Vine copula regional dry-wet evaluation index incorporating baseflow and XGBoost-SHAP early warning system in Huzhou, China

Study region: Huzhou City, Southeastern China. Study focus: This study applied the Lyne-Hollick digital filter to separate baseflow and compute the baseflow index (BFI). We then used a three-dimensional vine copula to couple BFI with precipitation and evaporation, constructing a comprehensive dry-wet index (CDI). We further developed a monthly dry-wet early warning model using XGBoost and employed the SHapley Additive exPlanations (SHAP) framework to interpret the driving mechanisms. Results revealed a notable drying tendency at four of the five stations. Despite the spatially heterogeneous trend pattern, wet events across the basin consistently coincided with high baseflow periods during the rainy season. The early warning model achieved an overall accuracy of 63.5–76.0%, while SHAP analysis indicated spatial heterogeneity in the dominant driving factors across sub-regions. New hydrological insights for the region: The CDI exhibits strong correlations with the standardized precipitation index (SPI), standardized evapotranspiration index (SEI), and standardized baseflow index (SBI) at multiple temporal scales. Dry-wet conditions in the basins have shifted from seasonal alternation toward more persistent states: dry periods are concentrated mainly in winter (December to February), and wet spells largely coincide with high baseflow during the rainy season. These findings provide a scientific basis for upgrading the integrated drought and flood early warning system, optimizing water resource allocation, and formulating region-specific disaster prevention strategies in Huzhou City.