Assessing drought resilience and identifying its driving factors in Loess Plateau grain production systems
Аннотация
Amidst escalating drought risks under climate change, enhancing resilience in grain production systems is critical. Yet few studies have investigated the characteristics and underlying mechanisms of resilience responses to drought in these systems at a macro scale. To narrow this gap, this study establishes a novel disturbance–response-based framework using resilience curves to overcome limitations in static indicator-based assessments. County-level drought resilience across China’s ecologically fragile Loess Plateau was assessed using spatial autocorrelation analysis and Bayesian hierarchical modeling with irrigation-stratified random slopes to identify spatial patterns and their drivers. Results show that: (1) County-level drought resilience predominantly clusters within the moderate-to-high range (mean DRGP = 233.5), where high-resilience systems show lower drought sensitivity, and resilience-level differentials in yield loss scale exponentially with drought intensification; (2) Spatially, drought resilience manifests a steep west-to-east gradient. Hotspots concentrate in northwestern irrigation districts and southwestern alpine foothills, whereas coldspots prevail in the northeastern agro-pastoral ecotone. Significant positive spatial autocorrelation substantiates this pronounced spatial differentiation (Moran’s I = 0.50, p < 0.01); (3) Irrigation infrastructure emerges as a core effect modifier, reversing the directional associations of nine out of eleven drivers. In irrigated districts, annual precipitation, mean temperature, soil organic carbon density, fertilizer application rate per unit area, cropland proportion, total power of agricultural machinery, and per capita disposable income of rural households are positively associated with drought resilience, whereas grassland proportion and cropland slope are negatively associated. These associations reverse in non-irrigated areas. These findings indicate that drought resilience follows irrigation-dependent pathways. Irrigation districts should pursue synergy-based intensification, whereas rainfed regions require damage-mitigation strategies and ecological buffers. These context-specific pathways provide theoretical and practical paradigms for climate-resilient agriculture in global drylands.