Tracking NO2 and CO pollution hotspots at provincial scale in China with TROPOMI observations and image segmentation method

Satellite measurements of short-lived gaseous pollutants such as nitrogen dioxide (NO₂) and carbon monoxide (CO) provide an unprecedented opportunity to infer global distribution of their emission sources. However, the great differences in the magnitude of anthropogenic emissions have hampered the identification of pollution sources with satellite observations. In this study, we identified NO₂ and CO pollution hotspots in China with oversampled TROPOspheric Monitoring Instrument (TROPOMI) concentration data by utilizing an image segmentation method at provincial scale. We also determined the major emitting sectors associated with the pollution hotspots in each province. Based on NO₂ pollution hotspots, we can find major emission sources, such as ports, airports, industrial and urban areas. In the hotspots of most provinces, Steel plant, Coke plant, and Power plant were identified as the main NO_x emission sources. CO pollution hotspots were closely linked to Steel plant, with additional contributions from Coke plant, Chemical industry, and Non-ferrous metal smelting in certain provinces. The pollution hotspots and their corresponding concentrations yielded a relatively strong correlation with Multi-resolution Emission Inventory for China (MEIC) (For NO₂, the R² was 0.937 in provinces with small land area and 0.654 in others, while for CO, the R² was 0.849 and 0.659, respectively). Our results can provide an observational constraint on the spatial locations of NO_x and CO emission sources for emission inventories.