Spatiotemporal Analysis of Air Pollution Using GIS and Geographically Weighted Regression: A Case Study in the Eastern Marmara Basin, Türkiye

Natural hazards and environmental risks, including air pollution, pose significant threats to human health and urban sustainability, underscoring the need for comprehensive spatial databases to support effective monitoring and management. This study presents a framework for constructing a district-level environmental hazard database for PM₁₀ (particulate matter with an aerodynamic diameter of 10 µm or less) concentrations in Eastern Marmara Basin, Türkiye, from 2014 to 2024, highlighting the challenges posed by high-volume, heterogeneous data. Meteorological, land-use, and topographic variables were integrated into a Geographic Information Systems (GIS) framework to create a spatially explicit database for analysis. Geographically Weighted Regression (GWR) was applied to assess the spatially varying influences of these variables on PM₁₀ levels. Results indicate that meteorological factors are the dominant drivers of PM₁₀ variability, while land use and topography exert moderate but locally significant effects. Local R² values reveal spatial heterogeneity and identify hotspots where pollutant levels are most sensitive to environmental and anthropogenic drivers. This approach demonstrates a replicable GIS-based methodology for building environmental hazard databases to support risk assessment and evidence-based decision-making.