年际增量方法在西南夏季降水预测中的应用

利用中国西南地区80站逐月降水资料及NCEP/NCAR再分析资料等，采用降水预测新方法——年际增量法，考察影响中国西南地区夏季降水年际增量的前期冬、春季大气环流年际增量状况，并选取5个关键影响因子，采用多元回归法建立中国西南夏季降水年际增量预测模型。对降水年际增量进行预测，在1971—2010年的建模阶段，预测模型的拟合率为0.78，在2011—2017的后报检验7年中，有6年与实况值同位相。后报检验2011—2017年的降水距平百分率，均方根误差为16%。为考察对降水异常分布型的预报效果，逐站建立回归方程，并进行趋势预报检验，近5年的趋势异常综合评分高于发布预测，预报效果较好。因此，该方法的应用及模型的建立对提高西南地区夏季降水预测水平有重要意义。 

Application of an interannual increment method for summer precipitation forecast in Southwest China

Based on monthly precipitation observations collected at 80 stations in Southwest China and NCEP/NCAR reanalysis data, the method of annual increment is implemented to investigate interannual increment of atmospheric circulation in the preceding winter and spring, which influences interannual increment of summer precipitation in Southwest China. Five key factors have been selected to establish an interannual incremental prediction model for summer rainfall in Southwest China by using multiple regression method. In the modeling stage from 1971 to 2010 for interannual precipitation increment forecast, the fitting rate of the prediction model is 0.78. In the post-test seven years from 2011 to 2017, the forecast is in phase with observations in six out of the seven years, and the root mean square error of percentage precipitation anomalies is 16% in the return test from 2011 to 2017. The regression equation is established station by station to investigate the prediction effect on various types of precipitation anomaly, and the PS score in the recent five years is higher than that of operational forecasts. Therefore, the establishment of the model and its application are of great significance for the improvement of summer precipitation forecast in Southwest China.