Genotype × environment interactions for phenological adaptation in narrow-leafed lupin: A simulation study with a parameter optimized model

Abstract

An understanding of genotype×environment (G×E) interactions for phenological adaptation in crops can help identify traits that facilitate genetic improvement and cultivar selection. Such understanding is also necessary to identify environments that facilitate identification of cultivars with better environmental adaptation. Process-based crop simulation models including APSIM are valuable tools to investigate G×E interactions and traits conferring adaption to particular conditions. However cultivar parameters relating to crop phenology in crop model are difficult to determine directly through field experiments. Within this paper, we present a simulation study to explore the environmental and genotypic control of time to flowering and maturity in lupin and its grain yield, by using the APSIM model. In the model, lupin cultivar parameters were optimized by combining prior knowledge and PEST (Parameter ESTimation), a model-independent parameter optimiser. The parameterisation was for nine lupin cultivars with contrasting maturity (early vs late) and vernalization requirement (non-responsive vs responsive) across 13 sites×years across Australia. We showed that phenological differences among the cultivars were well explained by the optimized set of parameters, mainly reflected in vernalization response and photoperiodic sensitivity before flowering. Simulated flowering days and grain yield agreed well with the observations, with RMSEs (root mean squared error) of 4days and 0.34tha−1 for all cultivars together. The simulation results showed that lupin yield was significantly affected by temperature and water availability. An early-maturity cultivar would yield better than the late-maturity one at the warm and low rainfall region, while at high rainfall or long-season regions, cultivar types should be chosen tactically based on the climatic conditions for that season. Thus, lupin breeding in Australia should therefore focus on temperature regimes and rainfall gradients. These results indicate that the parameter-optimized APSIM model could be used to explore the adaptation of genotypes to a particular environment to provide guidance for a more efficient breeding program and develop optimal agronomic management practices.