How does the local-scale relationship between ecosystem services and human wellbeing vary across broad regions?

Recent studies have shown that the relationships between ecosystem services (ES) and human wellbeing (HWB) can be positive, negative, or non-existent, but the underlying causes and processes remain unclear. This study aimed to investigate how and why the local level ES-HWB relationship would change geospatially and manifest on broad regions. Using data for Mainland China, we first calculated seven ES and Human Development Index (an indicator of HWB), then used geographically weighted regression and cluster analysis to quantify the county-level ES-HWB relationship, and finally adopted Wilcoxon test and random forest to investigate key influencing factors. We found that (1) the local-scale relationship between ES and HWB exhibited a great deal of spatial heterogeneity, varying from positive to negative or no correlations across broad regions; (2) the varying relationships merged spatially into three general types of regions: Positive Correlation-Dominant Region, Negative Correlation-Dominant Region, and No Correlation-Dominant Region; and (3) the variations and spatial patterns of the ES-HWB relationships were influenced by a number of social-ecological factors (e.g., population density and land cover compositions), and generally corresponded to different stages of land use transition and socioeconomic development: a positive ES-HWB relationship was found mainly in socioeconomically underdeveloped (rural or agricultural) regions with low ES production levels; a negative ES-HWB relationship occurred mostly in intermediately developed regions with abundant non-food ES; and ES and HWB had no relationships in socioeconomically well-developed (intensive agriculture/urbanized) societies with ample provisioning ES. These findings suggest that neither the "environmentalist's paradox" nor the "environmentalist's expectation" adequately accounts for the complexity of the ES-HWB relationship.