Comparison of wheat simulation models under climate change. II. Application of climate change scenarios

A comparison of the performance of 5 wheat models (AFRCWHEAT2, CERES, NWHEAT, SIRIUS and SOILN) was carried out for 2 sites in Europe Rothamsted, UK, and Seville, Spain. The aims of this study were (1) to compare predictions of wheat models for climate change scenarios, and (2) to investigate the effects of changes in climatic variability in climate change scenarios on model predic-tions. Simulations were run for climate change scenarios derived from a number of 2 × CO2 equilibrium and transient GCM (global circulation model) experiments. For most climate change scenarios the model results were broadly similar. Where results differed, much of the difference could be explained by model sensitivity to climate and differences in initial conditions. Transient scenarios without changes in climatic variability usually resulted in large yield increases for Rothamsted and in nil to large yield increases for Seville. Incorporation of changed climatic variability in the transient scenario had a more profound effect on grain yield and resulted in a substantial decrease in mean yield with a strong increase in yield variation at Seville. This was associated with the changes in the duration of dry spells and a redistribution of precipitation over the vegetation period. The results show that future studies of the effect of climate change on crop yields must consider changes in climatic variability as well as changes in mean climate.