降水的统计降尺度预报及其空间相关性和时间连续性重建

Statistical downscaling forecast and reconstruction of spatial and temporal correlation of the precipitation

利用欧洲中期天气预报中心(ECMWF)、日本气象厅(JMA)、美国国家环境预报中心(NCEP)以及英国气象局(UKMO)四个中心1~7 d日累计降水量集合预报资料,以中国降水融合产品作为"观测值",对我国地面降水量进行统计降尺度预报,并对预报降水的空间相关性和时间连续性进行重建。对降水量进行分级后,建立各个量级的回归方程进行统计降尺度预报。此外,还利用Schaake Shuffle方法重建丢失的空间相关性和时间连续性。结果表明,分级回归比未分级回归后的预报结果相关系数更高,预报误差更小,更接近观测值。Schaake Shuffle方法可以有效地改进降水预报的空间相关性和时间连续性,使之更接近实况观测,集合成员间的相关性也更好。

Based on the ensemble forecasts of 1-7 day daily accumulated precipitation from the European Centre for Medium-Range Weather Forecasts(ECMWF),Japan Meteorological Agency(JMA),National Centers for Environmental Prediction(NCEP)and UK Met Office(UKMO),with the hourly merged precipitation product over China as the observed data,the forecast of daily precipitation in China by means of statistical downscaling and the reconstruction of spatial and temporal correlation of the precipitation forecast were conducted.The statistical downscaling based on different categories of the rainfall was applied to improve the precipitation forecast.The Schaake Shuffle was used to reconstruct the spatial correlation and temporal persistence of the precipitation forecast.The results show that the forecasts after the regression based on the different categories of the rainfall are more accurate than the ones after the regression.Classifying the rainfall into different categories was not considered,due to the fact that the anomaly correlation coefficient becomes larger and the root-mean-square error becomes smaller.The spatial and temporal correlations after the reconstruction are quite close to the observed ones.