Reanalysis-driven multi-RCM high-resolution simulation of precipitation within CORDEX East Asia Phase II

Under the Coordinated Regional Downscaling Experiment-East Asia second phase (CORDEX-EA-II) framework, the performances of three regional climate models (RCMs), including Weather Research and Forecasting (WRF) model, RegCM4 and Consortium for Small-Scale Modelling model in Climate Mode (CCLM), in simulating the climatic distribution, diurnal variation and extremes of precipitation during 1989-2009 are evaluated. The RCMs are driven by the ERA-Interim reanalysis and the evaluation is based on both the gridded gauge-based and the satellite-derived precipitation data. In general, the RCMs can reproduce the spatial pattern of the mean daily precipitation, with the spatial correlations (SCORs) larger than 0.66, and CCLM has the lowest root-mean-square error over East Asia. For the diurnal cycle of precipitation, WRF outperforms the other two RCMs despite the peak time about 3 hr ahead in some areas. RegCM4 overestimates the amplitude of the diurnal curve, and CCLM has limited ability to simulate the diurnal variation over most areas. All the three RCMs can reproduce the spatial patterns of the extreme indices, with the SCORs larger than 0.57. WRF and CCLM perform the best for the frequency of wet days (R1mm) and moderate rain (R20mm), respectively. The intensity indices are overestimated over most land areas, with relatively higher SCORs in WRF and relatively lower biases in RegCM4. The consecutive wet days (CWD) can be well simulated by WRF and CCLM; however, RegCM4 overestimates it over most regions in East Asia. The overestimation of the extreme intensity and frequency indices may be related to the higher water vapour content and the weaker water vapour divergence in the lower layer, as well as the higher mid-layer geopotential height, especially in CCLM.