Jingyuan Luo

Urban surface water networks are closely associated with the urban land surface thermal environment, but the spatial non-stationarity of their relationships with land surface temperature (LST) remains insufficiently understood. This study examined relationships between water network indicators and LST in Fuzhou. Following scale-sensitivity analysis to select the analytical grid scale, this study used global correlation and regression analyses to identify key indicators and applied multiscale geographically weighted regression (MGWR) to examine spatial non-stationarity in their relationships with LST. Based on the selected 150 m grid scale, the results showed the following: (1) Water coverage ratio (WCR), water body shape index (SI), and water network density (D) were retained as key explanatory variables, with WCR showing the strongest negative association with LST. (2) The relationships between water network indicators and LST showed spatial non-stationarity. WCR showed the most stable negative local association with LST, especially in built-up areas with higher background heat loads, whereas the local associations of SI and D were more context-dependent. These findings suggest that thermal-environment-oriented planning for urban surface water networks should prioritize water coverage while considering water network structure, water body morphology, the surrounding built environment and land cover conditions. These associations can inform spatially differentiated planning of urban surface water networks in comparable subtropical coastal cities.

In the context of contemporary urbanization and urban renewal, blue–green spaces (BGS) serve not only as vital open spaces and regional characteristic scenery but also as critical ecological infrastructure underpinning urban sustainability