Modeling Genetic Improvement among Peanut (Arachis Hypogaea L.) Cultivars in Ghana and Burkina Faso

MISSING IMAGE

Material Information

Title:
Modeling Genetic Improvement among Peanut (Arachis Hypogaea L.) Cultivars in Ghana and Burkina Faso
Physical Description:
1 online resource (161 p.)
Language:
english
Creator:
Narh, Stephen
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Agronomy
Committee Chair:
Boote, Kenneth J
Committee Members:
Tillman, Barry
Mylavarapu, Rao S
Erickson, John E
Jones, James W

Subjects

Subjects / Keywords:
crop-modeling -- cropgro-model -- genetic-improvement -- genotype-environment
Agronomy -- Dissertations, Academic -- UF
Genre:
Agronomy thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Yield trials are an integral part of plant breeding programs and are used to evaluate the yield potential and stability of selected lines. The objectives of this study were to evaluate yield potential,G × E interactions, stability of peanut genotypes, calibrate the CROPGRO-Peanut model for the cultivars screened in field trials, and investigate simulated cultivar traits contributing to genetic improvement of pod yield using the CSM-CROPGRO Peanut model. Twenty peanut genotypes were tested at two sites in Ghana and two sites in Burkina Faso in 2010 and 2011. Among the lines with broad adaptability, ICGV-IS 96814 was considered the best because it produced the highest pod yield (1755 kg/ha) over all environments and had a regression coefficient close to unity (b =1.06). However, released cultivar NKATESARI was considered equivalent in some respects because it had pod yield equal to ICGV-IS 96814, but with a higher regression coefficient. Cultivar coefficients were solved by optimization procedure with the CROPGRO-Peanut model using data from the Ghana sites. The solved cultivar coefficients provided simulated pod yield that agreed quite well with the observed pod yield. There was considerable genetic variation in solved cultivar traits among cultivars. Model evaluation with independent pod yield data from the Burkina Faso sites, not used in model calibration, provided a good test of how well model predictions worked in an independent validation. The derived cultivar coefficients allowed the CROPGRO-Peanut model to mimic yield ranking quite well. Model sensitivity analysis showed that LFMAX, XFRT, PODUR, SFDUR, and THRSH influenced yield, and percent yield enhancement due to increases in XFRT and LFMAX were largest in general.The farmer check cultivars CHINESE and TS 32-1 had the lowest yield of 1199 and 1254 kgha-1, respectively. Improved cultivars NKATESARI and ICGV-IS 96814were among the top yielding cultivars, indicating that the analyzed virtual traits are all feasible and were achieved to a large extent in these improved cultivars compared to the farmer check cultivars. Higher yielding cultivars had longer life cycle, greater leaf spot resistance, higher photosynthesis, and higher partitioning.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Stephen Narh.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Boote, Kenneth J.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-08-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045814:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 MODELING GENETIC IMPROVEMENT AMONG PEANUT ( Arachis hypogaea L.) CULTIVARS IN GHANA AND BURKINA FASO By STEPHEN NARH A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Stephen Narh

PAGE 3

3 To my mom and dad

PAGE 4

4 ACKNOWLEDGMENTS I am highly indebted to Dr. Kenneth Boote (committee chair) for giving me the opportunity to study at the University of Florida, and for his support and encouragement during my program I would also like to thank my committee members Dr. Barry Tillman, Dr. Jim Jones, Dr. Rao Mylavarapu and Dr. John Erickson for their contributions. I would like to thank Dr. Jesse Naab for hosting me at the Savannah Agricultural Research Station (SARI Wa), Ghana for my field research and for his contributions and support. I would like to thank Dr Mumuni Abudulai, SARI Nyankpala, Dr. ZagreBertin, and Dr. Philippe Sankara both at the University of Ouagadougou for their role in multi location trials and data collection. To the staff of the Savannah Agricultural Research Statio n, Wa Ghana I say a big thank you for your help with field work Many thanks also to my wife Charlotte Narh for coming all the way from Accra on a number of occasions to help with data collection both at Wa and Nyankpala. I would like to acknowledge the support of my parents during my studies. To all who contributed in various ways towards the success of this work, I am very grateful. Finally, I give thanks to God whose love and abundant grace sustained me throughout this program.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 8 LIST OF FIGURE S ........................................................................................................ 11 LIST OF ABBREVIATIONS ........................................................................................... 12 ABSTRACT ................................................................................................................... 13 CHAPTER 1 INTRODUCTION .................................................................................................... 15 2 LITERATURE REVIEW .......................................................................................... 18 Genotype, Environment, and Genotype by Environment Interactions .................... 18 Adaptation and Yield Stability ................................................................................. 20 Assessment of Yield Stability .................................................................................. 21 Diseases of Peanut in West Africa .......................................................................... 22 West African Climatology ........................................................................................ 23 West African Climate Variability ....................................................................... 24 Climate Change Scenarios for West Africa ...................................................... 26 Peanut Growth and Yield Responses to Elevated Temperature, CO2, and Drought ................................................................................................................ 27 Elevated Temperature, CO2, and Drought Tolerance Traits in Crops ..................... 30 Physiological Traits and Yield Improvement ........................................................... 33 Virtual Crop Cultivars .............................................................................................. 35 3 SCREENING 20 PEANUT CULTIVARS FOR YIELD AND GENOTYPE BY ENVIRONMENT STABILITY IN MULTI LOCATION TRIALS IN BURKINA FASO AND GHANA ................................................................................................ 37 Background ............................................................................................................. 37 Materials and Methods ............................................................................................ 40 Experimental Site and Design .......................................................................... 40 Measurement of Disease Incidence and Yield ................................................. 41 Statistical Analysis ............................................................................................ 42 Results and Discussion ........................................................................................... 44 Growth Environment ......................................................................................... 44 Plant Stand ....................................................................................................... 45 Disease Assessment ........................................................................................ 46 Yield and Yield Components ............................................................................ 47 Pooled Analysis of Variance ............................................................................. 49

PAGE 6

6 Genotype by Environment Interaction Analysis ................................................ 50 4 GROWTH AND YIELD POTENTIAL OF DIFFERENT PEANUT CULIVARS ACROSS MULTIPLE SITES ANALYZED WITH THE CROPGRO PEANUT MODEL ................................................................................................................... 61 Background ............................................................................................................. 61 Materials and Methods ............................................................................................ 65 Experimental Sites and Design ........................................................................ 65 Measurement of Disease Incidence and Yield ................................................. 66 Measurement of Crop Growth .......................................................................... 67 The CROPGRO Peanut Model and Model Inputs ............................................ 68 Calibration of Cultivar Coefficients ................................................................... 69 Optimization Procedure .................................................................................... 70 Results and Discussion ........................................................................................... 74 Reproductive Growth Stages ............................................................................ 74 Optimization ..................................................................................................... 74 Variation in Cultivar Coefficients ....................................................................... 75 Independent Evaluation of Derived Cultivar Coefficients .................................. 76 Relationship of Yield to Cultivar Coefficients .................................................... 77 Evaluation of Calibration Procedure ................................................................. 78 5 YIELD RESPONSES TO HYPOTHETICAL VARIATION IN GENETIC TRAITS OF DIFFERENT PEANUT CULTIVARS WITH THE CSM CROPGRO PEANUT MODEL ................................................................................................................... 91 Background ............................................................................................................. 91 Materials and Methods ............................................................................................ 93 Results and Discussion ........................................................................................... 96 Yield Responses to Life Cycle .......................................................................... 96 Yield Responses to Disease ............................................................................. 97 Linking SFDUR to SDPM ................................................................................ 98 Yield Responses to Genetic Traits other than Life Cycle and Disease ........... 100 Effect of Trait Combinations on Pod Yield ...................................................... 101 Maximum Possible Yield ................................................................................ 102 Geno type by Environment Interaction Analysis .............................................. 102 Weather induced Variability in Pod Yield as Indicator of Traits for Yield Stability ........................................................................................................ 103 6 SUMMARY AND CONCLUSIONS ........................................................................ 119 APPENDIX A LEAFSPOT RATING SCALES ............................................................................. 125 B SOIL PROPERTIES AT THE EXPERIMENTAL SITES ........................................ 126 C LEAFSPOT DISEASE DATA FROM EXPERIMENTAL SITES ............................. 127

PAGE 7

7 D SOIL FERTILITY FACTOR OF SOILS .................................................................. 143 LIST OF REFERENCES ............................................................................................. 144 BIOGRAPHICAL SKETCH .......................................................................................... 161

PAGE 8

8 LIST OF TABLES Table page 3 1 Names of peanut genotypes used in the screening trials conducted in 2010 and 2011. ........................................................................................................... 53 3 2 Climate data during the growing season (JuneOctober) for the locations where the trials were conducted in 2010 and 2011. ........................................... 53 3 3 Water holding characteristics of the soils for Wa, Nyankpala, Farakoba and Gampela. ............................................................................................................ 54 3 4 Mean plant stand for 19 genotypes over 4 locations in 2010 and 2011. ............. 54 3 5 ICRISAT score, leafspot (percent necrosis) at 60 and 90 days after sowing and necrosis progress (slope) for 20 genotypes ................................................ 55 3 6 Pooled mean performance for pod yield, seed yield, unit seed weight, shelling percentage and Pod Harvest Index for 19 p eanut genotypes ............... 56 3 7 Contribution of the individual sources of variation in the pooled analysis of variance for pod yield of the peanut genotypes over all environments. .............. 56 3 8 Mean performance for pod yield (kg/ha) for 19 peanut genotypes at 4 locations in 2010 and 2011. ................................................................................ 57 4 1 List of peanut cultivars screened in 2010 and 2011 at four sites and used to analyze growth and yield potential with the CROPGRO Peanut model .............. 81 4 2 Climate data dur ing the growing season (JuneOctober) for Wa, Nyankpala, Farakoba and Gampela in 2010 and 2011. ........................................................ 82 4 3 Days from sowing to specif ic reproductive growth stages for peanut cultivars at Wa, Ghana in 2010. ........................................................................................ 82 4 4 Steps in the calibration of cultivar coefficients with an optimizer using data from Wa and Nyankpala in Ghana in 2010 and 2011. ........................................ 83 4 5 Cultivar coefficients of individual peanut cultivars derived using the CROPGRO Peanut model with an optimizer. ..................................................... 84 4 6 Observed (obs) and simulated (sim) pod yield, and measures of agreement for baseline starting point, assuming all 19 c ultivars were CHINES E ................ 85 4 7 Observed (obs) and simulated (sim) pod yield, and measures of agreement after calibration of corr ect life cycle ................................................................... 85

PAGE 9

9 4 8 Observed (obs) and simulated (sim) pod yield, and measures of agreement after calibration of life cycle and entering disease data ..................................... 85 4 9 Observed (obs) and simulated (sim) pod yield, and measures of agreement after calibration of l ife cycle, disease, and genetic coefficients .......................... 86 5 1 Cultivar coefficients of individual peanut cultivars derived using the CROPGRO Peanut model with an optimizer. ................................................... 106 5 2 Life cycle traits of three standard life cycle peanut cultivars for Wa Farakoba, Gampela and Nyankpala. ................................................................................. 107 5 3 ICRISAT score, leafspot (percent necrosis) at 60 and 90 days after sowing and necrosis progress (slope) for 20 genotypes .............................................. 108 5 4 Cultivar coefficients (low, median, and high for five traits) of individual virtual peanut cultivars derived for short cycle cultivars. ............................................. 109 5 5 Percent necrosis der ived from disease data at each site and assigned to susceptible, moderately susceptible, and resistant peanut cultivars ................ 110 5 6 Percent defoliation der ived from disease data at each site and assigned to susceptible, moderately susceptible, and resistant peanut cultivars ................ 111 5 7 Simulated pod yield (kg/ha) of short, medium, and long life cycles at susceptible, moderately susceptible and resistant disease. ............................. 112 5 8 Simulated pod yield of short, medium, and long life cycles at susceptible, moderately suscept ible and resistant disease comp ared with two versions .... 112 5 9 Pod yield for 5 traits as percent change from standard (median) values for three life cycle cultivar s at three disease l evels. .............................................. 113 5 10 Simulated pod yield of hypothetical and actual peanut cultivars averaged over 30 years of weather data over Nyankpala, Wa, Farakoba and Gampela. 114 5 11 Simulated possible pod yield averaged over 30 years of weather data for a long cycl e virtual cultivar with perfect disease control, and high partitioning. ... 115 5 12 Weather induced coefficient of variation of pod y ield for different cultivar traits at Wa, Nyankpala, Gampela, and Farakoba. .................................................... 116 A 1 ICRISAT 1 9 rating scale (based on Subrahmanya m et al., 1995) ................... 125 B 1 Some physical and chemical properties of the soils used in the model. ........... 126 C 1 Data of leaf spot score and corresponding percent necrosis and percent defoliation for Farakoba, 2010. ......................................................................... 127

PAGE 10

10 C 2 Data of leaf spot score and corresponding percent necrosis and percent defoliation for Farakoba, 2011. ......................................................................... 129 C 3 Data of leaf spot score and corresponding percent necrosis and percent defoliation for Gampela, 2010. .......................................................................... 1 31 C 4 Data of leaf spot score and corresponding percent necrosis and percent defoliation for Gampela, 2011. .......................................................................... 133 C 5 Data of leaf spot score and corresponding percent necrosis and percent defoliation for Nyankpala, 2010. ....................................................................... 135 C 6 Data of leaf spot score and corresponding percent necrosis and percent defoliation for Nyankpala, 2011. ....................................................................... 137 C 7 Data of leaf spot score and corresponding percent necrosis and percent defoliation for Wa, 2010. ................................................................................... 139 C 8 Data of leaf spot score and corresponding percent necrosis and percent defoliation for Wa, 2011. ................................................................................... 141 D 1 Soil fertility factor (SLPF) of the soils at the different locations used in the model. ............................................................................................................... 143

PAGE 11

11 LIST OF FIGURES Figure page 3 1 Adaptive responses to environment for pod yield of 19 peanut genotypes. ........ 58 3 2 Relationship between mean pod yield and regression coefficient of 19 peanut genotypes. .......................................................................................................... 59 3 3 Genotype pod yield (kg/ha) against site mean yield for genotype NC7, NKATESARI, and ICGVIS 96814. ..................................................................... 60 3 4 Genotype pod yield (kg/ha) against site mean yield for genotype GM 123 and ICGV IS 92101. .................................................................................................. 60 4 1 Simulated (lines) and observed (points) pod yield, biomass, and Pod Harvest Index for NKATESARI over 2 sites over 2 years. ............................................... 87 4 2 Simulated (lines) and observed (points) Pod Harvest Index of ICGV IS 96814, GUSIE BALIN (92099), GM 57, and NKATESARI for Wa, 2011. ........... 88 4 3 Simulated versus observed pod yield for Gampela and Farakoba, using cultivar coefficients derived from dat a collected at two sites in Ghana. .............. 89 4 4 Simulated versus observed pod yield for Nyankpal a and Wa using cultivar coefficients derived from data collected at Nyankpala and Wa, Ghana. ............. 89 4 5 Relations hip between pod yield and XFRT and LFMAX of 19 cultivars. Yield is averaged over 2 years over 4 sites. ................................................................ 90 4 6 Relationship betw een pod yield and SFDUR and life cycle of 19 cultivars. Yield is averaged over 2 years over 4 sites. ....................................................... 90 5 1 Adaptive responses to environments for pod yield of 11 peanut genotypes ..... 117 5 2 Relationship between mean pod yield and regression coefficient of 19 peanut genotypes ......................................................................................................... 118

PAGE 12

12 LIST OF ABBREVIATIONS ANOVA Analysis of V ariance DAS Days After Sowing DSSAT Decision Support System for Agrotechnology Transfer ICRISAT International Crops Research Institute for the Semi Arid Tropics

PAGE 13

13 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MODELING GENETIC IMPROVEMENT AMONG PEANUT ( Arachis hypogaea L.) CULTIVARS IN GHANA AND BURKINA FASO By Stephen Narh August 2013 Chair: Kenneth J. Boote Major: Agronomy Yield trials are an integral part of plant breeding program s and are used to evaluate the y ield potential and stability of selected lines. The objectives of this study were to evaluate yield potential, G E inter actions, stability of peanut ge notypes, calibrate the CROPGRO P eanut model for the cult ivars screened in field trials, and investigate simulated cultivar traits contributing to genetic improvement of pod yield using the CSM CROPGRO Peanut model. Twenty peanut genotypes were tested at two sites in Ghana and two sites in Burkina Faso in 2010 and 2011. Among the lines with broad adaptability ICGV IS 96814 was considered the best because it produced the highest pod yield (1755 kg/ha) over all environments and had a regression coeff icient close to unity (b =1.06) However, released cultivar NKATESARI was considered equivalent in some respects because it had pod yield equal to ICGV IS 96814, but with a higher regression coefficient Cultivar coefficients were sol ved by optimization procedure with the CROPGRO Peanut model using data from the Ghana sites. The solved cultivar coefficients provided simulated pod yield that agreed quite well with the observed pod yield. There was considerable genetic variation in solved cultivar t raits among cultivars Model evaluation with independent pod yield data from the Burkina

PAGE 14

14 Faso sites, not used in model calibration, provided a good test of how well model predictions work ed in an independent validation. The derived cultivar coe fficients allowed the CROPGRO Peanut model to mimic yield ranking quite well. Model sensitivity analysis showed that LFMAX, XFRT, PODUR, SFDUR, and THRSH influenced yield, and percent yield enhancement due to increases in XFRT and LFMAX were largest in general. The farmer check cultivars CHINESE and TS 321 had the lowest yield of 1199 and 1254 kg ha1, respectively. Improved cultivars NKATESARI and ICGV IS 96814 were among the top yielding cultivars, indicating that the analyzed virtual traits are all feas ible and were achieved to a large extent in these improved cultivars compared to the farmer check cultivars. Higher yielding cultivars had longer life cycle, greater leaf spot resistance, higher photosynt hesis, and higher partitioning.

PAGE 15

15 CHAPTER 1 INTRODUCTION Peanut or groundnut (Arachis hypog aea L) is an important food and cash crop due to its high protein (0.250.30) and edible oil (0.480.50) content, compared with other oil seed crops. The average yields of groundnut in most parts of West Africa are lower (903 kg/ha) than those in South Afric a (2000 kg/ha), Asia (1798 kg/ha) or the rest of the world (1447 kg/ha) (FAO 2005). The lower yields in West Africa are attributed considerably to leaf spot disease (Naab et al. 2005) as well as low soil fertility and water limitation. Early leaf spot caus ed by Cercospora arachidicola and late leaf spot caused by Cercosporidium personatum are critical yield limiting diseases of groundnut in W est Africa (Waliyar 1991; Waliyar et al. 2000), accounting for yield reductions of 5070 % where fungicides are n ot used (Waliyar 1991; Shokes and Culbreath, 1997). Production of peanuts is mainly rain fed. As a result, high variability in the amount and distribution of rainfall, both within and between seasons cause droughts of varying lengths and severity. Farmers yields are therefore reduced below expected yields. Also, the soils in the peanut growing regions of West Africa have low fertility, especially phosphorus and low water holding capacity which further compounds the problems contributing to low yields. Climate change is expected to result in gradual increases in temperature, rainfall variability and prevalence of extreme events such as droughts and floods. Climate change therefore poses a serious and continuing threat to development. Scholes and Biggs (200 4), referring to Sub Saharan Africa as the food crisis epicenter of the world, concluded that projected climate change during the first half of the twenty first century will make this situation worse. Climate change will add burdens to those who are

PAGE 16

16 already poor and vulnerable (IPCC, 2007). Overall, crop yields in Africa may fall by 10 20 % by 2050 because of warming and drying, but there are places where yield losses may be much more severe, as well as areas where crop yields may increase (Jones a nd Thornt on, 2003). Many developing countries in Africa are seen as being highly vulnerable to climate variability and change, in part because they have only a limited capacity to adapt to changing circumstances (Thomas and Twyman, 2005). As a result, management an d genetic improvement of different crop species and cultivars may be effective options to decrease small scale farmer vulnerability to climate. Judicious cultivation of crop species and cultivars that differ in their maturity cycle and sensitivity to speci fic stresses like drought and heat stress should provide a better buffering mechanism against climate variability Crop performance depends on the genotype, the environment and the interaction between genotype and environment. Identifying superior cultivars with a high and stable yield across a range of environments is the primary aim of breeding program s. However, the inherent differential responses of crop genotypes to varying environmental conditions or genotype by environment interactions require that ne w lines be evaluated in multienvironment trials (METs) over a set of sites, years or both to assess their yield performance and stability (Kang, 1990). This assessment is the primary basis for identifying the best performers and their range of adaptabilit y, both temporal ly and spatial ly By evaluating the performance of different peanut cultivars in multi environments, new cultivars adapted to these environments can be identified. The objectives of this study were to evaluate yield potential, G E intera ctions, stability of peanut ge notypes, calibrate the CROPGRO P eanut model for the cultivars

PAGE 17

17 screened in field trials, investigate simulated cultivar traits contributing to genetic improvement of pod yield and evaluate the yield response of different peanut cultivars to variation in genetic traits under different weather and soil conditions using the CSM CROPGRO Peanut model.

PAGE 18

18 CHAPTER 2 LITERATURE REVIEW Genotype, Environment, and Genotype by Environment Interactions With regard to the comparison of plant material in a set of multi environment yield trials, the term genotype refers to a cultivar (i.e. with material genetically homogeneous, such as pure lines or clones, or heterogeneous, such as openpollinated populations) rather than to an individuals genetic make up. The term environment relates to the set of climatic, soil, biotic (pests and diseases) and management conditions in an individual trial carried out at a given l ocation in one or several years. In particular, an environment identifies a given locationyear combination in the analysis of trials repeated over time. Genotype by environment interaction refers to the inherent differential responses of crop genotypes to varying environmental conditions which may lead to changes in the performance ranking of test genotypes (Kang, 1990; Cooper and Delacy, 1994; Delacy et al., 1996a; Annicchiarico, 2002a, b). Crop yield is a quantitative trait that generally exhibits large genotype by environment interactions. Consequently, differences in yield performance between genotypes may vary widely among environments (Delacy et al., 1990; Annic chiarico, 1997). When genotype by environment (G E) interacti on effects are absent environmental effects reflecting the different ecological potential of sites and managem ent conditions are not of direct concern for the breeding or recommendation of plant varieties. In that case, g enotypic main effects (i.e. differences in mean yield between genotypes) provide the only relevant information. However, differences between genotypes may vary widely among environments in the presence of GE interaction effects as large as those reported in extensive investigations by DeLacy et

PAGE 19

19 al. (1990) and Annicchiarico (1997a). In general, G E interactions are considered a hindrance to crop improvement in a target region (Kang, 1998). Moreover, such effects may contribute, together with purely environmental effects, to the temporal and spatial instability of cr op yields. Temporal instability mostly due to weather variability has a negative effec t on farmers income and, in the case of staple crops, contributes to food insecurity at national and household level s. On the other hand, G E interactions may offer opportunities, especially in the selection and adoption of genotypes showing positive interaction with the location and its prevailing environmental conditions (exploitation of specific adaptation) or of genotypes with low frequency of poor yield or crop failure (exploitation of yield stability) (Simmonds, 1991; Ceccarelli, 1996). Growing awar eness of the importance of G E interactions has led crop genotypes to be assessed in multi environment, regional trials for cultivar recommendation or for the final stages of elite breeding material selection. The most important G E effects for targeting c ultivars or for selection of material are the crossover types affecting top yielding genotypes. Major interaction can be expected when there is: on the one hand, wide variation between genotypes for morphophysiological characters conferring resistance to ( or avoidance of) one or more stresses, and, on the other, wide variation between environments f or incidence of the same stress (es) (as determined by climatic, soi l, biotic and management factors) Large GE interactions have frequently been reported between pairs of environments with contrasting levels of one major stress (Ceccarelli, 1989; Bramel Cox, 1996), with environments defined as favorable when characterized by low stress and high mean yield and unfavorable with high stress and low yield. However, large interactions may

PAGE 20

20 also occur between pairs of unfavorable environments and even between pairs of moderately favorable environments possessing similar mean yield but with differing combinations of stresses or patterns of one major stress (Annicchiar ico, 1997a) Adaptation and Yield Stability In a plant breeding context, adaptation is the ability of the plant material to be high yielding with respect to a given environment or given conditions to which it is adapted (Cooper and Byth, 1996). In breeding for wide adaptation (i.e. adaptability), the aim is to obtain a variety which performs well in nearly all environments; in breeding for specific adaptation, the aim is to obtain a variety which performs well in a definite subset of environments within a target region. The adaptive response of a variety is assessed with respect to other genotypes and tends to undergo modification when better performing germplasm becomes available. Breeding for wide adaptation and for high yield stability have sometimes been considered the same, insofar as the two terms indicate a consistently good yield response across environments. Some authors, however, have applied the yield stability concept with respect to consistency in time of genotype performance, using the adaptation concept in relation to consistency in space (Barah et al. 1981; Lin and Binns, 1988; Evans, 1993). It has also been widely acknowledged (Ghaderi et al. 1980; Becker, 1984; Lin and Butler, 1988; Annicchiarico, 1992; Romagosa and Fox, 1993; Piepho et al. 1998) that genotype location (GL) interaction, rather than all kinds of GE interaction, is useful for depicting adaptation patterns, as only this interaction can be exploited by selecting for specific adaptation or by growing specifically adapted genoty pes. The size of the GL effects is also relevant: in particular, if the variation due to GL interaction (although statistically significant) is small compared to other

PAGE 21

21 components, particularly the genotypic one, it increases the possible advantage of breeding for wide adaptation. Generally, the adaptive response of a genotype is assessed with respect to other genotypes utilizing multi environment data (Annicchiarica, 2002a). To capture much of the GL interaction that exists in the target area and to be able to separate it from the yearly inconsistencies, data are needed on yield performance of the test genotypes for a number of sites that cover the target area and for multiple years. Assessment of Yield Stability High yield stability usually refers to a genotypes ability to perform consistently, whether at high or low yield levels, across a wide range of environments. High yield stability may be associated with low mean yield (or low stability with high mean yield), which complicates genotype selection or recommendation. Most stability measures relate to either of two contrasting concepts of stability: static (Type 1) and dynamic (Type 2) (Becker and Lon, 1988; Lin et al. 1986). According to the static stability, a stable genotype tends to maintain a constant yield across environments or tends to maintain a relatively consistent ranking among other genotypes across environments. Dynamic stability implies for a stable genotype a yield response in each environment that is always parallel to the mean response of the tested genotypes, i.e. zero GE interaction. Type 1 stability relates to consistency both in time and in space, i.e. across environments belonging to the same or different sites. The practical interest of combining high levels of mean yield and yi eld stability has led to the development of the yield reliability concept (Eskridge, 1990; Kang and Pham, 1991; Evans, 1993). A reliable genotype is characterized by consistently high yield across environments. This facilitates genotype selection or recomm endation, as the mean yield and the yield

PAGE 22

22 stability traits are combined into a unique measure of genotype merit. Considering yield stability in conj unction with mean yield may provide a more appropriate comparison of genotypes than when using only mean yield (Kang, 1993). Domitruck et al (2000) indicated that the analysis of variance procedure is a useful tool for estimating the existence and magnitude of G x E interactions. Yates & Cochran (1938) proposed a purely statistical analysis, which was later used by Finlay and Wilkinson (1963) to analyze and measure the magnitude of G x E interactions. Diseases of Peanut in West Africa Diseases are major yieldlimiting factors for groundnut in West Africa (Gillier, 1982; Naab et al., 2005). A number of fungal, vi ral and nematode diseases of groundnut have been reported from West Africa (Emechebe, 1980; Khan and Misari, 1987; Subrahmanyam et al., 1988a). Most of the diseases are widespread in the region, but only a few are economically important throughout the regi on. Diseases that are considered to be regionally important are leaf spots, rust, and rosette. The lower yields in West Africa are attributed considerably to leaf spot disease (Naab et al. 2005) as well as low soil fertility and water limitation. Early lea f spot caused by Cercospora arachidicola and late leaf spot caused by Cercosporidium personatum are critical yield limiting diseases of groundnut in W est Africa (Waliyar 1991; Waliya r et al 2000), accounting for yield reductions of 5070 % where fungi cides are not used (Waliyar 1991; Shokes and Culbreath, 1997). The pathogens may survive from season to season on plants and in infected crop debris. No authentic host species are known outside the genus Arachis. Long distance distribution of the pathogens include air borne conidia, movemen t of infected crop debris, or movement of pods or seeds that are surface contaminated with conidia or infected crop debris. There is no evidence of

PAGE 23

23 either pathogen being internally seed borne (McDonald et al ., 1985). Leaf spots damage the plant by reducing the available photosyntheti c area by lesion formation, damaging the photosynthetic apparatus of neighbouring, apparently healthy tissues and inducing early leaf abscission They also stimulate leaflet abscission (Boote e t al., 1983). In addition to causing leaf spots, the two pathogens also produce lesions on petioles, stems and pegs. When disease attack is severe the affected leaflets first become chlorotic, then necrotic; lesions often coalesce; and leaflets are shed (S mith, 1984). West African Climatology West Africa is that part of Africa that lies approximately between 5 N and 20 N and occupies an area of approximately 5 million km2. It is bounded on the west and south by the Atlantic Ocean and on the north by the S ahara desert. The eastern border lies on a line running from the Cameroon Mountains to Lake Chad. West Africa covers four tropical regions which include humid areas such as the equatorial forests in the south, semi humid areas such as the Guinea savanna, s emi arid areas such as the Sudan savanna, and the arid Sahel in the north. The climate of West Africa is characterized by wet and dry seasons. The weather pattern is associated with the northward and southward migration of a narrow zone of reversal in the meridional wind, called the Intertropical Discontinuity (ITD). It is a region of tradewind confluence, which produces weak horizontal pressure gradients responsible for weak winds at the surface. Another commonly used term in the literature is the Intertr opical Convergence Zone (ITCZ), associated with the zone of maximum convection. Both the ITD and ITCZ exhibit seasonal migration following the seasonal movement of the overhead sun. The wind systems associated with the ITD are characterized mainly by the northeasterly and southwesterly trade winds. During the wet season, the moist southwest monsoon with

PAGE 24

24 its maritime characteristics from the Gulf of Guinea invades the region, bringing with it cool breezes. It is often associated with convection and cloudines s. The northeast trade wind, which characterizes the dry period, on the other hand is continental, hot, dry and dust laden because of its long track from the Sahara desert. West African Climate Variability The inhabitants of West Africa depend largely on rain fed agriculture for sustenance. As a result, climatic factors play an important role in determining the production of food crops in West Africa. The region is characterized by low rainfall that is highly variable over space and time, which constitutes a limiting potential for crop yields (Graef et al., 2001; Tesfaye et al., 2004) and for increased agricultural production necessary for food security. The amount of rainfall and its distribution over the year (especially during the farming season) greatly affects the productivity of agriculture in these regions. The inter annual and inter decadal rainfall variability in West Africa, as in other African countries, has been documented by several investigators in numerous publications. From one year to the next, there may be more than a 30 % variation in the length of the rainy season (Roudier et al., 2011). There has been a substantial reduction in rainfall in West Africa over the second half of the 20th century, with a clear break between 1968 and 1973. The reduction is extremely clear in the Sahel. The Sudan and Guinea zones were also affected by the drop in precipitation during this period (OECD, 2009). Since the mid1990s, a return to better rainfall conditions has been noted, in particular in continental Sahel (Niger, Northern Nigeria and Chad), though, these conditions went hand in hand with greater inter annual rainfall variability. There is a strong link between inter annual rainfall variability in West Africa and patterns of sea surface temperature (SS T) anomalies in the tropical Atlantic, Pacific and

PAGE 25

25 Indian Oceans (Ward et al., 1990; Folland et al., 1991; Ward, 1992; Shinoda and Kawamura, 1994; Nicholson and Grist, 2001; Rowell, 2001; Bader and Latif, 2003; Giannini et al., 2005). The El NioSouthern Oscillation (ENSO) in the tropical Pacific Ocean (e.g., Nicholson and Kim, 1997; Nicholson et al., 2000; Nicholson and Selato, 2000; Hulme et al., 2001) has been confirmed as one of the more important factors influencing rainfall variability for some regions in Africa. Hulme et al. (2001) in their detailed analysis of African climate change observed a strong ENSO relationship for equatorial east Africa (high rainfall during a warm ENSO event) and southern Africa (low rainfall during a warm ENSO event), cons istent with earlier studies. Elsewhere in Africa, West Africa in particular, there has been a controversy on the influence of the ENSO on rainfall. While there is a general consensus among researchers on ENSOs influence in some regions, for instance the G uinea coast, where it tends to increase rainfall (Nicholson, 2001), there is a controversy over its influence in the Sahel regions of West Africa. A weak correlation between ENSO and Sahelian JuneAugust drying has been observed (Ward et al., 1990; Folland et al., 1991; Ward, 1992; Shinoda and Kawamura, 1994; Nicholson and Grist, 2001; Rowell, 2001; Bader and Latif, 2003; Giannini et al., 2005), which is consistent with observations by Ropelewski and Halpert (1987). The different opinions among several auth ors are due to the complex nature of ENSO's influence in the region. In West Africa there is very little variation in temperatures which are generally high throughout the year. However, day night temperature variations are greater, as much as 1015oC (even more in the Sahel of West Africa), whereas inter annual variations remain between 6 and 10 oC (OECD, 2009). Significant increasing

PAGE 26

26 temperature trends have also been found in each of the following regions: all of Africa, Northern Hemisphere Africa, Souther n Hemisphere Africa, tropical Africa, and subtropical Africa (Collins 2011). However, even without ascertaining the specific causes, Collins (2011) demonstrated that a significant rise in African temperatures occurred between 1979 and 2010. Climate Change Scenarios for West Africa Africa is considered the worlds most vulnerable region with respect to the effect of climate change. However, it is still difficult to assess the extent and nature of such changes in the future partly due to the complexity of it s regional climates and geography. Climate models are relatively useful when it comes to forecasting temperature changes in Africa. It has been confirmed that in the 21st century, global warming will be more intense in Africa than in the rest of the world. The average rise in temperature b etween 1980/99 and 2080/99 will be between 3 and 4 oC for the continent as a whole, 1.5 times greater than at the global level (IPCC, 2007). The increase will be less marked in coastal and equatorial areas (+3 oC ) and the highest increase will take place in the Western region of Africa (+4 oC). Climate change scenarios for the West African region include an anticipated increase in mean surface temperature of up to 0.5 oC per decade (OECD, 2009). On the other hand, there is still a fair amount of uncertainty in rainfall related projections for West Africa. North Africa, Southern Africa and East Africa are some of the regions where there is less uncertainty. Africas Mediterranean coast has been projected to experience a decrease in precipitation ( 15 to 20 %) between 1980/99 to 2080/2 0 99. Less rainfall is also expected in Southern Africa during the winter and especially in spring. Results from models show an increase in rainfall in East Africa. However, General Circulation Models have shown a limited

PAGE 27

27 capacity to forecast West African climate, especially rainfall (OECD, 2009). As a result, no conclusions can be drawn regarding rainfall in West Africa. Peanut Growth and Y ield R esponses to Elevated T emperature, CO2, and D rought Global climate change has emerged as an important environmental challenge due to its potential impact on biological systems of planet Earth (Walther et al., 2002). Human activities such as deforestation and burning of fossil fuel are mainly responsible for the recent rapid increases in atmospher ic concentrations of greenhouse gases including CO2 (Kaufmann and Stern, 1997; Houghton et al., 2001; Stott et al., 2001). At the present rate of emission, CO2 concentration is projected to be in the range of 540 970 ppm by the end of this century, which will potentially increase global near surface temperatures by 1.4 5.8 oC (Houghton et al., 2001). Peanut or groundnut ( Arachis hypogaea L.) is an important grain legume crop and is grown as a principal source of edibl e oil and vegetable protein. About 90% of the worlds peanut production occurs in the tropical and semi arid tropical regions, which are characterized by high temperature and low or erratic rainfall. The mean optimal air temperature range for vegetative gr owth of peanut is between 25 and 30 oC, which is warmer than the optimum range for reproductive growth, which is between 22 and 24 oC (Wood, 1968; Cox, 1979; Ong, 1984). Both short and long term exposure to air and soil temperatures above optimum can caus e significant yield loss in peanut (Dreyer et al., 1981; Ketring, 1984; Ong, 1984; Golombek and Johansen, 1997; Prasad et al., 1999a, b, 2000a, b). Effects of short term (1 6 days) exposure to daytime temperature between 28 and 48 oC during reproductive development and yield were thoroughly investigated (Prasad et al., 1999a, b, 2000a, 2001; Craufurd et al., 2002, 2003). It was observed that day temperature >34 oC

PAGE 28

28 decreased fruit s et and resulted in fewer pods (Prasad et al., 1999b, 2000a). Decreased fruit set at high temperatures was mainly due to poor pollen viability, reduced pollen production and poor pollen tube growth, all of which lead to poor fertilization of flowers (Prasad et al., 1999b, 2000a, 2001). Increasing daytime temperature from 26 30 to 34 36 oC significantly reduced the number of subterranean pegs and pods, seed size and seed yield by 30 50% (Cox, 1979; Ketring, 1984; Ong, 1984). Prasad et al. (2000b) investigated the effects of daytime soil and air temperature of 28 and 38oC respectively from start of flowering to maturity, and reported 50% reduction in pod yield at high temperatures. Further studies by Prasad et al. (2003) show ed that increasing temperature decreased peanut seed yield as a result of lower seedset due to poor pollen viability, and smaller seed size due to decreased seed growth rates and decreased shelling percentages. Elevated CO2 was observed to increase the photosynthetic rate and pod yield of peanut under both fully irrigated and drought conditions (Chen and Sung, 1990 ; Clifford et al., 1993, 2000; Mortley et al., 1997; Stanciel et al., 2000). Increased photosynthetic rates at elevated CO2 is a common phenomenon in most C3 plant species, including peanut grown under irrigated or water stress conditions (Chen and Sung 19 90; Clifford et al., 1993, 2000; Stronach et al., 1994). Stomatal opening is highly sensitive to changes in CO2; increasing CO2 significantly decreased stomatal conductance and leaf transpiration. Decreases in stomatal conductance and leaf transpiration rates at elevated CO2 have been reported for peanut (Stronach et al., 1994; Clifford et al., 1993, 2000).

PAGE 29

29 Peanut is mostly grown under rain fed conditions and so is frequently subjected to drought stresses of different duration and intensities. Drought stres s has adverse influence on water relations (Babu and Rao, 1983) as well as growth and yield of groundnut (Suther and Patel 1992). Bell et al. (1993) studied the factors influencing dry matter partitioning in four diverse groundnut cultivars. Rates of dry matter accumulation in pods ( and pod addition) varied significantly with both cultivar and sowing date. Root growth of groundnut was also influenced by soil moisture as water stress stimulated the growth of roots into deeper soil. Meisner and Karnok (1992) observed root growth on rhizotron glass every week and found that the groundnut root system, regardless of water stress, did not exhibit signs of senescence. The ability of groundnut to maintain a viable root system during water stress may contribute to t he crops drought resistance ( Sanders et al. 1993). In peanut the rate of flower production is reduced by drought stress during flowering but the total number of flowers per plant is not affected due to an increase in the duration of flowering (Gowda and Hegde, 1986; Meisner and Karnok 1992). Boote and Ketring (1990) observed that the start of flowering is not delayed by drought stress. However, peg elongation, which is turgor dependent, is delayed due to drought stress. Adequate pod zone moisture is crit ical for the development of pegs into pods. Bennett et al. (1990) reporte d that pod formation is reduced by a dry pod zone. However, Boote et al. (1992) reported that Florunner and Robut 33 1 produced pods in air dry soil although at a slower rate. Pod and seed development are progressively inhibited by drought stress due to insufficient plant turgor and lack of assimilates. The progress to these stages can also be inhibited by lack of soil water in the pod zone (Boote and Ketring 1990; Stirling and Black 1991). Pod dry weights were significantly

PAGE 30

30 reduced by a 30day water stress during the pod development stage (Meisner and Karnok 1992). The number of pods per plant can be low due to increases in soil resistance associated with low soil water content (Sha rma and Sivakumar 1991) Water deficits during seed development reduce pod and seed weight. Shelling percentage is also reduced by moisture stress during seed development (Janamatti et al. 1986). Elevated Temperature, CO2, and Drought Tolerance Traits in Crops Plants show a wide range of compensating, escape, or tolerance traits or mechanisms for drought and heat stress through various molecular, biochemical, physiological, developmental or growth adaptations. One of the important components of tolerance to d rought is enhanced soil water capture, which is possible by increased exploration of soil by the roots. Deeper roots, increased root length density in deeper soil layers as opposed to upper layers, faster root depth progression or a shift in root profi le shape will enable water absorption from greater depth provided the soil is deep and is recharged by rain. Cultivars with deeper root systems, when compared with shallow root systems, are generally more tolerant to drought stress during critical stages o f crop development. J ordon and Miller (1980) suggested that sorghum cultivars with greater root length density at greater depths would increase water uptake and encounter less water stress during grainfilling stages. Cultivars with larger root length dens ities and deeper rooting systems were found to be more tolerant to drought stress conditions in soybean (Sponchiado et al., 1989; Sloane et al., 1990; Hudak and Patterson, 1996). However, it is also important to consider that a plant with greater root leng th density may have greater access but use up the water rapidly, grow rapidly, and deplete the shallow water profile faster. Such cultivars may become severely stressed

PAGE 31

31 during later stages of reproductive development, which may result in lower yields. This is particularly important if the soil is shallow Plant water use efficiency (WUE) is a n important trait associated with drought and is defined as the amount of biomass or yield accumulated per unit of water used. Several studies have shown positive r elat ions between WUE and yield in peanut (Wright et al., 1988; Wright, 1989) and cowpea (Craufurd et al., 1998) Genetic variation for WUE has been observed in several crops such as wheat (Farquhar and Richards, 1984; Merah et al., 2001), barley (Hubick and Farquhar, 1989), sorghum (Peng and Krieg, 1992), and peanut (Hubick et al., 1986). Improvement of cultivars using this trait has not been very successful because of the complexity of the trait and also the difficulty in measuring WUE under field conditions ( Ismail and Hall, 1992). With the advent of molecular breeding, identification of QTL and their use in the breeding programs may help develop new cultivars with drought tolerance. Recently, the ERECTA gene has been associated with transpiration efficiency ( Masle et al., 2005). They showed that expression of ERECTA gene resulted in reduced stomatal frequency and conductance and greater photosynthetic rates, resulting in increased WUE under a wide range of water regimes Several studies have shown that shorter duration cultivars escape drought because they complete their life cycle before the occurrence of drought, whereas long duration cultivars have greater chances of being exposed to severe drought or heat stress, particularly, during the later stages of crop development. With respect to the importance of phenology to drought and heat stress, studies have shown that short duration peanut varieties developed at the International Crops Research Institute for the

PAGE 32

32 Semi Arid Tropics (ICRISAT) had higher pod yield over local longer duration cultivars when exposed to terminal drought in India (ICRISAT, unpublished data) Leaves senesce early in response to drought and heat stress particularly when these stresses occur during the post flowering stages of seed filling. Some genotypes tolerate drought during grain filling by keeping their leaves green; these cultivars are termed as stay green types. Similarly, the stay green trait and chlorophyll retention in leaves under heat stress conditions is considered an expression of heat tolerance (Fokar et al., 1998). Stay green genotypes retain chlorophyll in their leaves and maintain the ability to carry out photosynthesis longer than the senescent types, and are often shown to have a yield benefit (Borrell et al., 2001; Jordan et al., 2003). Borrell et al. (2001) showed that stay green types assimilate more nitrogen and have greater specific leaf nitrogen content, suggesting the link between the stay green trait and nitrogen and there is often no yield penalty associated with t he stay green trait (Borrell et al., 2001). Drought tolerance can often be associated with heat tolerance, since decreased transpirational cooling leads to increased tissue temperatures. Heat tolerance is a helpful auxiliary to drought tolerance under many conditions, since low water potential causes stomatal closure that leads to decreased transpiration, which in turn increases tissue temperatures. Studying the performance of drought tolerant cultivars under heat stress, Kakani et al. (2002) observed that peanut genotypes that were grown in the tropics and semiarid tropics (e.g., 55 437 grown in subSaharan Africa) showed tolerance to drought stress. Furthermore, genotype Kadiri 3, a known drought tolerant cultivar, was highly susceptible to high temperature stress, whereas cultivar ICGV -

PAGE 33

33 86015, which is susceptible to drought, was tolerant to heat stress. Similar results were also observed in wheat where a cultivar exhibiting greater performance under heat stress was highly susceptible to drought. These studies indicate that even though drought and heat stress can occur together in most of the regions (particularly semiarid tropics), the possible physiological or biochemical mechanisms operating to induce escape or tolerance to each of these stresses may be different. Furthermore, the characteristics of traits associated with drought and heat stress could be different. Therefore, selection breeding of genotypes for tolerance to combined stresses of drought and heat mus t be performed under stress conditions that include both of these stresses. An increase in CO2 is beneficial for crops especially for legumes, because legumes are CO2responsive C 3 species that also fix nitrogen, which removes the N fertility constraint t hat limits CO2 responsiveness of nonlegumes. The primary effect of higher CO2 is to stimulate leaf and canopy photosynthesis leading to increases in mass of leaf, stem, root, pod, and seed, as well as an increase in C supply for N fixation (Allen and Boote, 2000). Increase in specific leaf weight is a trait that can help maintain yield under elevated CO2. Elevated CO2 generally increases specific leaf weight (SLW), which by itself would normally increase photosynthesis capacity. However, elevated CO2 may ca use a downregulation of the amount of rubisco protein, especially in cereals. Downregulation in soybean is minimal as leaf N concentration shows only a slight decrease (Campbell et al. 1988; Allen et al. 1988). Physiological Traits and Yield Improvement Selection during crop improvement requires that there is sufficient genetic variability within the plant population. Substantial variation in physiological traits

PAGE 34

34 associated with heat and drought tolerance such as flowering time, root traits, water use eff iciency (WUE), amount of water transpired, stay green, and specific leaf area (SLA) have been reported in cereals and legumes (Dwivedi et al., 2007a). Also, reliable and easily measurable phenotypic screens are available for many of these traits (Dwivedi et al., 2007a). Studies at ICRISAT (unpublished) have shown genetic variation in flowering time for peanut cultivars. Harvest index (HI) is directly related to yield as it represents the proportion of total biomass partitioned into grain. As a result, i ncre ased HI has been a major factor in the improvement of grain yield in many crops ( Richards, 2000). Pod HI is a useful trait to assess progress in improving crop yield potential (Shorter et al., 1991) In peanut, high WUE and HI can lead to improvement in cr op yiel d (Sinclair, 1998). Nigam et a l. ( 1991) reported that selection based on a combined index of HI, WUE and water transpired (T) was effective in improving yield of peanut under drought stress conditions. In a study to identify photosynthetic traits wi th potential to be used in breeding programs designed to increase peanut yield, James (1982) measured net photosynthesis (Pn) on 32 diverse genotypes. 2s 1PAR) eight genotypes had Pn rates of 20 to 30, seventeen 31 to 40 and seven 41 to 43 mg CO2 dm 2 h 1. At any given light level tested, Pn rates of a given genotype were significantly different from other genotypes Duncan et al. (1978) using five peanut cultiv ars, studied the physiological changes made in yield improvement that were responsible for the great increase in peanut yield potential over the years. Quantitative estimates of the physiological factors responsible for the dry weight differences were made by computer simulation using the PENUT Z model. Differences in three physiological processes explained most of the yield variation among th e five peanut cultivars; the partitioning of

PAGE 35

35 assimilate between vegetative and reproductive parts, the length of the filling period, and the rate of fruit establishment. Of these, the partitioning of assimilate had the greatest effect on fruit yield. Specific leaf area is another trait of interest and it has been used in a largescale screening program for improved drought resistance in Australia (Nageswara Rao et al. 2000a). This research group has demonstrated progress using physiological trait s to indirectly select for high pod yield of peanut under water limited conditions. After two cycles of selection, they selected progenies that yielded 30 % more than their parents under drought stressed conditions (Nageswara Rao et al. 2000b). Virtual Crop Cultivars B reeding for higher yield over the years has been base d mainly on empirical selection for yield per se. This approach has been successful in increasing yield of various crops in the past, but further progress is becoming increasingly diffi cult (Araus et al., 2002). However, more eff ort has gone into the ident ifi cation of traits that breeders might select for to increase yield indirectly (Donald, 1968; Blum, 1988; Lawn, 1989; Ludlow and Muchow, 1990; Evans, 1993; Jackson et al., 1996; Fukai et al., 1999; Monneveux et al., 2008; Phakamas et al., 2008a). Although the concept of plant ideotype has long been proposed, it has never been used directly (Rasmusson, 1991; Sedgley, 1991). This is because the desired trait or combination of traits actually has to be incorporated into a genotype to be able to evaluate their eff ects on yield performance (Marshall, 1991). Crop simulation models have potential for cre ating virtual crop cultivars, for further assisting the breeders selection criteria, and for genetic enhancement of important traits that contribute to yield impr ovement in different target environments

PAGE 36

36 (Hammer and Jordan, 2007) The ability of these models to predict growth and yield as infl uenced by local environmental conditions, agronomic practices, and cultivar traits off ers an opportunity to evaluate the eff ects of a trait or a combination of traits on yield through model simulation without an actual incorporation of these traits into a genotype. Using simulation m odels, the eff ect of traits can be assessed with sensitivi ty analysis in which the coeffi cients that determine a trait are varied and the eff ects on simulated growth or yield observ ed ( White, 1998). These models thus can be used to identify a desirable trait or a combination of traits leading to the design of a crop ideotype for a specific environment (White and Hoogenboom, 2003). The target environments necessarily must be described in terms of weather (over multiple seasons), water availability, soil physical and chemical constraints, desired crop life cycle, and management. Since these models incorp orate cultivar specific parameters that represent genetic traits of cultivars, their effects on crop performance can be evaluated in target environments. These traits express the genetic potential of each cultivar to determine their yield in a given environment

PAGE 37

37 CHAPTER 3 SCREENING 20 PEANUT CULTIVARS FOR YIELD AND GENOTYPE BY ENVIRONMENT STABILITY IN MULTI LOCATION TRIALS IN BURKINA FASO AND GHANA Background Peanut is widely used as an oilseed crop and as a direct source of human food around the world due to its high protein content (0.250.30) and edible oil (0.480.50) content, compared with other oil seed crops. Peanut is insensitive to daylength and so flowers over a wide range of daylengths. It is usually grown during the rainy season under rain fed conditions. Peanut is a self pollinated plant and its normal method of commercial propagation is by seed. Inflorescences are borne in the axils of leaves on both primary and secondary branches. Peanut is now cultivated on 24 million ha in more than 100 countries between 40o N and 40o S, generating an annual production of nearly 37 Tg (FAO, 2011). The average yields of groundnut in most part of West Africa are lower (903 kg/ha) than those in South Africa (2000 kg/ha), Asia (1798 kg/ha) or the rest of the wor ld (1447 kg/ha) (FAO 2005). The lower yields in West Africa are attributed considerably to leaf spot disease (Naab et al. 2005) as well as low soil fertility and water limitation. Early leaf spot caused by Cercospora arachidicola and late leaf spot caused by Cercosporidium personatum are critical yield limiting diseases of groundnut in West Africa (Waliyar 1991; Waliyar et al. 2000), accounting for yield reductions of 5070 % where fungicides are not used (Waliyar 1991; Shokes and Culbreath, 1997). Un ited Nations projections estimate that the world population will continue to grow to about 10 billion by 2050 (FAO, 1996). The increase in population and the subsequent rise in the demand for agricultural produce are expected to be greater in

PAGE 38

38 regions where production is already insufficient, in particular in subSaharan Africa and south Asia (PinstrupAndersen et al 1999). The necessary increase in agricultural production represents a huge challenge to local farming systems and must come mainly from increased yield per unit area, given the limited scope for extension of cultivated land worldwide (Evans, 1998). In less favorable areas with poor ecological potential for crop production, food insecurity also depends on the marked climatic fluctuations from year to year (McCown et al 1992). To increase the availability and stability of agricultural production in the future in view of projected climate variability such as high temperatures and drought, breeding program s need to produce improved germplasm capable of maximizing the agricultural potential of specific areas and farming systems, and of minimizing the occurrence of crop failures or very low yields in unfavorable years. Many developing countries in Africa are seen as being highly vulnerable to climate variability and change, in part because they have only a limited capacity to adapt to changing circumstances (Thomas and Twyman, 2005). As a result, management and genetic improvement of different crop species and cultivars may be an effective option to decrease small scale farmer vulnerability to climate. Judicious cultivation of crop species and cultivars that dif fer in their maturity cycle and sensitivity to specific stresses like drought and heat stress should provide a better buffering mechanism against climate variability. Improved adaptation and yield stability may be derived from the appropriate choice of cul tivars (whether indigenous or foreign, and either traditional or released from public or private breeding institutions) by using plant material with increased tolerance to prevailing biotic and abiotic stresses. Successful new genotypes must have high yiel d and other essential agronomic traits

PAGE 39

39 A major goal of plant breeding programs is to increase and stabilize crop yield across a range of environments. Genotype by environment (G E) interactions make it difficult to compare genotypes over a range of year s and locations. They affect the extent of genetic progress through plant breeding, and they also affect the efficiency of varietal choice for farmers (Becker, 1981). However, it is also essential to evaluate the G E interaction, because a good and effic ient genotype must remain good, whatever the year and location or environmental conditions (Hill, 1975). The G E interaction, which is associated with the differential performance of materials tested at different locations and in different years, influences selection and recommendation of genotypes (Lin et al., 1986; Annicchiarico, 1997). Several analysis methods have been proposed, with the aim of explaining G E interaction (Becker and Leon, 1988; Lin et al., 1986; Piepho, 1994; Truberg and Huehn, 200 0). These methods can be divided into two major groups, univariate and multivariate stability statistics (Lin et al., 1986). The most widely used one is the regression method, which is based on regressing the mean value of each genotype on the environmental index or marginal means of environments (Yates and Cochran, 1938; Finlay and Wilkinson, 1963; Eberhart and Russell, 1966). Finlay and Wilkinson (1963) concluded that stability was defined by the regression coefficient and adaptability was defined by the relative mean yield of the cultivar. The study of genotypes according to their slope through joint regression analysis provides information on both stability and adaptation (Rharrabti et al., 2003). The objective of this study was to conduct cultivar screening trials to identify cultivars with superior performance among test entries and to identify the range of

PAGE 40

40 adaptability of the different peanut cultivars in the screening trials on the basis of yield stability. Materials and Methods Experimental Site and Design Twenty peanut genotypes (Table 31) were tested in two sites in Ghana and two sites in Burkina Faso in 2010 and 2011. The experiments in Burkina Faso were conducted at the Environmental and Agricultural Research Institute in Gampela, Ouagadougou (120 2551N, 10 2218W) and Farakoba, Bobodilaso (110 560N, 40960W). The Ghana experiments were conducted at the Savanna Agr iculture Research Institute sites at Nyankpala (90 42 N, 00 92W) and Wa (100 3 N, 20 50 W) A randomized complete block d esign with three replicates was used in all trials. Seeds were sown in each plot (4 m 2 m) at a spacing of 0.1 m within rows and 0.5 m between rows, resulting in a planned plant population of 20 plants m2 and four rows. One seed was sown per hole. Gap f illing wa s done where necessary to improve uniform crop establishment. Actual stand count was recorded at harvest. Before sowing, soil tests were carried out for pH, phosphorus (P) and potassium (K) to determine the need for liming as well as levels of P and K. In some sites, t he preemergence herbicide Pendimethalin was used in combination with hand weeding to control weeds. There was no fungicide application to allow screening for leafs pot resistance and/or tolerance, and to mimic farmer production practices. Table 32 shows the climate data during the growing season (JuneOctober) of the locations where the trials were conducted in 2010 and 2011. Table 33 shows the water holding characteristics of the soils at the four locations for this study. The soil at Wa was a sandy loam and it is 50 cm deep. It has plant available water of 60.0 mm The low moisture characteristics and shallow nature

PAGE 41

41 of this soil likely will confer some water stress to peanut grown on this soil. The soil at N yankpala was a sandy loam and it is 113 cm deep. It has a plant available water of 116.1 mm The experiment at Gampela was conducted on loam y sand and it is 210 cm deep. This soil has a plant available water of 184.0 mm In the case of Farakoba, the soil w as a sandy loam and it is 200 cm deep. Plant available water for this soil is 184.8 mm The soil at Farakoba is characterized by a subsurface accumulation of clay (> 45 %) which can lead to poor internal drainage. Aluminum content also increases down the p rofile to over 2 ppm. Both of these characteristics may affect root development of crops. Measurement of Disease Incidence and Yield Crop management data was collected on planting date, plant population and row spacing. Using the ICRISAT scale which ranges from 1 to 9 (Subrahmanyam et al., 1995) disease score was recorded, which depends on visual estim ate of necrosis and defoliation (Appendix A 1) A disease rating of 1 means no disease and a r ating of 2 means lesions present on lower leaves with no defoliation. Ratings of 3 to 8 are associated with increasing levels of defoliation and necrosis. A rating of 9 implies defoliation of almost all leaves leaving bare stems, with any leaflets present having severe leafspots. The ICRISAT leafspot disease rating was converted to percent necrosis (eq.31) and to percent defoliation (eq.32), both equations being modified forms of the equations developed by Maninder Singh et al. ( 201 1 ) between percent necrosis and the Florida 110 visual rating scale and between percent defoliation and the Florida 110 visual rating scale. The modified equations are: ( % ) = 1 36 10 9 1 45 (3 1)

PAGE 42

42 ( % ) = 12 5 12 5 (3 2) where IS represents the ICRISAT visual score. At the end of the season, pods from the two bordered inner rows were handharves ted and pods removed manually. A sub sample of six plants was taken from the large plot yield sample. The plants were depodded and the fresh weight of pods, leaf and stem was recorded (excluding taproot) The subsamples were oven dried and the dry weights of pods, leaves and stems as well as the number of full sized pods not collapsed after drying were obtained. The dry matter concentrations and ratios of the sub samples were then used to calculate pod yi eld, stem and leaf weight, total biomass and the number of pods for each plot. Pod yield was computed from large plot harvest on a dry weight basis, after adding the subsampled pods back in. Pod Harvest Index was obtained by dividing the subsample pod ma ss by the total biomass of the subsample. The whole plot biomass was calculated by dividing t he whole plot pod yield by the P od Harvest Index. One hundred ovendried pods were randomly selected from the sub sample and shelled. Shelling percentage was meas ured by dividing the seed dry weight by the weight of seed plus shell, expressed as a percentage. One hundred seeds were randomly selected to determine the weight of onehundred seeds and the weight per seed. Analysis of variance was conducted on cultivar pod yields, seed yields, plant stands, unit seed weight and shelling percentage. Repeated Measures analysis of variance was conducted for the time series disease data. Statistical Analysis Statistical analyses were performed using analysis of variance proc edures in the GLIMMIX procedure of SAS (SAS Institute, 2009). Cultivar was considered fixed effect and replicate, year and location as random effects. Degrees of freedom were

PAGE 43

43 determined using the KenwardRoger method. Where significant ( P < 0.05) effects w ere observed, pairwise comparisons were made using t he LSMEANS statement with TUKEY method. Homogeneity of variance test was not significant, so a combined analysis of variance was conducted for pod yield for the four locations The yield response Rijkr of the genotype i in the location j year k and block r is: = + + + + ( ) + ( ) + + ( ) + (3 3) Where is the overall mean of all plots in all environments, Gi is the effect of genotype i Lj is the effect of location j Yk is the effect of year k, (GY)ik is the interaction between genotype i and year k, (GL)ij is the interaction between genotype i and location j ( L Y )jk is t he interaction between location j and year k, (GYL)ijk is the interaction between genotype i location j and year k, eijkr is random error This analysis was to help determine the relative contributions of the various sources of variation as well as the significance of the genotype by environment interaction. Once the significance of the genotype by environment interaction was established, conventional linear regression analysis was used to determine whether any of the genotypes or pairs of genotypes exhibited a characteristic response to environmental change. Statistical analysis of plant stand, seed yield, unit seed weight and shelling per centage followed equation 33. The leafspot disease response Rijkr t of the genotype i in the location j year k, block r and time t is: = + + + + + ( ) + ( ) + ( ) + ( ) + ( ) + ( ) + (3 4) Where (GT )it is the interaction between genotype i and time t Tt is the effect of time t ( L T )j t is the interaction between location j and time t (YT )k t is the interaction betw een

PAGE 44

44 year k and time t and eijkrt is random error All other letters have the same meaning as stated earlier Determination of the adaptive responses of peanut lines to environmental conditions for the different locations was based on stability parameters derived from yield stability analysis. In this approach, each environment was characterized by the mean yield of all cultivars or genotypes and the mean yield was used as an index of productivity of the site ( Finlay and Wilkinson, 1963). Yields of individual cultivars were regressed against the index, and the slope of the regression lines determined. Regression coefficients were obtained for the regression of genotype mean yield for each location upon the mean of all genotypes for each location (site mean yield). A regression coefficient greater than 1.0 indicates the adaptability of the cultivar to high yielding environments, a v alue less than 1.0 indicates the adaptability of the line to low yielding environments Cultivars with high mean yields and a regression coefficient close to unity are considered stable with broad adaptation to all environments. Results and Discussion Grow th Environment Environm ental conditions during the 2010 and 2011 growing seasons (JuneOctober) for the four locations are provided (Table 33 ). In 2010, Wa recorded a mean temperature of 27.1 oC and Nyankpala 27.3 oC. Farakoba and Gampela had mean temperatures of 26.3 oC and 28.5 oC respectively. Total rainfall from June October for Wa in 2010 was 696.3 mm which occurred in 59 rainfall events compared to 995.4 mm for Nyankpala in 68 rainfall events. A total of 720.5 mm of rainfall was recorded at Gampela comprising of 57 rainfall events. Farakoba had the highest amount of rainfall (1017.8 mm) and number of rainfall events (78) during the 2010 growing season. During

PAGE 45

45 the 2011 growing season, a mean temperature of 27.2 oC, 27.5 oC, 26.8 oC and 29.1oC was obs erved for Wa, Nyankpala, Farakoba, and Gampela, respectively. In 2011, Wa recorded 60 rainfall events resulting in a total growing season rainfall of 734.6 mm. In the case of Nyankpala, total growing season rainfall amounted to 964.6 mm which occurred in 54 rainfall events. Farakoba had a total of 644 mm of rainfall in 76 rainfall events and Gampela received 629.5 mm of rainfall in 56 rainfall events. Plant Stand Table 34 shows the mean plant stand of 19 peanut cultivars over all environments (LocationYear) Plant stand across all environments ranged from 11.5 plants/m2 for CHINESE to 6.8 plants/m2 for ICGV IS 96895. The plant stand of 11.5 plants/m2 for CHINESE was the highest and this was significantly higher ( P < 0.05) than all the other cultivars. This was followed by TS 321 with a plant stand of 10.6 plants/ m2. Cultivar ICGV IS 96895 recorded the lowest plant stand of 6.8 plants/ m2. The early maturing check cultivars had relatively higher plant stands with CHINESE, TS 32 1 and DOUMBA LA having 11.5, 10.6 and 9.6 plants/m2 respectively compared to the longer season cultivars. The plant stands for most of the cultivars across all environments were not significantly different ( P < 0.05). The low plant stands were mostly due to the failure of some seeds to germinate at all. In some cases, seeds germinated, but wilted as the season progressed. Germination tests carried out before planting in Wa did show that some seeds had reduced or lower seed viability in 2011 and lower seedling vigor. The lower seed viability and seedling vigor was responsible for the very low plant stands in some cultivars like ICGV IS 96895, G122TX95 and GUSIE BALIN (92099) in 2011.

PAGE 46

46 Disease Assessment Data on ICRISAT score, leafspot (percen t necrosis) and slope of necros is progress at 60 and 90 days after sowing (DAS) for 20 peanut cultivars over four locations and two years are presented (Table 35). At 60 DAS, DOUMBALA, which is one of the early maturing check cultivars, recorded the highest percent necrosis of 4.1 % ac ross all environments. Percent necrosis for the short season check cultivars DOUMBALA, CHINESE, and TS 321 did not differ (P< 0.05). Apart from DOUMBALA, the early season check cultivars CHINESE and TS 321 as well as the GM cultivars GM 123, GM 57, and G M 515 had similar percent necrosis (P< 0.05) at 60 DAS. Cultivar G122TX95 which is a longer season cultivar had the lowest percent necrosis of 1.7 % at 60 DAS. The early season cultivars generally had higher percent necrosis compared to the longer season c ultivars at 60 DAS. This can be attributed to the earlier onset of the leaf spot disease in the shorter season check cultivars At 90 DAS, the highest percent necrosis was observed for the shorter season check cultivars with TS 321, CHINESE and DOUMBALA r ecording values of 10.6 %, 10.5 % and 10.3 % respectively. Cultivars DOUMBALA, CHINESE, and TS 32 1 did not differ (P< 0.05) in percent necrosis at 90 DAS. However, percent necrosis for the three shorter season check cultivars DOUMBALA, CHINESE, and TS 321 was significantly higher (P< 0.05) than that for the longer season cultivars at 90 DAS. Again, the lowest percent necrosis of 6.0 % was recorded for cultivar G122TX95. Leafspot disease progress, estimated as the slope of percent necrosis from 60 to 90 D AS showed that the shorter season check cultivars DOUMBALA, TS 321, and CHINESE had significantly higher (P< 0.05) percent necrosis slope values than the longer season cultivars. DOUMBALA had a slope value of 0.137, TS 321 recorded a slope of 0.147

PAGE 47

47 and a slope of 0.152 was observed for CHINESE. The higher slopes in the shorter season cultivars are an indication of a faster leaf spot disease progression. Cultivars G122TX95, PC 79 79 and B106TX95 had the lowest slopes with values of 0.097, 0.089 and 0.081, respectively. Leaf spot disease progression was therefore slowest for cultivars G122TX95, PC 7979 and B106TX95 between 60 and 90 DAS. However, among the longer season cultivars, improved cultivars F MIX, ICGV IS 96814 and NKATESARI recorded quite low dise ase slope values of 0.098, 0.098 and 0.102, respectively. The higher percent necrosis at 90 DAS and higher slope values between 60 and 90 DAS for the shorter season check cultivars can be attributed to the earlier onset of the disease and a faster disease progression. The shorter season check cultivars reached maturity around 90 DAS and the disease effect on yield was likely to be more pronounced. Among the longer season cultivars, the GM cultivars GM 123, GM 515, and GM 57 generally had higher percent necr osis at 60 and 90 DAS and also recorded higher slope values between 60 and 90 DAS compared to the other longer season cultivars. The GM cultivars are therefore more susceptible to leaf spot disease and had a faster leaf spot disease progression than other longer season cultivars. The generally lower percent necrosis and slope values of the longer season cultivars may be due to the fact that, the longer life cycle resulted in a slower disease progression for these cultivars, or they were less susceptible to leaf spot disease. Yield and Yield Components Table 36 shows the seed yield, unit seed weight, shelling percentage, pod harvest index (over 3 locations) and pod yield (over 4 locations) for 19 cultivars in 2010 and 2011. Cultivar F MIX was not included as the seed source was poor for 2010. The longer season cultivars generally had higher yield than the early maturing ones. This

PAGE 48

48 may be attributed to the fac t that the longer season cultivars flowered and matured later than the early maturing ones. This would have provided more vegetative growth (leaf area) for the longer season cultivars to increase photosynthesis. Also, the longer season cultivars generally are more likely to have longer duration for seed filling because they had a longer duration from first pod stage to maturity, though t here is a lag period after first pod before seed development begins. These cultivars also had less leaf spot disease. The highest pod yield was observed for ICGV IS 96814 with a yield of 1755 kg/ha follo wed by NKATESARI with a pod yield of 1722 kg/ha. Cultivars ICGV IS 96814 and NKATESARI were not different (P < 0.05) in pod yield. Cultivars ICGV IS 92093, F MIXSINK 24, an d NC 7 also did not differ from each other in pod yield (P < 0.05) with values of 1581 kg/ha, 1524 kg/ha and 1496 kg/ha, respectively. Leaf spot resistance may have contributed to longer seed fill and higher seed filling rate, which may have resulted in the higher yield observed for the longer season cultivars The lowest pod yield of 805 kg/ha was observed for cultivar GM 515. Poor plant stands may also be responsible for the low pod yield observed for some of the cultivars Seed yield generally followed the same trend as pod yield. The longer season cultivars generally were the higher yielding cultivars in terms of seed yield. Cultivar NKATESARI recorded the highest seed yield of 1203 kg/ha followed by ICGV IS 96814 with a seed yield of 1124 kg/ha. These two cultivars did not differ in seed yield (p < 0.05). The lowest seed yield of 488 kg/ha was observed for cultivar ICGV IS 96895 Unit seed weight ranged from 0.328 g for c ultivar G122TX95 to 0.431 g for cultivar NKATESARI. Unit seed weight for cultivar NKATESARI was higher than the unit seed weight of all other cultivars (P < 0.05). Cultivars NC 7, GM 123 and ICGV IS

PAGE 49

4 9 92101 were the only cultivars besides NKATESARI with unit seed weight over 0.400 g, with values of 0.407 g, 0.404 g, and 0.401 g respectively Shelling percentage, expressed as the ratio of seed to seed plus shell is also shown in Table 36. Cultivar GM 57 recorded the highest shelling percentage of 67.6 % followed by cultivar G122TX95 with a value of 67.1 %. The lowest shelling percentage of 49.9 % was observed for cultivar ICGV IS 96895. Apart from cultivar ICGV IS 96895, all other cultivars had shelling percentage above 60 %. The very low shelling percentage for cultivar ICGV IS 96895 may be due to a low partitioning to seeds during the seed filling period. Another reason could be that the seasons were not long enough for seed filling to be completed before harvest at the various locations. It also had the poorest germination judged from the plant stand which could also be associated with th e low shelling percentage of this cultivar Cultivar ICGV IS 92101 recorded the highest pod harvest index (HI) of 0.425. This was followed by NC 7, ICGV IS 96814 and NKATESARI with pod HI of 0.420, 0.405 and 0.400, respectively. Pod HI for these four culti vars did not differ (P >0.05). Cultivars G122TX95, ICGV IS 96895 and B106TX95 recorded the lowest pod HI of 0.240, 0.240 and 0.220, respectively. High pod HI means there was more partitioning to reproductive plant parts in such cultivars. Low pod HI cultiv ars therefore, had low partitioning to pods. Pooled Analysis of Variance Test for homogeneity of error variance was used and the nonsignificance (P > 0.05) of the test was an indication of variance homogeneity, so combined analysis of variance (ANOVA) was performed. The combined ANOVA was performed to determine the significance or otherwise of the variance components. Results of the combined

PAGE 50

50 analysis of variance (Table 37) showed highly significant differences among genotypes (G) and locations (L) (P < 0.0 5) for pod yield. However, Years (Y) was not significant (P< 0.05). The results of the combined ANOVA for pod yield also showed that Cultivar Location, Year Location and Cultivar Year Location interactions were highly significant (P < 0.05). Signif icant Y x L interaction indicated that location means were inconsistent across the two years. Significance of genotype x location interaction indicated that it may be due to either by crossover (qualitative) interaction, in which a significant change in ranking occurs from one environment to another (Singh et al ., 1998; Akram et al ., 1999) or a non crossover interaction (Quantitative), G x L interaction, in which case the ranking of genotypes remains constant across environments and the interaction is significant because of change in the magnitude of response (Cooper, 1999; Honarnejad, 2003) Genotype by Environment Interaction Analysis G enotype and environment mean pod yield data of 19 peanut cultivars over all environments (LocationYear) is provided in Table 38. There was a large variation in both genotype and environment mean pod yield. The highest yielding environment was Gampela with a pod yield of 1754 kg/ha followed by Wa and Nyankpala with pod yield of 1036 kg/ha and 963 kg/ha, respectively. Farakoba was the lowest performing environment with mean pod yield of 938 kg/ha. The low perf ormance observed for Farakoba may be attributed to the high clay content of the soils, which created flooded conditions (poor internal drainage) after heavy rainfall especially in 2010. This soil was also characterized by a subsurface accumulation of Alumi num which may have affected root growth. Genotype mean yield ranged from 805 kg/ha for GM 515 to 1755 kg/ha for ICGV IS 96814. Figure 3 1 indicates the adaptive responses of the individual cultivars

PAGE 51

51 or genotypes to environments Regression coefficients wer e obtained for the regression of genotype mean yield for each location upon the mean of all genotypes for each location A regression coefficient greater than 1.0 indicates the adaptability of the cultivar to high yielding environments, a value less than 1.0 indicates the adaptability of the cultivar to low yielding environments Cultivars with high mean yields and a regression coefficient close to unity are considered stable with broad adaptation to all environments. Figure 3 2 illustrates the pod yie ld of each cultivar across environments against the regression coefficients. This relationship measures the spatial stability (adaptability) of the lines to mostly soil conditions and associated climatic environments The broken lines represent 1 standard error. Genotypes ICGV (FDRS) 20 F MIX 39, GUSIE BALIN (92099 ) ICGV IS 92093, ICGV IS 92101 and ICGV IS 96814 are considered cultivars with broad adaptability, as they have above average mean yield across sites and a regression coefficient close to 1.0. Geno type NC 7 had pod yield above the mean pod yield but with regression coefficient farther away from 1.0 ( b < 1), indicating its better responsiveness to unfavorable environmental conditions. Genotype NKATESARI however, was more responsive to the favorable e nvironmental conditions ( b > 1) with above average performance. The remaining genotypes all had pod yields lower than the grand mean pod yield, which indicated their inferior performance. For example, genotypes GM 123, GM 57 and ICGV IS 96895 all had regression coefficients close to 1.0. However, they all recorded pod yields lower than the grand mean pod yield. As a result, such genotypes were not considered to be superior performing genotypes. The patterns of genotypic responses to environments for select ed cultivars are shown in F igures 3 3, and 34. These f igures illustrate the different patt erns of

PAGE 52

52 cultivar response to environment that could be identified. In F igure 3 3, genotypes NC7, NKATESARI, and ICGVIS 96814 respond differently to the change in environment This response gives rise to a genotype by env ironment interaction of the crossover type due to a change in the ranking of cultivars under different environments. Genotypes ICGV IS 92101 and GM 123 as seen from F igure 34 give the non crossover type of genotype by environment interaction I n conclusion, this study provided an evaluation of the genotypic and environmental performance of 19 peanut cultivars over a range of environments. Genotypes ICGV (FDRS) 20 F MIX 39, GUSIE BALIN (92 099 ) ICGV IS 92093, ICGV IS 92101 and ICGV IS 96814 are considered cultivars with broad adaptability, as they have above average mean yield across sites and a regression coefficient close to 1.0. Among these four cultivars, ICGV IS 96814 was considered the best because it adequately demonstrated wide or broad adaptation across environments. This is because it recorded the highest pod yield and had a regression coefficient close to unity. Therefore, genotype ICGV IS 96814 is less responsive to changed envir onmental conditions and can be grown over a range of environments in West Africa. However, cultivar NKATESARI could be considered equivalent in some respects because it had pod yield equal to ICGV IS 96814, but with a higher regression slope. Cultivar G122 TX95 was the worst with a mean yield of 824 kg/ha and a stabilit y regression coefficient of 0.72.

PAGE 53

53 Tabl e 31. Names of peanut genotypes used in the screening trials conducted in 2010 and 2011. Entry Identity 1 CHINESE 2 DOUMBALA 3 TS 32 1 4 ICGV IS 92101* 5 PC 79 79 (disease resistant check) 6 GM 515 = 43 09 03 02** 7 ICGV IS 96814* 8 G204TX95** 9 G122TX95** 10 GUSIE BALIN (92099*) 11 ICGV IS 96895* 12 GM 57 = BC3.60 02 07 03** 13 ICGV (FDRS) 20 F MIX 39 (Ghana cross) 14 ICGV IS 92093* 15 GM 123 = BC3.41 10 09 02** 16 NKATESARI (Ghana release) 17 NC 7 (Not same as N. Carolinas NC7) 18 B106TX95** 19 F MIX SINK 24 (Ghana cross) 20 F MIX (Ghana release) CHECKS : CHINESE, TS 32 1, DOUMBALA *ICGV genotypes from ICRISAT, tested over past 10+ years in Ghana. **GM, G, and B genotypes are result of crosses from Mark Burows program at Texas A&M, College Station. Table 32. Climate data during the growing season (JuneOctob er) for the locations where the tria ls were conducted in 2010 and 2011. Wa Nyankpala Farakoba Gampela 2010 Mean temperature ( o C) 27.1 27.3 26.3 28.5 Total rainfall (mm) 696.3 995.2 1017.8 720.5 Rainfall events 59 68 78 57 2011 Mean temperature ( o C) 27.2 27.5 26.8 29.1 Total rainfall (mm) 734.6 964.6 644 629.5 Rainfall events 60 54 76 56

PAGE 54

54 Table 33. Water holding characteristics of the soils for Wa, Nyankpala, Farakoba and Gampela. Location Depth (cm) Available water (mm) Wa 50 62.0 Nyankpala 90 116.1 Farakoba 200 184.0 Gampela 210 184.8 Table 34. Mean plant stand for 19 genotypes over 4 locations in 2010 and 2011. Identity Plant stand/m 2 CHINESE 11.5a TS 32 1 10.6b ICGV (FDRS) 20 F MIX 39 10.5b NKATESARI 10.4bc GM 57 = BC3.60 02 07 03 10.2bcd ICGV IS 96814 10.0bcd GM 123 = BC3.41 10 09 02 9.9bcde F MIX SINK 24 9.7cdef DOUMBALA 9.6defg G204TX95 9.2efgh GM 515 = 43 09 03 02 9.0fghi NC 7 8.9ghij ICGV IS 92093 8.9ghij B106TX95 8.8hij ICGV IS 92101 8.3ijk G122TX95 8.1jk PC 79 79 7.8kl GUSIE BALIN (92099) 7.2lm ICGV IS 96895 6.8m Means followed by the same letter are not significantly different at P < 0.05

PAGE 55

55 Table 35. ICRISAT score, leafspot (percent necrosis) at 60 and 90 days after sowing and necrosis progress (slope) for 20 genotypes over 4 locations in 2010 and 2011. ICRISAT Score Percent necrosis Genotype Day 60 Day 90 Day 60 Day 90 Slope DOUMBALA 3.7a 7.8a 4.1a 10.3a 0.137bc TS 32 1 3.6ab 8.0a 4.0ab 10.6a 0.147ab CHINESE 3.4bc 7.9a 3.7ab 10.5a 0.152a GM 123 3.3cd 7.0b 3.5abc 9.1b 0.121de GM 57 3.2cd 6.8bc 3.4abcd 8.9b 0.123de GM 515 3.1de 6.6c 3.2bcde 8.5b 0.112efgh NKATESARI 2.9ef 5.9e 2.9cdef 7.5dc 0.102hijk G204TX95 2.8fg 6.1d 2.8defg 7.8c 0.108ghij ICGV (FDRS) 20 F MIX 39 2.7fgh 6.6c 2.6efgh 8.5b 0.128cd ICGV IS 96895 2.6ghi 5.7efg 2.5fgh 7.2cde 0.105hijk PC 79 79 2.6ghi 5.2hij 2.5fgh 6.4fg 0.089ml B106TX95 2.6ghi 5.0j 2.5fgh 6.1g 0.081m ICGV IS 92093 2.5hij 5.8ef 2.3fgh 7.3cde 0.110fghi F MIX SINK 24 2.4ijk 5.5fgh 2.2fghi 6.9def 0.102hijk GUSIE BALIN (92099) 2.4ijk 5.4ghi 2.2fghi 6.7fe 0.100ijk NC 7 2.3jkl 5.9e 2.0ghi 7.5cd 0.119def ICGV IS 96814 2.3jkl 5.4ghi 2.0ghi 6.7fe 0.098jkl F MIX 2.2kl 5.1ij 1.9hi 6.3fg 0.098kl ICGV IS 92101 2.2kl 5.8ef 1.9hi 7.3cde 0.118defg G122TX95 2.1l 4.9j 1.7i 6.0g 0.097jkl Means followed by the same letter in a column are not significantly different at P < 0.05

PAGE 56

56 Table 36. Pooled mean performance for pod yield, seed yield, unit seed weight, shelling percentage and Pod Harvest Index for 19 peanut genotypes in 2010 and 2011. Cultivar Pod yield Seed yield Unit seed Shelling Pod Harvest (kg) (kg) (g) Percent (%) Index ICGV IS 96814 1755a 1124ab 0.376cdef 61.9efg 0.405a NKATESARI 1722a 1203a 0.431a 66.2abc 0.400ab ICGV IS 92093 1581b 1036bc 0.393bc 64.2bcde 0.425a F MIXSINK 24 1524b 954cd 0.374def 61.3fg 0.365c NC 7 1496b 944cd 0.407b 62.8defg 0.420a ICGV IS 92101 1359c 870de 0.401b 60.9g 0.345cd GUSIE BALIN (92099) 1323c 833e 0.392bcd 61.3fg 0.360cd ICGV(FDRS)20F MIX 39 1309c 926de 0.334i 64.1bcde 0.355cd PC 79 79 1078d 664fgh 0.353gh 63.6cdef 0.245f ICGV IS 96895 1048de 488i 0.359fg 49.9k 0.240f CHINESE 979def 724f 0.338hi 65.6abc 0.330de DOUMBALA 960defg 708f 0.363efg 61.6efg 0.355cd G204TX95 947defg 658fgh 0.377cde 65.6abc 0.365c B106TX95 935efgh 574hi 0.330i 66.3ab 0.220f GM 57 898fgh 680f 0.396b 67.6a 0.305f TS 32 1 889fgh 644fgh 0339hi 66.4ab 0.370bc GM 123 881fgh 674fg 0.404b 66.3ab 0.340cd G122TX95 824gh 540i 0.328i 67.1a 0.240f GM 515 805h 578ghi 0.397b 65.2abcd 0.355cd Table 37. Contribution of the individual sources of variation in the pooled analysis of variance for pod yield of the peanut genotypes over all environments. Source df F Value Pr > F Year 1 0.61 0.4365 NS Location 3 14.22 <.0001* Cultivar 18 42.79 <.0001* YearCulltivar 18 5.73 <.0001* LocationCulltivar 54 4.71 <.0001* YearLocation 3 17.73 <.0001* Rep(YearLocation) 16 0.14 1.000 NS YearLocationCultivar 54 4.04 <.0001* Error 288 Total 455 Significant at P< 0.05. NS = not significant at P<0.05

PAGE 57

57 Table 38. Mean performance for pod yield (kg/ha) for 19 peanut genotypes at 4 locations in 2010 and 2011. Location Cultivar Wa Gampela Farakoba Nyankpala Mean ICGV IS 96814 1642ab 2368a 1519a 1492a 1755a NKATESARI 1751a 2513a 1230bc 1396a 1722a ICGV IS 92093 1438bc 2226ab 1230bc 1429a 1581b F MIXSINK 24 1359cd 1988bc 1387ab 1363ab 1524b NC 7 1756a 1723cde 933de 1571a 1496b ICGV IS 92101 1131de 1861cd 1384ab 1058dc 1359c GUSIE BALIN (92099) 1329cd 1912bcd 894def 1158bc 1323c ICGV(FDRS) 20F MIX 39 1504bc 1860cd 1022cd 850de 1309c PC 79 79 541gh 1655cde 940de 1175bc 1078d ICGV IS 96895 472h 1673cde 914def 1133c 1048de CHINESE 914ef 1725cde 669efg 608fg 979def DOUMBALA 907ef 1842cd 641fg 450g 960defg G204TX95 800f 1594def 680efg 713ef 947defg B106TX95 490h 1677cde 704efg 871de 935efgh GM 57 = BC3.60 02 07 03 835f 1492ef 746efg 521fg 898fgh TS 32 1 863f 1421ef 601g 671ef 889fgh GM 123 = BC3.41 10 09 02 829f 1456ef 803defg 478g 881fgh G122TX95 397h 1277f 792defg 858de 824gh GM 515 = 43 09 03 02 782fg 1247f 690efg 500fg 805h Mean 1036 1754 938 963 1174 Means followed by the same letter in a column are not significantly different

PAGE 58

58 Figure 3 1 Adaptive responses to environment for pod yield of 19 peanut genotypes Entry 1= CHINESE Entry 2 = DOUMBALA Entry 3 = TS 321 Entry 4 = ICGV IS 92101 Entry 5 = PC 7979 Entry 6 = GM 515 Entry 7 = ICGV IS 96814 Entry 8 = G204TX95 Entry 9 = G122TX95 Entry 10 = GUSIE BALIN (92099) Entry 11 = ICGV IS 96895 Entry 12 = GM 57 Entry 13 = ICGV (FDRS) 20 F MIX 39 Entry 14 = ICGV IS 92093 Entry 15 = GM 123 Entry 16 = NKATESARI Entry 17 = NC 7 Entry 18 = B106TX95 Entry 19 = F MIX SINK 24 0 500 1000 1500 2000 2500 3000 3500 4000 0 500 1000 1500 2000 2500Cultivar pod yield (kg/ha) Environment mean pod yield (kg/ha) Entry 1 Entry 2 Entry 3 Entry 4 Entry 5 Entry 6 Entry 7 Entry 8 Entry 9 Entry 10 Entry 11 Entry 12 Entry 13 Entry 14 Entry 15 Entry 16 Entry 17 Entry 18 Entry 19

PAGE 59

59 Figure 3 2. Relationship between mean pod yield and regression coefficient of 19 peanut genotypes. G 1= CHINESE G 2 = DOUMBALA G 3 = TS 321 G 4 = ICGV IS 92101 G 5 = PC 7979 G 6 = GM 515 G 7 = ICGV IS 96814 G 8 = G204TX95 G 9 = G122TX95 G 10 = GUSIE BALIN (92099) G 11 = ICGV IS 96895 G 12 = GM 57 G 13 = ICGV (FDRS) 20 F MIX 39 G 14 = ICGV IS 92093 G 15 = GM 123 G 16 = NKATESARI G 17 = NC 7 G 18 = B106TX95 G 19 = F MIX SINK 24 400 600 800 1000 1200 1400 1600 1800 2000 2200 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2Mean Pod yield (kg/ha) Regression Coefficient

PAGE 60

60 Figure 3 3 Genotype pod yield (kg/ha) agains t site mean yield for genotype NC7, NKATESARI, and ICGVIS 96814 Figure 3 4 Genotype pod yield (kg/ha) against site mean yield for genotype GM 123 and ICGV IS 92101. 0 500 1000 1500 2000 2500 3000 3500 4000 0 500 1000 1500 2000 2500Cultivar yield (kg/ha) Environment mean pod yield (kg/ha) NC 7 NKATESARI ICGV-IS 96814 0 500 1000 1500 2000 2500 3000 0 500 1000 1500 2000 2500Cultivar yield (kg/ha) Environment mean pod yield (kg/ha) ICGV-IS 92101 GM 123

PAGE 61

61 CHAPTER 4 GROWTH AND YIELD POTENTIAL OF D IFFERENT PEANUT CULIVARS ACROSS MULTIPLE SITES ANALYZED WITH THE CROPGRO PEANUT MODEL Background Greenhouse gases (GHGs) concentrations in the atmosphere continue to rise and this is causing increase in global air temperature and variability in the amount and distribution of rainfall (IPCC, 2007). This is gradually changing the environmental conditions in areas where food crops are currently grown. Further increases in climate variability in future may reduce the productivity of crops in regions such as West Africa. This poses a serious threat to food security of this region. It is therefore important that new cultivars from genetic improvement are identified to cope with these stresses, thus helping to provide food security to the regions and people most likely to be affected by variable climate. Crop simulation models can be used as tools for evaluating the consequences of climate variability on production, as well as evaluating the effect of sowing date, cultivars, and management practices for adapting to climate variability. They can also be used to evaluate the possibilities of genetic improvement in yield, especially if evaluated against past growth and yield data across different cultivars. Prior to successful use of crop models as tools for such evaluations, it is important that the crop models are calibrated and tested for their responses to CO2, drought, and temperature effects on growth and development Crop models include many temperaturedependent processes such as rate of leaf appearance, leaf expansion, internode elongation, root depth progression, leaf photosynthesis, respiration, fruit set, and single seed growth rate. Pilumwong et al. (2007) st udied the effect of temperature on peanut growth at 25/15 C and 35/25C

PAGE 62

62 day/night air temperatures at ambient and elevated CO2. They observed that increase in temperature enhanced root and shoot growth, but decreased pod dry weight. The beneficial effects of increased CO2 on photosynthesis and growth were overwhelmed by the negative effect of high temperature on reproductive growth. Studying the effect of temperature on reproductive processes and yield of peanuts grown at ambient CO2, Prasad et al. (2003) observed a 14 %, 59 % and 90 % decrease in seed yield as temperature increased from 32/22, 36/26, 40/30 and 44/34C daytime maximum/nighttime minimum temperatures respectively. Seed harvest index also decreased from 0.41 to 0.05 as temperature increased f rom 32/22 to 44/34 under both ambient and elevated CO2. Mortley et al. (2004) studied peanuts grown at 20/16C, 24/20C, 28/24C, and 32/28C day/night air temperatures to evaluate effects on growth and yield. Vegetative growth was substantially higher at increasingly warmer temperatures. Pod yield increased with temperature up to 28/24C but declined by 15 % at the hi ghest temperature (32/28C). Seed yield and harvest index were also highest at 28/24C. Boote et al. (2010) tested the CROPGRO Peanut model against data from sunlit, controlledenvironment chambers over a wide range of elevated temperatures at ambient and elevated CO2. The model was found to adequately predict the sharp decline in yield (observed by Prasad et al., 2003) from high yield at optim um temperature of 32/22C to nearly zero yield as temperature increased to 44/34C. The predicted response to doubling of CO2 for biom ass and pod yield was similar to the observed response. Most experiments todate have not shown any beneficial interaction of CO2 with temperature on grain yield (Boote et al 2005; Prasad et al 2002; 2003; 2006).

PAGE 63

63 Soil water deficit and drought are anticipated to increase under many future climate change scenarios in certain regions. It is difficult to adequately assess cl imate variability responses to future rainfall conditions with confidence for several reasons: 1) the general circulation models themselves are highly uncertain as to rainfall, and 2) the effects of rainfall amount and frequency are greatly influenced by t he soil type, climatic conditions, and crop. Dangthaisong et al. (2006) tested the capability of the CROPGRO Peanut model in simulating the responses of two peanut cultivars to three levels of soil moisture regimes under field conditions and the model predicted the relative yield reductions from drought stress of the peanut cultivars quite accurately. Cecilia et al. (2012) also tested the capability of the CROPGRO Peanut model to simulate growth and development of peanut under different moisture regimes in experiments conducted in automated rainout shelters. The CROPGRO Peanut model was able to accurately simulate growth and development of peanut grown under different irrigation treatments when compared to the observed data. Biomass and seed yield incr eases were also similar to reported values In a study to determine the effect of elevated CO2 on peanut, Stanciel et al. (2000) noted that plants grown at 800 ppm CO2 had net photosynthetic rates that were 29 % greater than those of plants grown at 400 ppm CO2. The number of pods, pod weight and seed dry weight per area all increased with atmospheric CO2 e nrichment from 400 to 1200 ppm. The harvest i ndex, for example, was 19 and 31 % greater at 800 and 1200 ppm CO2, respectively, than it was at 400 ppm CO2. On the other hand, Prasad et al. (2003) observed no enhancement of harvest index attributable to elevated CO2. Boote et al. (2010) observed that the CROPGRO Peanut model predicted canopy

PAGE 64

64 assimilation against field measurements with good success for peanut. They found that the leaf level version of the model predicted 34% response to doubled CO2, which was close to the observed 30 % response to doubled CO2. The (CSM)CROPGRO Peanut model is included in the Decision Support System for Agrotechnology Trans fer (DSS AT) (Tsuji et al., 1994; Hoogenboom et al., 1999; Jones et al. 2003; Hoogenboom et al., 2004). The model is physiologically based and its potential to simulate yields of peanut cultivars based on weather conditions and soil characteristics has been demonst rated (Singh et al., 1994a; Boote et al., 1998a, 1998b). Recent studies also suggest that the CSM CROPGRO Peanut model can be used in assisting multi environment evaluation of advanced peanut breeding lines (Banterng et al., 2004, 2006). Prior to model application, one first has to determine these cultivar coefficients where the genotypes used are new breeding lines or local cultivars that have not been used previously with the model. Using a crop model to resolve which genetic coefficients contributed to y ield gain among the cultivars in past experiments, is additionally a good exercise that can lead to designing ideotypes for genetic improvement (Boote et al., 2001, 2003). Once a crop model is tested and evaluated for a given site, it can be used with long term historical weather data to simulate crop performance under varying environmental and soil conditions. Simulations with genetic hypotheses using selected future climate scenarios can lead to identification of cultivars or traits that are likely to per form well under such future climate conditions (Singh et al., 2012).

PAGE 65

65 The objectives of this study were to calibrate the CROPGRO peanut model for cultivars screened in field trials and simulate genetic improvement of pod yield among different cultivars unde r different environmental and soil conditions. Materials and Methods Experimental Sites and Design Twenty peanut genotypes (Table 41) were tested at two sites in Ghana and two sites in Burkina Faso in 2010 and 2011. The experiments in Burkina Faso were conducted at the Environmental and Agricultural Research Institute in Gampela, Ouagadougou (120 2551N, 10 2218W) and Farakoba, Bobodilaso (110 560N, 40960W). The Ghana experiments were conducted at the Savanna Agriculture Research Institute site s at Nyankpala (90 42 N, 00 92W) and Wa (100 3 N, 20 50 W). These experiments were part of a larger study conducted to screen the twenty peanut genotypes for yield and to determine the range of adaptability of the peanut cultivars. A randomized complete block design with three replicates was used in all trials Seeds were sown in each plot (4 m 2 m) at a spacing of 0.1 m within rows and 0.5 m between rows, resulting in a planned plant population of 20 plants m2 and four rows. One seed was sown per hole. Gap filling was done where necessary to improve uniform crop establishment. Actual stand count was recorded at harvest The preemergence herbicide Pendimethalin was used in combination with hand weeding to control weeds. There was no fungicide application to allow screening for leaf spot resistance and/or tol erance, and to mimic farmer production practices. Table 42 shows the climate data during the growing season (JuneOctober) of the locations where the trials were conducted in 2010 and 2011. Some physical and chemical properties of the soils used in this s tudy are shown in Appendix B 1. The soil at Wa was a sandy loam and it is 50

PAGE 66

66 cm deep The low moisture characteristics and shallow nature of this soil likely will confer some water stress to peanut grown on this soil. The soil at Nyankpala was a sandy loam and it is 113 cm deep. The experiment at Gampela was conducted on l oamy sand and it is 210 cm deep. In the case of Farakoba, the soil was a s andy loam and it is 200 cm deep The soil at Farakoba is characterized by a subsurface accumulation of clay (> 45 %) which can lead to poor internal drainage. Aluminum content also increases down the profile to over 2 ppm. Both of these characteristics may affect root development of crops. Measurement of Disease Incidence and Yield Disease score was recorded in 2010 and 2011 for all four sites using the ICRISAT scale which ranges from 1 t o 9 (Subrahmanyam et al., 1995). The disease rating depends on visual estimate of necrosis and defoliation A disease rating of 1 means no disease and a rating of 2 means lesions present on lower leaves with no de foliation. Ratings of 3 to 8 are associated with increasing levels of defoliation and necrosis. A rating of 9 implies defoliation of almost all leaves leaving bare stems, with any leaflets present having severe leaf spots For use with the crop model, t he ICRISAT leaf spot disease rating was converted to percent necrosis (eq.41) and to percent defoliation (eq.42), using modified forms of the equations developed by Maninder Singh et al. (2011) between percent necrosis and the F lorida 110 visual rating scale and between percent defoliation and the Florida 110 visual rating scale. The modified equations are: ( % ) = 1 36 10 9 1 45 (4 1) ( % ) = 12 5 12 5 (4 2)

PAGE 67

67 where IS represents the ICRISAT visual score. The ICRISAT disease scores and derived percent necrosis and percent defoliation for all dates for all cultivars at all sites and years is available in A ppendices C 1 through C 8. At the end of the season, pods from the two bordered inner rows (4 m2 land area) were handharvested and pods removed manual ly for Wa and Nyankpala sites. At Wa, a sub sample of six plants was taken from the large plot yield sample. The plants were depodded and the fresh weight of pods, leaf and stem was recorded (excluding taproot). The subsamples were ovendried and the dry weights of pods, leaves and stems were obtained. The dry matter concentrations and ratios of the subsamples were then used to calculate pod yield, stem and leaf weight and total biomass for each plot. Pod yield was computed from large plot harvest on a dry weight basis, after adding the subsampled pods back in. For Gampela and Farakoba sites, pods were handharvested from all rows in 2010 from an area of 9.375 m2. In 20 11, pods were handharvested from the inner three rows from a 5.625 m2 land ar ea Pods were removed manually. Measurement of Crop Growth Crop management data of planting date, plant population and row spacing was recorded. Inseason phenology data was obtained nondestructively at Wa in the 2010 season only for the developmental stages of first flower (R1), first peg (R2), first pod (R3) and fully expanded pod (R4). Each stage was considered to have occurred if at least 50 % of the plants in a plot had reac hed that stage. The R3 and R4 stages were determined by careful inspection below ground regularly in the two border rows. At Wa and Nyankpala, inseason data was collected in 2010 and 2011 at 60, 80, 100 days and at harvest from single plant samples (1 plant per replicate because of limited sample area) and analyzed for total biomass, pod mass, Pod Harvest Index, and percent of total

PAGE 68

68 dry matter in leaf. Time series pod yield on a landarea basis was estimated using plant population at harvest. The CROPGRO P eanut Model and Model Inputs The CROPGRO peanut model is a mechanistic and process oriented crop growth simulation model (Boote et al ., 1998a, 1998b; Maninder Singh et al., 2011). It includes crop carbon balance, crop and soil N balance, and soil water balance. The model computes canopy photosynthesis at hourly time steps using leaf level photosynthesis parameters. The model dynamically responds to daily weather inputs (temperature, radiation, rainfall, wind speed and relative humidity), soil water, cult ura l practices and cultivar choice. This model has coupling points and procedures for entering pest damage to simulate growth and yield reductions associated with foliar diseases such as late leaf spot (Batchelor et al., 1993; Boote et al., 1993; Teng et al., 1998). Crop development includes the rates of vegetative and reproductive development (expressed as physiological days as a function of temperature, photoperiod, water, and N deficit) that governs dry matter partitioning to plant organs over time. Crop N balance includes daily soil N uptake, N2 fixation, N mobilization from vegetative tissues, and N loss from abscised parts. Soil water balance includes infiltration of irrigation and rainfall, runoff, drainage, root uptake, soil evaporation, and plant tr anspiration. Crop C balance includes daily photosynthesis, growth and maintenance respiration, conversion and condensation of C to crop tissues, and C lo sses to abscised parts. The data required for the model include crop management data, soil data, weathe r data and the cultivar coefficients. All the data collected on weather, soil, crop management, disease (percent necrosis and percent defoliation from Appendices C1C8) crop growth and yield were entered in the standard DSSAT file

PAGE 69

69 formats (*.PNX, *.PNA, *.PNT, *.WTH, *.PST, and SOIL.SOL). Late leaf spot (LLS) disease first occurs as necrotic lesions on peanut leaflets and subsequently induces leaflet abscission. This defoliation commonly lowers LAI values below the optimum value of 3.0 determined by Dunc an et al. (1978) which reduces light interception and can cause significant loss of canopy carbon assimilation and yield (Bourgeois and Boote, 1992; Boote et al., 1983a). Calibration of Cultivar Coefficients The CROPGRO peanut model requires genetic coeffi cients that describe durations of phases of the crop life cycle, vegetative growth traits, and reproductive traits unique to a given cultivar (Boote et al., 1998). Genetic coefficients were determined by iteration of model simulation against data in 2010 a nd 2011 over two sites, following the procedure of Boote (1999). The calibr ation procedure was stepwise, and this made it possible to successively follow the improvement in the calibration of cultivar coefficients from one step to the next using the root m ean square error (RMSE) and d statistic The steps used for the optimization of cultivar coefficients are shown in Table 44. The first step was to select from the genotype file, the standard short season cultivar CHINESE which was calibrated by Naab et al (2004) for the Guinean Savanna zone of Ghana. This cultivar was then used as the baseline starting point for calibration of all cultivars. Then, the coefficients for duration to flowering (EM FL), beginning pod (FL SH), beginning seed (FLSD), and maturi ty (SD PM) were adjusted to predict the observed life cycle for each cultivar for year 2010 in Wa, Ghana (data in Table 43 ). The next step involved simulating the observed pod yield averaged over all cultivars and two seasons per site, and adjusting the s oil fertility factor (SLPF) in the soil file such that the simulated pod yield was near the observed pod yield (n=38) for each

PAGE 70

70 given site. Data for leaf spot disease (as percent necrosis and percent defoliation ) for the two years for each cultivar were ent ered into the time series files per site (Appendix C ) With disease effects turned on, t he SLPF (Appendix D 1) was again adjusted per site, such that mean pod yield over all 19 cultivars was close to observed pod yield averaged over all cultivars The SLPF controls the rate of dry matter production, and accounts for site specific soil nutrient effects other than nitrogen. Toxic effects of aluminium or nematodes would also reduce SLPF. Optimization Procedure The following steps (Table 44) were aimed to solv e productivity and partitioning traits affecting pod yield where life cycle traits were alr eady set as above from Table 43 data T he model was simulated for each cultivar, one at a time, using a staged maximum likelihood optimization method, using the tim e series growth analysis data from Wa and Nyankpala from 2010 and 2011. In the first stage, thre e cultivar coefficients : maximum leaf photosynthesis rate (LFMAX) maximum fraction of daily growth partitioned to seed plus shell (XFRT), and duration of adding pods (PODUR) were estimated, by optimization to time series data on biomass, pod mass, and pod harvest index. The maximum likelihood optimization method used is a Bayesian approach, and a Bayesian approach to parameter estimation treats parameters as ran dom variables, each with its own distribution. The objective of this approach is to estimate the mean of this distribution based on prior knowledge about the parameters in question, a set of observed data, and a likelihood function. Prior knowledge comes f rom previous studies and scientific literature and provides information about the possible distribution of each parameter being estimated, called the prior distribution. In the first step, uniform distributions of LFMAX, XFRT, and PODUR were used, with no limitations

PAGE 71

71 on range of XFRT or LFMAX, although the lower range for PODUR was constrained to 5.0 photothermal days. The maximum likelihood approach estimates cultivar coe fficients by selecting the set of values of the parameters (cultivar coefficients) that maximize the likelihood function, that is, the set of cultivar coefficients that are most likely to have produced the observed data. After the optimization, simulations were evaluated using the statistical packages in GBUILD, to evaluate the root mean square error (RMSE) and d statistic, averaged over the four seasons (two sites by two years). The trait LFMAX helps reproduce the correct slopes of biomass and pod mass, wh ile XFRT sets the partitioning between pod and biomass as reflected in timeseries thus affecting slope and final value of pod harvest index. The onset of pod harvest index is affected by the cultivar trait PODUR as well as XFRT The second staged optimization involved solving for maximum potential weight per seed (WTPSD), the maximum ratio of (seed/seed +shell) (THRSH), seed filling duration for individual pod cohort (SFDUR) and PODUR based on observed average weight per grain and shelling percentage, agai n with data from the two Ghana sites over two years The third staged optimization was to solve again for LFMAX, PODUR and XFRT keeping THRSH, WTPSD and SFDUR fixed. The third staged process was iterative and repeated for some cultivars, because unsatisfac torily low values for PODUR or SFDUR from stage two were subjectively screened out for some cultivars (where PODUR and SFDUR were falling below the lower limit of reported values for previously modeled cultivars in the peanut model). In these cases, the values were fix e d at more reasonable values, and the optimization process was repeated with only LFMAX and XFRT (for

PAGE 72

72 exam ple if PODUR was fix ed). Another requirement after the three staged optimizations was a good prediction of mean biomass, pod mass and pod harvest index, as well as retaining a good RMSE and dstatistic for these three predicted variables To determine the accuracy of the calibration procedure for estimating the cultivar coefficients, the mean simulated values of pod mass, biomass and pod harvest index were compared to the corresponding observed values, and to the index of agreement ( d ) (Willmott et al., 1985) and the root mean square error (RMSE). Mean values of d near one and low values of RMSE over two locations and two years were used as targets for estimating the cultivar coefficients. The d value was calculated using the following equation: = 1 ( ) 0 1 (4 3) Where n = number of observations, Pi = predicted value for the i th measurement, Oi = observed value for the i th measurement, = the overall mean of observed values, = and = The RMSE was computed using the following equation = ( ) (4 4) Rescaling LFMAX When simulations of final yield were made for the sites, t he values of estimated cultivar trait LF MAX at this point were too high compared to the standard Spanish or Runner peanut cultivars in the genotype file. The suspected reason is that solving for LFMAX (and other trai ts) in the above optimization, was based only on 1plant growth analysis samples, which are likely to be inflated in magnitude,

PAGE 73

73 compared to final pod yield samples (and the prior SLPF values per site had been calibrated on final pod yields at the sites ). A s a result, LFMAX was adjusted (re scaled) to bring it closer to the desired range among standard Spanish or Runner peanut cultivars in the genotype file. T o re scale LFMAX, the LFMAX of the standard short season cultivar in the genotype file (1.28 mg CO2 m2 s1) was divided by the mean LFMAX of all cultivars (1.41 mg CO2 m2 s1) to give a factor of 0.9078. The LFMAX of each cultivar was then multiplied by this factor to give new LFMAX values. Then the SLPF was recalibrated to give the mean pod yield over all cultivars per site for Wa and Nyankpala. Slight changes were made to PODUR and XFRT where necessary to refit pod harvest index (PHI) to obtain a good fit for the time series data of each cultivar. The reason for this is that PHI is a ratio and was c learly more reliable observed data than were biomass or pod mass, and PHI is nicely calibrated with XFRT and PODUR The final step involved working with all four sites and solving for remaining cultivar effects on yield that the timese ries data from Wa an d Nyankpala did not reveal in that optimization. Before doing so, the SLPF was adjusted per site to obtain the mean pod yield over all cultivars per site, for all four sites. Up to this point, the two Burkina Faso sites had not been used in model calibrati on, and they provided a good test of how well model predictions worked in an independent validation, (a point that will be made later in the results), given the optimization to life cycle from Wa, and optimization of genetic coefficients to timeseries dat a from two years at Wa and Nyankpala. This last step was done one cultivar at a time, where the mean simulated values of pod mass, biomass and pod harvest index of e ach individual cultivar were compared to the

PAGE 74

74 corresponding observed final values at harvest T he cultivar coefficients LFMAX, XFRT, PODUR, and SFDUR were adjusted to mimic the observed yield ranking of each cultivar over all four sites over two years. Most of the optimization at this stage focused on LFMAX changes, with constraints placed on XFR T, PODUR, SFDUR, etc, because these latter traits had been well solved from pod harvest index ratio of the timeserie s. Results and Discussion Reproductive Growth Stages Table 43 shows the days from sowing to specific reproductive growth stages for 19 peanut cultivars at Wa, Ghana in 2010. Beginning bloom (R1) ranged from 27 days after sowing (DAS) for DOUMBALA to 34 DAS for PC 7979. The farmer check cultivars CHINESE, DOUMBALA a nd TS 321 flowered earlier than all the remaining cultivars. Beginning peg (R2) also occurred earlier in the farmer check cultivars and ranged from 33 DAS for CHINESE to 40 DAS for G204TX 95 and F MIX. There was over 2 weeks variation in beginning pod (R3) with the farmer check cultivars starting earlier. Beginning pod occurred at 44 DAS for TS 321 and was as late as 60 DAS for B106TX95. The farmer check cultivars reached full sized pod (R4) first with values ranging from 49 DAS for DOUMBALA to 68 DAS for G122TX95. The earliest maturing cultivars were the farmer check cultivars CHINESE, DOUMBALA and TS 321 which matured at 90 DAS. The remaining cultivars all matured between 108 and 118 DAS. Optimization It is not feasible to show simulated timeseries outc omes for all cultivars; however, the outcome of the staged optimization procedure with the timeseries data will be illustrated for several examples. Figure 4 1a and 4 1b show s the observed pod mass and biomass respectively for cultivar NKATESARI over Wa a nd Nyankpala sites

PAGE 75

75 over two years. Because the data was derived from 1plant samples, there was considerable variability in biomass and pod mass and thus more error associated with prediction of biomass and pod mass, than for PHI (Figure 41c). Similar obs ervations were made for other cultivars. The ratio of pod mass to total biomass is pod harvest index, which is a good indicator of onset and intensity of partitioning to pods. Figure 1c shows the pod harvest index for cultivar NKATESARI Variation among cultivars was clearly observed for pod harvest index of cultivars ICGV IS 96814, GUSIE BALIN (92099), and GM 57 as shown in Figure 4 2. The differences in simulated pod mass in Figure 41 are due to differences in disease, SLPF, soil water traits and weather For all cultivars, there was less error associated with prediction of pod harvest index in comparison with pod mass and biomass prediction. Therefore, more emphasis was placed on pod harvest index data during the calibration of cultivar coefficients Var iation in Cultivar Coefficients Estimates of the cultivar coefficients obtained from model calibration for the individual peanut cultivars are given in Table 45 For most of the cultivar coefficients, considerable variations among cultivars were observed. The cultivar trait PODUR varied from 5 photothermal days (PTD) for B106TX95 to 20.7 PTD for ICGV IS 96814. Maximum leaf photosynthesis rate (LFMAX) ranged from 1.04 for GM 515 to 1.60 mg CO2 m2s1 for ICGV IS 96814. Cultivar ICGV IS 96814 had the highest value of 0.72 for XFRT while PC 79 79 had the lowest XFRT value of 0.42. The cultivar trait SFDUR showed less variation (possibly because shorter values were constrained for some cultivars) and ranged from 20.0 to 25.6 PTD. DOUMBALA had the lowest duration to flowering (EM FL) of 19.0 PTD and PC 7979 had the highest EM FL of 24.5 PTD. DOUMBALA had the lowest beginning seed (FLSD) value of 20.0 PTD and B106TX95

PAGE 76

76 had the highest FLSD of 33.0 PTD. The lowest duration from first seed to maturity (SD P M) of 35.14 PTD was recorded for CHINESE and GM 123 had the highest SD PM of 53.0 PTD. The farmer check cultivars CHINESE, DOUMBALA and TS 321 had the shortest life cycle coefficients ( sum of EM FL, FL SD, and SD PM). Independent Evaluation of Derived Cul tivar Coefficients Model simulations with t he derived cultivar coefficients were evaluated against pod yield data not used in model calibration for two sites (Gampela and Farakoba) over two years, to provide a good test of how well the model ed genetic coef ficients predicted yield in an independent validation. The derived cultivar coefficients were used to simulate the pod yield of each cultivar for two years in each of these locations. Evaluation of model derived cultivar coefficients was carried out by com paring the simulated values of pod yield of each cultivar to the corresponding observed values. Figure 4 3a and 4 3b show a plot of simulated pod yield against observed pod yield for two years at Gampela and Farakoba, respectively. Pod yield at Gampela in 2011 was generally lower than in 2010 due to water limitation in 2011 For Farakoba, pod yield in 2010 was generally lower due to excess water stress in 2010. The results from Figure 4 3 indicate that transfer of the derived cultivar coefficients to these two independent sites succeeded in predict ing cultivar effects (and year effects) on pod yield in these sites with an appreciable level of accuracy where disease effect for those locations were also input Figure 44 shows the simulated versus observed po d yield for Nyankpala and W a using data from Wa and Nyankpala. The ability of the model to reasonably predict pod yield f or Gampela is indicated by the d statistic (0.8 2 ) and RMSE ( 510. 1 kg ha1). A reasonably good agreement between simulated and observed

PAGE 77

77 pod yield was also obtained for Farakoba with a d statistic value of 0.78 and a RMSE of 443.0 kg ha1. Relationship of Yield to Cultivar Coefficients The CROPGRO Peanut model has cultivar traits that can create yield differences, even after solving for the correct maturity date, and these traits are all not equal in importance (Boote et al., 2003). The maximum fraction of daily growth partitioned to seed plus shell (XFRUIT) is a trait that affects partitioning to pods and seeds. As a result, a higher XFRT m eans more growth is channeled to seeds and pods, and this tends to lead to higher yield. Maximum leaf photosynthesis rate (LFMAX) also af fects yield potential such that higher values of this trait are associated with higher rate of photosynthesis /biomass a ccumulation The refore, higher LFMAX is generally associated with higher yield. A shorter duration of pod addition ( PODUR) means pods are added more quickly and pod addition is completed sooner. Seed formation is therefore likely to start sooner which allows more time for grain filling to occur. This generally leads to higher yield potential. The trait SFDUR is the duration of growth of individual seeds and an increase in SFDUR gives a slower rate of growth on a per seed bas is. This slower growth per seed a llows a longer period for seed growth as well as allowing more pods /seeds to be carried which tends to lead to higher yield. Cultivars with relatively longer life cycle tend to flower later than shorter seas on cultivars, and this increases the leaf area index (LAI) of longer season cultivars before the onset of reproductive growth. A higher LAI ensures more radiation capture and longer life cycle allows leaves to stay on for a longer time. This allows photosynthesis to be prolonged, leading to higher yield. The life cycle range of calibrated cultivars (77.7100.9 PTD) is in range of other similar cultivars in DSSAT (87.5106.5 PTD). However, the range of

PAGE 78

78 XFRT of calibrated cultivars (0.420.72) is relatively lower than similar cultivars in DSSAT (0.76 0.84). The LFMAX of calibrated cultivars (1.041.60 mg CO2 m2s1) covers a larger range, being lower at one end and higher at the other end compared to the range of similar cultivars in DSSAT (1.201.28 mg CO2 m2s1). The SFDUR of similar cul tivars in DSSAT (2 9 PTD) is long er than that for the calibrated cultivars (20.025.6 mg CO2 m2s1). The value of PODUR of calibrated cultivars (5.020.7) has a larger range, being lower at one end and high at the other end in relation to PODUR of similar short to moderate cycle cultivars in DSSAT (14 16 PTD ). The relationship between pod yield (over two years over four sites) and the derived cultivar traits XFRT and LFMAX of 19 peanut cultivars is shown in Figure 45 As seen in Figure 45 a, higher yielding cultivars were associated with higher XFRT (partitioning to pods) Figure 45 b shows that higher yield potential was associated with higher LFMAX (leaf photosynthesis rate). Figure 4 6 shows the relationship between pod yield (over two years over four sites) and the culti var trait SFDUR and total life cycle. H igher yielding cultivars required a relatively longer life cycle ( Figure 4 6 a ). However, there were cultiv ars with longer life cycle, with lower yield. The latter cultivars had lower partitioning to pods (XFRT). Higher yielding cultivars had relatively longer SFDUR (Figure 4 6 b ); h owever, there were cultivars with longer SFDUR and lower yield where the cultivars also had lower partitioning to pods. Evaluation of Calibration Procedure The estimation of cultivar coe fficients was a systematic, stepwise procedure which involved selection of baseline starting point cultivar, calibration of life cycle and other cultivar coefficients, and finally calibration of cultivars over all sites to mimic yield potential (as outlined in Table 44) Table 4 6 shows the observed and simulated pod

PAGE 79

79 yield and the corresponding RMSE and d statistic per site using the baseline starting point cultivar, CHINESE. Starting with the baseline CHINESE cultivar resulted in a low d statistic per sit e with values ranging between 0.33 for Wa to 0.56 for Farakoba. Nyankpala and Gampela both had d statistic of 0.52. The lowest RMSE of 346.5 kg ha1 was re corded for Nyankpala and Gampela had the highest RMSE of 632.9 kg ha1. Tables 47, 4 8, and 4 9 show the reductions in RMSE and improvements in d statistic during the successive calibration steps of 19 cultivars ov er two years per site. Table 47 shows the observed and simulated pod yield and measures of agreement (RMSE and d statistic averaged over 19 c ultivars per site) after calibration of correct life cycles for cultivars The d stat istic ranged from 0.51 for Wa to 0.77 for Gampela. Nyankpala recorded the lowest RMSE of 336.6 kg ha1 and Gam pela had the highest RMSE of 485.8 kg ha1. Table 48 shows t he observed and simulated pod yield and RMSE and d statistic (averaged over 19 cultivars per site) after cal ibration of correct life cycle and entering disease data for cultivars. The lowest d statistic of 0.58 was observed for W a with Gampel a recording the highest d statistic of 0.79 The RMSE ranged from 231.9 kg ha1 for Wa to 432.8 kg ha1 for Gampel a. The observed and simulated pod yield as well as RMSE and d statistic of the final step (calibration of XFRT, PODUR, LFMAX, SFDUR, WTPSD and THRSH for the 19 cultivars) in the calibration pro cedure are shown in Table 49 Gampela had the highest d statistic of 0.92 and Farakoba had the lowest d statistic of 0.85. Nyankpala recorded the lowest RMSE of 218.3 kg ha1 and the highest RMSE of 378.7 kg ha1 was observed for Farakoba. Apart from Wa where the d statistic remained the same (0.65) in Table 47 and Table 48 there was always an increase in d statistic and a reduction in RMSE over 19 cultivars per site from one step

PAGE 80

80 to th e next (Tables 46, 4 7, 4 8 and 4 9 successively). Therefore, the success of the calibration procedure over 19 cultivars per site was shown by the reduction in RMSE and increase in d stati stics from one step to the next In conclusion, the optimization procedure estimate d cultivar coefficients that provided simulated pod yield that agreed quite well with the observed pod yield. There was also considerable genetic variation in cultivar traits among peanut cultivars. Model evaluation with independent pod yield data not used in model calibration (Gampela and Farakoba over two years ), provided a good test of how well model predictions worked in an independent validation. The derived cultivar coefficients over four sites and two years allowed the CROPGRO Peanut model to mimic y ield ranking quite well and suggests value in using the model to hypothesize genetic improvement (combinations of traits for best yield and stability) for target environments where long term weather data is available

PAGE 81

81 Table 41. List of p eanut culti vars screened in 2010 and 2011 at four sites: Wa, Ghana, Nyankpala, G h ana, F arakoba, Burkina Faso, and Gampela, Burkina Faso, and used to analyze growth and yield potential with the CROPGRO Peanut model Entry Identity 1 CHINESE 2 DOUMBALA 3 TS 32 1 4 ICGV IS 92101* 5 PC 79 79 (disease resistant check) 6 GM 515 = 43 09 03 02** 7 ICGV IS 96814* 8 G204TX95** 9 G122TX95** 10 GUSIE BALIN (92099*) 11 ICGV IS 96895* 12 GM 57 = BC3.60 02 07 03** 13 ICGV (FDRS) 20 F MIX 39 (Ghana cross) 14 ICGV IS 92093* 15 GM 123 = BC3.41 10 09 02** 16 NKATESARI (Ghana release) 17 NC 7 (Not same as N. Carolinas NC7) 18 B106TX95** 19 F MIX SINK 24 (Ghana cross) 20 F MIX (Ghana release) CHECKS : CHINESE, TS 32 1, DOUMBALA *ICGV genotypes from ICRISAT, tested over past 10+ years in Ghana. **GM, G, and B genotypes are result of crosses from Mark Burows program at Texas A&M, College Station.

PAGE 82

82 Table 42. Climate data during the growing season (JuneO ctober) for Wa, Nyankpala, Farakoba and Gampela in 2010 and 2011. Wa Nyankpala Farakoba Gampela 2010 Mean temperature ( o C) 27.1 27.3 26.3 28.5 Total rainfall (mm) 696.3 995.2 1017.8 720.5 Rainfall events 59 68 78 57 2011 Mean temperature ( o C) 27.2 27.5 26.8 29.1 Total rainfall (mm) 734.6 964.6 644 629.5 Rainfall events 60 54 76 56 Table 43 Days from sowing to specific reproductive growth stages for peanut cultivars at Wa, Ghana in 2010. Cultivar R1 R2 R3 R4 Maturity ICGV IS 96814 30 37 46 55 108 NKATESARI 29 35 47 58 108 ICGV IS 92093 30 34 47 57 108 F MIXSINK 24 30 37 47 56 108 NC 7 30 35 47 58 110 ICGV IS 92101 31 37 48 59 110 GUSIE BALIN (92099) 33 38 49 57 110 ICGV(FDRS) 20F MIX 39 30 34 51 57 108 PC 79 79 34 39 52 66 110 ICGV IS 96895 31 36 52 67 115 CHINESE 28 33 44 51 90 DOUMBALA 27 34 41 49 90 G204TX95 33 40 54 61 110 B106TX95 33 39 60 69 108 GM 57 = BC3.60 02 07 03 32 38 52 61 108 TS 32 1 28 33 44 51 90 GM 123 = BC3.41 10 09 02 30 35 50 60 115 G122TX95 33 39 59 68 108 GM 515 = 43 09 03 02 30 34 52 61 110 F MIX 33 40 52 64 118 R1= beginning bloom, R2 = beginning peg, R3 = beginning pod, R4 = full sized pod, Maturity = actual harvest but this was determined for Wa in 2010 by hull scrape and approximate achievement of 7080 % darkened hulls.

PAGE 83

83 Table 44. Steps in the calibration of cultivar coefficients with an optimizer using data from Wa and Nyankpala in Ghana in 2010 and 2011. Step Procedure 1 Chinese as baseline starting point for calibration of all cultivars 2 Calibration of life cycle (per cultivar) 3 Adjusting soil fertility factor (SLPF) per site mean (across years) 4 Entering leaf spot disease (as percent necrosis and percent defoliation) per site year 5 Adjusting soil fertility factor (SLPF) per site mean (across years) 6 Solving for LFMAX, XFRT and PODUR against time series data on biomass, pod mass, and pod harvest index per cultivar 7 Solving for WTPSD, THRSH, SFDUR, and PODUR based on observed average weight per seed and shelling percentage per cultivar 8 Solving again for LFMAX, XFRT, and PODUR against time series data on biomass, pod mass, and pod harvest index, keeping THRSH, WTPSD, and SFDUR fixed per cultivar 9 Rescaling LFMAX 10 Adjusting soil fertility factor (SLPF) per site mean (across years) 11 Adjusting PODUR and XFRT where necessary to re fit pod harvest index per cultivar [To this point, all optimization was done with data from Wa and Nyankpala] 12 Working with all four sites and solving for remaining cultivar effects on yield adjusting only LFMAX, XFRT, PODUR, and SFDUR per cultivar

PAGE 84

84 Table 45 Cultivar coefficients of individual peanut cultivars derived from data at Wa and Nyankpala in Ghana in 2010 and 2011 using the CROPGRO Peanut model with an optimizer. Cultivar traits LFMAX XFRT PODUR SFDUR THRSH EM FL FL SD SD PM Cultivar (mg CO 2 m 2 S 1 ) (fraction) (PD) (PD) (%) (PD) (PD) (PD) CHINESE 1.17 0.59 7.5 20.0 71.5 19.9 22.7 35.14 DOUMBALA 1.16 0.58 9.5 22.5 70.0 19.0 20.0 39.50 TS 32 1 1.10 0.59 8.0 24.0 72.0 20.0 20.5 38.20 ICGV IS 92101 1.44 0.59 15.0 22.0 70.0 21.8 25.0 50.00 PC 79 79 1.32 0.42 5.0 24.0 75.8 24.5 28.5 42.80 GM 515 1.04 0.52 7.0 25.6 74.8 21.0 28.0 47.00 ICGV IS 96814 1.60 0.72 20.7 25.0 72.8 21.0 23.4 50.70 G204TX95 1.13 0.54 9.0 23.0 71.6 23.8 25.5 47.40 G122TX95 1.09 0.49 7.0 22.0 74.5 23.8 28.8 38.40 GUSIE BALIN (92099) 1.36 0.63 15.5 24.0 70.2 23.8 23.4 49.50 ICGV IS 96895 1.23 0.49 8.0 25.0 70.5 21.5 30.6 47.40 GM 57 1.15 0.58 12.0 20.0 73.5 22.5 26.8 45.00 ICGV(FDRS) 20F MIX 39 1.42 0.62 11.0 24.0 72.1 21.0 27.0 47.00 ICGV IS 92093 1.55 0.71 18.0 24.0 74.9 21.0 26.0 48.00 GM 123 1.12 0.58 11.0 20.0 74.5 21.0 26.9 53.00 NKATESARI 1.57 0.70 15.0 23.0 73.9 20.0 25.0 50.00 NC 7 1.44 0.69 15.0 23.0 71.4 21.0 24.0 51.50 B106TX95 1.20 0.44 5.0 20.0 76.0 23.8 33.0 37.00 F MIXSINK 24 1.56 0.61 14.4 22.0 71.0 21.0 25.5 49.00 PD, Photothermal days

PAGE 85

85 Table 46 Observed (obs) and simulated (sim) po d yield and measures of agreement for baseline starting point assuming all 19 cultivar s were same as cultivar CHINESE for Wa Nyankpala, Farakoba and Gampela over two years (calibrated for site means only) Obs pod Sim pod RMSE Location (kg ha 1 ) (kg ha 1 ) d statistic (kg ha 1 ) Farakoba 938 868 0.56 506.8 Gampela 17 53 1791 0.52 632.9 Nyankpala 961 986 0.52 346.5 Wa 1036 1081 0.33 473.8 Table 47 Observed (obs) and simulated (sim) pod yield, and measures of agreement after calibration of correct life cycle averaged over 19 cul tivars over two years for Wa, Nyankpala, Farakoba and Gampela. Obs pod Sim pod RMSE Location (kg ha 1 ) (kg ha 1 ) d statistic (kg ha 1 ) Farakoba 938 940 0.67 461.9 Gampela 1753 17 68 0.77 485 .8 Nyankpala 961 970 0.65 336.6 Wa 1036 10 84 0.51 410.3 Table 48 Observed (obs) and simulated (sim) pod yield, and measures of agreement after calibration of l ife cycle and entering disease data averaged over 19 cultivars for Wa, Nyankpala, Farakoba and Gampela over two year s. Obs pod Sim pod RMSE Location (kg ha 1 ) (kg ha 1 ) d statistic (kg ha 1 ) Farakoba 938 944 0.7 6 41 4.0 Gampela 1753 1780 0. 79 432.8 Nyankpala 961 974 0. 65 296.0 Wa 1036 1049 0. 58 231.9

PAGE 86

86 Table 49 Observed (obs) and simulated (sim) po d yiel d and measures of agreement afte r calibration of life cycle, entering disease data, then calibrating XFRT, PODUR, LFMAX, SFDUR, WTPSD and THRSH for all four s ites a veraged over 19 cultivars for Wa, Nyankpala, Far akoba and Gampela over two years. Obs pod Sim pod RMSE Location (kg ha 1 ) (kg ha 1 ) d statistic (kg ha 1 ) Farakoba 938 958 0.85 378.7 Gampela 1753 1780 0.92 357.0 Nyankpala 961 995 0.91 218.3 Wa 1036 1052 0.91 240.0

PAGE 87

87 Figure 41. Simulated (line s) and observed (point s) pod yield (a) biomass (b) and Pod Harvest Index (c) for NKATESARI over 2 sites over 2 years. 0 1000 2000 3000 4000 5000 6000 0 50 100 150Biomass (kg/ha) Days after sowing TL2011 TL2011 TL2010 TL2010 WA2011 WA2011 WA2010 WA2010 0 500 1000 1500 2000 2500 0 50 100 150Pod yield (kg/ha) Days after sowing TL2011 TL2011 TL2010 TL2010 WA2011 WA2011 WA2010 WA2010 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0 50 100 150Pod Harvest Index Days after sowing WA-2010 WA-2010 Wa-2011 Wa-2011 TL-2011 TL-2011 TL-2010 TL-2010

PAGE 88

88 Figure 42. Simulated (lines) and observed (point s) Pod Harvest Index of ICGV IS 96814, GUSIE BALIN (92099), GM 57, and NKATESARI for Wa, 2011. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 40 60 80 100 120Pod Harvest Index Days after sowing ICGV-IS 96814 ICGV-IS 96814 GISIE BALIN (92099) GUSIE BALIN (92099) GM 57 GM 57 NKATESARI NKATESARI

PAGE 89

89 Figure 4 3. Simulated versus observed pod yield for Gampela (a) and Farakoba (b), using cultivar coefficients derived from data collected at Wa and Nyankpala, Ghana. Figure 44. Simulated versus observed pod yield for Nyankpala (a) and Wa (b) using cultivar coefficients derived from data collected at Nyankpala and Wa, Ghana. 0 1000 2000 3000 4000 0 1000 2000 3000 4000Simulated pod yield (kg/ha) Observed pod yield (kg/ha) 2011 2010 0 1000 2000 3000 4000 0 1000 2000 3000 4000Simulated pod yield (kg/ha) Observed pod yield (kg/ha) 2011 2010 0 1000 2000 3000 4000 0 1000 2000 3000 4000Simulated pod yield (kg/ha) Observed pod yield (kg/ha) 2010 2011 0 1000 2000 3000 4000 0 1000 2000 3000 4000Simulated pod yield (kg/ha) Observed pod yield (kg/ha) 2010 2011 (b)Farakoba (a)Nyankpala

PAGE 90

90 Figure 45 Relationship between pod yield and XFRT ( maximum fraction of daily growth partitioned to seed plus shell ) and LFMAX (maximum leaf photosynthesis rate) of 19 cultivars. Yield is averaged over 2 years over 4 sites. Figure 4 6 Relationship between pod yield and SFDUR ( duration of growth of individual seeds ) and life cycle of 19 cultivars. Yield is averaged over 2 years over 4 sites. 500 700 900 1100 1300 1500 1700 1900 0.3 0.5 0.7 0.9Pod yield (kg/ha) XFRT 500 700 900 1100 1300 1500 1700 1900 0.9 1.4 1.9Pod yield (kg/ha) LFMAX (mg CO 2 m2 S1) 500 700 900 1100 1300 1500 1700 1900 70 90 110Pod yield (kg/ha) Life cycle (Photothermal days) 500 700 900 1100 1300 1500 1700 1900 18 23 28Pod yield (kg/ha) SFDUR (Photothermal days)

PAGE 91

91 CHAPTER 5 YIELD RESPONSES TO HYPOTHETICAL VARIATION IN GENETIC TRAITS OF DIFFERENT PEANUT CULTIVARS WITH THE CSM CROPGRO PEANUT MODEL Background Breeding for highyielding crop cultivars for specific environments is a major challenge t o feed growing world populations. Through extensive selection, based largely on empirical field observations, breeders have been successful in creating highyielding cultivars. Enhancing crop yields through sciencebased breeding has occurred over several decades, a task that has been further accelerated in recent years through molecular technologies (using DNA based markers as well as transgenes). Plant breeders, for many years, have attempted, and succeeded in many cases, to model plant ideotypes that res ult in higher yields (Donald, 1968) starting in the 1960s with the shorter semi dwarf rice cultivars that did not lodge under increased N supply (Chandler, 1969), continuing to the New Plant Type for rice of the past decade (Peng et al., 2008). There are s till important new demands for crop improvement programs to produce highyielding crop cultivars to ensure food security in regions with low yield potential (Dwivedi et al., 2008b). Breeding programs must develop cropspecific and regionspecific strategie s to increase world food supply, in view of the continued increase in world population. Breeders must identify new sources of variation to increase yield or must find germplasm with traits that can be used to develop highyielding crop cultivars With the recent advancements in dynamic crop growth simulation, crop models have excellent potential for evaluating genetic improvement, for analyzing past genetic improvement from experimental data, and for proposing plant ideotypes for target environments. Crop simulation models have potential for creating virtual crop cultivars and for further assisting the breeders selection criteria, and for genetic enhancement of

PAGE 92

92 important traits that contribute to yield improvement in different target environments (Hammer an d Jordan, 2007). Large multi national plant breeding firms are beginning to use crop simulation models for this purpose (Loffler et al., 2005). The target environments necessarily must be described in terms of weather (over multiple seasons), water availability, soil physical and chemical constraints, desired crop life cycle, and management. Crop models are very useful for these purposes, but they often contain many empirical elements in which parameters may have little biological meaning, and they may be l imited in physical, chemical, and physiological components needed to represent resource availability and the crop responses to resource limitations (e.g., water, nitrogen, phosphorus, salinity) and biotic stresses. Since these models incorporate cultivar specific parameters that represent genetic traits of cultivars, their effects on crop performance can be evaluated in target environments. These traits express the genetic potential of each cultivar to determine their yield in a given environment For peanut, the CSM CROPGRO Peanut model has been developed and its potential to simulate yields of peanut cultivars based on weather conditions and soil characteristics has been demonstrated (Singh et al., 1994; Boote et al., 1998). The CSM CROPGRO Peanut model has the ability to mimic yield differences due to cultivar variation in photosynthesis, life cycle allocation to different phases, vegetative, partitioning and reproductive attributes. The CSM CROPGRO Peanut model has been used widely to evaluate genetic improvement of peanut (Boote and Jones, 1986; Suriharn et al., 2008; Anothai et al., 2009; Putto et al., 2009, Singh et al., 2012). The peanut model was tested and found to work well for groundnut crops in India (Singh et

PAGE 93

93 al., 1994a, 1994b). Crop models can be used to evaluate the quantitative response over the entire possible range of single genetic trait, such as rate of root depth progression (Boote et al., 2003) For example, Yin et al. (1999) evaluated rice yield response to variation in specific leaf area (SLA). Crop models can also be used to evaluate response to specific multiple combinations of traits. In such evaluations, it is essential that the feasible genetic range for each given genotypespecific trait be considered relative to reported literatur e. Therefore, model s can be used to assess the value of genetic traits on crop performance under differing environments and management (Boote and Jones, 1986; Elwell et al., 1987; Boote and Tollenaar, 1994; Boote et al, 2003; Singh et al., 2012). An assess ment of the effect of different cultivar characteristics on yield could help identify traits that improve crop yield in relation to the environment. The objective of this study was to evaluate the yield response of different virtual peanut cultivars to ide ntify combination of traits that increase yield under different weather and soil conditions using the CSM CROPGRO Peanut model. Materials and M ethods The data required for the model include crop management data, soil data, weather data and the cultivar coefficients. All the data on weather, soil, disease, crop management, crop growth and yield were entered in the standard DS SAT file formats (*.PNX, *.PNA, *.PNT, and SOIL.SOL). The soils were the same soils used in Chapter 4 and so had the same soil fertility factor (SLPF) as in Chapter 4. Observed w eather data including rainfall, daily maximum and minimum temperature and solar radiation for 30 years were used for Gampela ( 120 2551N, 10 2218W) Farakoba (110 560N, 40960W), Nyankpala (90 42 N, 00 92W) and Wa (100 3 N, 20 50 W) sites The weather data was used to create weather files using the Weatherman utility in DSSAT Three

PAGE 94

94 peanut cultivars with life cycles of 90, 103, and 116 days were created by assuming that shorter life cycle cult ivars tend to have relatively shorter duration from emergence to flowering, start seed formation earlier, and have shorter duration from start of seed filling to maturity Respective phase durations increased proportionately in medium and longer cycle cult ivars. The life cycle traits modified include duration from emergence to start of flowering (EM FL), flowering to beginning pod (FLSH), flowering to beginning seed (FLSD), and beginning seed to physiological maturity (SD PM). The life cycles were obtaine d by adding the EM FL, FL SD and SD PM to give 90, 103, and 116 day life cycles. Table 52 shows the life cycle traits of three standard life cycles developed from repr oductive and growth traits of peanut cultivars from Wa and Nyankpala in Ghana in 2010 and 201. Disease progress curves: Leaf spot disease plays a major role in reducing yield of peanut. To account for the effect of leaf spot necrosis and defoliation on yield, three leaf spot disease progress curves were developed from the data to represent susceptible, moderately susceptible and resistant. In the first step, the mean leafspot score over two years (2010 and 2011) for four sites (Gampela, Farakoba, Nyankpala and Wa) was obtained at days 60 and 90 for 19 peanut cultivars in Table 51. The ICRISAT leaf spot disease rating (Subrahmanyam et al., 1995) was converted to percent necrosis (eq.31) and to percent defoliation (eq.32), both equations being modified for ms of the equations developed by Maninder Singh et al. (2011) between percent necrosi s and the Florida 110 visual rating scale and between percent defoliation and the Florida 110 visual rating scale. The modified equations are:

PAGE 95

95 ( % ) = 1 36 10 9 1 45 (5 1) ( % ) = 12 5 12 5 (5 2) where IS represents the ICRISAT visual score. Then necrosis progress, estimated as the slope of the necrosis scores between 60 and 90 days over 2 years and 4 locations was obtained for eac h of the 19 cultivars (Table 53 ). The slope values were classified into 3 groups to represent susceptible, moderately susceptible and resistant. Cultivars with slope values less or equal to 0.160 were considered resistant to leaf spot disease. Cultivar s with slope values between 0.16 1 and 0.200 were considered cultivars with m oderate resis tance to leaf spot disease while cultiva rs with slope values greater than 0.200 were classified as cultivars susceptible to leaf spot Based on the three disease groupings (using slope values), the 19 cultivars (over 2 years over 4 locations) were each classified as susceptible, moderately susceptible and resistant to leaf spot disease. The next step involved estimating for each location, the mean leaf spot score of the cultivars in each group (resistant, moderately susceptible and resistant) t o give a single leaf spot score for each group (resistant, moderately susceptible and resistant). As a result, the leaf spot disease scores for susceptible, moderately susceptible and resistant differed for all four locations since the scores were based on the actual disease score recorded per location. The three levels of necrosis (susceptible, moderately susceptible and resistant) in Table 55 and three levels of defoliation (susceptible, moderately susceptible and resistant) in Table 56 wer e then assigned to each of the three developed life cycles (three life cycles by three disease levels). Genetic traits: Five growth traits namely maximum leaf photosynthesis rate (L FMAX), maximum fraction of daily growth partitioned to seed plus shell (XFRT),

PAGE 96

96 duration of adding pods (PODUR) seed filling duration for individual pod cohort (SFDUR) and the maximum rati o of seed/(seed +shell) (THRSH) were selected from the genetic coefficients previously calibrated for 19 peanut cultivars (Table 51). The range (m inimum and maximum) and median values of each of the five growth traits were used for each of the three standard life cycle cultivars developed to give a total of 45 virtual cultivars The c ul tivar coefficients of 15 peanut cultivars derived for the s hort cycle cultivars are shown in Table 54. The same cultivar coefficients in Table 54 were created for the medium cycle and long cycle cultivars (not shown) to give a total of 45 cultivars. A n experimental crop management file was created with data on planti ng date, cultivars, soils etc. A row spacing of 0.5 m between rows and 0.1 m within rows was used giving a plant population of 20 plants/m2 at seeding a nd emergence. The experiment was run under rain fed conditions with no fertilization (but simulated N fixation provides needed N) One sowing date per site was used for the study To obtain a measure of yield variability across seasons, the coefficient of variation was estimated for each location over the weather years. Results and D iscussion Yield Responses to Life Cycle The simulated pod yield (at median trait) of short, medium, and long life cycles at susceptible, moderately susceptible and resistant disease for Wa, Nyankpala, Gampela, and Farakoba is sh own in Table 57. Generally, Gampela was the highest yielding environment followed by Farakoba and Wa, with Nyankpala having the lowest simulated pod yield. For Wa, increasing life cycle for susceptible disease cultivars from short to long cycl e increased pod yield from 1068 to 1249 kg ha1 representing a 16.9 % increase in yield. Also, increasing life cycle for disease resistant cultivars increased pod

PAGE 97

97 yield from 1304 to 1505 kg ha1 representing a 15.4 % yield increase. For Nyankpala, when life cycle was increased from short to long for susceptible cultivar, pod yield increased from 840 to 929 kg ha1. When disease resistance was present, increasing life cycle from short to long increased pod yield from 986 to 1139 kg ha1. For Gampela, pod yield increased from 2394 to 2665 kg ha1 under susceptible disease case when life cycle increased from short to long. Increasing life cycle for a disease resistant cultivar increased pod yield from 3040 kg ha1 to 3292 kg ha1 representing a yield increase of 8.3 %. In the case of Farakoba, increasing life cycle from short to long for susceptible disease cultivars increased yield from 1288 to 1477 kg ha1. With disease resistance, increasing life cycle from short to long increased yield from 1757 kg ha1 to 1901 kg ha1 representing a yield increase of 8.2 %. Generally, pod yield increase attributed to increase in life cycle was smaller than anticipated for all four locations. This was hypothesized because of the low ratio of SFDUR to SD PM for the cultivars (which was an outcome of trait calibration). When SFDUR is too small a fraction of SD PM, the simulated yield can be limited because seed filling duration is short and seeds reach the limit of their pod cavity (caused by model structure). The SFDUR as a fraction of SD PM is 0.495 for the 37 standard DSSAT cultivars. The values of SFDUR from optimization for the 19 calibrated cultivars were a smaller fraction of SD PM and affected the 45 virtual cultivars. The SFDUR to SDPM fraction of the 45 virtual cultivars were 0.360, 0.377, and 0.392 for the long, medium and short life cycles respectively. Yield Responses to Disease The pod yield response to disease resistance trait is shown in Table 57. For Wa, increasing disease resistance from susceptible to resistant under shor t life cycle increased pod yield from 1068 kg ha1 to 1304 kg ha1 representing a 22.1 % increase in

PAGE 98

98 yield. Under long life cycle, susceptible cultivar had a yield of 1249 kg ha1 and resistant cultivar had a yield of 1505 kg ha1. At Nyankpala, pod yield increased from 840 kg ha1 for susceptible to 986 kg ha1 for resistant cultivar under short life cycle, while under long life cycle, pod yield increased from 929 for susceptible to 1139 kg ha1 for resistant cultivar. For short life cycle at Gampela, pod yield was 2394 kg ha1 for susceptible and 3040 kg ha1 for the resistant cultivar, giving a 27.0 % increase in pod yield. Under long life cycle, pod yield for susceptible and resistant cultivars were 2665 kg ha1 and 3292 kg ha1 respectively. When life c ycle was short at Farakoba, susceptible yielded 1288 kg ha1 and resistant cultivar yielded 1757 kg ha1. This represents a 36.4 % increase in pod yield. Finally, increasing disease resistance from susceptible to resistant under long cycle at Farakoba increased yield from 1477 kg ha1 to 1901 kg ha1. In all cases, disease resistant cultivars yielded considerably higher than susceptible cultivars and shows the importance of selecting for leafspot resistance. Linking SFDUR to SD PM Considering the fact that SFDUR of the virtual cultivars was shorter (than the DSSAT standards) and potentially limiting to yield, the correlation between SFDUR and SD PM of the standard DSSAT cultivars was estimated. The relationship between SFDUR and SDPM of the standard DSS AT cultivars was strong (r = 0.85, n = 37), and significant (P< 0.05). Similar correlation analysis using SFDUR and SD PM of the 19 calibrated cultivars gave a strong (r = 0.71, n = 19)) and significant (P< 0.05) relationship. A significant positive correl ation between SFDUR and SD PM indicates that both traits increase together in the same direction. This strong and significant relationship between SFDUR and SD PM, perhaps, is an indication that increases in single traits such as life cycle or SD PM are mo st often accompanied by a longer

PAGE 99

99 SFDUR. In addition to SFDUR, other traits like PODUR (pod adding duration) may also have some relationship with SD PM. This requires further inv estigation. After establishing this relationship, a constant fraction of SFDUR to SD PM for the calibrated 19 cultivars (0.377) and for the standard DSSAT cultivars (0.495) were used to estimate SFDUR values for the corresponding SD PM values of short, medium, and long cycle cultivars (rather than the median values previously used). These values were then used to simulate pod yield for the four locations. Simulated pod yield (at median trait) of short, medium, and long life cycles at susceptible, moderately susceptible and resistant disease averaged over Wa, Nyankpala, Gampela, and Fa rakoba with and without SFDUR linked to SDPM is shown in Table 5 8 When SFDUR was not linked to SD PM, pod yield for susceptible cultivar increased from 1398 kg ha1 for short life cycle to 1580 kg ha1 for long life cycle. With disease resistance, pod yi eld increased from 1772 kg ha1 for short cycle to 1959 kg ha1 for long cycle. When SFDUR of the 19 calibrated cultivars was linked to SD PM (using a ratio of 0.377), pod yield for susceptible cultivar increased from 1288 kg ha1 for short life cycle to 1 708 kg ha1 for long life cycle. With disease resistance, pod yield at short and long life cycle was 1698 kg ha1 and 2074 kg ha1 respectively. This represents a 22.1 % increase in yield. This confirms that short SFDUR (median from calibration) was reduci ng the effect of life cycle on pod yield. Linking SFDUR to SD PM using the ratio from the standard DSSAT cultivars for susceptible cultivar increased pod yield 30 % from 1477 kg ha1 for short life cycle to 1920 kg ha1 for long life cycle. With disease re sistance, pod yield at short life cycle was 1871 kg ha1 and pod yield at long life cycle was 2317 kg ha1.

PAGE 100

100 Overall, linking SFDUR to SD PM resulted in higher simulated pod yield. This also led to higher percent increase in pod yield when life cycle was increased from short to long. In addition to SFDUR, other traits such as PODUR (pod adding duration) also have a positive relationship with SD PM, but this was not evaluated. Yield Responses to Genetic Traits other than Life Cycle and Disease T he simulated p od yield of short medium, and long cycle cultivars at susceptible, moderately susceptible, and resistant disease for 5 traits over Wa, Nyankpala, Gampela, and Farakoba.is shown in Table 59. The pod yields are expressed as percent change from standard (median) cultivar. Maximum leaf photosynthesis rate (LFMAX) affects yield potential, because higher values of this trait are associated with higher rate of photosynthesis. Increasing LFMAX from 1.04 to 1.60 mg CO2 m2s1 increased simulated pod yield for all three (short, medium and long ) life cycles over four locations. A shorter duration of pod addition (PODUR) m eans pods are added faster and pod addition is completed sooner. Seed formation is therefore likely to start sooner which allows more time for grain filling to occur. For all three life cycles, increasing the rate of pod addition (decreasing PODUR from 20.7 to 5 photothermal days) resulted in higher simulated pod yield (by 10.0 %, on average). Higher PODUR resulted in decreased simulated pod yield. Th e trait SFDUR is the duration of growth of individual seeds and an increase in SFDUR gives a slower rate of growth on a per seed basis. The slower growth per seed allows mor e pods to be carried Higher simulated pod yield was observed for increased SFDUR f or all three life cycles. The maximum fraction of daily growth partitioned to seed plus shell (XFRUIT) affects partitioning to pods and seeds. As a result, a higher XFRT means more growth is channeled to seeds and pods Increasing XFRT from 0.42 to 0.72 al so increased simulated yield for all three life cycles

PAGE 101

101 over four locations. Shelling percentage is the ratio of seed to seed plus shell (seed/seed + shell). Increasing shelling percentage from 70.0 to 76.0 % decreased pod yield for all three life cycles, presumably because less shell was needed per unit of seed mass. The effect was relatively small (usually less than 4 % yield change). Similar observations were made at moderately susceptible and resistant disease where higher trait values of LFMAX, SFDUR, X FRT and lower trait value of PODUR all resulted in higher yield change. Pod yield changes due to changes in XFRT and LFMAX were largest in general, whereas effects of changes in THRSH were minimal. Effect of Trait Combinations on Pod Yield Table 510 sh ows the simulated pod yield, percent change from prior step and cumulative percent of different peanut cultivars averaged over 30 years of observed weather data and over Nyankpala, Wa, Farakoba and Gam pela. This shows the pod yield that is possible from combi nations of cultivar traits, and it also contrasts this possible yield to what exists now. The farmer check cultivars CHINESE and TS 321 were simulated with the derived susceptible disease and improved cultivars NKATESARI and ICGVIS 96814 were simulated w ith the derived resistant disease. The farmer check cultivars TS 321 and CHINESE had the lowest yield of 1199 and 1254 kg ha1 respectively. For yield based on possible trait combination, the susceptible short cycle cultivar with median rest traits had th e lowest yield of 1398 kg ha1. This yield, however, was higher than that of the farmer check cultivars which also had short life cycle, were susceptible, but with minimum value for other traits. Generally, the resistant long cycle cultivars had higher yield. The long cycle resistant cultivar with maximum trait value for XFRT, SFDUR, LFMAX and minimum trait value for PODUR had the highest yield of 2898 kg ha1. Improved cultivars NKATESARI had a pod yield of

PAGE 102

102 2454 kg ha1 and ICGV IS 96814 had pod yield of 2501 kg ha1. Therefore improved cultivars NKATESARI and ICGV IS 96814 were among the top yielding cultivars, indicating that these traits are all possible and have been achieved to a large extent compared to the farmer check cultivars. In fact, there is ro om for further yield improvement with combinations from existing variation in traits, as yield variation among the hypothetical cultivars is more than twofold. Maximum Possible Yield One virtual peanut cultivar was evaluated to obtain a measure of the max imum yield that is possible for each of four environments Table 511 shows the maximum possible pod yield averaged over 30 years of weather data for Farakoba, Gampela, Nyankpala, and Wa using a cultivar that has a long cycle (129 days), high XFRT (0.85), and no disease. This virtual cultivar was simulated with LFMAX and SFDUR at their maximum solved values and PODUR at its minimum solved value, along with a 13 day longer cycle and 18.1 % higher XFRT, feasible values from DSSAT Gampela had the highest possible pod yield of 5967 kg ha1 and Nyankpala had the lowest yield of 2633 kg ha1. Farakoba and Wa had yield of 3566 and 3449 kg ha1, respectively. The mean pod yield of this virtual cultivar over all four locations was 3904 kg ha1. These high yields are comparable to reported high yields in other regions and represent yields possible with current weather and soils in West Africa, with the best genetic traits along with good fungicide control of disease or perfect leaf spot resistance. Genotype by Environment Interaction Analysis Figure 51 indicates the adaptive responses of four existing cultivars (G1, G2, G10, and G11) and seven virtual cultivars (G3, G4, G5, G6, G7, G8, and G9) to environments (2 years by 4 locations) The existing cultivars were simulated using their

PAGE 103

103 cultivar coefficients that were derived in Chapter 4. Genotypes G1 (CHINESE) and G2 (TS 32 1) are farmer check cultivars and G10 (NKATESARI) and G11 (ICGV IS 9614) are improved cultivars. Regression coefficients were obtained for the regression of genotype mean yield for each location upon the mean of all genotypes for each location. A regression coefficient greater than 1.0 indicates the adaptability of the cultivar to high yielding environments, a value less than 1.0 indicates the adaptability of the cultivar to low yielding environments Cultivars with high mean yields and a regression coefficient close to unity are considered stable with broad adaptation to all environments. Figure 52 illustrates the pod yie ld of each cultivar across environments against the regression coefficients. This relationship measures the spatial stability (adaptability) of the lines to mostly soil conditions and associated climatic environments The broken lines represent 1 standard error. This relationship indicates that the regression coefficient is intimately linked to yield potential, with higher yielding cultivars having higher regression coefficients, and less yield stability. The short season farmer cultivars had both low yields along with yield stability (low regression coefficient ) These simulations if correct highlight a problem with over reliance on regression analyses to evaluate stability and adaptability. Weather induced Variability in Pod Yield as Indicator of Traits for Yield Stability The variability in simulated pod yield caused by weather variability at Wa, Nyankpala, Gampela, and Farakoba is shown for the differe nt cultivar traits in Table 512. This will be an important statement about whether modifying a given trait increases or decreases yield variability (stability). The coefficient of variation (CV) for the standard (median) trait was 7.1 % at Wa and 5.0 % at Nyankpala. The CV of the standard cultivar at Gampela and Farakoba was 3.6 % and 5.1 % respectively. The CVs were generally

PAGE 104

104 low, possibly reflecting minimal variation in solar radiation in weather years, along with warm nonlimiting temperatures and generally sufficient rainfall (which highlights concern for true variability in the used weather data). Increasing LFMAX from 1.04 to 1.60 mg CO2 m2s1 increased the CV from 6.9 % to 7.4 % at Wa. Similar observations were made for Nyankpala and Gampela where increased LFMAX increased the CV for pod yield. A lower CV was observed at higher LFM AX for Farakoba. Decreasing PODUR (adding pods faster) from 20.7 to 5 photothermal days (PD) increased the CV for pod yield at Wa from 6.7 % to 7.5 %. Similar observations were made at Nyankpala, Farakoba, and Gampela, where decreasing PODUR increased the CV for pod yield. Increasing SFDUR from 20.0 to 25.6 PD increased the CV at Wa, Nyankpala, and Gampela, but reduced the CV at Farakoba slightly. Increasing THRSH from 70.0 to 76.0 % decreased the CV at Wa, Nyankpala and Gampela but increased the CV at Farakoba. When XFRT was increased from 0.42 to 0.72, CV increased at Wa, Nyankpala, Gampela and Farakoba with increase in XFRT. Generally, increasing LFMAX, SFDUR, and XFRT as well as decreasing PODUR gave relatively higher CVs. Not much variation was observed with THRS H. Most important is that yield improving traits (higher LFMAX, SFDUR, XFRT, and lower PODUR) all tended to increase CV meaning higher yielding cultivars (with improved traits) are more susceptible to weather variability Similar observations were made for solved cultivars in this experiment where higher yielding cultivars ( with those same higher value traits) were associated with higher CV This is consistent with prior simulations (Boote and Jones, 1986), and is expected, as ultimate yield stability and lowest CV (of zero) would be zero yield which is obtainable with 0.0 XFRT or 0.0

PAGE 105

105 LFMAX The farmer check cultivars had relatively lower CV than the improved cultivars, but is that a reason to ignore yield improvement? As shown in Figure 52, it a ppears that the search for higher yield (via yield improvement traits) is linked and associated with steeper regression coefficients (and potentially less yield stability). In conclusion, this study provided a hypothetical evaluation of crop performance of different peanut cultivars to variation on genetic traits under different weather and soil conditions. Simulated pod yield using the CROPGRO Peanut model showed that increasing LFMAX, SFDUR and XFRT increased pod yield for all life cycl es. Lower values of PODUR were associated with higher simulated pod yield while increasing THRSH reduced pod yield slightly. Life cycle also influenced pod yield such that longer life cycle increased pod yield. In the case of leafspot disease, cultivars wi th higher level of disease resistance had higher simulated pod yield. S election for these traits can be incorporated in to peanut breeding programs for higher yield by measuring these traits. The XFRT influences pod harvest index (HI) and therefore selection for XFRT can be done by selecting for high HI. Selecting for LFMAX can be done by measuring light saturated leaf photosynthesis or measuring biomass and yield, which is easier to do. Life cycle and disease resistance are traits that can also be m easured.

PAGE 106

106 Table 51. Cultivar coefficients of individual peanut cultivars derived with data from Wa and Nyankpala in Ghana in 2010 and 2011 using the CROPGRO Peanut model with an optimizer. Cultivar traits LFMAX XFRT PODUR SFDUR THRSH EM FL FL SD SD PM Cultivar (mg CO 2 m 2 s 1 ) (fraction) (PD) (PD) (%) (PD) (PD) (PD) CHINESE 1.17 0.59 7.5 20.0 71.5 19.9 22.7 35.14 DOUMBALA 1.16 0.58 9.5 22.5 70.0 19.0 20.0 39.50 TS 32 1 1.10 0.59 8.0 24.0 72.0 20.0 20.5 38.20 ICGV IS 92101 1.44 0.59 15.0 22.0 70.0 21.8 25.0 50.00 PC 79 79 1.32 0.42 5.0 24.0 75.8 24.5 28.5 42.80 GM 515 1.04 0.52 7.0 25.6 74.8 21.0 28.0 47.00 ICGV IS 96814 1.60 0.72 20.7 25.0 72.8 21.0 23.4 50.70 G204TX95 1.13 0.54 9.0 23.0 71.6 23.8 25.5 47.40 G122TX95 1.09 0.49 7.0 22.0 74.5 23.8 28.8 38.40 GUSIE BALIN (92099) 1.36 0.63 15.5 24.0 70.2 23.8 23.4 49.50 ICGV IS 96895 1.23 0.49 8.0 25.0 70.5 21.5 30.6 47.40 GM 57 1.15 0.58 12.0 20.0 73.5 22.5 26.8 45.00 ICGV(FDRS) 20F MIX 39 1.42 0.62 11.0 24.0 72.1 21.0 27.0 47.00 ICGV IS 92093 1.55 0.71 18.0 24.0 74.9 21.0 26.0 48.00 GM 123 1.12 0.58 11.0 20.0 74.5 21.0 26.9 53.00 NKATESARI 1.57 0.70 15.0 23.0 73.9 20.0 25.0 50.00 NC 7 1.44 0.69 15.0 23.0 71.4 21.0 24.0 51.50 B106TX95 1.20 0.44 5.0 20.0 76.0 23.8 33.0 37.00 F MIXSINK 24 1.56 0.61 14.4 22.0 71.0 21.0 25.5 49.00 PD, Photothermal days

PAGE 107

107 Table 52. Life cycle traits of three standard life cycle peanut cultivars for Wa Farakoba, Gam pela and Nyank pala. Trait Short cycle Medium cycle Long cycle EM FL 19.0 20.0 21.0 FL SH 9.0 10.0 11.0 FL SD 20.0 22.0 24.0 SD PM 51.0 61.0 73.0

PAGE 108

108 Table 53 ICRISAT score, leafspot (percent necrosis) at 60 and 90 days after sowing and necrosis progress (slope) for 20 genotypes over 4 locations in 2010 and 2011. ICRISAT Score Percent necrosis Genotype Day 60 Day 90 Day 60 Day 90 Slope DOUMBALA 3.7a 7.8a 4.1a 10.3a 0.207 bc TS 32 1 3.6ab 8.0a 4.0ab 10.6a 0.220 ab CHINESE 3.4bc 7.9a 3.7ab 10.5a 0.227 a GM 123 3.3cd 7.0b 3.5abc 9.1b 0.187 de GM 57 3.2cd 6.8bc 3.4abcd 8.9b 0.183 de GM 515 3.1de 6.6c 3.2bcde 8.5b 0.177 efgh NKATESARI 2.9ef 5.9e 2.9cdef 7.5dc 0.153 hijk G204TX95 2.8fg 6.1d 2.8defg 7.8c 0.167 ghij ICGV (FDRS) 20 F MIX 39 2.7fgh 6.6c 2.6efgh 8.5b 0.197 cd ICGV IS 96895 2.6ghi 5.7efg 2.5fgh 7.2cde 0.157 hijk PC 79 79 2.6ghi 5.2hij 2.5fgh 6.4fg 0.130 ml B106TX95 2.6ghi 5.0j 2.3 fgh 6.1g 0.127 m ICGV IS 92093 2.5hij 5.8ef 2.3fgh 7.3cde 0.167 fghi F MIX SINK 24 2.4ijk 5.5fgh 2.2fghi 6.9def 0.157 hijk GUSIE BALIN (92099) 2.4ijk 5.4ghi 2.2fghi 6.7fe 0.150 ijk NC 7 2.3jkl 5.9e 2.0ghi 7.5cd 0.183 def ICGV IS 96814 2.3jkl 5.4ghi 2.0ghi 6.7fe 0.157 jkl F MIX 2.2kl 5.1ij 1.9hi 6.3fg 0.147 kl ICGV IS 92101 2.2kl 5.8ef 1.9hi 7.3cde 0.180 defg G122TX95 2.1l 4.9j 1.7i 6.0g 0.143 jkl Means followed by the same letter in a column are not significantly different at P < 0.05

PAGE 109

109 Table 54 Cultivar coefficients (low, median, and high for five traits) of individual virtual peanut cultivars derived for s hort cycle cultivars. Cultivar traits EM FL FL SH FL SD SD PM LFMAX XFRT PODUR SFDUR THRSH Cultivar PD PD (PD) (PD) (mg CO 2 m 2 S 1 ) (PD) (PD) (PD) (%) LFMAX 1.04 19.0 9.0 20.0 51.0 1.04 0.59 11.0 23.0 72.1 LFMAX 1.23 19.0 9.0 20.0 51.0 1.23 0.59 11.0 23.0 72.1 LFMAX 1.60 19.0 9.0 20.0 51.0 1.60 0.59 11.0 23.0 72.1 XFRT 0.42 19.0 9.0 20.0 51.0 1.23 0.42 11.0 23.0 72.1 XFRT 0.59 19.0 9.0 20.0 51.0 1.23 0.59 11.0 23.0 72.1 XFRT 0.72 19.0 9.0 20.0 51.0 1.23 0.72 11.0 23.0 72.1 PODUR 5.0 19.0 9.0 20.0 51.0 1.23 0.59 5.0 23.0 72.1 PODUR 11.0 19.0 9.0 20.0 51.0 1.23 0.59 11.0 23.0 72.1 PODUR 20.7 19.0 9.0 20.0 51.0 1.23 0.59 20.7 23.0 72.1 SFDUR 20.0 19.0 9.0 20.0 51.0 1.23 0.59 11.0 20.0 72.1 SFDUR 23.0 19.0 9.0 20.0 51.0 1.23 0.59 11.0 23.0 72.1 SFDUR 25.6 19.0 9.0 20.0 51.0 1.23 0.59 11.0 25.6 72.1 THRSH 70.0 19.0 9.0 20.0 51.0 1.23 0.59 11.0 23.0 70.0 THRSH 72.1 19.0 9.0 20.0 51.0 1.23 0.59 11.0 23.0 72.1 THRSH 76.0 19.0 9.0 20.0 51.0 1.23 0.59 11.0 23.0 76.0 PD, Photothermal days

PAGE 110

110 Table 55. Percent necrosis derived from disease data collected at each site and assigned to susceptible, moderately susceptible, and resistant peanut cultivars for Gampela, Farakoba, Nyankpala and Wa. LOCATION Leafspot resistance 60 DAS 80 DAS 90 DAS 103 DAS 116 DAS Necrosis (%) Susceptible 3.6 6.0 10.1 10.1 10.1 GAMPELA Moderately susceptible 2.8 3.7 5.9 5.9 5.9 Resistant 2.7 3.2 4.5 4.5 4.5 Susceptible 2.3 5.4 10.4 10.4 10.4 FARAKOBA Moderately Susceptible 1.9 3.1 6.9 6.9 6.9 Resistant 0.5 2.6 5.5 5.5 5.5 Susceptible 3.2 6.6 8.4 8.4 8.4 NYANKPALA Moderately Susceptible 2.9 5.1 6.7 7.0 7.0 Resistant 2.2 4.3 6.1 6.5 6.5 Susceptible 3.8 7.0 9.9 9.9 9.9 WA Moderately susceptible 2.5 5.9 7.3 8.0 8.0 Resistant 1.6 5.6 6.8 7.6 7.6 DAS, days after sowing used same value as at 90 DAS. used same value as at 100 DAS.

PAGE 111

111 Table 56. Percent defoliation derived from disease data collected at each site and assigned to susceptible, moderately susceptible, and resistant peanut cultivars for Gampela, Farakoba, Nyankpala and Wa. LOCATION Leafspot resistance 60 DAS 80 DAS 90 DAS 103 DAS 116 DAS Defoliation (%) Susceptible 28.1 46.9 78.9 78.9 78.9 GAMPELA Moderately susceptible 21.9 28.9 46.1 46.1 46.1 Resistant 21.1 25.0 35.2 35.2 35.2 Susceptible 18.0 42.2 81.3 81.3 81.3 FARAKOBA Moderately Susceptible 14.8 24.2 53.9 53.9 53.9 Resistant 3.9 20.3 43.0 43.0 43.0 Susceptible 25.0 51.6 65.6 65.6 65.6 NYANKPALA Moderately Susceptible 22.7 39.8 52.3 57.4 57.4 Resistant 17.2 33.6 47.7 53.3 53.3 Susceptible 29.7 54.7 77.3 77.3 77.3 WA Moderately susceptible 19.5 46.1 57.0 65.6 65.6 Resistant 12.5 43.8 53.1 62.3 62.3 DAS days after sowing used value at 90 DAS used values at 100 DAS

PAGE 112

112 Table 57. Simulated pod yield (kg/ha) of short, medium, and long life cycles at susceptible, moderately susceptible and resistant disease for Wa, Nyankpala, Gampela, and Farakoba, using median values for other traits. Life Cycle Location Disease Short Medium Long Yield (kg/ha) Yield (kg/ha) Yield (kg/ha) Susceptible 1068 1160 1249 Wa Mod. susceptible 1216 1327 1437 Resistant 1304 1407 1505 Susceptible 840 852 929 Nyankpala Mod. susceptible 910 949 1039 Resistant 986 1089 1139 Susceptible 2394 2480 2665 Gampela Mod. susceptible 2622 3038 3070 Resistant 3040 3236 3292 Susceptible 1288 1398 1477 Farakoba Mod. susceptible 1620 1654 1749 Resistant 1757 1818 1901 Table 58. Simulated pod yield of short, medium, and long life cycles at susceptible, moderately susceptible and resistant disease averaged over four sites, with median calibrated values for other traits, compared with two versions where SFDUR was linked by a ratio to SDPM Life Cycle SFDUR SDPM Disease Short Medium Long Yield (kg/ha) Yield (kg/ha) Yield (kg/ha) Not linked Susceptible 1398 1473 1580 Not linked Mod. susceptible 1592 1742 1824 Not linked Resistant 1772 1888 1959 Linked* Susceptible 1288 1473 1708 Linked* Mod. susceptible 1587 1742 1960 Linked* Resistant 1698 1888 2074 Linked** Susceptible 1477 1695 1920 Linked** Mod. susceptible 1746 1960 2128 Linked** Resistant 1871 2109 2317 Linked*, SFDUR was 0.377 of SD PM of 19 calibrated cultivars Linked**, SFDUR was 0.495. of SD PM of DSSAT standard cultivars

PAGE 113

113 Table 59. Pod yield for 5 traits as percent change from standard (median) values for short, medium, and long cycle cultivars at susceptible, moderately susceptible, and resistant disease levels averaged over Wa, Nyankpala, Gampela, and Farakoba. SFDUR change is compared to median values. Disease Trait Short life cycle Medium life cycle Long life cycle Change in pod yield (%) Change in pod yield (%) Change in pod yield (%) Min Max Min Max Min Max Susceptible LFMAX 14.5 +21.3 17.5 +17.2 17.2 +17.9 Susceptible PODUR +9.1 3.8 +9.7 5.9 +10.9 7.5 Susceptible SFDUR 5.3 +6.7 7.8 +5.7 7.0 +27.6 Susceptible THRSH + 0.4 0.3 +0.5 0.6 +0.7 0.9 Susceptible XFRT 27.5 +19.0 10.1 +15.8 35.0 +16.0 Change in pod yield (%) Change in pod yield (%) Change in pod yield (%) Min Max Min Max Min Max Mod.susceptible LFMAX 10.3 +19.7 17.4 +15.3 16.1 +14.9 Mod.susceptible PODUR +9.7 7.3 +18.8 +0.6 +13.4 3.1 Mod.susceptible SFDUR 1.0 +4.2 3.6 +3.6 5.2 +4.7 Mod.susceptible THRSH +2.6 +0.4 +2.3 3.1 +0.5 1.6 Mod.susceptible XFRT 23.8 +17.0 28.5 +12.7 31.6 +14.4 Change in pod yield (%) Change in pod yield (%) Change in pod yield (%) Min Max Min Max Min Max Resistant LFMAX 15.1 +16.1 14.8 +14.4 16.0 +17.7 Resistant PODUR +4.8 3.9 +13.8 +4.8 +13.9 1.1 Resistant SFDUR 4.2 +6.1 3.8 +4.7 5.6 +3.5 Resistant THRSH +3.0 2.8 +3.9 3.7 +1.0 2.1 Resistant XFRT 27.4 +12.8 24.3 +12.0 29.5 +13.0 Min minimum trait value Max, maximum trait value

PAGE 114

114 Table 510. Simulated pod yield of hypothetical and actual peanut cultivars averaged over 30 years of weather data over Nyankpala, Wa, Farakoba and Gampela. Cultivar Yield Change *** C um**** percent Kg/ha % % CHINESE 1199 0.0 0.0 TS 32 1 1254 4.6 4 .6 Short cycle, susceptible, median rest of traits 1398 11.5 16. 6 Long cycle, susceptible, median rest of traits 1582 13.2 31.9 Long cycle, resistant, median rest of traits 1958 23.8 63.3 Long cycle, resistant, LFMAX max median rest of traits 2382 21.7 98.6 NKATESARI 2453 3.0 104.6 ICGV IS 96814 2501 2.0 108.6 Long cycle, resistant, LFMAX max XFRT max median rest of traits 2656 6.2 121.5 Long cycle, resistant, LFMAX max XFRT max SFDUR max median rest of traits 2791 5.1 132.8 Long cycle, resistant, LFMAX max XFRT max SFDUR max PODUR min ** median rest of traits 2898 3.8 141.7 farmer check cultivar improved cultivar LFMAX, maximum leaf photosynthesis rate ( mg CO2 m2s1) PODUR, duration of adding pods (photothermal days) XFRT, maximum fraction of daily growth partitioned to seed plus shell SFDUR, seed filling duration for individual pod cohort (photothermal days) max*, maximum trait value: LFMAXmax = 1.60, XFRTmax = 0.72, SFDURmax = 25.6 min**, minimum trait value: PODURmin ** = 5.0 NKATESARI: long cycle, resistant, LF M AX (1.57), XFRT (0.70 ), PODU R (15.0) SFDUR (23.0 ) ICGV IS 96814: long cycle, resistant, LFMAX (1.60), XFRT (0.72), PODUR (20.7) SFDUR (25 .0) CHINESE: short cycle, susceptible, LFMAX (1.17), XFRT (0.59), PODUR (7.5) SFDUR (20.0) TS 32 1: shor t cycle, susceptible, LFMAX (1.10), XFRT (0 .59 ), PODUR (8.0) SFDUR (24.0) Change***, percent change from prior yield Cum**** percent cumulative percent compared to Chinese

PAGE 115

115 Table 511. Simulated possible pod yield averaged over 30 years of weather data for Farakoba, Gampela, Nyankpala and Wa for a long cycle virtual cultivar with perfect disease control, and high partitioning. Location Pod yield (kg/ha) Farakoba 3566 Gampela 5967 Nyankpala 2633 Wa 3449 Mean 3904

PAGE 116

116 Table 512. Weather induced c oefficient of variation of pod yield for different cultivar traits at Wa, Nyankpala, Gampela, and Farakoba. Locations Cultivar Wa Nyankpala Gampela Farakoba CV* (%) CV* (%) CV* (%) CV* (%) Standard (median trait) 7.1 5.0 3.6 5.1 LFMAX (min) 6.9 4.5 3.5 5.7 LFMAX (max ) 7.4 6.8 5.5 4.7 PODUR (min) 7.5 6.0 4.6 7.5 PODUR (max ) 6.7 4.2 3.8 5.6 SFDUR (min) 6.9 4.7 4.1 5.4 SFDUR (max ) 7.1 7.4 5.3 5.1 THRSH (min) 7.1 4.2 3.2 5.2 THRSH (max ) 6.9 3.9 3.1 5.4 XFRT (min) 6.7 4.1 3.5 4.8 XFRT (max ) 7.7 6.1 5.1 5.5 CV** (%) CV** (%) CV** (%) CV** (%) TS 32 1 6.2 4.6 2.7 5.3 CHINESE 6.5 4.2 2.7 5.0 Short cycle, susceptible, median rest of traits 6.8 4.6 2.7 5.3 Long cycle, susceptible, median rest of traits 7.2 3.9 3.0 5.8 Long cycle, resistant, median rest of traits 6.5 5.4 4.3 5.8 Long cycle, resistant, LFMAXmax median rest of traits 6.6 5.2 3.8 5.5 NKATESARI 7.6 5.0 3.0 6.7 ICGV IS 96814 8.8 5.1 4.9 8.4 Long cycle, resistant, LFMAXmax XFRTmax median 7.2 5.1 4.6 5.8 rest of traits Long cycle, resistant, LFMAXmax XFRTmax SFDURmax 8.0 6.2 4.0 6.3 median rest of traits Long cycle, resistant, LFMAXmax XFRTmax SFDURmax 9.8 6.1 4.9 6.2 PODURmin median rest of traits min minimum trait value m ax maximum trait value CV*, average of 9 CVs from 3 life cycles by 3 disease levels CV**, CV from 30 years of weather data.

PAGE 117

117 Figure 51. Adaptive responses to environments for pod yield of 11 peanut genotypes G1= CHINESE G2= T2 321 G3= short cycle, susceptible, median rest of traits G4= long cycle, susceptible, median rest of traits G5= long cycle, resistant, median rest of traits G6= long cycle, resistant, LFMAXmax *, median rest of traits G7= long cycle, resistant, LFMAXmax *, XFRTmax *, median rest of traits G8= Long cycle, resistant, LFMAXmax, XFRTmax, SFDURmax, median rest of traits G9= Long cycle, resistant, LFMAXmax *, XFRTmax *, SFDURmax *, PODURmin **, median rest of traits G10= NKATESARI G11= ICGV IS 96814 max*, maximum trait value: LFMA Xmax* = 1.60, XFRTmax* = 0.72, SFDURmax* = 25.6 min**, minimum trait value: PODURmin** = 5.0 0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000Cultivar pod yield (kg/ha) Environment mean pod yield (kg/ha) G1 G2 G3 G4 G5 G6 G7 G8 G9 G10 G11

PAGE 118

118 Figure 52. Relationship between mean pod yield and regression coefficient of 19 peanut genotypes G1= CHINESE G2= T2 321 G3= short cycle, susceptible, median rest of traits G4= long cycle, susceptible, median rest of traits G5= long cycle, resistant, median rest of traits G6= long cycle, resistant, LFMAXmax, median rest of traits G7= long cycle, resistant, LFMAXmax, XFRTmax, median rest of t raits G8= Long cycle, resistant, LFMAXmax, XFRTmax, SFDURmax, median rest of traits G9= Long cycle, resistant, LFMAXmax, XFRTmax, SFDURmax, PODURmin, median rest of traits G10= NKATESARI G11= ICGV IS 96814 400 900 1400 1900 2400 2900 3400 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2Mean pod yield (kg/ha) Regression coefficient

PAGE 119

119 CHAPTER 6 SUMMARY AND CONCLUSIONS The objectives of this study were to conduct cultivar screening trials to identify cultivars that are superior (top ranking performance) among test entries and to identify the range of adaptability of the different peanut cultivars in the screening trials on the basis of yield stability. Additional objectives were to calibrate the CROPGRO peanut model for the cultivars screened in field trials, to investigate the causes or traits associated with the simulate d genetic improvement of yield among these cultivars and to evaluate the yi eld response of the different peanut cultivars to variation in genetic traits under different weather and soil conditions using the CSM CROPGRO Peanut model Twenty peanut genotypes wer e tested at two sites in Ghana and two sites in Burkina Faso in 2010 and 2011. The experiments in Burkina Faso were conducted at the Environmental and Agricultural Research Institute in Gampela, Ouagadougou and Farakoba, Bobodilaso. The Ghana experiments were conduct ed at the Savanna Agriculture Research Institute sites at Nyankpala and Wa. Pod yield, shelling percentage and plant stand were measured at all four locations at the end of the season. Timeseries disease data based on the ICRISAT disease score was collect ed at all four locations. Timeseries data on pod harvest index, total crop biomass, and pod mass were measured at Wa and Nyankpala in 2010 and 2011. Data was also obtained on days from sowing to specific reproductive growth stages at Wa, Ghana in 2010. Ge notype by environment interaction analysis was conducted on pod yield over 8 environments (2 years by 4 locations). In addition, crop model simulations were conducted with an optimizer to estimate (solve) cultivar specific traits for each of the cultivars by comparison to data collected from the field experiments at Wa and

PAGE 120

120 Nyankpala in 2010 and 2011. Model simulations were also performed for hypothetical evaluation of yield response to variation in genetic traits of peanut cultivars. Field research showed t hat there was large variation in both genotype and environment (site) mean pod yield. The highest yielding environment was Gampela with a pod yield of 1754 kg/ha followed by Wa and Nyankpala with pod yield of 1036 kg/ha and 963 kg/ha, respectively. Farakoba was the lowest performing environment with mean pod yield of 938 kg/ha. The low performance observed for Farakoba may be attributed to the high clay content of the soils, which created flooded conditions (poor internal drainage) after heavy rainfall especially in 2010. This soil was also characterized by a subsurface accumulation of Aluminum which may have affected root growth. Genotype mean yield ranged from 805 kg/ha for GM 515 to 1755 kg/ha for ICGV IS 96814, and by contrast to the short season farmer check cultivars which yielded 889 to 979 kg/ha. From the genotype by environment interaction analysis, genotypes ICGV (FDRS) 20 F MIX 39, GUSIE BALIN (92099 ) ICGV IS 92093, ICGV IS 92101 and ICGV IS 96814 are considered cultivars with broad adaptability as they have above average mean yield across sites and a regression coefficient close to 1.0. Among these four cultivars, ICGV IS 96814 was considered the best because it adequately demonstrated wide or broad adaptation across environments. This is because it produced the highest pod yield and had a regression coefficient close to unity. Therefore, genotype ICGV IS 96814 is less responsive to changed environmental conditions and can be grown over a range of environments in West Africa. However, released c ultivar NKATESARI could be considered equivalent in some respects because it had pod yield equal to ICGV IS 96814, but with a higher regression slope.

PAGE 121

121 Cultiva rs G122TX95 was the worst with a stability regression coefficient of 0.72 and mean yield of 824 kg /ha The farmer check cultivars CHINESE, DOUMBALA and TS 321 flowered earlier than all the remaining cultivars. Beginning peg (R2) also occurred earlier in the farmer check cultivars. There was more than 2 weeks variation in beginning pod (R3) with the fa rmer check cultivars starting earlier. Beginning pod occurred at 44 DAS for TS 321 and was as late as 60 DAS for B106TX95. The farmer check cultivars reached full sized pod (R4) first with values ranging from 49 DAS for DOUMBALA to 68 DAS for G122TX95. The earliest maturing cultivars were the farmer check cultivars CHINESE, DOUMBALA and TS 32 1 which matured at 90 DAS. The remaining cultivars all matured between 108 and118 DAS Genotypespecific traits (for the model) were optimized from timeseries data on 1 plant total biomass, pod mass, and pod harvest index collected in 2010 and 2011 at Wa and Nyankpala. Because the data for optimization of cultivar coefficients was derived from 1plant samples, there was considerable variability in biomass and pod mass and thus more error was associated with prediction of biomass and pod mass, than for pod harvest index. T he optimization procedure estimated cultivar coefficients that provided simulated pod yield that agreed quite well with the observed pod yield. There was also considerable genetic variation in solved cultivar traits among peanut cultivars. Model evaluation with independent pod yield data not used in model calibration (Gampela and Farakoba over two years ), provided a good test of how well model predictions worked in an independent validation as Gampela had a d statistic of 0.88 and RMSE of 433 kg ha1. Farakoba had a d statistic and RMSE of 0.77 and 444

PAGE 122

122 kg ha1, respectively Evaluation of the calibration procedure showed that there were increases in d statistic and a reduction in RMSE over 19 cultivars per site, after successively accounting for life cycle effects, disease input, and solving for genetic traits. The derived cultivar coefficients over four sites and two years allowed the CROPGRO Peanut m odel to mimic yield ranking quite well and suggests value in using the model to hypothesize genetic improvement (combinations of traits for best yield and stability) for target environments where long term weather data is available. Evaluation of crop perf ormance of different peanut cultivars to variation in solved genetic traits showed that higher yielding cultivars were associated with higher XFRT higher LFMAX, longer SFDUR and shorter PODUR. The trait THRSH had minimal effect on pod yield. Life cycle also influenced pod yield such that longer life cycle increased pod yield. Generally, pod yield increase due to increase in life cycle was relatively smaller than anticipated when SFDUR was not linked to SD PM. Correlation coefficient estimated between SFD UR and SDPM of the standard DSSAT cultivars. was strong (r = 0.85, n = 37), and significant (P< 0.05). Similar correlation analysis using SFDUR and SD PM of the 19 calibrated cultivars gave a strong (r = 0.71, n = 19)) and significant (P< 0.05) relationship. When the SFDUR is linked to SDPM, with a constant ratio of 0.377, yield increased from 1788 to 1914 kg ha1. But when SFDUR is linked to SD PM (based on the ratio of 0.495 of DSSAT cultivars) there was a greater yield increase with longer life cycle. Using the constant ratios of SFDUR to SD PM also led to higher percent increase in yield from short to long life cycle. When SFDUR was not linked to SD PM, percent increase in yield from short to long life cycle ranged between 10.6 and 14.6 %. When the SFD UR was linked to SD PM with a constant ratio

PAGE 123

123 of 0.377, percent increase in yield from short to long life cycle was in the range of 18.6 to 32.6 %. When SFDUR was linked to SD PM (based on the ratio of 0.495 of DSSAT cultivars), there was a 21.9 to 30.0 % y ield increase from short to long life cycle. In a comparison of the benefits of leaf spot disease, cultivars with higher level of disease resistance had higher simulated pod yield for all life cycles. Varying the cultivar traits LFMAX, XFRT, PODUR, SFDUR, and THRSH from minimum to maximum had different percent yield responses. Percent yield response due to changes in XFRT and LFMAX were largest in general, whereas effects of changes in THRSH were minimal. Evaluation of possible cultivar yield from combinati ons of cultivar traits showed that the farmer check cultivars CHINESE and TS 321 had the lowest yield of 1199 and 1254 kg ha1, respectively. Generally, the resistant long cycle cultivars had higher yield. The resistant long cycle cultivar with maximum tr ait value for XFRT, SFDUR, LFMAX and minimum trait value for PODUR had the highest yield of 2898 kg ha1, larger than even the improved cultivar NKATESARI with a pod yield of 2454 kg ha1 and ICGV IS 96814 with pod yield of 2501 kg ha1. Therefore improved cultivars NKATESARI and ICGV IS 96814 were among the top yielding cultivars, indicating that these traits are all possible and have been achieved to a large extent compared to the farmer check cultivars. T he hi ghest yielding peanut lines had high values for LFMAX, XFRT, and SFDUR but a low value of PODUR. S election for these traits can be incorporated in to peanut breeding programs for higher yield by measuring these traits. The XFRT influences pod harvest index (HI) and therefore selection for XFRT can be done by selecting for high HI. Selecting for LFMAX can be done by measuring light saturated leaf photosynthesis or measuring biomass and yield. Life cycle and disease resistance

PAGE 124

124 are traits that can also be measured. Yield improving traits (higher XFRT, LF MAX, SFDUR, and lower PODUR) all tended to increase CV in pod yield caused by weather variability. The results show that high yielding cultivars, mostly from ICRISAT lines or ICRISAT derived crosses, yielded nearly 80 % more than the standard farmer check cultivars like CHINESE. Pod yield for CHINESE was 979 kg ha1 and ICGV IS 96814 was 1755 kg ha1. These higher yielding cultivars had longer life cycles, higher disease resistance, and higher partitioning than the farmer check cultivars. It appears that i mproved cultivars are already available. It is recommended that these improved cultivars be multiplied and distributed to farmers in a system that includes onfarm demonstrations to enhance farmer acceptability.

PAGE 125

125 APPENDIX A LEAFSPOT RATING SCALES Table A 1. ICRISAT 1 9 rating scale (based on Subrahmanyam et al., 1995) Rating Description Disease severity (%) 1 No disease 0 2 Lesions largely on lower leaves; no defoliation 1 5 3 Lesion largely on lower leaves; very few lesions on middle leaves; defoliation of some leaflets evident on lower leaves 6 10 4 Lesions on lower and middle leaves; but severe on lower leaves; defoliation of some leaflets evident on lower leaves 11 20 5 Lesions on all lower and middle leaves; ov er 50 % defoliation of lower leaves 21 30 6 Lesions severe on lower and middle leaves; lesions on top leaves but less severe; extensive defoliation of lower leaves; defoliation of some leaflets evident on middle leaves 31 40 7 Lesions on all leaves but less severe on top leaves; defoliation of all lower and some middle leaves 41 60 8 Defoliation of lower and middle leaves; lesions severe on top leaves and some defoliation of top leaves evident 61 80 9 Defoliation of almost all leaves leaving bare stems; some leaflets may be Present, but with severe leafspot 81 100

PAGE 126

126 APPENDIX B SOIL PROPERTIES AT THE EXPERIMENTAL SITES Table B 1. Some physical and chemical properties of the so ils used in the model. Depth SLL L SDU L SSAT (cm) (cm 3 cm 3 ) (cm 3 cm 3 ) (cm 3 cm 3 ) pH (H 2 O) Al (ppm) Wa 0 5 0.054 0.187 0.309 5.8 5 10 0.052 0.183 0.309 5.8 10 15 0.059 0.184 0.309 5.8 15 20 0.052 0.180 0.313 5.5 20 40 0.055 0.167 0.323 5.5 40 50 0.055 0.167 0.323 5.5 Nyankpala 0 5 0.095 0.257 0.359 6.7 5 15 0.060 0.227 0.359 6.7 15 30 0.06 0.228 0.340 6.4 30 45 0.105 0.229 0.342 6.6 45 60 0.120 0.205 0.342 6.1 60 90 0.130 0.200 0.347 6.0 6.0 Farakoba 0 14 0.068 0.142 0.403 5.5 0.0 14 23 0.145 0.230 0.388 6.1 0.0 23 43 0.265 0.363 0.421 5.3 0.6 43 62 0.261 0.359 0.418 4.8 1.6 62 90 0.243 0.342 0.411 4.9 2.0 90 120 0.229 0.326 0.400 5.1 1.7 120 140 0.179 0.260 0.379 5.3 1.1 140 174 0.219 0.319 0.408 5.4 1.7 174 200 0.225 0.326 0.408 5.4 2.3 Gampela 0 15 0.088 0. 80 0.402 4.7 0.0 15 29 0.108 0.200 0.388 4.6 0.0 29 44 0.166 0.2 76 0.388 4.5 0.0 44 65 0.222 0.3 59 0.395 4.5 0.3 65 88 0.217 0.3 49 0.392 4.5 0.1 88 110 0.210 0. 339 0.393 4.7 0.0 110 135 0.203 0. 326 0.393 4.8 0.0 135 155 0.194 0. 314 0.390 4.9 0.0 155 180 0.176 0.2 77 0.389 5.0 0.0 180 210 0.157 0.249 0.393 5.3 0.0 SDUL, upper limit, drained SLL, lower limit, drained SSAT, upper limit, saturated

PAGE 127

127 APPENDIX C LEAFSPOT DISEASE DATA FROM EXPERIMENTAL SITES Table C 1. Data of leaf spot score and corresponding percent necrosis and percent defoliation for Farakoba, 2010. ICRISAT Cultivar DOY score PDLA (%) PCLA (%) CHINESE 256 1.7 1.1 8.8 CHINESE 276 3.7 4.1 33.8 CHINESE 286 8.7 11.7 96.3 DOUMBALA 256 2.0 1.6 12.5 DOUMBALA 276 3.0 3.1 25.0 DOUMBALA 286 8.0 10.6 87.5 TS 32 1 256 1.7 1.1 8.8 TS 32 1 276 3.3 3.5 28.8 TS 32 1 286 8.3 11.1 91.3 ICGV IS 92101 256 1.7 1.1 8.8 ICGV IS 92101 276 2.3 2.0 16.3 ICGV IS 92101 286 6.0 7.6 62.5 PC 79 79 256 1.7 1.1 8.8 PC 79 79 276 2.3 2.0 16.3 PC 79 79 286 4.7 5.7 46.3 GM 515 256 2.0 1.6 12.5 GM 515 276 3.0 3.1 25.0 GM 515 286 6.7 8.7 71.3 ICGV IS 96814 256 1.3 0.5 3.8 ICGV IS 96814 276 2.0 1.6 12.5 ICGV IS 96814 286 5.7 7.2 58.8 G204TX95 256 1.3 0.5 3.8 G204TX95 276 2.7 2.6 21.3 G204TX95 286 6.0 7.6 62.5 G122TX95 256 1.3 0.5 3.8 G122TX95 276 2.3 2.0 16.3 G122TX95 286 5.7 7.2 58.8 GUSIE BALIN (92099) 256 1.0 0.1 0.0 GUSIE BALIN (92099) 276 2.0 1.6 12.5 GUSIE BALIN (92099) 286 5.7 7.2 58.8 ICGV IS 96895 256 1.3 0.5 3.8 ICGV IS 96895 276 2.7 2.6 21.3 ICGV IS 96895 286 6.7 8.7 71.3 GM 57 256 2.0 1.6 12.5 GM 57 276 3.3 3.5 28.8 GM 57 286 6.7 8.7 71.3 ICVG (FDRS) 20 F MIX 39 256 1.3 0.5 3.8 ICVG (FDRS) 20 F MIX 39 276 3.3 3.5 28.8 ICVG (FDRS) 20 F MIX 39 286 7.0 9.1 75.0 ICGV IS 92093 256 1.2 0.4 2.5 ICGV IS 92093 276 2.0 1.6 12.5 ICGV IS 92093 286 6.7 8.7 71.3 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 128

128 Table C 1 Continued ICRISAT Cultivar DOY SCORE PDLA PCLA GM 123 256 2.0 1.6 12.5 GM 123 276 3.3 3.5 28.8 GM 123 286 7.0 9.1 75.0 NKATESARI 256 2.0 1.6 12.5 NKATESARI 276 2.0 1.6 12.5 NKATESARI 286 6.7 8.7 71.3 NC 7 256 1.3 0.5 3.8 NC 7 276 2.0 1.6 12.5 NC 7 286 6.0 7.6 62.5 B106TX95 256 1.3 0.5 3.8 B106TX95 276 3.0 3.1 25.0 B106TX95 286 4.3 5.1 41.3 F MIX SINK 24 256 1.0 0.1 0.0 F MIX SINK 24 276 2.0 1.6 12.5 F MIX SINK 24 286 6.7 8.7 71.3 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 129

129 Table C 2. Data of leaf spot score and corresponding percent necrosis and percent defoliation for Farakoba, 2011. ICRISAT Cultivar DOY score PDLA (%) PCLA (%) CHINESE 248 3.0 3.1 25.0 CHINESE 268 6.0 7.6 62.5 CHINESE 278 7 .7 10.2 83.8 DOUMBALA 248 3.3 3.5 28.8 DOUMBALA 268 5.0 6.1 50.0 DOUMBALA 278 6.3 8.1 66.3 TS 32 1 248 3.3 3.5 28.8 TS 32 1 268 6.0 7.6 62.5 TS 32 1 278 8.0 10.6 87.5 ICGV IS 92101 248 2.7 2.6 21.3 ICGV IS 92101 268 3.5 3.8 31.3 ICGV IS 92101 278 4.3 5.1 41.3 PC 79 79 248 2.0 1.6 12.5 PC 79 79 268 3.7 4.1 33.8 PC 79 79 278 4.7 5.7 46.3 GM 515 248 3.3 3.5 28.8 GM 515 268 3.7 4.1 33.8 GM 515 278 4.7 5.7 46.3 ICGV IS 96814 248 3.0 3.1 25.0 ICGV IS 96814 268 3.0 3.1 25.0 ICGV IS 96814 278 3.7 4.1 33.8 G204TX95 248 3.0 3.1 25.0 G204TX95 268 3.3 3.5 28.8 G204TX95 278 5.3 6.6 53.8 G122TX95 248 2.3 2.0 16.3 G122TX95 268 2.3 2.0 16.3 G122TX95 278 3.3 3.5 28.8 GUSIE BALIN (92099) 248 2.7 2.6 21.3 GUSIE BALIN (92099) 268 2.7 2.6 21.3 GUSIE BALIN (92099) 278 3.7 4.1 33.8 ICGV IS 96895 248 2.7 2.6 21.3 ICGV IS 96895 268 3.3 3.5 28.8 ICGV IS 96895 278 4.3 5.1 41.3 GM 57 248 3.0 3.1 25.0 GM 57 268 4.3 5.1 41.3 GM 57 278 5.3 6.6 53.8 ICVG (FDRS) 20 F MIX 39 248 3.0 3.1 25.0 ICVG (FDRS) 20 F MIX 39 268 3.7 4.1 33.8 ICVG (FDRS) 20 F MIX 39 278 5.0 6.1 50.0 ICGV IS 92093 248 3.0 3.1 25.0 ICGV IS 92093 268 3.0 3.1 25.0 ICGV IS 92093 278 4.7 5.7 46.3 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 130

130 Table C 2 Continued ICRISAT Cultivar DOY SCORE PDLA PCLA GM 123 248 3.0 3.1 25.0 GM 123 268 4.0 4.6 37.5 GM 123 278 4.7 5.7 46.3 NKATESARI 248 3.0 3.1 25.0 NKATESARI 268 3.0 3.1 25.0 NKATESARI 278 3.7 4.1 33.8 NC 7 248 2.3 2.0 16.3 NC 7 268 3.3 3.5 28.8 NC 7 278 4.7 5.7 46.3 B106TX95 248 3.0 3.1 25.0 B106TX95 268 3.3 3.5 28.8 B106TX95 278 4.3 5.1 41.3 F MIX SINK 24 248 3.0 3.1 25.0 F MIX SINK 24 268 3.7 4.1 33.8 F MIX SINK 24 278 4.3 5.1 41.3 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 131

131 Table C 3. Data of leaf spot score and corresponding percent necrosis and percent defoliation for Gampela, 2010. ICRISAT Cultivar DOY score PDLA (%) PCLA (%) CHINESE 250 1.3 0.5 3.8 CHINESE 270 3.0 3.1 25.0 CHINESE 280 9 0 12.1 100.0 DOUMBALA 250 1.7 1.1 8.8 DOUMBALA 270 3.7 4.1 33.8 DOUMBALA 280 9.0 12.2 100 TS 32 1 250 1.7 1.1 8.8 TS 32 1 270 3.7 4.1 33.8 TS 32 1 280 9.0 12.2 100 ICGV IS 92101 250 1.0 0.0 0.0 ICGV IS 92101 270 2.0 1.6 12.5 ICGV IS 92101 280 5.7 7.2 58.8 PC 79 79 250 1.0 0.0 0.0 PC 79 79 270 2.0 1.6 12.5 PC 79 79 280 5.7 7.2 58.8 GM 515 250 1.0 0.0 0.0 GM 515 270 2.0 1.6 12.5 GM 515 280 8.3 11.1 91.3 ICGV IS 96814 250 1.0 0.0 0.0 ICGV IS 96814 270 2.3 2.0 16.3 ICGV IS 96814 280 5.0 6.1 50.0 G204TX95 250 1.0 0.0 0.0 G204TX95 270 2.3 2.0 16.3 G204TX95 280 6.0 7.6 62.5 G122TX95 250 1.0 0.0 0.0 G122TX95 270 2.0 1.6 12.5 G122TX95 280 5.3 6.6 53.8 GUSIE BALIN (92099) 250 1.0 0.0 0.0 GUSIE BALIN (92099) 270 2.0 1.6 12.5 GUSIE BALIN (92099) 280 5.7 7.2 58.8 ICGV IS 96895 250 1.0 0.0 0.0 ICGV IS 96895 270 2.7 2.6 21.3 ICGV IS 96895 280 4.7 5.7 46.3 GM 57 250 1.3 0.5 3.8 GM 57 270 3.0 3.1 25.0 GM 57 280 8.3 11.1 91.3 ICVG (FDRS) 20 F MIX 39 250 1.0 0.0 0.0 ICVG (FDRS) 20 F MIX 39 270 2.7 2.6 21.3 ICVG (FDRS) 20 F MIX 39 280 7.7 10.2 83.8 ICGV IS 92093 250 1.0 0.0 0.0 ICGV IS 92093 270 2.3 2.0 16.3 ICGV IS 92093 280 6.3 8.1 66.3 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 132

132 Table C 3 Continued ICRISAT Cultivar DOY SCORE PDLA PCLA GM 123 250 1.7 1.1 8.8 GM 123 270 2.3 2.0 16.3 GM 123 280 8.7 11.7 96.3 NKATESARI 250 1.0 0.0 0.0 NKATESARI 270 2.0 1.6 12.5 NKATESARI 280 7.3 9.6 78.8 NC 7 250 1.0 0.0 0.0 NC 7 270 2.7 2.6 21.3 NC 7 280 6.0 7.6 62.5 B106TX95 250 1.0 0.0 0.0 B106TX95 270 1.7 1.1 8.8 B106TX95 280 4.7 5.7 46.3 F MIX SINK 24 250 1.0 0.0 0.0 F MIX SINK 24 270 2.7 2.6 21.3 F MIX SINK 24 280 5.3 6.6 53.8 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 133

133 Table C 4. Data of leaf spot score and corresponding percent necrosis and percent defoliation for Gampela, 2011. ICRISAT Cultivar DOY score PDLA (%) PCLA (%) CHINESE 255 3.0 3.1 25.0 CHINESE 275 4.3 5.0 41.3 CHINESE 285 8 0 10.6 87.5 DOUMBALA 255 3.7 4.1 33.8 DOUMBALA 275 4.3 5.0 41.3 DOUMBALA 285 8.0 10.6 87.5 TS 32 1 255 3.7 4.1 33.8 TS 32 1 275 4.0 4.6 37.5 TS 32 1 285 7.7 10.2 83.8 ICGV IS 92101 255 2.7 2.6 21.3 ICGV IS 92101 275 3.0 3.1 25.0 ICGV IS 92101 285 5.0 6.1 50.0 PC 79 79 255 2.7 2.6 21.3 PC 79 79 275 3.0 3.1 25.0 PC 79 79 285 3.3 3.5 28.8 GM 515 255 2.7 2.6 21.3 GM 515 275 3.3 3.5 28.8 GM 515 285 4.7 5.7 46.3 ICGV IS 96814 255 2.7 2.6 21.3 ICGV IS 96814 275 3.0 3.1 25.0 ICGV IS 96814 285 4.7 5.7 46.3 G204TX95 255 3.0 3.1 25.0 G204TX95 275 3.0 3.1 25.0 G204TX95 285 5.0 6.1 5.0 G122TX95 255 2.0 1.6 12.5 G122TX95 275 3.0 3.1 25.0 G122TX95 285 3.7 4.1 33.8 GUSIE BALIN (92099) 255 2.7 2.6 21.3 GUSIE BALIN (92099) 275 3.3 3.5 28.8 GUSIE BALIN (92099) 285 4.0 4.6 37.5 ICGV IS 96895 255 2.7 2.6 21.3 ICGV IS 96895 275 3.7 4.1 33.8 ICGV IS 96895 285 6.0 7.6 62.5 GM 57 255 3.0 3.1 25.0 GM 57 275 3.7 4.1 33.8 GM 57 285 5.0 6.1 50.0 ICVG (FDRS) 20 F MIX 39 255 3.0 3.1 25.0 ICVG (FDRS) 20 F MIX 39 275 3.7 4.1 33.8 ICVG (FDRS) 20 F MIX 39 285 7.0 9.1 75.0 ICGV IS 92093 255 2.3 2.0 16.3 ICGV IS 92093 275 3.3 3.5 28.8 ICGV IS 92093 285 5.3 6.6 53.8 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 134

134 Table C 4 Continued ICRISAT Cultivar DOY SCORE PDLA PCLA GM 123 255 2.7 2.6 21.3 GM 123 275 3.3 3.5 28.8 GM 123 285 5.3 6.6 53.8 NKATESARI 255 2.7 2.6 21.3 NKATESARI 275 3.3 3.5 28.8 NKATESARI 285 5.3 6.6 53.8 NC 7 255 3.0 3.1 25.0 NC 7 275 3.7 4.1 33.8 NC 7 285 4.7 5.7 46.3 B106TX95 255 2.7 2.6 21.3 B106TX95 275 3.0 3.1 25.0 B106TX95 285 3.3 3.5 28.8 F MIX SINK 24 255 2.7 2.6 21.3 F MIX SINK 24 275 3.3 3.5 28.8 F MIX SINK 24 285 4.3 5.0 41.3 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 135

135 Table C 5. Data of leaf spot score and corresponding percent necrosis and percent defoliation for Nyankpala, 2010. ICRISAT Cultivar DOY score PDLA (%) PCLA (%) CHINESE 236 3.0 3.1 25.0 CHINESE 256 6.0 7.6 62.5 CHINESE 266 7.5 9.9 81.3 DOUMBALA 236 3.0 3.1 25.0 DOUMBALA 256 5.0 6.1 50.0 DOUMBALA 266 7.7 10.2 83.8 TS 32 1 236 3.0 3.1 25.0 TS 32 1 256 5.0 6.1 50.0 TS 32 1 266 7.8 10.3 85.0 ICGV IS 92101 236 2.0 1.6 12.5 ICGV IS 92101 256 4.0 4.6 37.5 ICGV IS 92101 266 5.0 6.1 50.0 ICGV IS 92101 276 5.6 7.0 57.5 PC 79 79 236 2.0 1.6 12.5 PC 79 79 256 3.0 3.1 25.0 PC 79 79 266 5.0 6.1 50.0 PC 79 79 276 4.5 5.4 43.8 GM 515 236 3.0 3.1 25.0 GM 515 256 5.0 6.1 50.0 GM 515 266 6.0 7.6 62.5 GM 515 276 6.6 8.5 70.0 ICGV IS 96814 236 2.0 1.6 12.5 ICGV IS 96814 256 4.0 4.6 37.5 ICGV IS 96814 266 5.0 6.1 50.0 ICGV IS 96814 276 4.7 5.7 46.3 G204TX95 236 2.0 1.6 12.5 G204TX95 256 4.0 4.6 37.5 G204TX95 266 5.0 6.1 50.0 G204TX95 276 6.1 7.8 63.8 G122TX95 236 2.0 1.6 12.5 G122TX95 256 4.0 4.6 37.5 G122TX95 266 5.0 6.1 50.0 G122TX95 276 3.2 3.4 27.5 GUSIE BALIN (92099) 236 2.0 1.6 12.5 GUSIE BALIN (92099) 256 4.0 4.6 37.5 GUSIE BALIN (92099) 266 5.0 6.1 50.0 GUSIE BALIN (92099) 276 5.1 6.3 51.3 ICGV IS 96895 236 2.0 1.6 12.5 ICGV IS 96895 256 4.0 4.6 37.5 ICGV IS 96895 266 5.0 6.1 50.0 ICGV IS 96895 276 4.6 5.5 45.0 GM 57 236 3.0 3.1 25.0 GM 57 256 5.0 6.1 50.0 GM 57 266 6.0 7.6 62.5 GM 57 276 7.3 9.6 78.8 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 136

136 Table C 5 Continued ICRISAT Cultivar DOY SCORE PDLA PCLA ICVG (FDRS) 20 F MIX 39 236 3.0 3.1 25.0 ICVG (FDRS) 20 F MIX 39 256 5.0 6.1 50.0 ICVG (FDRS) 20 F MIX 39 266 6.0 7.6 62.5 ICVG (FDRS) 20 F MIX 39 276 7.4 9.7 80.0 ICGV IS 92093 236 2.0 1.6 12.5 ICGV IS 92093 256 4.0 4.6 37.5 ICGV IS 92093 266 5.0 6.1 50.0 ICGV IS 92093 276 4.5 5.4 43.8 GM 123 236 3.0 3.1 25.0 GM 123 256 5.0 6.1 50.0 GM 123 266 6.0 7.6 62.5 GM 123 276 7.2 9.4 77.5 NKATESARI 236 2.0 1.6 12.5 NKATESARI 256 4.0 4.6 37.5 NKATESARI 266 5.0 6.1 50.0 NKATESARI 276 5.5 6.9 56.3 NC 7 236 2.0 1.6 12.5 NC 7 256 4.0 4.6 37.5 NC 7 266 6.0 7.6 62.5 NC 7 276 6.4 8.2 67.5 B106TX95 236 2.0 1.6 12.5 B106TX95 256 4.0 4.6 37.5 B106TX95 266 5.0 6.1 50.0 B106TX95 276 4.0 4.5 37.5 F MIX SINK 24 236 2.0 1.6 12.5 F MIX SINK 24 256 4.0 4.6 37.5 F MIX SINK 24 266 5.0 6.1 50.0 F MIX SINK 24 276 4.7 5.7 46.3 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 137

137 Table C 6. Data of leaf spot score and corresponding percent necrosis and percent defoliation for Nyankpala, 2011. ICRISAT Cultivar DOY score PDLA (%) PCLA (%) CHINESE 221 3.5 3.8 31.3 CHINESE 241 4.9 6.0 62.5 CHINESE 251 6.0 7.6 62.5 DOUMBALA 221 2.9 2.9 23.8 DOUMBALA 241 5.8 7.3 60.0 DOUMBALA 251 6.0 7.6 62.5 TS 32 1 221 2.9 2.9 23.8 TS 32 1 241 5.6 6.6 53.8 TS 32 1 261 6.0 7.6 62.5 ICGV IS 92101 221 2.9 2.9 23.8 ICGV IS 92101 241 3.3 3.5 28.8 ICGV IS 92101 251 5.0 6.1 50.0 ICGV IS 92101 261 6.0 7.6 62.5 PC 79 79 221 3.1 3.2 26.3 PC 79 79 241 3.8 4.3 35.0 PC 79 79 251 5.0 6.1 50.0 PC 79 79 261 5.1 6.3 51.3 GM 515 221 3.5 3.8 31.3 GM 515 241 4.7 5.6 46.3 GM 515 251 6.0 7.6 62.5 GM 515 261 6.0 7.6 62.5 ICGV IS 96814 221 3.2 3.4 27.5 ICGV IS 96814 241 3.7 4.1 37.5 ICGV IS 96814 251 5.0 6.1 50.0 ICGV IS 96814 261 5.0 6.1 50.0 G204TX95 221 3.2 3.4 27.5 G204TX95 241 4.3 5.1 41.3 G204TX95 251 5.0 6.1 50.0 G204TX95 261 5.0 6.1 50.0 G122TX95 221 2.6 2.5 20.0 G122TX95 241 3.2 3.4 27.5 G122TX95 251 5.0 6.1 50.0 G122TX95 261 5.0 6.1 50.0 GUSIE BALIN (92099) 221 2.3 2.0 16.3 GUSIE BALIN (92099) 241 3.3 3.5 28.8 GUSIE BALIN (92099) 251 5.0 6.1 50.0 GUSIE BALIN (92099) 261 5.0 6.1 50.0 ICGV IS 96895 221 2.7 2.6 21.3 ICGV IS 96895 241 3.7 4.1 33.8 ICGV IS 96895 251 5.0 6.1 50.0 ICGV IS 96895 261 5.0 6.1 50.0 GM 57 221 4.1 4.8 38.8 GM 57 241 4.6 5.5 45.0 GM 57 251 6.0 7.6 62.5 GM 57 261 6.0 7.6 62.5 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 138

138 Table C 6 Continued ICRISAT Cultivar DOY SCORE PDLA PCLA ICVG (FDRS) 20 F MIX 39 221 3.3 3.5 28.8 ICVG (FDRS) 20 F MIX 39 241 4.1 4.8 38.8 ICVG (FDRS) 20 F MIX 39 251 5.0 6.1 50.0 ICVG (FDRS) 20 F MIX 39 261 5.0 6.1 50.0 ICGV IS 92093 221 3.5 3.8 31.3 ICGV IS 92093 241 4.3 5.1 41.3 ICGV IS 92093 251 5.0 6.1 50.0 ICGV IS 92093 261 5.0 6.1 50.0 GM 123 221 4.2 4.9 40.0 GM 123 241 5.1 6.3 51.3 GM 123 251 6.0 7.6 62.5 GM 123 261 6.0 7.6 62.5 NKATESARI 221 3.2 3.4 27.5 NKATESARI 241 4.3 5.1 41.3 NKATESARI 251 5.0 6.1 50.0 NKATESARI 261 5.0 6.1 50.0 NC 7 221 2.9 2.9 23.8 NC 7 241 3.9 4.4 36.3 NC 7 251 5.0 6.1 50.0 NC 7 261 5.0 6.1 50.0 B106TX95 221 3.2 3.4 27.5 B106TX95 241 4.2 4.9 40.0 B106TX95 251 5.0 6.1 50.0 B106TX95 261 5.0 6.1 50.0 F MIX SINK 24 221 3.6 4.0 32.5 F MIX SINK 24 241 4.0 4.6 37.5 F MIX SINK 24 251 5.0 6.1 50.0 F MIX SINK 24 261 5.0 6.1 50.0 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 139

139 Table C 7. Data of leaf spot score and corresponding percent necrosis and percent defoliation for Wa, 2010. ICRISAT Cultivar DOY score PDLA (%) PCLA (%) CHINESE 247 3.3 3.5 28.8 CHINESE 267 6.0 7.6 62.5 CHINESE 277 7.0 9.1 75.0 DOUMBALA 247 3.0 3.1 25.0 DOUMBALA 267 5.3 6.6 53.8 DOUMBALA 277 7.0 9.1 75.0 TS 32 1 247 3.0 3.1 25.0 TS 32 1 267 5.3 6.6 53.8 TS 32 1 277 7.0 9.1 75.0 ICGV IS 92101 247 2.0 1.6 12.5 ICGV IS 92101 267 3.3 3.5 28.8 ICGV IS 92101 287 5.0 6.1 50.0 ICGV IS 92101 297 6.0 7.6 62.5 PC 79 79 247 2.0 1.6 12.5 PC 79 79 267 3.3 3.5 28.8 PC 79 79 287 5.0 6.1 50.0 PC 79 79 297 6.0 7.6 62.5 GM 515 247 3.0 3.1 25.0 GM 515 267 5.0 6.1 50.0 GM 515 287 6.0 7.6 62.5 GM 515 297 7.0 9.1 75.0 ICGV IS 96814 247 2.0 1.6 12.5 ICGV IS 96814 267 4.6 5.5 45.0 ICGV IS 96814 287 5.0 6.1 50.0 ICGV IS 96814 297 6.0 7.6 62.5 G204TX95 247 2.3 2.0 16.3 G204TX95 267 4.7 5.7 46.3 G204TX95 287 5.0 6.1 50.0 G204TX95 297 6.0 7.6 62.5 G122TX95 247 2.0 1.6 12.5 G122TX95 267 4.7 5.7 46.3 G122TX95 287 5.0 6.1 50.0 G122TX95 297 6.0 7.6 62.5 GUSIE BALIN (92099) 247 2.0 1.6 12.5 GUSIE BALIN (92099) 267 4.7 5.7 46.3 GUSIE BALIN (92099) 287 5.0 6.1 50.0 GUSIE BALIN (92099) 297 6.0 7.6 62.5 ICGV IS 96895 247 3.0 3.1 25.0 ICGV IS 96895 267 4.7 5.7 46.3 ICGV IS 96895 287 5.0 6.1 50.0 ICGV IS 96895 297 6.0 7.6 62.5 GM 57 247 3.3 3.5 28.8 GM 57 267 5.0 6.1 50.0 GM 57 287 6.0 7.6 62.5 GM 57 297 7.0 9.1 75.0 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 140

140 Table C 7 Continued ICRISAT Cultivar DOY SCORE PDLA PCLA ICVG (FDRS) 20 F MIX 39 247 3.0 3.1 25.0 ICVG (FDRS) 20 F MIX 39 267 4.7 5.7 46.3 ICVG (FDRS) 20 F MIX 39 287 5.3 6.6 53.8 ICVG (FDRS) 20 F MIX 39 297 6.0 7.6 62.5 ICGV IS 92093 247 2.0 1.6 12.5 ICGV IS 92093 267 4.3 5.0 41.3 ICGV IS 92093 287 5.0 6.1 50.0 ICGV IS 92093 297 6.0 7.6 62.5 GM 123 247 3.6 4.0 32.5 GM 123 267 5.0 6.1 50.0 GM 123 287 6.0 7.6 62.5 GM 123 297 7.0 9.1 75.0 NKATESARI 247 2.6 2.5 20.0 NKATESARI 267 4.7 5.7 46.3 NKATESARI 287 5.0 6.1 50.0 NKATESARI 297 6.0 7.6 62.5 NC 7 247 2.0 1.6 12.5 NC 7 267 4.7 5.7 46.3 NC 7 287 5.0 6.1 50.0 NC 7 297 6.0 7.6 62.5 B106TX95 247 2.0 1.6 12.5 B106TX95 267 4.7 5.7 46.3 B106TX95 287 5.3 6.6 53.8 B106TX95 297 6.0 7.6 62.5 F MIX SINK 24 247 2.0 1.6 12.5 F MIX SINK 24 267 4.7 5.7 46.3 F MIX SINK 24 287 5.0 6.1 50.0 F MIX SINK 24 297 6.0 7.6 62.5 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 141

141 Table C 8. Data of leaf spot score and corresponding percent necrosis and percent defoliation for Wa, 2011. ICRISAT Cultivar DOY score PDLA (%) PCLA (%) CHINESE 262 4.0 4.6 37.5 CHINESE 282 6.0 7.6 62.5 CHINESE 292 8.0 10.6 87.5 DOUMBALA 262 3.7 4.1 33.8 DOUMBALA 282 5.7 7.2 58.8 DOUMBALA 292 8.0 10.6 87.5 TS 32 1 262 4.0 4.6 37.5 TS 32 1 282 5.3 6.6 53.8 TS 32 1 292 8.0 10.6 87.5 ICGV IS 92101 262 2.0 1.6 12.5 ICGV IS 92101 282 5.0 6.1 50.0 ICGV IS 92101 292 6.0 7.6 62.5 ICGV IS 92101 302 7.0 9.1 75.0 PC 79 79 262 2.0 1.6 12.5 PC 79 79 282 5.0 6.1 50.0 PC 79 79 292 5.7 7.2 58.8 PC 79 79 302 7.0 9.1 75.0 GM 515 262 3.7 4.1 33.8 GM 515 282 5.7 7.2 58.8 GM 515 292 7.0 9.1 75.0 GM 515 302 8.0 10.6 87.5 ICGV IS 96814 262 2.0 1.6 12.5 ICGV IS 96814 282 5.0 6.1 50.0 ICGV IS 96814 292 6.0 7.6 62.5 ICGV IS 96814 302 7.0 9.1 75.0 G204TX95 262 3.0 3.1 25.0 G204TX95 282 5.0 6.1 50.0 G204TX95 292 6.0 7.6 62.5 G204TX95 302 7.0 9.1 75.0 G122TX95 262 2.0 1.6 12.5 G122TX95 282 5.0 6.1 50.0 G122TX95 292 5.7 7.2 58.8 G122TX95 302 7.0 9.1 75.0 GUSIE BALIN (92099) 262 2.0 1.6 12.5 GUSIE BALIN (92099) 282 5.0 6.1 50.0 GUSIE BALIN (92099) 292 6.0 7.6 62.5 GUSIE BALIN (92099) 302 7.0 9.1 75.0 ICGV IS 96895 262 3.0 3.1 25.0 ICGV IS 96895 282 5.0 6.1 50.0 ICGV IS 96895 292 5.7 7.2 58.8 ICGV IS 96895 302 7.0 9.1 75.0 GM 57 262 4.0 4.6 37.5 GM 57 282 5.7 7.2 58.8 GM 57 292 7.0 9.1 75.0 GM 57 302 8.0 10.6 87.5 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 142

142 Table C 8 Continued ICRISAT Cultivar DOY SCORE PDLA PCLA ICVG (FDRS) 20 F MIX 39 262 3.0 3.1 25.0 ICVG (FDRS) 20 F MIX 39 282 5.0 6.1 50.0 ICVG (FDRS) 20 F MIX 39 292 6.0 7.6 62.5 ICVG (FDRS) 20 F MIX 39 302 7.3 9.6 78.8 ICGV IS 92093 262 2.0 1.6 12.5 ICGV IS 92093 282 5.0 6.1 50.0 ICGV IS 92093 292 6.0 7.6 62.5 ICGV IS 92093 302 7.0 9.1 75.0 GM 123 262 3.7 4.1 33.8 GM 123 282 5.7 7.2 58.8 GM 123 292 7.0 9.1 75.0 GM 123 302 7.7 10.2 83.8 NKATESARI 262 3.0 3.1 25.0 NKATESARI 282 5.0 6.1 50.0 NKATESARI 292 6.0 7.6 62.5 NKATESARI 302 7.0 9.1 75.0 NC 7 262 2.0 1.6 12.5 NC 7 282 5.0 6.2 50.0 NC 7 292 6.0 7.6 62.5 NC 7 302 7.0 9.1 75.0 B106TX95 262 2.0 1.6 12.5 B106TX95 282 4.6 5.5 45.0 B106TX95 292 5.7 7.2 58.8 B106TX95 302 7.0 9.1 75 F MIX SINK 24 262 2.0 1.6 12.5 F MIX SINK 24 282 5.0 6.1 50.0 F MIX SINK 24 292 6.0 7.6 62.5 F MIX SINK 24 302 7.0 9.1 75.0 DOY, day of year PDLA, percent necrosis PCLA, percent defoliation

PAGE 143

143 APPENDIX D SOIL FERTILITY FACTOR OF SOILS Table D 1. Soil fertility factor (SLPF) of the soils at the different locations used in the model. Locaion SLPF WA 0.69 NYANKPALA 0.63 FARAKOBA 0.70 GAMPELA 0.9 9

PAGE 144

144 LIST OF REFERENCES Adomou, F.M., and A. Traore. 2000. Rational use of fungicide applications to maximize peanut yield under foliar disease pressure in West Africa. Plant Disease 84:1203 1211. Akram M., S.S. Ali, and F.M. Abassi. 1999. Genot ype environment interaction and stability parameters for paddy yield. Pak. J. Sci. and Industr. Res 42(5): 279 281. Allen, L.H. Jr., and K.J. Boote. 2000. Crop ecosystem responses to climatic change, Soybean. In: K.R. Reddy and H.F. Hodges (eds) Climate change and Global Crop Productivity Chapter 7, pp. 133 160. CAB International, New York, NY. Allen, L.H. Jr., J.C.V. Vu, and R.R. Valle. 1988. Nonstructural carbohydrates and nitrogen of soybean grown under carbon dioxide enrichment. Crop Science 28: 84 94. Annicchiarico, P. 2002a. Genotype by environme nt interactions : Challenges and opportunities for plant breeding and cult ivar recommendations. FAO plant production and protection paper No. 174. FAO, Rome. Annicchiarico, P. 2002b. Defining adaptation strategies and yieldstability targets in breeding programs. In Kang M.S. (Ed.), Quantita tive Genetics Genomics and Plant Breeding. CABI Publishing, Wallingford, UK, pp: 365383. Annicchiarico, P. 1997a. Joint regression vs AMMI analysis of genotypeenvironment interactions for cereals in Italy. Euphytica 94: 5362. Anni cchiarico, P. 1992. Cultivar adaptation and recomm endation from alfalfa trials in northern Italy. J. Genet. Breed 46: 269278. Anothai, J., A. Patanothai, K. Pannangpetch, S. Jogloy, K. J. Boote, and G. Hoogenboom. 2009. Multi environment eval uation of peanut lines by model simulation with the cultivar coefficients derived from a reduced set of observed field data. Field Crops Res. 110:111 121. Araus, J.L., G.A. Slafer, M.P. Reynolds, and C. Royo. 2002. Plant breeding and drought in C3 cereals: What should we breed for? Ann. Bot. (London) 89:925 940. doi:10.1093/aob/mcf049 Babu, V.R., and D.V.M. Rao. 1983. Water stress adaptations in groundnut ( Arachis hypogaea L.) foliar characteristics and adaptations to moisture stress. Plant Physiol. Biochem. 10 (1): 64 80. Bader, J., and M. Latif. 2003: The impact of decadal scale Indian Ocean sea surface temperature anomalies on Sahelian rainfall and the North Atlantic Oscillation. Geophys. Res. Lett 30: 2169, doi:10.1029/2003GL018426.

PAGE 145

145 Banterng, P., A. Patanothai, K. Pannangpetch, S. Jog loy, and G. Hoogenboom. 2006. Yield stability evaluation of peanut lines: a comparison of an experimental versus a simulation approach. Field Crops Res. 96:168 175. Banterng P ., A. Patanothai K. Pannangpetch, S. Jogloy, and G. Hoogenboom. 2004.Determination of genetic coefficients for peanut lines for breeding applications. Eur. J. Agron. 21:297 310. Barah, B.C., H.P. Binswanger, B.S. Rana, and G.P. Rao. 1981. The use of risk aversion in plant breeding; concept and application. Euphytica 30:451458. Batchelor, W.D., J.W. Jones, K.J. Boote, and H.O. Pinnschmidt. 1993. Extending the use of crop models to study pest damage. Trans. ASAE 36:551558 Becker, H.C. 1984. Analysis of genotype x environment i nteraction with partitioning of environmental effects into effects of locations and years. Vortr. Pflanzenzchtg ., 7: 209218. Becker, H.C. 1981. Correlations among some statistical measures of phenotypic stability. Euphytica 30:83584 0. Becker, H.C., and J. Lon. 1988. Stability analysis in plant breeding. Plant Breed. 101: 1 23. Bell, M.J., G.C. Wright, and G. Harch. 1993. Environ mental and agronomic effects on the growth of four peanut cultivars in a subtropi cal environment. II. Dry matter partitioning. Exp. Agric. 29 (4):491 501. Bennett, J.M., P.J. Sexton, and K.J. Boote. 1990. A r oot tubepegging pan apparatus: preliminary observations and effects of soil water in the pegging zone. Peanut Sci. 17:68 72. Blum, A. 1988. Plant breeding for stress environments. CRC Press, Boca Raton, FL. Boote, K.J. 1999. Concepts for calibrating crop growth models. In Hoogenboom, G., Wilkens, P.W., Tsuji, G.Y. (Eds.), DSSAT Vers ion 3, 4. University of Hawaii, Honolulu, Hawai i, pp. 179 200. Boote, K.J., J.W. Jones, G. Hoogenboom, and N.B. Pickering. 1998. The CROPGRO model for grain legumes. p. 99128. In G.Y. Tsugi, G. Hoogenboom, and P.K. Thornthon (ed). Understanding options for agricultural production. Kluwer Academic Publ ., Dordrecht, the Netherlands. Boote, K.J., and D.L. Ketring. 1990. Peanut. In: Stewart B.A. and Niel son D.R. (eds), Irrigation of Agricultural Crops. ASA CSSA SSSA, Madison. Boote, K.J., and J.W. Jones. 1986. Applications of, and limitation to, crop growt h simulation models to fit crops and cropping systems to semi arid environments.

PAGE 146

146 p. 63 75. In F.R. Bidinger and C. Johansen (eds). Drought research priorities for the dryland tropics, pp. 6375. International Crops Research Institute for the Semi Arid Trop ics, Patancheru, A.P. 502324, India. Boote, K. J., L. H. Allen, Jr., P. V. Vara Prasad, and J. W. Jones. 2010. Testing effects of climate change in crop models. In: D. H illel and C. Rosenzweig (eds.), Handbook of Climate Change and Agroecos ystems. Imperial College Press, London UK. pp. 109129. Boote, K.J., L.H. Allen, P.V.V. Prasad, J.T. Baker, R .W. Gesch, A.M. Snyder, D. Pan, and J.M.G. Thomas. 2005. Elevated temperature and CO2 impacts on pollination, reproductive growth, and yield of several globally im portant crops. J. Agric. Meteorol 60:469 474. Boote, K.J., J.W. Jones, W.D. Batchelor, E.D. Nafzig er, and O. Myers. 2003. Genetic coefficients in the CROPGRO Soybean model: Links to field performance and genomics. Agron. J. 95:3251. Boote, K.J., M.J. Krop ff, and P.S. Bindraban. 2001. Physiology and modelling of traits in crop plants: Implications for genetic improvement. Agric. Syst. 70:395 420. Boote, K.J., J.W. Jones, and G. Hoogenboom. 1998a. Simulation of crop growth: CROPGRO model. p. 651692. In R.M. Peart and R.B. Curry (ed.) Agricultural Systems Modeling and Simulation. Marcel Dekker, New York. Boote, K.J., J.W. Jones, G. Hoogenboom, and N.B Pickering. 1998b. The CROPGRO model for grain legumes. p. 99128. In G.Y. T suji et al. (ed.) Understanding O ptions for Agricultural Production. Kluwer Academic Publishers, Dordrecht The Netherlands. Boote, K.J., J.W. Jones, and N.B. Pickering. 1996. Potential uses and limitations of crop models. Agron. J. 88:704 716. Boote, K.J., W.D. Batchelor, J.W. Jones, H. Pinnschmidt, and G. Bourgeois. 1993. Pest damage relations at the field level. p. 277296. In F.W.T. Penning de Vries et al. (ed.) Systems approaches for agricultural development. Kluwer Acad. Publ., Dordrecht, the Netherlands. Boote, K.J., J.W. Jones, and P. Singh. 1992. M odeling growth and yield of groundnut. In: Nigam S.N. (ed.), Groundnut A P erspective. International Crops Research Institute for the Semi Arid Tropics, Patancheru, India, pp. 331343. Boote, K. J., J.W. Jones, J.W. Mishoe, and R.D Ber ge r. 1983a. Coupling pests to crop growth stimulators to predict yield reductions. Phytopathology 73:15811587. Bourgeois, G., and K. J. Boote. 1992. Leaflet and canopy photosynthesis of peanut affected by late leaf spot. Agron. J. 84:359366.

PAGE 147

147 Borrell, A .K., G. Hammer, and E. van Oosterom. 200 1. Stay green: A consequence of the balance between supply and demand for nitr ogen during grain filling. Ann. Appl. Biol. 138:91 95. Bidinger, F.R., and C. Johansen (ed.). 1988. Drought research priorities for the drylan d tropics. Int. Crops Research Inst. for the Semi Arid Tropics, Patancheru, India. Bramel Cox, P.J. 1996. Breeding for reliability of p erformance across unpredictable environments. In M.S. Kang and H.G. Gauch, eds. Genotypeby environment interaction. p. 309 339. Boca Raton, FL, CRC Press. Campbell, W.J., L.H. Allen, Jr, and G. Bowes. 1988. Effects of CO2 concentration on rubisco activity, amount, and photosynthesis in soybean leaves. Plant Physiology 88:1310 1316. Ceccarelli, S. 1996. Positive interpr etation of genotype by environment interaction in relation to sustainability and biodiversity. In M. Cooper and G.L. Hammer, eds. Plant adaptation and crop improvement p. 467 486. Wallingford, UK, CABI. Ceccarelli, S. 1989. Wide adaptation: How wide? Euphytica 40:197205. Cecilia, M., T. Soler, A. Suleiman, J. Anothai, I. Flitc roft, and G. Hoogenboom. 2012. Scheduling irrigation with a dynamic crop g rowth model and determining the relation between simulated drought stress an d yield for peanut. Irrig. Sci. Springer Verlag 113. Chandler, R. F., Jr. 1969. Plant morphology and stand geometry in relation to nitrogen. Chapter 12. Pp. 265285. IN J. D. Eastin, F. A. Haskins, C. Y Sullivan, and C.H.M. Van Bavel. Physiological aspects of crop yield. ASA CSSASSSA, Madison, WI. Chen, J.J., and J.M. Sung. 1990. Gas exchange rate and yield responses of Virginia type peanut to carbon dioxide enrichment. Crop Science 30:1085 1089. Chin Choy, E.W., J.F. Stone, R.S. Stone, R.S. Matlock, and G.N. McCauley. 19 82. Plan t population and irrigation effects on Sp anish peanuts (Arachis hypogaea L.). Peanut Science 9:73 76. Clifford, S.C., I.M. Stronach, and C.R. Black. 2000. Effects of elevated CO2, drought and temperature on the water relations and gas exchange of groundnut (Arachis hypogaea L.) stands grown in controlled environment glasshouses. Physiologia Plantarum 110:7888. Clifford, S.C., I.M. Stronach, and A.D. Mohamed. 1993. The effects of elevated atmospheric carbon dioxide and water stress on light int erception, dr y matter production and yield in stands of groundnut (A rachis hypogaea L.). Journal of Experimental Botany 44:17631770.

PAGE 148

148 C ollins, J.M. 2011: Temperature v ariability over Africa. J. Climate 24 :3649 3666. Cooper, M. 1999. Concepts and strategies for plant ada ptation research in rainfed lowland rice. Field Crops Res 64(2):1334. Cooper, M., and D.E. Byth. 1996. Understanding plant adaptation to achieve systematic applied crop improvement A fundamental challenge. In M. Cooper & G.L.Hammer, eds. Plant adaptati on and crop improvement p. 5 23. Wallingford, UK, CABI. Cooper, M., and I.H. DeLacy. 1994. Relationships a mong analytical methods used to study genotypic variation and genotypeby environment interaction in plantbreeding multi environment experiments. Theor. Appl. Genet 88:561 572. Cox, F.R. 1979. Effect of temperature treatment on pean ut vegetative and fruit growth. Peanut Science 6: 14 17. Craufurd, P.Q., P.V.V. Prasad, and V.G. Kakani. 2003. Heat tolerance in groundnut. Field Crops Research 80: 63 77. Craufurd, P.Q., P.V.V. Prasad, and R.J. Summerfield. 2002. Dry matter production and harvest index at high temperature in peanut. Crop Science 42:146 151. Craufurd, P.Q., M. Bojang, T.R. Wheeler, and R.J. Su mmerfield. 1998. Heat tolerance in cowpea: Effec t of timing and duration of heat s tress. Ann. Appl. Biol.133:257267. Dangthaisong, P., P. Banterng, S. Jogloy, N. Vor asoot, A. Patanothai, and G. Hoogenboom. 2006. Evaluation of the CSM CROPGRP Peanut model in simulating responses of two peanut cultivars to different moisture regimes. Asian J. Pl. Sci. 6:913922 DeLacy, I.H., K.E. Basford, M. Cooper, J.K. Bull, and C.G. McLaren. 1996a. Analysis of multienvironment data An historical perspective. In M. Cooper & G.L. Hammer, eds. Plant adaptation and c rop improvement p. 39 124. Wallingford, UK, CABI. DeLacy, I.H., R.L. Eisemann, and M. Cooper. 1990. The importance of genotypeby environment interaction in regional variety trials. In M.S. Kang, ed. Genotypeby environment interaction and plant breeding p. 287 300. Baton Rouge, LA, Louisiana State Univ. Domitruk, D.R., B.L. Duggah, and D.B. Fowler. 2001. Geno type environment interaction of no till winter wheat in Western Canada. Canadian J. Pl. Sci 81:716. Donald, C. M. 1968. The breeding of crop ideot ypes. Euphytica 17:193211. Dreyer, J., W.G. Duncan, and D.E. McCloud. 1981. F ruit temperature, growth rates, and yield of peanuts. Crop Science 21: 68 6 688.

PAGE 149

149 Dwivedi, S.L., E. Perotti, and R. Ortiz. 2008b. Towards molecular breeding of reproductive traits in cereal crops. Plant Biotech. J. 6(6):529559. Dwivedi, S.L., J.H. Crouch, D.J. Mackill, Y. Xu, M.W. B lair, M. Ragot, H.D. Upadhyaya, and R. Ortiz. 2007a. The molecularization of public sector crop breeding: progress, problems, and prospects. Adv. Agron. 95: 163 318. Duncan, W. G., D. E. McCloud, R. L. McGraw, and K. J. Boote. 1978 Physiological aspects of peanut yield improvement. Crop Sci. 6: 10151020. Eberhart, S.A., and W.A. Russell. 1966. Stability parameters for comparing varieties. Crop Sci. 6:36 40. Elwell, D.L., R.B. Curry, and M.E. Keener. 1987. Determinat ion of potential yield limiting factors of soybeans using SOYMOD/OARDC. Agric. Syst. 24:221 242. Emechebe, A. M. 1980. A check list of diseases of cro ps in the savanna and semi arid areas of Nigeria. Samaru Miscellaneous Paper no. 100. Institute of Agricultural Research, Ahmadu Bello University, Samaru, Zaria, Nigeria, pp. 36. Eskridge, K.M. 1990. Selection of stable cultivars using a safety first rule. Crop Sci 30: 369374. Evans, L.T. 1 998. Feeding the ten billion: plants and population growth. New York, Cambridge Univ. Press. Evans, L.T. 1993. Crop evolution, adaptation, and yield. New York, Cambridge Univ. Press. FAO. 2011. FAOSTAT. Available at http://faostat.fao.org/default.aspx (verified 6 June 2011) FAO, Rome, Italy. FAO. 2005. Crop Production Statistics. Rome, Italy:Food and Agricultural Organization of United Nations http://faostat.fao.org (verified 9 April 2008). FAO 1996. Food requirements and population growth. Technical Background Document No. 4. Rome. Folland, C. K., J. A. Owen, M. N. Ward, and A. W. Colman, 1991: Predi ction of seasonal rainfall in the Sahel region using empi rical and dynamical methods. J. Forecasting 10: 215 6. Farquhar, G.D., and R.A. Richards. 1984. Isotopic composition of plant carbon correlates with water use efficiency of wheat genotypes. Aust. J. Plant Physiol. 11:539 552. Fokar, M., A. Blum, and H.T. Nguyen. 1998. Heat toler ance in spring wheat. II. Grain filling. Euphytica 104:9 15.

PAGE 150

150 Fontes, J., and S. Guinko. 1995. Vegetation and l and use map of Burkina Faso. An explanatory note. Ministere de la cooperation francaise, Toulouse. Finlay, K.W., and G.N. Wilkinson. 1963. The analysis of adaptation in a pla nt breeding programme. Aust. J. Agric. Res 14: 742 754. Fukai, S., G. Pantuwanb, B. Jongdeec, and M. Cooper. 1999. Screening for drought resistance in rainfed lowland rice. Field Crops Res. 64:61 74. doi:10.1016/S03784290(99)000519 Ghaderi, A., E.H. Ev erson, and C.E. Cress. 1980. Classification of environm ents and genotypes in wheat. Crop Sci 20: 707 710. Giannini, A., R. Saravanan, and P. Chang. 2005: Dynamics of the boreal summer African monsoon in the NSIPPI atmospheric model. Clim. Dyn. 25: 517 535. Gillier, P. 1982. Influences on groundnut production in West Africa. In Proceedings of the International Symposium in Africa on Produc tion, World Oilseeds Market andIntra African Trade in Groundnuts and Products, 7 11 June, Banjul, Gambia, pp. 111 129. Golo mbek, S.D., and C. Johansen. 1997. Effect of soil t emperature on vegetative and reproductive growth and development in three Spanish genotypes of peanut (Arachis hypogaea L.). Peanut Science 24: 67 72. Gomez, K.A., and A.A. Gomez. 1984. Statistical Procedur es for Agricultural Research. John Wiley and Sons, New York. Gowda, A., and B.R. Hegde. 1986. Moisture stres s and hormonal influence on the flowering behavior and yield of groundnut ( Arachis hypogaea L.). Madras Agric. J. 73(2): 82 86. Graef, F., and J. Haigis. 2001. Spatial and temporal rainfall variability in the Sahel and its effects on farmers management strategies Journal of Arid Environments, 48:221213. Hammer, G. L., and D. R. Jordan. 2007. An integrated systems approach to crop improvement. pp. 4561. In : J. H. J. Spiertz, P C. Struik, and H. H. van Laar (eds.) Scale and complexity in plant systems research: geneplant crop relations Springer. Honarnejad, R. 2003. Stability and productivity of Iranian rice cultivars. Pl. Breed. Abst .,12382( 12): 73. Hoogenboom, G., J.W. Jones, P.W. Wilkens, C.H. Porter, W.D. Batchelor, L.A. Hunt, K.J. Boote, U. Singh, O. Uryasev, W.T. Bowen, A.J. Gijsman, A. du Toit, J.W. White, and G.Y. Tsuji. 2004. Decision Su pport System for Agrotechnology Transfer Version 4.0 [CD ROM] Universi ty of Hawaii, Honolulu, Hawaii. Hoogenboom, G., P.W. Wilkens, and G.Y. Tsuji. (Eds.), 1999. DSSAT Version 3, vol. 4. IBSNAT Project. University of Hawaii, Honolulu, Hawaii.

PAGE 151

151 Houghton, J.T., Y. Ding, and D.J. Griggs (eds). 2001. Climate Chan ge 2001: The Scientific Basis (Contribution of Working Group to the Third Assessment Report to the Intergovernmental Panel on Climate Change (IPCC). Cambridge University Press, UK. Hubick, K., and G.D. Farquhar. 1989. Carbon isotope discrimination and the ratio of carbon gained to water lost in barley cultivars. Plant Cell Environ. 12:795 804. Hubick, K.T., G.D. Farquhar, and R. Shorter. 1986. Correlation between water use efficiency and carbon isotope discriminat ion in diverse peanut (Arachis) germplasm. A ust. J. Plant Physiol. 13:803 816. Hudak, C.M., and R.P. Patterson. 1996. Root distribution and soil moisture depletion of a drought resistant soybean plant introduction. Agron. J. 88:478 485. Hulme, M., R. Doherty, T. Ngara, M. New, and D. Lister 2001. A frican climate change: 1900 2100. Clim. Res. 17: 145 168. Hunt, L.A., and K.J. Boote. 1998. Data for model operati on, calibration and evaluation. In Tsuji, G.Y., Hoogenboom, G., Thornton, P.K (Eds.), Understanding Options for Agricultural Production. Kluwer Acad emic Publishers, Dordrecht, The Netherlands, pp. 9 39. IPCC, 2007. IPCC (Intergovernmental Panel on Climate Change), 2007. Climate Change 2007: Impacts, Adaptation and Vulnerability. Summary for policy makers. Online at http://www.ipcc.cg/SPM13apr07.pdf Ismail, A.M., and A.E. Hall. 1992. Correlation between water use efficie ncy and carbon isotope discrimination in diverse cowpea genotypes and isogenic lines. Crop Sci. 32:7 12. Jackson, P.A., M. Robertson, M. Cooper, and G. Hammer. 1996. The role of physiological understanding in plant breeding: From a breeding perspective. Field Crops Res. 49:11 37. doi:10.1016/S03784290(96)01012X James, E. P. Jr. 1982. Photosynthetic traits of selected peanut genotypes. Peanut Science, Vol. 9, No. 1, pp. 1417. Janamatti, V.S., V.R. Sashidhar, I.G. Prasad, and K.S.K. Sastry. 1986. Effect of cycles of moisture stress on flowering pattern, flower production, gynophore length andtheir relationship to pod yield in bunch types of groundnut. Narendra Deva J. Agric. Res. 1 (2):136 142. Jones, J.W., G. Hoogenboom, C.H. Porter, K.J. Boote, W.D. Batchelor, L.A. Hunt, P.W. Wilkens, U. Singh, A.J. Gijsman, and J.T. Ri tchie. 2003. The DSSAT cropping system model. Eur. J. Agron. 18: 235 265. Jones, P.G., and P.K. Thornton. 2003. The potential impacts of climate change in tropical agriculture: the case of maize in Africa and Latin America in 2055, Global

PAGE 152

152 Environmental Change 13 (2003), pp. 51 59 Article | PDF (1636 K) | View Record in Scopus | Cited By in Scopus (48). Jordan, D.R., Y. Tao, I.D. Godwin, R.G. Henzell, M. C ooper, and C.L. McIntyre. 2003. Prediction of hybrid performance in grain sor ghum using RFLP markers. Theor. Appl. Genet. 106:559 567. Jordon, W.R., and F.R. M iller. 1980. Genetic variability in sor ghum root systems: Implications for drought tolerance. p. 383 399. In N.C. Turner and P. J. Kramer (ed.) Adaptation of plants to water and high tem perature stress. John Wiley and Sons, New York. Kakani, V.G., P.V.V. Pr asad, P.Q. Craufurd, and T.R Wheeler. 2002. Response of in vitro pollen germination and pollen tube growth of groundnut ( Arachis hypogaea L.) genotypes to temperature. Plant Cell Environ. 25:1651 1661. Kang, M.S. 1998. Using genotypeby environmen t interaction for crop cultivar development. Adv. Agron 62: 199252. Kang, M.S. 1993. Simultaneous selection for yield and stability i n crop performance trials: consequences for growers. Agron. J 85: 754757. Kang, M.S. 1990. Genotypeby environment interaction and plant breeding Baton Rouge, LA, Louisiana State Univ. Kang, M.S., and H.N. Pham. 1991. Simultaneous selection for high yiel ding and stable crop genotypes. Agron. J 83: 161 165. Kaufmann, R.K., and D.I. Stern. 1997. Evidence of human influence on climate from hemispheric tem perature relations. Nature, 388: 39 44. Ketring, D.L. 1984. Temperature effects on vegetative and reproductive development of peanut. Crop Science, 24: 877 882. Khan, E. F., and S. M. Misari. 1987. Effect of soil treatment with insecti cides and nematicides on the seasonal development of plant parasitic nematodes on groundnuts. Legumes and Oilseeds Research Pr ogram, Cropping Scheme Meeting 1987, Institute of Agricultural Res earch, Ahmadu Bello University, Samaru, Zaria, Nigeria. Lawn, R. J. 1989. Agronomic and physiological constraints to productivity in the tropical grain legumes and opportunities for improvement. Exp. Agric. 25:509 528. doi:10.1017/ S0014479700015143 Lin, C.S. and G. Butler. 1988. A databased approach for selecting loca tions for regional trials. Can. J. Plant Sci 68: 651 659.

PAGE 153

153 Lin, C.S., and M.R. Binns. 1988. A method for analy zing cultivar x location x year experiments: a new stability parameter. Theor. Appl. Genet 76: 425 430. Lin, C.S., M.R. Binns, and L.P. Lefkovitch. 1986. Stabili ty analysis: where do we stand? Crop Sci 26: 894 900. Lo ffler, C. M., J. Wei, T. Fast, J. Gogerty, S. Langton, M. Bergman, B. Merrill, and M. Cooper. 2005. Classification of maize environments using crop simulation and geographic information s ystem. Crop Sci. 45:17081716. Ludlow, M.M., and R.C. Muchow. 1990. A critical evaluation of traits for improving crop yields in water limited environments. Adv. Agron. 43:107 153. doi:10.1016/S00652113(08)604770 Maninder, P.S., J.E. Erickson, K.J. Boot e, B.L. Tillman, J.W. Jones, and A.H.C. van Bruggen. 2011. Late leafspot effects on growth, photosynthesis, and yield in peanut cultivars of differing resistance. Agron. J. 103:85 91. Marshall, D.R. 1991. Alternative approaches and perspectives in breeding for higher yield. Field Crops Res. 26:171 190. doi:10.1016/03784290(91)90034S Masle, J., S.R. Gilmore, and G.D. Farquhar. 2005. The erecta gene regulates plant transpiration efficiency in Arabidopsis. Nature 436:866 870. McCown, R.L., B.A. Keati ng, M.E. Probert, and R.K. Jones. 1992. Strategies for sustainable crop production in semi arid Africa. Outlook on Agric 21: 2131. McD onald, D., P. Subrahmanyam, R. W. Gibbons, and D .H. Smith. 1985. Early and late leaf spots of groundnut. Information B ulletin no. 21, ICRISAT Center, Patancheru, AP, India. Meisner, C.A., and K.J. Karnok. 1992. Peanut root response to drought stress. Agron. J.84(2): 159 165. Metelerkamp, H.R.R. 1988. Review of crop research relevant to the semiarid areas of Zimbabwe. In Cropp ing in the Semiarid Areas of Zimbabwe Proceedings of a workshop. Harare, Zimbabwe: Agritex, Department of Research and Specialist Services (DR&SS), and Gemeinschaft fr T echnische Zusammenarbeit (GTZ). Pp.190315. Merah, O., E. Deleens, I. Souyris, M. Nac hit, and P Monneveux. 2001. Stability of carbon isotope discrimination and grain yield i n durum wheat. Crop Sci. 41:677681. Monneveux, P., C. Sanchez, and A. Tiessen. 2008. Future progress in drought tolerance in maize needs new secondary traits and cros s combinations. J. Agric. Sci. 146:287 300. doi:10.1017/ S0021859608007818

PAGE 154

154 Mortley, D.G., C.K. Bonsi, W.A. Hill, and C.E. Morri s. 2004. Temperature influences yield, reproductive growth, harvest index and oil content of hydroponically grown Georgia Red p eanut plants. HortScience 39(5):975978. Mortley, D.G., P.A. Loretan, and J.H. Hill. 1997. CO2 enrichment influences yield of Florunner, Georgia Red and New Mexico peanut cultivars. Advances in Space Research 20:19051908. Naab, J.B., F.K. Tsigbey, P.V.V. Prasad, K.J. Boote, J. E. Bailey, and R.L. Brandenburg. 2005. Effects of sowing date and fungicide application of yield of early and late maturing peanut cultivars grown under rainfed conditions in Ghana. Crop Protection 24: 325 332. Naab, J.B., P. Singh, K.J. Boote, J.W. Jones, and K.O. Marfo. 2004. Using the CROPGRO peanut model to quantify yield gaps of peanut in the Guinean Savanna zone of Ghana. Agron. J. 96:12311242. Nageswara Rao, R.C., G.C. Wright, and A .L. Cruickshank. 2000a. Genetic enhancement o f drought resistance in Australian peanuts. P roc. Am. Peanut Res. Educ. Soc. 32:71. Nageswara Rao, R. C., G.C. Wright, A. W. Cruickshank, M. S. Basu, and S. N. Nigam. 2000b. Genetic enhancement of drought resis tance in peanuts. Proc. Intern. Crop Sci. Soc. Euro pean Soc. Agronomy, Germany. p. 71. (abstr.). Nicholson, S. E. 2001. Climatic and environmental chang e in Africa during the last two centuries. Clim. Res. 17 : 123 144. Nicholson, S. E., and J. P. Grist. 2001: A conceptual m odel for understanding rainfall variability in the West Africa Sahel on interannual and interdecadal times cales. Int. J. Climatol. 21 : 17331757. Nicholson, S. E., and J. Kim. 1997. The relationship of t he El Nio Southern Oscillation to Afri can rainfall. I nt. J. Climatol. 17 : 117 135. Nicholson, S. E., and J. C. Selato. 2000. The influence of La Nia on Afric an rainfall. Int. J. Climato l. 20 : 17611776. Nicholson, S. E., B. Some, and B. Kone. 2000. An analysi s of recent rainfall conditions in West Africa, inc luding the rainy seasons of t he 1997 El Nio and the 1998 La Nia years. J. Climate 13: 26282640. Nigam, S.N., S.L. Dwivedi and R.W. Gibbons. 1991. G roundnut breeding, constraints, achievements and future possibil ities. Plant Breeding Abst. 61: 11271136. O ECD. 2009. OECD (Organization for Economi c Co operation and Development) Annual Report, 2009.West African Studies. Regional Atlas on West Africa.Paris: OECD/SWAD.

PAGE 155

155 Ong, C.K. 1984. The influence of temperature and water deficit on the partitioning of dry mat ter in groundnut (Arachis hypogaea L.). Journal of Experimental Botany 35 746 755. Pan, D. 1996. Soybean responses to elevated temperature and doubled CO2. Ph.D. dissertation. University of Florida, USA. 227p. Phakamas, N., A. Patanothai, S. Jogloy, K. Pannangpetch, and G. Hoogenboom. 2008a. Physiological determinants for pod yield of peanut lines. Crop Sci. 48:2351 2360. doi:10.2135/ cropsci2008.02.0102 Piepho, H.P. 1994. Empirical best linear unbiased prediction i n cultivar trials using factor analytic variancecovariance structures. Theor. Appl. Genet. 97:195201. Pilumwong, J., C. Senthong, S. Srichuwong, and K .T. Ingram. 2007. Effects of temperature and elevated CO2 on shoot and root growth of peanut ( Arachis hypogeae L) grown in controlled environment chambers. ScienceAsia, PinstrupAndersen, P., R. PandyaLorch, and M.W. Rosegrant. 1999. World food prospects: critical issues for the twenty first century Washington, IFPRI. Peng, S., and D.R. Krieg. 1992. Gas exchange traits and their relationship to water use efficiency of grain sorghum. Crop Sci. 32:386 391. Peng, S., G. S. Khush, P. Virk, Q. Tang, and Y. Zou. 2008. Progress in ideotypebreeding to increase rice yield potential. Field Crops Res. 108:3238. Piepho, H.P., J.B. Denis, and F.A. van Eeuwijk. 1998. Predicting cultivar differences using covariates. J. Agric. Biol. Environ. Stat 3:151 162. Prasad, P.V.V., K.J. Boote, and L.H. Allen, Jr. 2006. A dverse high temperature effects on pollen viability, seedset, seed yield and harvest index of grainsorghum [Sorghum bicolor (L.) Moench] are more severe at elevated carbon dioxide due to high tissue temperature. Agric. For. Met 139:237 251. Prasad, P.V.V., K.J. Boote, L.H. Allen, Jr., and J.M .G. Thomas. 2003. Super optimal temperatures are detrimental t o peanut ( Arachis hypogaea L.) reproductive processes and yield at ambient and elevated carbon dioxide. Global Change Biol. 9:1775 1787. Prasad, P.V.V., K.J. Boote, L.H. Allen, Jr., and J.M.G. Thomas. 2002. Effects of elevated temperature and carbon dioxide on seedset and yield of kidney bean ( Phaseolus vulgaris L.). Global Change Biol 8:710 721. Prasad, P.V.V., P.Q. Craufurd, and V.G. Kakani. 2001. Influence of high temperature during pre and post anthesis stages of floral development on fruit set and pollen germination in peanut. Austral ian Journal of Plant Physiology 28: 233 240.

PAGE 156

156 Prasad, P.V.V., P.Q. Craufurd, and R.J. Summerfield. 1999a. Sensitivity of peanut to timing of heat stress during reproduc tive development. Crop Science 39: 13521357. Prasad, P. V.V., P.Q. Craufurd, and R.J. Summerfield. 1999b. Fruit number in relation to pollen production and viability in groundnut ex posed to short episodes of heat stress. Annals of Botany 84: 381 386. Prasad, P.V.V., P.Q. Craufurd, and R.J. Summerfield. 2000a. Ef fects of shor t episodes of heat stress on flower production and fruitset of groundnut ( Arachis hypogaea L.). Journal of Experimental Botany 51: 777 784. Prasad, P.V.V., P.Q. Craufurd, and R.J. Summerfield. 2000b. Effect of high air and soil temperature on dry matter production, pod yield and yield components of groundnut. Plant and Soil 222: 231 239. Putto, C., A. Pathanothai, S. Jogloy, K. Pa nnangpetch, K. J. Boote, and G. Hoogenboom. 2009. Determination of efficient test sites for evaluation of peanut breeding lines using the CSM CROPGRO peanut model. Field Crop Res. 110:272281. Rasmusson, D.C. 1991. A plant breeders experience with ideotype breeding. Field Crops Res. 26:191 200. doi:10.1016/03784290(91)90035T Rharrabti, Y., D.M. Garcia del Moral, L.F. Villegas, and C. Royo. 2003. Durum wheat quality in Mediterranean environments. III: Stabi lity and comparative methods in analyzing G E interaction. Field Crops Res. 80:141146. Richards, R.A., 2000. Selectable traits to increase crop photos ynthesis and yield of grain crops. J. Exp. Bot. 51: 447 458. Romagosa, I., and P.N. Fox. 1993. Genotype x environm ent interaction and adaptation. In M.D. Hayward, N.O. Bosemark & I. Romagosa, eds. Plant breeding: principlesand prospects p. 373390. London, Chapman & Hall. Ropelewski, C. F., and M. S. Halpert. 1987. Global a nd regional scale precipitation patterns associated with the El Nio/Southern Oscillation. Mon. Wea. Rev. 115: 16061626. Roudier, P., S. Benjamin, Q. Philippe, and B. Alexis. 2011. The i mpact of future climate change on West African crop yields: what does the recent literature say? Global Environmental Change 21:10731083. Rowell, D. P. 2001. Teleconnections between the tropical Pacific and the Sahel. Quart. J. Roy. Meteor.Soc. 127 : 1683 1 706.

PAGE 157

157 Sanders, T.H., R.J. Cole, P.D. Balnkenship, a nd J.W. Dorner. 1993. Aflatoxin contamination of peanuts from plants drought stressed in pod and root zones. Peanut Sci. 20(1): 5 8. SAS Institute. 2009. The SAS system for Windows. Release 9.2. SAS Inst., Cary, NC. Scholes, R.J., and R. Biggs. (Eds.), 2004. Ecosystem services in southern Africa: A regional assessment. Millennium Ecosystem Assessment. Online at http://www.millenniumassessment.org Sedgley, R.H. 1991. An appraisal of the Donald ideotype after 21 years. Field Crops Res. 26:93 112. doi:10.1016/03784290(91)90031P Sharma, P.S., and M.V.K. Sivakumar. 1991. Penetromet er soil resistance, pod number and yield of peanuts as influenced by drought st ress. Indian J. Plant Physiol. 34 (2): 147 152. Shinoda, M., and R. Kawamura. 1994: Tropical rainbel t, circulation, and sea surface temperatures associated with the Sahelian rainfall trend. Journal of the Meteorological Society of Japan 72 : 341 357. Shokes, F.M., and A.K. Culbreath. 1997. Early and late leaf spots. In Compendium of Peanut Diseases, 2nd edition (Eds N. K. Burelle, D. M. Porter, R. R. Kabana, D. H. Smith and P. Subrahmanyam), p p. 17 20. St Paul, MN: American Phytopathology Society. Sho rter, R., R.J. Lawn, and G.L. Hammer. 1991. Improvin g genetic adaptation in crops: A role for breeders, physiologists and modelers. Exp. Agric. 27:155175. Simmonds, N.W. 1991. Selection for local adaptation in a plant breeding programme. Theor. Appl. Gen et 82: 363367. Sinclair, T.R., 1998. Historical changes in harvest index and crop ni trogen accumulation. Crop Sci. 38: 638643. Singh P., K.J. Boote, U. Kumar, K. Srinivas, S.N. Nigam, and J.W. Jones. 2012. Evaluation of genetic traits for improving prod uctivity and adaptation to climate change in India. Journal of Agronomy and Crop Science. 198:399 413 Singh, A., P.S. Sabharwal, D.V.S. Panwar, and I.S. Mehl a. 1998. Stability of yield and seed parameters in rice. Pl. Breed. Abst 4559(5): 68. Singh, P., K .J. Boote, A.Y. Rao, M.R. Iruthayaraj, A.M. Sheikh, S.S. Hundal, R.S. Narang, and P. Singh. 1994a. Evaluation of the groundnut model PNUTGRO for cropresponse to water availability, sowing dates and seasons. Field Crops Res. 39: 147 162

PAGE 158

158 Singh, P., K. J. Bo ote, and S. M. Virmani. 1994b. Ev aluation of the groundnut model PNUTGRO for crop response to plant populati on and row spacing. Field Crops Res. 39:163 170. Sloane, J.F., R.P. Patternson, and T.E. Cater, Jr. 1990. Field drought tolerance of a soybean plant introduction. Crop Sci. 30:118 123. Smith, D. H. 1984. Early and late leaf spots. In Compendium of Peanut Diseases, (D. M. Porter, D. H. Smith and R. R odriguez Kabana, Eds), American Phytopathological Society, St Paul, MN, USA, pp. 5 7. Sponchiado, B.N., J.W. White, J.A. Castillo, and P.G. J ones. 1989. Root growth of four common bean cultivars in relation to drought tolerance in environments with contrasting soil types. Exp. Agric. 25:249 257. Stanciel, K., D.G. Mortley, and D.R. Hileman. 2000. Growth, pod, and seed yield, andgas exchange of hydroponically grown peanut in response to CO2 enrichment. HortScience 35:49 52. Stirling, C.M., and C.R. Black. 1991. Stages of reproductive develop ment in groundnut ( Ar a chis hypogaea L.) most susceptible to env iro nmental stress. Trop. Agric. (Trinid ad) 68 (3): 296 300. Stott, P.A., S.F.B. Tett, and G.S. Jones. 2001. A ttribution of twentieth century temperature change to natural and anthropogenic c auses. Climate Dynamics 17: 1 21. Stronach, I.M., S.C. Clifford, and A.D. Mohamed 1994. The effects of elevated atmospheric carbon dioxide, temperature, and soil moisture on w ater use of stands of groundnut (Arachis hypogaea L.). Journal of Experimental 45:16331638. Subrahmanyam, P., D. McDonald, F. Waliyar, L.J. Re ddy, S.N. Nigam, R.W. Gibbons, V. Ramanatha Rao, A.K. Singh, S. Pande, P .M. Reddy, and P.V. Subba Rao. 1995. Screening methods and sources of resistance to rust and late leaf spot of groundnut. ICRISAT Information Bulletin No. 47. ICRISAT, Patancheru, And hra Pradesh, India. Subrahmanyam, P., H. Hassane, A. Mounkaila, and D.H. SMITH. 1988a. A groundnut rosette epidemic in Niger. International Arachis Newsletter, 3, 3 9. Suriharn, B., A. Patanothai, K. Pannangpetch, S. Jogloy, and G. Hoogenboom. 2008. Yield performance and stability evaluation of peanut breeding lines with the CSM CROPGRO Peanut model. Crop Sci. 48:13651372. Suther, D.M., and M.S. Patel. 1992. Yield and nutrient a bsorption by groundnut and iron availability in soil as influenced by lime and soil water. J Indian Soc. Soil Sci. 40 (3):594 596.

PAGE 159

159 Teng, P.S., W.D. Batchelor, H.O. Pinnschmidt, and G.G. Wilkerson. 1998. Simulation of pest effects on crops using coupled pest crop models: the potential for decision support. p 221266. In G.Y. Tsuji et al (ed.) Understanding options for agricultural production. Kluwer Academic Publishers, Dordrecht The Netherlands. Tesfaye, K., and S. Walker. 2004. Matching of crop and environment for optimal water use: the case of Ethiopia. Phy sics and Chemistry of th e Earth 29: 1061 1067. Tewolde, H., C.J. Fernandez, and C.A. Erickson. 2006. Wheat cultivars adapted to post heading high temperature stress. J. Agron. Crop Sci. 192:111 120. Thomas, D.S.G., and C. Twyman. 2005. Equity and justi ce in climate change adaptati on amongst natural resource dependent societies, Global Environmental Change 15 (2005), pp. 115124. Truberg, B., and M. Huehn. 2000. Contribution to the analysis of genotype by environment interactions: Comparison of different parametric and nonparametr ic tests for interactions with emphasis on crossover interactions. J. Agron. Crop Sci. 185:267274. Tsuji, G.Y., G. Uehara, and S. Balas. 1994. DSSA T Version 3: vols. 1, 2, and 3. IBSNAT Project. Department of Agronomy and Soil Science, University of Hawaii, Honolulu, Hawaii. Narang, and P. Singh. 1994. Evaluation of the groundnut model PNUTGRO for crop response to water availability, sowing dates, and s easons. Field Crops Res. 39:147 162. Waliyar, F. 1991. Evaluation of yield losses due to groundnut leaf diseases in West Africa. In Summary Proceedings of the Second ICRISAT Regional Groundnut Meeting for West Africa, 11 14 September 1990, Niamey, Niger (Eds B. J. Ndug uru, F. Waliar & B. R. Ntare), pp. 32 33. Patancheru, India: ICRISAT. Wal iyar, Waliyar, F., M. Adomou and A. Traore. 2000. Rational use of fungicide applications to maximize peanut yield under foliar disease pressure in West Africa. Plant Dis. 84: 1203 1211 Walther, G.R., E. Post, and P. Convey. 2002. Ecologi cal responses to recent climate change. Nature 416: 389 395. Ward, M. N. 1992: Provisionally corrected surface wind data, worldwide ocean atmosphere surface fields, and Sahelian rainfall variability. J. Climate 5 : 454475. Ward, M. N., C. K. Folland, K. Maskell, D. Rowell, a nd A. Colman. 1990: Understanding and predicting seasonal rainfall in s ubSaharan Africa. WMO Tropical Meteorology Research Programme Report Series 36: 157 161. Wendt, J.W., R.B. Jones, and O.A. Itimu. 1994. An integrated approach to soil fertility improvem ent in Malawi, including agroforestry. In E.T. Craswell and J. Simpson

PAGE 160

160 (eds.), Soil Fertility and Climatic Constraints in Dryland Agriculture. ACIAR Proceedings No.54. Canberra, Australia: Australian Council for International Agricultural Research (ACIAR). Pp. 7479. White, J.W. 1998. Modeling and crop improvement. p. 179 188. In G.Y. Tsuji, G. Hoogenboom, and P.K. Thornton (ed.) Understanding options for agricultural production. Kluwer Academic Publishers, Dordrecht, The Netherlands. White, J.W., and G. Ho ogenboom. 2003. Genebased approaches to crop simulation: Past experiences and future opportunities. Agron. J. 95:52 64. doi:10.2134/agronj2003.0052. Willmott, C.J., S.G. Ackleson, R.E. Davis, J.J. Feddema, K.M. Legates, D.R. Legates, J. OConnell, and C.M Rowe. 1985. Statistics for the evaluation and comparison of models. J. Geophys. Res. 90(C5): 8995 9005. Wood, I.M.W. 1968. The effect of temperature at early flowering on the growth and development of peanuts (Arachis hypogaea). Australian J ournal of Agricultural Research 19: 241 251. Wright, G.C. 1989. Effect of pod zone moisture content on reproductive growth in three cultivars of peanut ( Arachis hypogaea L.). Plant Soil 116:111 114. Wright, G.C., K.T. Hubick, and G.D. Farquhar. 1988. Disc rimination of carbon isotope of leaves correlates with water use efficiency of fiel d grown peanut cultivars. Aust. J. Plant Physiol. 15:815 825. Yates, F. & Cochran, W.G. 1938. The analysis of groups of experiments. J. Agric. Sci ., 28: 556 580. Yin, X., M. J. Kropff, and P. Stam. 1999. The role of ecophysiological models in QTL analysis: the example of specific leaf area in barley. Heredity 82:415421

PAGE 161

161 BIOGRAPHICAL SKETCH As the only son of a man who was conscientious in the discharge of his duties, enjoyed the study of books and was refreshed and fulfilled after hours of working in his nearby farm, Stephen Narh had experience with growing crops at an early age. After graduating from Anum Presbyterian and Bishop Herman High School, he earned a Bachelor of Science degree in the year 2000 and Master of Philosophy degree in soil science from the University of Ghana, Legon in 2006. As part of his Master of Philosophy degree, he had his first international exposure on a oneyear program in science and engineering at th e Tokyo University of Agriculture and Technology, Japan in 2004. In the Fall of 2009, he joined the University of F lorida for his doctorate degr ee.