统计降尺度法在黄河源区未来降水变化分析中的应用

Research and Analysis of Change of Precipitation in Headstream Region of Yellow River by Using Statistical Downscaling Model

利用统计降尺度方法对黄河源区的未来气候变化中可能产生的降水变化进行了模拟分析,采用黄河源区的10个水文站点共41a的日降水资料为预报量,以NCEP再分析资料和英国HadCM3模式的大气变量为预报因子,研究了黄河源区月降水的降尺度方法。结果表明:采用建立回归关系和逐步线性回归分析方法可很好地选出有效的预报因子;SDSM模型可有效模拟黄河源区月降水过程,但SDSM模型对降水模拟的量值偏小;在2021～2050年间,黄河源区的降水将会出现一个较大变化,夏季降水呈上升趋势,降水集中于8月附近,上半年降水减少更易发生干旱,下半年发生极端降水事件的频率增加更易发生洪水。

The rainfall in Headstream Region of Yellow River may change due to future climate change and the possible change of rainfall is simulated and analyzed by using statistical downscaling model.The daily precipitations which collected by 10 hydrological stations in the Headstream Region of Yellow River during 41 years are adopted as the prediction variable and the predictors are selected from reanalyzed data of NCEP and climate variable of British HadCM3 mode.On the basis of these,the downscaling of monthly precipitation is studied.The results show that the predictors can be selected more efficiently by using the methodology of establishing regression relationship and stepwise multilinear regression method;SDSM can simulate monthly precipitation process in Headstream Region of Yellow River,but the simulation is underestimating the value of the precipitation;In the future,between 2021 and 2050,the rainfall in the Headstream Region of Yellow River will appear a large change;the rainfall tends to increase in summer and generally happens in August,moreover,the rainfall will reduce in first half year,which means drought may more likely to happen;however,the frequency of extreme precipitation events will increase in second year,which indicates that flood may more likely to take place.