Incorporating drought and marker associated traits into the DSSAT model

Abstract

Most agronomic traits are genetically complex and often show genotype × environment (G × E) interactions which complicate selection and slows down the breeding progress. As model parameters can represent certain genetic characteristics, crop modeling has been commonly used to identify and test desirable traits, to evaluate genetic improvement, and to design the optimum ideotype adapted to future climates. In this work, the DSSAT model was modified to evaluate the potential benefits of drought-tolerant traits in peanut under water-limited conditions, to incorporate genomic information to compute maize model inputs using marker-based information and combined with representative spatial model inputs to evaluate drought tolerant peanut performance under climate change. In Chapter two, the trait of maintaining photosynthesis under water deficit was observed in rainout shelter experiments and incorporated into the DSSAT model as a new drought tolerance cultivar coefficient. This specific trait was shown to be an advantageous trait for peanut varieties, which produced higher simulated rainfed yield with enhanced seasonal evapotranspiration and grain water use efficiency, especially for dry seasons. In Chapter three, we extend this approach to simulate the performance of drought tolerant peanuts for several important peanut production counties in the Southeastern USA. Results showed that a single set of cultivar coefficients and soil parameters could be calibrated to simulate historic peanut growth duration and county-level yields reasonably well. The simulation for future climate change indicated that the rainfed yields will suffer from increasing daytime temperature and an irrigation strategy could potentially offset the heat and drought stress to main higher peanut production in the Southeastern USA. Finally, In Chapter four we developed a methodology to use marker-based prediction of maize model inputs to assess new hybrid performance for plant breeding, and quantitatively assess the effect of genes by explicitly accounting for G × E interactions. These findings provided a promising insight into the use of crop model in drought-tolerant simulation, marker-based modelling, and regional scale simulation.