利用2013、2014年5-10月ECMWF集合预报系统（EPS）输出的降水预报资料，CMORPH（NOAA Climate Prediction Center Morphing Method）卫星与全国30000余个自动气象观测站的逐小时降水量融合资料，基于Brier评分、Talagrand分布、ROC（Relative operating characteristic）分析等方法，研究ECMWF集合预报对我国中部降水的概率预报性能。主要结论如下（1）模式对小雨具有稳定的预报技巧，但预报概率偏大；对大雨以上量级降水的分辨能力不足，概率预报偏小。（2）Talagrand图整体表现为“U”型，模式对中等强度以上量级降水预报频次偏少；就不同量级降水的预报和观测频次一致程度来看，时效越长预报性能越稳定；第11天模式集合成员的发散度发生明显变化，Talagrand图呈“钟”型分布，预报评分显著下降。（3）预报日数越短，降水概率预报效果越好，平均而言，预报日数超过2天后，模式对12 h（24 h）小于30%（40%）的暴雨概率预报技巧低于气候预测。（4）模式24 h小雨命中率较12 h偏高，空报率偏低，预报技巧优于12 h；对暴雨来说，前6天12 h降水概率预报的ROC面积较24 h表现较好。
Using precipitation forecasting data of ECMWF ensemble forecast from May to October in 2013 and 2014, hourly fusion precipitation data of CMORPH (NOAA Climate Prediction Center Morphing Method) satellites and gauged rainfall from about 30000 automatic weather stations, based on Brier score, Talagrand distribution and ROC (Relative operating characteristic) analysis, the precipitation probability forecast ability of ECMWF ensemble system in central China is studied.Results show that:ECMWF ensemble system has stable forecast skill for light rain, but probability forecast is large; the resolution for heavy rain or rainfall magnitude greater than heavy rain is low, and probability is small.Talagrand distribution overall show "U" shape, forecast frequency for moderate precipitation is less; as for the consistency of forecast and observation in different magnitude rainfall, the longer the forecast time is, forecast ability more stable; ensemble members' divergence have apparent changes in 11th forecast day, Talagrand distribution show "clocks" shape, forecast score decreased significantly.The shorter the forecast days, the better the precipitation probability forecast, on average, forecast time more than 2 days, forecast skill of 12 hours (24 hours) rainstorm probability less than 30% (40%) is lower than climatic prediction.24 hours precipitation forecast has higher hit rate and lower false rate compared with 12 hours forecast, forecast skill is better; in previous six days, ROC area of 12 hours precipitation probability forecast is better than 24 hours.
[1]Hamill T M, Colucci S J.1997.Verification of Eta-RSM short range ensemble forecasts[J].Mon Wea Rev, 125(6):711-724.
[2]Hamill T M.2001.Interpretation of rank histograms for verifying ensemble forecasts[J].Mon Wea Rev, 129(3):550-560.
[3]Kharin K, Zwiers F W.2007.Improved Seasonal Probability Forecasts[J].J Climate, 16(11):1684-1701.
[4]Murphy A H.1973.A new vector partition of the probability score[J].J Appl Meteor, 12(4):595-600.
[5]Murphy A H.1993.What is a good forecast? An essay on the nature of goodness in weather forecasting[J].Wea Forecasting.8(2):281-293.
[6]Stephenson D B, Coelho C A, Jolliffe I T.2005.Two Extra Components in the Brier Score Decomposition[J].Wea forecasting, 23(8):752-757.
[7]Wilks D S.2005.Statistical Methods in the Atmospheric Sciences[M].Pittsburgh:Academic Press, 234-251.
[8]Du Jun, Cheng Jing.2010.The corner stone in facilitating the transition from deterministic to probabilistic forecasts-ensemble forecasting and its impact on numerical weather prediction[J].Meteor Mon, 36(11):1-11.<br/>杜均, 陈静.2010.单一值预报向概率预报转变的基础-谈谈集合预报及其带来的变革[J].气象, 36(11):1-11.
[9]Du Jun, Deng Guo.2010.The utility of the transition from deterministic to probabilistic weather forecasts-verification and application of probabilistic forecasts[J].Meteor Mon, 36(12):10-18.<br/>杜均, 邓国.2010.单一值预报向概率预报转变的价值-谈谈概率预报的检验和应用[J].气象, 36(12):10-18.
[10]Duan Mingkeng, Wang Panxing, Li Chiqin.2012.Comparisons on forecast performance of NCEP ensemble prediction system between Eurasia and North America regions[J].Trans Atmos Sci, 35(6):641-651.<br/>段明铿, 王盘兴, 李驰钦.2012.NCEP集合预报系统在亚欧和北美区域的预报效果对比[J].大气科学学报, 35(6):641-651.
[11]Hao Ming, Xu Zhifang, Tao Shiwei, et al.2011.Simulation study and application experiment of GRAPES RUC system[J].Plateau Meteor, 30(6):1573-1583.<br/>郝民, 徐枝芳, 陶士伟, 等.2011.GRAPES RUC系统模拟研究及应用试验[J].高原气象, 30(6):1573-1583.
[12]He Xiaofeng, Zhou Rongwei, Sun Yihan.2014.Verification on surface wind speed of three global circulation models in China[J].Plateau Meteor, 33(5):1315-1322.DOI:10.7522/j.issn.1000-0534.2013.00093.<br/>何晓凤, 周荣卫, 孙逸涵.2014.3个全球模式对近地层风场预报能力的对比检验[J].高原气象, 33(5):1315-1322.
[13]He You, Yang Kun, Yao Chandong, et al.2012.Numerical simulation of a heavy precipitation in Qinghai-Xizang Plateau based on WRF model[J].Plateau Meteor, 31(5):1183-1191.<br/>何由, 阳坤, 姚檀栋, 等.2012.基于WRF模式对青藏高原一次强降水的模拟[J].高原气象, 31(5):1183-1191.
[14]Liu Jia, Xu Jinxia, Ma Zhengfeng, et al.2014.Evaluation of flood season prediction in southwest region by the second-generation monthly dynamic extended-range forecast[J].Plateau Meteor, 33(6):1468-1479.DOI:10.7522/j.issn.1000-0534.2013.00168.<br/>刘佳, 徐金霞, 马振峰, 等.2014.第二代月动力延伸预报产品对西南汛期降水的预报检验[J].高原气象, 33(6):1468-1479.
[15]Liu Yonghe, Yan Zhongwei, Feng Jinming, et al.2013.Predictability of 6-hour precipitation in the Yishu River Basin based on TIGGE data[J].Chinese J Atmos Sci, 37(3):539-551.<br/>刘永和, 严中伟, 冯锦明, 等.2013.基于TIGGE资料的沂沭河流域6小时降水集合预报能力分析[J].大气科学, 37(3):539-551.
[16]Niu Jinlong, Huang Chuhui, Liguoping, et al.2015.Moist helicity indexes selection and checking in Chengdu region for heavy rainfall based on the fine data[J].Plateau Meteor, 34(4):942-949.DOI:10.7522/j.issn.1000-0534.2014.00053.<br/>牛金龙, 黄楚惠, 李国平, 等.2015.基于高分辨率资料的湿螺旋度指标及其对成都强降水的预报应用[J].高原气象, 34(4):942-949.
[17]Pan Liujie, Zhang Hongfang, Wang Jianpeng, et al.2014a.An Objective Verification of Forecasting Ability of Japan High-Resolution Model Precipitation in China[J].Plateau Meteor, 33(2):483-494.DOI:10.7522/j.issn.1000-0534.2012.00188.<br/>潘留杰, 张宏芳, 王建鹏, 等.2014a.日本高分辨率模式对中国降水预报能力的客观检验[J].高原气象, 33(2):483-494.
[18]Pan Liujie, Zhang Hongfang, Wang Jianpeng.2014b.Progress on verification methods of numerical weather prediction[J].Adv Earth Sci, 29(3):327-335.<br/>潘留杰, 张宏芳, 王建鹏.2014b.数值天气预报检验方法研究进展[J].地球科学进展, 29(3):327-335.
[19]Xue Chunfang, Pan Liujie.2016.Object based diagnostic analysis of Japan thin-grid modes precipitation forecasting[J].Plateau Meteor, 35(2):406-418.DOI:10.7522/j.issn.1000-0534.2015.00010.<br/>薛春芳, 潘留杰.2016.基于MODE方法的日本细网格模式降水预报的诊断分析[J].高原气象, 35(2):406-418.
[20]Pan Yang, Shen Yan, Yu Jingjing, et al.2012.Analysis of the combined gauge-satellite hourly precipitation over China based on the OI technique[J].Acta Meteor Sinica, 70(6):1381-1389.<br/>潘旸, 沈艳, 宇婧婧, 等.2012.基于最优插值方法分析的中国区域地面观测与卫星反演逐时降水融合试验[J].气象学报, 70(6):1381-1389.
[21]Wang Min, Li Xiaoli, Fan Guangzhou, et al.2012.Calibrating 2 m temperature forecast for the regional ensemble prediction system at NMC[J].J Appl Metero Sci, 23(4):395-401.<br/>王敏, 李晓莉, 范广洲, 等.2012.区域集合预报系统2 m温度预报的校准技术[J].应用气象学报, 23(4):395-401.
[22]Xu JianWei, Gao Yan Hong.2014.Validation of summer surface air temperature and precipitation simulation over Heihe River Basin[J].Plateau Meteor, 33(4):937-946.DOI:10.7522/j.issn.1000-0534.2013.00149.<br/>许建伟, 高艳红.2014.WRF模式对夏季黑河流域气温和降水的模拟及检验[J].高原气象, 33(4):937-946.
[23]Zhang Lihong, He Guangbi.2014.Verification for GRAPES_Meso model in eastern Qinghai-Xizang Plateau and its surroundings in summer of 2011[J].Plateau Meteor, 33(1):14-25.DOI:10.7522/j.issn.1000-0534.2012.00175.<br/>张利红, 何光碧.2014.GRAPES_Meso模式对2011年夏季青藏高原东部及周边区域的预报检验[J].高原气象, 33(1):14-25.
[24]Zhao Linna, Liu Lin, Liu Ying, et al.2010.Assessment of probabilistic precipitation forecasts for the Huaihe Basin using TIGGE data[J].Meteor Mon, 36(7):133-142.DOI:10.7519/j.issn.1000-0526.2010.7.020.<br/>赵琳娜, 吴昊, 田付友, 等.2010.基于TIGGE资料的流域概率性降水预报评估[J].气象, 36(7):133-142.
