Detection of hotspot areas using Sentinel-5P and GEMS imagery for evaluating bottom-up emission inventorie

Emissions from power plants contribute significantly to the overall greenhouse gas and air pollutant levels. Satellite observations of compounds such as NO2 and CO can help improve CO2 emission estimates, as proposed in the CORSO (CO2 Monitoring and Verification Support Research on Supplementary Observations) project. One of the key objectives of CORSO is to improve global and local capabilities in using observations of co-emitted species (NO2 and CO) to better estimate anthropogenic CO2 emissions. In this presentation, we will discuss a methodology developed to detect hotspots associated with power plants using NO2 data from the TROPOMI instrument on the Copernicus Sentinel-5P satellite, as well as observations from the Geostationary Environment Monitoring Spectrometer (GEMS) over Eastern and Southeast Asia. The final goal is to improve the accurate detection of hotspot locations to identify missing sources in emission inventories and refine their geolocation. It is important to note that the detectability of the anthropogenic signal from co-emitted species is generally much higher than that of CO2. Various statistical methods have been tested to identify high-probability hotspots in the NO2 tropospheric column densityfrom TROPOMI and GEMS. This includes an exploratory spatial data analysis for cluster detection (Getis-Ord Gi*), which evaluates each spatial variable’s neighborhood to determinewhether its values are significantly higher or lower than those in the surrounding area. The results indicate agreement between the hotspots identified through the Gi* method and the locations of power plants from the literature. These identified hotspot coordinates can be used to enhance the quanification of emissions and address mislocation in power plant emissions withinemission inventories.