Linking agricultural drought to compound hot–dry risk: A process-based compound index for crop and water management in the North China Plain

Against the backdrop of global warming, compound drought–heatwave events have become more frequent, severe, and spatially extensive, posing substantial threats to agricultural production, water-resource security, and ecosystem stability. Existing studies have largely focused on the superimposed identification of meteorological drought and high-temperature anomalies, while paying insufficient attention to agricultural drought as a key process linking meteorological moisture deficits with surface disaster responses; limitations also remain in the integrated characterization of multidimensional hydrothermal anomalies, the depiction of event evolution, and the identification of phase-dependent drivers. To address these gaps, this study focuses on the North China Plain and adopts the Standardized Precipitation Evapotranspiration Index (SPEI), Standardized Soil Moisture Index (SSI), and Standardized Temperature Index (STI) to represent meteorological drought, agricultural drought, and high-temperature anomalies, respectively; based on a unified probability transformation, a Copula-based Compound Hot–Dry Index (CHDI) is developed and combined with typical historical-event validation, Gross Primary Productivity (GPP) response assessment, Mann–Kendall trend analysis, improved three-threshold run theory, and a machine-learning–SHAP (SHapley Additive exPlanations) framework to systematically reveal the spatiotemporal evolution and phase-dependent driving mechanisms of compound hot–dry events during 1980–2022. The results show that CHDI can accurately identify the occurrence timing, spatial extent, and month-to-month evolution of typical events, and exhibits stronger consistency with anomalies in agricultural vegetation productivity; compared with existing indices, CHDI shows higher correlations across most pixels, key growing-season months, and short-term lag scales, and produces a clearer severity–GPP response gradient. Over the past four decades, compound hot–dry conditions in the North China Plain have intensified overall, with CHDI decreasing at a rate of approximately −0.111 per decade, the spatial extent of high-grade events continuously expanding, and the most pronounced aggravation occurring in the southern and eastern regions; event frequency peaks in autumn, while intensity generally exhibits a south-strong and north-weak pattern. The dominant factors gradually shifted from precipitation and runoff in the early period toward relative humidity and evaporative demand. Overall, CHDI enhances the integrated characterization of compound hot–dry events and the identification of their impacts on agricultural vegetation, providing scientific support for monitoring and early warning of compound extremes, agricultural risk assessment, and regional adaptive management.