Landscapes and planning determine the spatiotemporal heterogeneity of urban soil environmental capacity

Soil environmental capacity (SEC), a key indicator of the soil's ability to sequester pollutants and maintain ecosystem services, is vital for effective urban environmental management. However, traditional methods for evaluating SEC often overlook the complex spatial heterogeneity of environmental drivers. This study introduced an integrated approach combining eXtreme Gradient Boosting, SHapley Additive exPlanations, and geographically weighted regression (GWR) to compare the spatiotemporal dynamics and key drivers of SEC in Zhenhai and Tianjin, China, representing distinct landscape and socioeconomic contexts. Results revealed steep SEC gradients in Zhenhai, driven by industrial concentration, and a more uniform distribution in Tianjin, influenced by urbanization. From 1997 to 2021, Tianjin's SEC remained stable, whereas Zhenhai experienced a sharp decline after 2013. The analysis identified land use and cover change (LUCC) as key drivers in Zhenhai, whereas gross domestic product dominated in Tianjin. The GWR results revealed spatial heterogeneity in SEC drivers, with LUCC dominating Zhenhai's urban and industrial zones, while in Tianjin, economic fluctuations had contrasting impacts, reflecting regional differences in economic activities and pollutant dispersion. These findings highlighted the need for tailored soil management strategies to mitigate pollution risks in rapidly urbanizing areas.