Probabilistic multi-model ensemble prediction of interdecadal variability of East Asian surface air temperature based on CMIP5 data

Abstract

Based on the CMIP5 runs of 10 models for climate hindcast of the surface air temperature, the prediction based on the method of Bayesian model average (BMA) has been conducted in the study. The results show that the ensemble system exhibited high performance in hindcasting the interdecadal (1981–2010) mean of temperature anomalies. In the RCP4.5 scenario, the decadal projection initialized on 2005 is used to estimate the temperature changes for the period from 2006 to 2035. Moreover, the temperature averaged over 2006-2035 increases conspicuously compared with 1961-1990. The results show that the temperature rising in the south is relatively slow, which reaches 0.7°C, while the temperature rises 1.3°C in the north. The temperature rise in the Western Pacific is fairly modest, while the temperature rise in the northwestern part of East Asia is more considerable, and the inland temperature rises more significantly than the coastal and offshore areas.