NEX-GDDP降尺度数据对中国极端降水指数模拟能力的评估

利用1986—2005年中国地面气象台站观测的格点化逐日降水数据(CN05.1)评估了NASA高分辨率降尺度逐日数据集NEX-GDDP中21个全球气候模式在0.25?(约25 km×25 km)分辨率下对中国极端降水的模拟能力.选取年最大日降水量(RX1D)、年最大5 d降水量(RX5D)、湿日总降水量(PRCPTOT...

Evaluation of extreme precipitation indices performance based on NEX-GDDP downscaling data over China

Taking the grid daily precipitation data (CN05.1) observed by China surface meteorological stations from 1986 to 2005 as the observation data, the performance of 21 global climate models were evaluated based on the high-resolution downscaling daily dataset NASA Earth Exchange/Global Daily Downscaled Projections (NEX-GDDP) with the resolution of 0.25° (~25 km×25 km). Six intensity indices, annual maximum daily precipitation (RX1D), the largest consecutive precipitation for five days (RX5D), total wet-day precipitation (PRCPTOT), simple daily precipitation intensity (SDII), cumulative precipitation in the 95 and 99 quantiles (R95p, R99p), and five frequency indices, heavy rain days (R50), cumulative precipitation days in the 95 and 99 quantiles (R95T, R99T), consecutive wet days (CWD), consecutive dry days (CDD), were selected for evaluation. The results show that: (1) It is difficult for models to capture the linear variation of extreme precipitation indices. Even for the best performance model, GFDL-ESM2G, only 45% of the simulated indices present the positive correlation with the observation. (2) The performance of models on the climatological means is better. CSIRO-MK3-6-0, NorESM1-M and MRI-CGCM3 have better performance on the intensity indices. The inmcm4, IPSL-CM5A-MR and MIROC5 have better performance on the frequency indices. The three best synthetical performance models are CSIRO-MK3-6-0, inmcm4 and MRI-CGCM3. (3) Considering the performance of 11 extreme precipitation indices in the climatological means and trend, GFDL-ESM2G, CSIRO-MK3-6-0 and ACCESS1-0 have relatively higher performance.