Using typhoon data from the Shanghai Typhoon Institute of the China Meteorological Administration, the Japan Meteorological Agency’s annual Pacific decadal oscillation (PDO) index, and NCEP/NCAR reanalysis data from 1951 to 2021, correlation and composite analyses were carried out to study the relationship between the variability among tropical cyclones of different magnitudes affecting South China and the PDO. The results show that there is an obvious out-of-phase relationship between the proportion of tropical cyclones reaching a typhoon-level intensity or above in South China and the PDO index. When the PDO is in a cold (warm) phase, the sea surface temperature in the eastern and central equatorial Pacific is cold (warm), similar to the eastern Pacific La Niña (El Niño) phenomenon, and the SST in the eastern and western tropical Pacific Ocean shows a negative (positive) gradient; the subtropical high in the western Pacific Ocean is weaker (stronger) than normal, with the western ridge point to the east (west), and the 500 hPa geopotential height in the South China Sea and the area east of the Philippines is weaker (stronger), which is conducive to (unfavorable to) the formation of a monsoon trough; and the westerly (easterly) winds at high altitudes and the southwesterly (northeasterly) winds at low altitudes from the South China Sea to the Philippines are abnormally strong, and a positive (negative) vorticity at low altitudes, a low (high) sea level pressure, and strong (weak) convection are shown. These conditions are favorable (unfavorable) for the intensification of typhoons affecting South China, and as a result, the number of tropical cyclones reaching the level of typhoons or above account for a greater (smaller) proportion of those affecting South China.

Accurate prediction of tropical cyclone‐induced extreme rainfall (TCER) is of utmost importance for disaster mitigation in coastal regions. However, it remains a formidable challenge due to the intricate interactions among multi‐scale meteorological factors and the inherent data imbalances. This study presented an interpretable machine learning (ML) framework aimed at predicting both the occurrence and magnitude of TCER in Guangxi (GX), China. The framework integrated three supervised learning algorithms, namely XGBoost, Random Forest, and AdaBoost, along with feature selection techniques and an explainable method. A total of 202 experiments were conducted to comprehensively evaluate the framework's performance. Genetic Algorithm (GA) optimization and Shapley additive explanations (SHAP) were utilized to identify the optimal subsets of features and accurately quantify the contributions of each variable. Results showed that the optimized XGBoost model exhibited outstanding performance, integrating 18 predictors across dynamic, thermodynamic, moisture, and precursor variables, with a Threat Score of 0.41 for the classification of TCER occurrence and a Threat Score of 0.49 for the regression of rainfall magnitude, outperforming the TIGGE ensemble data in case studies of typhoons Chaba (2022) and Doksuri (2023). SHAP analysis revealed that Distance to Track is the most crucial factor for TCER occurrence. It also unveiled the existence of nonlinear interactions. For instance, an increase in vertical wind shear, favorable thermal conditions, ascending motion, and subtropical high activity can substantially amplify the likelihood of TCER when coupled with low‐level humidity accumulation. Moreover, time‐lagged variables and time‐evolution variables demonstrated their ability to capture the precursor signals of TCER events, like humidity accumulation, circulation adjustment, and typhoon intensity changes, highlighting the model's effectiveness in considering these factors. Therefore, this study showcases the great potential of ML in enhancing TCER prediction while maintaining physical interpretability. Additionally, it offers a valuable reference for addressing imbalance issues in similar research fields.

This study investigates the main climatological features of extreme precipitation (TCER) induced by tropical cyclones (TCs) affecting Guangxi (GX), South China using multiple datasets and a 99th percentile threshold during 1981–2020, with an emphasis on the rainfall diversities of different high-impact TC groups and their associated mechanisms. Results show that there are large regional differences and a seasonal imbalance in the climatological features of TCER in GX. In summer (fall), TCs with TCER events primarily move northward or eastward (northwestward or westward), namely, S-NWTCs and S-ETCs (F-WTCs and F-NWTCs). The rainfall centers exhibit asymmetrical features with S-NWTCs and F-NWTCs located in the northeast quadrant, but S-ETCs and F-WTCs in the southwest and northeast quadrants, respectively. Comparisons of atmospheric circulations and environmental factors indicate that the intense rainfall of F-WTCs is mainly attributed to the trough–TC interaction, which is accompanied by stronger upper-level westerly jet and cold air intrusion, thus increasing baroclinic energy and uplifting for the strongest rainfall among these four groups. This interaction is absent for other groups due to a greater South Asian high and western North Pacific subtropical high. Instead, the increased rainfall in S-NWTCs and F-NWTCs can mainly be attributed to the stronger low-level southwesterly jet, which, in combination with low-level warm advection and convergence induced by land–sea friction, promotes the release of latent heat through moisture condensation. S-ETCs differ from S-NWTCs and F-NWTCs in that moisture convergence is weaker due to the much-weakened TC circulation.

The southern Tibetan Plateau (TP) is snow covered during cold season but exhibits faster snow melting in early summer. Using in situ observations and improved satellite-derived data, the present study indicates that the snow depth (SD) over the southern TP exhibits distinction characteristics between late spring (i.e., P1: April 16th–May 15th) and early summer (i.e., P2: May 16th–June 14th). In terms of climate states, the snow melting rate over the southern TP in P2 is faster than that in P1. The acceleration of snow melting during P2 is mainly found over high elevation areas caused by the increase of local air temperature. Diagnoses of the thermodynamic equation further demonstrate that the warming over the southern TP during the two periods is mainly attributed to the meridional temperature advection and diabatic heating in situ. On the interannual time scale, the SD over the southern TP is closely related to diabatic heating over South Asia. During P1, the diabatic cooling from the southern Bay of Bengal eastward to the western South China Sea suppresses convection over the Bay of Bengal and southern TP and has resulted in an upper-level anomalous cyclone and cold temperature anomalies from the surface to 200 hPa over the southern TP, favoring the above-normal SD over the southern TP. On the other hand, SD over the southern TP in P2 is closely related to diabatic cooling over the northern Indochina Peninsula and diabatic heating over the southern China. But we could not prove that these diabatic heating anomalies can affect the SD over the southern TP by modulating local surface air temperature. This may be limited by the quality of the data and the simulation capability of the simple model.