1961—2004年新疆降水极值概率分布特征  

Probability Distribution of Precipitation Extremes over Xinjiang during 1961-2004

作　　者：张延伟[1,2,3] 姜逢清[1] 魏文寿[1,4] 王雯雯[5] 刘明哲[1,3] 韩茜[1,2,3] 洪雯[1,2,3] 陆恒[1,2,3] 

机构地区：[1]中国科学院新疆生态与地理研究所,新疆乌鲁木齐830011 [2]中国科学院研究生院,北京100049 [3]中国科学院新疆积雪雪崩站,新疆伊宁835000 [4]中国气象局新疆沙漠气象研究所,新疆乌鲁木齐830002 [5]新疆大学数学与科学学院,新疆乌鲁木齐830046

出　　处：《中国沙漠》2012年第2期503-508,共6页Journal of Desert Research

基　　金：国家科技支撑计划项目(2007BAC03A0604);中国科学院新疆生态与地理研究所"绿洲学者"计划项目(40871027);中国科学院"西部之光"项目(RCPY200902);科技部公益行业科研专项(GYHY200806011);国家自然科学基金面上项目(41171066);国家自然科学基金项目(41071072)资助

摘　　要：根据1961—2004年新疆地区55个气象站逐日降水观测资料和PRECIS(Providing Regional Climates for Impacts Studies)区域气候影响模式(1961—2004年)逐日降水模拟资料,建立年最大降水AM(annual maximum)序列及日降水量小于0.05mm的年连续最长干旱天数AMCD(Annual Maximum Consecutive Dry Days)序列,分析了新疆地区降水极值序列的时空分布特征和概率分布模式。结果表明:①新疆地区降水事件的强度和概率最大的地区位于阿合奇、巴里坤、昭苏、乌鲁木齐等地,干旱事件强度和概率最大的地区位于且末、若羌、吐鲁番等地;②PRECIS区域气候影响模式模拟的新疆地区AM事件的多年平均值普遍高于观测值,且离差系数也普遍高于观测值;③PRECIS区域气候影响模式模拟结果与观测的降水极值空间分布有一定的差异,需要进行改进,但具有实际参考价值。对于实际观测的降水极值概率分布的拟合,证明了GEV分布函数能够较好地拟合降水极值的概率分布。Based on daily precipitation data from 55 meteorological stations in Xinjiang,China during 1961-2004 and the daily precipitation data simulated by PRECIS(Providing Regional Climates for Impacts Studies),the series of precipitation extremes,which contain annual maximum precipitation and annual maximum consecutive dry days,were constructed,and the distribution features of precipitation extremes were analyzed on the base of the two data series.Results show that: 1) the intensity and probability of annual maximum precipitation are the highest in Aheqi,Barkul,Zhaosu and Urumqi,and the intensity and probability of annual maximum consecutive dry days are the highest in Qiemo,Ruoqiang and Turpan;2) compared with observational data,inter-annual average maximum precipitations simulated by PRECIS are higher,and the discrete coefficients of annual maximum precipitation simulated by PRECIS are higher too;3) in spite of certain differences of the spatial distributions between observed and simulated precipitation extremes,the PRECIS simulation data have reference value.Fitting of precipitation extremes observation data shows that general extreme value distribution can fit the probability distribution of precipitation extremes quite well.

关 键 词：降水极值 广义极值分布 Gumbel分布 区域气候影响模式 

分 类 号：P426[天文地球—大气科学及气象学]