Simulation and evaluation of multi-model dynamical downscaling of temperature extreme indices over the Middle and East China

The variable-resolution atmospheric general circulation model LMDZ4 that is nested into five global climate models (GCMs) (BCC-csm1.1-m, CNRM-CM5, FGOALS-g2, IPSL-CM5A-MR and IPSL-CM5A-MR) is used to conduct an ensemble dynamical downscaling simulation for China during 1961-2005. The performances of the above GCMs and the downscaling results for extreme temperature in China are evaluated comprehensively. Compared with GCMs, LMDZ4 shows its superiority by better depicting the terrain characteristics at regional scale like the Tibetan Plateau and Sichuan Basin and spatial distribution of extreme temperature in China. However, the improvement of dynamical downscaling shows significant regional differences. For mean minimum temperature, mean maximum temperature and frost days, the simulation by dynamical downscaling is mainly improved in Northeast China, Northwest China, the Tibetan Plateau and Southwest China. The correlation coefficients are increased to 0.95, the normalized root mean square errors are decreased to below 0.5℃ (0.5 d), and the improvements of the correlation coefficient for mean minimum temperature and mean maximum temperature both increase with terrain height. The improvement of heat wave duration index in Northeast China, North China and Southwest China is significant, but there are large differences between various models. Furthermore, compared with GCMs, the downscaling model is able to reproduce, to a certain extent, the spatial distributions of the trends of mean maximum temperature and mean minimum temperature in China, and reduce the trend errors of mean minimum temperature and frost days in Northeast China, North China, the Tibetan Plateau and Southwest China. The downscaling model ensemble also performs well in reproducing the observed spatial patterns of climate state and trends of temperature extremes in China. Dynamical downscaling can improve the simulation capability of GCMs for extreme temperature, which can be applied to the projection of future extreme temperature.