Spatiotemporal Variation and Driving Mechanisms of Methane Concentration in Tianjin Based on the DCNM-GTWR Hybrid Model

To examine the spatiotemporal dynamics of methane in Tianjin and their interactions with multiple drivers, this study employed the DCNM and GTWR models to identify determinants of concentration variability. The findings indicate that methane concentrations showed a significant upward trend, with a citywide mean of 1,907.93ppb during 2019–2024 and an average annual increase of 11.94ppb. Concentrations exhibited a clear seasonal cycle—lower in spring and higher in autumn—with the minimum in March and the maximum in October. Elevated concentrations occurred primarily in coastal areas, heavy-industrial zones, and densely populated districts; values in the 1,915–1,930ppb range represented 54.68% of observations. Spatially, summer and autumn displayed higher concentrations in the north and lower in the south, whereas winter showed higher concentrations in the center and lower toward the periphery. DCNM results indicate that methane concentration variability is primarily associated with trends in D-LST, LAI, and the NDVI. The majority of variables exhibit pronounced non-stationarity and seasonal characteristics, substantiating that D-LST and the NDVI play a decisive role in the spatiotemporal variation of methane concentration. The GTWR model achieved a strong fit (R² = 0.822). It identified lagged D-LST and N-LST, NDVI, and prior methane concentration as key predictors; the corresponding coefficients display significant spatial non-stationarity.