The Yangtze River Basin. Understanding hydrometeorological risk from large-scale climate drivers (LSCDs) is often complicated by nonlinear, lagged, and asymmetric teleconnections. This study addresses the challenge of linking combined LSCD phases to extreme precipitation clusters in the monsoon-dominated Yangtze River Basin. We introduce a unified framework analyzing six monthly extreme precipitation indices (1961–2022). By employing Random Forest lag screening, seasonally grouped PIT-calibrated copulas, and scenario-conditioned singular value decomposition (SVD), we identify dominant drivers with optimal lags, quantify tail-specific risks, and map spatial hotspot shifts driven by joint driver phases. Random Forest screening identifies Arctic Oscillation (AO), Niño3.4, and Atlantic Multidecadal Oscillation (AMO) as dominant drivers. Niño3.4 peaks at 1–2 month lags and strengthens lower tail risks, while AO peaks at 4–6 months and intensifies the upper tail. Scenario-conditioned SVD reveals a robust basin-scale dipole. The leading mode accounts for an average of 95.7% of the squared covariance fraction and is driven primarily by AO intensity. Its spatial polarity shifts hotspots between the southeast lowlands and the upper basin based on ENSO phases. These findings link phase, lag, and location to provide a probabilistic framework for diagnosing compound drought or flood risks and supporting resilience-oriented preparedness. • Developed an integrated framework to decode complex hydroclimatic teleconnections. • Measured upper and lower tail risks across six extreme precipitation indices. • The joint-phase SVD reveals a sign-flipping basin-wide dipole in Yangtze River Basin. • The framework enables month-ahead, tail-specific risk guidance for basin hotspots.
Li et al. (Tue,) studied this question.