夏季西太平洋副热带高压指数的统计预测模型

Statistical Forecast Model of Western Pacific Subtropical High Indices in Summer

利用国家气候中心提供的1959-2013年74项环流指数和美国NOAA气候预测中心提供的Nino 3区海温指数等气候指数,采用相关分析和最优子集的原理选择预测因子,建立统计预测模型,并对夏季西太平洋副热带高压(西太副高)各特征指数的年际增量进行预测,进而预测实际特征指数。为检验该预测模型的预测效果,对2001-2013年西太副高指数进行了独立样本回报检验,结果表明:该预测模型能够较好地回报出2001-2013年夏季西太副高脊线、西伸脊点以及强度指数,回报指数与观测的距平相关系数分别为0.64,0.67和0.78,预测准确率分别达到84.63%,76.92%,69.23%。因此,该预测模型对汛期气候预测具有潜在的应用价值。

By using the 74 circulation indices provided by National climate center,the Nino 3 indices and other climate indices provided by NOAA climate forecast center,a physical and statistical forecast model is established to predict the western pacific subtropical high(WPSH) indices in summer.The correlation analysis and the optimal subset regression are used to choose the predictors.The predictand is the year-to-year increment of the WPSH indices in summer.Then the real indices can be calculated.The hindcast of WPSH indices during 2001-2013 is carried out to evaluate the performance of this forecast model.The results show that the forecast model can predict the ridge line,the western ridge point,and the intensity indices of the WPSH successfully in most years.The anomaly correlation coefficients between the observed ridge line,the western ridge point and the intensity indices and the prediction during 2001-2013 are 0.64,0.67 and 0.78,respectively.The accuracy rate of the predicted ridge line,the western ridge point and the intensity indices during these 13 years reach up to84.63%,76.92%and 69.23%.Therefore,this forecast model will be of great value in the climate prediction during the flood season.