Evaluation of Variation and Uncertainty in the Potential Yield of Soybeans in South Korea Using Multi-model Ensemble Climate Change Scenarios

Abstract

Recently, information provided by various Global Climate Models (GCMs) has been applied to various research fields. A Multi-model Ensemble (MME) approach, which assesses the impact of climate change on agricultural crop production using one or more climate datasets from GCMs, has been widely used. We estimated the changes in soybean potential yield at 16 sites using the climate change scenarios, and then predicted the relative change in predicted potential yield for each single GCM, producing an observation climate-based simulated potential yield. Lastly, we assessed the degree of uncertainty for changes in potential yield predicted from MME approach. In the results, although there were differences in the values themselves, the Standard Deviations (SD) of predicted soybean potential yield for each individual GCM were not significantly different from the SD of observation climate-based simulated potential yield, and there were no correlations between the predicted soybean potential yield for each individual GCMs and observation climate-based simulated potential yield in most sites. The estimation error decreased as the number of participating GCMs in the MME increased, but it did not decrease to zero. The means, but not the variance, of the MME of potential yield of soybean was similar to that of the observation climatebased simulated potential yield. The relative changes for predicted soybean potential yield for individual GCMs values of the Representative Concentration Pathways 4.5 and 8.5 scenarios increased in the northern regions of South Korea, such as Chuncheon and Hongcheon. In contrast, differences between them were not significant in most southwestern regions.