Interpreting seasonal droughts over the Yangtze River Basin utilizing anomalies of local-scale atmospheric circulation

Study region: The Yangtze River Basin (YRB), China. Study focus: This study quantifies the contribution of local-scale factors to seasonal droughts over the YRB, with a particular emphasis on specific humidity, vertical velocity, horizontal divergence, and relative vorticity at pressure levels from 100 to 925hPa. Using the 3D Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm, we identified 43 seasonal drought events in the YRB from 1971 to 2022. An Explainable Artificial Intelligence framework was then applied to assess the influence of these local-scale factors on the Standardized Precipitation Index (SPI3) at the grid scale. New hydrological insights for the region: Low-level specific humidity significantly influences drought conditions in the northern middle and lower reaches, while mid-level vertical velocity is the dominant factor in the southern middle and lower reaches. Mid-level specific humidity plays a key role in the upper reaches. The Random Forest model, using these four local factors as predictors, accurately simulates the spatiotemporal evolution of seasonal droughts over the YRB, providing new insights into the mechanisms behind these events.