Spatial heterogeneity in the nexus of blue-green infrastructure, urban heat, and population distribution within Hangzhou

To mitigate urban heat, planners must understand the interdependencies between blue-green infrastructure (BGI), land surface temperature (LST), and population distribution. However, previous studies have rarely integrated these three elements within a single, multi-scale analytical framework. This study investigates these complex relationships in Hangzhou across various urban functional areas (UFAs), employing an innovative Python Geographical Random Forest (PyGRF) model to capture non-linear dynamics missed by traditional methods. Our city-wide macro-analysis found BGI's landscape percentage (PLAND) explained over 35% of LST variance. Meso-scale bivariate Moran's I analysis revealed significant heterogeneity: commercial areas showed 'High-LST, High-Population' clusters, while residential areas showed 'Low-LST, High-Population' clusters, proving strategic BGI can decouple heat from population density. The key finding is that a 'one-size-fits-all' BGI approach is ineffective. Our research provides a precise, spatially-explicit blueprint for tailoring BGI design to unique UFA functions, fostering cities that are both thermally resilient and socially vibrant.