ECMWF细网格TP产品在北疆降雪天气中的预报性能检验

摘要

基于升级后的EC细网格TP降水量预报产品,对北疆2015年17场降雪天气的12h累积降雪量,主要运用平均误差、平均绝对误差和均方根误差进行了检验.结果表明,该模式在北疆降雪天气预报中小雪的预报准确率最高,暴雪最小;预报准确率随时效的延长并非都是减小的;小雪空报率较高、暴雪漏报率较高,中雪和大雪空报率和漏报率都不容忽视;小雪和中雪的3种误差均较小,随时效的延长变化较小;强降雪(大雪及以上量级)的误差较大,随时效的延长有增大的趋势;模式对小雪的预报总体为系统性偏大,对强降雪预报则为明显的系统性偏小,对中雪的预报系统性偏向不稳定.%Based on the upgraded ECMWF fine grid TP products,17 snowfall cases during 2015 in north Xinjiang were selected to evaluate 12 h accumulated snowfall by using statistical and synoptic methods.Results showed that the forecast accuracy (Ts) of light snow in North Xinjiang is the highest,and heavy Snowstorm is the lowest;Ts is not always increasing with forecast time length,but false alarm rate of light snow and under-predicting rate of snowstorm are higher.False and missed alarm rate of moderate snow and heavy snow also cannot be ignored;Three kinds errors of lights snow and moderate snow is small,and change little with increase of forecast time length;On contrast,error of Heavy Snowfall is greater and increase with forecast time length.Light snow is systemic over-predicted,Heavy Snowfall is systemic under-predicted obviously,moderate snow is instability.