基于全球模式模拟的山西省极端降水变化特征及评估  

Characteristics and Assessment of Extreme Precipitation Variation Based on the Simulation of the Global Model in Shanxi Province

作　　者：王艳 李晶晶 赵尚卓 杨艳平 

机构地区：[1]长治市气象局,山西长治046000

出　　处：《山西气象》2016年第4期11-17,28,共8页Shanxi Meteorological Quarterly

摘　　要：基于1961—2013年山西省地面观测逐日降水数据计算得到6个极端降水指数，对山西省极端降水的时空变化特征进行了分析。并且对CMIP5中31个全球气候模式的模拟效果进行评估。结果表明：①山西地区极端降水指数空间分布具有较大差异，日降水强度（pint）和极端降水阈值（pq90）南高北低，大雨日数（pnl90）和连续干日（pxcdd）东西差异明显，极端降水指标的年际变化显示山西省降水极端事件增加。②全球模式对山西省极端降水指数的模拟能力差异较大，其中对平均日降水强度（pint）和极端降水阈值（pq90）空间分布模拟能力较好，有一半模式的空间相关系数超过了0．6。同一模式对不同指数的模拟能力也存在较大的差异。③模式模拟相关系数和均方根误差有较好的一致性，模式排序的相关系数超过了0．8。ACCESS1．0、ACCESS1．2、CMCC—CM、CMCC—CMS以及MRIOC4h是模拟山西极端指数最好的5个全球模式。相比于多模式集合平均，优选模式集合能够显著降低模拟的湿偏差。The the spatial and temporal variation characteristics of extreme precipitation in Shanxi province were analyzed with Six extreme precipitation indexes which were calculated based on the daily precipitation datas of the ground observation in Shanxi province from 1961 to 2013. The paper evaluated the simulation results of 31 global climate models in CMIP5. The results showed that： （1） The spatial distribution of extreme precipitation index in Shanxi region existed differences： the daily precipitation intensity （pint） and the threshold value of extreme precipitation （ pq90 ） were higher in south region than north region, the number of heavy rain days （pnl90） and the continuous dry days （PXCDD） existed obvious differences between east region and west region, and the annual change of extreme precipitation index showed the increase of precipitation extreme events in Shanxi Province ; （2） The simulated abilities of the global model existed differences： the ability of spatial distribution simulation were better in both？average daily precipitation intensity（ pint ） and threshold value of extreme precipitation（ pq90 ）, the spatial correlationcoefficient of half of the model were more than 0.6, and the simulated ability of the same model to different indexes were also different significantly; （3） The simulation correlationcoefficient and the root mean square error were consistent, and the correlationcoefficient of model sort was more than 0.8. ACCESS1.0, ACCESS1.2, CMCC-CM, CMCC-CMS and MRIOC4h were the five best global models in simulaing the extreme indexs of shanxi. Compared with the multi-model ensemble mean, the set of preferred model could reduce the simulation deviationsignifieantly.

关 键 词：极端降水 全球气候模式 山西省 

分 类 号：P426.61[天文地球—大气科学及气象学]