一种预测我国汛期降水统计模型的建立  

Establishment of Statistical Model Predicting the Precipitation in the Flood Season in China

作　　者：程智[1,2,3] 

机构地区：[1]兰州大学大气科学学院 [2]中国科学院大气物理研究所东亚区域气候-环境重点实验室 [3]93381部队

出　　处：《安徽农业科学》2011年第16期9877-9878,9930,共3页Journal of Anhui Agricultural Sciences

基　　金：科技支撑项目(2007BAC294);国家自然科学基金(40775048;41075058)

摘　　要：[目的]建立预测我国汛期降水的回归模型。[方法]基于吴志伟等的统计模型,以北大西洋涛动指数、ENSO发展阶段和衰落阶段的海温指数预测东亚夏季风指数,将站点归为16个区域后,用同样因子建立预测我国汛期站点降水的回归模型,并与先预测季风指数再估计降水的模型进行了比较。[结果]对比各模型在2005~2009年夏季降水的预测结果发现,因子间接预测区域降水的模型好于间接预测站点的,同时因子直接预测区域的模型好于间接预测区域的模型;最优模型3个因子直接预测区域的预报评分P平均达74.2,距平相关系数ACC平均为0.219。综合5年的预测值与观测值的降水距平百分率正负区分布的对比情况来看,在东北南部、华北东部、江南部分地区、华南沿海以及新疆大部分地区的预测效果较好,降水距平百分率的预报与观测的正/负分布较为吻合。[结论]该模型能较好地预测我国夏季降水。[Objective] To establish the regression model which was used to predict the precipitation in the flood season in China.[Method] Based on the statistical model of Wu Zhiwei et al.,North Atlantic oscillation index and the sea temperature index in ENSO development and decline stages were used to predict the East Asian summer monsoon index.After the stations were divided into 16 zones,the same factors were used to establish the regression model predicting the precipitation in the station in the flood season in China.Moreover,it was compared with the model that firstly predicting the monsoon index,then estimating the precipitation.[Result] The prediction results of summer precipitation during 2005-2009 by every model were contrasted.It was found that the model that the factor indirectly predicted the regional precipitation was better than that indirectly predicted in the station.Meanwhile,the model that the factor directly predicted regional precipitation was better than that of indirectly prediction.The forecast score P of optimum model that three factors directly predicted the regional precipitation averagely reached 74.2,and the anomaly correlation coefficient ACC averagely reached 0.219.Seen from the comparison situation of positive and negative zone distribution of precipitation anomaly percentage between the predicted and observation values in 5 years,the prediction effects in the south and east of Northeast China,some areas in the south of Yangtze River,the coast of South China and most areas of Xinjiang were good.The positive and negative distribution of precipitation anomaly percentage prediction and observation tallied.[Conclusion] The model could well predict the summer precipitation in China.

关 键 词：东亚季风指数 线性回归 夏季降水 ENSO 北大西洋涛动指数 

分 类 号：P468.024[天文地球—大气科学及气象学]