Characteristics of regional heavy rainfall in the pre-flood season in South China and prediction skill of NCEP S2S

South China stands out as one of the regions experiencing the longest flood season and the highest frequency of floods in China. Understanding the characteristics of heavy precipitation during the flood season and its potential mechanisms is crucial for predicting and preventing flood disasters in South China. In this study, we classify regional heavy precipitation during the pre-flood season in South China into frontal heavy rainfall (FR) and warm-sector heavy rainfall (WR). FR is mainly controlled by the southwest monsoon and the westerly jet in the northwest of the West Pacific Subtropical High (WPSH). The westerly wind brings abundant water vapor to the front, creating an unstable environment conductive to heavy rainfall events. In instances of WR, precipitation occurs in warmer regions due to the northward movement of warm and humid air from the western side of the WPSH, coupled with the westerly jet transporting warm and humid air to the mainland. These atmosphere processes contribute to ascent, condensation, and the subsequent formation of heavy rainfall. Investigating the extended-range period predictability of the National Centers for Environmental Prediction (NCEP) sub-seasonal to seasonal (S2S) forecast system for these two types of heavy rainfall reveals a diminishing predictability for FR and WR as forecast lead time increases. Notably, NCEP S2S demonstrates superior prediction skill for FR compared to that for WR. The diminished prediction skill for WR within NCEP S2S primarily stems from model inadequacies related to forecasting atmospheric circulation and water vapor transport anomalies associated with WR. When forecasting WR, the easterly wind anomaly has influence the upper troposphere since the model's prediction 10 days in advance. Simultaneously, the model faces challenges in accurately predicting the characteristics of the WPSH and water vapor transport anomalies originating from the Bay of Bengal. Moreover, the model is unable to predict the influence of the Eurasian (EU) wave train on WR beyond the lead time of 20 days. These issues can be ascribed (attributed) to the absence of a distinct weather-scale boundary (front) for WR in the model, rendering the forecast of WR inherently more challenging compared to that of FR.