Progress on Mechanism and Prediction Methods for Persistent Extreme Precipitation in the Yangtze-Huai River Valley

Persistent extreme precipitation (PEP) results in severe floods in China, especially in the Yangtze-Huai River Valley (YHRV), making it one of the main weather disasters in China. There exists an urgent need to enhance understandings on the formation mechanism and developing rules of PEP and extend forecast valid time of the PEP for the scientific decision of government.In recent years, progress has been achieved from related studies on the formation mechanism and forecast method of PEP in the YHRV which has caught wide attention. The method of automatically identifying regional PEP events is established which is named as RePEEI (Regional Persistent Extreme Event Identifier). Conceptual model is established on the large-scale circulation patterns responsible for PEP events, revealing that concurrent anomalies of the key influential systems are important causes for the occurrence and maintenance of PEP, and precursor signals (about 1-2 weeks prior to the onset of PEP) are investigated. Taking East Asia/Pacific teleconnection pattern (EAP) as a point of penetration, the mechanism of its effects on PEP is explored. Moreover, it indicates that whether the PEP will occur in YHRV is decided by the north-south location of high systems at low latitudes. Schematics for precursor circulation features of typical EAP patterns responsible for persistent extreme precipitation events in the YHRV is established. And corresponding precursor signals are also obtained, the feasibility of predicting PEP on the use of EAP is discussed. Furthermore, based on the key influential systems and precursor signals found above in characteristic large-scale circulation patterns, the physical statistical forecast model for the prediction of PEP is established, which is named as KISAM (Key Influential Systems based Analog Model), with the idea of parameter optimization method and ensemble mean introduced, using different predictors and cosine angular analog method with weight assigned.However, the forecast of PEP is still a challenge, especially when the forecast lead time extends to medium range or even extended range. The performance of numerical models in predicting the occurrence and location of PEP still leaves much to be solved. How to further improve direct outputs of numerical models and combine model outputs with physical statistical methods to improve the forecast of PEP is a research area that needs much more study.