资源环境科技发展态势分析平台

In this study, the extreme temperature in high-resolution hindcast simulations conducted by three reanalysis-driven regional climate models (WRF, RegCM4 and CCLM) is evaluated during the period of 1989–2009. We select 10 extreme temperature indices recommended by the Expert Team on Climate Change Detection and Indices to compare the simulations with the CN05.1 dataset. In addition, apparent temperature (AT) combining relative humidity and air temperature is applied to define the humid heat wave indices. The results show that the regional climate models (RCMs) can skillfully reproduce the spatial distribution and the probability density function of daily maximum temperature (Tmax) and minimum temperature (Tmin) in spite of some biases in certain ranges and regions. For the intensity indices (TXx, TNn, TXn, TNx), the RCMs demonstrate very high spatial correlations with the observation, but they have difficulties in reproducing the interannual variations of these indices. CCLM best simulates the interannual variabilities of the percentile-based indices, while WRF shows better skill in reproducing their spatial patterns (TX10p, TX90p). In regard to the duration indices, RegCM4 outperforms WRF and CCLM in simulating the consecutive cold events (CSDI), while WRF performs best for the continuous hot days (WSDI). The humid heat wave, which is characterized by both high temperature and humidity, is skillfully reproduced by the RCMs as well. The overestimation of the magnitude and intensity of humid heat waves in WRF can be attributed to the favorable large-scale circulation and water vapor condition.