See more from this Session: Symposium--Incorporating Genomic Knowledge Into Crop Simulation Models
Tuesday, November 2, 2010: 1:30 PM
Long Beach Convention Center, Room 103A, First Floor
Genetic improvements of grain yield (GY) and protein concentration (GPC) are impeded by large genotype by environment by management (G x E x M) interactions and by compensations (tradeoffs) between traits. Here global uncertainty and sensitivity analyses of SiriusQuality2, a robust process-based wheat simulation, were conducted to (i) identify candidate traits to increase GY and GPC, and (ii) identify genotypic parameters for model calibration. Three contrasted European sites were considering and simulations were performed on long term weather data and at two N supplies in order to quantify the effect of parameter uncertainty on GY and GPC under variable environments. The overall influence of all of the 75 parameters of the model was first analyzed using the Morris method. Forty one influential parameters were identified and their individual and total effects on the model outputs were quantified using the extended Fourier amplitude sensitivity test. The overall effect of the parameters was dominated by their interactions with other parameters. While under non-limiting N supply a few influential parameters with respect to GY were identified (e.g. the light use efficiency, the potential duration of grain filling or the phyllochron), under limiting N more than ten parameters showed equivalent effects on GY and GPC and their influence on GY and GPC was relatively limited. All the parameters had antagonist effects on GY and GPC, but leaf and stem N storage capacity appeared as good candidate traits to shift the negative GY-GPC relationship. These information on parameters’ uncertainty and sensitivity were then use to select the minimum set of parameters to describe G x E x M interactions. The model was calibrated for 16 elite bread wheat cultivars with a step-by-step procedure using a modified genetic algorithm. The ability of the model to simulate the observed G x E x M interactions was then analyzed using results from a two year-four site network.
See more from this Division: A03 Agroclimatology & Agronomic ModelingSee more from this Session: Symposium--Incorporating Genomic Knowledge Into Crop Simulation Models