East Asian summer rainfall projection and uncertainty under a global warming scenario

Abstract

AbstractThis work investigated the simulation and future changes in East Asian summer mean rainfall, interannual variability, and associated uncertainty based on the historical and representative concentration pathway (RCP) 4.5 runs of 26 CMIP5 models. It is noted that the present‐day mean rainfall was well simulated by most of the models, though underestimation is over the Meiyu/Baiu/Changma (MBC) rainfall belt, and overestimation is in the western North Pacific (WNP). Nevertheless, it is still a challenge for the CMIP5 models to well simulate the interannual variability. The approach of the multi‐model ensemble greatly improves the credibility of model simulations, especially for the rainfall variability. The East Asian summer rainfall is projected to a relatively consistent increase in the near future, while the spatial distributions of the summer rainfall changes by the end of the 21st century are similar to those in the near future, with twice magnitude of the mean rainfall amount increase. The rainfall changes are associated with the strengthened low‐level southerly wind over East China and intensified western North Pacific high. Also, the variation of rainfall interannual variability by the end of the 21st century over East Asia is larger than that in the near future, with similar spatial patterns. It may imply that the stronger rainfall interannual variability is associated with the increased mean rainfall amount in the future. The inter‐model spreads of both the mean and interannual variability changes are larger in the low latitudes than in the high latitudes over East Asia. The signal‐noise‐ratio (SNR) is largely determined by the rainfall change (signal) in the future. Lower SNR of interannual variability than that of mean rainfall over East Asia implies that the CMIP5 models are more robust to a certain extent in simulating the summer mean rainfall change in East Asia than the change of interannual variability.