. International Rice Research Institute, Los Ban˜os, Philippines. Glaszmann, J. C. (1987). Isozymes and classification of Asian rice varieties. Theor. Appl. Genet. 74, 21–30. Granier, C., Aguirrezabal, L., Chenu, K., Cookson, S. J., Dauzat, M., Hamard, P., Thioux, J. J., Rolland, G., Bouchier-Combaud, S., Lebaudy, A., Muller, B., Simonneau, T., et al. (2006). Phenopsis, an automated platform for reproducible phenotyping of plant responses to soil water deficit in Arabidopsis thaliana permitted the identification of an accession with low sensitivity to soil water deficit. New Phytol. 169, 623–635. Hammer, G. L., and Jordan, D. (2008). An integrated systems approach to crop improvement. In ‘‘Drought Frontiers in Rice—Crop Improvement for Increased Rainfed

92

R. Serraj et al.

Production’’ (R. Serraj, J. Bennett, and B. Hardy, Eds.), World Scientific Publishing, Singapore. Heinemann, A. B., Dingkuhn, M., Luquet, D., Combres, J. C., and Chapman, S. (2008). Characterization of drought stress environments for upland rice and maize in central Brazil. Euphytica 162(3), 395–410. Herve´, P., and Serraj, R. (2009). GM technology and drought: A simple solution for a complex trait?. Afr. J. Biotechnol. 8(9), 1740–1749. Horie, T., Matsuura, S., Takai, T., Kuwasaki, K., Ohsumi, A., and Shiraiwa, T. (2006). Genotypic difference in canopy diffusive conductance measured by a new remotesensing method and its association with the difference in rice yield potential. Plant Cell Environ. 29, 653–660. Hu, H., Dai, M., Yao, J., Xiao, B., Li, X., Zhang, Q., and Xiong, L. (2006). Overexpressing a NAM, ATAF, and CUC (NAC) transcription factor enhances drought resistance and salt tolerance in rice. Proc. Natl. Acad. Sci. USA 103, 12987–12992. Impa, S. M., Nadaradjan, S., Boominathan, P., Shashidhar, G., Bindumadhava, H., and Sheshshayee, M. S. (2005). Carbon isotope discrimination accurately reflects variability in WUE measured at a whole plant level in rice. Crop Sci. 45, 2517–2522. IRGSP (International Rice Genome Sequencing Project). (2005). The map-based sequence of the rice genome. Nature 436, 793–800. Islam, T. M. (2008). Modeling of drought for Aman rice in the Northwest region of Bangladesh, p. 194. Bangladesh Agricultural University, Mymensingh, Ph.D. Thesis. Ito, Y., Katsura, K., Maruyama, K., Taji, T., Kobayashi, M., Seki, M., Shinozaki, K., and Yamaguchi-Shinozaki, K. (2006). Functional analysis of rice DREB1/CBF-type transcription factors involved in cold-responsive gene expression in transgenic rice. Plant Cell Physiol. 47, 141–153. Iuchi, S., Kobayashi, M., Taji, T., Naramoto, M., Seki, M., Kato, T., Tabata, S., Kakubari, Y., Yamaguchi-Shinozaki, K., and Shinozaki, K. (2001). Regulation of drought tolerance by gene manipulation of 9-cis-epoxycarotenoid dioxygenase, a key enzyme in abscisic acid biosynthesis in Arabidopsis. Plant J. 27, 325–333. Jang, I. C., Oh, S. J., Seo, J. S., Choi, W. B., Song, S. I., Kim, C. H., Kim, Y. S., Seo, H. S., Do Choi, Y., Nahm, B. H., and Kim, J. K. (2003). Expression of a bifunctional fusion of the Escherichia coli genes for trehalose-6-phosphate synthase and trehalose-6 phosphate phosphatase in transgenic rice plants increases trehalose accumulation and abiotic stress tolerance without stunting growth. Plant Physiol. 131, 516–524. Ji, X. M., Raveendran, M., Oane, R., Ismail, A., Lafitte, R., Bruskiewich, R., Cheng, S. H., and Bennett, J. (2005). Tissue-specific expression and drought responsiveness of cell-wall invertase genes of rice at flowering. Plant Mol. Biol. 59(6), 945–964. Jones, H. G. (1992). Plants and Microclimate 2. Cambridge University Press, Cambridge, UK. Jongdee, B., Fukai, S., and Cooper, M. (2002). Leaf water potential and osmotic adjustment as physiological traits to improve drought tolerance in rice. Field Crop Res. 76, 153–163. Jongdee, B., Pantuwan, G., Fukai, S., and Fischer, K. (2006). Improving drought tolerance in rainfed lowland rice: An example from Thailand. Agric. Water Manage. 80, 225–240. Jung, K. H., An, G., and Ronald, P. C. (2008). Towards a better bowl of rice: Assigning function to tens of thousands of rice genes. Nat. Rev. Genet. 9, 91–101. Kamoshita, A., Babu, R. C., Boopathi, N. M., and Fukai, S. (2008). Phenotypic and genotypic analysis of drought-resistance traits for development of rice cultivars adapted to rainfed environments. Field Crop Res. 109, 1–23. Kathiresan, A., Lafitte, H. R., Chen, J., Mansueto, L., Bruskiewich, R., and Bennett, J. (2006). Gene expression microarrays and their application in drought stress research. Field Crop Res. 97, 101–110.

Improvement of Drought Resistance in Rice

93

Khowaja, F. S., and Price, A. H. (2008). QTL mapping rolling, stomatal conductance and dimension traits of excised leaves in the Bala Azucena recombinant inbred population of rice. Field Crop Res. 106, 248–257. Kumar, R., Venuprasad, R., and Atlin, G. N. (2007). Genetic analysis of rainfed lowland rice drought tolerance under naturally-occurring stress in eastern India: Heritability and QTL effects. Field Crop Res. 103, 42–52. Kumar, A., Bernier, J., Verulkar, S., Lafitte, H. R., and Atlin, G. N. (2008). Breeding for drought tolerance: Direct selection for yield, response to selection and use of droughttolerant donors in upland and lowland-adapted populations. Field Crop Res. 107, 221–231. Lafitte, H. R. (2002). Relationship between leaf relative water content during reproductive stage water deficit and grain formation in rice. Field Crop Res. 76, 165–174. Lafitte, H. R. (2003). Managing water for controlled drought in breeding plots. In ‘‘Breeding Rice for Drought-Prone Environments’’ (K. S. Fischer, R. Lafitte, S. Fukai, G. Atlin, and B. Hardy, Eds.), pp. 23–26. International Rice Research Institute, Los Ban˜os, Philippines. Lafitte, H. R., and Courtois, B. (2002). Interpreting cultivar environment interactions for yield in upland rice: Assigning value to drought-adaptive traits. Crop Sci. 42, 1409–1420. Lafitte, H. R., Price, A. H., and Courtois, B. (2004). Yield response to water deficit in an upland rice mapping population: Associations among traits and genetic markers. Theor. Appl. Genet. 109, 1237–1246. Lafitte, H. R., Li, Z. K., Vijayakumar, C. H. M., Gao, Y. M., Shi, Y., Xu, J. L., Fu, B. Y., Ali, A. J., Domingo, J., Maghirang, R., Torres, R., and Mackill, D. (2006). Improvement of rice drought tolerance through backcross breeding: Evaluation of donors and selection in drought nurseries. Field Crop Res. 97, 77–86. Lanceras, J. C., Pantuwan, G. P., Jongdee, B., and Toojinda, T. (2004). Quantitative trait loci associated with drought tolerance at reproductive stage in rice. Plant Physiol. 135, 384–399. Leinonen, I., and Jones, H. G. (2004). Combining thermal and visible imagery for estimating canopy temperature and identifying plant stress. J. Exp. Bot. 55, 1423–1431. Leinonen, I., Grant, O. M., Tagliavia, C. P. P., Chaves, M. M., and Jones, H. G. (2006). Estimating stomatal conductance with thermal imagery. Plant Cell Environ. 29, 1508–1518. Leung, H. (2008). Stressed genomics—Bringing relief to rice fields. Curr. Opin. Plant Biol. 11, 201–208. Levitt, J. (1980). Water stress, dehydration and drought injury. In ‘‘Responses of Plants to Environmental Stresses’’ (T. T. Kozlowski, Ed.), Vol. II, Academic Press, London. Lian, H. L., Yu, X., Ye, Q., Ding, X. S., Kitagawa, Y., Kwak, S. S., Su, W. A., and Tang, Z. C. (2004). The role of aquaporin RWC3 in drought avoidance in rice. Plant Cell Physiol. 45, 481–489. Lilley, J. M., and Fukai, S. (1994a). Effect of timing and severity of water-deficit on 4 diverse rice cultivars. 1. Rooting pattern and soil-water extraction. Field Crop Res. 37, 205–213. Lilley, J. M., and Fukai, S. (1994b). Effect of timing and severity of water-deficit on 4 diverse rice cultivars. 2. Physiological responses to soil-water deficit. Field Crop Res. 37, 215–223. Liu, J. X., Liao, D. Q., Oane, R., Estenor, L., Yang, X. E., Li, Z. C., and Bennett, J. (2005). Genetic variation in the sensitivity of anther dehiscence to drought stress in rice. Field Crop Res. 97, 87–100. Liu, H., Mei, H., Yu, X., Zou, G., Liu, G., and Luo, L. (2006). Towards improving the drought tolerance of rice in China. Plant Genet. Res. 4, 47–53. Ludlow, M. M., and Muchow, R. C. (1990). A critical-evaluation of traits for improving crop yields in water-limited environments. Adv. Agron. 43, 107–153.

94

R. Serraj et al.

Mackill, D. J., Coffman, W. R., and Garrity, D. P. (1996). Rainfed Lowland Rice Improvement International Rice Research Institute, Los Ban˜os, Philippines, p. 242. McCouch, S., Teytelman, L., Xu, Y., Lobos, K., Clare, K., Walton, M., Fu, B., Maghirang, R., Li, Z., Xing, Y., Zhang, Q., Kono, I., et al. (2003). Development and mapping of 2240 new SSR markers for rice (Oryza sativa L.). DNA Res. 9, 199–207. McLaren, C. G., Bruskiewich, R. M., Portugal, A. M., and Cosico, A. B. (2005). The International Rice Information System: A platform for meta-analysis of rice crop data. Plant Physiol. 139, 637–642. McNally, K. L., Bruskiewich, R., Mackill, D., Buell, R. C., Leach, J. E., and Leung, H. (2006). Sequencing multiple and diverse rice varieties: Connecting whole-genome variation with phenotypes. Plant Physiol. 141, 26–31. McNally, K. L., Childs, K. L., Bohnert, R., Davidson, R. M., Zhao, K., Ulat, V. J., Zeller, G., Clark, R. M., Hoen, D. R., Bureau, T. E., Stokowski, R., Ballinger, D. G., et al. (2009). Genome-wide SNP variation reveals relationships among landraces and modern varieties of rice. Proc. Natl. Sci. USA (in press). Merlot, S., Mustilli, A. C., Genty, B., North, H., Lefebvre, V., Sotta, B., Vavasseur, A., and Giraudat, J. (2002). Use of infrared thermal imaging to isolate Arabidopsis mutants defective in stomatal regulation. Plant J. 30, 601–609. Michelmoore, R. W., Paran, I., and Kesseli, R. V. (1991). Identification of markers linked to disease resistance genes by bulked segregant analysis: A rapid method to detect markers in specific genomic regions by using segregating populations. Proc. Natl. Acad. Sci. USA 88, 9828–9832. Miflin, B. (2000). Crop improvement in the 21st century. J. Exp. Bot. 51, 1–8. Missaoui, A. M., Ha, B. K., Philips, D. V., and Boerma, H. R. (2007). Single nucleotide polymorphism detection of the Rcs3 gene for resistance to frogeye leaf spot in soybean. Crop Sci. 45, 1681–1690. Monneveux, P., and Ribaut, J. M. (2006). Secondary traits for drought tolerance improvements in cereals. In ‘‘Drought Adaptation in Cereals’’ ( J. M. Ribaut, Ed.), pp. 97–144. Haworth Press, New York. Monneveux, P., Rekika, D., Acevedo, E., and Merah, O. (2006). Effect of drought on leaf gas exchange, carbon isotope discrimination, transpiration efficiency and productivity in field grown durum wheat genotypes. Plant Sci. 170, 867–872. Monteith, J. L. (1973). Principles of Environmental Physics Edward Arnold, London. Nakamura, T., Yokota, S., Muramoto, Y., Tsutsui, K., Oguri, Y., Fukui, K., and Takabe, T. (1997). Expression of a betaine aldehyde dehydrogenase gene in rice, a glycinebetaine nonaccumulator, and possible localization of its protein in peroxisomes. Plant J. 11, 1115–1120. Navabi, A., Mather, D. E., Bernier, J., Spaner, D. M., and Atlin, G. N. (2008). QTL detection with bidirectional and unidirectional selective genotyping: Marker-based and trait-based analyses. Theor. Appl. Genet. 10.1007/s00122-008-0904-2. Nguyen, H. T., Babu, R. C., and Blum, A. (1997). Breeding for drought resistance in rice: Physiological and molecular genetics considerations. Crop Sci. 37, 1426–1434. Nigam, S. N., Chandra, S., Sridevi, K. R., Bhukta, M., Reddy, A. G. S., Rachaputi, N. R., Wright, G. C., Reddy, P. V., Deshmukh, M. P., Mathur, R. K., Basu, M. S., Vasundhara, S., et al. (2005). Efficiency of physiological trait-based and empirical selection approaches for drought tolerance in groundnut. Ann. Appl. Biol. 146, 433–439. Oh, S. J., Song, S. I., Kim, Y. S., Jang, H. J., Kim, S. Y., Kim, M., Kim, Y. K., Nahm, B. H., and Kim, J. K. (2005). Arabidopsis CBF3/DREB1A and ABF3 in transgenic rice increased tolerance to abiotic stress without stunting growth. Plant Physiol. 138, 341–351. Ohyanagi, H., Tanaka, T., Sakai, H., Shigemoto, Y., Yamaguchi, K., Habara, T., Fujii, Y., Antonio, B. A., Nagamura, Y., Imanishi, T., Ikeo, K., Itoh, T., et al. (2006). The Rice Annotation Project Database (RAP-DB): Hub for Oryza sativa ssp. japonica genome information. Nucl. Acids Res. 34, D741–D744.

Improvement of Drought Resistance in Rice

95

Olsen, K. M., Caicedo, A. L., Polato, N., McClung, A., McCouch, S., and Purugganan, M. D. (2006). Selection under domestication: Evidence for a sweep in the rice waxy genomic region. Genetics 173, 975–983. O’Toole, J. C. (2004). Rice and water: The final frontier. First International Conference on Rice for the Future, Bangkok, Thailand, 31 August–2 September 2004. O’Toole, J. C., and Moya, T. B. (1978). Genotypic variation in maintenance of leaf water potential in rice. Crop Sci. 18, 873–876. O’Toole, J. C., and Namuco, O. S. (1983). Role of panicle exsertion in water stress induced sterility. Crop Sci. 23, 1093–1097. Ouk, M., Basnayake, J., Tsubo, M., Fukai, S., Fischer, K. S., Cooper, M., and Nesbitt, H. (2006). Use of drought response index for identification of drought tolerant genotypes in rainfed lowland rice. Field Crop Res. 99, 48–58. Pandey, S., and Bhandari, H. (2008). Drought: Economic costs and research implications. In ‘‘Drought Frontiers in Rice—Crop Improvement for Increased Rainfed Production’’ (R. Serraj, J. Bennett, and B. Hardy, Eds.), World Scientific Publishing, Singapore. Pandey, S., Bhandari, H., and Hardy, B., (Eds.) (2007). In ‘‘Economic Costs of Drought and Rice Farmers’ Coping Mechanisms: A Cross-Country Comparative Analysis from Asia’’, p. 203. International Rice Research Institute (IRRI), Los Ban˜os, Philippines. Pantuwan, G., Fukai, S., Cooper, M., Rajatasereekul, S., and O’Toole, J. C. (2002). Yield response of rice (Oryza sativa L.) genotypes to different types of drought under rainfed lowlands. Part 3. Plant factors contributing to drought resistance. Field Crop Res. 73, 181–200. Peng, S., Laza, R. C., Khush, G. S., Sanico, A. L., Visperas, R. M., and Garcia, F. V. (1998). Transpiration efficiencies of indica and improved tropical japonica rice grown under irrigated conditions. Euphytica 103, 103–108. Price, A. H., Young, E. M., and Tomos, A. D. (1997). Quantitative trait loci associated with stomatal conductance, leaf rolling and heading date mapped in upland rice (Oryza sativa). New Phytol. 137, 83–91. Price, A. H., Cairns, J. E., Horton, P., Jones, H. G., and Griffiths, H. (2002). Linking drought-resistance mechanisms in upland rice using a QTL approach: Progress and new opportunities to integrate stomatal and mesophyll responses. J. Exp. Bot. 53, 989–1004. Pritchard, J. K., Stephens, M., Rosenberg, N. A., and Donnelly, P. (2000). Association mapping in structured populations. Am. J. Hum. Genet. 37, 170–181. Quarrie, S., Lazic-Jancic, V., Kovacevic, D., Steed, A., and Pekic, S. (1999). Bulk segregant analysis with molecular markers and its use for improving drought resistance in maize. J. Exp. Bot. 50, 1299–1306. Rabbani, M. A., Maruyama, K., Abe, H., Khan, M. A., Katsura, K., Ito, Y., Yoshiwara, K., Seki, M., Shinozaki, K., and Yamaguchi-Shinozaki, K. (2003). Monitoring expression profiles of rice genes under cold, drought, and high-salinity stresses and abscisic acid application using cDNA microarray and RNA gel-blot analysis. Plant Physiol. 133, 1755–1767. Raghavan, C., Naredo, M. E. B., Wang, H. H., Atienza, G., Liu, B., Qiu, F. L., McNally, K. L., and Leung, H. (2007). Rapid method for detecting SNPs on agarose gels and its application in candidate gene mapping. Mol. Breed. 19, 87–101. Ramanjulu, S., and Bartels, D. (2002). Drought- and desiccation-induced modulation of gene expression in plants. Plant Cell Environ. 25, 141–151. Ray, J. D., and Sinclair, T. R. (1997). Stomatal closure of maize hybrids in response to drying soil. Crop Sci. 37, 803–807. Rebetzke, G. J., Condon, A. G., Richards, R. A., and Farquhar, G. D. (2002). Selection for reduced carbon isotope discrimination increases aerial biomass and grain yield of rainfed bread wheat. Crop Sci. 42, 739–745. Reshmidevi, T. V., Jana, R., and Eldho, T. I. (2008). Geospatial estimation of soil moisture in rain-fed paddy fields using SCS-CN-based model. Agric. Water Manage. 95, 447–457.

96

R. Serraj et al.

Richards, R. A. (2006). Physiological traits used in the breeding of new cultivars for waterscarce environments. Agric. Water Manage. 80, 197–211. Rosielle, A. A., and Hamblin, J. (1981). Theoretical aspects of selection for yield in stress and non-stress environments. Crop Sci. 21, 943–946. Roy, M., and Wu, R. (2001). Arginine decarboxylase transgene expression and analysis of environmental stress tolerance in transgenic rice. Plant Sci. 160, 869–875. Sadras, V. O., and Milroy, S. P. (1996). Soil-water thresholds for the responses of leaf expansion and gas exchange. Field Crop Res. 47, 253–266. Saijo, Y., Hata, S., Kyozuka, J., Shimamoto, K., and Izui, K. (2000). Overexpression of a single Ca2+-dependent protein kinase confers both cold and salt/drought tolerance on rice plants. Plant J. 23, 319–327. Saini, H. S., and Westgate, M. E. (2000). Reproductive development in grain crops during drought. Adv. Agron. 68, 59–96. Salekdeh, G. H., Siopongco, J., Wade, L. J., Ghareyazie, B., and Bennett, J. (2002). Proteomic analysis of rice leaves during drought stress and recovery. Proteomics 2, 1131–1145. Sarkarung, S., and Pantuwan, G. (1999). Improving rice for drought-prone rainfed lowland environments. In ‘‘Genetic Improvement of Rice for Water-Limited Environments’’ (O. Ito, J. C. O’Toole, and B. Hardy, Eds.), pp. 57–70. International Rice Research Institute, Los Ban˜os, Philippines. Septiningsih, E. M., Pamplona, A. M., Sanchez, D. L., Neeraja, C. N., Vergara, G. V., Heuer, S., Ismail, A. M., and Mackill, D. J. (2009). Development of submergence tolerant rice cultivars: The Sub1 locus and beyond. Ann. Bot. 103, 151–160. Serraj, R., and Atlin, G. (2008). Drought-resistant rice for increased rainfed production and poverty alleviation. In ‘‘Drought Frontiers in Rice—Crop Improvement for Increased Rainfed Production’’ (R. Serraj, J. Bennett, and B. Hardy, Eds.), World Scientific Publishing, Singapore. Serraj, R., and Cairns, J. E. (2006). Diagnosing drought. Rice Today 5(3), 32–33. Serraj, R., and Sinclair, T. R. (1997). Variation among soybean cultivars in dinitrogen fixation response to drought. Agron. J. 89, 963–969. Serraj, R., and Sinclair, T. R. (2002). Osmolyte accumulation: Can it really help increase crop yield under drought conditions?. Plant Cell Environ. 25, 335–341. Serraj, R., Allen, H. L., and Sinclair, T. R. (1999). Soybean leaf growth and gas exchange response to drought under carbon dioxide enrichment. Glob. Change Biol. 5, 283–292. Serraj, R., Krishnamurthy, L., Kashiwagi, J. W., Kumar, J., Chandra, S., and Crouch, J. H. (2004). Variation in root traits of chickpea (Cicer arietinum L.) grown under terminal drought. Field Crop Res. 88, 115–127. Serraj, R., Hash, C. T., Rizvi, S. M. H., Sharma, A., Yadav, R. S., and Bidinger, F. R. (2005). Recent advances in marker-assisted selection for drought tolerance in pearl millet. Plant Prod. Sci. 8(3), 332–335. Serraj, R., Liu, D. C., Hong, H., Sellamuthu, R., Impa, S., Cairns, J., Dimayuga, G., and Torres, R. (2008a). Novel approaches for integration of physiology, genomics and breeding for drought resistance improvement in rice. Proceedings of the 5th International Crop Science Congress, pp. 14–18. Jeju, Korea, April. Serraj, R., Dimayuga, G., Gowda, V., Guan, Y., He, H., Impa, S., Liu, D. C., Mabesa, R. C., Sellamuthu, R., and Torres, R. (2008b). Drought-resistant rice: Physiological framework for an integrated research strategy. In ‘‘Drought Frontiers in Rice— Crop Improvement for Increased Rainfed Production’’ (R. Serraj, J. Bennett, and B. Hardy, Eds.), World Scientific Publishing, Singapore. Shashidhar, H. E., Vinod, M. S., Sudhir, G. V., Sharma, N., and Krishnamurthy, K. (2005). Markers linked to grain yield using bulked segregant analysis approach in rice (Oryza sativa L.). Rice Genet. Newsl. 22, 69–71.

Improvement of Drought Resistance in Rice

97

Shen, L., Courtois, B., McNally, K. L., Robin, S., and Li, Z. (2001). Evaluation of nearisogenic lines of rice introgressed with QTLs for root depth through marker-aided selection. Theor. Appl. Genet. 103, 75–83. Shen, X., Zhou, M., Lu, W., and Ohm, H. (2003). Detection of Fusarium head blight resistance QTL in a wheat population using bulked segregant analysis. Theor. Appl. Genet. 106, 1041–1047. Shinozaki, K., and Yamaguchi-Shinozaki, K. (2007). Gene network involved in drought stress response and tolerance. J. Exp. Bot. 58, 221–227. Sinclair, T. R. (2005). Theoretical analysis of soil and plant traits influencing daily plant water flux on drying soils. Agron. J. 97, 1148–1152. Sinclair, T. R., and Ludlow, M. M. (1986). Influence of soil water supply on the plant water balance of four tropical grain legumes. Aust. J. Plant Physiol. 13, 329–341. Sinclair, T. R., and Muchow, R. C. (2001). System analysis of plant traits to increase grain yield on limited water supplies. Agron. J. 93, 263–270. Sinclair, T. R., Purcell, L. C., and Sneller, C. H. (2004). Crop transformation and the challenge to increase yield potential. Trends Plant Sci. 9, 70–75. Sinclair, T. R., Zwieniecki, M. A., and Holbrook, N. M. (2008). Low leaf hydraulic conductance associated with drought tolerance in soybean. Physiol. Plant. 132, 446–451. Siopongco, J. D. L. C., Sekiya, K., Yamauchi, A., Egdane, J., Ismail, A. M., and Wade, L. J. (2008). Stomatal responses in rainfed lowland rice to partial soil drying; evidence for root signals. Plant Prod. Sci. 11, 28–41. Smirnoff, N. (1993). The role of active oxygen in the response of plants to water deficit and desiccation. New Phytol. 125, 27–58. Steele, K. A., Price, A. H., Shashidhar, H. E., and Witcombe, J. R. (2006). Marker-assisted selection to introgress rice QTLs controlling root traits and aroma into an Indian upland rice variety. Theor. Appl. Genet. 112, 208–221. Steele, K. A., Virk, D. S., Kumar, R., Prasad, S. C., and Witcombe, J. R. (2007). Field evaluation of upland rice lines selected for QTLs controlling root traits. Field Crop Res. 101, 180–186. Su, J., Hirji, R., Zhang, L., He, C. K., Selvaraj, G., and Wu, R. (2006). Evaluation of the stress-inducible production of choline oxidase in transgenic rice as a strategy for producing the stress-protectant glycine betaine. J. Exp. Bot. 57, 1129–1135. Tardieu, F. (2003). Virtual plants: Modeling as a tool for the genomics of tolerance to water deficit. Trends Plant Sci. 8, 9–14. Tardieu, F., and Davies, W. J. (1993). Integration of hydraulic and chemical signaling in the control of stomatal conductance and water status of droughted plants. Plant Cell Environ. 16, 341–349. Thornsberry, J. M., Goodman, M. M., Doebley, J., Kresovich, S., Nielsen, D., and Buckler, E. S. (2001). Dwarf polymorphisms associate with variation in flowering time. Nat. Genet. 28, 286–289. Turner, N. C., Shahal, A., Berger, J. D., Chaturvedi, S. K., French, R. J., Ludwig, C., Mannur, D. M., Singh, S. J., and Yadava, H. S. (2007). Osmotic adjustment in chickpea (Cicer arietinum L.) results in no yield benefit under terminal drought. J. Exp. Bot. 58, 187–194. Umezawa, T., Fujita, M., Fujita, Y., Yamaguchi-Shinozaki, K., and Shinozaki, K. (2006). Engineering drought tolerance in plants: Discovering and tailoring genes to unlock the future. Curr. Opin. Biotechnol. 17, 113–122. Venuprasad, R., Lafitte, H. R., and Atlin, G. N. (2007a). Response to direct selection for grain yield under drought stress. Crop Sci. 47, 285–293. Venuprasad, R., Zenna, N., Choi, I. R., Amante, M., Virk, P. S., Kumar, A., and Atlin, G. N. (2007b). Identification of marker loci associated with tungro and drought tolerance in near-isogenic rice lines derived from IR64/Aday Sel. Int. Rice Res. Notes 32(1), 27–29.

98

R. Serraj et al.

Venuprasad, R., Sta Cruz, M. T., Amante, M., Magbanua, R., Kumar, A., and Atlin, G. N. (2008). Response to two cycles of divergent selection for grain yield under drought stress in four rice breeding populations. Field Crop Res. 107, 232–244. Venuprasad, R., Dalid, C., Del Valle, M., Bool, E., Zhao, D., Espiritu, M., Sta Cruz, M. T., Amante, M., Atlin, G., and Kumar, A. (2009). Identification and characterization of large-effect quantitative trait loci (QTL) for grain yield under lowland drought stress and aerobic conditions in rice using bulk-segregant analysis (BSA). Theor. Appl. Genet., (in press). Vigouroux, Y., McMullen, M., Hittinger, C. T., Houchins, K., Schulz, L., Kresovich, S., Matsuoka, Y., and Doebley, J. (2002). Identifying genes of agronomic importance in maize by screening microsatellites for evidence of selection during domestications. Proc. Natl. Acad. Sci. USA 99, 9650–9655. Wade, L. J., Fukai, S., Samson, B. K., Ali, A., and Mazid, M. A. (1999a). Rainfed lowland rice: Physical environment and cultivar requirements. Field Crop Res. 64, 3–12. Wade, L. J., McLaren, C. G., Quintana, L., Harnpichitvitaya, D., Rajatasereekul, S., Sarawgi, A. K., Kumar, A., Ahmed, H. U., Singh, S. A.K, Rodriguez, R., Siopongco, J., et al. (1999b). Genotype by environment interactions across diverse rainfed lowland rice environments. Field Crop Res. 64, 35–50. Wanchana, S., Thongjueu, S., Ulat, V. J., Anacleto, M., Mauleon, R., Conte, M., Rouard, M., Ruiz, M., Krishnamurthy, N., Sjolander, K., Van Hintum, T., and Bruskiewich, R. (2008). The Generation Challenge Programme comparative plant stress-responsive gene catalogue. Nucl. Acids Res. 36, D943–D946. Wang, F. Z., Wang, Q. B., Kwon, S. Y., Kwak, S. S., and Su, W. A. (2005). Enhanced drought tolerance of transgenic rice plants expressing a pea manganese superoxide dismutase. J. Plant Physiol. 162, 465–472. Wassmann, R., Jagadish, S. V. K., Sumfleth, K., Pathak, H., Howell, G., Ismail, A., Serraj, R., Redon˜a, E., Singh, R. K., and Heuer, S. (2009). Regional vulnerability of climate change impacts on Asian rice production and scope for adaptation. Adv. Agron. 102, 91–133. Wilson, L. M., Whitt, S. R., Ibanez, A. M., Rocheford, T. R., Goodman, M. M., and Buckler, E. S. (2004). Dissection of maize kernel composition and starch production by candidate gene association. Plant Cell 16, 2719–2733. Wu, R. L., and Lin, M. (2006). Functional mapping: How to map and study the genetic architecture of dynamic complex traits. Nat. Rev. Genet. 7(3), 229–237. Wu, H. L., Ni, Z. F., Yao, Y. Y., Guo, G. G., and Sun, Q. X. (2008). Cloning and expression profiles of 15 genes encoding WRKY transcription factor in wheat (Triticum aestivum L.). Prog. Nat. Sci. 18, 697–705. Xiao, B. Z., Huang, Y. M., Tang, N., and Xiong, L. Z. (2007). Overexpression of a LEA gene in rice improves drought resistance under field conditions. Theor. Appl. Genet. 115, 35–46. Xiao, B. Z., Chen, X., Xiang, C. B., Tang, N., Zhang, Q. F., and Xiong, L. Z. (2009). Evaluation of seven function-known candidate genes for their effects on improving drought resistance of transgenic rice under field conditions. Molec. Plant 2, 73–83. Xiong, L., and Yang, Y. (2003). Disease resistance and abiotic stress tolerance in rice are inversely modulated by an abscisic acid-inducible mitogen-activated protein kinase. Plant Cell 15, 745–759. Xiong, L., Ishitani, M., Lee, H., and Zhu, J. K. (2001). The Arabidopsis LOS5/ABA3 locus encodes a molybdenum cofactor sulfurase and modulates cold stress- and osmotic stressresponsive gene expression. Plant Cell 13, 2063–2083. Xu, Y., and Crouch, J. H. (2008). Marker-assisted selection in plant breeding: From publication to practice. Crop Sci. 48, 391–407.

Improvement of Drought Resistance in Rice

99

Xu, D., Duan, X., Wang, B., Hong, B., Ho, T., and Wu, R. (1996). Expression of a late embryogenesis abundant protein gene, HVA1, from barley confers tolerance to water deficit and salt stress in transgenic rice. Plant Physiol. 110, 249–257. Xu, K., Xia, X., Fukao, T., Canlas, P., Maghirang-Rodriguez, R., Heuer, S., Ismail, A. M., Bailey-Serres, J., Ronald, P. C., and Mackill, D. J. (2006). Sub1A is an ethylene response factor-like gene that confers submergence tolerance to rice. Nature 442, 705–708. Yamaguchi-Shinozaki, K., and Shinozaki, K. (2005). Organization of cis-acting regulatory elements in osmotic- and cold-stress-responsive promoters. Trends Plant Sci. 10, 88–94. Yin, X., Kropff, M., and Stam, P. (1999). The role of ecophysiological models in QTL analysis: The example of specific leaf area in barley. Heredity 82, 415–421. Yu, J. M., Holland, J. B., McMullen, M. D., and Buckler, E. S. (2008). Genetic design and statistical power of nested association mapping in maize. Genetics 178, 539–551. Yuan, Q., Ouyang, S., Liu, J., Suh, B., Cheung, F., Sultana, R., Lee, D., Quackenbush, J., and Buell, C. R. (2003). The TIGR rice genome annotation resource: Annotating the rice genome and creating resources for plant biologists. Nucl. Acids Res. 31, 229–233. Yue, B., Xue, W., Xiong, L., Yu, X., Luo, L., Cui, K., Jin, D., Xing, Y., and Zhang, Q. (2006). Genetic basis of drought resistance at reproductive stage in rice: Separation of drought tolerance from drought avoidance. Genetics 172, 1213–1228. Zhu, B. C., Su, J., Chan, M. C., Verma, D. P. S., Fan, Y. L., and Wu, R. (1998). Overexpression of a Delta(1)-pyrroline-5-carboxylate synthetase gene and analysis of tolerance to water- and salt-stress in transgenic rice. Plant Sci. 139, 41–48. Zhu, Q. H., Ramm, K., Shivakkumar, R., Dennis, E. S., and Upadhyaya, N. M. (2004). The anther indehiscence1 gene encoding a single MYB domain protein is involved in anther development in rice. Plant Physiol. 135, 1514–1525.

C H A P T E R

T H R E E

Problems, Challenges, and Strategic Options of Grain Security in China Huixiao Wang,* Minghua Zhang,† and Yan Cai* Contents 1. Introduction 1.1. Current worldwide grain crisis 1.2. Present grain production and supply in China 2. Connotation of Grain Security in China 2.1. Concept of food security and grain security 2.2. Characteristics of grain security in China 2.3. Assessment indicators for grain security in China 3. Development Stages of Grain Production in China 3.1. Stage 1: 1949–1957 3.2. Stage 2: 1958–1978 3.3. Stage 3: 1979–1984 3.4. Stage 4: 1985–2003 3.5. Stage 5: 2004–now 4. Achievements and Experiences of Grain Production in China 4.1. Achievements 4.2. Experiences 5. Problems and Challenges of Grain Security in China 5.1. Cultivated land loss and degradation 5.2. Water resources problems 5.3. Inferior climatic conditions and severe natural disasters 5.4. Vulnerable ecosystems 5.5. Global climatic change 5.6. Changes in population growth and the standard of living 5.7. Small-scale agricultural economy and relatively low profit 5.8. Old agricultural capital construction and weak risk resistance 6. Strategy Options and Countermeasures for Grain Security in China 6.1. To reinforce agricultural infrastructures 6.2. To increase the fiscal input in agriculture production

* {

103 103 105 106 106 107 108 109 110 111 111 112 114 114 114 116 120 120 123 124 125 126 127 127 128 129 129 130

Key Laboratory for Water and Sediment Sciences, Ministry of Education, College of Water Sciences, Beijing Normal University, Beijing, People’s Republic of China Department of Land, Air and Water Resources, University of California, California, USA

Advances in Agronomy, Volume 103 ISSN 0065-2113, DOI: 10.1016/S0065-2113(09)03003-X

#

2009 Elsevier Inc. All rights reserved.

101

102

Huixiao Wang et al.

6.3. To maximize the role of science and technology in grain production 6.4. To protect the quantity and quality of cultivated land 6.5. To protect water resources and improve water-use efficiency 6.6. To protect and enhance the farmers’ initiative for grain production 6.7. To response climatic change 6.8. To develop grass agriculture 6.9. To effectively control the growth rate of population 7. Case Studies 7.1. Science and technology engineering for grain high yield 7.2. 2008 activity for establishing high yield of grain in China 8. Concluding Remarks Acknowledgment References

131 133 135 136 137 137 138 139 139 141 143 144 144

Abstract Given China’s large population size, the attainment of grain security has become both a national and global issue. In the last decade, the concerted efforts of the Chinese government and its people to meet its grain demand have been highly successful. Despite the current global grain crisis, China’s grain market has remained relatively stable, allowing it to contribute to world grain security. The purpose of this chapter was to give an overview of grain security in China, with an analysis of the present national and global grain security, as well as a historical accounting of the most important factors that have led to success. The Chinese grain production process was analyzed in detail, discussing in-depth achievements and experiences, which can then be used for reference in other countries. Although currently stable, grain security for China faces many long-term challenges, such as loss of cultivated land from degradation and urbanization, limited water resources, frequent natural disasters, impacts of climate change, vulnerable ecosystems, increased demand from population growth and improved standard of living, a small-scale agricultural economy, and outdated aging agricultural infrastructure, among others. The chapter then in details suggested countermeasures that should be taken to guarantee grain security, including the improvement of agricultural infrastructure, increased fiscal input, greater use of science and technology, protection of cultivated land and water resources, support for the farmers’ livelihood, adaptation to climatic change, improvements in grassland agriculture, and controls on population growth, etc. The chapter concludes with two case studies serving to tie the chapter together with the idea that China’s present successes and proactive long-term plans are working, thus presenting a highly favorable outlook for the future.

Problems, Challenges, and Strategic Options of Grain Security in China

103

1. Introduction 1.1. Current worldwide grain crisis As a fundamental provider of food, fuel, and fiber, agriculture will always be a uniquely essential industry for human kind. It serves as a foundation to human life, social stability, and economic development. Grains are one of the world’s most important staple crops. However due to increases in global population, reductions in cultivated land, and regional differences in economic development, the world has experienced a severe grain crisis in recent years. The current grain crisis has taken on a global scale and has heightened worldwide awareness of the urgent need to address the concerned agricultural problems. Since 2005, the world has experienced a dramatic surge in the price of many staple food commodities. The price of maize increased by 80% between 2005 and 2007, and has since risen further. Many other commodity prices also rose sharply over this period: wheat by 70% and rice by about 25% (Ivanic and Martin, 2008). The FAO estimated that 37 countries currently face grain crises requiring external assistance, with over 20 countries implementing some policies to control grain exports (FAO, 2008a). The rapid increase in the international grain price not only affected the consumers of grain and the supply of agricultural products relying on grain, but also deeply affected the world economy, disrupting political stability and creating social unrest in many countries. In June 2008, skyrocketing global food prices spurred riots in Haiti, resulting in the removal of the prime minister. The FAO reported similar problems in several other countries around the world (FAO, 2008b). The global estimate of famine victims increased to 925 million at the beginning of 2008. Grain security also affected developed countries, with many seeing rates of inflation as high as 40–50%. Decisive resolutions are urgently needed. Without them, the combination of the newly implemented export controls, the effects of climatic change, and the speculations of the futures market will together push the world into an even more dangerous condition. Therefore, sustainable, practical solutions to this global crisis must be found, and a partnership needs to be established between countries with fiscal, management, and technological resources, and those with land, water, and human resources, to set up an internationally sustainable agricultural development plan. According to the FAO, global grain yield has been steadily increasing in recent years, reaching a record yield of 2.2 billion tons in 2008 (http:// www.news.cn 2008-07-14). However, this increase in supply has not been enough to prevent the grain crisis now occurring. There are many reasons attributed to the current crisis, such as increased demand from population growth, higher consumption by countries of grain-demanding meat animals,

104

Huixiao Wang et al.

and high demand for biofuel production. Production costs are also higher due to the effects of elevated crude oil prices on transportation costs. Finally, the rapid growth of technology during the previous century has slowed, while natural disasters attributed to climate change have increased. Compared to the ‘‘Green Revolution’’ of the last century which increased grain yield through advanced techniques such as the creation of specialized varieties through plant breeding and the improvement of irrigation and fertilization technology, there have been far fewer radical breakthroughs in the agricultural technology of this century, thus slowing the rate of yield increase. Population growth has endangered the food security of more and more developing countries. According to the statistics, the world’s population is over 6.5 billion, with almost 94% of the latest increment (78 million) residing in developing countries. As the living standard for these countries improves and the diet structure changes, grain consumption inevitably increases through the higher consumption of animal proteins. The ratio of meat, eggs, and milk in the diet of developing country populations has largely increased as per capita income rose. The grain used for forage reached a global level of 760 million tons in 2007. In recent years, climate change and related natural disasters have caused fluctuations in global grain yield. Abnormally torrid weather has been causing agricultural losses in America. Australia has been experiencing the worst super drought of a century, greatly shocking agriculture and grain production. In addition, the price of energy has risen sharply in recent years, resulting in high production costs, which translate into high consumer prices for agricultural products. Furthermore, the rising price of oil has caused a large demand for grain to be transformed into fuel. In 2007, 100 million tons of grain was used for fuel, which is about 5% of the total grain yield. It is this extra demand of grain for biofuel that many consider to have broken the balance of global grain supply and demand. For example, to reduce dependence on overseas oil, America used 28% of its total corn yield in 2008 for producing ethanol. Sixty percent of the oilseed rape harvest in the European Union was also used for producing fuel (http://www. chinaview.cn 2008–7–14). According to international institutions such as the International Monetary Fund (IMF), the strong demand for biofuel has pushed up the international grain price by 15–30%. Trade barriers of agricultural products further aggravated the worldwide grain crisis, as many nations adopted measures limiting grain exports or imposing high tariffs on imports to reduce pressure on national markets. These new regulations disrupted the ability of the global market to naturally adjust for supply and demand of grain in the international market, resulting in increased global grain price. Since many countries were not willing or able to conserve grain on a large scale due to high costs, world grain storage dramatically decreased in recent decades, with grain stock reduced from 30% in 2002 to 14.7% in 2008.

Problems, Challenges, and Strategic Options of Grain Security in China

105

1.2. Present grain production and supply in China China is both a big grain producer and consumer. Supply and demand of grain in China not only affects the food security of its population of 1.3 billion, but also has a strong bearing on international markets, strongly influencing world grain prices. In recent years, the annual average international grain trade is around 240 billion kg, which is about 50% of China’s total grain yield. One percent of China’s grain trade is imports, which amounts to 2% of the entire international grain trade amount. Therefore, any increase in grain imports by China correlates to a significant reduction of world grain supply, and thus higher global prices. Consequently, China’s grain security policy is geared toward promoting self-sufficiency, and participation in the international grain market is limited to the occasional regulation of surpluses and deficiencies. In 2007, China’s average grain yield was 5175 kg/hm2, 58% higher than the world average. Total grain yield reached 502 million tons, an increase of 71 million tons since 2003. Domestic grain consumption was around 510 million tons, almost balancing the supply with demand. Grain reserves reached 150–200 billion tons, at a stocking level two times the world average. The ratio of stocking to consumption was higher than the international accepted grain security line of 17–18%. In rural areas, grain storage was about 500 kg per person, equivalent to grain consumption for 2 years (People’s Daily, 2008). In 2008, grain yield reached 525 million tons, despite serious snow and frost in the south, a powerful earthquake in Sichuan Province, and frequent typhoons in the southeastern coastal regions. In the face of frequent natural disasters and an ever-changing complex international environment, China has had good grain harvests since 2004. This 5-year trend of good harvests is the first since China’s reform in 1978, when it opened to the outside world. It represents a major achievement for China, which is currently feeding 20% of the world population using only 9% of the world’s cultivated land. China produces surpluses in most major grain crops, such as wheat, rice, and corn. According to the international standards of grain, since 1997, China was only a net importer of grain once in 2004, which was a reflection of a grain shortage in 2003. In contrast, when grain yields began reaching 500 million tons, supply often exceeded demand, and China became a net grain exporter. Soybean, however, proved to be an exception: soybean yield in China is only one third of its consumption. As a result, China has become the biggest buyer in the international soybean market, importing 31 million tons in 2007, representing nearly half of all international sales. At present, the annual soybean requirement is more than 45 million tons in China, most of which is used to produce edible oil. To create effective grain policy, therefore, China must distinguish between soybean and other grains, so as to put resources where they are most needed.

106

Huixiao Wang et al.

In the long run, grain security in China is far from certain. First, grain consumption is rapidly rising as the population grows and living standards improve. Furthermore, increasing production costs, natural disasters, fuelrelated demands, market disruptions, and the slower growth of needed technology will all play important roles in the future balance between supply and demand. The idea of ‘‘having grain in hands’’ is necessary to ensure national grain security in China, as well as contributing to grain security at a global level.

2. Connotation of Grain Security in China 2.1. Concept of food security and grain security Food security, a concept that emerged in the 1970s relating to international food issues, was defined by the 1974 World Food Summit as: ‘‘availability at all times of adequate world food supplies of basic grains to sustain a steady expansion of food consumption and to offset fluctuations in production and prices’’ (United Nations, 1975). In 1983, the FAO refined the concept by specifying that food security must include secure access by vulnerable people to available supplies, and implying that attention should be balanced between the demand and supply side of the food security equation (FAO, 1983). In 1986, the highly influential World Bank report, ‘‘Poverty and Hunger,’’ elaborated on the concept, defining food security to mean ‘‘access of all people at all times to enough food for an active, healthy life,’’ thus introducing a quality idea into the definition (World Bank, 1986). By the mid-1990s, food security was recognized as a significant concern. As a result, the 1996 World Food Summit adopted a much more complex definition: ‘‘Food security, at the individual, household, national, regional and global levels is when all people, at all times, have physical and economic access to sufficient, safe and nutritious food to meet their dietary needs and food preferences for an active and healthy life’’ (FAO, 1996). The connotation of FAO food security concept includes five aspects: equal access, sufficient quantity, adequate quality, long-run stability, and ample purchasing power in times of surpluses and deficits (Long, 2007b). The food security concept was introduced in China in the mid-1970s and translated into ‘‘grain security,’’ since food consumption at that time was almost entirely grain based. Grains include cereal (wheat, rice, corn, etc.), legumes, and tuber crops, while the broader term, food, is defined as including grains, edible vegetable oil, meat, poultry, eggs, milk, and aquatic products. Grain security is a prerequisite for the ultimate goal of food security. However, only the supply of all grains, vegetables, meats, and other food products together will ultimately guarantee ‘‘food security.’’

Problems, Challenges, and Strategic Options of Grain Security in China

107

The FAO’s definition of food security prompted discussion among scholars in China, resulting in many different ideas. Zhu (1998) thought that the background and characteristics of grain security in China differed from those of other countries, and that therefore simply copying the concept of FAO for use in China was not advisable. He defined China’s grain security as needing to meet increasing demand as well as being able to endure diverse, unexpected events during national industrialization. He emphasized that the aim of studying improvements to grain security should be to boost the process of industrialization, and could be facilitated using the following four indicators: fluctuating coefficient of grain yield, grain reserve rate, grain self-sufficiency rate, and per capita grain availability. Lou (2003) defined grain security as a goal ‘‘to meet people’s direct and indirect grain consumption at reasonable price and with the ability to resist diversified grain risks including natural disasters, embargo blockage, unbalanced grain structure, and wrong strategic decisions.’’ The Department of Regulation, State Administration of Grain (2004) proposed that grain security is the responsibility of a country to meet its grain requirement, while resisting diverse, unexpected events related to national economic development and foreign trade. The country should be able to guarantee access to production materials, while regulating self-sufficiency and import levels, reserve capacities, and should be able to guarantee the consuming capacity via the total effective grain demand and consuming preference and structure matching with the economic development and income level. The country should also be able to ensure grain supply via an efficient supply chain system from producer to consumer and price policies. Since food security was defined by FAO, there has been much development and perfecting of the concept, increasing the richness of its connotations. The notion of grain security now includes quantity and quality standards, production, supply chain, and consumption issues, as well as short- and longrun planning. Grain security should be understood in a systematic, dynamic and developing way.

2.2. Characteristics of grain security in China The objective of grain security in China is to reach a balance between demand and supply of per capita grain rations, feed grain, industrial grain, and seed grain, with self-sufficiency of the people’s grain ration being the most important. Given China’s large population, self-sufficiency is a much more reliable option than looking to the international grain market to meet demand. Grain security for a nation is not just a production issue, but a complex distribution problem as well, including issues regarding production, supply chain, transportation, import and export, subsidies for the low-income groups, etc. Therefore, grain security should include the following aspects.

108

Huixiao Wang et al.

(1) Agricultural production: the ability to produce sufficient quantities of grain to meet demand is the fundamental issue of grain security. (2) Supply chain process: the supply chain from the producer to the consumer can strongly affect the price of grain. Grain price rises as the number of middle men needed to get a commodity from producer to consumer increases. Grain security is therefore dependent on an efficient supply chain that can eliminate all unnecessary middle links. (3) Transportation: transportation can be especially important when the regional distribution of grain production and demand does not overlap. For example, northeast China often has a rice surplus, while the southern areas need rice. Because of rising fuel prices, however, the cost of transportation between the north and south makes market transactions between the two areas unprofitable. Therefore, the northeast could not distribute its surplus, and its rice price dropped in the first quarter of 2008, despite skyrocketing prices worldwide. (4) Grain reserve: a plentiful grain reserve can ensure grain security when sudden natural disasters happen. (5) Regulation of grain surplus and deficit through import and export: China’s grain security cannot rely on grain import and export and the role of grain import and export in China is just to regulate grain surplus and deficit. Whether you can have the rights for determining grain price or not is the most important thing in the international import and export trade. For example, China was the first biggest country of importing soybean in 2007, 30.8 million tons, but the price of soybean in the international market was determined by several multinational corporations. Therefore, it is difficult for China to gain soybean security, which is a great lesson to learn. (6) Grain price: higher grain prices can limit the grain security of lower income portions of the population. In 2008, the total world grain yield was substantial enough to feed the world population; however, many people in Africa and other developing countries, and low-income people everywhere, were strongly impacted by the crisis because their incomes were too low to afford the food. In response, some countries provided subsidies or coupons to ensure a grain supply to affected portions of their populations (Zhang, 2008).

2.3. Assessment indicators for grain security in China The assessment indicators for grain security include grain yield level, stock level, import and export trade level, and the living level of the most impoverished people. The FAO’s criteria for grain security are as follows: a country should produce over 95% of its national grain, and the yield should be sufficient for an annual 400 kg per capita supply, with a grain reserve equal to 18% of annual consumption, never dropping below 14%. Expanding on the FAO criteria, many other scientific appraisal methods and indicators have been developed. The Department of Rural Social Economy Survey, National Bureau of Statistics of China (2005) carried

Problems, Challenges, and Strategic Options of Grain Security in China

109

out a systematic study on grain security evaluating indicators. From the four subsystems of grain supply, demand, market, and stock, the following 14 indicators concerned with grain security were established: (1) the per capita area sowed in grain must be over 0.08 hm2, setting a lower limit on land resources dedicated to grain production; (2) effective irrigation area ratio higher than 60% showing 60% of farmland should be under irrigation, as a capital investment in production; (3) climate change and natural disasters should never cause more than 20% of cropland to be damaged; (4) a minimum of 385.7 US$/hm2 should be spent on grain inputs to ensure high yields; (5) agricultural technology should contribute to a 60% increase in grain production; (6) the gap between supply and demand should never be more than 50 billion kg, reflecting the relationship between present productivity and grain requirement; (7) the per capita grain ration should be 100% secure; (8) the percentage of households lacking grain should be less than 3%; (9) provinces such as cities or municipalities that lack grain production should never have grain shortages of more than 40%; (10) the grain consumer price index (CPI) should never be greater than 103.9% (reference year: 2004); (11) the agricultural production material price index should never exceed 106.3% (reference year: 2004); (12) dependence on foreign trade should never exceed 5%; (13) national grain reserves should be at 20% of annual consumption, to shelter against unforeseen events; and (14) the average stock of grain per household should be around 250 kg. Long (2007a) considered grain supply and demand balance, grain production, supply chain and consumption, present grain security and future development and then proposed the following seven indicators: (1) a baseline of 0.08 hm2 of area sown with grain per person; (2) a baseline per capita share of grain of 390 kg; (3) a maximum yield fluctuation range of 5%; (4) a baseline grain stock rate at 20% of annual consumption; (5) a baseline domestic production level of 92% to ensure self-sufficiency; (6) zero percent of the population lacking grain; and (7) limiting price increases to a maximum of 4%. Using the seven indicators, Long (2007a) calculated grain security coefficients over the time period 2000–2006. The average security coefficient was 0.98, showing that China has had high grain security during the twenty-first century (Long, 2007a).

3. Development Stages of Grain Production in China Grain production and supply are affected by many factors. Policy, technology, and natural disasters are the most important, causing significant yield fluctuations over time. Yield fluctuations affect not only the present, but also the trend of grain supply and demand in the future. Therefore,

110

Huixiao Wang et al.

much can be learned about ensuring grain security today by understanding historically influential factors and trends (Long, 2007a). Since the People’s Republic of China began in 1949, Chinese grain production has steadily increased (Fig. 1). While the frequency of yield fluctuations was less before the economic reform at the end of 1970s, the magnitude of the fluctuations was greater (Li et al., 2001; Lin and Yu, 2006). The grain yield nearly quadrupled from 113 million tons in 1949 to 512 million tons in 1998, with grain per capita doubling from 208 to 412 kg. Following 1998, however, yields began to decrease, with a total yield of 430 million tons and 334 kg per capita in 2003. If a stage is defined as change in yield of 100 million tons, grain yields have climbed four stages from 1949 to 1996, increasing from 100 to 500 million tons. Stage 1 spanned the 9-year period from 1949 to 1957, at an annual increment of 9.1 million tons, followed by stage 2 over the 19-year period from 1958 to 1978 at a yearly increment of 5.4 million tons. Stage 3 was only 5 years, from 1979 to 1984, with an annual increment of 20.5 million tons, while stage 4 covered the 11 years through 1996, at an annual increment of 8.8 million tons (Zhang, 2005). The policy, technology, and natural conditions that were influential on grain production during each of these stages are described below.

3.1. Stage 1: 1949–1957 Prior to 1949, when the People’s Republic of China was founded, land reform had already begun in the liberated area. Its completion between 1949 and 1952 ended the feudal system which had governed China for several thousand years, as well as the exploitation of peasants by landlords through high land rent and usury. After the land reform was complete, and 15

600

10 5 400

0 −5

300

−10 −15

200 Grain yield Variation coefficient

100 0 1949

−20 −25

1954

Figure 1

1959

1964

1969

1974 1979 Year

1984

1989

1994

1999

−30 2004

Total annual grain yield and its variation coefficient in China.

Variation coefficient (%)

Grain yield (million tons)

500

Problems, Challenges, and Strategic Options of Grain Security in China

111

peasants had ownership of land for the first time, their motivation and enthusiasm resulted in strong increases in yields; however, grain production was still at a very low level. During the initial period of new China, much of the grain production was destroyed by a long-term war, resulting in a large disparity between supply and demand, and strong fluctuations in grain price. At that time, the main challenge to grain security was achieving sufficient supply, and the primary task of government was to stimulate grain production, guarantee a normal supply of grain, and stabilize the grain market and price. To alleviate pressure of grain scarcity, the government created a policy called ‘‘Taking grain as the key link,’’ which prioritized grain production importance. Under this policy, the government paid more attention to grain production and gradually strengthened control so as to be very effective in the short term. As a result, in the period from 1949 to 1957, total national grain yield rose from 113 to 195 million tons, while the market price of grain remained basically steady.

3.2. Stage 2: 1958–1978 From 1958 to 1978, China experienced 3 years of natural disasters, as well as political campaigns such as the Great Leap Forward, Mass Movement of Steel and Iron Making, the People’s Commune Movement, and the Cultural Revolution, which completely destroyed the whole economy and production system. Thus, in stage 2, grain production increased very slowly, at about 6 million tons every year. From 1958 to 1961, grain yield declined by around 31% from 177 to 137 million tons. One reason was natural disaster: for three consecutive years, China suffered from severe drought and bad weather, resulting in declining yields, with affected area of 45, 66, and 62 million hm2, and damaged area of 14, 25, and 29 million hm2, for each of the 3 years, respectively. Both natural and man-induced disasters restricted the incentive of farmers to cultivate, causing large declines in grain yields (Long, 2007b). Food supply was not able to keep up with demand, resulting in severe famine—the most serious of which was during the Great Leap Forward, with some 30 million or more famine-related deaths. The political philosophy to reorganize the farmers into very large people’s communes of about 5400 households and 25,000 members led to devastating results for agricultural production (Heilig, 1999; Peng, 1987).

3.3. Stage 3: 1979–1984 Since 1978, which saw the reintroduction of family farming and the dissolution of large collective production units operating on collective land, China’s agricultural sector rapidly began to increase in productivity (Heilig, 1999). In 1978, the household contract responsibility system

112

Huixiao Wang et al.

emerged, which assigned land and farm work to individual households in rural areas, and gradually defined farmers as a separate economic unit, able to benefit individually, which greatly increased motivation and enthusiasm for production. In less than 4 years, more than 90% of the rural areas followed this new system (Wang and Lu, 2006). Due to the combination of the household responsibility system, price policy, technology advancements, and an increase of agricultural inputs, grain production dramatically rose, defining this time frame as the golden period of Chinese agricultural production. Yield increased from 2527 to 3607 kg/hm2, with a total yield increase of 34% from 305 to 407 million tons. For the first time, yields were greater than the baseline requirements, and a surplus was accrued (Lu, 1999).

3.4. Stage 4: 1985–2003 3.4.1. 1985–1995 During this stage, China carried out a grain market reform, so policy, especially changes in price policy, appeared to have the most influence on grain yield changes. In China, grain has been bought and sold in the universal way for a long time. By the mid-1980s, the buying price had dramatically risen above the selling price, with the country compensating the consumer with the difference. Since the financial burden of the country continued to rise with the buying price increasing, a Chinese grain market reform was initiated in 1985. In the previous grain market system, if a peasant sold more than a prescribed amount, the additional sales went for a higher price, benefiting the farmer. Under the new reform system, the price of the additional sales was decided by the market, which greatly lessened the buying price. At the same time, the costs of production materials rose, further reducing the anticipated profits of the farmers, who were then less inclined to cultivate grain crops with high amounts of chemical fertilizer and other inputs (Guo, 1990). In addition, as a response to the surpluses of grain in the previous stage, the government adjusted the planting structure by reducing the area sowed in grain crop: in 1986, it was 4 million hm2 less than that of 1985. These economic policies, especially the grain buying price reform, led to a drop in grain yield for 4 years. To motivate a yield increase, China increased the buying price by 27% in 1989, resulting in a rapid development of grain production and increased yields reaching 446 million tons in 1990 with an annual increment of 39 million tons from 1989 to 1990. Until 1993, China decided to have a completely free grain market. To coordinate grain selling and buying price and to drive grain supply chain system reform, China reduced the grain buying price in 1990 and 1991, resulting in a yield decline of 5% in 1991. Then after the grain price was free, to compensate and motivate the farmers growing crops, the

Problems, Challenges, and Strategic Options of Grain Security in China

113

government greatly raised the buying price by 47% and 29% in 1994 and 1995, respectively. In addition, a subsidy was given to enterprises dealing with grain production and other supply chain links in order to assist in a supply chain system reform. As a result, grain yield increased steadily from 1991 to 1995 (Mao, 2005). 3.4.2. 1996–1999 Starting in 1996, China established a firm buying price to purchase all excess grain to store as a surplus protection against unforeseen events, thus ensuring that farmers can sell all the grain that they can grow at a high price to protect grain-growing farmers. In the same year, supply chain regulation and grain market management policies were implemented, stimulating further yield increases. From 1996 to 1998, yields remained high at around 500 million tons, rising to a historical peak of 512 million tons in 1998. 3.4.3. 2000–2003 After a continuing trend of good harvests from 1996 to 1999, all of the main agricultural products were well stocked. Thus, the financial burden of buying grain surplus as a protection policy became too heavy, and the government reduced the grain price. With the lower price came a growing cost of production, sometimes to the point where costs exceeded the price that could be gotten for the grain, and hence farmers were unwilling to cultivate. Additionally, the central government adjusted agricultural structure, reducing much of the grain sowing area and turning some farmland to forest and grass for ecological restoration. Grain production continued to decrease from 1999 to 2003, with the grain sowing area reduced to 99 million hm2 and a yield of 431 million in 2003. The proportion of grain deficit to yield reached 13% in 2003, normally about 5%. Each big price inflation event has been directly related to a shortage of grain supply. When grain production declines, yet consumption steadily increases and imports and exports remain balanced, then the gap between grain production and demand must be compensated by using stock. Faced with an 84% drop in yield since 1998, the government raised grain price at the end of 2003 and the beginning of 2004. The Grain-for-Green program (also known as Sloped Land Conservation Program) was implemented in 1999 by China as a cropland set-aside program to increase forest cover and prevent soil erosion (Xu et al., 2002). By setting aside more than 7 million hm2, China’s program became the largest of its kind in the developing world. The program was designed to curtail soil erosion in China’s major river basins, and to reduce the rising incidence of floods that were thought to be caused by the increased siltation build-up in the country’s river system. However, it also served to dramatically reduce the amount of land available for grain cultivation, and thus production and yield sharply fell. The program was then blamed as the source of an unprecedented fall in China’s domestic grain production, and

114

Huixiao Wang et al.

leadership severely curtailed the progress of the program in 2004. But many questioned how closely the two trends were actually connected. Xu et al. (2006) reported that the effects of the Grain-for-Green program on grain price and imports were almost nonexistent: only the lowest quality land is retired through the program, lessening overall effects on production. As prices rise, farmers respond by increasing production intensity on the remaining higher quality land. The study of Sun et al. (2006) showed that the Grain-for-Green program could not only improve China’s ecological quality, but also influence grain supplies in the short term, but at the national level should not cause a grain shortage or threaten the grain security criteria.

3.5. Stage 5: 2004–now By 2003, the grain stock was reduced to rock bottom. To intercept the trend of production declining, the central government issued a series of measures to promote grain production, including several No. 1 documents of the Central Committee, reductions or exemptions to the agricultural and agricultural specialties tax, and subsidies for farmers to buy better seed varieties and tools. The subsidy was 6.4 billion US$ in 2004, and 10.1 billion US$ in 2005. Whether these policies are sustainable or not in the long term, they did have obvious immediate effects on motivating production. Grain yield reached 470 million tons in 2004 and increased by 39 million tons, or 9%, from the previous year, which was the highest annual increment in history. Yields continued to rise by around 15 million tons annually on the average, reaching 511 million tons in 2007. Grain yield increased for five consecutive years, signifying that the comprehensive grain production capacity had reached a new level. The great achievements during these years can be attributed to the government’s strong focus on grain production, resulting in the launch of engineering projects and other activities dedicated to improving grain production nearly nationwide, as described in detailed case studies of Section 7.

4. Achievements and Experiences of Grain Production in China 4.1. Achievements 4.1.1. Disputation of China’s grain security Due to its large population size, China has received much international attention concerning issues of grain security. One of the most controversial and widely debated studies was a 1995 book by the American agricultural economist, Lester Brown, entitled ‘‘Who will feed China? Wake-up call for a small planet’’ (Brown, 1995). The book predicted that China would see a

Problems, Challenges, and Strategic Options of Grain Security in China

115

grain yield reduction of 20% by 2030 if diet remained constant, resulting in a need to import 200–369 million tons of grain, the current amount of total world grain trade. Brown also published several other papers on this issue, speculating on gaps between supply and demand of different grain commodities in China, the rates of depletion of reserves, and the influence on the global economy if China became an importer. For example, Brown determined that the gap between wheat supply and demand in China was around 19 million tons in 2003, and that the wheat reserve would be used up by 2005, so that China would then need to depend on imports. Similarly, he estimated the global annual rice-exporting volume to be around 26 million tons in 2003, which is just barely enough to meet the needs of a predicted gap between rice supply and demand in China of 20 million tons, thus causing turbulence in the world rice economy. The study predicted that China will soon also be an importer of maize, with a gap between supply and demand around 15 million tons, and most of the reserve already gone. With grain yield declining each year, population growth increasing by 11 million annually, and rapidly improving incomes, Brown predicted that China’s requirement for grain would continue to increase over the long term. To meet this requirement, Brown speculated that China will enter the global market in an unprecedented geostrategic situation, where America controls half of the global grain exports and China must scramble to meet the needs of its 1.3 billion population. If China raises its total grain consumption from the present per capita rate of just under 300 to 400 kg (roughly half of the per capita rate of the United States) by the year 2030, Brown predicts that it will need to import around 369 million tons of grain in 2030. If China’s rapid industrialization continues, its import demand will soon overwhelm the export capacity of the United States and other grain-exporting countries, resulting in shortages as well as serious impacts to the environment and the world’s natural resources. 4.1.2. Successful self-sufficiency In contrast to Brown’s predictions, China has proven itself to be very capable of self-sufficient grain production. In fact, China successfully feeds 22% of the world’s population using only 9% of the arable land and 8% of the freshwater of the world, an achievement attracting worldwide attention. China has had little need to rely on imports, and has played an important role in helping many developing countries in Africa and Southeast Asia to improve their grain production and reduce poverty. Thus, instead of imposing a threat to the world’s grain security, China plays an important role in securing it (State Council, 1996). By 1995, the per capita share of grain had reached the average world level. At present, the grain reserve volume is 150–200 million tons, which is double the world average, with a reserving rate of 20%. For the last decade, China has been able to produce over 95% of its grain requirement, and in

116

Huixiao Wang et al.

the last two decades, food and fiber problems have been remedied for at least 200 million impoverished people living in rural areas. Furthermore, the 25-year grain assistance of the United Nation’s World Food Programme (WFP) to China was ended in 2005. In summary, China has been very successful in meeting its primary grain security indicators. China has also played an important role in contributing to global grain security. Twenty percent of commodity aid to countries in need came from China in last 25 years. According to the WFP report (2007), China is now the third biggest supplier of aid in the form of grain, following America and the European Union. China provided 577,000 tons of grain as aid in 2005, 2.6 times the amount of aid given in 2004. The use of super hybrid rice not only solved China’s own grain problem, but also provided help to other developing countries in preventing famine and grain shortage (Zhang and Duan, 2008).

4.2. Experiences 4.2.1. Government taking the responsibility for grain security Grain security forms the basis for the national economy and political security, and therefore is an important indicator for judging the degree of national economic security of a country. After the establishment of the People’s Republic of China, the Central Committee and the State Council attached high importance to grain security, prioritizing agriculture in national economic development, improving grain production, solving the food problem of its people, and making a great contribution to global grain security. The central government therefore took full responsibility for ensuring grain security, especially in rural areas, with the creation of 10 No. 1 documents concerning the three dimensions of agriculture, rural areas, and the farmer. From 1982 to 1986, a series of five No. 1 documents about the policies and measures of rural reform were issued, resulting in dramatic changes in the rural areas of China. Based on these documents, a land policy was widely adopted that allowed rural farmers to contract with the collectives for the right to work on public-owned lands and keep the yield, in exchange for a fee paid to the central and local governments. This household contract responsibility system was first started in a small village called Xiaogang in Anhui province in 1978, where it greatly improved agricultural production by strengthening the motivation and enthusiasm of the farmers. But the system did not spread widely very quickly because there was a lack of policy supporting it. The No. 1 document in 1982 affirmed the responsibility system as a part of socialist economy, which allowed it to rapidly extend to a wide range of rural areas throughout China. The focus of the No. 1 document of 1983 was transformation from traditional full or partial sustenance farming to a modern agriculture of large-scale commodity production. The No. 1 document of 1984 further refined and stabilized

Problems, Challenges, and Strategic Options of Grain Security in China

117

the household contract responsibility system, stipulating the time frame of the land contracts to be 15 years. In 1985, the No. 1 document cancelled the fixed quota purchasing system of farm products by the state, a method that had been in place for the past 30 years, to solve the difficulties in selling grain because of the lower grain price. Finally, the topic of the 1986 No. 1 document was to increase agricultural inputs and to regulate the relationships between industry and agriculture, and urban and rural areas. The five No. 1 documents from 1982 to1986 stressed solutions to the problem of food and clothing facing 800 million farmers, laying a substantial foundation for the coming urban reform. Because of the balanced development in the rural areas, since 1987 the same kind of No. 1 document stopped to be issued by the government. The declining yields between 1999 and 2003, the resulting shortages of grain, and a low 4% rate of increase in farmers’ incomes during this time together spurred a significant income gap between urban and rural areas. This income gap led to unbalanced development between cities and the countryside, hindering the goal of a nationwide well-off society. So after 18 years since 1986, another five No. 1 documents of the Central Committee were issued again from 2004 to 2008, focused again on the three dimensional rural issues of agriculture, rural area, and the farmer, which concentrated on improving farmer’s income, increasing agricultural comprehensive productive capacity, building a new socialist countryside, developing modern agriculture, and reinforcing agricultural infrastructure construction. The overall goal of the last five No. 1 documents was to narrow the differences between urban and rural areas by establishing a long-term mechanism for improving the integration of urban and rural social and economic development. In addition to the documents, additional policies were adopted in an effort to mitigate the farmers’ burden, such as reductions or exemptions to the agricultural and agricultural specialties tax, and subsidies for farmers to buy better seed varieties and tools. The tax and fee exemption would have minimal financial impact on the government, while promoting goodwill among farmers (Tso, 2004). As a result of these new policies, the past 5 years is considered to be one of the best time periods in China’s history for agricultural development, positive change in rural areas, and increased benefits to farmers, with the annual rate of increase in agricultural incomes improving from 4.8% in 2002 to 9% in 2007. In view of the world financial crisis, a stable economy and grain security are more important than ever. At an executive meeting on July 2, 2008, China’s cabinet agreed that the country faced tremendous challenges in increasing output to ensure grain security and that the sprawling industrialization and urbanization had increased demand while land and water shortages and climate change had diminished supply. As a result, the Medium- to Long-Term Plan for National Grain Security was approved (State Council of China, 2008). The plan sets a target of grain self-

118

Huixiao Wang et al.

sufficiency rate at 95% or more over the next 12 years, relying on domestic production. The country’s grain yield is expected to reach 500 and 540 million tons in 2010 and 2020, respectively. To meet this expectation, effort needs to be made to ensure that a minimum of 120 million hm2 of land is cultivated and that the agriculture and water infrastructure is improved (State Council of China, 2008). Plans such as grain production subsidies, which increase annually to boost farmers’ incomes, and a plan to help the northwestern Jilin Province, one of the major grain-producing provinces, to increase grain yield by more than 5 million tons in 5 years were approved, as a matter of strategic importance to national grain security. The Chinese government’s unwavering efforts to ensure grain supply, including its restrictions on grain-consuming bioethanol projects, were behind the stable prices at present time. Since diverting grain from consumption purposes to the production of biofuel will enlarge the gap between grain supply and demand, China has strictly limited the amount that corn and other oil plants can be used for biofuel production, thus maintaining a principle of ‘‘biofuel will not have to scramble for grain with people, and will not have to scramble for land with grain.’’ Instead, China has promoted the use of agricultural wastes, such as straw and livestock and poultry manure, or nongrain crops such as cassava, sweet potato, and sweet sorghum, to produce bioenergy. These nongrain crops can produce fairly high yields at low costs on lower quality alkaline, sandy, or sloped lands. The yield of sweet sorghum or root crops from 1 hm2 can produce 3–10 tons of ethanol fuel. In China, 50 million hm2 of low-quality land could be used for biofuel crops. In summary, the development of biofuel from nongrain crops and agricultural waste can help ensure grain security, while promoting sustainable energy development way as well (Shi, 2008). 4.2.2. The importance of maintaining a high rate of self-sufficiency Despite its large population size, China should not rely on the help of other countries to feed its 1.3 billion people. Dependency on other countries for food can result in unwanted power structures and control relationships. China must always bear in mind the historical lessons learned from time periods when agricultural issues were ignored. For the last 10 years, China has produced over 95% of its grain supply, meeting the self-sufficiency criteria for grain security. In addition, for most of this time period, China has been a net grain exporter, contributing to the world’s grain security. This high self-sufficiency and stability strongly lend itself to insurance of future grain security for China, especially within the framework of the current global grain crisis.

Problems, Challenges, and Strategic Options of Grain Security in China

119

4.2.3. Advanced agricultural techniques To improve grain production, the most important factors are policy, science, technology, and production inputs. Grain yield is the product of the amount of land in production and the yield per unit of area, the latter of which is determined predominantly by the technical level of the agricultural production system. According to the Ministry of Agriculture, science and technology accounts for 49% of grain security in China. High-yielding, good-quality grain varieties have been bred through advanced technology, allowing total yields to increase while the area planted was held constant. The super hybrid rice breeding technique developed in China is a good example, having been praised as the third revolution of rice production. The super hybrid rice promotion project began in China in 1996, the end result of many decades of research on super hybrid rice led by Yuan Longping, an academician from the Chinese National Hybrid Rice Research and Development Center, also known as the ‘‘father of hybrid rice.’’ Guided by Yuan Longping’s research, China is now leading the world in the process of super rice selection and breeding. The yield of the super hybrid rice broke historical records, at over 18,000 kg/hm2 when grown in a small area. The project was divided into three phases: The aim of first stage was to breed a rice variety with a high average yield of 10,500 kg/hm2 over a large area, which was accomplished in 2000. The goal of the second stage was a yield of 12,000 kg/hm2, which was fulfilled in 2004, and two years in advance. Finally, the aim of the third stage is a yield of 13,500 kg/hm2, which is expected by 2010. The third phase of the project has started and will be using predominantly molecular plant breeding technology. In some demonstration areas in Hunan province, there are already yields of 13,500 kg/hm2. China’s government has approved a plan for the development of new crop varieties through biotechnology starting in July 2008, stressing that it could increase agricultural competitiveness and raise farmers’ incomes. The new crop varieties will be developed for disease and stress resistance, high quality and yield, and efficient use under sustainable agricultural development. It is thought that the use of these crops, such as super hybrid rice, will feed nearly 30 million more people. 4.2.4. Grain storage in the farmer’s house The longstanding grain policy of China is aimed at attaining and keeping a high degree of self-sufficiency in grain production. The maintenance of large grain reserves, both at the household and the state level, requires that attention be paid to grain security management. Besides planting grain crops, grain farmers also act as grain depositors. Nearly, every household stores a large amount of grain, contributing to grain security in conjunction with the national barns dispersed throughout the country. According to an investigation by the National Grain Bureau, about 60% of grain reserves is

120

Huixiao Wang et al.

stored in farmers’ home, although that rate has slightly declined in recent years. However, due to the use of simple crude grain storing appliances, farmers have had difficulty protecting the grain from mildew, mice, and moths, among other pests. Currently, many farmers have migrated to urban areas for employment, and have had little time or thought for effectively storing the grain produced rurally. China’s government is very aware of the importance of the grain stored by farmers to national grain security, and has thus focused on helping farmers with these issues. In 2007, the National Grain Bureau of China implemented a program for reducing grain loss after harvest in Shandong, Sichuan, and Liaoning provinces. The grain departments plan to improve the storing equipment of 5 million households between 2007 and 2017, as well as disseminate information to farmers regarding more secure grain storing techniques. The program should prevent grain losses of around 11–18 million tons each year.

5. Problems and Challenges of Grain Security in China Although many achievements have been made in the area of grain production, with current supplies meeting demand, there are still many uncertainties that could derail China’s grain security. Population growth and the increase in meat consumption that comes with economic development will rapidly increase demand, while natural disasters, resource availability, technology growth, and policy can all affect supply. The stability of grain security therefore presents a great challenge, with a need to consider many complex factors simultaneously.

5.1. Cultivated land loss and degradation The comparatively small amount of cultivated land in China is an inherent disadvantage compared to other countries, with cultivated land representing only 13.68% of the national land area and only 9% of the entire world’s cultivated land. Roughly two third of the China’s land is mountainous or desert, unsuitable for crop production. The cultivated land in China can be divided into three categories: high, medium, and low yield. The high-yield land represents only one fourth of the total arable land, located in the southeastern coastal areas and the North China Plain, near the urban and industrial areas of China. At present, the per capita share of cropland in China is only 0.093 hm2, less than 40% of the world average. According to statistics, 666 counties or regions currently have a per capita share of cropland lower than the threshold level of 0.053 hm2 set by the United Nations (Sun, 2008). The opposing forces of population growth and loss of

Area of arable land (million hm2)

Problems, Challenges, and Strategic Options of Grain Security in China

121

140 130 120 110 100 90 1949 1954 1959 1964 1969 1974 1979 1984 1989 1994 Year

Figure 2

Cultivated land area of China from 1949 to 2007.

cultivated land to industrialization and urbanization is expected to further exacerbate this problem. Furthermore, in recent years, the quality of much of the world’s cultivated land has been seriously degraded. Production methods are greatly needed that can increase yield in a sustainable manner. Both the quantity and quality of land for cultivation must be preserved. Figure 2 shows changes in the cultivated land area of China from 1949 to 2007. The statistics on cultivated land during 1949–1960 could basically express the area and its changing trend of the Chinese cultivated land. Due to a lack of data from 1960 to 1985, the cultivated land area was calculated from grain yield data. Data from 1986 to1995 came from the Chinese land survey of 1996, and data after 1996 were directly adopted from the China Land Resources Yearbook (Feng et al., 2005). While the amount of cultivated land in China generally increased after the establishment of the People’s Republic of China in 1949, different periods of time reflected very different rates of change, and sometimes decline. From 1949 to 1957, the cultivated land in China increased quickly, especially for the first few years which showed average annual increases of around 3 million hm2. The rapid increase was in large part due to the land reform movement, which motivated farmers to reclaim land that had been lost during the war. From 1957 to 1961, the Great Leap Forward and the People’s Commune policies resulted in serious loss of cultivated land, further exacerbated by 3 years of natural disasters. The decline was halted by the central government, and the amount of cultivated land increased until 1979. From 1979 to 1999, the amount of cultivated land declined slowly, at an annual rate of 263,000 hm2, due to the Chinese economic transition and township enterprises development. However, steps were taken to protect the cultivated land, such as the first ‘‘Land Management Ordinance’’ initiated in 1986 (ECCLY, 1997). After 1999, the amount

122

Huixiao Wang et al.

of available arable land declined sharply again, with a yearly reduction of 1.5 million hm2. This decline was thought to be due to the ecological land retirement program in conjunction with agricultural structure adjustments and construction. The Chinese government has already begun to remedy the declining, however, adopting many land protection measures and prioritizing land protection as a national policy. As a result, the fulfillment of the ecological land retirement plan will continue at a slower pace, and it is expected that the amount of cultivated land in China is predicted to be stable by 2010 (Feng et al., 2005). To summarize, the main causes of the reduction in cultivated land area can be attributed to agricultural structure adjustments, the ecological land retirement plan, construction encroachments, and natural disasters. Following the enactment of the household contract responsibility system, farmers pursued maximum efficiency of land use, often retiring or changing the use of lowquality land. In addition, many farmers migrated to urban areas, leaving rural land unused. Under the ecological land retirement plan, low-quality land was retired, returning it to forest or grassland. From 1997 to 2004, agricultural structure adjustment and ecological land retirement accounted for 78.61% of the decline in arable land, while construction and natural disaster accounted for 15.69% and 5.70%, respectively (Zhu et al., 2007). In addition to quantity, land quality is also low. Taking yield per unit area as the grading standard, high-, middle-, and low-yield farmland accounts for 21%, 37%, and 42% of the total, respectively. Most land has a moderate level of soil fertility, with only 1–2% organic matter content. Sixty percent of farmland is subject to drought, sloped aspect, flood, salinization, and/or other natural factors. Furthermore, farmland quality is declining due to man-induced factors: most construction has been taking place on highquality land. The reserved land that remains for cultivation generally produces only one third of the yield that could be grown on higher quality land (Yu and Hu, 2003). In addition, farmers have been replacing organic fertilizers with chemical forms, thus leading to imbalances in soil nutrients. These effects are compounded by soil erosion, salinization, soil pollution and acid rain, which further lower quality. Any further intensification of growing practices may threaten the long-term sustainability of agricultural production (Verburg et al., 2000). Meanwhile, China’s remaining farmland was approximately 4.9 million hm2 in 2003, 74% of which was located in the west, which has poorer natural conditions for supporting agricultural production (Ministry of Land and Resources, 2003). In summary, cultivated land security is the basis of grain security. Therefore, to achieve grain security, land security must include the following three aspects: (1) quantity, determined by per capita land area; (2) quality, reflecting productivity per unit area; and (3) ecological security, represented by the environmental impacts of the production process (Zhang, 2005).

Problems, Challenges, and Strategic Options of Grain Security in China

123

5.2. Water resources problems Water shortage is a serious problem worldwide. In China, per capita water availability is below 25% of the world average. Water resource shortages have become a bottleneck to Chinese economic development. The agriculture sector is the largest user of water, although its share in total annual water consumption has dropped from 97% in 1949 to 69% in recent times. Limited water resources impede not only agricultural production, but also the growth of the entire economy (Zhu et al., 2008). Regional distribution of water resources in China does not match agricultural irrigation demand. While some 44% of the population and some 58% of the cultivated land are in the northern and northeastern provinces, only 14.4% of the total water resources (surface runoff and groundwater) can be found in those regions (Heilig, 1999). With the demand for water in China increasing sixfold since 1949, the northern half of the country has become a water-deficit region. Throughout much of this region, the deficit is being filled by pumping down aquifers. At some point, pumping necessarily will be reduced to the rate of aquifer recharge, reducing irrigated area accordingly (Brown, 1998). Excessive exploitation of groundwater causes groundwater table decline and related ecological and environmental troubles. Overextraction of groundwater in the North China Plain has expanded to an area of nearly 9 million hm2, 70% of the northern part of the North China Plain. Since the 1950s, the accumulated overextracted groundwater amount has totaled 90 billion m3. Water-use efficiency is very low in all sectors, but particularly in irrigation. Experts have estimated that up to 60% of water evaporates from open canals and from fields with traditional flooding irrigation. There are also significant water losses due to outdated water supply infrastructure, bad maintenance, and poor management practices (Heilig, 1999). At the same time, more and more agricultural water is being diverted to meet the needs of industry development and urban expansion. Thus, the annual water deficit in agriculture is 30 billion m3 in China (Wang, 2004). Water pollution aggravates water scarcity even further. According to the statistical data of the Ministry of Environmental Protection of the People’s Republic of China (2008), surface water pollution was still a very serious concern in China in 2007. Different levels of water pollution were found in the seven biggest water systems (Yangtze River, Yellow River, Zhujiang River, Songhua River, Huaihe River, Haihe River, and Liaohe River). In 407 cross sections of 197 rivers, 49.9% had water quality of I–III, 26.5% were at IV–V, and 23.6% were lower than V. Water quality of the Yangtze River and Zhujiang River was considered good, while there was slight pollution in Songhua River, moderate pollution in the Yellow River and Huaihe River, and severe pollution in Haihe River and Liaohe River (Table 1).

124 Table 1

Huixiao Wang et al.

The pollution degree of major rivers and lakes in China (2007) Water quality (%)

River or lake

Pollution degree

Sections

II

Yangtze River Zhujiang River Songhua River Yellow River Huaihe River Haihe River Liaohe River Tai lake Dianchi Chao lake

Little

103

81.5

3.9

Little

33

81.8

15.2

Slight

42

23.8

52.4

4.8

19.0

Medium

44

63.7

9.1

4.5

22.7

Medium

86

25.6

39.5

9.3

25.6

Severe

62

25.9

9.7

11.3

53.1

Severe

37

43.2

10.8

5.5

40.5

Severe Severe Severe

21 8 12

2.3 12.5 8.3

22.7 12.5 41.7

22.7

34.1 62.5 50.0

III

18.2 12.5

IV

V

7.8

Lower than V

6.8 3.0

5.3. Inferior climatic conditions and severe natural disasters China has relatively harsh climatic conditions compared with much of the world. Most of China has a continental monsoon climate, with greater seasonal temperature variations than other areas at the same latitude, such as North America and Western Europe. In most parts of China, it is cold in winter and hot in summer with extremely low and high temperatures. Precipitation in China is unevenly distributed both seasonally and spatially. Most of the precipitation occurs in summer and varies greatly among regions. There is a lack of harmony between the timing of the crop water requirement and the occurrence of the summer monsoon (Bruins and Bu, 2006). In addition, China frequently suffers from meteorological disasters, which are unusual in terms of the scope of affected areas, the number of disasters, the gravity of the disaster, and the amount of affected population. Natural disasters are the main factor responsible for reductions in grain yield in China. From 1950 to 2002, 26% of all farmland was affected by natural disasters, with 12% seriously damaged, and an average annual grain loss approaching 30 billion kg (Li et al., 2005). Drought and flood are the natural disasters with the greatest threat to grain security. Between 1988 and

125

Problems, Challenges, and Strategic Options of Grain Security in China

60 Affected area Area (million hm2)

50

Damaged area

40 30 20 10 0

1950s

1960s

1970s

1980s

1990s

2000–2002

Year

Figure 3 The amount of affected and damaged farmland by natural disasters in China from 1950 to 2002.

1995, China lost around 856,000 hm2 of cropland due to disasters— primarily flooding. Currently, around 33,000 medium and smaller dams and dykes need urgent repairs, better maintenance, or even reconstruction for flood control. Annually, grain harvests harmed by drought have reached more than 33 million hm2 (Wang, 2008b). In recent years, the rate of damage to agriculture from natural disasters has been very high: grain loss from disasters in 2004, 2005, and 2006 was 31, 35, and 45 billion kg, respectively. The annual amount of farmland affected or damaged by natural disasters since the 1950s is graphed in Fig. 3, showing a trend of increasing impact over time.

5.4. Vulnerable ecosystems There are many vulnerable ecosystems throughout China. In 2005, national forest area represented 18.21% of total land, at 175 million hm2. China’s grassland area was 400 million hm2, represented mostly by high-cold prairie and desert steppe. A smaller amount of temperate grasslands located in Northern China are on the verge of degradation and desertification due to drought and environmental deterioration. The Comprehensive Scientific Investigation of China’s Erosion and Ecological Security report, covering the 3-year time span from 2005 to 2008, stated that 360 million hm2, or 37.2% of China’s land is impacted by erosion: 161 million hm2 is due to water erosion, while 196 hm2 resulted from aeolian erosion. Ninety percent of natural grassland has been degraded in varying degrees, with an annual degradation increment of 2 million hm2. Soil desertification in China impacts 174 million hm2, distributed over 30 provinces (municipalities or cities). The annual sediment in the Yellow

126

Huixiao Wang et al.

River Basin is 1.6 billion tons and the annual soil loss in the Yangtze River Basin is 2.4 billion tons. Soil erosion also causes nutrient loss. The annual total amount of N, P, and K losses in the Yellow River Basin was estimated to be 40 million tons, which is more than the total fertilizer requirement of 39.9 million tons in the whole country in 2003. Analysis on the data of 2000 showed that the economic loss from erosion each year is over 28.6 billion US$, equivalent to 2.3% of the national GDP. Given the present erosion rate, it is estimated that in 50 years the black soil layer of 0.9 million hm2 cultivated land in Northeastern black earth region will be lost, resulting in a 40% grain yield reduction. The area of rock desert in the Southwestern karst region is predicted to be double in about 35 years, with around 100 million people losing their means to survive. In recent years, an annual increment of over 1.5 million hm2 of land is lost to erosion due to man-induced factors. The report showed that sloping cultivated land, representing 17.5% of all cultivated land, is the main source of erosion, with an annual erosion amount of around 1.5 billion tons, or one third of the total erosion amount in China. The sloping land is mainly distributed in the upper reaches of the Yangtze River region, the Loess Plateau, the rock desert, and the Northeastern black soil region. It is estimated that 1 kg of grain produced in the Loess Plateau will result in a soil loss of 40–60 kg.

5.5. Global climatic change Climate change is a major global issue. There is an increasing trend of warming in China. The nationwide annual mean air temperature is predicted to increase by 1.3 2.1 C from 2002 to 2020 and 2.3 3.3 C by 2050. Climate change has already started to impact agriculture in China. Future climate change can increase instability in agricultural production, and the yields of three main crops, that is, wheat, rice, and maize, are likely to decline if no proper adaptation measures are taken (National Development and Reform Commission of China, 2007; Xiong et al., 2007). Climate change can also benefit agriculture through increased effective accumulated temperatures at crop growth period, and increased concentrations of CO2, which can improve crop yield. However, climate warming can also accelerate crop growth rates which can lower yields, counteracting any climate change benefits. It is estimated that grain yield will probably reduce by 5–10% as a result of global climate change (Chen, 2008). Crops will require more water because soil water will evaporate at a faster pace under higher temperatures. With a 1 C rise in air temperature, irrigation water consumption is expected to increase by 6–10% (Guo et al., 2008). Therefore, it is expected that the gap between supply and demand for water will worsen with climate change. A 2008 Greenpeace International report entitled Climatic Change and China’s Grain Security warned that China may not be able to meet its grain

Problems, Challenges, and Strategic Options of Grain Security in China

127

requirement after 20 years if no actions are taken to mitigate or adapt to climatic change. Climate change is expected to cause tremendous changes in temperature, water resources, frequency of extreme weather events, soil degradation, crop disease, and insect pests. The report predicted that an average temperature rise of 2.5 3.0 C will result in a continuous decline in rice, wheat, and corn yields, reaching a 23% overall reduction by 2050. China is therefore considered to be one of the most sensitive regions to climatic change. While short-term climatic change may cause both positive and negative effects on grain production, long-term influences are expected to be predominantly negative.

5.6. Changes in population growth and the standard of living The Chinese population was 1.3 billion in 2007, and it is estimated that in the first half of the twenty-first century, the population will rise to 1.4 billion by 2010. The population is expected to peak at 1.6 billion by 2050, with a gradual decline thereafter (Feng, 2007; State Council Information Office of China, 2000). The grain consumption per person has increased in recent years. From 1990 to 2002, per capita consumption rose slowly from less than 370 to 375 kg, despite fairly rapid growth in income. Rates then began to increase more rapidly, reaching 388 kg per person in 2007. It is estimated that grain consumption rates will reach 389 kg by 2010 and 395 kg by 2020, with a total grain requirement of 525 and 573 million tons for 2010 and 2020, respectively. The structure of grain demand is expected to change significantly, with direct dietary consumption representing 49% of total consumption by 2010 and 43% in 2020. In contrast indirect consumption through meat is expected to increase to 36% of total consumption in 2010 and 41% in 2020 (National Development and Reform Committee, 2008). Given that indirect consumption through meat requires more grain than direct grain consumption, the future total grain requirement is expected to be higher.

5.7. Small-scale agricultural economy and relatively low profit China has followed the household responsibility system since the 1980s. For more than 20 years, this system has met the requirements for Chinese agricultural economic development. But with the development of a market economy and agricultural commercialization, disadvantages of small-scale family management become increasingly evident. Given the cheap price of labor, the adoption of machines and other new technologies is low. In addition, the small, dispersed plots of land are not suitable for mechanization. Therefore, most agriculture remains at a household sustenance level.

128

Huixiao Wang et al.

Due to low crop output and little collateral, financial institutions are reluctant to assist farmers with capital acquisition, thus further limiting their ability to modernize production. The poor financial prospects associated with farming have resulted in a migration of talented farmers to nonagricultural employments, leaving much land unused and wasted. With more and more of the farming labor force turning to nonagriculture employment, agricultural labor is left predominantly to women and the elderly, most of whom lack a basic education, making the adoption of new technologies difficult. All of these factors adversely affect grain production and reduce grain yield per unit area. Incomes of grain farmers are low compared to that of nonagricultural work. Taking 2005 as an example, the income of growing grain was 521 US$/hm2, and the income per capita was only 69 US$ for a farmer growing grain on 0.13 hm2 of farmland. Lack of an educated labor force limits the ability of farmers to benefit from science and technology, resulting in low adoption of new varieties and techniques in grain production. In recent years, national policies have resulted in stable increases in the incomes of grain farmers; however, they are still significantly lower than incomes from crops such as cotton and vegetables. At present, the profit ratio of grain to cotton is 1:5, grain to vegetable 1:4, so the farmer’s enthusiasm for growing grain crops is still quite low. It is often the case that the areas producing the most grain are also the most impoverished. For example, Henan, the biggest province in China, produces 1/10 of China’s total grain yield. A quarter of Henan’s yield comes from 12 counties in the eastern Henan plain, which have all been designated as impoverished by the government. Therefore, the government should create long-term policies for the support of grain production, as is done in many developed countries.

5.8. Old agricultural capital construction and weak risk resistance Agricultural capital investment is an important cornerstone of modern agriculture, reducing natural and market risks through an increase in capacity to deal with uncertain and unforeseeable events. In China, the long-term investment in agricultural capital has been insufficient. From the first 5-year plan to the ninth 5-year plan, only 3.4–17% of the total national financial budget was invested in agriculture, and only 5.8–18.8% of total capital investment was for agriculture. The national payout on agricultural capital construction shared only 2% of total financial payout in 2005. Most investment goes into hydraulic engineering, while only a small annually unstable amount is used for directly improving agricultural production. Improvement of production at the field level has been far from sufficient, especially regarding field water conservation, which is the basis for improving grain yields. Insufficient investment has resulted in an aging, out of repair

Problems, Challenges, and Strategic Options of Grain Security in China

129

agricultural infrastructure, which cannot support needed grain production (Fang et al., 2004). In 2006, the amount of irrigated farmland with satisfactory irrigation equipment was 55 million hm2, only 45% of the total area of the farmland.

6. Strategy Options and Countermeasures for Grain Security in China China’s grain security is a complex problem, which requires simultaneous consideration of the multitude of diverse factors affecting it to devise an effective solution. There are several dimensions that play a major role in China’s grain prospects, such as population growth, diet change, urbanization, quantity and quality of cultivated land, water supply, policies and economic arrangements, and scientific and technological developments. Each of these dimensions must be taken into account for political planning and scientific research to successfully ensure China’s grain security.

6.1. To reinforce agricultural infrastructures Although agricultural production is currently thriving, agriculture remains an important issue for long-term economic and social development, with improvement of infrastructure a primary concern. Currently, only 45% of farmland has irrigation equipment, most of which is low quality in the face of drought. Consequently, only 13 million hm2 of farmland can ensure stable yields. Lack of adequate infrastructure for coping with natural disasters has resulted in an annual grain loss of around 50 million tons, or 1/10 of the total grain yield. Investment in field infrastructure needs to be strengthened in order to improve grain production. Considering the effects of the nationwide drought of 2006, the ecological crisis of 2007, and the snow disaster of 2008, it is obvious that the lack of agricultural infrastructure contributed to the inability of the nation to meet an emergency. In response to this issue, the focus of the Central Government No. 1 document of 2008 is solely about agricultural infrastructure, with a goal of improving and remedying deficiencies in the agricultural foundational establishment. The goal will be met through the following objectives. (1) Accelerate the construction of supporting facilities for large-scale water-saving irrigation areas: upgrade aging electromechanical equipment, and improve irrigation and drainage systems. (2) Continue to expand demonstrations on water-saving irrigation techniques: conduct pilot projects in the main grain production areas, develop dryland water-saving agricultural production methods, disseminate techniques of dryland farming to farmers in arid areas, and increase the effective irrigation coefficient from its level of 0.45

130

Huixiao Wang et al.

in 2005 to 0.50 by 2010, and over 0.55 by 2020. (3) Conduct small-scale hydraulic engineering focused on field irrigation and drainage, as well as watershed-scale projects for fighting drought. (4) Restore and prevent the further salinization and alkalinization of medium- and low-yield fields in the main grain production areas. (5) Enforce investment in quality agricultural infrastructure that has the highest potential for improving fertility and sustainable output capacity of farmland. (6) Accelerate the engineering of rainwater collection and utilization techniques in mountain areas and other arid regions.

6.2. To increase the fiscal input in agriculture production Although there have been many recent achievements in agriculture due to prioritization in importance, increases in inputs, and improvements of agricultural production conditions, there is still more to be done, especially for improving grain production: Agricultural infrastructure remains inadequate, especially for protection from natural disasters. There are insufficient funds for inputs in grain-producing areas, resulting in low yields. There is a progressive loss in all agricultural productive elements such as land, funds, and labor force, as these resources are redirected toward industrialization and urbanization. Farmers are unwilling to invest in grain production, due to decreasing grain prices and increasing production costs. The recent growth in grain yield can be attributed to three main sources: one is the private investments in labor, fertilizer, machinery, seeds, and other necessary production materials; the second is institutional reforms; and the third is public inputs such as improvements in technology, education, and rural infrastructure. However, institutional reforms are likely to have a limited, one-time effect, and private investments can significantly increase a farmer’s cost, as well as impact the environment if investments result in overuse or misuse of an input. In contrast, studies have shown that public investment in agricultural research, irrigation, education, and other rural infrastructures has been a major driving force in the growth of grain production in China (Fan, 2000). The study found that increasing public investment in agricultural research leads to higher output in a cost effective manner. Increasing public investment in agricultural research and development should therefore be considered a high priority in the future to be utilized as a policy tool to enhance long-term food security (Zhu, 2004). To improve grain production, the first key measure is to increase financial investment. Twenty percent to fifty percent of the gross financial expenditures in developed countries goes to agriculture, compared to 10–20% in most developing countries, and only 8–11% in China. In recent years, China has poured a huge investment into agriculture, reducing the burden on farmers. The national government financial input into agriculture reached 31 billion US$ in 2003. By 2007, it was 62 billion US$, an

Problems, Challenges, and Strategic Options of Grain Security in China

131

increase of 101.4%, while the total financial revenue in 2007 was over 714 billion US$. The accumulated financial input to agriculture from 2003 to 2007 reached 229 billion US$. The total subsidies for farmers to buy better seed varieties and tools reached 15 billion US$ in 2008. However, the proportional share of agricultural investment is still low, with much room for improvement. An increased agricultural investment can help to further modernize agriculture, promote steady growth of grain output, increase farmers’ incomes, improve field irrigation, consolidate farmland, reclaim damaged land, create monitoring and warning systems for agricultural meteorological disaster prevention, and prevent and control crop diseases and insect pests. The financial support should give prominence to the main grain-producing regions and the biggest grain-producing counties, which can serve as pilot regions for basic farmland protection.

6.3. To maximize the role of science and technology in grain production According to a State Council Information Office white paper, the population will reach 1.6 billion, with a total grain requirement of 640 million tons based on a per capita grain consumption of 400 kg (State Council Information Office of China, 1996). There are three ways for solving the grain security problem: reducing total demand, enlarging grain import, and increasing the total grain production output. The first two solutions are infeasible, as the first significantly compromises the population’s standard of living, while the second results in too large of an upset to international trade. Therefore, increasing the total grain production output is the best choice, either through increasing the grain sowing area or improving the grain yield per unit of area. However there are many problems associated with increasing the amount of cultivated land, given the decline in both quantity and quality, the difficulty in reclaiming damaged land, and the unbalanced distributions between land and water resources (Zhang, 2005). Currently, 123 million hm2 of land are cultivated, with 100 million hm2 in grain production. Therefore, the potential of increasing grain yield by area enlargement is rather small. Especially given that in recent years the contradiction between more people and less land in China has become much more acute. The per capita share of cultivated land decreased to 0.093 hm2 in 2006. Therefore improving yield per unit of area is the most feasible way to ensure grain security. Adequate investment in science and technology will be needed to achieve many goals, including the increase of grain yield per unit of area, the melioration of agricultural product quality, the strengthening of agricultural overall production capacity, and the construction of a modern agricultural infrastructure. But at present time, the investment in agricultural research only accounts for 0.1% of the total agricultural productive value in China, only 1/10 of that in the developed

132

Huixiao Wang et al.

countries and also lower than the average level of 0.26% in the developing countries. Currently, science and technology is thought to contribute less than 50% to grain production in China and the conversion rate of scientific research to actual agricultural production achievements is estimated at 30% (State Council Information Office of China, 1996). So policies and measures for driving agricultural science and technology innovation, strengthening agricultural science and technology service system construction to speed the extension of the advanced practical techniques and good varieties are urgently needed so as to turn the experimental yields of the expert into the field yields of farmers. China has devoted great effort to developing its science and technology base, and has made remarkable progress (Tso, 2004). The country has introduced high-yield crops, increased the use of agrochemicals and agricultural machinery, and expanded irrigation. But the plentiful input of fertilizer and pesticide is at the cost of environmental quality. At present, the fertilizer and pesticide amounts applied in China are the largest in the world. In 2003, fertilizer consumption in China accounted for 28% of the world total. Nitrogen-based fertilizers made up 30% of China’s total consumption, a type of fertilizer closely linked not only to high yields, but also to environmental pollution. As a result, China’s cereal output was 19.1% of the world total. However, the high yield was gained at the price of agricultural pollution, including water body eutrophication, nitrate concentration levels in groundwater higher than the drinking water standard, high ammonia concentrations in the rural atmosphere, and greenhouse gas N2O emissions increasing. Therefore, improving grain yield through increased chemical inputs is not considered to be a sustainable long-run solution. In the last 5 years, 40% of grain yield increase was due to an enlargement of sowing area, while 60% was from increases in yield per unit of area. However, the FAO predicts that in the future, only 20% of total world grain increase will come from sowing area enlargement with 80% from yield per unit of area improvements. The studies showed that to guarantee China’s grain security, the annual grain yield increase should be not less than 1%, a goal that can potentially be fulfilled through the use of agricultural biotechnology. In recent years, the Chinese government has given high priority to advanced research in molecular biology, plant genetics, biotechnology, and related fields, all aimed at increasing crop yields. By the mid-1980s, a major national biotechnology program had been initiated. Today, China is a leader in agricultural biotechnology and one of the most advanced countries in terms of using genetic markers and tools in rice breeding. In the future, it will be advanced breeding methods that will help to further increase the productivity of crop plants in China (Heilig, 1999). Transgenic technology is expected to be the leading form of agricultural biotechnology in the world (Wu et al., 2007). The integration of agricultural science and technology has

Problems, Challenges, and Strategic Options of Grain Security in China

133

already played an important role in grain production, with core techniques including the creation and distribution of productive varieties, the determination of fertilizer application rates through soil testing, the implementation of integrated scientific field management, and the implementation of sowing and harvesting by machine.

6.4. To protect the quantity and quality of cultivated land With the continuous increase in population and an ever-growing demand for food, the pressure on cultivated land has been mounting. A series of policy measures were implemented to stabilize the domestic grain supply and to halt the decline in cultivated land. The Provincial Governor Responsibility System for grain supply was introduced in early 1995. Provincial governments were required to take responsibility for maintaining a dynamic balance of cultivated land with other land uses in their jurisdictions: losses of cultivated land had to be compensated by reclaiming new land, rehabilitating damaged land, and/or reusing deserted land. Policies such as this are essential to maintain the ‘‘red line’’ threshold minimum of 120 million hm2 of cultivated land (Yang and Li, 2000). After implementing the contract responsibility system, where remuneration was linked to the household, a farming policy of equally dividing the land among the people was adopted. While greatly stimulating production, the system was not suited for modern agricultural production, as it still favored sustenance-level, small-scale production. Agriculture under this system was incapable of utilizing modern infrastructure, capital, technology, scientific knowledge, and management in an optimal manner. With the ability to transfer land-use rights, farmers had a form of capital to improve land-use efficiency, enlarge production scale, implement mechanized farming, introduce new techniques and varieties, and regulate cropping structure. Given the correlation between land-use efficiency and income, this policy will likely result in more efficient management and land concentration so as to achieve economy of scale production and allow farmers to reap the all of the benefits of modern agricultural technology. The core of the rural land system in China is a dualistic philosophy which separates ownership from right of use. While the actual ownership of land belongs to the collectives in a town or village, the rights of land use are distributed among the households. On the 30th anniversary of China’s reform, it was very important to push for rural reform development. The Third Plenary Session of the Seventeenth Central Committee of the Communist Party of China was convened in October 2008, and the Central Committee discussed and passed A Decision on Several Important Issues for Advancing Rural Reform Development, which became a guiding principle for rural reform in China. In the decision, it was mandated that the Management in Contract with Families and the Unified and Decentralized

134

Huixiao Wang et al.

Management System would be continued, they are the foundation of the rural policy of the Communist Party of China. In addition, the decision advocated that even more full contracting management rights of rural land be given to the farmers, to ensure long-term stability. The decision decreed that according to the legal, willing and being paid principles, the farmers are allowed to circulate the contracting management rights of land as follows: through subcontracting, renting, interchanging, attorning, and joint-stock cooperation. While larger-scale farms such as professional family, family farm, and professional farmer cooperative are allowed, under the conditions of no change of land usage and no harm to farmers’ rights and interests. Moderate concentration of cultivated lands in order to scale up production is the only way for farmers to raise their standard of living in the twenty-first century. One feasible choice is to reconstruct the Agricultural Development Bank of China, letting it to take on the function of a national land bank. It then must follow a principle of ‘‘deposit high and loan low,’’ which serves to encourage farmers to deposit unused land into the bank so that the farmers who want to cultivate more can take out land loans. A land bank as an agency institution has three advantages: one is to improve the orderliness of land circulation, which is beneficial for land market management; the second is to facilitate ease in land concentration to achieve economies of scale; and the last is to alleviate risk to farmers making deposits or taking loans through national credit assurance. To set up a national land bank, there is a technical issue that needs to be solved regarding who will fill the gap between deposit and loan. Since the government should pay for grain security, this money should come from the government financial budget. Given that it is commonly agreed that grain security is a priority, the government should be willing to cede a portion of the financial budget to fund this enterprise. If the national financial situation ever becomes critical, the money could be raised by collecting a grain security tax from the residents of cities and towns, and from the industrial enterprises who occupy arable land (Wang, 2005). The low quality of land has affected China’s agricultural production and grain security through the need to increase the amount of agricultural inputs and labor, which raises production costs and lowers the economic benefits. To deal with declining land quality, soil fertility must be improved. Restoring soil quality through improvements in the soil organic carbon pool is essential to increasing agronomic yields (Lal, 2007). Three potential measures for improving land quality are as follows: through policy, through evaluation, and through technology.

Problems, Challenges, and Strategic Options of Grain Security in China

135

6.5. To protect water resources and improve water-use efficiency Water is a precious resource for grain production. China’s supply of groundwater is limited, and the surface water is unevenly distributed, with only 20% of water collecting in the northern plain where 64% of arable land is located. Large parts of the existing agricultural areas in the North cannot be cultivated to their full potential due to insufficient rainfall. Water scarcity has increasingly become a constraint to economic development, particularly grain production. So water is certainly an important factor for China’s grain security, particularly in northern China, where water resources for agriculture are becoming increasingly exhausted and diverted to urban and industrial consumption. The government has addressed this problem by various means. One answer would be to fully and wisely use all water resources by water conservation and diversion, but south to north water transfer from the Yangtze River would not benefit agriculture (Tso, 2004). China must prioritize sound management over further construction of new water projects. Water management involves many aspects, such as policy, legislation (law and regulations) institutions, organization, personnel, finance, and operation. Water management must focus on rational use, with water conservation as a core goal. Therefore, water resource management should derive a national policy and water law, which potentially creates market incentives through water fees, as well as reforms agricultural water management in irrigated districts. Water conservation in agriculture can be achieved through highly efficient irrigation conveyance canal systems and better field water use in cropping systems. This should include projects to increase water supply in agricultural regions with water deficits, projects to improve water quality and wastewater treatment, and in particular, projects to improve irrigation efficiency. Various authors have pointed out the problem of water waste in open irrigation canals and on floodirrigated fields. The estimated water loss is in the range of up to 60%. This is a significant loss of water resources for the draught-affected North China Plain, a problem which could be remedied in a relatively short time through better maintenance of irrigation infrastructure and more advanced irrigation technology. Water conservation in irrigation is therefore critical for China’s grain security (Blanke et al., 2007; Heilig, 1999). Human use and pollution of freshwater have reached a level where water scarcity will potentially limit grain production, ecosystem function, and urban supply in the decades to come. Feasible expansion of irrigated agriculture will be able to accommodate only a portion of increased demand for grain, while the rest must come from an increase in the productivity of rainfed agriculture ( Jury and Vaux, 2007).

136

Huixiao Wang et al.

6.6. To protect and enhance the farmers’ initiative for grain production It is important to consolidate and fine-tune policies that are beneficial to agriculture in order to ensure the continued motivation and initiative of local governments and farmers to prioritize grain production. The lower profits associated with producing grain has resulted in a lack of initiative among farmers to produce grain, as well as financial difficulties in the core grain-producing areas. The markup of agricultural materials has kept production costs high recently. Despite an increase in the minimum purchasing prices of rice and wheat, it is not enough to counteract the markup of chemical fertilizer, diesel oil, agricultural machinery, and seed. Some measures should therefore be taken to stabilize the agricultural materials market, and effectively control the climbing cost of growing grain crops. The actual incomes (including subsidies) for producing grain were 454, 819, 705, and 686 US$ per hectare during 2003–2006. To guarantee national grain security, the farmer must benefit from grain production, or there will be no motivation for planting crops. Benefits can be increased either through change in grain price, or through government subsidies. For developed countries like Europe and America, normally over 60% of farmer’s profit comes from government subsidy. But in China that is unpractical at present time due to limited financial support and a large number of farmers. Therefore, price appears to be the primary source of maintaining the farmer’s initiative to produce grain. The experiences in the years preceding 2004 confirm this idea: when grain prices were too low, grain yield fluctuated for a long time, sometimes decreasing. After 2004, with the abolishment of agricultural tax and the increase in subsidies to reduce the cost of producing grain, grain yields began successively increasing each year (Deng, 2008). China’s grain security is determined by the farmer’s enthusiasm to grow grain crops. As long as farmers cultivate grain and no land is left fallow, China should be able to maintain grain security. To sustain continuing high motivation of farmers to plant grains, the subsidy policies should be reevaluated to ensure clear objectives, direct benefits to the farmer, and ease in facilitation. Encouragement and assistance should be given to the big grainproducing counties having financial difficulties, to support the construction of core grain-producing regions. In addition, a grain price mechanism should be created which reflects market supply and demand, considering benefits accrued to both the producer and consumer, while protecting profits of grain-producing farmers through a detailed understanding of costs and returns. A minimum grain purchasing price policy should be continuously implemented to prevent grain price declining and to lead farmers to extend grain production consciously. A mechanism to compensate farmers for losses should be established to increase their capacity to elude risks associated with grain production. Agricultural disaster insurance should be

Problems, Challenges, and Strategic Options of Grain Security in China

137

gradually developed for commodity grain production bases. Finally, a warning and prediction system for extreme meteorological disasters should be implemented to help grain farmers avoid disaster risk (Deng, 2008).

6.7. To response climatic change China has already taken actions for coping with climatic change. Promoting sustainable development, the Chinese government has implemented a national environmental protection policy, using science and technology to reduce energy consumption. China’s National Climate Change Programme was instigated in June, 2007, targeting a reduction in energy consumption per unit of GDP by 20%, a reduction in total discharge of major pollutants by 10%, an increase in the proportion of renewable energy supply up to 10%, and an increase in forest coverage of 20% by 2010 (compared with 18.2% coverage in 2005). Regarding agricultural production, the agricultural structure and patterns of planting and breeding will be regulated, so that good varieties and advanced production techniques can improve the ability of agriculture to adapt to climatic change (National Development and Reform Commission of China, 2007). Greenpeace International’s ‘‘Climatic Change and China’s Grain Security’’ report discussed the need for conservation tillage and planting land with forests and grass as a method to discharge greenhouse gases and thus combat climatic change. Both traditional agriculture and modern agriculture, with its overreliance on agrochemicals, should be replaced by ecological agriculture. Ecological agriculture is based on maintaining soil fertility naturally, with the resulting reductions in external fertilizer inputs helping to protect the environment, while not sacrificing yield, quality, and efficiency. However, the development of ecological agriculture still faces many difficulties, such as high costs, input scarcity, questionable efficacy of technique, and high labor demand. Thus, ecological agriculture cannot currently be employed widely in China. Using national finance, a subsidy should be available for the further development of ecological agriculture. Exchanges and development of technology could be reinforced through international cooperation, as a positive response to climatic change.

6.8. To develop grass agriculture For a long time, the target of grain security has been to balance the total supply and demand of grain ration, feed grain, grain used for industry, and seed grain. The proportion of the per capita grain ration gradually reduced from 280 million tons in 1995 to 264 million tons in 2005, while during the same period, feed grain and grain used for industry gradually increased from 117 and 38 million tons in 1995 to 165 and 58 million tons in 2005. The main reason for this change is the increasing meat consumption of the

138

Huixiao Wang et al.

population as the standard of living improved. At present, feed grain accounts for more than one third of the total grain consumption, and it is estimated that feed grain will represent 38%, 43%, and 50% of total consumption in 2010, 2020, and 2030, respectively. China is the second biggest country in terms of grassland, with a grassland area of approximately 400 million hm2, three times that of its cultivated land. However, many issues affect China’s grasslands. Management of predators resulted in a high overloading of herbivores above the carrying capacity of the land. People also ignored artificial grassland cultivation with very low input even no input. The quality of grassland in China is therefore very poor, and the environment is badly destroyed, resulting in grassland degradation or desertification. For example, the productivity of livestock on the grassland in Inner Mongolia was only 1/20–1/80 of those in developed countries in 1994. However, in the northern grassland area of China, if only 30 kg/hm2 of urea are applied, grassland can increase output by about 35%. Given that the productivity of farmed grass is double that of natural grassland, the development of artificial grassland is an important consideration for livestock development. Agricultural production of artificial grassland accounts for at least 10% of the total grasslands in developed countries, while it only represents 2% of the grasslands in China. About 20 million hm2 of land in the northern pasturing areas are suitable for farming grass, with around 12 million hm2 in southern areas, which means China has great potential for cultivating grassland (Ren, 2008).

6.9. To effectively control the growth rate of population The government’s national policy on family planning has brought the excessive population growth trend under effective control. According to the United Nations, China’s fertility rate was lower than that of other developing countries and the world average. In 2005, the birth rate in China was 12.40%, with a population growth rate of 5.89%, making China’s fertility rate one of lowest in the world. As a country with an underdeveloped economy, China has accomplished a historical transition in population reproduction pattern from one featuring a high birth rate, low death rate and high growth rate to one featuring a low birth rate, low death rate and low growth rate. This change was accomplished in a relatively short period of time, compared to the decades it has taken many developed countries to accomplish similar transitions in the past. Since the implementation of the family planning program in 1981, over 300 million births were averted by 2005. It is a significant contribution to controlling world population. China’s grain security greatly depends on the continuation of an effective family planning program. An increase in average fertility would result in a big increase in the number of people due to the large number of people at reproductive age.

Problems, Challenges, and Strategic Options of Grain Security in China

139

Stabilizing population growth is probably the single most important measure for increasing China’s grain security (Heilig, 1999).

7. Case Studies 7.1. Science and technology engineering for grain high yield 7.1.1. Background Grain production continuingly decreased from 1999 to 2003, resulting in a gap between grain supply and demand of 40 million tons in 2003, despite the price of grain rising since 2002. With the grain requirement increasing, there is a strong need to produce more grain and to heighten selfsufficiency. Natural disaster, policy, less available farmland, and other factors resulted in the decline in grain production. In addition, the intensive use of inputs results in high production costs. While China’s farmland accounts for only 9% of the world’s total cultivated land, its fertilizer use represents 33% of the world’s fertilizer consumption. Therefore, production costs are often high, with low profits, substantial risk, and unstable yields. Technology only contributes 48% to production, while the world average level is 71%. While technology has not played a strong role in Chinese agricultural production yet, it means there is still great potential for improvements (Wang, 2008a). Land and water resources are rapidly reducing, so increases in yield and guaranteeing security must depend on the support of technology. At the present time, technology is not able to support agriculture in China. Technology growth has slowed compared to past achievements, such as precision wheat sowing with machines, improved varieties like hybrid maize, dwarfing wheat, and hybrid rice. In China, the potential for heightening production efficiency through the integration of different crops and cultivating techniques is often neglected (Lan et al., 2006). Due to several barriers, agricultural technology has not been extended widely. Rice yield in large experimental plots has exceeded 12,000 kg/hm2, but the national average is only 6300 kg/hm2 over the past 10 years and it was only 4721 kg/hm2 in 2006 (Wang, 2008c). Furthermore, farmers are not getting needed information about new agricultural technology, high-yield breeds and effective cultivation for saving water and fertilizer. Therefore, grain production efficiency is currently lower than it should be. 7.1.2. Objectives of the activity To restore and improve grain production, the Ministry of Science and Technology, the Ministry of Agriculture, and the Ministry of Finance and National Grain Bureau together carried out ‘‘Science and Technology Engineering for Grain High Yield,’’ with the purpose of providing technical support to increase national grain yield and farmer’s income. The project

140

Huixiao Wang et al.

was conducted in the middle lower reaches of the Yangtze River Plain, North China Plain, and Northeast Plain, where provide over 90% of the national commodity grain, with rice, wheat, and corn as the commodities of interest. The project devised innovative solutions such as development of key technology for sustaining high grain yield and reducing postharvest loss, technology integration and transformation, technology demonstration over a large area, monitoring techniques, and grain security strategies. Through the combination of super high-yield varieties, high-yield techniques, and technology demonstrations in large areas, the goal of this program is to restore productive capacity from 440 million tons to 500 million tons. With advances in key techniques, it is expected that technology can provide powerful support to potentially achieve a productive capacity approaching to 540 million tons in 2010 (Shao, 2005). 7.1.3. Main contents of the activity The science and technology behind the engineering for high grain yield was put into effect in stages. The first stage was from 2004 to 2006, the second stage was from 2007 to 2010. The main tasks are to integrate existing techniques with new achievements in grain production, to demonstrate the techniques in a large region, and to make breakthroughs on the key techniques concerning grain yield and profit increases. The 667 hm2 of core experimental area, 67,000 hm2 of demonstration area, and 667,000 hm2 of radicalized affecting area will be used, with an extension area reaching 8.3 million hm2 in the main grain production regions, such as Hunan, Hubei, Jiangsu, Jiangxi, Sichuan, Hebei, Henan, Shandong, Heilongjiang, Jilin, and Liaoning Provinces. In the earlier stage, the program optimized and integrated 20 regional cultivation cost-saving and yield-increasing techniques, fine-tuning the processes for different yield levels, and/or different planting systems. This stage could provide effective technological support for restoring the grain production capacity in China. Average yield was expected to increase by 20%. Furthermore, 10–20 complete sets of highyield cultivation schemes would be formed, providing technological support and demonstration models for national high-yield production (Dong, 2008). For each province, new crop varieties of high quality were selected along with efficient cultivation technology suited to the local environmental conditions. For example, Hebei Province is located in the North China Plain, where winter wheat and summer maize are the main crops and scarce water resources is the main limiting factor of production. Therefore, saving water and heightening water-use efficiency are important aspects in technical integration, including choosing water-saving, drought resistant and high-quality varieties, optimizing and integrating straw return, shrinking rows, implementing water-saving irrigation, and efficient fertilizer application (Kong et al., 2006; Zhang and Zhao, 2007).

Problems, Challenges, and Strategic Options of Grain Security in China

141

7.1.4. Achievements The Science and Technology Engineering for Grain High-Yield program has been active for 5 years now from 2004 with great progress. The demonstrating area of the program was 0.4 billion hm2. More than 100 new grain varieties of high quality and yield have been developed, about 100 cultivation techniques have been fine-tuned to regional environmental conditions, and 22.7 million hectare demonstration areas in 12 provinces have been built, accounting for 10.1% of the total grain production area. Grain yield per unit area was improved by more than 10%. Over the 5 years from 2004 to 2008, grain production increased 37 million tons and economic benefit increased by 7.1 billion US$. In addition, use efficiency of fertilizer and irrigation water in demonstration areas rose by more than 10%, pesticide use was reduced by about 25%. Some grain yield records have been created, yield of double rice harvest reached 19.6 ton/hm2, single rice harvest at 12.8 ton/hm2, maize at 21.0 ton/hm2, wheat at 11.0 ton/hm2, and wheat and corn together at 25.9 ton/hm2. Finally, in the maize area of northeast plain under rainfed conditions, yield was recorded at 17.8 ton/hm2. The study of high-yield technology on three main crops with demonstrations in large areas resulted in grain production increasing, and a heightened integration of regional and national grain production effort. The engineering focus on reducing the amount of agricultural chemicals used and saving resources relieved pressure on the environment. The engineering program improved the technological quality base for extension agents and farmers, who rapidly disseminated the practices and promoted its applications (Dong, 2008).

7.2. 2008 activity for establishing high yield of grain in China 7.2.1. Background At the end of 2007, agriculture production was facing several important problems, including increasing grain requirements, climatic change, frequent occurrence of natural calamity, relative low yield per unit area, and uneven regional distribution of natural resources. Most of China experienced a great drought in autumn and winter of 2007, in addition to a frost and snow disaster in early 2008 followed by a season of heavy pest pressure. The national level of yield per unit area was currently lower than the advanced levels in the developed countries. Rice, wheat, and soybean yields per unit area only accounted for 85% of the international advanced level, with maize and potato at less than 50%. Furthermore, there was a regional imbalance in grain production. For example, Shandong and Anhui Province are located in the same ecological zone with similar agricultural production environments, but in 2007, the average yield of wheat in Shandong was 810 kg/hm2, more than that of Anhui. There is thus much

142

Huixiao Wang et al.

potential for yield increases to bridge the gap between regions. The Ministry of Agriculture in China decided to ordain 2008 as the ‘‘activity year for establishing high yield of grain and oil,’’ organizing multiple activities aimed at increasing national grain yield. Within a framework of increasingly limited agricultural resources, Chinese agricultural production needed a way to increase yield per unit area relying on science and technology. 7.2.2. Objectives of the activity The aim of 2008 Activity for Establishing High Yield of Grain is to evenly increase grain yield in regions by integrating and spreading advanced and practical technology. Technology integration for establishing high yield is to unite the existing cultivation technology, including selecting good breeds, precision seeding of wheat, formula fertilization by soil testing, comprehensive pest control, and others. The activity would build 500 of demonstration areas of good quality and high yield in national main grain production regions. In the end wheat, single season rice, maize, and double-season rice’s yield would reach 9, 10.5, 12, and 13.5 ton/hm2, respectively. 7.2.3. Main contents of the activity The basic idea in this activity was to promote a unified consistent methodology for practices such as tillage, seeding, regulation of fertilizer and water, pest control, and harvesting by machine. Objectives included selecting appropriate varieties, demonstrating advanced practical technology, setting up cultivation system fine-tuned to different regions and different crops, monitoring and forecasting disease and insect pest to make prevention and control plan, applying fertilizer according to soil tests and crop requirements, and, finally, training farmers to fertilize in scientific way (Ministry of Agriculture, 2008). The main task of the activity was to create some 33–667 hm2 demonstration plots to showcase high-quality, high-yield varieties, with 150 sown in wheat. The activity organized experts to disseminate technology to farmers at key seasonal times and locations. 7.2.4. Achievements The 2008 Activity for Establishing High Yield of Grain resulted in very good harvests. The unit yield and total yield both broke historical records. The activity also accelerated the distribution of good varieties and dissemination of practical technology, significantly improving production capacity. According to the data of Ministry of Agriculture, many 667 hm2 demonstration areas of high yield emerged in the activity and there were 143 demonstration areas with an average wheat yield of 8100 kg/hm2, 3315 kg/hm2 higher than national average level. In addition, there were

Problems, Challenges, and Strategic Options of Grain Security in China

143

63 667 hm2 demonstration areas of early rice with an average yield of 7785 kg/hm2, 2250 kg/hm2 more than national average level. The further dissemination of new cultivation technology resulting in good quality and high yields would improve grain production and increase economic benefits from cultivation. It was estimated that the 602 demonstration areas promoted an increase in grain production of 14.5 million tons through the dissemination of information. The Bureau of Agriculture of Heilongjiang Province concluded that the average benefit per area of rice, maize, potato, and soybean in the demonstration areas were, respectively, 23.6%, 30.0%, 45.5%, and 79.2% higher than that in local farmland ( Jiang, 2008).

8. Concluding Remarks Given that China’s population is the largest in the world, grain security is not only a national interest, but is a global concern as well. Even small changes in China’s grain balance may seriously affect world grain trade. China is able to feed 22% of the world’s population on only 9% of the world’s cultivated lands. However, to achieve this, China uses over 10% of the world’s agricultural labor, 30% of the world’s fertilizer consumption, and 25% of the world’s irrigation water. China’s achievements can be attributed to policy, technology, and use of quality agricultural inputs. The central government of China prioritizes grain security and agriculture in national economic development, taking responsibility and a leading role in improving grain production, and solving food security problems, both nationally and globally. In the last decade, China has been self-sufficient in grain, producing over 95% of its domestic needs. Furthermore, China has predominantly been a net exporter of grain, not only meeting the demand of its own people, but also significantly contributing to global demand. Stable self-sufficiency has played an important role in shielding China from the effects of the global grain crisis. However in the long term, China’s grain security remains uncertain. Many challenges must be faced, including increasing demand as the population grows and the standard of living rises, setting within a framework of limited natural and agricultural resources, climate change and frequent natural disasters, slow growth in technology, and a smallscale agricultural economy. Thus, the tight balance between grain supply and demand will be a long-term issue in China. Four trends related to grain security are not expected to change in China: an increase in population, a decline in per capita land shares, an increase in total grain requirement, and the need to maintain self-sufficiency in grain production. Based on these trends, the following four issues appear

144

Huixiao Wang et al.

to be the major challenges for improving grain production capacity: improving economic benefits associated with growing grain, consolidating small parcels of land to achieve economy of scale production, maintaining a sufficient amount of land for cultivation in the face of urban encroachment, and dissemination of knowledge about new varieties and technology. To ensure long-term grain security, several measures should be instigated to improve grain production capacity. First, efforts should focus on protecting and improving the production capacity in the main grainproducing regions. The agricultural comprehensive development program, which concentrates on improving medium- and low-yield fields, should be continued. Advances in science and technology should be employed to improve grain yield per unit area. There is much potential for the application and extension of new techniques such as precision agriculture and gene engineering. Lasting policies must be implemented to ensure benefits to farmers for grain production, including insurance of grain price stability. The supply chain from the producer to the consumer needs to be reformed to improve efficiency. All levels of government must take responsibility for land protection, grain production, distribution, and security. China’s current success in feeding its population, in conjunction with a long-term proactive plan for meeting challenges, thus inspires confidence for the future domestic and national grain security.

ACKNOWLEDGMENT This chapter was supported by Specialized Research Fund for the Doctoral Program of Higher Education (SRFDP, 20060027018).

REFERENCES Blanke, A., Rozelle, S., Lohmar, B., Wang, J., and Huang, J. (2007). Water saving technology and saving water in China. Agric. Water Manage. 87, 139–150. Brown, L. (1995). ‘‘Who Will Feed China? Wake-Up Call for a Small Planet’’. Worldwatch Institute, Washington, DC. Brown, L. (1998). ‘‘China’s Water Shortage Could Shake World Food’’. Worldwatch Institute, Washington, DC. Bruins, H., and Bu, F. (2006). Food security in China and contingency planning: The significance of grain reserves. J. Contingencies Crisis Manage. 14(3), 114–124. Chen, L. (2008). Impact of global climate change on China’s agriculture. Seed Technol. 4, 38–40. Deng, Y. (2008). Low grain can not ensure grain security. Study Times, July 23. Department of Regulation. (2004). State administration of grain consideration on China’s grain security. Macroeconomics 9, 6–9. Department of Rural Social Economy Survey, National Bureau of Statistics of China (2005). Study on grain security evaluating indicator system in China. Stat. Res. b(8), 3–9.

Problems, Challenges, and Strategic Options of Grain Security in China

145

Dong, W. (2008). Applying grain high yield science and technology engineering, guarantee China’s grain security. Crops 4, 23–25. ECCLY. (1997). “China Land Yearbook”. China Land Press, Beijing. Fan, S. (2000). Research investment and the economic returns to Chinese agricultural research. J. Prod. Anal. 14(92), 163–180. Fang, F., Qian, Y., and Liu, S. (2004). An empirical analysis on China’s agricultural. J. Finance Econ. 30(2), 89–96. FAO. (1983). In ‘‘World Food Security: A Reappraisal of the Concepts and Approaches.’’ Director General’s Report, Rome. FAO. (1996). In ‘‘Rome Declaration on World Food Security and World Food Summit Plan of Action.’’ World Food Summit, Rome, 13–17 November 1996. FAO. (2008a). In ‘‘Crop prospects and food situation. Economic and Social Department’’, FAO, No. 2, Rome, April 2008. FAO. (2008b). ‘‘Declaration of the High-Level Conference on World Food Security: The Challenges of Climate Change and Bioenergy,’’ Food and Agricultural Organization of the United Nations, Rome, 5 June. Feng, Z. (2007). Future food security and cultivated land guarantee for population development in China. Popul. Res. 31(2), 15–29. Feng, Z., Liu, B., and Yang, Y. (2005). A study of the changing trend of Chinese cultivated land amount and data reconstructing: 1949–2003. J. Nat. Resour. 20(1), 35–43. Guo, S. (1990). ‘‘China’s Grain: Study and Thinking from Multi Angles,’’ Chinese Economic Press, Beijing, pp. 14. Guo, M., Xie, L., Cao, M., and Liu, J. (2008). Impact of measurement of climate change on agricultural production and country development. Agric. Econ. 10, 8–10. Heilig, G. K. (1999). ‘‘China Food: Can China Feed Itself?’’ IIASA, Laxenburg; CD-ROM Vers. 1.1. Ivanic, M, and Martin, W (2008). Implications of Higher Global Food Prices for Poverty in Low-Income Countries. Policy Research Working Paper 4594, The Development Research Group Trade Team of the World Bank, April 2008. Jiang, N. (2008). Establishing high yield drive unit yield and total yield exceeding history. Farmer Daily, 9 December. Jury, W., and Vaux, H. (2007). The emerging global water crisis: Managing scarcity and conflict between water users. Adv. Agron. 95, 1–76. Kong, L., Hu, Y., Wang, G., and Li, J. (2006). Improve comprehensive grain yield capability of Anhui province depending on advancement of science and technology. Manage. Agric. Sci. Technol. 25(5), 24–26. Lal, R. (2007). Anthropogenic influences on world soil and implications to global food security. Adv. Agron. 93, 69–93. Lan, S., Yan, X., Xu, S., and Li, W. (2006). Study of grain storage in rural area of China. Grain Storage 4, 54–56. Li, Y., Jiang, N., and Guo, Z. (2001). ‘‘Theory about Fluctuation of China’s Grain,’’ Chinese Agricultural Press, Beijing, pp. 76. Li, M., Li, Z., Wang, D., Yang, X., Zhong, X., Li, Z., and Li, Y. (2005). Impact of natural disaster change on grain yield in China in the past 50 years. J. Nat. Disasters 14(2), 55–60. Lin, Y., and Yu, L. (2006). A study on the fluctuation of China’s grain yield. J. Jilin Agric. Univ. 28(3), 345–351. Long, F. (2007a). ‘‘Study on China’s Grain Security in New Century,’’ Chinese Economic Press, Beijing, pp. 116. Long, F (2007b). Research on China Food Security of the New Century; Thesis for a Ph.D. Degree of Hunan Agriculture University. Lou, Y. (2003). Macroscopical analysis and compared study on China’s grain security. Agric. Econ. Study 3, 30–32.

146

Huixiao Wang et al.

Lu, F. (1999). Grain surplus of three times (1984–1998)—Alternative explanation of policy evolution of China’s grain circulation. Studying Data Econ. Worker 60, 33–42. Mao, H. (2005). ‘‘Study on China’s Grain in New Stages,’’ Chinese Economic Press, Beijing. Ministry of Agriculture. (2008). ‘‘Working Planning of the Year of Establishing High Yield of Grain in China’’. Ministry of Environmental Protection of the People’s Republic of China. (2008). Bull. Environ. Situations; 4 June. Ministry of Land and Resources. (2003). ‘‘Land and Resources Bulletin (1997–2003).’’ China Statistical Press, Beijing. National Development and Reform Commission of China. (2007). In ‘‘China’s National Climate Change Programme.’’ Prepared under the Auspices of National Development and Reform Commission People’s Republic of China, Printed in June. National Development and Reform Committee. (2008). In ‘‘National Mid Long Term Planning Outline of Grain Security (2008–2020).’’ Peng, X. (1987). Demographic consequences of the Great Leap Forward in China’s provinces. Popul. Dev. Rev. 13(4), 639–670. People’s Daily. (2008). “Explaining Grain Developing Trend in China by numbers”. December 28. Ren, J. (2008). To develop grass agriculture: New ideas on grain security in China. China Animal Science and Veterinary Daily; July 15. Shao, L. (2005). Study on China’s grain comprehensive and productivity and grain safety. Chin. J. Agric. Resour. Regional Planning 26(1), 23–26. Shi, Y. (2008). Thinking for solving two big issues of petroleum and three rural problems (Farm, Farming & Farmer) in China. Science and Technology Daily; August 8. State Council. (1996). The white paper on the grain issue in China. State Council Information Office of China. (1996). The white paper on food security in China. State Council Information Office of China. (2000). In ‘‘Population and Its Development in China in the 21st Century’’, Beijing, December 2000. State Council of China. (2008). In ‘‘A Medium to Long-Term Plan for National Grain Security’’, July 2008. Sun, Z. (2008). Not allowing ignore: Agriculture and lasting basic business. Chin. Inf. Daily; July 18. Sun, D., Li, H., Dawson, R., Tang, C., and Li, X. (2006). Characteristics of steep cultivated land and the impact of the Grain-for-Green policy in China. Pedosphere 16(2), 215–223. Tso, T. (2004). Agriculture of the future. Nature 428, 215–217. United Nations. (1975). In ‘‘Report of the World Food Conference.’’ The World Food Conference, Rome, 5–16 November, 1974 (New York). Verburg, P., Chen, Y., and Veldkamp, T. (2000). Spatial explorations of land use change and grain production in China. Agric. Ecosyst. Environ. 82, 333–354. Wang, D. (2005). Insuring grain security urgently need land scaled management. Study Times; May 16. Wang, H. (2004). Water being an important influence factor of sustainable development. Sci. Times; November 8. Wang, J. (2008a). Sustainable utilization of water and land resources is the key of grain security. World Environ. 4, 45–46. Wang, Y. (2008b). Potential crisis analysis of China’s food security. Res. Agric. Mod. 29(5), 537–541. Wang, Y. (2008c). Potential crisis of China’s grain security. Res. Agric. Mod. 29(5), 537–540. Wang, J., and Lu, W. (2006). ‘‘Analysis and Solution to China’s Food Security Problem from Marketization and Globalization Perspective.’’ Zhejiang University Press, Hangzhou.

Problems, Challenges, and Strategic Options of Grain Security in China

147

World Bank. (1986). ‘‘Poverty and Hunger: Issues and Options for Food Security in Developing Countries.’’ The World Bank, Washington, DC. Wu, L., Shan, L., Shan, H., and Cui, C. (2007). Current situation and prediction of grain security in China. Agro Food Industry Hi-Tech 18(4), 16–17. Xiong, W., Lin, E., Ju, H., and Xu, Y. (2007). Climate change and critical thresholds in China’s food security. Climatic Change 81, 205–221. Xu, J., Katsigris, E., White, T.A., (Eds.) (2002). In ‘‘Implementing the Natural Forest Protection Program and the Sloping Land Conversion Program: Lessons and Policy Implications’’ CCICED-WCFGTF, China Forestry Publishing House, Beijing. Xu, Z., Xu, J., Deng, X., Huang, J., Uchida, E., and Rozelle, S. (2006). Grain for Green versus grain: Conflict between food security and conservation set-aside in China. World Dev. 34(1), 130–148. Yang, H., and Li, X. (2000). Cultivated land and food supply in China. Land Use Policy 17, 73–88. Yu, Z., and Hu, X. (2003). Research on the relation of food security and cultivated land’s quantity and quality in China. Geogr. Geo-Inform. Sci. 19(3), 45–49. Zhang, S. (2005). Cultivated land resources and grain security; Thesis for a Ph.D. Degree of Chinese Academy of Agricultural Sciences. Zhang, X. (2008). In ‘‘Focusing economic hotspot: True and false of grain security, Interview and Talk with Experts’’. CCTV, 24 July. Zhang, Z., and Duan, Z. (2008). China not only feeding itself but also improving world grain security, Viewpoint of Scholars. http://www.cas.cn; 3 June. Zhang, Y., and Zhao, G. (2007). Evaluation of effect on science and technology project of food production in Hebei Province. Chin. Agric. Sci. Bull. 23(2), 465–468. Zhu, Z. (1998). ‘‘China Grain Security Issues: Case Study and Policy Options.’’ Hunan Science and Technology Press, Beijing. Zhu, J. (2004). Public investment and China’s long-term food security under WTO. Food Policy 29, 99–111. Zhu, Y., Huang, X., Peng, Y., Lv, X., and Ma, Y. (2007). Another thinking of farmland protection and grain security. Natl. Land Resour. 6, 28–33. Zhu, S., Herzfeld, T., Glauben, T., Zhang, Y., and Hu, B. (2008). Factors affecting Chinese farmers’ decisions to adopt a water-saving technology. Can. J. Agric. Econ. 56, 51–61.

C H A P T E R

F O U R

Weed Management in Rice-Based Cropping Systems in Africa J. Rodenburg* and D. E. Johnson† Contents 150 150 153 155 155 163

1. Introduction 1.1. Rice in Africa 1.2. Importance of weeds 2. Weed Species in Rice in Africa 2.1. Major problem weeds 2.2. The usefulness of weeds 3. Weed Management Practices in African Rice-Based Cropping Systems 3.1. Cultural weed control 3.2. Mechanical weed control 3.3. Rice varietal development for improved weed control 3.4. Biological weed control 3.5. Chemical weed control 3.6. Integrated weed control 4. Emerging Weed Problems and Weed Management Issues 4.1. Likely effects of demography on weed management 4.2. Effect of changing climates on weed management 4.3. Herbicide resistance in weeds 5. A Strategic Vision for Weed Management and Research in African Rice Production Systems 5.1. Weed management strategy 5.2. Weed research strategy 6. Concluding Remarks Acknowledgments References

* {

165 165 172 174 178 179 185 186 186 186 188 189 189 193 200 200 201

Africa Rice Center (WARDA), Dar es Salaam, Tanzania Crop, Soil and Water Sciences Division, International Rice Research Institute (IRRI), Metro Manila, Philippines

Advances in Agronomy, Volume 103 ISSN 0065-2113, DOI: 10.1016/S0065-2113(09)03004-1

#

2009 Elsevier Inc. All rights reserved.

149

150

J. Rodenburg and D. E. Johnson

Abstract Weed competition is a major constraint in all the rice production systems in Africa. In addition to the costs of weed control, weeds account for yield losses estimated to be at least 2.2 million tons per year in sub-Saharan Africa, valued at $1.45 billion, and equating to approximately half the current total imports of rice to this region. Important weeds in upland rice include the perennial species Cyperus rotundus, Imperata cylindrica and Chromolaena odorata, the annual species Euphorbia heterophylla, Digitaria horizontalis, and the parasitic weeds Striga spp. In lowland rice the perennial weeds: Cyperus rotundus, C. esculentus and Oryza longistaminata and annual weeds Sphenoclea zeylanica, Echinochloa spp., Cyperus difformis, C. iria, Fimbristylis littoralis, Ischaemum rugosum, and O. barthii cause serious losses. Common weed management practices in rice-based cropping systems include soil tillage, clearance by fire, hand- or hoe-weeding, herbicides, flooding, fallow and crop rotations, and these are often used in combination. Labor shortages and lack of access to information, inputs, and credits are widespread constraints for African farmers. To optimize financial, social and environmental costs and benefits, integrated and ecological management approaches are advocated. Locally adapted and affordable combinations of preventive measures and interventions should be targeted. Future weed research should aim to deliver the information and tools for the implementation of these approaches. This requires the generation of knowledge on weed biology and ecology and on the consequences of changes in management and the environment on weed populations. To address the diversity of rice-based cropping systems in Africa, priorities need to be set and products and information delivered that take full account of local conditions. This will require farmer participatory approaches that are inclusive with respect to resource-poor farmers and gender.

1. Introduction 1.1. Rice in Africa Rice is the fifth most important cereal in Africa in terms of area harvested and the fourth in terms of production (FAO, 2008b). Rice production in Africa is increasing at the fastest rate of any cereal, and over the past three decades, harvested area has risen by 105% and production by 170% (Table 1). With respect to production and area statistics, important rice producing countries in Africa are Nigeria, Madagascar, Guinea, Sierra Leone, Egypt, Congo DR, Mali, Coˆte d’Ivoire, Tanzania, and Mozambique (Table 2). The majority of rice is consumed in West Africa, although the region is not self-sufficient in rice, and increasing import costs are a concern. Rice-based cropping systems are diverse and vary among subregions, ecosystems, management input levels, farm scales, and traditional practices. Topography and hydrology are among the most important variables

151

Weed Management in Rice-Based Cropping Systems in Africa

Table 1 The five major cereals grown in Africa (data 2006) in terms of harvested area (1000 ha), area under rice compared to total area under cereals (Area Share; %), production (1000 t), rice production compared to total cereal production (Production Share; %), and increases in area (D Area; %) and production (D Production; %) between 1976 and 2006

Cereal

Area

Maize Sorghum Millet Wheat Rice All

26,118 25,137 20,196 10,175 8825 98,746

Area share

26 25 20 10 9

Production

46,260 26,113 17,788 25,096 21,131 145,892

Production share

D Area

D Production

32 18 12 17 14

36 64 54 11 105 45

74 125 120 144 170 107

Source: FAOSTAT (2008).

determining differences in rice-based cropping systems in Africa. Five main rice ecosystems can be distinguished based on water supply and topography (Windmeijer et al., 1994): (1) rain-fed upland rice on plateaus and hydromorphic slopes, (2) lowland rain-fed rice in valley bottoms and floodplains, (3) irrigated rice in deltas and floodplains, (4) deep-water floating rice along major rivers, and (5) mangrove-swamp rice in lagoons and deltas. Upland rice ecosystems roughly represent 39% of the total area under rice in SSA leaving 33% to rain-fed lowlands, 19% to irrigated lowlands and around 9% to deep water and mangroves (Balasubramanian et al., 2007; updated with data from FAO, 2008b). Upland, rain-fed and irrigated lowland rice systems are widely distributed while deep water and mangrove systems are of only local importance such as the Niger River flood plains of Guinea, Mali, and Nigeria (deep water) or coastal zones of Sierra Leone, Liberia, and The Gambia (mangroves). Further distinction of the rice production systems can be based on the agroecological zones that differ in the length of the growing season (i.e., Sahel and Guinea savanna, derived savanna, and the humid forest zones). Upland rice cropping systems are in the forest, savanna, and derived savanna zones. If the rainy season is of sufficient duration or if residual moisture is adequate after rice, subsequent crops like maize, cowpea, or soybean may be grown. In East African highlands, upland rice is rotated with wheat, maize, or potato. On hydromorphic areas, where the perched water table is within 50 cm of the soil surface for the majority of the growing season, rice and cash crops such as cotton are grown. In upland cropping systems of subsistence farmers, input levels are generally low, and low yields (mean: <1 t ha 1; range: 0.1–3.5 t ha 1) are commonly due to poor soil fertility and weed competition (Balasubramanian et al., 2007;

Table 2 Major rice countries in Africa (2006 data) in terms of harvested area (1000 ha), area under rice compared to total area under cereals (Area Share; %), production (1000 t), rice production compared to total cereal production (Production Share, %), self-sufficiency (SS), and ranking (R) among 51 African countries (only top 10 countries by rice area shown and in decreasing order of importance)

Country

Area

R

Area share

R

Production

R

Nigeria Madagascar Guinea Sierra Leone Egypt Congo DR Mali Coˆte d’Ivoire Tanzania Mozambique

2725 1250 758 730 613 418 401 370 355 180

1 2 3 4 5 6 7 8 9 10

14 94 89 83 20 16 12 47 10 9

12 2 3 4 10 11 13 5 14 17

3924 3485 1340 1062 6500 316 1019 700 784 174

2 3 4 5 1 9 6 8 7 13

a ( ) Production < consumption; (+) production > consumption; (?) no consumption statistics available. Source: FAOSTAT (2008)

SS a

+ + + + + ? +

Production share

R

14 92 55 92 29 21 30 50 15 10

18 3 5 2 10 12 9 6 16 21

Weed Management in Rice-Based Cropping Systems in Africa

153

Windmeijer and Andriesse, 1993). In rain-fed lowlands in the savanna and forest zones, fields are often unbunded and poorly leveled. Rice is grown during the wet season and land is often left fallow during the dry season. If the rainy season is long enough (>5 months) or if residual soil moisture in the valley bottom suffice, farmers use the dry season to grow groundnut, soybean, maize, or vegetables (Kent et al., 2001; Windmeijer and Andriesse, 1993). In irrigated lowlands, wet season rice is often followed by a second rice crop, or vegetables (Kent et al., 2001), or left fallow in the dry season. Some irrigated areas in the Sahel are double cropped with rice usually followed by a short fallow period (Defoer et al., 2004b). Double cropping is also practiced in the irrigated highlands of Central and East Africa (including Madagascar) either as rice–rice or rotated with vegetables, wheat, potatoes, or soybean (Balasubramanian et al., 2007).

1.2. Importance of weeds Throughout Africa, from Senegal to Madagascar, weeds are cited among the main production constraints in any of the rice ecosystems (Adesina et al., 1994; e.g., Ampong-Nyarko, 1996; Becker and Johnson, 1999a; Diallo and Johnson, 1997). Common agronomic factors that contribute to weed problems are inadequate land preparation (soil tillage, soil leveling in lowland areas), rice seed contamination with weed seeds, use of poor quality rice seeds, broadcast seeding in lowlands, use of old rice seedlings for transplanting, inadequate water management, inadequate fertilizer management, mono-cropping, labor shortages for hand weeding and delayed herbicide applications and other interventions (Becker and Johnson, 1999a, 2001b; Diallo and Johnson, 1997). In the upland systems, crop intensification and inadequate fallow management are also contributory factors (Becker and Johnson, 2001b). Worldwide, weeds are estimated to account for 32% potential and 9% actual yield losses in rice (Oerke and Dehne, 2004). The nature and severity of weed problems, however, vary according to the rice ecosystem. Likewise, weed management practices and the available options are often a function of biophysical and socioeconomic factors which, in turn, are determined by the agroecosystem. Weeds are the major constraints in rainfed uplands and in the unbunded lowlands, for instance, where they cannot be controlled by flooding the soil surface. Similarly, rice in the deep-water rice systems along the major rivers can be severely affected by weeds prior to flooding as the crop is direct-seeded and farmers rely on hand weeding and use relatively little herbicides (Akobundu, 1987; Ampong-Nyarko and De Datta, 1991). In irrigated production systems where rice is directseeded, weeds are the major yield constraints (Becker et al., 2003; Diallo and Johnson, 1997). Uncontrolled weed growth is reported to cause yield losses in the range of 28–74% in transplanted lowland rice, 28–89%

154

J. Rodenburg and D. E. Johnson

in direct-seeded lowland rice and 48–100% in upland ecosystems in West Africa (Akobundu, 1980; Diallo and Johnson, 1997; Enyinnia, 1992; Imeokparia, 1994; Johnson et al., 2004). In irrigated rice in West Africa, from the forest zone to the Sahel, poor weed management by farmers was estimated to be responsible for a yield reduction by at least 1 t ha 1 and it was demonstrated that better weed control by farmers could raise yields by 15% (Becker et al., 2003; Haefele et al., 2000). In areas of rain-fed lowland rice, without bunds, yields could be increased by 23% through improved weed control, while in the most widespread upland rice systems, yields could be raised by 16% (Becker and Johnson, 2001a,b). These estimates indicate that in sub-Saharan Africa weeds account for rice yield losses of at least 2.2 million tons per year at a value of $1.45 billion (Table 3), in addition to the costs of weed control. These estimated losses equate approximately to half the current imports of rice to the region.1 This chapter discusses the major weeds of African rice ecosystems and the various weed management options available to farmers. The objectives of this chapter are to provide an overview and source of reference with respect to weed problems and weed management, to assist with the identification of knowledge gaps and to contribute to the formulation of strategies for research to improve weed management options for rice farmers.

Table 3 Potential annual rice import savings from improved weed management in sub-Saharan Africa, 2008

Rice area (‘000 ha)a Rice area share SSA (%) Average yield (t ha 1)b Annual production (103 t) Weed-inflicted yield loss (%)c Annual production loss (103 t) Potential annual import savings (M $)d a b c d

1

Irrigated lowland

Rain-fed lowland

Rain-fed upland

1559 19 3.62 5648 15

2707 33 1.22 3316 23

3118 38 1.28 3998 16

842

756

648

543

488

418

Total area under rice in SSA is estimated at 8.2 million ha (FAO, 2008b). Calculated from Balasubramanian et al. (2007) and updated with FAO (2008b). Based on Becker and Johnson (2001a,b) and Becker et al. (2003). Based on a world rice price (Thai 25%) of $645 t 1 in September 2008 (FAO, 2008a).

FAO Rice Market Monitor (2008a); estimated annual rice imports in 2007 in SSA: 2.7 to 3.0 billion USD.

Weed Management in Rice-Based Cropping Systems in Africa

155

2. Weed Species in Rice in Africa 2.1. Major problem weeds Each rice production system harbors weed species well adapted to the environment and management practices. While the weed flora of a specific production system (e.g., lowland or upland) may be similar across different agroecological zones, the abundance of individual species can differ substantially (Akobundu and Fagade, 1978). A review of the literature on weeds in rice-based cropping systems in Africa yielded 130 different weed species (upland: 61; hydromorphic: 31; lowland: 74), 57 of which were reported more than once (upland: 26; hydromorphic: 13; lowland: 30), and 12 were observed in more than one rice ecosystem. These 57 species are listed in Table 4. Most cited weed species of upland areas were Rottboellia cochinchinensis (Lour.) W. Clayton, Digitaria horizontalis Willd., Ageratum conyzoides L., and Tridax procumbens L., while A. conyzoides and Panicum laxum Sw. were most cited in the hydromorphic areas and Cyperus difformis L., Sphenoclea zeylanica Gaertner, Fimbristylis littoralis Gaudich, Oryza longistaminata A. Chev. & Roehr., Echinochloa colona (L.) Link and E. crus-pavonis (Kunth) Schultes were the most cited weeds of lowland rice. Gramineae (43%) and Cyperaceae (37%) were the most prevalent weeds of lowland rice while, in the uplands, weed species composition tended to be more diverse with Gramineae (36%) and Compositae (16%) most prevalent. Weed populations of upland rice are reported to be more dynamic than those of lowland rice areas (Johnson and Kent, 2002). Perennial species accounted for more than 45% of the weed species of lowland rice and only 31% in the upland or hydromorphic rice ecosystems (Table 4). Commonly only a few weed species dominate the population in each rice ecosystem ( Johnson and Kent, 2002) and consequently only a relative small proportion of the species found in rice are considered problem weeds (Diallo and Johnson, 1997). Characteristics that distinguish such species are high competitiveness, high multiplication rates, similarity in appearance with rice, and, in the lowland systems, submergence tolerance. Some problem weeds in rice are annuals with short growth cycles such as C. difformis and D. horizontalis (40–80 days) and are able to reproduce before rice harvest even when they emerge after the first weeding operation ( Johnson, 1997). Such species, if not controlled, are able to build up populations very rapidly. Annual weeds causing problems in upland rice production are Euphorbia heterophylla (L.), D. horizontalis and the parasitic weeds Striga spp. (S. hermonthica [Del.] Benth. and S. asiatica [L.] Kuntze). Perennials Cyperus rotundus and C. esculentus as well as the annual A. conyzoides are frequently encountered on hydromorphic areas. Perennial weeds of rice in the upland forest and derived upland savanna zones tend to be those that are able to reestablish rapidly after disturbance and

156

J. Rodenburg and D. E. Johnson

Table 4 Weed species in rice production ecosystems in Africa: species (in decreasing order of citation) names, family, biology, and utility Species

Upland Rottboellia cochinchinensis (Lour.) W. Clayton syn. R. exaltata Digitaria horizontalis Willd. Ageratum conyzoides L. Tridax procumbens L. Eleusine indica (L.) Gaertner Euphorbia heterophylla (L.) syn. E. geniculata Ortega Imperata cylindrica (L.) Raeuschel Paspalum scrobiculatum L. Mariscus cylindristachyus Steudal Trianthema portulacastrum L. Striga hermonthica (Del.) Benth. Striga asiatica (L.) Kuntze Cynodon dactylon (L.) Pers. Amaranthus viridis L. Euphorbia hirta (L.) syn. E. pilulifera (L.); Chamaesyce hirta (L.) Millsp. Commelina benghalensis L. Brachiaria lata (Schum.) C.E. Hubb. Dactyloctenium aegyptium (L.) Willd. Cyperus rotundus L. Chromolaena odorata (L.) King & Robinson syn. Eupatorium odoratum L. Panicum laxum Sw. Calopogonium mucunoides Desv. Aspilia bussei O. Hoffm. & Muschler Pennisetum purpureum Schum. Boerhavia erecta L. Striga aspera (Willd.) Benth. Hydromorphic Ageratum conyzoides L. Panicum laxum Sw. Leersia hexandra Sw. Cyperus rotundus L. Digitaria horizontalis Willd. Eclipta prostrata (L.) L. syn. E. alba (L.) Hassk. Spilanthes uliginosa Sw. syn. S. acmella A. Chev.

Fam.

Biologya

Useb

GRAM

A;C4

GRAM COMP COMP GRAM EUPH

A A A A;C4 A

GRAM GRAM CYPE AIZO OROBc OROB GRAM AMAR EUPH

P;C4 P P A;C4 A/ohp A/ohp P;C4 A;C4 A;C4

M AF

COMM GRAM GRAM CYPE COMP

A A A;C4 P;C4 P

M

GRAM LEGU COMPd GRAMe NYCT OROB

A P A A;C4 P;C4 A/ohp

COMP GRAM GRAM CYPE GRAM COMP

A A P P;C4 A A

M

COMP

A

M

M, CP M F

F/M

M M, CP

L

M

x

157

Weed Management in Rice-Based Cropping Systems in Africa

Table 4 (continued) Species

Fam.

Biologya

Useb

Commelina benghalensis L. Fimbristylis littoralis Gaudich. syn. F. miliacea Vahl Echinochloa colona (L.) Link syn. E. colonum (L.) Link; Panicum colonum L. Cyperus esculentus L. Cynodon dactylon (L.) Pers. Rhamphicarpa fistulosa (Hochst.) Benth. Lowland Sphenoclea zeylanica Gaertner Cyperus difformis L. Fimbristylis littoralis Gaudich. syn. F. miliacea Vahl Oryza longistaminata A. Chev. & Roehr. Echinochloa colona (L.) Link syn. E. colonum (L.) Link; Panicum colonum L. Echinochloa crus-pavonis (Kunth) Schultes syn. E. rostrata (Stapf) Michael Leersia hexandra Sw. Oryza barthii A. Chev. syn. O. breviligulata Cyperus iria L. Bolboschoenus maritimus (L.) Palla syn. Scirpus maritimus L., Schoenoplectus maritimus (L.) Lye Ischaemum rugosum Salisb. Panicum laxum Sw. Ludwigia abyssinica A. Rich. syn. Jussiaea abyssinica (A. Rich.) Dandy & Brenan Ammania prieureana Guill. & Perr. Heteranthera callifolia Rchb. ex Kunth Ipomea aquatica Forssk. syn. I. reptans Poiret Echinochloa pyramidalis Hitch & Chase

COMM CYPE

A A;(C4)

M

GRAM

A;C4

AF/ F/T

CYPE GRAM OROB

P;C4 P;C4 A/fhp

SPHE CYPE CYPE

A A A;(C4)

GRAM GRAM

P A;C4

GRAM

A

GRAM GRAM CYPE CYPE

P A A;C4 P;(C4)

GRAM GRAM ONAG

A A A

LYTH PONT CONV

A A P

GRAM

P

CYPE CYPE GRAM

P;C4 P P

GRAM GRAM GRAM

A P P;C4

Cyperus esculentus L. Cyperus halpan L. syn. C. haspan L. Sacciolepis africana C. E. Hubb. & Snowden Acroceras amplectans Stapf Diplachne fusca (L.) P. Beauv. ex Stapf Panicum repens L.

M

AF/ F/T

I

M

AF/ F/T F

(continued)

158

J. Rodenburg and D. E. Johnson

Table 4 (continued) Species

Fam.

Biologya

Eleocharis spp. (E. complanata Boeck; E. acutangula (Roxb.) Schultes; E. mutata (L.) Roemer & Schultes; E. dulcis (Burm. f.) Henschel) Fimbristylis ferruginea (L.) Vahl Pycreus macrostachyos (Lam.) Raynal syn. P. tremulus (Poiret) C. B. Clarke; P. albomarginatus Nees Schoenoplectus senegalensis (Steudel) Raynal syn. Scirpus jacobii C. Fischer Ludwigia adscendens (L.) Hara syn. Jussiaea repens L. Eclipta prostrata (L.) L. syn. E. alba (L.) Hassk. Rhynchospora corymbosa (L.) Britton syn. R. aurea Vahl; Scirpus corymbosus L.

CYPE

A/P

CYPE CYPE

P;(C4) A

CYPE

A;(C4)

ONAG

P

COMP

A

CYPE

P;(C4)

Useb

a

A, annual; P, perennial; fhp, facultative hemiparasitic; ohp, obligate hemiparasitic; C4, C4 photosynthetic pathway; (C4), uncertainty about photosynthesis pathways; some species of the genus are C3 some are C4. b AF, fodder (animal feed); CP, crop protection (bio-pesticides); F, food; I, insecticide (mosquito control); L, legume (green manure/improved fallow); M, medicinal; T, Thatching (roof material). c Formerly: Scrophulariaceae. d Papilionoideae. e Poaceae. Source: Akobundu and Fagade (1978), Ampong-Nyarko (1996), Becker and Johnson (1998, 1999a, 2001b), Buddenhagen and Bidaux (1978), Burkill (2004), Diallo and Johnson (1997), Dzomeku et al. (2007), Elliot et al. (1993), Elmore and Paul (1983), Haefele et al. (2000), Harahap et al. (1993), Hillocks (1998), Hong et al. (2004), Johnson (1997), Johnson and Kent (2002), Johnson et al. (1997, 1998a, 2004), Kent et al. (2001), Kent and Johnson (2001), Mallamaire (1949), Nyoka (1982), Okafor (1986), Parkinson (1989), Reneaud (1980), Schwartz et al. (1998), and Xuan et al. (2004).

these include C. rotundus L., C. esculentus L., Imperata cylindrica (L.) Raeuschel, and Chromolaena odorata (L.) King & Robinson. Parasitic species Striga aspera (Willd.) Benth. and Rhamphicarpa fistulosa (Hochst.) Benth. are locally important annual weed species in rice in hydromorphic areas. In forest and savanna lowlands in Africa S. zeylanica, E. colona, C. difformis, C. iria, and F. littoralis are important annual weeds (Kent et al., 2001). In irrigated rice in the Sahel, important grass weeds are E. colona and Ischaemum rugosum and the wild rice species O. longistaminata and O. barthii (Diallo and Johnson, 1997). Notable sedges in irrigated rice in the Sahel are the annuals C. difformis, C. iria and Pycreus macrostachyos and the perennial Cyperus haspan (Diallo, 1999). Some of these species are discussed in more detail below.

Weed Management in Rice-Based Cropping Systems in Africa

159

2.1.1. Problem weeds of uplands and hydromorphic zones a. Cyperus spp. In moist to hydromorphic upland areas some of the most intractable weed problems in rice are due to the perennial sedges C. rotundus L. (Purple nutsedge) and C. esculentus L. (Yellow nutsedge). Tubers and seeds can remain dormant to survive periodic flooding or dry seasons. These species are able to multiply rapidly through tubers which can be greatly accelerated by soil tillage (Holm et al., 1991). The tubers can grow from soil depths of more than 0.5 m ( Johnson, 1997). Biomass of roots, tubers, and rhizomes of C. rotundus can be up to 40 t ha 1 (Holm et al., 1991). The abovementioned characteristics make these species typical weeds of intensely cultivated lands and very difficult to control. b. Imperata cylindrica. The perennial grass I. cylindrica (L.) Raeuschel (Speargrass) is a common and persistent weed in many upland crops like cassava, maize, sorghum, and rice. For more than 50% of the farmers surveyed by Chikoye et al. (1999) in West Africa, I. cylindrica was the most important weed. It reproduces through seeds and rhizomes. The species is particularly difficult to control as it is tolerant to fires and shallow cultivation due to the extensive underground network of rhizomes. The weed tends to be abundant where fields are regularly cultivated and burnt, as it recovers rapidly from disturbance, and burning induces flowering. It exerts great competition on crops (Chikoye et al., 2000; Johnson, 1997). The grass is common in the forest to savanna transition zone (Chikoye et al., 1999) and is widely adapted (Townson, 1991) but growth is suppressed by shade. c. Chromolaena odorata. The perennial woody shrub C. odorata (L.) R. King & H. Robinson (Siam weed) of the Compositae (Asteraceae) family produces large quantities of seeds and is capable of rapid regrowth after being cut ( Johnson, 1997). It is a common and dominant species in the upland fields and fallow vegetations of the forest zone and reported by many as a troublesome weed (e.g., Anthofer and Kroschel, 2007; Becker and Johnson, 2001a; Johnson, 1997; Kent et al., 2001). C. odorata gradually replaces indigenous species in fallow vegetation (Weise, 1995) and this in turn has consequences for the weed community in subsequent crops as it causes a predominance of broad-leaved species such as A. conyzoides, T. procumbens, and Phyllanthus amarus (Ikuenobe and Anoliefo, 2003). d. Digitaria horizontalis. The annual grass species D. horizontalis Willd. ( Jamaican crabgrass) is wide spread in upland and hydromorphic areas in the savanna and forest zones of Africa ( Johnson, 1997; Mallamaire, 1949). It is capable of rapid growth and has become a dominant species in intensively cultivated fields ( Johnson, 1997). D. horizontalis was observed to replace I. cylindrica over 5 years of rice cropping following fallow in Coˆte d’Ivoire ( Johnson and Kent, 2002). e. Euphorbia heterophylla. The annual species E. heterophylla L. (Mexican fireplant) of the Euphorbiaceae family is a common and very competitive weed of upland rice in the savanna zones of Africa. It can rapidly form a

160

J. Rodenburg and D. E. Johnson

closed canopy, and it has a life cycle of only about 60 days from germination to seed setting contributing to a rapid buildup of the population. Seeds of E. heterophylla are dispersed explosively through its dehiscent seed capsules (Wilson, 1981). Germination occurs throughout the cropping season due to the variable dormancy of the seeds. E. heterophylla is particularly problematic in mechanized cropping systems as contamination of fields frequently occurs through machinery; other sources of infestation are seed supply and wild animals ( Johnson, 1997). E. heterophylla was one of the species that was observed to increase with duration of rice cropping after fallow in Coˆte d’Ivoire ( Johnson and Kent, 2002). f. Ageratum conyzoides. The annual species A. conyzoides L. (Billy goat weed or Tropical white weed) is a member of the Compositae family. The species is widespread in moist uplands, hydromorphic and temporary, shallow flooded lands ( Johnson, 1997). The tolerance to temporary flooding, abundant seed production, and rapid germination of this species makes it a successful weed in rain-fed rice cropping systems in Africa. A. conyzoides has been reported to have medicinal and bioherbicidal applications (Xuan et al., 2004). Such uses however are not widespread. g. Striga spp. Parasitic weeds of the family Orobanchaceae (formerly: Scrophulariaceae) are locally important biotic constraints to upland rice production in the savanna zone of Africa, and these include Striga hermonthica (Del.) Benth. (Purple or Giant witchweed), S. asiatica (L.) Kuntze (Asiatic or Red witchweed), and S. aspera (Willd.) Benth. The first two are almost entirely found in free draining uplands, while the latter is also found on hydromorphic areas (e.g., Ampong-Nyarko, 1996; Buddenhagen and Bidaux, 1978; Johnson, 1997). In West Africa, S. hermonthica and S. aspera are the most important Striga species in rice (e.g., Dugje et al., 2006; Johnson et al., 1997; Parkinson, 1989), while S. hermonthica and S. asiatica are the dominant species in East African countries like Tanzania, Kenya, and Madagascar (e.g., Elliot et al., 1993; Fujisaka, 1990; Harahap et al., 1993; Mbwaga, 1996; Reneaud, 1980). Striga spp. (witchweeds) are annual, obligate hemiparasitic weeds on tropical cereal crops like maize, sorghum, and rice. Despite the presence of chlorophyll and photosynthetically active leaves (Press et al., 1991), these parasitic plants can severely reduce growth and development of the infected host plant. They parasitize host roots through a xylem-to-xylem connection made by a special organ, the haustorium (Parker and Riches, 1993). Through this connection, the parasite subtracts host-plant metabolites, water, nutrients, and amino acids (e.g., Press et al., 1987; Rogers and Nelson, 1962) and alters the hormone balance (Drennan and El Hiweris, 1979; Frost et al., 1997; Taylor et al., 1996). Most of the studies showing these effects have been conducted with C4 hosts (e.g., sorghum and maize), but similar interactions are expected with C3 hosts such as rice. Susceptible and sensitive rice varieties show stunted growth, low biomass production

Weed Management in Rice-Based Cropping Systems in Africa

161

and failure to flower (Riches et al., 1996). In a study by Cechin and Press (1994), rice plants infected with S. hermonthica did not produce any grain. This was partly attributed to reduced shoot growth and partly to reductions of up to 44% in photosynthesis compared to uninfected control plants. Severe crop damage (60–100%) has been reported in cases of heavy infestation with the parasitic weed Striga asiatica in Madagascar (e.g., Elliot et al., 1993). Striga spp. have a very successful life-cycle strategy including a wide host range, an out-crossing nature (with exception of S. asiatica), and prolific production of very small (0.2–0.35 mm) seeds (5000–85,000 per plant) with high off-season survival rates (e.g., Andrews, 1945; Krause and Weber, 1990; Parker and Riches, 1993; Stewart, 1990; Webb and Smith, 1996). Their germination depends on the availability of germination signals (xenognosins) (Saunders, 1933; Vallance, 1950; Yoder, 2001) that are exuded by suitable host plants, although some species (so-called false hosts) may provoke germination without supporting parasitism. Van Delft et al. (1997) calculated that it would only take 2–3 seeds producing Striga plants per m2 to fully replenish the annual seed-bank losses. There is evidence that parasitic weed problems are increasing in Africa and this is reported for Striga spp. in Nigeria (Dugje et al., 2006) and Ghana (Aflakpui et al., 2008). In Tanzania too, rice farmers witnessed a progressive decline in yields associated with an increased severity of S. asiatica infestations (Mbwaga and Riches, 2006). 2.1.2. Problem weeds in rain-fed and irrigated lowlands a. Sphenoclea zeylanica. The annual broad-leaved plant S. zeylanica Gaertner (Goosweed or Chickenspike) of the Sphenocleaceae family is very common, widespread (observed from West Africa to Madagascar), and often serious weed, typical of lowland rice (Elliot et al., 1993; Holm et al., 1991; Johnson, 1997; Kent et al., 2001). The species can be very competitive and this may be because of efficient uptake of nitrogen (Biswas and Sattar, 1991), is able to emerge from flooded soils (Kent and Johnson, 2001), and produces large numbers of miniscule seeds. b. Cyperus difformis. The annual C. difformis L. (Variable flatsedge or Smallflower umbrella sedge) of the Cyperaceae family is one of the most important weeds in lowland rice in Africa (Diallo and Johnson, 1997; Johnson, 1997). In the irrigated rice areas in Senegal, it is the most common weed species together with E. colona. C. difformis can be particularly abundant where fields are only intermittently flooded or where land leveling is poor. The weed is well adapted to direct-seeded rice production methods ( Johnson, 1997; Rao et al., 2007). It is a problem weed because of its short growth cycle, and it can form dense stands in the rice crop ( Johnson et al., 2004) and produce large quantities of seed. c. Echinochloa spp. Species of the genus Echinochloa constitute some of the most important and widespread grass weeds in rice worldwide

162

J. Rodenburg and D. E. Johnson

(Holm et al., 1991). In Africa, the most important species are E. colona (L.) Link ( Jungle rice or Awnless barnyard grass) and E. crus-pavonis (Kunth) Schultes (Gulf barnyard or Gulf cockspur grass). Distinguishing between E. crus-pavonis and E. crus-galli is difficult due to the morphological similarity and the high level of variation in the species which in turn can vary as a function of environment (Danquah et al., 2002; Holm et al., 1991). E. colona and E. crus-pavonis can both be found in rain-fed and irrigated lowlands and are often the most dominant species in a rice crop (Diallo and Johnson, 1997; Kent and Johnson, 2001). Like C. difformis, E. colona thrives on hydromorphic or lowland soils that are only temporarily flooded while E. crus-pavonis favors flooded conditions. These are problem weeds due to their close resemblance with rice at the early stages of growth which often causes confusion during transplanting or hand weeding. They are also highly competitive and have short life cycles (seed production within 70 days) and prolific seed production ( Johnson, 1997). A single E. crus-pavonis plant, for example, can produce more than 20,000 seeds. E. crus-pavonis is reported as a host of a number of pathogens including Rice Yellow Mottle Virus and could therefore contribute to survival and spread of this important disease in African rice (Abo et al., 2002). d. Oryza spp. Wild and weedy rices are important weeds in the lowland rice growing areas in Africa. Weedy rices are weedy biotypes of the cultivated rice species O. sativa L. and O. glaberrima Steudel (Delouche et al., 2007), while wild rices comprise the group of noncultivated rice species. Important wild rice species in Africa are the perennial O. longistaminata A. Chev. & Roehr., and the annuals O. barthii A. Chev. and O. punctata Kotschy ex Steud. ( Johnson et al., 1999). In Egypt, weedy/red rices (O. sativa) are more common (Delouche et al., 2007). Wild rice species constitute problems to lowland rice production due to their resemblance with the crop in the early stages and their competitiveness. In later stages, these species are easier to identify because they are tall, vigorous, and awned. Seeds shatter readily and they have variable seed dormancies, making them particularly difficult weeds to control in rice (Delouche et al., 2007). The perennial O. longistaminata is difficult to control because of the well-developed underground rhizome system ( Johnson et al., 1999). Besides the competition for resources with the crop, wild and weedy rices are the only alternative hosts for the African Rice Gall Midge and important hosts for Rice Yellow Mottle Virus ( Johnson et al., 1999), two of Africa’s most worrisome biological constraints in rice. Riches et al. (2005) reported problems with perennial wild rice (O. longistaminata) to be severe in Tanzania, where water cannot be fully controlled such as in the floodplains of the districts of Ifakara (east) and Kyela (south). In Mali, heavy infestation by O. longistaminata was reported to reduce rice yields on farmers’ fields by up to 85% and also forced farmers to change management practices or abandon fields

Weed Management in Rice-Based Cropping Systems in Africa

163

( Johnson et al., 1999). In Senegal, where rice is mainly grown in irrigated areas, it is estimated that 50% of the area sown to rice are infested by wild rice species O. longistaminata and O. barthii (Diallo, 1999). Yield reductions due to the annual wild rice O. barthii in Senegal are reported to be as high as 97% (Davies, 1984). e. Rhamphicarpa fistulosa. R. fistulosa (Hochst.) Benth is an annual facultative hemiparasitic weed in hydromorphic and rain-fed lowland rice ecosystems of tropical Africa ( Johnson et al., 1998b). The plant has white flowers that only open at night time. Flowers are pollinated by moths (Cisse´ et al., 1996; Parker and Riches, 1993). The seeds are very small (200-550 mm), numerous, and after ripening, the seeds are dormant for 6 months and then only germinate when exposed to humid conditions and daylight (Ouedraogo et al., 1999). Flowering occurs between 70–140 days after sowing the crop (Ouedraogo et al., 1999; Zossou, 2008). R. fistulosa is not a common weed but has been widely observed in West Africa (from Senegal to Benin) as well as in East and Southern Africa (e.g., Tanzania and Zimbabwe) (Bouriquet, 1933; Cisse´ et al., 1996; Johnson et al., 1998b; Kuijt, 1969; Ouedraogo et al., 1999; Parker and Riches, 1993). Locally, it causes important yield reductions in rice, millet, sorghum, maize, and even cowpea (Cisse´ et al., 1996; Gbehounou and Assigbe´, 2003; Hoffmann et al., 1997; Kuijt, 1969; Maiti and Singh, 2004; Neumann et al., 1998; Ouedraogo et al., 1999). As with Striga spp. it is difficult to reduce soil seed banks of this species due to its prolific seed production. In addition, R. fistulosa is a facultative hemi-parasite, and the control options are seriously limited by its wide host range and relative host independence. Different studies have reported or predicted increasing problems with R. fistulosa in crops, including rice, in West and East Africa (e.g., Gbehounou and Assigbe´, 2003; Gworgwor and Ndahi, 2004; Johnson et al., 1998b; Raynal Roques, 1994) and more recently in Benin (Zossou, 2008) and Tanzania ( J. Kayeke, personal communication).

2.2. The usefulness of weeds There is a growing understanding of the importance of ecosystem services to human well-being, and that the world’s poor has a disproportionate and direct reliance on these ecosystem services (MEA, 2005). Biodiversity has an important role in supporting and regulating ecosystem services such as nutrient cycling, pest and disease regulation, and pollination (UNEPWCMC, 2007). Biodiversity describes the abundance and diversity of genes and species, ecosystems, and habitats within a region. Biological interactions are important in maintaining ecosystem services (e.g., relation of predators and prey) and, in this context, weeds have critical, yet poorly understood, roles in the landscape. Weeds are the first stage in the vegetative

164

J. Rodenburg and D. E. Johnson

succession after land clearance or disturbance and, as such, recycle nutrients and protect the soil from erosion. Many weeds found in rice also have one or more direct benefits for African farmers in terms of domestic uses, crop protection and insect (mosquito) control (e.g., Akobundu, 1987; Burkill, 2004; Hillocks, 1998; Schwartz et al., 1998) (Table 4). Some weeds of rice production systems in Africa have been reported useful for the control of other weeds. Examples are a bioherbicide based on A. conyzoides (Xuan et al., 2004) or an improved fallow system with Calopogonium mucunoides Desv. (Akanvou et al., 2001) or C. odorata (Ngobo et al., 2004). Rice farmers often leave useful species untouched or keep them apart during hand weeding operations (Both, 2006). Wild rice species (e.g., O. longistaminata) are sometimes purposely left in the field and harvested by rice farmers to complement cultivated rice production particularly when food is scarce (e.g., Nyoka, 1983). Wild rice species (including O. barthii and O. punctata) also possess potentially useful traits against a variety of biotic (e.g., drought) and abiotic (e.g., bacterial blight, brown planthopper, and green leafhopper) production constraints (Khush, 1997) and can as such be useful in rice breeding programs. It has been speculated that wild rice species, and weed species like E. colona can, in theory, alleviate intensity of bird attack on rice (Treca, 1985), but the feasibility needs to be established before it could be advocated as a management practice. Weeds and mulches can have marked effects on the population dynamics of arthropods either through the influence of vegetation directly or through attraction of predator populations (Altieri et al., 1985). ‘‘Reliable natural enemy action’’ is in part dependent for continuity on the proximity of year round nonrice habitats such as vegetation covered bunds (Way and Heong, 1994). Studies in West Africa showed that ants were the most abundant predators in the rice canopy and abundance of these were greater in areas where weeds occurred (Afun et al., 1999b). Further, there was greater spider activity and beetles were more abundant where there was weed trash on the soil surface in rice fields (Afun et al., 1999a). To enhance abundance and efficiency of natural enemies, some noncultivated plants can be planted as intercrop or in field margins. Examples are grasses, such as molasses grass (Melinis minutiflora Beauv.), Napier grass (Pennisetum purpureum Schum.), and Sudan grass (Sorghum sudanensis (Piper) Hitch.), that attract natural enemies (e.g., Cotesia sesamiae) of stem borers (Chilo spp.) (Khan et al., 1997; 2000). Managing Paspalum scrobiculatum in the field margins may aid the control of African Rice Gall Midge by encouraging parasitoids (Nwilene et al., 2008). Weeds may however serve as hosts and sources of infection for fungal and viral diseases of rice (Ou, 1985). A good understanding of such relations is therefore required to achieve greater regulation of pests, through the maintenance of natural enemies, while avoiding these deleterious effects.

Weed Management in Rice-Based Cropping Systems in Africa

165

3. Weed Management Practices in African Rice-Based Cropping Systems 3.1. Cultural weed control 3.1.1. Planting methods Crop establishment is a key factor in determining the outcomes of weed–crop interactions and preventive weed management measures. A vigorous rice crop with a closed canopy denies weeds space and light. Crop establishment involves several steps of land preparation and sowing or planting depending on the agroecosystem. Crop establishment can be improved through soil tillage, land leveling, use of ‘‘clean seed,’’ transplanting with healthy seedlings and timely flooding and nutrient management. Such integrated crop management (ICM) practices can reduce the weed problems in lowland rice fields and were shown to increase productivity by 4–25%, depending on the level of water control (Becker and Johnson, 1999a, 2001b; Haefele et al., 2000). In the following section, rice establishment methods relevant to rice cropping systems in Africa are discussed; whereas land preparation is discussed in Section 3.2. In the traditional rice systems in Africa, particularly in the forest zones, shifting cultivation is still common practice (Balasubramanian et al., 2007). The crop is sown into the ashes after slash-and-burn, through direct-seeding either by broadcasting or dibbling in the uplands (Ampong-Nyarko, 1996) or by direct-seeding (dibbling) or transplanting on ridges or heaps in the inland valleys (Moormann and Juo, 1986; Windmeijer and Andriesse, 1993). In irrigated lowland rice, direct-seeding or transplanting is practiced. In the Sahel, the areas under these establishment methods are approximately equal and, for example, in the lower- and middle delta of the Senegal River Valley in Senegal rice is mainly direct-seeded while in the upper delta it is largely transplanted (Diallo and Johnson, 1997). In the humid forest and savanna zones, irrigation schemes are usually smaller, compared to those in the Sahel, and rice is mostly transplanted (Balasubramanian et al., 2007). Transplanting is usually done with 25- to 30-day-old rice seedlings, although often much older, and these are either planted in rows or at random. Compared to direct-seeding, transplanting saves seed, reduces the period the field is occupied and, importantly, it provides the crop with a competitive (size) advantage over weeds. Further, the soil can be flooded immediately after transplanting which suppresses the emergence of the majority of the potential weed species (see Section 3.1.2). Transplanting in rows facilitates the use of labor- and time-saving weeding equipment such as a hoe or a push weeder. Moreover, grasses that have similar appearance as rice, especially in the early stages, are easier to recognize if they occur outside the planting pattern.

166

J. Rodenburg and D. E. Johnson

Transplanting can cause a shock to rice seedlings which leads to longer growth periods and lower yield potentials (Poussin, 1997). Labor shortages in many areas however motivate farmers to continue to seed directly (Becker and Johnson, 1999a, 2001b). In direct-seeding, seeds may be sown dry as ungerminated seeds or ‘‘wet’’ as pregerminated seeds which are often sown into shallow water to reduce weed problems. Unless fields are well leveled, however, this may not result in effective weed control as the water layer will be of variable depth (Diallo and Johnson, 1997). Direct-seeded and transplanted rice have equivalent yields when weeds are properly controlled (De Datta et al., 1968), and direct-seeding can save labor compared to transplanting (Akobundu and Fagade, 1978). An agroeconomic study on rain-fed lowland rice in Southern Senegal, however, concluded that, overall, transplanting is less time-consuming, fits in better with other farm activities, and requires less fertilizer than direct-seeding of rice (Posner and Crawford, 1991). The ‘‘plasticity’’ of plants with respect to the available resources implies that there is a wide range of planting densities with more or less constant crop yield levels (Harper, 1977; Radosevich, 1987). Increasing the plant density within this range would in theory only increase a crop’s competitive advantage over weeds with no concomitant negative consequences for crop yield. This is the case with rice, and varying the plant population density is an option for improving its competitiveness. Increased seeding rates have been proposed and tested as a component for improved weed management (e.g., Akobundu and Ahissou, 1985; Cousens, 1985; Fagade and Ojo, 1977; Kristensen et al., 2008; Mohler, 1996). In the irrigated rice production schemes in the Sahel, (direct) sowing densities of up to 200 kg seed ha 1 have been observed (Diallo and Johnson, 1997). As seeding density surpasses a certain level, increased intraspecific competition may result in a poor crop growth (e.g., Rao et al., 2007). 3.1.2. Flooding Flooding is one of the most important weed management options in lowland rice (Diallo and Johnson, 1997) as many weeds will not germinate in anaerobic conditions. Maintaining a flood layer of 5–10 cm or more suppresses the growth of most species (Akobundu, 1987) and it is this means to limit weed growth that has enabled the sustainability of transplanted lowland rice. Even superficial flooding (2 cm of flood water) can reduce growth of one of the most noxious weeds, Echinochloa crus-pavonis (Kent and Johnson, 2001). This may require however that the soil remains flooded for prolonged periods throughout crop establishment as drainage or shallow flooding may encourage the emergence of grass weeds such as Leptochloa chinensis and Echinochloa spp. (Hill et al., 2002). It is the timing, duration, and depth of flooding that determines the extent of weed suppression by flooding (Mortimer et al., 2005). Weeds tend

Weed Management in Rice-Based Cropping Systems in Africa

167

to be recruited in the early stages of the rice crop and management of water at these stages can be critical in determining the nature and abundance of the weed flora. In a study of wet-seeded rice sown on puddled soil, where the soil was flooded 10–15 DAS after seeding, the recruitment of sedges and broadleaves occurred in the early stages of the crop while grass weeds continued to increase in density up to 60 days after sowing (Hill et al., 2002). In dry-seeded rice, the pattern of germination is likely to be determined by the moisture regime and the timing of flooding. As weed seedlings are reliant largely on the seed reserves to enable them to emerge from flooded conditions, seed size will influence the ability of species to establish under flooded conditions. C. difformis for instance might already be suppressed by 0.8 cm of turbid water while for suppression of L. chinensis 1.5 cm or more would be necessary and Echinochloa crus-galli has sufficient seed reserves to emerge from 8 cm of water (Chauhan and Johnson, 2008e; Mortimer et al., 2005). Another variable is dormancy and while this may be pronounced or variable in some species, others (e.g., Fimbristylis miliacea and E. colona) exhibit no dormancy and germinate rapidly on the surface of puddled soil (Kim and Moody, 1989). Farmers require appropriate field infrastructure to precisely manage flooding and drainage in the field to exploit differentials between rice and weeds. For effective control of weeds by flooding, fields need to be well leveled to ensure uniform water depth. Good land leveling requires skills and equipment not commonly available to resource-poor farmers in this region. As a result, uniform flooding is often difficult to achieve and therefore other control methods need to be integrated to provide adequate weed control (Akobundu, 1987). 3.1.3. Soil fertility management Weeds have been observed to have less effect on adequately fertilized crops compared to unfertilized crops due to the more vigorous crop growth (e.g., McKenzie, 1996; Tollenaar et al., 1994). It has also been demonstrated however that, for example, Echinochloa sp. responded more to fertilizer than the rice (Gibson et al., 1999). Timing of fertilizer application may be very important with respect to its influence on the outcome of competition. Early fertilizer applications stimulate weed growth especially of weeds with small seed sizes that have little reserves (Liebman and Davis, 2000). In upland rice in the forest zone of West Africa, N application tended to increase weed growth and only benefited yields when it was accompanied by improved weed control (Becker and Johnson, 2001a). While farmers commonly recognize the importance of combining improved weed management with fertilizer applications, data from Africa on the impact of improved rice crop nutrition on competition with weeds are very limited. Improved soil fertility is important for the effective management of parasitic weeds (e.g., Ransom, 2000). The application of urea at 3 weeks after sowing reduced the number on S. asiatica infections in rice in Tanzania

168

J. Rodenburg and D. E. Johnson

(Riches et al., 2005). Further, application of 90–120 kg N ha 1 proved an adequate method to delay and reduce S. hermonthica infestation and ensure satisfactory crop yields in upland rice in Nigeria (Adagba et al., 2002a). Improved yields may be the result of reduced and delayed parasitism rather than improved host performance (Cechin and Press, 1994), but the costs of these applications may be a constraint to farmer adoption (Riches et al., 2005). 3.1.4. Mulching Mulching is a feasible option in upland rice but not widely practiced in Africa. Mulching with residues from trees (Budelman, 1988; Kamara et al., 2000; MacLean et al., 2003) or crops (e.g., Iwuafor and Kang, 1993; Singh et al., 2007) has shown to suppress weeds in cereal crops, including rice. Mulching can inhibit weed seed germination by shading and in some cases through the release of allelopathic substances (e.g., Akobundu, 1987; Singh et al., 2003). Rice straw proved an effective mulch material to reduce weed growth (Lal, 1975). A limitation of this practice is the limited availability of suitable mulching material in many parts of Africa. Rice straw, for instance, also has an economic value in many areas as it is often used for forage or fuel. Mulching can also hinder rice establishment, encourage pests such as termites (Akanvou et al., 2000) or, in the case of rice straw, facilitate the survival and spread of rice diseases over the off-season. 3.1.5. Mixed cropping, rotations, and fallow Intensifying cropping systems can result in an increase in weed growth, increasing losses due to weeds and, with inadequate control, larger soil seed banks (Akobundu et al., 1999; Ekeleme et al., 2000). Traditionally, rain-fed rice farmers in Africa use fallow and rotations to interrupt the buildup of weeds. Rotations with noncereal crops like cowpea and soybean in the savanna and forest uplands and groundnut, soybean, cassava, potato, sweet potato, or vegetables in the rain-fed lowlands are often practiced in subsistence rice-based production systems. Changes in management and crop rotations help prevent the buildup of crop-specific weeds (Akobundu, 1987). Improved fallows and intercropping can be effective measures but their introduction, where these are not already traditional practice, has met with limited success (see below). In the humid forest zones, African rice farmers traditionally manage weeds and other stresses through long fallow periods in shifting cultivation systems. In the Taı¨ forest in Coˆte d’Ivoire, for instance, a single crop of rice may be followed by many (up to 20) years of fallow (de Rouw, 1995). In such extensive systems, weed population buildup is limited and farmers need little effort for additional control (de Rouw, 1995). Such practice is still common in some areas (Ampong-Nyarko, 1996; Johnson, 1997) but becoming less frequent. With human population growth and concomitant

Weed Management in Rice-Based Cropping Systems in Africa

169

increased pressure on land, fallow lengths are progressively being reduced (Braimoh, 2006; de Rouw, 1995; Demont et al., 2007). Cropping intensification in these systems may lead to an increase in the losses due to weeds (Becker and Johnson, 2001a). Consequently, there is a clear trade-off between fallow length and weeding labor requirements (Dvorak, 1992). The significance of weed management in the traditional upland rice production systems is underscored by the fact that weeding can require 29% of all the time needed for cropping operations including clearing, soil tillage, planting, harvest, and transport (Windmeijer and Andriesse, 1993). To facilitate intensification in upland rice systems, relay cropping with weed-suppressing legumes may be a viable alternative to reduce weed growth and improve soil fertility (Becker and Johnson, 1999b). Species choice and planting date need to be carefully chosen to avoid the threat of severe competition for resources between the legume and the rice. Sowing of Cajanus cajan 56 days after rice proved an appropriate management practice in this respect (Akanvou et al., 2002). Legumes may continue to grow after rice harvest and thereby suppress weeds during the off-season. Such ‘‘short fallow’’ rotation systems were shown to increase rice yields (20–30% across agroecosystems) and lower weed pressure in the forest and savanna zones of West Africa (Akanvou et al., 2000; Becker and Johnson, 1998, 1999b). Legume species proposed for improved fallows in rice-based cropping systems in Africa are summarized in Table 5. Best-bet legumes for upland cropping systems in different agroecological zones have been proposed (Becker and Johnson, 1998, 1999b). Fallow species including Cassia occidentalis L. (Mallamaire, 1949) and Aeschynomene histrix (Merkel et al., 2000) were reported to control S. hermonthica. A. histrix acted as ‘‘trap crop’’ in infested fields by stimulating the germination of Striga seeds without supporting parasitism. Intercropping with a legume crop is another cultural practice to reduce weed problems. For instance in India, Sengupta et al. (1985) showed that intercropping upland rice with black gram (Vigna mungo [L.] Hepper) contributed to improved weed suppression compared to rice monoculture. Successful examples of intercropping in rice-based systems in Africa are scarce. For the control of Striga spp., rotations or the use of intercrops especially with ‘‘trap crops’’ like cowpea (e.g., Carsky et al., 1994b), yellow gram (Oswald et al., 2002), pigeon pea (e.g., Oswald and Ransom, 2001), soybean (e.g., Carsky et al., 2000; Robinson and Dowler, 1966), groundnut (Carson, 1989), or cotton (e.g., Murdoch and Kunjo, 2003) has been proposed. Most of these crops grow in similar environments as upland rice and could fit existing cropping systems. For example, rotations with Crotalaria ochroleuca or pigeon pea in Tanzania resulted in improved rice productivity in S. asiatica infested fields, reduced Striga infection and the number of weeding operations required (Riches et al., 2005). Intercropping with the fodder legumes Silverleaf (Desmodium uncinatum) and Greenleaf

170 Table 5

Suitable legume species for weed-suppressive fallow rotations or intercrops in African rice-based cropping systems Spatial and temporal arrangement

Characteristics and traits

Rain-fed lowland

Off-season fallow

Aeschynomene histrix

Savanna and forest upland

Relay seeding or offseason fallow, burning of residues

Cajanus cajan

Forest and savanna upland Savanna and forest upland

Off-season fallow, burning or mulching of residues Off-season fallow

Biomass accumulation Weed suppressive High N accumulation Forage value Weed suppression Striga hermonthica trap crop High N accumulation Weed suppressive High N accumulation Forage value Weed suppression

Species

Ecosystem

Aeschynomene afraspera

Canavalia ensiformis

Sources

Becker and Johnson (1999b)

Becker and Johnson (1999b), Becker and Johnson (1998), and Merkel et al. (2000)

Becker and Johnson (1998) and Akanvou et al. (2000) Becker and Johnson (1999b), Becker and Johnson (1998), and Akanvou et al. (2000)

Cassia occidentalis

Upland

1 year fallow

Mallamaire (1949)

Off-season fallow

Rhamphicarpa fistulosa and Striga spp. control Weed suppressive

Crotalaria anagyroides Crotalaria juncea

Forest upland

Crotalaria ochroleuca Mucuna spp.

Savanna upland and rain-fed lowland Upland

Off-season fallow

Weed suppressive

Becker and Johnson (1999b) and Becker and Johnson (1998)

Rotation

Striga asiatica control High N accumulation Weed suppressive Biomass accumulation Weed suppressive High N accumulation Weed suppressive

Riches et al. (2005)

Savanna upland

Off-season fallow

Sesbania rostrata

Rain-fed lowland

Off-season fallow

Stylosanthes guianensis

Savanna and forest upland

Relay seeding or offseason fallow, burning of residues

Becker and Johnson (1998)

Becker and Johnson, (1999b) and Becker and Johnson (1998) Becker and Johnson (1999b)

Becker and Johnson (1999b) and Becker and Johnson (1998)

171

172

J. Rodenburg and D. E. Johnson

(D. intortum) was also shown to reduce S. hermonthica infestations in maize in East Africa (Khan et al., 2006). Desmodium spp. is suggested to have an allelopathic effect on Striga spp. (Khan et al., 2002). Further, a modeling study showed that the use of a ‘‘trap crop’’ to reduce the Striga soil seed bank was more effective when the legume was being intercropped rather than grown separately in a rotation (van Mourik et al., 2008). Legumes or cover crop species need to be rotated to avoid the buildup of detrimental weeds and pests (Teasdale, 2003). At the end of the fallow period, legumes may be cut and burnt, removed, incorporated in the soil, or mulched in order to avoid unnecessary competition to the subsequent rice crop. The best residue management, from a weed management perspective, appeared to be burning in the forest zones (Akanvou et al., 2000; Tonye et al., 1997) and incorporation in the soil in the savanna zones (Akanvou et al., 2000). Residue burning results in less weed infestation in subsequent crops than soil incorporation (Akanvou et al., 2000). Burning, however, causes substantial loss of nitrogen to the atmosphere (e.g., Juo and Mann, 1996), and destruction of the protective surface litter layer increases the risk of erosion (e.g., Alegre and Cassel, 1996). Intercropping, improved fallow systems, relay cropping or rotations with legumes have had low farmer adoption rates in Africa. Some reasons for this are the additional labor and energy required for clearing and incorporation of the legume into the soil, a poor fit with traditional cropping systems, lack of land tenure, subsequent poor crop establishment and additional costs of inputs (Faulkner, 1934; Langyintuo and Dogbe, 2005; Tarawali et al., 1999). Desmodium spp., for instance, may have little potential for adoption as the species has proved difficult to establish, has a limited geographical range, and would only have any economic value as a fodder in mixed farming systems (Gressel and Gebrekidan, 2007). Direct economic benefit proved imperative for legumes to be acceptable for farmers in West Africa (Becker and Johnson, 1999b).

3.2. Mechanical weed control Mechanical weed control can be applied as an intervention within the crop, and as a preventative measure as part of preseason land preparation or as off-season dry-soil tillage. Preventive mechanical weed control options can be differentiated as either off-season soil tillage between harvest and establishment of the next crop or land preparations prior to crop establishment that may include tillage, leveling, and puddling. Off-season dry-soil tillage at sufficient depth may help breaking and drying subsoil rhizomes of perennial weeds. Tillage in dry-soil tillage is often however too superficial to bury weed seeds or control perennial species (Diallo and Johnson, 1997) particularly where mechanization is limited. When soil is sufficiently moist, for instance, after the first rains at the onset of the rainy season, several tillage

Weed Management in Rice-Based Cropping Systems in Africa

173

passes with sufficient time intervals enable weeds to germinate can limit following weed growth (Diallo and Johnson, 1997). Land preparation on small-scale farms in rain-fed systems is usually undertaken manually and commonly with a short-handled hand hoe. In some inland valleys (e.g., in Sierra Leone, Coˆte d’Ivoire) animal traction or small power tillers have been introduced (Ampong-Nyarko, 1996) but many rice farmers in Africa are restricted by scarce resources and limited availability of animals. The latter may be determined by the presence of the Tsetse fly. In larger, irrigated schemes as found in Egypt, Madagascar, and Nigeria, however, land preparation is often mechanized (Akobundu, 1987) with medium and large, twin-axle tractors (Wanders, 1986). In these situations, land may be prepared by wet rotovation (Ampong-Nyarko, 1996) or by using disc ploughs or harrows (van der Meijden, 1998). Due to a common lack of equipment and mechanization in the rain-fed lowland production systems, fields are often inadequately tilled, bunded, and leveled. Unleveled land and the absence of bunds in the inland valleys result in uneven flooding and patchy conditions which favors weed growth and increases weed control costs (Akobundu and Fagade, 1978; Ampong-Nyarko, 1996). Puddling, or the thorough tillage of flooded soil, besides controlling any established weeds, promotes vigorous rice growth and enhances crop competitiveness with weeds (De Datta and Baltazar, 1996). Soil puddling is not widely practiced in Africa, as it is in Asia, which is perhaps primarily due to the lack of draught animals and small power tillers noted above. Hand weeding is the most widely practiced intervention against weeds on small-scale rice farms in Africa (e.g., Adesina et al., 1994), yet this is labor demanding, and requires 250–780 man h ha 1 (Akobundu, 1987; Akobundu and Fagade, 1978; Stessens, 2002). Using this labor requirement, and assuming eight working hours a day at a daily wage of 1.5 per person, weeding costs range from 48 up to 149 per ha. This, however, assumes that farmers have alternative opportunities for employment that pay 1.5 per person or more. Vissoh et al. (2004) showed that hand weeding costs ( 57 per ha) were comparable to the costs of applying herbicide (Garil) to rice ( 58 per ha) in Benin. On subsistence farms, weeding is mostly carried out by women from the farm household and involvement of children is common. On larger farms, labor for hand weeding may be hired from outside the farm family and, in these cases, costs can exceed those for herbicide use. Provided adequate labor is available, hand weeding is an effective method to prevent weeds from producing seeds. In deep-water rice, for instance, it is suggested as the most effective management practice for O. barthii (Catling, 1992). However, for most perennial weeds, such as O. longistaminata and I. cylindrica, hand weeding alone is unlikely to provide adequate control (Akobundu, 1987) as these are capable of rapid regrowth from rhizomes. A further disadvantage of hand weeding is that weeds need

174

J. Rodenburg and D. E. Johnson

to grow tall enough to be hand pulled, by which time competition for resources, extraction of metabolites, or phytotoxic effects in case of parasitic weeds has already taken place. Hand hoes or push weeders are often used in row sown crops providing rows are spaced wide enough (Rijn, 2001), and the implements are available to farmers. A shortcoming of such devices is that it does not target weeds in the row and when used close to the rice plant they may also cause crop damage (Navasero and Khan, 1970). The use of power tillers or tractors for mechanical weeding is not common in West Africa and, for instance, only 4% of the rice area is mechanized in Senegal (van der Meijden, 1998). In irrigated systems in river deltas, such as the Senegal and Niger rivers, the clay soils seriously limit the effectiveness of mechanized weeding during the cropping season. Attempts at mechanization in the Senegal River Valley have failed due at this constraint in addition to the limited financial resources of most rice farmers (Diallo and Johnson, 1997). To prevent weed induced yield losses, two to three weeding operations are required for upland and three for hydromorphic and flooded rice (Ampong-Nyarko and De Datta, 1991). Despite recommendations to the contrary however, weeding is frequently inadequate or delayed, often due to labor shortages or conflicts between on- and off-farm activities ( Johnson et al., 1998a).

3.3. Rice varietal development for improved weed control In rice systems where farmers have scarce resources and use few external inputs, as often found in Africa, rice varieties that suppress weeds maintain high yields under weedy conditions and are well adapted to the local conditions would bring considerable advantages to resource-poor farmers ( Johnson et al., 1998a). In morphological terms, weed competitive rice varieties are suggested to be those that are tall and have a high tillering ability, a high specific leaf area (SLA = leaf area per leaf dry weight), erect to droopy leaves and relative long crop durations to compensate from losses suffered during early weed competition (Asch et al., 1999; Dingkuhn et al., 1998, 1999; Fofana and Rauber, 2000). Cultivars of the African rice species Oryza glaberrima have shown yield advantages under weedy conditions compared to the Asian O. sativa varieties ( Johnson et al., 1998a). There are possible trade-offs between various competitive characteristics (e.g., Dingkuhn et al., 1999; Perez de Vida et al., 2006) or between competitive traits and yield potential (e.g., Jannink et al., 2000; Jennings and Aquino, 1968; Kropff et al., 1997). Although some studies showed that such trade-offs are no general phenomena (e.g., Garrity et al., 1992; Haefele et al., 2004; Pernito et al., 1986), many desirable morphological characteristics with respect to weed competitiveness may have negative effects on yield potential. For instance, characteristics associated with high yielding modern varieties,

Weed Management in Rice-Based Cropping Systems in Africa

175

such as short stature and erect leaves, are considered to be unfavorable for weed suppression ( Johnson et al., 1998a). Droopy leaves, on the other hand, may shade out weeds but limit light penetration to lower rice leaves, while tall rice plants may compete for light more effectively than shorter plants but these may be more prone to lodging (Bastiaans et al., 1997). While O. glaberrima can be competitive with weeds, they have low yield potentials and yield losses are incurred due to lodging and grain shattering (Dingkuhn et al., 1998; Jones et al., 1997; Koffi, 1980). Interspecific hybrids of O. sativa and O. glaberrima were developed with higher yield potential and without the seed shattering characteristic. Varieties derived from these interspecific crosses were named New Rice for Africa (NERICA) and currently comprise 18 upland and 60 lowland varieties (Rodenburg et al., 2006b), of which 17 upland and 11 lowland varieties have been released in SSA (I. Akentayo, personal communication). Early observations on these varieties, developed for the upland areas, have shown that some putative traits of the O. glaberrima parent, contributing to weed suppressiveness, and traits of the O. sativa parent, contributing to yielding ability, are heritable (Dingkuhn et al., 1999; Johnson et al., 1998a; Jones et al., 1997). In a recent study carried out in two upland environments in Nigeria, compared to the popular check variety ITA150 and the NERICA parents (WAB56-104 and CG14), NERICA-1, -2, and -4 generally had slightly higher weed infestation levels and relative yields losses due to weed competition (Ekeleme et al., 2009). In the same study, however, all three NERICA varieties had higher yields than CG14 and ITA150 when the crop was weeded one or two times. Another recent study carried out in a lowland environment in Benin, showed that nine lowland varieties of NERICA (NERICA-L-6, -32, -35, -37, -42, -53, -55, -58, and 60) have significant higher yields than both lowland NERICA parents under weedy and weed-free conditions, and comparable yield performances as the high yielding and weed competitive check variety Jaya (Rodenburg et al., 2009). Even though varietal differences in weed competitiveness have been found in rice (Fischer et al., 2001; Garrity et al., 1992; Zhao et al., 2006a), so far, only a limited number of varieties are confirmed to combine superior weed competitiveness with good adaptation to African rice ecosystems. In upland fields in Coˆte d’Ivoire, O. glaberrima varieties IG10 (Fofana and Rauber, 2000), CG14, and CG20 ( Jones et al., 1996) were found to be superior in suppressing weeds but also had low yield potential. On hydromorphic soils in Nigeria, the tall variety OS6, incurred 24% less yield reductions from weed competition than the semidwarf cultivar ANDNY11 (Akobundu and Ahissou, 1985). In Senegal, Haefele et al. (2004) reported that lowland rice variety Jaya was weed competitive and high yielding compared to a range of varieties. Jaya incurred lower yield losses due to weeds (<20%) compared to popular Sahel 108 (>40%). Superior performance of Jaya under both weedy and weed-free conditions

176

J. Rodenburg and D. E. Johnson

was confirmed in a study carried out in Benin (Rodenburg et al., 2009). This study also identified nine superior lowland NERICA varieties as noted above. Table 6 lists some O. glaberrima, O. sativa, and interspecific rice varieties that have shown to be weed competitive in Africa. Varieties with superior levels of weed competitiveness have been confirmed in other regions, such as Apo and UPLRi-7 in Asia (Zhao et al., 2006a; 2007), Oryzica Sabana 6 in Latin America (Fischer et al., 2001), and M-202 in north America (Gibson et al., 2001), and these could be tested under African rice production conditions in the future. It was suggested, but not demonstrated, that the weed-suppressive ability of IG10 (O. glaberrima) may be, in part, due to allelopathy (Fofana and Rauber, 2000). A large number of reviews has already been published on crop allelopathy (e.g., Belz, 2007; Olofsdotter et al., 2002; Singh et al., 2003; Weston and Duke, 2003; Xuan et al., 2005). Crop allelopathy refers to the process of the release of chemical compounds by living and intact roots of crop plants that affect plants of other species (Belz, 2007; Olofsdotter et al., 1999a; Weston and Duke, 2003). Allelopathy is suggested by many as one of the potential mechanisms to suppress weeds and as a possible component in integrated weed management (IWM) (e.g., Belz, 2007; Fofana and Rauber, 2000; Jordan, 1993; Olofsdotter et al., 2002; Weston, 1996; Weston and Duke, 2003). Weed suppressiveness and allelopathy may, however, be confounded and they may coexist in the same variety (e.g., Olofsdotter et al., 1999a). Indeed, as Rao et al. (2007) suggest, the significance of allelopathy for weed management in rice will remain conjecture until it is clearly demonstrated that differences observed in bioassays also occur in the field. The use of weed competitive varieties is unlikely to be feasible as a standalone technology but rather it may be a valuable component of integrated measures. Suitable varieties should, in addition to weed competitiveness, also possess other traits (Dingkuhn et al., 1999) like resistance or tolerance to other biotic and abiotic stresses. Furthermore, a suitable variety needs to be well adapted to the environment and have the specific characteristics desired by farmers and consumers. Rice varietal development may contribute to the management of parasitic weeds in rice. Differences among O. sativa and O. glaberrima in the interaction with Striga spp. have been observed, and a selection of African rice species (O. glaberrima) showed greater Striga resistance than O. sativa varieties ( Johnson et al., 1997, 2000; Riches et al., 1996). An O. glaberrima cultivar, CG14, showed resistance against S. hermonthica and S. aspera ( Johnson et al., 1997; Kaewchumnong and Price, 2008). This was not expressed in the progenies (F7) from interspecific hybrid of CG14 with the O. sativa WAB56-104 and it appeared that the resistance to Striga may have been lost during the repeated back-crossing ( Johnson et al., 2000). In another study carried out by Gurney et al. (2006), many of the O. glaberrima varieties (including CG14) that showed resistance in the field

Table 6

A selection of rice varieties with proved superior levels of weed competitiveness in African production systems

Ecosystem

Variety

Species

Main superior traits

Sources

Upland

IG10

O. glaberrima

CG14

O. glaberrima

Biomass; Tiller number; LAI; SLA; Early vigor; Yield under weedy conditions; Root length density SLA; Tillering; Early vigor; Weed suppression

CG20

O. glaberrima

Johnson et al. (1998a) and Fofana and Rauber (2000) Asch et al. (1999), Dingkuhn et al. (1998), and Jones et al. (1996) Jones et al. (1996)

ACC102257 WAB96-1-1 SP4

O. glaberrima O. sativa O. sativa

Jaya

O. sativa

TOG5681 NERICA-L -6, -32, -35, -37, -42, -53, -55, -58, and -60

O. glaberrima interspecific

Lowland

SLA: Tillering; Early vigor; Weed suppression Root length density Height; Weed suppression Height; Weed suppression Yields under weedy and weed-free conditions; Weed suppression Weed suppression Yields under weedy and weed-free conditions

Fofana and Rauber (2000) Jones et al. (1996) Jones et al. (1996) Haefele et al. (2004) and Rodenburg et al. (2009) Rodenburg et al. (2009) Rodenburg et al. (2009)

178

J. Rodenburg and D. E. Johnson

in Coˆte d’Ivoire were found to be susceptible in the lab. This could be due to differences in Striga species and strains, growing conditions, or differences in screening methods used (detail or moment of observation). It has been observed earlier that the expression of Striga resistance or tolerance may differ between pot and field trials (e.g., Omanya et al., 2004; Riches et al., 1996; Rodenburg et al., 2005). As even highly resistant crop varieties have been shown to be susceptible to infection by Striga spp., breeders should aim at incorporation of tolerance into the resistant material (Haussmann et al., 2001; Pierce et al., 2003; Rodenburg et al., 2005).

3.4. Biological weed control No published evidence is available on farmer adoption of biological weed control in rice in Africa. The possible reasons for this are yet to be validated but the implementation of biological control has a number of intrinsic constraints and, further, smallholder farmers have limited access to the technologies. Biological control agents are generally very host specific and their use requires a relative high skill level to put them in practice. This knowledge is often lacking among poor farmers in Africa (e.g., Abate et al., 2000; Riches et al., 1993). Outside of Africa, suitable candidate pathogens have been identified for the biological control of weeds that also occur in African rice systems, such as Dactylaria higginsii against C. rotundus and C. iria (Kadir and Charudattan, 2000) and Alternaria alternata to control S. zeylanica (Masangkay et al., 1999). Hong et al. (2004) found allelopathic properties in some wild plants in Vietnam that could be used for biological control. Two of these plants (Bidens pilosa and Euphorbia hirta) are also found as upland rice weeds in Africa and therefore may have relevance for biological control in rice cropping systems. Another rice weed with putative potential for use in a bioherbicide is A. conyzoides (Xuan et al., 2004). The leaf-feeding moth Pareuchaetes pseudoinsulata/insulata has been released in Ghana, Nigeria, and South Africa for the control of C. odorata in plantations (Gunasekera and Rajapakse, 1994; Kluge and Caldwell, 1993). No reports are available however on the use of such biological agents in rice cropping systems in Africa. Biological control may have a role in the management of invasive weeds and, for example, some biological control methods tested on Striga spp. in maize and sorghum could be applied in rice-based cropping systems. Larvae of the weevil Smicronyx spp. (Coleoptera: Curculionidae) feed on Striga spp. seeds inside their capsules and prevent seed production (Pronier et al., 1998) but their effectiveness as biological control agent is limited (Smith et al., 1993). Pathogenic fungi like Fusarium spp. may be used as biological control against Striga spp. (Ahmed et al., 2001). The low virulence of some plant pathogens is reported as a constraint to the application of biological control, and control agents are rarely able to eradicate an established weed population

Weed Management in Rice-Based Cropping Systems in Africa

179

or reduce the invasion of weed species into new areas (Sands et al., 2007). This may in part be due to the evolutionary necessity of allowing some host plants to persist in order for the pathogen itself to survive (Gressel et al., 2007; Sands et al., 2007). Some promising results with fungal pathogens have however been obtained and, for example, Fusarium (i.e., F. oxysporum and F. solani) was shown to reduce Striga emergence by up to 98% (Abbasher et al., 1995; Kroschel et al., 1996). The use of rhizobacterial strains such as Pseudomonas fluorescens and P. putida isolates, as seed treatments, may be useful as biological control agents (Ahonsi et al., 2002) if phytotoxic effects on the crop can be precluded. Field inoculation with arbuscular mycorrhizal fungi was found to be an effective biological control method against S. hermonthica in sorghum (Lendzemo et al., 2005). These options could have useful applications in rice-based cropping systems although there are major ‘‘bottlenecks’’ for these technologies that include the lack of availability of the pathogens and their limited suitability for smallholder systems. For example, inoculum needs to be brought to the field and incorporated into the soil in sufficiently high quantities. This could be achieved either through seed coatings (Ciotola et al., 2000) or via granular formulations (Elzein, 2003; Marley and Shebayan, 2005) at time of crop sowing. The formal seed-supply systems in sub-Saharan Africa are, however, weak (Balasubramanian et al., 2007) and specialized agro-industries limited, and hence technologies dependent on such infrastructures are unlikely to become widely available to farmers in the near future. Establishment of public–private partnerships or training programs for on-farm production of inoculum and application-media might provide a necessary shortcut for such developments.

3.5. Chemical weed control 3.5.1. Conventional chemical weed control Herbicides are important control methods in the lowlands, and in upland rice grown in rotation with cotton ( Johnson, 1997). The use of herbicides is economically attractive as it requires less overall weeding time and it enables the farmer to use time- and labor-saving planting methods such as direct (broadcast) seeding (e.g., Akobundu and Fagade, 1978; Babiker, 1982; Riches et al., 2005). Herbicides are likely to be particularly useful in areas where labor is in short supply. Farmers should also have sufficient financial resources to invest in herbicides and the return of such investments should be high enough. In the rain-fed rice production systems in the Casamance (South Senegal) herbicides were found to be a profitable investment on fertile soils (Posner and Crawford, 1991). Herbicides are often used in combination with other control options and, for example, in the irrigated rice systems in Senegal, most farmers rely on chemical weed control followed by hand weeding (e.g., Haefele et al., 2002).

180

J. Rodenburg and D. E. Johnson

For effective and safe herbicide use, the appropriate product, application equipment and application rates are important (Zimdahl, 2007). Moreover, herbicide application requires good timing with respect to crop and the growth stage of weeds (King and Oliver, 1992), weather conditions (Hammerton, 1967) and flooding. Interactions between flooding and herbicides tend to be product specific (Ampong-Nyarko and De Datta, 1991). Good chemical weed control under conditions of imperfect water management has been reported with different mixtures of propanil with thiobencarb, oxadiazon, and fluorodifen (Akobundu, 1981). Farmers require the knowledge on exactly how and when to apply herbicides to achieve effective control (Haefele et al., 2000; Hill et al., 2002). In Africa, where farmers generally have limited access to information and where literacy rates are low, the knowledge of proper herbicide use is often inadequate. Due to this, it is common that herbicide applications are too late, the herbicides poorly applied, the rates incorrect or the applications rendered ineffective by improper water management. This may result in inefficient weed control (Haefele et al., 2000), increased costs and phytotoxicity damage to the crop (e.g., Gitsopoulos and Froud-Williams, 2004; Johnson et al., 2004; Riches et al., 2005). In turn, this may cause reduced crop vigor or plant population densities and increased weed competition. In addition to limited access to information on weed biology, herbicide action, and proper application methods, African farmers often have limited market access. The markets are also often characterized by an insufficient range of products and intermittent supplies. In addition, African farmers often lack sufficient financial means for the purchase of the product and application and protection equipment (Balasubramanian et al., 2007). The incorrect use of herbicides, caused by the above cited problems, may accelerate the evolution of herbicide resistance in weeds ( Johnson, 1995). As mentioned above, good water control in lowland rice is important for effective herbicide use. The combination of a preemergence herbicide with effective water management can provide season-long weed control (Ampong-Nyarko, 1996). In fields prone to uncontrolled flooding, such as in hydromorphic areas and unimproved inland valleys, herbicide efficiency may however be very low (Akobundu, 1987). Formulations that can be applied directly to the irrigation or flood water rather than spraying, and hence not requiring equipment, may be particularly suitable for resource-poor rice farmers ( Johnson, 1995). For parasitic weeds, in addition to the above constraints, the use of postemergence herbicides has two major limitations. Firstly, detrimental effects on the crop have already occurred before the parasite emerges aboveground and, secondly, there are few effective herbicides available. The herbicide 2,4-D was found to be effective against S. hermonthica (Carsky et al., 1994a) and S. asiatica (Delassus, 1972). However, 2,4-D has a low selectivity and, like many other herbicides, requires multiple applications to

Weed Management in Rice-Based Cropping Systems in Africa

181

affect Striga. These constraints may be overcome by rice seed treatment with herbicide. In upland rice in Nigeria, cinosulfuron (0.2–0.6 g l 1) and CG152005 (0.064 g l 1) delayed and reduced S. hermonthica infection, and it was suggested this could be used in combination with resistant varieties (Adagba et al., 2002b). Possible drawbacks to such approaches are unfavorable effects on the environment as, for instance, Ahonsi et al (2004) found that the ALS-inhibitors imazaquin and nicosulfuron have negative impacts on soil biology and natural suppression of Striga spp. Commonly used herbicides in rice in Africa can be found in Table 7. Herbicide use in rice in Africa is poorly documented and recent publications covering currently used products are not available. Herbicides targeting broad-leaved weed species in rice in Africa are 2,4-D and MCPA, while, butachlor, molinate, oxadiazon, and thiobencarb are commonly used against grass weeds ( Johnson, 1997; Rao et al., 2007). Glyphosate, a herbicide used in land preparation for rice, is effective against O. longistaminata and O. barthii as preemergence treatment (Davies, 1984; Riches et al., 2005). Propanil is a popular herbicide for use in tank mixtures and, for example, one of the most frequently used combinations in rice production schemes of the Senegal River Valley is propanil and 2,4-D + dichlorprop (e.g., Haefele et al., 2000). Postemergence applications of propanil mixed with piperophos (Imeokparia, 1994), molinate (Babiker, 1982), thiobencarb, fluorodifen, or oxadiazon (Akobundu, 1981; Okafor, 1986) proved successful in irrigated rice in various other African countries. In irrigated direct-seeded rice, good weed control was obtained with preemergence applications of dymrone or thiobencarb in the Lake Chad Basin in Nigeria (Okafor, 1986), and bifenox or oxadiazon in Sudan (Babiker, 1982). In upland rice in Nigeria, good weed control has been reported by using mixtures of pretilachor with dimethametryne and piperophos with cinosulfuron (Enyinnia, 1992; Ishaya et al., 2007). Chemical weed control is best used in conjunctions with other weed management components within an IWM approach (Rijn, 2001). In this respect, however, herbicides may not lend themselves to be combined with other practices such as mixed cropping systems (Akobundu and Fagade, 1978) or biological pest control (e.g., Afun et al., 1999b; Taylor et al., 2006). 3.5.2. Herbicide resistant rice technologies Rice varieties with resistance against postemergence nonselective or broad-spectrum herbicides could facilitate improved weed management in some situations (e.g., Fernandez-Quintanilla et al., 2008). These may be particularly useful for the control of problem weeds like wild and weedy rice species. Worldwide there are three herbicide resistance (HR) technologies. One of them, known under the commercial name ClearfieldÒ , was developed through mutagenesis and ClearfieldÒ rice possesses resistance to broad-spectrum imidazolinone herbicides (e.g., Sha et al., 2007; Tan et al.,

182 Table 7 Herbicides (alphabetic order) used in rice in Africa: common name, product names, application rates, timing, target weeds, and production ecosystem Common name

Example of product

Rates (kg a.i. ha

2,4-D

Dacamine Fernoxone Herbazol

Weedone

2,4-D+ ○

Dichlorprop

Bensulfuron Bentazon Bifenox Butachlor

Londax Basagran As a mixture = Foxpro D Machete

Cinosulfuron Dymrone (K-223) Fluorodifen Glyphosate MCPA Molinate Oxadiazon

Set off 20WG Dymrone Preforan Round-up Herbit Ordram Ronstar 25EC Ronstar 12 L Gramoxone

Paraquat

1

)

Timing

Target

Ecol.

0.5–1.5

Late post

B/S

U/L

1–1.5 (l ha 1) 0.05–1.0 1.0–3.0 1.5–2.4 1.0–2.5

Post

B/S

U/L

B/S B/S B/(G) AG/(B)b

L U/La U/L U/L

0.05–0.08 3.0–5.0 2.0–3.5 1.5–3.0 0.5–1.5 1.5–4.0 0.6–1.5

Post Post Pre Pre/early post Post Pre Pre Pre/post Post Pre/early post Pre/early post

S/B S/(G/B) AB G B/S G/S/(B) G/B/S

U L U/L L U/L L U/L

0.5–1.0

Pre/post

A

L

Pendimethalin Piperophos Piperophos + ○

Cinosulfuron

○

Dimethametryne

Pretilhachlor + Propanilc

Propanil +

Bentazon ○ Triclopyr ○ Piperophos ○ Oxadiazon Quinclorac Thiobencarb Triclopyr ○

a

Stomp 500 Prowl

0.5–1.5

Pre

G/B/S

U/L

Rilof 500

0.5–2.0

Pre/early post

G/S

U/L

Pipset 35 WP Rifit extra 500 EC

1.5 1.5/0.5

Post Pre

G/S/B G/B

U U/L

2.5–4.0

Early post

A

U/L

6–8 (l ha 1) 5 (l ha 1) 1.5 5 (l ha 1) 0.25–0.5 1.5–3.0 0.36–0.48

Post Post

B/S G/S/(B)

Post Pre/post Pre/early post Post

G/B/S G G/B/S B/S

U/L U/L U U/L L U/L U

Stam F34 Propanil Surcopur Rogue

Basagran PL2 Garil Rilof S80 g l 1 Ronstar PL Facet Saturn Garlon

L, lowland; U, upland; B, broad-leaved weeds; S, Sedges; G, Grasses; A, Annuals. Weed types between brackets indicate that the product may control some species of that group or at some (early) stages. c Propanil is most often applied as a mixture with other products such as MCPA, molinate, oxadiazon, 2,4-D, fluorodifen, thiobencarb, bentazone, and butachlor. Sources: Akobundu (1987), Akobundu and Fagade (1978), Ampong-Nyarko (1996), Babiker (1982), Diallo and Johnson (1997), Grist (1968), Ishaya et al. (2007), Johnson (1997), Okafor (1986), Rijn (2001), Wopereis et al. (2007), and Zimdahl (2007). b

183

184

J. Rodenburg and D. E. Johnson

2005). Development of transgenic rice has led to two additional HR rice technologies, namely Liberty LinkÒ (compatible with glufosinate) and Roundup ReadyÒ (compatible with glyphosate) which are currently awaiting worldwide approval. HR rice technologies have the potential to control a wide range of weeds (broad leaf, grasses, and sedges) including problem weeds like Echinochloa spp. and weedy rices. Glyphosate and glufosinate are considered as relatively environmentally benign and, as postemergence herbicides, the application rates can be adjusted to the weed population (Olofsdotter et al., 1999b). In addition, the technology has a wider ‘‘window’’ for herbicide application compared to conventional technologies which is an attractive characteristic for farmers dealing with labor peaks (Olofsdotter et al., 1999b). A recent case study on the potential economic impact of HR rice in the irrigated rice production systems in Senegal, pointed out that farmers could substantially gain from access to these technologies (Demont et al., 2009). The authors concluded that introduction of HR rice should be combined with farmer training on the proper use of it to assure the long-term effectiveness. They also identified potential institutional constraints to introduction of this technology, such as the existing subsidy arrangements on chemical input and seed, which would result in very small marginal profits for commercial seed industries, which in turn would discourage private investments in the required biotechnology capacity. Establishment of effective public–private partnerships might therefore be a precondition to transfer of this technology (Demont et al., 2009). Despite the possible attractions of HR options, there are concerns regarding the likelihood of gene flow from HR rice to wild and weedy rice species. If HR rice is to be grown in close proximity to wild and weedy rice populations with overlapping periods of flowering the question has been raised as to how quickly fitness-enhancing transgenes will accumulate in these populations and whether unwanted environmental consequences will result from this (Chen et al., 2004; Lu and Snow, 2005). Field studies in the USA with HR rice (ClearfieldÒ ) have shown that there is outcrossing to red rice (O. sativa) resulting in resistant plants (Shivrain et al., 2007). A further concern is the evolution of herbicide tolerance or resistance in other weeds, which has widely occurred in rice systems (Rao et al., 2007), due to the repeated use of the same herbicide. The ability to control problem weed species efficiently makes HR rice an attractive technology and farmers may rapidly adopt it in many cases. The above considerations regarding gene-flow also suggest, however, that the reliance on HR technology for effective weed control in rice is likely to have a limited life, at a particular location, unless its introduction and use are carefully managed.

Weed Management in Rice-Based Cropping Systems in Africa

185

3.6. Integrated weed control IWM describes the integration of multiple control options based on the knowledge of weed biology and ecology, with other crop management practices (Sanyal et al., 2008). IWM may combine preventive measures with interventions, and short-term with long-term approaches, to sustainably reduce yield losses due to weeds. It is opined that this can contribute to reductions in input expenses and to the robustness of long-term weed management (Swanton and Weise, 1991). Rice cropping systems in Africa may often be suitable for integrated approaches to pest management. Farmers are often constrained by a lack of finance, information, and inputs and therefore are often reliant on traditional methods. Weed management practices based on cultural and integrated approaches may be more compatible with farmers’ resources than single-component technologies requiring high levels of external inputs ( Johnson, 1995). In Coˆte d’Ivoire, in lowland rice fields with poor water control, options such as transplanting of young seedlings and timely weed control interventions made investments in additional herbicides or improved water control less urgent (Becker and Johnson, 1999a). Combining a weed-suppressive genotype with an optimum seeding rate (e.g., 300 viable seeds m 2) can improve weed management (Zhao et al., 2007). Rice varieties may play an important role in an integrated approach, and besides improving weed competitiveness, plant breeding can contribute to better weed management through the development of shorter duration (90–100 days) rice varieties. These varieties may allow farmers to grow two crops a year and open up the possibility of introducing a weed suppressive fallow legume into the rotation (Balasubramanian et al., 2007). Short crop cycles allow crop diversification (rotations) which may improve weed management. Prerequisite to such approach is effective early weed control as short duration varieties which have little time to recover from early competition. Integrated approaches are particular useful to control weedy and wild rice in rice cropping systems in Africa. For instance, dry season tillage and the stale seedbed method using rotary cultivation can be used for the management of the perennial O. longistaminata ( Johnson et al., 1999). Farmers in Mali, when confronted with heavy infestations of O. longistaminata in lowland rice, were observed to burn rice straw in their fields right after crop harvest followed by thorough plowing prior to the next rainy season to destroy rhizomes (M. Dembele, personal communication). Manual weeding in addition to herbicides, preirrigation, the use of clean seed, transplanting in a standing water layer, and crop rotations were used by Senegalese rice farmers in fields with heavy infestations of wild rice (Diallo, 1999). Other examples of integrated practices in African rice systems are zero- or reduced tillage combined with herbicides (Kegode et al., 1999) and ‘‘under-water mowing’’ of O. longistaminata during the fallow periods in Mali (Nyoka, 1983).

186

J. Rodenburg and D. E. Johnson

4. Emerging Weed Problems and Weed Management Issues 4.1. Likely effects of demography on weed management Future weed management issues will be influenced by changing human and environmental factors, and perhaps most prominent among these will be demographic and climate changes. Africa has a higher population growth than any other continent. With a current annual average growth rate of 2.2% (compared to the world’s rate of 1.2%) the population is expected to increase by 170% between 2005 and 2030 (UN, 2007), hence there is a growing need to produce more food. As a consequence of population growth and changing consumption patterns, the rice area in Africa is increasing (Table 1) although, at the same time, rural to urban migration is causing labor shortages in some rural rice growing areas. These changes are likely to cause a shift away from labor-intensive practices and to favor the implementation of direct-seeding (Rao et al., 2007), increased herbicide applications, reduced soil tillage operations, and increased cropping intensities. As a result, there are likely to be shifts in the composition in weed flora, a greater risk of herbicide resistance, and an increase in the incidence of challenging weed problems.

4.2. Effect of changing climates on weed management There are different scenarios projected for future climate changes (Biasutti et al., 2008; Giannini et al., 2008), but whatever the changes, they are likely to affect weed problems in rice in Africa. Major global changes may comprise a rise in atmospheric greenhouse gases and an increase (>0.2 C decade 1) in temperature (IPCC, 2007). Trends suggest that the variability of rainfall will increase and the monsoon regions may become drier (Giannini et al., 2008) leading to a 5–8% increase in drought prone area in the Sahel and southern Africa by 2080 (IPCC, 2007). Equatorial zones of Africa may receive more intense rainfall (Christensen et al., 2007). The spatial distribution of future rainfall remains uncertain, however (Giannini et al., 2008), particularly for the Sahel for which there are a number of contrasting predictions (Biasutti et al., 2008; Cook and Vizy, 2006; Held et al., 2005; Hoerling et al., 2006). It has been suggested that higher temperatures alone may already be responsible for 10–40% yield losses (Tubiello et al., 2000). Combined with heavy precipitation events and increased frequency of droughts, rising temperatures are estimated to cause yield reductions of up to 50% in rain-fed agriculture in Africa by 2020 (IPCC, 2007). Increased atmospheric CO2 levels may have different consequences for species depending on their photosynthetic pathways (C3 vs C4). For a C3 crop like rice, elevated CO2 levels may have positive effects on crop growth rates, resource-use

Weed Management in Rice-Based Cropping Systems in Africa

187

efficiency, competitiveness with C4 weeds, and tolerance to Striga infection (Fuhrer, 2003; Patterson et al., 1999; Watling and Press, 2000). Many weed species in rice production systems in Africa, however, have the C3 pathway and so are likely to be favored by these changes (Table 3, adapted with information from Elmore and Paul, 1983). With elevated atmospheric CO2 levels, both C3 and C4 grasses showed increased biomass production, C3 species had a greater increase in tillering while C4 species had a greater increase in leaf area (Wand et al., 1999). Tillering and leaf canopy development are likely to be important traits affecting interspecific competition, and therefore, change in the outcome of this is anticipated. Increased CO2 levels are likely to be accompanied by higher temperatures favoring C4 weeds over C3 crops, accelerating plant development and increasing crop water consumption (Fuhrer, 2003). Drought (Liu et al., 2006) or increased temperatures ( Jagadish et al., 2007), combined with elevated CO2 levels (Matsui et al., 1997), may increase spikelet sterility in rice and consequently reduce crop yields. Certain weed species are likely to be better adapted to these environmental changes and, for instance, dry conditions are found to favor C4 weeds (Bjorkman, 1976). The net effect of changing climates on weeds is uncertain (Tubiello et al., 2007) and due to the interrelated factors difficult to predict. The outcome will depend on the species involved (Ziska, 2008), the photosynthetic pathways and the interaction effects between CO2, temperature, and water availability (Patterson et al., 1999). Changes in precipitation patterns will significantly impact crops (Tubiello et al., 2007) and weeds. Temperature will affect the geographic ranges of weeds (Patterson et al., 1999), with some species moving to higher latitudes (Patterson, 1995) and altitudes (Parmesan, 1996). Such changes are, for instance, likely for Sahelian species (Wittig et al., 2007). Witchweeds (Striga spp.) may extend their range as a result of climate change (Mohamed et al., 2006). Parasitic weeds of the Orobanchaceae family (including Striga spp.) have a wide host range with high genetic variability enabling rapid adaptation to changing environments (Kroschel, 1998), production systems, and control methods. Parasitic weeds that thrive in erratic and low rainfall environments (e.g., S. hermonthica) or temporary flooded conditions (e.g., R. fistulosa) could be favored by future climate extremes. Striga spp. problems are associated with low soil fertility (Ejeta, 2007; Kroschel, 1998; 1999; Vogt et al., 1991), and hence if climate extremes indeed lead to greater soil degradation in Africa (IPCC, 2007) this might favor parasitic weeds. Aspects of climate change that will have the greatest effect on parasitic weeds are, as yet, however unknown. S. asiatica has been found to be relatively insensitive to temperature (Patterson et al., 1982) and distribution may be affected by changes in the geographic range of the host crop rather than directly by temperature (Cochrane and Press, 1997). Phoenix and Press (2005) argued that this could be true for parasitic weeds in general.

188

J. Rodenburg and D. E. Johnson

Water is becoming a scarcer resource in many parts of Africa (Seckler et al., 1999; UNDP, 2007) and rice varieties and cropping methods need to be adapted accordingly (Ingram et al., 2008). The challenges are likely to differ for upland and lowland rice production (Manneh et al., 2007). For upland rice, drought tolerance will be important not just to reduce losses due to moisture stress but also, under stress, rice may become less competitive with weeds (Asch et al., 2005). In irrigated rice, approaches to conserve irrigation water, such as aerobic rice and alternate wetting and drying (AWD), which is also an integral part of the system of rice intensification (SRI), may be adopted but these are likely to have consequences for weed management. Haden et al. (2007) observed weed populations to shift to an increased incidence of sedges under the reduced flooding regimes of the SRI. As the season-long flooding of lowland rice fields is replaced by only temporary flooding or aerobic conditions, increased weed infestations are anticipated (e.g., Morita and Kabaki, 2002). Hand weeding requirements increased by up to 35% with temporary rather than permanent flooding in lowland systems (Latif et al., 2005). Maintaining the flooding to suppress weeds is likely to be increasingly difficult in many areas as water becomes scarcer however and, as a consequence, farmers lacking the means for effective weeding are likely to suffer severe yield losses (Barrett et al., 2004). Increased temperatures affect herbicide persistence in the soil and the ‘‘windows’’ for herbicide effectiveness (Bailey, 2004). Extreme weather may affect the risk of herbicide by either causing crop damage or by reducing the efficacy (Patterson et al., 1999). With high rainfall events, for instance, herbicides may be diluted and cease to be effective (e.g., Kanampiu et al., 2003). High CO2 environments may increase belowground plant growth relative to aboveground shoot growth (Ziska, 2003) and favor root, rhizome and tuber growth of (in particular C3) perennial weeds (Oechel and Strain, 1985) rendering their control more difficult (Patterson, 1995; Patterson et al., 1999). Increased tillage, for instance, could then lead to a multiplication of vegetative propagation material (Ziska, 2008). For rice production in Africa, this could mean increasing problems with perennial weeds like the grasses O. longistaminata and Leersia hexandra in the lowlands. Other perennial weeds with difficult to control belowground structures (e.g., I. cylindrica, Cynodon dactylon, Cyperus esculentus and C. rotundus on upland and hydromorphic soils and Bolboschoenus maritimus in the lowlands) are all of the C4 type.

4.3. Herbicide resistance in weeds There are few confirmed reports of weeds in Africa that have evolved resistance to herbicides (ISHRW, 2008) and in none of these cases are rice implicated. In Egypt, Conyza bonariensis (Hairy fleabane) is reported to be resistant to paraquat, but this weed is of no significance in rice. Rice

Weed Management in Rice-Based Cropping Systems in Africa

189

farmers in Africa do however indicate that herbicides have become less effective against certain species. For instance, propanil has been observed to be less effective to control E. colona in Senegal than in the past (Haefele et al., 2000). A survey of rice farms in Benin in 2006 revealed disadoption of a once popular herbicide (Garil, a mix of propanil and triclopyr) as it became less efficient (Kouazounde´, 2006). Further investigation indicated that the herbicide had become ineffective against D. horizontalis perhaps as a result of propanil resistance. Resistance may be accelerated by the continuous use of a single product which in turn can be a consequence of the limited range of products available on the local market; a common constraint to small farmers in Africa. Despite the absence of confirmed resistance to herbicides in weeds of rice in Africa, the problem may already exist and is a threat for the near future. The few active weed scientists in Africa often lack facilities and resources to validate herbicide resistance. Little information is available to farmers and the choice of herbicides to manage the risks of resistance is not available. Herbicide use is expected to increase in the near future and, with it, resistance is likely to develop. Environmental changes can accelerate this and, for example, raised CO2 levels have been shown to increase the tolerance of weeds to herbicides (Ziska et al., 1999; Ziska and Teasdale, 2000). Reasons behind this effect are not clear, however, rising CO2 levels may alter transpiration, reduce the number of leaf stomata or alter the thickness of the weed leaf, and thereby reduce the absorption or uptake of the pesticide (Ziska, 2008). Increased leaf starch concentrations caused by elevated CO2, as found in C3 plants (Wong, 1990), might also affect herbicide activity (Patterson et al., 1999). Following with earlier suggestions (Haefele et al., 2000), the emergence of resistant weed populations in rice production systems in Africa needs to be monitored in areas where farmers are reporting herbicides to be less effective. In this way timely and effective coping strategies for farmers could be identified and introduced.

5. A Strategic Vision for Weed Management and Research in African Rice Production Systems 5.1. Weed management strategy Levels of literacy among rice farmers in Africa tend to be low which limit the options for effective information transfer to farmers. Furthermore, there are institutional constraints ranging from absent or malfunctioning commodity markets, seed-supply systems, agro-industries, and transport facilities to unfavorable subsidy and trade arrangements. Other important constraints

190

J. Rodenburg and D. E. Johnson

to weed control, typical of rice farms in Africa, are the limited access to capital intensive inputs and information, and a lack of time or manpower. Weed problems are particularly severe where controlled flooding for weed control is not an option, such as in rain-fed cropping systems. In Africa, this is the case for roughly two thirds of the rice growing area. Even in irrigated systems rice farmers may have limited control over water (Kent and Johnson, 2001). Rice farmers in Africa would be best served by effective weed control options that do not require substantial labor, that are affordable, easy to learn and apply and that are relatively independent of markets and (agro-) industries and of the level of water control. Effective weed management strategies are likely to build on knowledge on weed ecology, biology, competition mechanisms, and the effectiveness of control methods. Rather than achieving weed-free fields, the emphasis should be on optimizing resource use which may result in minor rice yield losses being incurred. Problematic weed species may be targeted with preventive management practices, back stopped with strategically timed weeding interventions. Some considerations for appropriate weed management strategies are outlined below. 5.1.1. Prioritization of weed species A range of wild species has value as medicinal, food, or other functions and is therefore not controlled (e.g., Hillocks, 1998). Moreover, in each rice production ecosystem, there are usually only a few problem weeds in terms of the severity of competition or difficulty of control ( Johnson and Kent, 2002). Targeting the limited number of problem weed species, rather than all species occurring in a field, could help reduce labor requirements without compromising farm profits. Certain weed species are likely to become significant constraints as cropping systems intensify. The emergence of a particular species as dominant weeds will be largely influenced by the environment and especially seasonal soil moisture regimes, and will be likely to reflect the crop management. E. colona, Digitaria spp., R. cochinchinensis, I. cylindrica, Dactyloctenium aegyptium, and Eleusine indica are considered to be some of the most important grass weeds in rice worldwide (Moody, 1991), and they all occur in Africa. The perennial sedges C. esculentus and C. rotundus will be favored where the systems intensify, particularly where there is regular soil tillage. There is also evidence from Asia that C. rotundus has become adapted to wetter environments (Pena-Fronteras et al., 2008). E. heterophylla is a broadleaved weed capable of rapid growth and multiplication that can become a serious problem in upland rice-based rotations, and Striga spp. and R. fistulosa can cause serious losses in certain environments. In the lowland rice systems it is likely that some of the same weeds that pose problems elsewhere in the world (Rao et al., 2007) will be an increasing constraint in Africa. These will likely include the annual grasses

Weed Management in Rice-Based Cropping Systems in Africa

191

Echinochloa spp., L. chinensis, and I. rugosum and the annual sedges Fimbristylis spp., C. difformis, and Cyperus iria among others. A number of these species are already established problems in Africa. Successful management of these will likely follow the application of practices that integrate cultural measures of land preparation, water management and rotations, and, especially in the direct-seeded systems, the judicious use of herbicides. Weed management strategies for particular species may consider the ecology and biology of the species. Certain management practices (e.g., seed burial or submergence depth due to soil tillage or flooding) are likely to provoke differential responses among the key weed species and this can be employed as a basis to develop more sustainable management practices. Many species are capable of prolific seed production (e.g., Echinochloa spp., Striga spp.) and reducing the number of weeds that produce seed or multiply after harvest can contribute to decreased weed problems in the subsequent seasons (e.g., Diallo and Johnson, 1997; Haefele et al., 2000). As an example, if environmental conditions are favorable, S. hermonthica continues to reproduce after crop harvest, contributing considerably to the total seed production (Rodenburg et al., 2006a). To prevent a buildup of the weed seed bank, a preflowering or a postharvest weed control operation may be required (e.g., Diallo and Johnson, 1997; Haefele et al., 2000). Farmers however may have other priorities after crop harvest or may not be aware of the longer term impact on weed populations. Greater understanding of species biology can contribute to weed management in different ways. Obligate hemiparasitic weeds like Striga spp., for instance, require a suitable host for survival and development after germination. Infestation levels of these parasitic weeds can therefore be reduced through rotations with nonhost crops. ‘‘Trap crops’’ like cowpea, bean, soybean, yellow gram, groundnut, or bambara groundnut (e.g., Oswald et al., 2002) are particularly effective as these provoke seed germination of Striga spp., but do not support subsequent infection and development of the parasite. Facultative hemiparasitic weeds, such as R. fistulosa and Buchnera hispida, require a different approach however as they are fairly independent of the presence of a suitable host to develop and multiply. Seeds of R. fistulosa and B. hispida require sunlight for germination (Nwoke and Okonkwo, 1980; Ouedraogo et al., 1999) and it is possible that a cover crop or mulch might prevent these from germinating. Perennial wild rice, O. longistaminata, has underground rhizomes that enable it to survive superficial weed control operations especially when the soil is moist, but deep tillage in the dry season brings rhizomes to the surface where they desiccate and die (e.g., Johnson, 1997; Johnson et al., 1999). Annual wild rice such as O. barthii, on the other hand, reproduce only through seed and their control requires the use of rice seed free of wild rice as a contaminant and hand rouging of wild rice in the field before seed setting and shattering occurs (Delouche et al., 2007). Knowledge of seed germination ecology of

192

J. Rodenburg and D. E. Johnson

different weed species with relevance to rice production systems in Africa (e.g., Chauhan and Johnson, 2008a–d) is useful for the design of targeted control options, such as tillage practices or crop residue management. The development of knowledge-based technologies is often hampered by limited understanding of weed ecology and biology (Fernandez-Quintanilla et al., 2008). There are also large gaps between ‘‘scientific understandings’’ and the information available to the farmer. Farmers often lack information on parasitic weeds for instance (Frost, 1995; Reichmann et al., 1995), and, in this respect, women often have less access to information than men (Gehri et al., 1999). This lack of available information is likely to limit implementation by farmers of any control methods that are developed. 5.1.2. Integrated crop and weed management approaches Weed management strategies that are financially, socially, and environmentally cost effective are only likely to be achieved through integrated approaches. Such approaches will combine crop management practices to achieve good crop establishment with optimum plant population densities and vigorous crop growth together with weed management options that prevent, suppress, and control weeds. The choice of management practices is largely predetermined by the type of production environment and system. Good crop establishment and vigorous growth can in part be achieved through good land preparation, adequate soil leveling to the best of farmers’ means, use of weed-free rice seeds of good quality, transplanting of young seedlings or sowing of pregerminated seeds in rows, maintaining soil flooding until maximum tillering, and split fertilizer applications (e.g., Becker and Johnson, 1999a, 2001b). Complementary weed management can include one or more of the following components:

Weed competitive or parasitic weed resistant and tolerant crop varieties (e.g., Fofana and Rauber, 2000; Johnson et al., 1997) Crop rotations with a noncereal crop (e.g., Oswald et al., 2002; Sengupta et al., 1985) Weed-suppressive fallows (e.g., Akanvou et al., 2000; Merkel et al., 2000) Weed-suppressive mulches (e.g., Iwuafor and Kang, 1993; Kamara et al., 2000) Postharvest weeding to prevent buildup of weed seed bank (e.g., Diallo and Johnson, 1997; Haefele et al., 2000) Increased plant densities and improved plant arrangements (e.g., Akobundu and Ahissou, 1985; Phuong et al., 2005)

5.1.3. Timing of weed control and crop management interventions Timing is critical to effective integration of crop and weed management. For instance, correct rates and timing of fertilizer provides adequate nutrients without a surplus soon after rice sowing or planting which would encourage

Weed Management in Rice-Based Cropping Systems in Africa

193

weed growth (e.g., Liebman and Davis, 2000). Another example is timing of transplanting; careful transplanting of young rice seedlings reduces transplanting shock and results in better crop establishment and a more competitive crop as compared to older seedlings (e.g., Poussin, 1997). Timing is imperative for effective weed control interventions. Herbicides, for instance, often have ‘‘windows’’ for application as, for example, postemergence applications usually need to be applied in the early stages of crop growth to be efficient and minimize crop damage (e.g., Haefele et al., 2000). Timing of interventions is important also with respect to crop–weed competition. Crops have critical periods during which weed competition affects yield and beyond which effects are minimal. In irrigated rice in the Sahel, this critical weed period varied markedly between the two seasons of study, but fell between 14 and 56 DAS ( Johnson et al., 2004), while in upland rice in the Guinea Savannah the critical weed period assessed with two varieties of NERICA, fell between 21 and 42 DAS (Dzomeku et al., 2007). Few of such studies have been undertaken and there is a dearth of information on the critical periods for weed competition for other production systems, rice varieties, and weed species.

5.2. Weed research strategy Despite weeds being the most widespread biotic production constraints of rice in Africa, data on distribution and importance of specific weed species are lacking. Such data are an initial requirement for improved priority setting for weed research in the context of these production systems. Moreover, as rice systems in Africa are diverse, general recommendations tend to be flawed. Weed research therefore needs to have regional relevance while also generating outputs (e.g., technologies, knowledge) that are locally applicable and validated with farmers. Improved approaches to weed management will discriminate between the uplands, and the rain-fed and irrigated lowlands in which hydrology and degree of water control have decisive impacts on weed species and the range of applicable management options. In each environment, compiling basic knowledge on the biology and ecology of the most troublesome weeds for each ecosystem will provide insights on which management options could be developed. Some of the options discussed in this review have yet to be validated for rice systems in Africa. Changing weed populations, water, and labor shortages will be key future issues for weed research. These can be addressed by genetic and management improvements, but such developments also require a thorough understanding of underlying ecological and biological principles and interactions between crop, weeds, management options, and the environment. The following sections discuss some research topics relevant for enhancing the effectiveness of weed management practices to meet the future challenges.

194

J. Rodenburg and D. E. Johnson

5.2.1. Climate change Many studies have been carried out on the effects of CO2 enrichment on plant species (Bunce, 2005; Navas et al., 1999; Wand et al., 1999; Ziska, 2001; 2003). Fewer studies have focused on temperature rise effects on weeds (Tungate et al., 2007) or on combined effects of CO2 and temperature increases (Coleman and Bazzaz, 1992; Nonhebel, 1996; Williams et al., 2007). Fewer still peer-reviewed studies report on specific investigations of the effect of drought or water-stress on crop–weed interactions (Moffett and McCloskey, 1998). Studies have not been undertaken on the combined effects of the three main anticipated climate change causes/effects (e.g., CO2, temperature and drought) on rice-weed competition. Controlled experiments with field crops and weeds are however difficult and costly to conduct, and the absence of the resources and facilities to conduct such experiments in Africa suggests that partnerships with institutes and universities elsewhere would be advantageous. Anticipated climate changes will bring changes in species distribution and predicting these changes is an important task for weed science (Fernandez-Quintanilla et al., 2008). The anticipated changes will need to be considered in the context of emerging constraints such as demographic changes and water shortages. Integrated management options should be developed that on one hand arrest potential increased losses to weeds due to changing climatic and environmental conditions and that prevent climate change to aggravate on the other hand (Ingram et al., 2008). 5.2.2. Crop management More detailed information on the mechanisms of competition (nutrients, water, light and space) is required to provide the basis on which to develop new elements of IWM. This might include investigating effects of management practices such as the quantity, method, and timing of fertilizer application on weed development. Detailed guidelines are needed on how to optimize crop growth without unduly favoring weeds. This could derive from a better understanding of the effects of fertilizer timing on problem weeds and crop–weed competition and how this in turn might be influenced by season and rice ecosystem. More nonchemical, labor-saving weed management technologies need to be explored and local innovations validated. This might include examples such as removing flower heads of weeds as observed in Bende, Abia state, Nigeria (E. A. Maji and M. Tokula, personal communication), the combined burning and off-season dry tillage in Ze´gue´sso, Sikasso, Mali (M. Dembele, personal communication) or the application of locally produced bioherbicides in Glazoue´, Collines in Benin (Kouazounde´, 2006). Such weed control practices are typically based on cultural and integrated approaches and inherently have a high compatibility with farmers’ resources and, as such, are likely to be more successful ( Johnson, 1995).

Weed Management in Rice-Based Cropping Systems in Africa

195

Cultural weed management practices such as relay cropping or rotations with legumes have often been proved technically sound but generally face low farmer adoption rates (e.g., Tarawali et al., 1999). Participatory approaches could help develop more appropriate options and help identify and alleviate reasons for low adoption rates of otherwise effective management practices. Relay cropping, intercropping and improved fallows should have additional benefits and good adaptation to agro-climatic and ecological ranges. These were typical weaknesses of the Striga control practice using Desmodium spp., for instance (Gressel and Gebrekidan, 2007). Legume species may be tested under local on-farm conditions and more research should be conducted on residue management practices to minimize competition effects and maximize weed suppression and reduction of the weed seed bank. Suitable and effective rice-intercrop practices, improved crop residue management like mulching (e.g., Iwuafor and Kang, 1993; Singh et al., 2007) and alternative establishment methods like line sowing and plant densities (Phuong et al., 2005) merit further study in rice systems in Africa. Further, ex ante analyses of the likely impacts and farmer adoption of potential technologies are required to guide strategic decisions on the allocation of scarce research resources. Finally, interactions of water management and weeds need further study. Weed germination could be reduced if the period between sowing/planting and flooding could be shortened by several days. This can be achieved by drill-planting imbibed or pregerminated rice seeds (Counce and Burgos, 2006), land leveling to allow earlier shallow flooding, or by transplanting in (shallow) flooded fields (Poussin, 1997). Differential responses to timing and depth of flooding among species, as shown by different studies (e.g., Kent and Johnson, 2001; Mortimer et al., 2005), provide opportunities to target weed species or groups of species through better water management. Knowledge and understanding is however required on the precise limits of the depth and timing of flooding for particular species and conditions to allow establishment of rice while suppressing weeds. 5.2.3. Varietal development Studies show that varieties with high yields under weed-free conditions are also likely to have superior yields under weed competition (e.g., Lemerle et al., 2001; Zhao et al., 2006a,b, 2007). Local adaptation is therefore an important characteristic for weed competitiveness (Lemerle et al., 2001) and weed-free yield may be an efficient indirect selection trait to find germplasm capable of high yield under weed competition (Zhao et al., 2006a). Combining yield potential with the ability to reduce losses to weeds would make a valuable contribution to IWM programs. There have only been limited efforts, compared to the challenges faced, to develop locally adapted rice varieties suitable for Africa, however. A preferred approach would be to intensify research and breeding activities to enlarge the range of

196

J. Rodenburg and D. E. Johnson

available germplasm with desirable traits in addition to increased weed competiveness or resistance and tolerance to parasitic weeds. It is likely that suitable rice varieties will combine different weed competitive or suppressive traits (Dingkuhn et al., 1999), including early vigor and efficient nutrient and water uptake, together with other desirable traits such as improved resistance or tolerance to other biotic or abiotic stresses and high yields and grain quality. To realize such advances in the near term may require that available screening measures (e.g., Haefele et al., 2004; Zhao et al., 2006b) are adapted or enhanced. Indirect selection measures and novel experimental designs may be utilized to screen desired traits, but the number of easily identifiable traits responsible for increased weed competitiveness needs to be enhanced (Pester et al., 1999). A better understanding of physiological traits conferring competitiveness may help to identify the related molecular markers and suitable parents for marker-assisted breeding programs. Few studies to date have reported advanced biomolecular methods such as employed by Gurney et al. (2006) and Kaewchumnong and Price (2008) for the improvement of resistance against parasitic weeds, or such as used by Jensen et al. (2001) for the development of weed competitive rice varieties. In this respect, the gene pools of the African wild and cultivated rice species and the currently available NERICA varieties are yet to be fully explored. 5.2.4. Herbicides Herbicide formulations that can be directly applied to the irrigation water or soil, rather than foliar applied postemergence applications might be particularly suited to rice systems in Africa ( Johnson, 1995). Such formulations (e.g., granular or dry flowable) do not require spraying equipment and they also carry lower risks of herbicide contamination compared to liquid formulations (Akobundu, 1987; Zimdahl, 2007). An inexpensive but effective preemergence and slow-release granular formulation of an herbicide would have great potential for these cropping systems. A promising approach may also be to focus on herbicide seed treatment. Seed treatment has the advantage that farmers do not need to apply herbicides in the field anymore. Moreover, through seed coatings, herbicide application is well targeted and doses are therefore relatively low. This option, combined with herbicide resistant germplasm has been successful in Striga control in maize in East Africa (e.g., Kanampiu et al., 2001). A possible disadvantage of such a technology is the increased dependency on the commercial seed systems and agro-industries that this technology would require. Furthermore the approach may not be effective in lowland rice or high rainfall areas as there may be rapid leaching of the product. Conversely, phytotoxicity may occur with low rainfall (Kanampiu et al., 2003). Further development may focus on seed treatment methods that can be applied by farmers and on products

Weed Management in Rice-Based Cropping Systems in Africa

197

with low toxicity to rice and coating methods that prevent the product from quickly dissolving and leaching away. Bioherbicides based on locally abundant weed species may be an attractive alternative to chemical herbicides for rice farmers in Africa. Potentially suitable weed species for production of bioherbicides are B. pilosa and E. hirta (Hong et al., 2004) and A. conyzoides (Xuan et al., 2004). It is important however to recognize the likely institutional, economic, and physical constraints to the use of bioherbicides in many of the rice environments in Africa. 5.2.5. Weed ecology and rice ecosystems Knowledge of the biology and ecological requirements of target species will be the starting point for the development of effective, appropriate, and affordable weed control technologies. Species-specific weed management options need to account for any negative consequences on the dynamics of the weed population and competition with the rice crop (Mortensen et al., 2000) as the removal of a key species may result in undesirable population shifts in the remaining species. The dynamics following species-specific weed management is as yet poorly understood. Wild and weedy rices (Oryza spp.), perennial weed species, particularly those with extensive subterranean rhizome systems (Cyperus spp. and I. cylindrica), parasitic weeds with wide geographic ranges and high genetic variation (e.g., Striga spp.), and annual species such as Digitaria spp., Echinochloa spp., and possibly also I. rugosum and Leptochloa spp., are expected to become more important in future rice production in Africa. Such species merit further research attention. Weed research needs to focus on elucidating biology, taxonomy, and control measures of wild and weedy rices. In addition, to overcome problems of contaminated seed supplies, possibilities for the initiation of community-based seed systems should be investigated in different countries. More research efforts should be undertaken to predict and anticipate spread of parasitic weeds. Parasitic weeds like Striga spp. in uplands and hydromorphic lands and R. fistulosa in hydromorphic lands and rain-fed lowlands are likely to become more important in rice in Africa in the near future. The minor parasitic weeds of today can be the major ones tomorrow (Raynal Roques, 1994). Weed research for rice systems in Africa may focus more on rain-fed and semi-irrigated lowland ecosystems (inland valleys) where flooding cannot, or only partially, be controlled. The inland valleys comprise a huge production potential that is yet underexploited. In these areas, weeds are a major constraint as water control is poor, and the soil is fertile and either wet or moist for much of the year. Rice production in the inland valleys is mainly for subsistence (Windmeijer and Andriesse, 1993). Land preparation is mostly done by hand and fields are often inadequately bunded and leveled resulting in uneven flooding and patchy conditions favoring weed growth (Akobundu and Fagade, 1978; Ampong-Nyarko, 1996). Uncontrolled flooding also

198

J. Rodenburg and D. E. Johnson

renders the use of herbicides less effective (Akobundu, 1987). The lack of a permanent and adjustable water layer favors weed infestations, such as observed with wild rice in Tanzania (Riches et al., 2005), and leads to severe crop–weed competition. Few suitable weed control technologies are yet available for farmers in these rice production ecosystems and developing these should be a priority. Rice production in inland valleys is considered very suitable for integrated management approaches. Relatively little is known about the interactions between weeds and other biotic constraints such as birds, insects (stem borers), and pathogens like Rice Yellow Mottle Virus and African Rice Gall Midge, and in selective cases these may merit further study. In a recent survey carried out in Senegal, farmers indicated that weeds attract birds (M. Diagne and Y. de Mey, personal communication). Weeds, like E. colona or wild rice species, provide shelter and food to grain-feeding birds such as Quelea quelea (Luder, 1985; Treca, 1985; Ward, 1965) especially in the period before rice grain filling and after harvest. More effective weed control could therefore contribute to reduced bird pressure (Luder, 1985; Treca, 1985). This would however require further study. In Asia, ducks are used to control weeds in rice and were shown to reduce herbicide use without compromising farm profits (Liu et al., 2004; Men et al., 2002). In rice–fish production systems, fish (mainly common carp, Cyprinus carpio, and Nile tilapia, Oreochromis niloticus) can control weeds through direct feeding and increased water turbidity, while the permanent flooding required for fish culture also adds to weed management (Halwart, 2001). Integrations of rice with poultry or fish may need to be adapted or tested for compatibility with prevailing rice farming systems. Concepts of the management of vegetation to regulate the natural enemies of insect pests, noted above (Afun et al., 1999b; Nwilene et al., 2008), warrant further investigation, particularly with regard to the practical implications for farmers and farm management in rice-based cropping systems. 5.2.6. Socioeconomics and gender Integration of socioeconomic sciences with agronomic practice is important if the impact and relevance of outcomes of weed science is to be increased. Such integration is currently often lacking (De Groote, 2007; FernandezQuintanilla et al., 2008). Socioeconomic perspectives may improve the targeting of the appropriate weed species and be necessary for setting priorities for the development of weed technologies. Data are lacking on distribution and economic losses caused by weeds in rice production systems, and further, impact studies (ex ante or ex post) on weed control technologies are scant. Social sciences could also have a greater role in identifying potential constraints to adoption and in developing approaches to overcome these. Weed management in the rice systems of Africa is mainly carried out by women. Including all crops, women in Africa collectively spend an

Weed Management in Rice-Based Cropping Systems in Africa

199

estimated 20 billion hours a year on (hand) weeding while still suffering crop yield losses to weeds of between 20 and 100% (Gianessi, 2008). Weeding is the most labor-intensive crop operation and therefore weighs heavily on women’s time which in turn impacts other economic activities (e.g., Gehri et al., 1999). Further, women are often also accompanied by children and therefore weeds impact across the generations as this impinges on opportunities for education. Farmers’ view on the importance of weed control relative to the overall farm operations is poorly understood. Information about these farm priorities, strategies, and choices could help identify the constraints to adoption, the development of weed technologies, and farmer extension efforts. Besides their effectiveness to control weeds, for technologies to be acceptable to farmers, they need to fit in the local social context and be affordable. Involvement of female farmers in weed research is extremely important to effectively reach the target group and get their input in the design and development of suitable weed management strategies (e.g., Gehri et al., 1999). There is a great need for farmers, extension agents and scientists to exchange views and identify expertise and knowledge gaps in order to better target problem weeds, and develop improved approaches and control options. Such interactions might already improve farmers’ decision making and consequently enhance weed management, as this is highly dependent on exposure to technologies and access to information (e.g., Becker et al., 2003; Haefele et al., 2002; Rao et al., 2007). Due to the diversity of rice systems in Africa, farmers require locally adapted solutions. An example of a successful participatory approach to improve farmers’ crop management, stimulating farmer experimentation, and identifying researchable issues is the curriculum for Participatory Learning and Action Research (PLAR) for Integrated Rice Management (IRM) in inland valleys of sub-Saharan Africa as developed by WARDA and IFDC. This method consists of a technical manual (Wopereis et al., 2007) and a facilitators’ guide (Defoer et al., 2004a) containing modules on water, crop, and pest management issues. Many of the integrated rice management practices discussed throughout this curriculum contribute directly (through modules on weed recognition, IWM, and the use of herbicides) or indirectly (through modules on land preparation, transplanting, and water management) to improved weed management. This approach could be expanded with modules for example on weed biology and ecology, rice-weed competition, and improved weed management. Currently, PLAR-IRM modules are converted into training videos in various local languages (WARDA, 2008). This medium ensures an easier and probably cheaper means of technology transfer than traditional extension, leading to rapid and massive dissemination among rice farmers (Van Mele, 2006). PLAR-IRM can contribute to more efficient and sustainable weed management in rice for resource-poor farmers and merits promotion and wider application through diverse media.

200

J. Rodenburg and D. E. Johnson

6. Concluding Remarks Critical research issues are implicated in the challenge to realize the potential productivity of some underutilized areas, such as the inland valleys of West Africa. For rice systems to be sustainable, ecological approaches to weed management must be applied. Components of such approaches may comprise effective land preparation and establishment of a competitive crop, managing flooding at critical stages in crop or weed growth, depletion of the soil seed bank, minimizing the ingression of undesirable species, and timely interventions against weeds that escaped the preventive measures. The implementation of such knowledge-based systems may build on experiences and results gathered from elsewhere and may also require adaptive research with farmers to identify and address local constraints and successful application of the options. Such initiatives are a likely prerequisite to the sustainable development of these underexploited areas. Approaches to improving weed management in rice farming systems in Africa can benefit from extending activities through participatory farmer learning and research activities. Farmers often lack knowledge on weed biology and control while at the same time, as daily practitioners, they may possess traditional knowledge that could provide useful insights for the development of sustainable management options. Enhancing exchange between farmers and scientist could lead to greater knowledge on some of the most important or troublesome species and management practices and is expected to achieve substantial gains. Parasitic weeds (in uplands) and weedy and wild rices (in lowlands) would be suitable subjects for pilot projects on participatory farmer training on weed biology and control. Finally, the economic importance of weeds in African rice production systems ($1.45 billion a year in addition to costs of weed control) is currently not reflected in the resources dedicated to reducing these losses and improving weed management. It is envisaged that weed research could generate a considerable impact on the lives of the rural poor, and economies of developing countries through the development of knowledge-based management practices. To realize this requires long-term investments in human and financial resources focused on the development of more productive farming systems.

ACKNOWLEDGMENTS Many of the insights presented here were gained in the course of implementation of various research and development projects. We thank our main donors, the Common Fund for Commodities (CFC), the Dutch Directorate-General for International Cooperation (DGIS), the Netherlands Foundation for the Advancement of Tropical Research (NWOWOTRO), and the Department for International Development (DFID), for their past and present financial support.

Weed Management in Rice-Based Cropping Systems in Africa

201

REFERENCES Abate, T., Van Huis, A., and Ampofo, J. K. O. (2000). Pest management strategies in traditional agriculture: An African perspective. Annu. Rev. Entomol. 45, 631–659. Abbasher, A. A., Kroschel, J., and Sauerborn, J. (1995). Micro-organisms of Striga hermonthica in northern Ghana with potential as biocontrol agents. Biocontrol Sci. Technol. 5, 157–161. Abo, M. E., Alegbejo, M. D., Sy, A. A., and Sere, Y. (2002). Field incidence of rice yellow mottle virus, genus sobemovirus on rice and a weed species in the field in Coˆte d’Ivoire. Acta Agron. Hung. 50(1), 85–89. Adagba, M. A., Lagoke, S. T. O., and Imolehin, E. D. (2002a). Nitrogen effect on the incidence of Striga hermonthica (Del.) Benth in upland rice. Acta Agron. Hung. 50(2), 145–150. Adagba, M. A., Lagoke, S. T. O., Singh, B. N., and Ndahi, W. B. (2002b). Potential of cinosulfuron and CGA152005 seed treatment for control of Striga hermonthica in upland rice. Acta Agron. Hung. 50(1), 7–18. Adesina, A. A., Johnson, D. E., and Heinrichs, E. A. (1994). Rice pests in the Ivory Coast, West Africa—Farmers perceptions and management strategies. Int. J. Pest Manag. 40(4), 293–299. Aflakpui, G. K. S., Bolfrey-Arku, G. E. K., Anchirinah, V. M., Manu-Aduening, J. A., and Adu-Tutu, K. O. (2008). Incidence and severity of Striga spp. in the coastal savanna zone of Ghana Results and implications of a formal survey. Outlook Agric. 37, 219–224. Afun, J. V. K., Johnson, D. E., and Russell-Smith, A. (1999a). The effects of weed residue management on pests, pest damage, predators and crop yield in upland rice in Coˆte d’Ivoire. Biol. Agric. Hortic. 17(1), 47–58. Afun, J. V. K., Johnson, D. E., and Russell-Smith, A. (1999b). Weeds and natural enemy regulation of insect pests in upland rice; A case study from West Africa. Bull. Entomol. Res. 89, 391–402. Ahmed, N. E., Sugimoto, Y., Babiker, A. G. T., Mohamed, O. E., Ma, Y., Inanaga, S., Nakajima, H., and Ma, Y. Q. (2001). Effects of Fusarium solani isolates and metabolites on Striga germination. Weed Sci. 49(3), 354–358. Ahonsi, M. O., Berner, D. K., Emechebe, A. M., and Lagoke, S. T. (2002). Selection of rhizobacterial strains for suppression of germination of Striga hermonthica (Del.) Benth. seeds. Biol. Control 24(2), 143–152. Ahonsi, M. O., Berner, D. K., Emechebe, A. M., and Lagoke, S. T. (2004). Effects of ALSinhibitor herbicides, crop sequence, and fertilization on natural soil suppressiveness to Striga hermonthica. Agric. Ecosyst. Environ. 104(3), 453–463. Akanvou, R., Becker, M., Chano, M., Johnson, D. E., Gbaka-Tcheche, H., and Toure´, A. (2000). Fallow residue management effects on upland rice in three agroecological zones of West Africa. Biol. Fertil. Soils 31, 501–507. Akanvou, R., Bastiaans, L., Kropff, M. J., Goudriaan, J., and Becker, M. (2001). Characterization of growth nitrogen accumulation and competitive ability of six tropical legumes for potential use in intercropping systems. J. Agron. Crop Sci. 187(2), 111–120. Akanvou, R., Kropff, M. J., Bastiaans, L., and Becker, M. (2002). Evaluating the use of two contrasting legume species as relay intercrop in upland rice cropping systems. Field Crops Res. 74(1), 23–36. Akobundu, I. O. (1980). Weed science research at the International Institute of Tropical Agriculture and research needs in Africa. Weed Sci. 28(439), 445. Akobundu, I. O. (1981). Weed control in direct-seeded lowland rice under poor water control conditions. Weed Res. 21(6), 273–278. Akobundu, I. O. (1987). Weed Science in the Tropics—Principles and Practices. Wiley, Chichester.

202

J. Rodenburg and D. E. Johnson

Akobundu, I. O., and Ahissou, A. (1985). Effect of interrow spacing and weeding frequency on the performance of selected rice cultivars on hydromorphic soils of West Africa. Crop Prot. 4(1), 71–76. Akobundu, I. O., and Fagade, S. O. (1978). Weed problems of African rice lands. In ‘‘Rice in Africa’’ (I. W. Buddenhagen and G. J. Persley, Eds.), pp. 181–192. Academic Press, London. Akobundu, I. O., Ekeleme, F., and Chikoye, D. (1999). Influence of fallow management systems and frequency of cropping on weed growth and crop yield. Weed Res. 39, 241–256. Alegre, J. C., and Cassel, D. K. (1996). Dynamics of soil physical properties under alternative systems to slash-and burn. Agric. Ecosyst. Environ. 58(39), 48. Altieri, M. A., Wilson, R. C., and Schmidt, L. L. (1985). The effects of living mulches and weed cover on the dynamics of foliage-arthropod and soil-arthropod communities in three crop systems. Crop Prot. 4, 201–213. Ampong-Nyarko, K. (1996). Weed management in rice in Africa. In ‘‘Weed Management in Rice’’ (B. A. Auld and K. U. Kim, Eds.), pp. 183–191. FAO, Rome. Ampong-Nyarko, K., and De Datta, S. K. (1991). A Handbook for Weed Control in Rice. IRRI, Los Banos. Andrews, F. W. (1945). The parasitism of Striga hermonthica Benth. on Sorghum spp. under irrigation. I. Preliminary results and the effect of heavy and light irrigation on Striga attack. Ann. Appl. Biol. 32, 193–200. Anthofer, J., and Kroschel, J. (2007). Effect of Mucuna fallow on weed dry matter and composition in succeeding maize. Biol. Agric. Hortic. 24(4), 397–414. Asch, F., Sow, A., and Dingkuhn, M. (1999). Reserve mobilization, dry matter partitioning and specific leaf area in seedlings of African rice cultivars differing in early vigor. Field Crops Res. 62(2–3), 191–202. Asch, F., Dingkuhn, M., Sow, A., and Audebert, A. (2005). Drought-induced changes in rooting patterns and assimilate partitioning between root and shoot in upland rice. Field Crops Res. 93(2–3), 223–236. Babiker, A. G. T. (1982). Chemical weed control in irrigated direct-seeded rice in the Sudan Gezira. Weed Res. 22(2), 117–121. Bailey, S. W. (2004). Climate change and decreasing herbicide persistence. Pest Manag. Sci. 60(2), 158–162. Balasubramanian, V., Sie, M., Hijmans, R. J., and Otsuka, K. (2007). Increasing rice production in sub-Saharan Africa: Challenges and opportunities. Adv. Agron. 94, 55–133. Barrett, C. B., Moser, C. M., Mchugh, O. V., and Barison, J. (2004). Better technology, better plots, or better farmers? Identifying changes in productivity and risk among Malagasy rice farmers. Am. J. Agric. Econ. 86(4), 869–888. Bastiaans, L., Kropff, M. J., Kempuchetty, N., Rajan, A., and Migo, T. R. (1997). Can simulation models help design rice cultivars that are more competitive against weeds?. Field Crops Res. 51(1–2), 101–111. Becker, M., and Johnson, D. E. (1998). Legumes as dry season fallow in upland rice-based systems of West Africa. Biol. Fertil. Soils 27(4), 358–367. Becker, M., and Johnson, D. E. (1999a). Rice yield and productivity gaps in irrigated systems of the forest zone of Coˆte d’Ivoire. Field Crops Res. 60(3), 201–208. Becker, M., and Johnson, D. E. (1999b). The role of legume fallows in intensified upland rice-based systems of West Africa. Nutr. Cycl. Agroecosys. 53(1), 71–81. Becker, M., and Johnson, D. E. (2001a). Cropping intensity effects on upland rice yield and sustainability in West Africa. Nutr. Cycl. Agroecosys. 59(2), 107–117. Becker, M., and Johnson, D. E. (2001b). Improved water control and crop management effects on lowland rice productivity in West Africa. Nutr. Cycl. Agroecosys. 59(2), 119–127.

Weed Management in Rice-Based Cropping Systems in Africa

203

Becker, M., Johnson, D. E., Wopereis, M. C. S., and Sow, A. (2003). Rice yield gaps in irrigated systems along an agro-ecological gradient in West Africa. J. Plant Nutr. Soil Sci. 166(1), 61–67. Belz, R. G. (2007). Allelopathy in crop/weed interactions—An update. Pest Manag. Sci. 63(4), 308–326. Biasutti, M., Held, I. M., Sobel, A. H., and Giannini, A. (2008). SST forcings and Sahel rainfall variability in simulations of the twentieth and twenty-first centuries. J. Climate 21(14), 3471–3486. Biswas, J. C., and Sattar, S. A. (1991). Effect of nitrogen uptake by weeds on rice yield. Int. Rice Res. Newsl. 16(5), 26. Bjorkman, O. (1976). Adaptive and genetic aspects of C4 photosynthesis. In ‘‘Metabolism and Plant Productivity’’ (R. H. Burris and C. C. Black, Eds.), pp. 287–309. University Park Press, Baltimore, MD. Both, J. (2006). Local community perception of the conservation, management and use of wild plant biodiversity in inland valleys of Benin and Togo. M.Sc. Thesis, Wageningen University, Wageningen, The Netherlands. Bouriquet, G. (1933). Une Scrophulariace´e parasite du riz a` Madagascar. Rev. Pathol. Ve´g. Entomol. Agric. 20, 149–151. Braimoh, A. K. (2006). Random and systematic land-cover transitions in northern Ghana. Agric. Ecosyst. Environ. 113(1–4), 254–263. Buddenhagen, I. W., and Bidaux, J. M. (1978). Striga root parasite—A problem of dryland rice in the African savanna environment. Int. Rice Res. Newsl. 3(3), 22. Budelman, A. (1988). The performance of the leaf mulches of Leucaena leucocephala, Flemingia macrophylla and Gliridia sepium in weed control. Agroforest. Syst. 6, 137–145. Bunce, J. A. (2005). Seed yield of soybeans with daytime or continuous elevation of carbon dioxide under field conditions. Photosynthetica 43(3), 435–438. Burkill, H. M. (2004). The Useful Plants of West Tropical Africa. Royal Botanic Gardens, Kew. Carsky, R. J., Singh, L., and Ndikawa, R. (1994a). Effect of herbicide and handweeding on current and subsequent season Striga hermonthica density on sorghum. Int. J. Pest Manag. 40(2), 111–116. Carsky, R. J., Singh, L., and Ndikawa, R. (1994b). Suppression of Striga hermonthica on sorghum using a cowpea intercrop. Exp. Agric. 30(3), 349–358. Carsky, R. J., Berner, D. K., Oyewole, B. D., Dashiell, K., and Schulz, S. (2000). Reduction of Striga hermonthica parasitism on maize using soybean rotation. Int. J. Pest Manag. 46(2), 115–120. Carson, A. G. (1989). Effect of intercropping sorghum and groundnuts on density of Striga hermonthica in the Gambia. Trop. Pest Manag. 35(2), 130–132. Catling, D. (1992). Rice in Deep Water. MacMillan Press Ltd, London. Cechin, I., and Press, M. C. (1994). Influence of nitrogen on growth and photosynthesis of a C-3 cereal, Oryza sativa, infected with the root hemiparasite Striga hermonthica. J. Exp. Bot. 45, 925–930. Chauhan, B. S., and Johnson, D. E. (2008a). Germination ecology of goosegrass (Eleusine indica): An important grass weed of rainfed rice. Weed Sci. 56(5), 699–706. Chauhan, B. S., and Johnson, D. E. (2008b). Germination ecology of southern crabgrass (Digitaria ciliaris) and India crabgrass (Digitaria longiflora): Two important weeds of rice in tropics. Weed Sci. 56(5), 722–728. Chauhan, B. S., and Johnson, D. E. (2008c). Germination ecology of two troublesome asteraceae species of rainfed rice: Siam weed (Chromolaena odorata) and coat buttons (Tridax procumbens). Weed Sci. 56(4), 567–573. Chauhan, B. S., and Johnson, D. E. (2008d). Influence of environmental factors on seed germination and seedling emergence of eclipta (Eclipta prostrata) in a tropical environment. Weed Sci. 56(3), 383–388.

204

J. Rodenburg and D. E. Johnson

Chauhan, B. S., and Johnson, D. E. (2008e). Germination ecology of Chinese Sprangletop (Leptochloa chinensis) in the Philippines. Weed Sci. 56(6), 820–825. Chen, L. J., Lee, D. S., Song, Z. P., Suh, H. S., and Lu, B. R. (2004). Gene flow from cultivated rice (Oryza sativa) to its weedy and wild relatives. Ann. Bot. 93(1), 67–73. Chikoye, D., Ekeleme, F., and Ambe, J. T. (1999). Survey of distribution and farmer perceptions of speargrass (Imperata cylindrica) in cassava based cropping systems in West Africa. Int. J. Pest Manag. 45, 305–312. Chikoye, D., Manyong, V. M., and Ekeleme, F. (2000). Characteristics of speargrass (Imperata cylindrica) dominated fields in West Africa: Crops, soil properties, farmer perceptions and management strategies. Crop Prot. 19(7), 481–487. Christensen, J. H., Hewitson, B., Busuioc, A., Chen, A., Gao, X., Held, I., Jones, R., Kolli, R. K., Kwon, W. T., Laprise, R., Magaa Rueda, V., Mearns, L., et al. (2007). Regional climate projections. In ‘‘Climate change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assesment Report of the Intergovernmental Panel on Climate Change’’ (S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K. B. Averyt, M. Tignor, and H. L. Miller, Eds.), pp. 847–940. Cambridge University Press, Cambridge. Ciotola, M., DiTommaso, A., and Watson, A. K. (2000). Chlamydospore production, inoculation methods and pathogenicity of Fusarium oxysporum M12-4A, a biocontrol for Striga hermonthica. Biocontrol Sci. Technol. 10(2), 129–145. Cisse´, J., Camara, M., Berner, D. K., and Musselman, L. J. (1996). Rhamphicarpa fistulosa (Scrophulariaceae) damages rice in Guinea. In ‘‘Advances in Parasitic Plant Research: 6th Parasitic Weeds Symposium’’ (M. T. Moreno, J. I. Cubero, D. K. Berner, D. Joel, L. J. Musselman, and C. Parker, Eds.), pp. 518–520. Cordoba, Spain. Cochrane, V., and Press, M. C. (1997). Geographical distribution and aspects of the ecology of the hemiparasitic angiosperm Striga asiatica (L.) Kuntze: a herbarium study. J. Trop. Ecol. 13(3), 371–380. Coleman, J. S., and Bazzaz, F. A. (1992). Effects of CO2 and temperature on growth and resource use of co-occurring C3 and C4 annuals. Ecology 73(4), 1244–1259. Cook, K. H., and Vizy, E. K. (2006). Coupled model simulations of the West African monsoon system: Twentieth- and twenty-first-century simulations. J. Climate 19(15), 3681–3703. Counce, P. A., and Burgos, N. R. (2006). Agronomic practices in addition to herbicides for improving weed control in dry-seeded flooded rice. J. Sustain. Agric. 28(3), 145–156. Cousens, R. (1985). An empirical model relating crop yield to weed and crop density and a statistical comparison with other models. J. Agric. Sci. 105, 513–521. Danquah, E. Y., Johnson, D. E., Riches, C., Arnold, G. M., and Karp, A. (2002). Genetic diversity in Echinochloa spp. collected from different geographic origins and within rice fields in Coˆte d’Ivoire. Weed Res. 42(5), 394–405. Davies, E. L. P. (1984). Mechanical and chemical methods for the control of annual wild rice. Trop. Pest Manag. 30(3), 319–320. De Datta, S. K., and Baltazar, A. M. (1996). Weed control technology as a component of rice production systems. In ‘‘Weed Management in Rice’’ (B. A. Auld and K. U. Kim, Eds.), pp. 25–52. FAO, Rome. De Datta, S. K., Tauro, A. C., and Balaoing, S. N. (1968). Effect of plant type and nitrogen level on the growth characteristics and grain yield of indica rice in the tropics. Agron. J. 60, 643–647. De Groote, H. (2007). Striga economics. In ‘‘Integrating new Technologies for Striga Control—Towards Ending the Witch-Hunt’’ (G. Ejeta and J. Gressel, Eds.), pp. 265–280. World Scientific, Singapore. de Rouw, A. (1995). The fallow period as a weed-break in shifting cultivation (tropical wet forests). Agric. Ecosyst. Environ. 54(1–2), 31–43.

Weed Management in Rice-Based Cropping Systems in Africa

205

Defoer, T., Wopereis, M. C. S., Idinoba, P. A., Kadisha, T. K. L., Diack, S., and Gaye, M. (2004a). Curriculum for Participatory Learning and Action Research (PLAR) for Integrated Rice Management (IRM) in Inland Valleys of Sub-Saharan Africa: Facilitators’ Manual.WARDA/IFDC, Cotonou, Benin/Muscle Shoals, USA. Defoer, T., Wopereis, M. C. S., Jones, M. P., Lancon, F., Erenstein, O., and Guei, R. G. (2004b). Rice-based production systems for food security and poverty alleviation in subSaharan Africa. Int. Rice Comm. Newsl. 53, 85–96. Delassus, M. (1972). Methods of controlling Striga spp. Agron. Trop. 27(2), 249–254. Delouche, J. C., Burgos, N. R., Gealy, D. R., Zorrilla de San Martin, G., and Labrada, R. (2007). Weedy Rices: Origin, Biology, Ecology and Control. Plant Production and Protection Paper 188. FAO, Rome. Demont, M., Jouve, P., Stessens, J., and Tollens, E. (2007). Boserup versus Malthus revisited: Evolution of farming systems in northern Coˆte d’Ivoire. Agric. Syst. 93(1–3), 215–228. Demont, M., Rodenburg, J., Diagne, M., and Diallo, S. (2009). Ex ante impact assessment of herbicide resistant rice in the Sahel. Crop Prot. in press. Diallo, S. (1999). Proble`me pose´ par le riz rouge en riziculture au Se´ne´gal. In ‘‘Global Workshop on Red Rice Control’’, pp. 45–49. FAO, Rome. Diallo, S., and Johnson, D. E. (1997). Les adventices du riz irrigue´ au Sahel et leur controˆle. In ‘‘Irrigated Rice in the Sahel: Prospects for Sustainable Development’’ (K. M. Mie´zan, M. C. S. Wopereis, M. Dingkuhn, J. Deckers, and T. F. Randolph, Eds.), pp. 311–323. WARDA, Dakar. Dingkuhn, M., Jones, M. P., Johnson, D. E., and Sow, A. (1998). Growth and yield potential of Oryza sativa and O. glaberrima upland rice cultivars and their interspecific progenies. Field Crops Res. 57(1), 57–69. Dingkuhn, M., Johnson, D. E., Sow, A., and Audebert, A. Y. (1999). Relationships between upland rice canopy characteristics and weed competitiveness. Field Crops Res. 61(1), 79–95. Drennan, D. S. H., and El Hiweris, S. O. (1979). Changes in growth regulator substances in Sorghum vulgare infected with Striga hermonthica. In ‘‘Proceedings of the Second International Symposium on Parasitic Weeds’’ (L. J. Musselman, A. D. Worsham, and R. E. Eplee, Eds.), pp. 144–155. North Carolina State University, Raleigh. Dugje, I. Y., Kamara, A. Y., and Omoigui, L. O. (2006). Infestation of crop fields by Striga species in the savanna zones of northeast Nigeria. Agric. Ecosyst. Environ. 116(3/4), 251–254. Dvorak, K. A. (1992). Resource-management by West-African farmers and the economics of shifting cultivation. Am. J. Agric. Econ. 74(3), 809–815. Dzomeku, I. K., Dogbe, W., and Agawu, E. T. (2007). Response of NERICA rice varieties to weed interference in the Guinea savannah uplands. J. Agron. 6(2), 262–269. Ejeta, G. (2007). The Striga scourge in Africa: A growing pandemic. In ‘‘Integrating new Technologies for Striga Control—Towards Ending the Witch-Hunt’’ (G. Ejeta and J. Gressel, Eds.), pp. 3–16. World Scientific, Singapore. Ekeleme, F., Akobundu, I. O., Isichei, A., and Chikoye, D. (2000). Influence of fallow type and land-use intensity on weed seed rain in a forest/savanna transition zone. Weed Sci. 48, 604–612. Ekeleme, F., Kamara, A. Y., Oikeh, S. O., Omoigui, L. O., Amaza, P., Abdoulaye, T., and Chikoye, D. (2009). Response of upland rice cultivars to weed competition in the savannas of West Africa. Crop Prot. 28(1), 90–96. Elliot, P. C., Clarisse, R. N., Beby, R., and Josue, H. R. (1993). Weeds in rice in Madagascar. Int. Rice Res. Notes 18(1), 53–54. Elmore, C. D., and Paul, R. N. (1983). Composite list of C4 weeds. Weed Sci. 31(5), 686–692.

206

J. Rodenburg and D. E. Johnson

Elzein, A. E. M. (2003). Development of a granular mycoherbicidal formulation of Fusarium oxysporum ‘‘Foxy 2’’ for the biological control of Striga hermonthica (Del.) Benth. In ‘‘Tropical Agriculture: Advances in Crop Research 2’’ ( J. Kroschel, Ed.), p. 174. Margraf Verlag, Weikersheim, Germany. Enyinnia, T. (1992). Chemical weed control in rainfed upland rice in Nigeria. Trop. Pest Manag. 38(4), 408–410. Fagade, S. O., and Ojo, A. A. (1977). Influence of plant density and nitrogen on yield and milling quality of lowland rice in Nigeria. Exp. Agric. 13(1), 17–24. FAO. (2008a). FAO Rice Market Monitor. http://www.fao.org/es/esc/en/index.html. July 2008. FAO. (2008b). FAO Statistical Databases. http://faostat.fao.org/. Faulkner, O. T. (1934). Some experiments with leguminous crops at Ibadan, Southern Nigeria 1925–1933. Emp. J. Exp. Agric. 2, 93–102. Fernandez-Quintanilla, C., Quadranti, M., Kudsk, P., and Barberi, P. (2008). Which future for weed science? Weed Res. 48(4), 297–301. Fischer, A. J., Ramı´rez, H. V., Gibson, K. D., and Da Silveira Pinheiro, B. (2001). Competitiveness of semidwarf upland rice cultivars against Palisadegrass (Brachiaria brizantha) and Signalgrass (B. decumbens). Agron. J. 93(5), 967–973. Fofana, B., and Rauber, R. (2000). Weed suppression ability of upland rice under low-input conditions in West Africa. Weed Res. 40(3), 271–280. Frost, H. M. (1995). Striga hermonthica surveys in western Kenya. In ‘‘Brighton Crop Protection Conference: Weeds. Proceedings of an International Conference’’, pp. 145–150. British Crop Protection Council, Farnham, UK. Frost, D. L., Gurney, A. L., Press, M. C., and Scholes, J. D. (1997). Striga hermonthica reduces photosynthesis in sorghum: The importance of stomatal limitations and a potential role for ABA. Plant Cell. Environ. 20(4), 483–492. Fuhrer, J. (2003). Agroecosystem responses to combinations of elevated CO2, ozone, and global climate change. Agric. Ecosyst. Environ. 97(1–3), 1–20. Fujisaka, S. (1990). Rice research priorities for Madagascar’s Middle West. IRRI Res. Pap. Ser. 144, 15. Garrity, D. P., Movillon, M., and Moody, K. (1992). Differential weed suppression ability in upland rice cultivars. Agron. J. 84(4), 586–591. Gbehounou, G., and Assigbe´, P. (2003). Rhamphicarpa fistulosa (Hochst.) Benth. (Scrophulariaceae): New pest on lowland rice in Benin. Results of a survey and immediate control possibilities. Ann. Sci. Agron. Be´nin 4(2), 89–103. Gehri, A., Kachelriess, S., and Fischer, P. A. (1999). Assessing women’s needs in Striga control. In ‘‘Advances in Parasitic Weed Control at On-Farm Level. Vol. I. Joint Action to Control Striga in Africa’’ ( J. Kroschel, H. Mercer Quarshie, and J. Sauerborn, Eds.), pp. 45–56. Margraf Verlag, Weikersheim, Germany. Gianessi, L. (2008). Solving Africa’s weed problem: Increasing crop production and improving the lives of women. In ‘‘5th International Weed Science Congress’’, Vancouver, Canada. Giannini, A., Biasutti, M., Held, I. M., and Sobel, A. H. (2008). A global perspective on African climate. Clim. Change 90(4), 359–383. Gibson, K. D., Foin, T. C., and Hill, J. E. (1999). The relative importance of root and shoot competition between water-seeded rice and Echinochloa phyllopogon. Weed Res. 39, 181–190. Gibson, K. D., Hill, J. E., Foin, T. C., Caton, B. P., and Fischer, A. J. (2001). Water-seeded rice cultivars differ in ability to interfere with watergrass. Agron. J. 93(2), 326–332. Gitsopoulos, T. K., and Froud-Williams, R. J. (2004). Effects of oxadiargyl on direct-seeded rice and Echinochloa crus-galli under aerobic and anaerobic conditions. Weed Res. 44(4), 329–334.

Weed Management in Rice-Based Cropping Systems in Africa

207

Gressel, J., and Gebrekidan, B. (2007a). Epilogue—Will there be integrated Striga control? In ‘‘Integrating new Technologies for Striga Control—Towards Ending the WitchHunt’’ (G. Ejeta and J. Gressel, Eds.), pp. 327–337. World Scientific, Singapore. Gressel, J., Meir, S., Herschkovitz, Y., Al-Ahmad, H., Babalola, O., and Amsellem, Z. (2007b). Transgenic biocontrol agents to overcome evolutionary barriers. In ‘‘Integrating new Technologies for Striga Control—Towards Ending the Witch-Hunt’’ (G. Ejeta and J. Gressel, Eds.), pp. 313–323. World Scientific, Singapore. Grist, D. H. (1968). Rice. Longman, London and New York. Gunasekera, T. G. L. G., and Rajapakse, C. N. K. (1994). Pareuchaetes pseudoinsulata as a possible biological control agent for Chromolaena odorata. Planter 70(822), 387–393. Gurney, A. L., Slate, J., Press, M. C., and Scholes, J. D. (2006). A novel form of resistance in rice to the angiosperm parasite Striga hermonthica. New Phytol. 169(1), 199–208. Gworgwor, N. A., and Ndahi, W. B. (2004). Parasitic weeds and their control/management in north-eastern Nigeria. In ‘‘Proceedings of the 4th International Weed Science Congress’’, Durban. Haden, V. R., Duxbury, J. M., DiTommaso, A., and Losey, J. E. (2007). Weed community dynamics in the system of rice intensification (SRI) and the efficacy of mechanical cultivation and competitive rice cultivars for weed control in Indonesia. J. Sustain. Agric. 30(4), 5–26. Haefele, S. M., Johnson, D. E., Diallo, S., Wopereis, M. C. S., and Janin, I. (2000). Improved soil fertility and weed management is profitable for irrigated rice farmers in Sahelian West Africa. Field Crops Res. 66(2), 101–113. Haefele, S., Wopereis, M. C. S., and Donovan, C. (2002). Farmers’ perceptions, practices and performance in a Sahelian irrigated rice scheme. Exp. Agric. 38, 197–210. Haefele, S. M., Johnson, D. E., M’ Bodj, D., Wopereis, M. C. S., and Mie´zan, K. M. (2004). Field screening of diverse rice genotypes for weed competitiveness in irrigated lowland ecosystems. Field Crops Res. 88(1), 39–56. Halwart, M. (2001). Fish as a component of integrated pest management in rice production. In ‘‘Integrated Agriculture–Aquaculture—A Primer. Fish as a Component of Integrated Pest Management in Rice Production’’, FAO/ICLARM/IIRR, Rome. Hammerton, J. L. (1967). Environmental factors and susceptibility to herbicides. Weeds 15(4), 330–336. Harahap, Z., Ampong Nyarko, K., and Olela, J. C. (1993). Striga hermonthica resistance in upland rice. Crop Prot. 12(3), 229–231. Harper, J. L. (1977). Population Biology of Plants.Academic Press, New York. Haussmann, B. I. G., Hess, D. E., Reddy, B. V. S., Mukuru, S. Z., Kayentao, M., Welz, H. G., and Geiger, H. H. (2001). Pattern analysis of genotype x environment interaction for Striga resistance and grain yield in African sorghum trials. Euphytica 122(2), 297–308. Held, I. M., Delworth, T. L., Lu, J., Findell, K. L., and Knutson, T. R. (2005). Simulation of Sahel drought in the 20th and 21st centuries. P. Natl. Acad. Sci. USA 102(50), 17891–17896. Hill, J. E., Mortimer, A. M., Namuco, O. S., and Janiya, J. D. (2002). Water and weed management in direct-seeded rice: Are we headed in the right direction. In ‘‘Rice Research for Food Security and Poverty Alleviation’’ (S. Peng and B. Hardy, Eds.), pp. 491–510. IRRI, Los Banos, Philippines. Hillocks, R. J. (1998). The potential benefits of weeds with reference to small holder agriculture in Africa. Integr. Pest Manag. Rev. 3, 155–167. Hoerling, M., Hurrell, J., Eischeid, J., and Phillips, A. (2006). Detection and attribution of twentieth-century northern and southern African rainfall change. J. Climate 19(16), 3989–4008. Hoffmann, G., Diarra, C., Ba, I., and Dembele, D. (1997). Parasitic plant species of food crops in Africa: Biology and impact, study in Mali. 1. Identification and biology of

208

J. Rodenburg and D. E. Johnson

parasitic plants. 2. Impact of parasitic plants based on the results of a study in Mali (1991– 1994). Agric. Dev. 13, 30–51. Holm, L. G., Plucknett, D. L., Pancho, P. V., and Herberger, J. P. (1991). The World’s Worst Weeds: Distribution and Biology. University Press, Hawaii, USA. Hong, N. H., Xuan, T. D., Eiji, T., and Khanh, T. D. (2004). Paddy weed control by higher plants from Southeast Asia. Crop Prot. 23(32), 255–261. Ikuenobe, C. E., and Anoliefo, G. O. (2003). Influence of Chromolaena odorata and Mucuna pruriens fallow duration on weed infestation. Weed Res. 43(3), 199–207. Imeokparia, P. O. (1994). Weed control in flooded rice with various herbicide combinations in the southern Guinea savanna zone of Nigeria. Int. J. Pest Manag. 40(1), 31–39. Ingram, J. S. I., Gregory, P. J., and Izac, A. M. (2008). The role of agronomic research in climate change and food security policy. Agric. Ecosyst. Environ. 126(1–2), 4–12. IPCC. (2007). Climate change 2007: Synthesis report.(Core Writing Team, R. K. Pachauri, and A. Reisinger, Eds.), IPCC, Geneva, Switzerland. Ishaya, D. B., Dadari, S. A., and Shebayan, J. A. Y. (2007). Evaluation of herbicides for weed control in three varieties of upland rice (Oryza sativa L.) in the Nigerian Savannah. Crop Prot. 26(10), 1490–1495. ISHRW. (2008). International Survey of Herbicide Resistant Weeds. http://www. weedscience.org. Iwuafor, E. N. O., and Kang, B. T. (1993). Effect of tillage and crop residue management on maize growth on an alfisol in the savanna zone. Agric. Syst. Africa 3, 71–83. Jagadish, S. V. K., Craufurd, P. Q., and Wheeler, T. R. (2007). High temperature stress and spikelet fertility in rice (Oryza sativa L. J. Exp. Bot. 58(7), 1627–1635. Jannink, J. L., Orf, J. H., Jordan, N. R., and Shaw, R. G. (2000). Index selection for weed suppressive ability in soybean. Crop Sci. 40(4), 1087–1094. Jennings, P. R., and Aquino, R. C. (1968). Studies on competition in rice. III. The mechanism of competition among phenotypes. Evolution 22(3), 529–542. Jensen, L. B., Courtois, B., Shen, L., Li, Z., Olofsdotter, M., and Mauleon, R. P. (2001). Locating genes controlling allelopathic effects against Barnyardgrass in upland rice. Agron. J. 93(1), 21–26. Johnson, D. E. (1995). Weed management strategies for smallholder rice production. In ‘‘Brighton Crop protection Conference: Weeds. Proceedings of an International Conference’’, pp. 1171–1180, British Crop Protection Council, Farnham, UK. Johnson, D. E. (1997). Weeds of Rice in West Africa. WARDA, Bouake´. Johnson, D. E., and Kent, R. J. (2002). The impact of cropping on weed species composition in rice after fallow across a hydrological gradient in west Africa. Weed Res. 42(2), 89–99. Johnson, D. E., Riches, C. R., Diallo, R., and Jones, M. J. (1997). Striga on rice in West Africa; Crop host range and the potential of host resistance. Crop Prot. 16(2), 153–157. Johnson, D. E., Dingkuhn, M., Jones, M. P., and Mahamane, M. C. (1998a). The influence of rice plant type on the effect of weed competition on Oryza sativa and Oryza glaberrima. Weed Res. 38(3), 207–216. Johnson, D. E., Riches, C., Camara, M., and Mbwaga, A. M. (1998b). Rhamphicarpa fistulosa on rice in Africa. Parasit. Plants Newsl, Haustorium, 33. Johnson, D. E., Riches, C. R., Kayeke, J., Sarra, S., and Tuor, F. A. (1999). Wild rice in subSaharan Africa: Its incidence and scope for improved management. In ‘‘Global Workshop on Red Rice Control’’ (R. Labrada, Ed.), pp. 87–93. FAO, Rome. Johnson, D. E., Riches, C. R., Jones, M. P., and Kent, R. (2000). The potential for host resistance to Striga on rice in West Africa. In ‘‘Breeding for Striga Resistance in Cereals: Proceedings of a Workshop Held at IITA’’ (B. I. G. Haussmann, D. E. Hess, M. L. Koyama, L. Grivet, H. F. W. Rattunde, and H. H. Geiger, Eds.), pp. 139–146. Margraf Verlag, Weikersheim, Germany.

Weed Management in Rice-Based Cropping Systems in Africa

209

Johnson, D. E., Wopereis, M. C. S., Mbodj, D., Diallo, S., Powers, S., and Haefele, S. M. (2004). Timing of weed management and yield losses due to weeds in irrigated rice in the Sahel. Field Crops Res. 85(1), 31–42. Jones, M. P., Johnson, D., Fofana, B., and Koupeur, T. (1996). Selection for weed competitiveness in upland rice. Int. Rice Res. Notes 21(1), 32–33. Jones, M. P., Dingkuhn, M., Aluko, G. K., and Semon, M. (1997). Interspecific Oryza sativa L. O. glaberrima Steud. progenies in upland rice improvement. Euphytica 94(2), 237–246. Jordan, N. (1993). Prospects for weed-control through crop interference. Ecol. Appl. 3(1), 84–91. Juo, A. S. R., and Mann, A. (1996). Chemical dynamics in slash-and-burn agriculture. Agric. Ecosyst. Environ. 58, 49–61. Kadir, J., and Charudattan, R. (2000). Dactylaria higginsii, a fungal bioherbicide agent for purple nutsedge (Cyperus rotundus). Biol. Control 17(2), 113–124. Kaewchumnong, K., and Price, A. H. (2008). A study on the susceptibility of rice cultivars to Striga hermonthica and mapping of Striga tolerance quantitative trait loci in rice. New Phytol. 180(1), 206–216. Kamara, A. Y., Akobundu, I. O., Chikoye, D., and Jutzi, S. C. (2000). Selective control of weeds in an arable crop by mulches from some multipurpose trees in Southwestern Nigeria. Agroforest. Syst. 50(1), 17–26. Kanampiu, F. K., Ransom, J. K., and Gressel, J. (2001). Imazapyr seed dressings for Striga control on acetolactate synthase target-site resistant maize. Crop Prot. 20(10), 885–895. Kanampiu, F. K., Kabambe, V., Massawe, C., Jasi, L., Friesen, D., Ransom, J. K., and Gressel, J. (2003). Multi-site, multi-season field tests demonstrate that herbicide seedcoating herbicide-resistance maize controls Striga spp. and increases yields in several African countries. Crop Prot. 22(5), 697–706. Kegode, G. O., Forcella, F., and Clay, S. (1999). Influence of crop rotation, tillage, and management inputs on weed seed production. Weed Sci. 47(2), 175–183. Kent, R. J., and Johnson, D. E. (2001). Influence of flood depth and duration on growth of lowland rice weeds, Coˆte d’Ivoire. Crop Prot. 20(8), 691–694. Kent, R. J., Johnson, D. E., and Becker, M. (2001). The influences of cropping system on weed communities of rice in Coˆte d’Ivoire, West Africa. Agric. Ecosyst. Environ. 87(3), 299–307. Khan, Z. R., Ampong-Nyarko, K., Chiliswa, P., Hassanali, A., Kimani, S., Lwande, W., Overholt, W. A., Pickett, J. A., Smart, L. E., and Woodcock, C. M. (1997). Intercropping increases parasitism of pests. Nature 388, 631–632. Khan, Z. R., Pickett, J. A., Berg, J., Wadhams, L. J., Woodcock, C. M., and van den Berg, J. (2000). Exploiting chemical ecology and species diversity: stem borer and striga control for maize and sorghum in Africa. Pest Manag. Sci. 56(11), 957–962. Khan, Z. R., Ahmed, H., Overholt, W., Khamis, T. M., Hooper, A. M., Pickett, J. A., Wadhams, L. J., and Woodcock, C. M. (2002). Control of witchweed Striga hermonthica by intercropping with Desmodium spp., and the mechanism defined as allelopathic. J. Chem. Ecol. 28(9), 1871–1885. Khan, Z. R., Midega, C. A. O., Hassanali, A., Pickett, J. A., Wadhams, L. J., and Wanjoya, A. (2006). Management of witchweed, Striga hermonthica, and stemborers in sorghum, Sorghum bicolor, through intercropping with greenleaf desmodium, Desmodium intortum. Int. J. Pest Manag. 52(4), 297–302. Khush, G. S. (1997). Origin, dispersal, cultivation and variation of rice. Plant Mol. Biol. 35, 25–34. Kim, S. C., and Moody, K. (1989). Germination of two rice cultivars and several weed species. Korean J. Weed Sci. 9(116), 122. King, C. A., and Oliver, L. R. (1992). Application rate and timing of acifluorfen, bentazon, chlorimuron, and imazaquin. Weed Technol. 6(3), 526–534.

210

J. Rodenburg and D. E. Johnson

Kluge, R. P., and Caldwell, P. M. (1993). Host-specificity of Pareuchaetes insulata (Lep. Arctiifae), a biological control agent for Chromolaena odorata (Compositae). Entomophaga 38(4), 451–457. Koffi, G. (1980). Collection and conservation of existing rice species and varieties of Africa. Agron. Trop. 34, 228–237. Kouazounde´, B. A. K. (2006). Innovations paysannes pour la gestion inte´gre´e des adventices dans les rizicultures des communes de Dassa-Zoume´ et de Glazoue´ au Be´nin Universite´ d’Abomey-Calavi, Cotonou, Benin. The`se de Diploˆme d’Inge´nieur. Krause, D., and Weber, H. C. (1990). SEM observations on seeds of Striga spp. and Buchnera americana (Scrophulariaceae). Plant Syst. Evol. 170(3/4), 257–263. Kristensen, L., Olsen, J., and Weiner, J. (2008). Crop density, sowing pattern, and nitrogen fertilization effects on weed suppression and yield in spring wheat. Weed Sci. 56, 97–102. Kropff, M. J., Bastiaans, L., and Lotz, L. A. P. (1997). Systems approaches in weed management and the design of weed suppressing crop varieties. In ‘‘Expert Consultation on Weed Ecology and Management’’ (R. Labrada, Ed.), pp. 73–85. FAO, Rome. Kroschel, J. (1998). Striga—How will it affect African agriculture in the future?—An ecological perspective. In ‘‘Agro-Ecology, Plant Protection and the Human Environment: Views and Concepts’’ (K. Martin, J. Muther, and A. Auffarth, Eds.), pp. 137–158. Margraf Verlag, Weikersheim, Germany. Kroschel, J. (1999). Analysis of the Striga problem, the first step towards further joint actions. In ‘‘Advances in Parasitic Weed Control at On-Farm Level’’ ( J. Kroschel, H. Mercer Quarshie, and J. Sauerborn, Eds.), pp. 3–26. Margraf Verlag, Weikersheim, Germany. Kroschel, J., Hundt, A., Abbasher, A. A., and Sauerborn, J. (1996). Pathogenicity of fungi collected in northern Ghana to Striga hermonthica. Weed Res. 36(6), 515–520. Kuijt, J. (1969). The Biology of Parasitic Flowering Plants. University of California Press, Berkeley. Lal, R. (1975). Role of Mulching Techniques in Tropical Soil and Water Management. IITA, Ibadan, Nigeria. Langyintuo, A. S., and Dogbe, W. (2005). Characterizing the constraints for the adoption of a Callopogonium mucunoides improved fallow in rice production systems in northern Ghana. Agric. Ecosyst. Environ. 110, 78–90. Latif, M. A., Islam, M. R., Ali, M. Y., and Saeque, M. A. (2005). Validation of the system of rice intensification (SRI) in Bangladesh. Field Crops Res. 93(2–3), 281–292. Lemerle, D., Verbeek, B., and Orchard, B. (2001). Ranking the ability of wheat varieties to compete with Lolium rigidum. Weed Res. 41(3), 197–209. Lendzemo, V. W., Kuyper, T. W., Kropff, M. J., and Ast, A. (2005). Field inoculation with arbuscular mycorrhizal fungi reduces Striga hermonthica performance on cereal crops and has the potential to contribute to integrated Striga management. Field Crops Res. 91, 51–61. Liebman, M., and Davis, A. S. (2000). Integration of soil, crop and weed management in low-external-input farming systems. Weed Res. 40, 27–47. Liu, X., Yang, Z., Huang, H., Hu, L., Chen, Y., and Wen, Z. (2004). Developing rules of field weeds in wetland rice-duck compounded system. J. Hunan Agric. Univ. 30(3), 292–294. Liu, J. X., Liao, D. Q., Oane, R., Estenor, L., Yang, X. E., Li, Z. C., and Bennett, J. (2006). Genetic variation in the sensitivity of anther dehiscence to drought stress in rice. Field Crops Res. 97(1), 87–100. Lu, B. R., and Snow, A. A. (2005). Gene flow from genetically modified rice and its environmental consequences. Bioscience 55(8), 669–678. Luder, R. (1985). Weeds influence red-billed quelea damage to ripening wheat in Tanzania. J. Wildlife Manag. 49(3), 646–647. MacLean, R. H., Litsinger, J. A., Moody, K., Watson, A. K., and Libetario, E. M. (2003). Impact of Gliricidia sepium and Cassia spectabilis hedgerows on weeds and insect pests of upland rice. Agric. Ecosyst. Environ. 94(3), 275–288.

Weed Management in Rice-Based Cropping Systems in Africa

211

Maiti, R. K., and Singh, V. P. (2004). Biotic factors affecting pearl millet [Pennisetum glaucum (L.) R. Br.] growth and productivity—A review. Crop Res. (Hisar) 27(1), 30–39. Mallamaire, A. (1949). Diseases, parasitic plants and weeds of cultivated rice in West Africa. Agron. Trop. 4, 77–80. Manneh, B., Kiepe, P., Sie´, M., Ndjiondjop, M. N., Drame, N. K., Traore´, K., Rodenburg, J., Somado, E. A., Narteh, L., Youm, O., Diagne, A., and Futakuchi, K. (2007). Exploiting partnership in research and development to help African rice farmers cope with climate variability. J. SAT Agric. Res. 4(1), 24. Marley, P. S., and Shebayan, J. A. Y. (2005). Field assessment of Fusarium oxysporum based mycoherbicide for control of Striga hermonthica in Nigeria. Biocontrol 50, 389–399. Masangkay, R. F., Mabbayad, M. O., Paulitz, T. C., and Watson, A. K. (1999). Host range of Alternaria alternata f. sp sphenocleae causing leaf blight of Sphenoclea zeylanica. Can. J. Bot. 77(1), 103–112. Matsui, T., Namuco, O. S., Ziska, L. H., and Horie, T. (1997). Effects of high temperature and CO2 concentration on spikelet sterility in indica rice. Field Crops Res. 51(3), 213–219. Mbwaga, A. M. (1996). Status of Striga species in Tanzania: Occurrence, distribution, and on-farm control packages. In ‘‘Drought tolerant crops for southern Africa. Proceedings of the SADC/ICRISAT Regional Sorghum and Pearl Millet Workshop’’ (K. Leuschner and C. S. Manthe, Eds.), pp. 195–200. International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Patancheru, India. Mbwaga, A. M., and Riches, C. R. (2006). Reversing the yield decline of upland-rice on Striga infested land. In ‘‘International Symposium on Integrating New Technologies for Striga Control: Towards Ending the Witch-hunt’’, Addis Ababa, Ethiopia. McKenzie, F. R. (1996). Influence of applied nitrogen on weed invasion of Lolium perenne pastures in a subtropical environment. Aust. J. Exp. Agric. 36, 657–660. MEA. (2005). Millennium Ecosystem Assessment—Ecosystems and Human Well-being: Synthesis. Island Press, Washington D.C. http://www.millenniumassessment.org/en/ synthesis.aspx. Men, B. X., Ogle, R. B., and Lindberg, J. E. (2002). Studies on integrated duck-rice systems in the Mekong Delta of Vietnam. J. Sustain. Agric. 20(1), 27–40. Merkel, U., Peters, M., Tarawali, S. A., Schultze Kraft, R., and Berner, D. K. (2000). Characterization of a collection of Aeschynomene histrix in subhumid Nigeria. J. Agric. Sci. 134(3), 293–304. Moffett, J. E., and McCloskey, W. B. (1998). Effects of soil moisture and yellow nutsedge (Cyperus esculentus) density on cotton (Gossypium hirsutum). Weed Sci. 46(2), 231–237. Mohamed, K. I., Papes, M., Williams, R., Benz, B. W., and Peterson, T. A. (2006). Global invasive potential of 10 parasitic witchweeds and related Orobanchaceae. Ambio 35, 281–288. Mohler, C. L. (1996). Ecological bases for the cultural control of annual weeds. J. Prod. Agric. 9, 468–474. Moody, K. (1991). Weed control in upland rice with the emphasis on grassy weeds. In ‘‘Tropical Grassy Weeds’’ (F. W. G. Baker and P. J. Terry, Eds.), CAB International, Wallingford, UK. Moormann, F. R., and Juo, A. S. R. (1986). Present land use and cropping systems in Africa. In ‘‘The Wetlands and Rice in Sub-Saharan Africa’’ (A. S. R. Juo and J. A. Lowe, Eds.), pp. 187–194. IITA, Ibadan, Nigeria. Morita, H., and Kabaki, N. (2002). Effects of soil moisture conditions on the emergence of weeds and rice plants from rainfed paddy soils in north-east Thailand. Weed Biol. Manag. 2(4), 209–212. Mortensen, D. A., Bastiaans, L., and Sattin, M. (2000). The role of ecology in the development of weed management systems: An outlook. Weed Res. 40(1), 49–62. Mortimer, A. M., Namuco, O., and Johnson, D. E. (2005). Seedling recruitment in directseeded rice: Weed biology and water management. In ‘‘Rice Is Life: Scientific

212

J. Rodenburg and D. E. Johnson

Perspectives for the 21st Century’’ (K. Toriyama, K. L. Heong, and B. Hardy, Eds.), pp. 202–205. IRRI, Tokyo and Tsukuba, Japan. Murdoch, A. J., and Kunjo, E. M. (2003). Depletion of natural soil seedbanks of Striga hermonthica in West Africa under different integrated management regimes. Asp. Appl. Biol. 69, 261–268. (Seedbanks: Determination, Dynamics and Management). Navas, M. L., Garnier, E., Austin, M. P., and Gifford, R. M. (1999). Effect of competition on the responses of grasses and legumes to elevated atmospheric CO2 along a nitrogen gradient: Differences between isolated plants, monocultures and multi-species mixtures. New Phytol. 143(2), 323–331. Navasero, N. C., and Khan, A. U. (1970). Use of mechanical power for rotary weeding. Pans 16, 87–92. Neumann, U., Salle´, G., and Weber, H. C. (1998). Development and structure of the haustorium of the parasite Rhamphicarpa fistulosa (Scrophulariaceae). Bot. Acta 111(5), 354–365. Ngobo, M., McDonald, M., and Weise, S. (2004). Impacts of type of fallow and invasion by Chromolaena odorata on weed communities in crop fields in Cameroon. Ecol. Soc. 9(2) http://www.ecologyandsociety.org/vol9/iss2/art1. Nonhebel, S. (1996). Effects of temperature rise and increase in CO2 concentration on simulated wheat yields in Europe. Clim. Change 34(1), 73–90. Nwilene, F., Togola, A., Agunbiade, T. A., Ogah, E. O., Ukwungwu, M. N., Hamadoun, A., Kamara, S. I., and Dakouo, D. (2008). Parasitoid biodiversity conservation for sustainable management of the African rice gall midge, Orseolia oryzivora (Diptera: Cecidomyiidae) in lowland rice. Biocontrol Sci. Technol. 18(10), 1075–1081. Nwoke, F. I. O., and Okonkwo, S. N. C. (1980). Photocontrol of seed-germination in the hemiparasite Buchnera hispida (Scrophulariaceae). Physiol. Plant. 49(4), 388–392. Nyoka, G. C. (1982). The influence of fallow period on weed vegetation and rice yields in Sierra Leone. Trop. Pest Manag. 28(1), 1–9. Nyoka, G. C. (1983). Weed problems and control practices in deepwater and floating rice in Mali. In ‘‘Proceedings of the Second Biannual Conference West African Weed Science Society’’ (M. Deat and P. Marnotte, Eds.), pp. 146–157. West African Weed Science Society, Abidjan, Coˆte d’Ivoire. Oechel, W. C., and Strain, B. R. (1985). Native species responses to increased atmospheric carbon dioxide concentration. In ‘‘Direct Effects of Increasing Carbon Dioxide on Vegetation’’ (B. R. Strain and J. D. Cure, Eds.), U.S. Department of Energy, Washington D.C. Office of Basic Energy Sciences. Carbon Dioxide Research Division. Available from the National Technical Information Service, US Department of Commerce, Springfield VA. Oerke, E.-C., and Dehne, H.-W. (2004). Safeguarding production-losses in major crops and the role of crop protection. Crop Prot. 23, 275–285. Okafor, L. I. (1986). Chemical weed control in direct-seeded irrigated rice in the Lake Chad Basin, Nigeria. Crop Prot. 5(3), 203–208. Olofsdotter, M., Navarez, D., Rebulanan, M., and Streibig, J. C. (1999a). Weed-suppressing rice cultivars—Does allelopathy play a role?. Weed Res. 39(6), 441–454. Olofsdotter, M., Valverde, B. E., and Madsen, K. H. (1999b). Herbicide resistant rice (Oryza sativa L.)—A threat or a solution. In ‘‘Global Workshop on Red Rice Control’’ (R. Labrada, Ed.), pp. 123–142. FAO, Rome. Olofsdotter, M., Jensen, L. B., and Courtois, B. (2002). Improving crop competitive ability using allelopathy—An example from rice. Plant Breed. 121(1), 1–9. Omanya, G. O., Haussmann, B. I. G., Hess, D. E., Reddy, B. V. S., Kayentao, M., Welz, H. G., and Geiger, H. H. (2004). Utility of indirect and direct selection traits for improving Striga resistance in two sorghum recombinant inbred populations. Field Crops Res. 89(2–3), 237–252.

Weed Management in Rice-Based Cropping Systems in Africa

213

Oswald, A., and Ransom, J. K. (2001). Striga control and improved farm productivity using crop rotation. Crop Prot. 20(2), 113–120. Oswald, A., Ransom, J. K., Kroschel, J., and Sauerborn, J. (2002). Intercropping controls Striga in maize based farming systems. Crop Prot. 21(5), 367–374. Ou, S. H. (1985). Rice Diseases. Commonwealth Mycological Institute, Kew, Surrey. Ouedraogo, O., Neumann, U., Raynal Roques, A., Salle, G., Tuquet, C., and Dembele, B. (1999). New insights concerning the ecology and the biology of Rhamphicarpa fistulosa (Scrophulariaceae). Weed Res. 39(2), 159–169. Parker, C., and Riches, C. R. (1993). Parasitic Weeds of the World: Biology and Control. Cab International, Wallingford, Oxon. Parkinson, V. O. (1989). A survey of infestation of crops by Striga spp. in Benin, Nigeria and Togo. Proc. Nova Scotian Inst. Sci. 39(1), 1–9. Parmesan, C. (1996). Climate and species’ range. Nature 382, 765–766. Patterson, D. T. (1995). Weeds in a Changing Climate. Weed Sci. 43(4), 685–701. Patterson, D. T., Musser, R. L., Flint, E. P., and Eplee, R. E. (1982). Temperature responses and potential for spread of witchweed (Striga lutea) in the United States. Weed Sci. 30(1), 87–93. Patterson, D. T., Westbrook, J. K., Joyce, R. J. V., Lingren, P. D., and Rogasik, J. (1999). Weeds, insects, and diseases. Clim. Change 43(4), 711–727. Pena-Fronteras, J. T., Villalobos, M. C., Baltazar, A. M., Merca, F. E., Ismail, A. M., and Johnson, D. E. (2008). Adaptation to flooding in upland and lowland ecotypes of Cyperus rotundus, a troublesome sedge weed of rice: Tuber morphology and carbohydrate metabolism. Ann. Bot. 103(2), 295–302. Perez de Vida, F. B., Laca, E. A., Mackill, D. J., Ferna´ndez, G. M., and Fischer, A. J. (2006). Relating rice traits to weed competitiveness and yield: A path analysis. Weed Sci. 54(6), 1122–1131. Pernito, R. G., Edillo, N. A., Baquiran, V. A., and Garrity, D. P. (1986). Differential weed suppression ability in rainfed lowland rice. Philipp. J. Crop Sci. 11(Suppl. 1), Abst. 3D–3a. Pester, T. A., Burnside, O. C., and Orf, J. H. (1999). Increasing crop competitiveness to weeds through crop breeding. J. Crop. Prod. 2(1), 59–76. Phoenix, G. K., and Press, M. C. (2005). Effects of climate change on parasitic plants: The root hemiparasitic Orobanchaceae. Folia Geobot. 40(2–3), 205–216. Phuong, L. T., Denich, M., Vlek, P. L. G., and Balasubramanian, V. (2005). Suppressing weeds in direct-seeded lowland rice: Effects of methods and rates of seeding. J. Agron. Crop Sci. 191(3), 185–194. Pierce, S., Mbwaga, A. M., Press, M. C., and Scholes, J. D. (2003). Xenognosin production and tolerance to Striga asiatica infection of high-yielding maize cultivars. Weed Res. 43, 139–145. Posner, J. L., and Crawford, E. W. (1991). An agro-economic analysis of field trials from a farming systems research perspective—Weed control in rain-fed lowland rice in Senegal. Exp. Agric. 27(3), 231–241. Poussin, J. C. (1997). Direct seeding or transplanting? The example of the Senegal middle valley. In ‘‘Irrigated Rice in the Sahel: Prospects for Sustainable Development’’ (K. M. Miezan, M. C. S. Wopereis, M. Dingkuhn, J. Deckers, and T. F. Randolph, Eds.), pp. 191–200. WARDA, Dakar. Press, M. C., Shah, N., Tuohy, J. M., and Stewart, G. R. (1987). Carbon isotope ratios demonstrate carbon flux from C4 host to C3 parasite. Plant Physiol. 85(4), 1143–1145. Press, M. C., Smith, S., and Stewart, G. R. (1991). Carbon acquisition and assimilation in parasitic plants. Funct. Ecol. 5(2), 278–283. Pronier, I., Pare, J., Traore, D., Vincent, C., and Stewart, R. K. (1998). A histological study of the effect of feeding by Smicronyx spp. (Coleoptera: Curculionidae) larvae on seed production by Striga hermonthica (Scrophulariaceae). Biol. Control 13(3), 152–157. Radosevich, S. R. (1987). Methods to study interactions among crops and weeds. Weed Technol. 1, 190–198.

214

J. Rodenburg and D. E. Johnson

Ransom, J. K. (2000). Long-term approaches for the control of Striga in cereals: Field management options. Crop Prot. 19(8–10), 759–763. Rao, A. N., Johnson, D. E., Sivaprasad, B., Ladha, J. K., and Mortimer, A. M. (2007). Weed management in direct-seeded rice. Adv. Agron. 93, 153–255. Reichmann, S., Kroschel, J., and Sauerborn, J. (1995). Distribution and infestation of Striga species in Shinyanga region of Tanzania and evaluation of control methods. In ‘‘Brighton Crop protection Conference: Weeds. Proceedings of an International Conference’’ (Brighton crop protection conference: Weeds, Ed.), pp. 151–156. British Crop Protection Council, Farnham, UK. Reneaud, H. (1980). Control of Striga asiatica, a rice weed in the Comoro islands. Agron. Trop. 35(1), 61–63. Riches, C., Shaxson, L., Logan, J., and Munthali, D. (1993). Insect and parasitic weed problems in southern Malawi and the use of farmer knowledge in the design of control measures. Netw. Pap. Agric. Adm. (Res. Ext.) Netw. 42, 1–17. Riches, C. R., Johnson, D. E., and Jones, M. P. (1996). The selection of resistance to Striga species in upland rice. In ‘‘Advances in Parasitic Research. Proceedings of the Sixth International Parasitic Weed Symposium’’ (M. T. Moreno, J. I. Cubero, D. Berner, D. Joel, L. J. Musselman, and C. Parker, Eds.), pp. 673–680. Junta de Andalucia, Sevilla. Riches, C. R., Mbwaga, A. M., Mbapila, J., and Ahmed, G. J. U. (2005). Improved weed management delivers increased productivity and farm incomes from rice in Bangladesh and Tanzania. Asp. Appl. Biol. 75, 127–138. Rijn, P. J. (2001). Weed Management in the Humid and Sub-humid Tropics Koninklijk Instituut voor de Tropen (KIT, Amsterdam, The Netherlands. Robinson, E. L., and Dowler, C. C. (1966). Investigations of catch and trap crops to eradicate witchweed (Striga asiatica). Weeds 14(3), 275–276. Rodenburg, J., Bastiaans, L., Weltzien, E., and Hess, D. E. (2005). How can field selection for Striga resistance and tolerance in sorghum be improved? Field Crops Res. 93(1), 34–50. Rodenburg, J., Bastiaans, L., Kropff, M. J., and van Ast, A. (2006a). Effects of host plant genotype and seedbank density on Striga reproduction. Weed Res. 46(3), 251–263. Rodenburg, J., Diagne, A., Oikeh, S., Futakuchi, K., Kormawa, P. M., Semon, M., Akintayo, I., Cisse´, B., Sie´, M., Narteh, L., Nwilene, F., Diatta, S., et al. (2006b). Achievements and impact of NERICA on sustainable rice production in sub-Saharan Africa. Int. Rice Comm. Newsl. 55, 45–58. Rodenburg, J., Saito, K., Glele, R., Toure´, A., Mariko, M., and Kiepe, P. (2009). Weed competitiveness of the lowland rice varieties of NERICA in the southern Guinea Savanna. Field Crops Res. under review. Rogers, W. E., and Nelson, R. R. (1962). Penetration and nutrition of Striga asiatica. Phytopathology 52, 1064–1070. Raynal-Roques, A. (1994). Major, minor and potential parasitic weeds in semi-arid tropical Africa: the example of Scrophulariaceae. In (Pieterse, A. H., Verkleij, J. A. C., and ter Borg, S. J., Eds.), Biology and management of Orobanche, Proceedings of the Third International Workshop on Orobanche and related Striga research. pp. 400-405. Royal Tropical Institute, Amsterdam. Sands, D. C., Pilgeram, A. L., Thompson, B. M., and Kirkpatrick, M. M. (2007). Genetically enhancing the virulence and efficacy of plant pathogens for biological control of parasitic plants. In ‘‘Integrating new Technologies for Striga Control—Towards Ending the Witch-Hunt’’ (G. Ejeta and J. Gressel, Eds.), pp. 301–311. World Scientific, Singapore. Sanyal, D., Bhowmik, P. C., Anderson, R. L., and Shrestha, A. (2008). Revisiting the perspective and progress of integrated weed management. Weed Sci. 56(1), 161–167. Saunders, A. R. (1933). Studies in phanerogamic parasitism, with particular reference to Striga lutea Lour. S. Afr. Dep. Agric. Sci. Bull. 128, 1–56.

Weed Management in Rice-Based Cropping Systems in Africa

215

Schwartz, A. M., Paskewitz, S. M., Orth, A. P., Tesch, M. J., Toong, Y. C., and Goodman, W. G. (1998). The lethal effects of Cyperus iria on Aedes aegypti. J. Am. Mosquito Contr. 14(1), 78–82. Seckler, D., Barker, R., and Amarasinghe, U. (1999). Water scarcity in the twenty-first century. Int. J. Water Resour. D 15, 29–42. Sengupta, K., Bhattacharyya, K. K., and Chatterjee, B. N. (1985). Intercropping of upland rice with blackgram [Vigna mungo (L)]. J. Agric. Sci. 104, 217–221. Sha, X. Y., Linscombe, S. D., and Groth, D. E. (2007). Field evaluation of imidazolinonetolerant Clearfield rice (Oryza sativa L.) at nine Louisiana locations. Crop Sci. 47(3), 1177–1185. Shivrain, V. K., Burgos, N. R., Anders, M. M., Rajguru, S. N., Moore, J., and Sales, M. A. (2007). Gene flow between Clearfield (TM) rice and red rice. Crop Prot. 26(3), 349–356. Singh, H. P., Batish, D. R., and Kohli, R. K. (2003). Allelopathic interactions and allelochemicals: New possibilities for sustainable weed management. Crit. Rev. Plant Sci. 22(3), 239–311. Singh, S., Ladha, J. K., Gupta, R. K., Bhushan, L., Rao, A., Sivaprasad, B., and Singh, P. Y. (2007). Evaluation of mulching, intercropping with Sesbania and herbicide use for weed management in dry-seeded rice (Oryza sativa L). Crop Prot. 26(4), 518–524. Smith, M. C., Holt, J., and Webb, M. (1993). A population model of the parasitic weed Striga hermonthica (Scrophulariaceae) to investigate the potential of Smicronyx umbrinus (Coleoptera: Curculionidae) for biological control in Mali. Crop Prot. 12(6), 470–476. Stessens, J. (2002). Analyse technique et e´conomique des syste`mes de production agricole au Nord de la Coˆte d’Ivoire. Katholieke Universiteit Leuven, Leuven. Stewart, G. (1990). Witchweed: a parasitic weed of grain crops. Outlook Agric. 19(2), 115–117. Swanton, C. J., and Weise, S. F. (1991). Integrated weed management—The rationale and approach. Weed Technol. 5(3), 657–663. Tan, S. Y., Evans, R. R., Dahmer, M. L., Singh, B. K., and Shaner, D. L. (2005). Imidazolinone-tolerant crops: History, current status and future. Pest Manag. Sci. 61(3), 246–257. Tarawali, G., Manyong, V. M., Carsky, R. J., Vissoh, P. V., Osei-Bonsu, P., and Galiba, M. (1999). Adoption of improved fallows in West Africa: Lessons from Mucuna and Stylo case studies. Agroforest. Syst. 47, 93–122. Taylor, A., Martin, J., and Seel, W. E. (1996). Physiology of the parasitic association between maize and witchweed (Striga hermonthica): Is ABA involved? J. Exp. Bot. 47 (301), 1057–1065. Taylor, R. L., Maxwell, B. D., and Boik, R. J. (2006). Indirect effects of herbicides on bird food resources and beneficial arthropods. Agric. Ecosyst. Environ. 116(3–4), 157–164. Teasdale, J. R. (2003). Principles and practices of using cover crops in weed management systems. In ‘‘Weed Management for Developing Countries.’’ (R. Labrada, Ed.), pp. 169–178. FAO, Rome. Tollenaar, M., Nissanka, S. P., Aguilera, A., Weise, S. F., and Swanton, C. J. (1994). Effect of weed interference and soil nitrogen on four maize hybrids. Agron. J. 86, 596–601. Tonye, J., Ibewiro, B., and Duguma, B. (1997). Residue management of a planted fallow on an acid soil in Cameroon: crop yields and soil organic matter fractions. Agroforest. Syst. 37, 199–207. Townson, J. (1991). Imperata cylindrica and its control. Weed Abstr. 40, 457–468. Treca, B. (1985). Les possibilite´s de lutte contre les oiseaux d’eau pour prote´ger les rizie`res en Afrique de l’Ouest. J. Agric. Tradit. Bot. Appl. 32, 191–213. Tubiello, F. N., Donatelli, M., Rosenzweig, C., and Stockle, C. O. (2000). Effects of climate change and elevated CO2 on cropping systems: Model predictions at two Italian locations. Eur. J. Agron. 13, 179–189.

216

J. Rodenburg and D. E. Johnson

Tubiello, F. N., Soussana, J., and Howden, S. M. (2007). Crop and pasture response to climate change. P. Natl Acad. Sci. USA 104(50), 19686–19690. Tungate, K. D., Israel, D. W., Watson, D. M., and Rufty, T. W. (2007). Potential changes in weed competitiveness in an agroecological system with elevated temperatures. Environ. Exp. Bot. 60(1), 42–49. UN. (2007). World Population Prospects: The 2006 Revision Population Database. http:// esa.un.org/unpp/. UNDP. (2007). UNDP Human Development Report 2007/2008. Fighting Climate Change: Human Solidarity in a Divided World United Nations Development Programme, New York. UNEP-WCMC. (2007). Biodiversity and Poverty Reduction: The Importance of Biodiversity for Ecosystem Services.United Nations Environmental Programme, World Conservation Monitoring Centre (UNEP-WCMC), Cambridge. Vallance, K. B. (1950). Studies on the germination of the seeds of Striga hermonthica I. The influence of moisture treatment, stimulant dilution, and after-ripening on germination. Ann. Bot. 14, 347–363. van Delft, G. J., Graves, J. D., Fitter, A. H., and Pruiksma, M. A. (1997). Spatial distribution and population dynamics of Striga hermonthica seeds in naturally infested farm soils. Plant Soil 195, 1–15. van der Meijden, G. (1998). Motorized soil tillage in West Africa. A survey on the current use and consequences of tillage done with engine-driven machinery FAO, Rome, Italy http://www.fao.org/ag/ags/agse/TILPAP.htm. Van Mele, P. (2006). Zooming-in zooming-out: A novel method to scale up local innovations and sustainable technologies. Int. J. Agric. Sustain. 4(2), 131–142. van Mourik, T. A., Bianchi, F. J. J. A., Van Der Werf, W., and Stomph, T. J. (2008). Longterm management of Striga hermonthica: strategy evaluation with a spatio-temporal population model. Weed Res. 48(4), 329–339. Vissoh, P. V., Gbehounou, G., Ahanchede, A., Kuyper, T. W., and Roling, N. G. (2004). Weeds as agricultural constraint to farmers in Benin: results of a diagnostic study. NJAS 52(3–4), 305–329. Vogt, W., Sauerborn, J., and Honisch, M. (1991). Striga hermonthica distribution and infestation in Ghana and Togo on grain crops. In ‘‘Proceedings of the 5th International Symposium of Parasitic Weeds’’ (J. K. Ransom, L. J. Musselman, A. D. Worsham, and C. Parker, Eds.), pp. 372–377. CIMMYT, Nairobi, Kenya. Wand, S. J. E., Midgley, G. Y. F., Jones, M. H., and Curtis, P. S. (1999). Responses of wild C4 and C3 grass (Poaceae) species to elevated atmospheric CO2 concentration: A metaanalytic test of current theories and perceptions. Glob. Change Biol. 5(6), 723–741. Wanders, A. A. (1986). Appropriate mechanization for rice cultivation in Africa. In ‘‘The Wetlands and Rice in Subsaharan Africa’’ (A. S. R. Juo and J. A. Lowe, Eds.), pp. 187–194. IITA, Ibadan, Nigeria. Ward, P. (1965). Feeding ecology of the black-faced dioch Quelea quelea in Nigeria. Ibis 107(173), 214. WARDA. (2008). Rice Videos. http://www.warda.org/warda/guide-video.asp. Watling, J. R., and Press, M. C. (2000). Infection with the parasitic angiosperm Striga hermonthica influences the response of the C3 cereal Oryza sativa to elevated CO2. Glob. Change Biol. 6(8), 919–930. Way, M. J., and Heong, K. L. (1994). The role of biodiversity in the dynamics and management of insect pests of tropical irrigated rice—A review. Bull. Entomol. Res. 84(4), 567–587. Webb, M., and Smith, M. C. (1996). Biology of Striga hermonthica (Scrophulariaceae) in Sahelian Mali: effects on pearl millet yield and prospects of control. Weed Res. 36(3), 203–211.

Weed Management in Rice-Based Cropping Systems in Africa

217

Weise, S. F. (1995). Distribution and significance of C. odorata (L.) R. King and H. Robinson across ecological zones in Cameroon. In ‘‘Proceedings of the 3rd International Workshop on Biological Control and Management of C. odarata’’, pp. 29–38. University of Guam, Abidjan. Weston, L. A. (1996). Utilization of allelopathy for weed management in agroecosystems. Agron. J. 88, 860–866. Weston, L. A., and Duke, S. O. (2003). Weed and crop allelopathy. Crit. Rev. Plant Sci. 22(3), 367–389. Williams, A. L., Wills, K. E., Janes, J. K., Schoor, J. K. V., Newton, P. C. D., and Hovenden, M. J. (2007). Warming and free-air CO2 enrichment alter demographics in four co-occurring grassland species. New Phytol. 176(2), 365–374. Wilson, A. K. (1981). Euphorbia heterophylla: A review of distribution, importance and control. Trop. Pest Manag. 27(1), 32–38. Windmeijer, P. N., and Andriesse, W. (1993). Inland Valleys in West Africa: An AgroEcological Characterization of Rice Growing Environments. ILRI, Wageningen. Windmeijer, P. N., Duivenbooden, N., and Andriesse, W. (1994). Characterization of ricegrowing agro-ecosystems in West Africa: Semi-detailed characterization of inland valleys in Coˆte d’Ivoire. In (Windmeijer, P. N., Duivenbooden, N., and Andriesse, W., Eds.). SC-DLO-Wageningen Agricultural University, Wageningen, The Netherlands. Wittig, R., Ko¨nig, K., Schmidt, M., and Szarzynski, J. (2007). A study of climate change and anthropogenic impacts in West Africa. Environ. Sci. Pollut. R 14(3), 182–189. Wong, S. C. (1990). Elevated atmospheric partial-pressure of CO2 and plant-growth. 2. Nonstructural carbohydrate content in cotton plants and its effect on growth-parameters. Photosynth. Res. 23(2), 171–180. Wopereis, M. C. S., Defoer, T., Idinoba, M. E., Diack, S., and Dugue´, M. J. (2007). Participatory learning and action research (PLAR) for integrated rice management (IRM) in inland valleys of sub-Saharan Africa: Technical manual.WARDA/IFDC, Cotonou, Benin/Muscle Shoals, USA. Xuan, T. D., Shinkichi, T., Hong, N. H., Khanh, T. D., and Min, C. I. (2004). Assessment of phytotoxic action of Ageratum conyzoides L. (billy goat weed) on weeds. Crop Prot. 23(10), 915–922. Xuan, T. D., Shinkichi, T., Khanh, T. D., and Min, C. I. (2005). Biological control of weeds and plant pathogens in paddy rice by exploiting plant allelopathy: An overview. Crop Prot. 24(3), 197–206. Yoder, J. I. (2001). Host-plant recognition by parasitic Scrophulariaceae. Curr. Opin. Plant Biol. 4, 359–365. Zhao, D. L., Atlin, G. N., Bastiaans, L., and Spiertz, J. H. J. (2006a). Cultivar weedcompetitiveness in aerobic rice: Heritability, correlated traits, and the potential for indirect selection in weed-free environments. Crop Sci. 46(1), 372–380. Zhao, D. L., Atlin, G. N., Bastiaans, L., and Spiertz, J. H. J. (2006b). Developing selection protocols for weed competitiveness in aerobic rice. Field Crops Res. 97(2–3), 272–285. Zhao, D. L., Bastiaans, L., Atlin, G. N., and Spiertz, J. H. J. (2007). Interaction of genotype management on vegetative growth and weed suppression of aerobic rice. Field Crops Res. 100(2–3), 327–340. Zimdahl, R. L. (2007). Fundamentals of Weed Science. Academic Press, London. Ziska, L. H. (2001). Changes in competitive ability between a C-4 crop and a C-3 weed with elevated carbon dioxide. Weed Sci. 49(5), 622–627. Ziska, L. H. (2003). Evaluation of the growth response of six invasive species to past, present and future carbon dioxide concentrations. J. Exp. Bot. 54, 395–404. Ziska, L. H. (2008). Rising atmospheric carbon dioxide and plant biology: The overlooked paradigm. DNA Cell Biol. 27(4), 165–172.

218

J. Rodenburg and D. E. Johnson

Ziska, L. H., and Teasdale, J. R. (2000). Sustained growth and increased tolerance to glyphosate observed in a C-3 perennial weed, quackgrass (Elytrigia repens), grown at elevated carbon dioxide. Aust. J. Plant Physiol. 27(2), 159–166. Ziska, L. H., Teasdale, J. R., and Bunce, J. A. (1999). Future atmospheric carbon dioxide may increase tolerance to glyphosate. Weed Sci. 47(5), 608–615. Zossou, N. (2008). Riziculture de bas-fonds au Be´nin: e´valuation des proble`mes phytosanitaires cause´s par Rhamphicarpa fistulosa dans les communes de Dassa et Glazoue´; roˆle des NERICA, autres riz de bas-fonds et de la fumure mine´rale dans le controˆle inte´gre´ du Rhamphicarpa fistulosa. The`se de Diploˆme d’Inge´nieur, Universite´ d’Abomey-Calavi, Cotonou, Benin.

Index

A African rice-based cropping systems. See also Weed management, Africa annual import savings, 154 biological weed control, 178–179 chemical weed control conventional methods, 179–181 herbicide resistant rice technologies, 181–184 integrated weed control, 185 cultural weed control flooding, 166–167 mixed cropping, rotations, and fallow, 168–172 mulching, 168 planting methods, 165–166 soil fertility management, 167–168 ecosystems, 151 legume species, improved fallows, 169–171 production, 150 rain-fed lowlands, 153 species, problem weeds, 155–158 superior lowland NERICA varieties, 176, 177 upland cropping systems, 151 varietal development, 174–178 weed management strategy, 189 integrated crop and weed management approaches, 192 prioritization, weed species, 190–192 timing, weed control and crop management, 192–193 weed research strategy, 193 climate change, 194 crop management, 194–195 ecosystems, 197–198 herbicides, 196–197 socioeconomics and gender, 198–199 varietal development, 195–196 weed species problematic weeds, 155–158 rain-fed and irrigated lowlands, 161–163 uplands and hydromorphic zones, 159–161 usefulness, 163–164 Ageratum conyzoides, 155, 156, 160 Anthropogenic emission, livestock agriculture and climatic change, 9–11 climatic change

in California, 5–7 in United States, 4–5 cultivated soil DAYCENT model, 24–25 direct and indirect N2O emissions, 25–26 desertification, 23–24 enteric fermentation carbon dioxide emission, 17–18 CH4 emission, 15–17 feed production and carbon emission BST hormone, technology application, 29 concentrates, 28 dual-utilization, cropland, 28 mineral fertilizer, 26–27 synthetic fertilizers, 27 technology application, 28–29 forest land transformation, 34 global contribution and climatic change, 3–4 greenhouse gas (GHG) sources, 3 land-use change definition, 21 LLS estimation, 21–22 life cycle assessment (LCA), 8–9 manure management CH4 and N2O emissions, 19 rice field and animal diet, 20 numerical suffix system, 35 on-farm fossile fuel, 29–30 postharvest activity hydroelectric vs. coal energy, 30 transportation, 31–32 waste and biomass, 32–33 scaling effect, 33 types and production systems grassland-based LPSs, 14 landless LPS, 13 mixed farming, 14 ruminants and nonruminants, 11–12 B Biological weed control, 178–179 Bovine somatotropin (BST) hormone, livestock emmision, 29 C Canopy temperature, 56–57 Carbon isotope discrimination (CID), 55

219

220

Index

Chemical weed control conventional methods, 179–181 herbicide resistant rice technologies, 181–184 integrated weed control, 185 Chromolaena odorata, 156, 158, 159 Comparative Mapping Tool (CMAP), 84 Comparative Plant Stress-Responsive Gene Catalog, 85 Cultivated land security, 122 Cultural weed control flooding, 166–167 mixed cropping, rotations, and fallow, 168–172 mulching, 168 planting methods, 165–166 soil fertility management, 167–168 Cyperus difformis, 155, 157, 161 D Dehydration-avoidance mechanism, stage III canopy temperature, 56–57 carbon isotope discrimination (CID), 55 photosythesis and stomata aperture traits, 54–55 plant water status, 52–53 root traits, 53–54 Digitaria horizontalis, 155, 156, 159 Drought resistant rice improvement bioinformatics and gene functional analysis application, 84–85 Comparative Plant Stress-Responsive Gene Catalog, 85–86 expressed sequence tag (EST) library, 84 International Crop Information System (ICIS), 85–86 International Rice Functional Genomics Consortium (IRFGC), 86 publicly accessible databases, 86 QTL position data, 84–85 breeding, conventional type developing breeding populations, 68 germplasm screening, 64–66 heritability, 66–67 high yield combination, 70 population size and selection intensity, 68–70 characterization drought-prone target environments, 47 parameters, drought-occurrence, 46–48 production systems and rainfall distribution, 45–46 systems analysis, 48 gene discovery, 78–79 genotype by environment (GE) interactions, 43 integrated strategy, 86–88 marker-assisted breeding and selection

application, 73–76 QTLenvironment interaction, 77 QTLgenetic background interaction, 77 quantitative trait loci (QTLs), grain yield, 71–72 selective genotyping and bulk segregant analysis, 72–73 phenotyping, concepts and tools field-managed drought screening, 57–59 FTSW dry-down approach, 59–60 gene expression and profiling, 63 model-based, 62–63 nondestructive methods, 61–62 soil moisture, control and monitoring, 60–61 responses dehydration-avoidance mechanisms, 52–57 plant water usage, 49–50 spikelet sterility and grain failure, 50–52 transformation approaches functional protein genes, 79–84 transcription factors and signaling genes, 82–84 E Echinochloa spp., 155, 157, 161–162, 167, 184, 191, 197 Euphorbia heterophylla, 155, 156, 159 F Food security concept. See Grain security, China Fraction of transpirable soil water (FTSW) dry-down approach, 59–60 plant gas exchange and soil drying, 49–50 G Generation challenge program (GCP), 85 Generic Model Organism Database (GMOD), 84–85 Grain-for-Green program, 113–114 Grain production, China achievements disputation of grain security, 114–115 self-sufficiency, 115–116 annual grain yield vs. variation coefficient, 110 average grain yield, 105 crisis, 103–104 development stages, 109 golden period, 1979–1984, 111–112 natural disasters and political campaign effect, 1958–1978, 111 price policy change effect, 1985–2003, 112–114 2004–till now, 114 war and production policy, 1949–1957, 110–111

221

Index

experiences advanced agricultural techniques, 119 government undertake, 116–118 self-sufficiency maintenance, 118 storage in farmer’s house, 119–120 super hybrid rice breeding technique, 119 supply, 105–106 Grain security, China assessment indicators, 108–109 Brown, Lester predictions, 114–115 case studies 2008 activity, high grain yield, 141–143 science and technology engineering, 139–141 characteristics, 107–108 concepts, 106–107 FAO, 107 land security, 122 problems and challenges cultivated land loss and degradation, 120–122 global climatic change, 126–127 inferior climatic conditions, 124–125 life standard change, 127 market risk resistance, 128–129 natural disasters, 124–125 old agricultural capital construction, 128–129 population growth change, 127 small-scale agricultural economy, 127–128 vulnerable ecosystems, 125–126 water resources problems, 123–124 strategy options and countermeasures agricultural infrastructures reinforcement, 129–130 climatic change, 137 farmers’ initiative, 136–137 fiscal input increase, 130–131 grass agriculture development, 137–138 population growth rate control, 138–139 quantity and quality protection, cultivated land, 133–134 science and technology role maximization, 131–133 water resources protection, 135 super hybrid rice breeding technique, 119 Greenhouse gas (GHG) emission, anthropogenic cultivated soil, 24–26 emission sources, 3 enteric fermentation carbon dioxide, 17–18 CH4 emission, 15–17 global contribution and climatic change, 3–4 postharvest activity hydroelectric vs. coal energy, 30 transportation, 31–32 waste and biomass, 32–33

I Imperata cylindrica, 156, 158, 159 International Crop Information System (ICIS), 85–86 International Rice Functional Genomics Consortium (IRFGC), 86 L Leaf water potential (LWP), 52 Life cycle assessment (LCA), 8–9 Livestock production systems (LPSs) grassland-based, 14 landless (LL), 13 mixed farming, 14 ruminants and nonruminants, 11–12 Livestock’s long shadow (LLS), 3 M Mulching, weed control, 168 N National Grain Bureau of China, 119–120 O Oryza spp., 155, 157, 162–163, 174, 197 P Phenotyping, concepts and tools field-managed drought screening control, initiation and severity, 59 screening sites and reasons, 58–59 FTSW dry-down approach, 59–60 gene expression and profiling, 63 model-based, 62–63 nondestructive methods, 61–62 soil moisture, control and monitoring, 60–61 R Relative water content (RWC), leaf, 52–53 Rhamphicarpa fistulosa, 157, 158, 163, 171 Rice breeding program, conventional type developing breeding populations, 68 germplasm screening environment stable genotypes, 65 upland and aerobic areas, 65 heritability (H), 66–67 high yield combination, 70 population size and selection intensity, 68–70 S Sloped Land Conservation Program. See Grain-for-Green program

222

Index

Sphenoclea zeylanica, 155, 157, 161 Striga spp., 155, 156, 158, 160–161 Super hybrid rice breeding technique, 119 W Way Rarem allele, 52 Weed management, Africa annual rice import savings, 154 biological weed control, 178–179 cereals, 150–151 chemical weed control conventional methods, 179–181 herbicide resistant rice technologies, 181–184 integrated weed control, 185 cultural weed control flooding, 166–167 mixed cropping, rotations, and fallow, 168–172 mulching, 168 planting methods, 165–166 soil fertility management, 167–168 importance, 153–154 problems changing climates, 186–188 demography, 186 herbicide resistance, 188–189

research, 193 climate change, 194 crop management, 194–195 ecosystems, 197–198 herbicides, 196–197 socioeconomics and gender, 198–199 varietal development, 195–196 rice ecosystems, 151 potential annual import savings, 154 production, 150 rain-fed lowlands, 153 upland cropping systems, 151 varietal development, 174–178 species, problem weeds, 155–158 strategy, 189 integrated crop and weed management approaches, 192 prioritization, weed species, 190–192 timing, weed control and crop management, 192–193 weed species problem weeds, 155–158 rain-fed and irrigated lowlands, 161–163 uplands and hydromorphic zones, 159–161 usefulness, 163–164