Determination of plant traits that affect genotype × location (G×L) interaction in peanut using the CSM-CROPGRO-Peanut model

Document Type : Research Paper

Authors

1 Department of Plant Science and Agricultural Resources, Faculty of Agriculture, Khon Kaen University, Khon Kaen 40002, Thailand.

2 Department of Agronomy, University of Florida, Gainesville, Florida 32611-0500, USA.

3 AgWeatherNet, Washington State University, Prosser, WA 99350-8694, USA.

Abstract

Genotype × environment (G×E) interaction complicates the identification of
superior genotypes. An understanding its causes is needed for a more effective
breeding strategy. The objective of this study was to determine the plant traits that
cause genotype × location (G×L) interaction for pod yield in peanut using a modeling
approach. The CSM-CROPGRO-Peanut model was used to simulate pod yield for
17 peanut genotypes for 14 locations representative of all peanut production areas in
Thailandusing 30 years of historical weather data. Sensitivity analysis was used to
determine the effects of individual and combinations of plant traits on pod yield and
yield response to environments by varying the value of one or more cultivar
coefficients and then evaluating their effects. The results showed that the cultivar
coefficients that showed major effects were the duration from first seed to
physiological maturity (SDPM), maximum leaf photosynthesis rate (LFMAX), the
maximum fraction of daily growth that is partitioned to seed and shell (XFRT),
single seed filling duration (SFDUR) and the duration of pod addition (PODUR).
Those having minor effects were the duration from emergence to first flower
(EMFL), maximum leaf size (SIZLF) and maximum seed weight (WTPSD). The
cultivar coefficients that caused the differences in both mean yield and yield response
to locations between peanut genotypes in different pairs included LFMAX, XFRT,
SDPM, SFDUR and PODUR, but the causal characters differed among pairs of
genotypes. It was concluded that changing the degree of genotypic response to
environments is possible through selection for a combination of some of these traits,
and that model simulation could be used to identify those traits.

Keywords