降尺度CMIP6模式对新疆降水模拟评估  

Evaluation of Precipitation Simulation of Downscaled CMIP6 Models in Xinjiang

作　　者：胡家晖 孙彩霞 董思言[2] 王磊斌 欧阳鸿翔 吴志文 钟校尧 HU Jiahui;SUN Caixia;DONG Siyan;WANG Leibing;OUYANG Hongxiang;WU Zhiwen;ZHONG Xiaoyao(Xinjiang Climate Center,Urumqi 830002,China;National Climate Center,Beijing 100081,China;Hebei Normal University,Shijiazhuang 050027,China;Jiangxi Meteorological Observatory,Nanchang 330029,China;Guangdong Electric Power Design and Research Institute Co.Ltd,China Energy Engineering Group Corporation,Guangzhou 510663,China;Dongyuan County Meteorological Bureau,Dongyuan 517500,China)

机构地区：[1]新疆维吾尔自治区气候中心,新疆乌鲁木齐830002 [2]国家气候中心,北京100081 [3]河北师范大学,河北石家庄050027 [4]江西省气象台,江西南昌330029 [5]中国能源建设集团广东省电力设计研究院有限公司,广东广州510663 [6]东源县气象局,广东东源517500

出　　处：《沙漠与绿洲气象》2024年第5期52-58,共7页Desert and Oasis Meteorology

基　　金：国家自然科学基金联合基金(U2342228);新疆气象局科技创新发展青年基金项目(QN202307);新疆气象局引导性项目(YD2022006);中国气象局创新发展专项(CXFZ2024J044);新疆气候中心科研基金课题(QHCX-2023-06)。

摘　　要：基于6个CMIP6模式的日降水量数据,采用降尺度方法将其统一分辨率到0.25°×0.25°,选取5个极端降水指数从降水气候态、极端性、季节性对新疆区域1961—2014年历史期降水模拟效果评估。结果表明:降尺度CMIP6模式能较好再现新疆区域降水的空间分布特征,最大年均降水量误差<30 mm,夏季降水模拟效果最佳,相关系数均高于0.8。模式在春、秋季对降水的模拟效果差异较小,标准差比值均在1.00~1.25,ACCESS-CM2模拟效果最佳。模式集合均值能模拟出观测降水增多趋势,但低估了降水的年际变率,模拟结果提示新疆20世纪80年代的降水转折可能与人类活动有关。在降水极端性和季节性方面,降尺度数据对新疆的极端降水和季节性降水均有较好的模拟性能,降尺度数据对季节性降水的模拟能力(与观测均值误差<0.001)比原始分辨率的数据(与观测误差>0.005)效果好。Based on the daily precipitation data from six CMIP6 models,a downscaling method was used to unify the resolution to 0.25°×0.25°,and five extreme precipitation indices were selected to evaluate the simulation effect of historical precipitation in Xinjiang from 1961 to 2014 from the perspectives of precipitation climate state,extreme,and seasonality.The results show that the downscaled models accurately capture the spatial distribution of precipitation in Xinjiang,with the highest annual precipitation error under 30 mm and a summer simulation correlation exceeding 0.8.The difference in the simulation effect of the model on precipitation during the spring and autumn seasons is small,with standard deviation ratios ranging from 1.00 to 1.25.and the ACCESS-CM2 has the best simulation effect.Although the ensemble mean simulated the increasing trend in observed precipitation,it underestimate interannual variability.The simulation results suggest that the precipitation turning point in Xinjiang in the 1980s may be related to human activities.In terms of extreme and seasonal precipitation,downscaled data has a good simulation performance for extreme and seasonal precipitation in Xinjiang.The ability of downscaled data to simulate seasonal precipitation(with an error of less than 0.001 compared to the observation mean)is better than that of original resolution data(with an error of more than 0.005 compared to the observation mean).

关 键 词：CMIP6模式 降水 模拟评估 降尺度 

分 类 号：P456.7[天文地球—大气科学及气象学]