A GIS-Based Methodology for Multi-Pollutant Analysis of Sentinel-5P TROPOMI Data with Spatial Intersection Masking

Urban air quality monitoring using satellite remote sensing requires consistent, spatially comparable datasets across multiple pollutants. This paper presents an end-to-end, open-source GIS-based methodology for processing Sentinel-5P TROPOMI Level-2 products to generate synchronized multi-pollutant datasets suitable for urban-scale analysis. The workflow integrates automated data acquisition via the Copernicus Data Space Ecosystem OData API, quality-controlled preprocessing using the HARP toolbox and a QGIS-based spatial analysis pipeline developed in Python. The main methodological contribution is а spatial intersections masking approach, in which a daily mask of valid pixels is calculated as a logical intersection of valid observations for carbon monoxide (CO), nitrogen dioxide (NO₂), and sulfur dioxide (SO₂). This ensures that all statistical comparisons between pollutants are derived from spatially coinciding sets of pixels, thereby eliminating location-based sampling biases arising from differences in swath geometry, cloud cover, and gas-specific retrieval sensitivity. The methodology was applied to four consecutive winter seasons (2019–2023) over Sofia, Bulgaria, a city with significant air pollution driven by residential solid-fuel combustion. The resulting synchronized daily database enables temporal trend analysis (Mann-Kendall test), inter-pollutant correlation assessment, and pollution hotspot identification. The workflow relies entirely on open-source software (QGIS, GDAL/OGR, NumPy, HARP) and is designed to be easily adaptable to any urban area and any combination of TROPOMI gas products with minimal parameter adjustments.