Spatiotemporal Patterns and Associations of Ecological Quality and Carbon Storage Across Three Major Agricultural Provinces in Central China

Ecological quality, vegetation productivity, and carbon storage represent different ecosystem properties and may respond differently to land-use change. This study assessed their spatiotemporal patterns and spatial relationships across Hubei, Henan, and Anhui, three major agricultural provinces in central China. Ecological quality was assessed using the Remote Sensing Ecological Index (RSEI), vegetation productivity was estimated as net primary productivity (NPP) using the Carnegie–Ames–Stanford Approach (CASA) model, and carbon storage was estimated using the Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) model. Spearman correlation, GeoDetector, and a descriptive coordination index were used to compare their spatial relationships. Ecological quality declined from 2000 to 2010 and recovered only modestly thereafter, with persistently lower values in the eastern and northern plains than in the western and southern mountains. From 1990 to 2020, built-up land expanded by 112.03%, while model-estimated carbon storage declined by 34.59 × 106 t. RSEI and NPP were predominantly positively correlated, but correlations varied substantially among provinces, ecological-quality classes, and land-use types. RSEI had the greatest explanatory power for spatial variation in carbon storage among the composite factors (q = 0.420–0.467), and NDVI was the strongest RSEI component. GeoDetector identified bivariate and nonlinear enhancement, indicating that paired factors explained more spatial variation than individual factors. Moderate and primary coordination together accounted for more than 99% of valid pixels, while severe imbalance declined from 0.88% to 0.44%. These findings show that the three indicators are spatially associated but not interchangeable. Together, the results support protecting forested mountain areas, limiting built-up expansion and cropland loss in plains and urban fringes, and using multiple indicators to identify areas requiring local assessment.