Application of dynamic spatiotemporal modeling to predict urban traffic–related air pollution changes.

Traffic-related urban air pollution is a pressing concern in Tehran, Iran, with severe health implications. This study aimed to create a dynamic spatiotemporal model to predict changes in urban traffic-related air pollution in Tehran using a land use regression (LUR) model. Two datasets were employed to model the spatiotemporal distribution of gaseous traffic-related pollutants—sulfur dioxide (SO₂), nitrogen dioxide (NO₂), and carbon monoxide (CO). The first dataset incorporated remote sensing data, including land surface temperature (LST), the normalized difference vegetation index (NDVI), apparent thermal inertia (ATI), population density, altitude, land use, road density, road length, and distance to highways. The second dataset excluded remote sensing data, relying solely on population density, altitude, land use, road density, road length, and distance to highways. The LUR model was constructed using both datasets at three different buffer distances: 250, 500, and 1000 m. Evaluation based on the R² index revealed that the 1000-m buffer distance achieved the highest accuracy. Without remote sensing data, R² values for CO, NO₂, and SO₂ pollutants were respectively spring (0.77, 0.79, 0.51), summer (0.59, 0.71, 0.59), and winter (0.41, 0.52, 0.59). With remote sensing data, R² values were respectively spring (0.82, 0.84, 0.74), summer (0.72, 0.87, 0.62), and winter (0.53, 0.59, 0.72). Incorporating remote sensing data notably improved the accuracy of modeling CO, NO₂, and SO₂ during all three seasons. The central, southern, and southeastern regions of Tehran consistently exhibited the highest pollutant concentrations throughout the year, while the northern areas maintained comparatively better air quality. [ABSTRACT FROM AUTHOR]