Near-Automated Estimate of City Nitrogen Oxides Emissions Applied to South and Southeast Asia

Abstract

Cities in South and Southeast Asia are developing rapidly, but lack routine, up-to-date, publicly available inventories of air pollutant precursor emissions such as nitrogen oxides (NOx). This data deficiency can be addressed by deriving city NOx emissions from satellite observations of nitrogen dioxide (NO2) tropospheric column densities. In this approach, the city plume is aligned along a consistent direction using wind rotation and a best-fit Gaussian applied to estimate NOx emissions. Issues that impact success of this approach is subjective selection of the sampling area around the city centre and the Gaussian fit often fails or yields physically implausible parameters. Here, we automate this top-down approach by defining many (54) sampling areas that we test with TROPOspheric Monitoring Instrument (TROPOMI) NO2 observations over 19 cities in South and Southeast Asia. Our approach is efficient, adaptable to a wide range of city sizes, eliminates the need for subjective sampling area selection, and increases success of deriving annual emissions from 40-60% with a single sampling area to 100% (all 19 cities) with 54 sampling areas. Annual emissions range from 16±5 mol s-1 for Yangon (Myanmar) to 118±39 mol s-1 for Dhaka (Bangladesh). A widely used global emissions inventory exhibits large (2-fold) discrepancies for 5 of the 19 cities. The increase in success achieved with our updated approach also enables derivation of monthly emissions, although all 12 months are only obtained for one city (Karachi in Pakistan). Seasonality in the monthly emissions matches seasonality in tropospheric column abundances of NO2 and is greater than can be reasoned with seasonality in anthropogenic activity in cities. This suggests that past annual emissions calculated using observations for a portion of the year or select days may be biased. Further refinement of this approach is needed to fully exploit the large sampling density of high-resolution low-Earth orbiting instruments such as TROPOMI and hourly measurements from geostationary instruments such as Geostationary Environmental Monitoring Spectrometer (GEMS), Tropospheric Emissions: Monitoring of Pollution (TEMPO), and Sentinel-4.