沙漠与绿洲气象2025，Vol.19Issue(4)：77-85,9.DOI:10.12057/j.issn.1002-0799.2311.21002
基于EOF_AP聚类模型的京津冀冬季强降水中期预报技术研究
Application of EOF_AP Clustering Model in Medium-Range Forecasting of Winter Heavy Precipitation Events in Beijing-Tianjin-Hebei
摘要
Abstract
This study utilizes precipitation data from the ensemble forecast of the European Centre for Medium-Range Weather Forecasts(ECMWF)to establish a clustering model,named EOF_AP,based on empirical orthogonal function decomposition and the affinity propagation clustering method.The EOF_AP model was applied to identify and classify the characteristics of medium-range forecasts for three heavy winter precipitation events in Beijing-Tianjin-Hebei region during the winters of 2019 to 2021.The results indicate that:(1)The EOF_AP clustering model is effective in identifying and classifying the main precipitation patterns and their corresponding occurrence probabilities for heavy precipitation events.This model works by comprehensively retaining the key forecast information from ensemble members while using a few representative types.It also successfully identifies characteristics of low-probability precipitation patterns,such as the extreme type.(2)For extreme winter precipitation events in Beijing-Tianjin-Hebei region,there were a few forecasting members that provided accurate predictions approximately 168 hours in advance.The number of ensemble members with reliable predictions increased significantly around 72 hours in advance.(3)The occurrence probability of the forecasted precipitation location was higher for the more accurate forecasts compared to those for precipitation intensity.关键词
EOF_AP聚类/京津冀/冬季强降水/集合模式/中期预报Key words
EOF_AP clustering model/Beijing-Tianjin-Hebei region/winter heavy precipitation/ensemble model/medium-range forecast分类
天文与地球科学引用本文复制引用
黄威,胡莹莹,马杰,李勇..基于EOF_AP聚类模型的京津冀冬季强降水中期预报技术研究[J].沙漠与绿洲气象,2025,19(4):77-85,9.基金项目
2023年度中国气象局水文气象重点开放实验室开放研究课题(23SWQXM030) （23SWQXM030）
国家自然科学基金联合基金项目(U2142207) （U2142207）
