GROWTH SIMULATION OUTPUTS FOR DETECTION OF DIFFERENTIAL CULTIVAR RESPONSE TO ENVIRONMENTAL-FACTORS

Despite considerable research on genotype-environment (GE) interaction s, breeders still need a simple way to describe the specificity of eac h genotype's response to environmental factors. A new approach is sugg ested based on (i) use of growth simulation outputs (simulated water d eficit, anthesis date, maximum leaf area) as environmental indices, (i i) use of simulated yield as a check, and (iii) use of simple correlat ion coefficients to describe the association between environmental ind ices and deviations from the average difference computed for each pair entry-check. Simulated grain yield can be considered a better indicat or of environmental adaptability, unaffected by factors like diseases, lodging, or winter-kill that can reduce the average yield in the best environments. Other outputs of simulation models, which integrate wea ther and soil factors with proper timing according to plant developmen t, can provide a better description of environments than the raw weath er data. Yield data of sixteen wheat (Triticum aestivum L.) genotypes, grown at three locations in Oregon for a 5-yr period (1988-1992) were analyzed. Correlation with water availability indices clearly differe ntiated the cultivars that were unable to adapt to improved environmen ts, because of lodging and/or disease susceptibility. Correlation with other environmental indices identified genotypes that responded more to low winter temperatures, to high temperatures after anthesis, or to delayed anthesis following cooler springs. The results indicated that the use of outputs from growth simulation models as covariates in ana lysis of GE Interaction could be a useful tool in characterizing diffe rential responses of genotypes to environmental factors.