Biases and improvements in three dynamical downscaling climate simulations over China

Hao Yang,Zhihong Jiang,Laurent Li
DOI: https://doi.org/10.1007/s00382-016-3023-9
IF: 4.901
2016-01-01
Climate Dynamics
Abstract:A dynamical downscaling is performed to improve the regional climate simulation in China. It consists of using a variable resolution model LMDZ4 nested into three global climate models (GCMs): BCC-csm1-1-m, FGOALS-g2 and IPSL-CM5A-MR, respectively. The regional climate from different simulations is assessed in terms of surface air temperature and rainfalls through a comparison to observations (both station data and gridded data). The comparison includes climatic trends during the last 40 years, statistical distribution of sub-regional climate, and the seasonal cycle. For surface air temperature, a significant part of the improvement provided by LMDZ4 is related to the effect of surface elevation which is more realistic in high-resolution simulations; the rest is related to changes in regional or local atmospheric general circulation. All GCMs and the downscaling model LMDZ4 are, more or less, able to describe the spatial distribution of surface air temperature and precipitation in China. LMDZ4 does show its superiority, compared to GCMs, in depicting a good regional terrain including the Tibetan Plateau, the Sichuan Basin and the Qilian Mountains.
What problem does this paper attempt to address?