东北地区一次夏季强降雨过程多模式预报与空间检验

Multi-model Forecasting and Spatial Verification of a Summer Heavy Rainfall Process in Northeastern China

基于对象的诊断评估方法(MODE)是一种空间检验方法,以传统气象预报检验方法和MODE方法对东北地区2023年8月1—5日的欧洲中期天气预报中心全球模式(ECMWF模式)和中国气象局全球模式(CMA-GFS模式)24 h时效降水预报结论进行了详细评估,选取了较为典型的两个24 h降水时段分别进行检验。结果表明:CMA-GFS模式在晴雨预测上优于ECMWF模式,尤其是在中雨及以下量级中有较好的效果,而ECMWF模式在大雨及以上的预测中更胜一筹,空间检验的对象目标相似度也更高。但两个模式在24 h时效的降水强度的预测上均较小。

The MODE(method for object-based diagnostic evaluation)is a spatial verification method.Using conventional meteorological forecast verification methods and MODE,this paper conducted a detailed evaluation of the 24-hour validity precipitation forecast conclusions of European Centre for Medium-Range Weather Forecasts global model(ECMWF model)and the China Meteorological Administration Global Forecast System(CMA-GFS model)during the period of August 1st to 5th,2023 in northeastern China.Two typical 24-hour precipitation periods during this rainfall event were selected for evaluation.The results indicated that the CMA-GFS model outperformed the ECMWF model in overall precipitation prediction,particularly in the forecasting of light to moderate rain.However,the ECMWF model demonstrated superior performance in predicting heavy rainfall and above,with higher object similarity in spatial verification.Both models exhibited relatively lower precipitation intensity in 24-hour validity predicting.