邻域空间检验技术在集合降水预报检验中的应用

常规降水检验受空间及时间微小差异所带来的"双重惩罚"影响严重,邻域空间检验FSS(Fraction Skill Score)方法在确定性预报中已体现出弥补这一不足的明显优势.随着集合预报分辨率的不断提高,集合降水预报同样存在与确定性预报相似的问题.本研究将FSS方法拓展至集合预报领域,构建适用于集合预报的降水空间检验指...

An extended spatial verification technique for ensemble precipitation forecasts

The established precipitation skill scores are influenced by the conventional duplex retribution problem which is induced by slight spatial and temporal contradiction between forecasts and observations.The FSS (Fraction Skill Score) is considered an approved scientific and problem-solving spatial technique that has been proposed in many articles for deterministic validation.In addition,this technique disclosed significant advantages in the directive of solving duplex retribution problem.In the consideration of high-resolution ensemble forecast,the ensemble precipitation forecasts also owned the corresponding problems with deterministic forecasts.In this research article,a novel ensemble precipitation verification skill score with spatial technique EFSS (Ensemble Fraction Skill Score) is developed based on extending FSS and deterministic ensemble forecasts.By using daily ensemble forecast obtained from ECMWF (European Centre for Medium-Range Weather Forecasts),QP and National Meteorological Information Center (NMIC) during June and August of 2018,the scoring consistency and the difference of traditional skill score in operational application have been analyzed.The consequences of this research indicated that EFSS is not affected by the ensemble members and the consistent evaluation can be obtained.The EETS (Ensemble Equitable Threat Score) was extended from deterministic traditional skill score that is suitable for ensemble forecasting.The comparison of EFSS and EETS indicates that the traditional skill scoring owned the \soutis limited skills in the directive to effective assessment of different characteristics of heavy precipitation processes.It can be concluded that,the EFSS technique can effectively improve the identification of heavy precipitation forecast verification.