Unmasking human-subsidized resilience through hydrological drivers of water use efficiency during compound droughts across the North China Plain

Provinces in the North China Plain (NCP). This study quantifies water use efficiency (WUE) in NCP provinces from MODIS datasets, and analyzed its resilience and vulnerability to meteorological (Standardized Precipitation Evapotranspiration Index, SPEI), agricultural (Standardized Soil Moisture Index, SSMI), hydrological (Standardized Water Availability Index, SWAI), and groundwater (GRACE-based Groundwater Drought Index, GGDI) droughts. Utilizing the Bayesian generalized additive models (BGAM) and reduced-BGAM frameworks at 1-, 3-, 6-, and 12-month scales, this study disentangles the multi-type drought nexus to isolate specific hydrological drivers and physical thresholds governing crop physiological stress. Results show non-linear propagation of drought, heavily altered by interventions temporarily decouple crop physiology from meteorological deficits. At 1-month scale, SSMI and SWAI explain 40–60% of WUE variance across several irrigated districts, creating anthropogenic masking effect. However, the compound drought assessment indicates the resilience is strictly conditional on groundwater dynamics. In deep groundwater depression cones in Hebei, we identified rigid physiological points that triggers a steep and nonlinear collapse in WUE anomalies. Positive proportions of drought to improve WUE range from 2.5% to 69.6%. At 6-month scale, both the positive proportion and mean slope increase significantly. Overall, transient WUE stability relies on continuous anthropogenic surface subsidies, long-term resilience of agro-ecosystem is primarily constrained by deep groundwater storage and structural integrity of the underlying aquifer. • We used BGAM to evaluate impact of compound droughts on water use efficiency (WUE). • BGAM compound drought analysis unmasks true drivers of ecosystem resilience. • Surface irrigation creates a temporary, human subsidized resilience in WUE. • GGDI explain the highest variance percentage (∼50%) in SPEI–SSMI–SWAI–GGDI nexus. • Breaching critical groundwater thresholds triggers rapid structural failure.