Precipitation-Related Heterogeneity in the Association Between Urban Vegetation Greenness and PM2.5: Evidence from Humid and Arid Regions in China

Precipitation may condition the relationship between urban vegetation and air quality across contrasting climates. Using a balanced panel of 14 cities in China’s Pearl River Delta and Hexi Corridor from 2015 to 2023, we examined associations between the Normalized Difference Vegetation Index (NDVI) and annual mean fine particulate matter (PM2.5). Two-way fixed-effects, interaction, and spatial panel models were estimated. Higher prefecture-scale NDVI was associated with lower PM2.5, and the positive NDVI–precipitation interaction indicated a weaker negative association at higher precipitation. Marginal estimates were significant at the observed 10th precipitation percentile but not at the median or 90th percentile. When NDVI and precipitation were recomputed over built-up land, however, the interaction was no longer statistically significant. In the preferred Spatial Durbin Model, the direct NDVI effect was insignificant, whereas the indirect and total effects were significantly negative; these patterns persisted under two alternative spatial-weight matrices. The findings reveal precipitation-related heterogeneity in the main specification, but sensitivity to measurement scale and the observational design preclude causal or physical pollution-removal interpretations. Vegetation planning should therefore account for local precipitation conditions, spatial measurement, and regional air-quality linkages.