ECMWF细网格模式降水在浙江省行政区域面雨量预报中的释用  

Verification and Interpretation of ECMWF Fine-grid Model in Areal Precipitation Forecast in Zhejiang Province

作　　者：吕劲文 申华羽 涂小萍 方艳莹 杨小荣 Lyu Jingwen;Shen Huayu;Tu Xiaoping;Fang Yanying;Yang Xiaorong(Ningbo Meteorological Bureau,Ningbo 315012,China)

机构地区：[1]宁波市气象局,浙江宁波315012

出　　处：《气象与环境科学》2023年第5期75-85,共11页Meteorological and Environmental Sciences

基　　金：宁波市科技计划项目(2019C50004);浙江省气象科技计划重点项目(2017ZD06);浙江省气象科技计划项目(2018YB03)共同资助。

摘　　要：区域面雨量预报是决策部门开展防灾减灾工作的重要参考依据。基于2012-2018年浙江省气象站逐日逐小时雨量和ECMWF细网格模式降水产品,探讨各地市中等及以上过程的面雨量月际分布特征,评估模式对上述面雨量过程的预报水平,综合考虑不同时效的降水预报,应用逐步回归方法建立分区域的面雨量解释应用预报方程。结果表明:浙江沿海暴雨个例相对较多,而内陆区域性明显降水发生概率呈现出由西南向东北减小的趋势,中等及以上面雨量过程主要集中在梅雨期和台汛期,台汛期暴雨平均面雨量是梅雨期的1.3倍。ECMWF细网格24 h面雨量预报一般较实况偏小,中雨普遍偏小3 mm左右,大雨偏小7~12 mm,暴雨偏小20~40 mm,并且大雨和暴雨预报偏小距平百分率明显高于中雨的。随着时效的临近,模式的预报效果逐渐改善,72 h以内的预报具有更好的参考价值。各地市订正后的预报方程预报的面雨量,较确定性预报均有不同程度的提高,其中中雨Ts评分提高最为明显,普遍超过50%,同时经释用后多数地市不同量级降水的RMSE和MAE较ECMWF细网格预报的有所减小,预报稳定性更好,预报偏差更小。Areal precipitation forecasting provides an important help for decision-making for disaster prevention and mitigation.Based on daily and hourly rainfall observations of automatic weather stations,as well as precipitation forecasts of ECMWF fine-grid model from 2012 to 2018,this study discussed the inter-monthly variation characteristic of moderate-and-above rainfalls in different cities of Zhejiang province,moreover,the model forecast skills for rainfall had been evaluated.Using rainfall forecasts of ECMWF model in different lead hours,areal rainfall forecasting equations for different cities were established by stepwise regression method.Results show that torrential rainfalls were more likely to happen in coastal cities,and probability of moderate-and-above rainfalls tended to declined from south-west to north-east areas in Zhejiang.Moderate-and-above rainfall mainly occurred in Meiyu periods and typhoon seasons,with average areal torrential rainfall of the later 1.3 times of the former.The model generally underestimated 24 h areal rainfalls,with a deviation of about 3 mm for moderate rainfall,and 7 to 12 mm and 20 to 40 mm for heavy and torrential rainfalls respectively.Furthermore,the deviation percentage of underestimated heavy and torrential rainfall was larger than that of moderate rainfall.The fewer the lead hours,the better the model performed,and forecast provided good guidance within 72 h.Forecast skills of areal precipitations for cities of Zhejiang improved to some extent by interpretation,with Ts score for moderate rainfall improved the most,which is over 50%for each city.The RMSE and MAE of post-interpretation areal rainfall forecasts also smaller than ECMWF,so the results were less deviated and thus more reliable.

关 键 词：面雨量 ECMWF细网格模式 释用 浙江 

分 类 号：P459.9[天文地球—大气科学及气象学]