应用统计学降尺度方法预测汉江流域降水变化

Application of statistical downscaling method in precipitation prediction for the Hanjiang river basin

统计降尺度方法是目前国内外研究气候变化的一个新途径。以汉江流域为例，选择全球气候观测NCEP再分析数据（1960～2000年）来率定和检验模型，利用主成分分析方法和多元线性回归模型建立大尺度GCMs模型的日降水统计降尺度方法，并应用全球气候模型CGCM2的A2气候情景来预测和分析汉江流域未来降水变化。相对于1961～2000年实测降水量均值，上游2001～2020、2121～2040年和中游2001～2020年的年降水量分别减少3．97％、4．85％和1，5％，其余统计时间年降水量大于实测值。

Statistical downscaling is a hot topic in the climate change research at present.For studying the future precipitation change in Hanjiang Basin,a large-size daily precipitation statistical downscaling model GCMs was proposed by using the multi-linear regression analysis and the principal component analysis method and the NCEP reanalysis data and the observation precipitation data(1960～2000) were used to calibrate the model.Based on the IPCC A2 scenario,the CGCM2 predictors were as input to the model to predict the precipitation change in future.Comparing to the period 1960～2000 the precipitation in the upper basin in period of 2001～2020 and 2021～2040 and in the middle basin in period of 2001～2020 will decrease by 3.97%,4.85%and 1.5% respectively.The results show that the daily precipitation statistical downscaling model can be further applied for predicting and analyzing the runoff change in Hanjiang basin in future.