Informing intra-city emission characteristics using satellite observations of CO2 and co-emitted species

Air pollutants (CO, NOx) are co-emitted with CO2 from many sources of combustion. While the joint use of satellite-based CO2 and co-emitted species has been advocated to better understand city-level emissions, few have examined the spatial allocation and sector attribution within cities. In particular, emission ratios (ERs) between tracers are not well constrained as they vary across combustion processes, sectors, cities, and years. Factors such as meteorological and chemical conditions, biospheric fluxes, and mismatches in the retrievals, are not fully accounted for in some prior tracer-tracer analyses. Our goal is to estimate ERs and inform sector characteristics for cities by combining spatially resolved XCO2 from OCO-3 with XCO and tropospheric NO2 (tNO2) data from TROPOMI. To interpret tNO2 data efficiently, we implement a simplified representation of the nonlinear NOx chemistry to an existing Lagrangian transport model and evaluate the system over several U.S. power plants using "true" emissions from EPA. For city cases (e.g., Baotou, China and Phoenix, USA), the model is coupled with a chemical inversion system (LETKF) to optimize EDGAR emissions. The model-data comparison of tNO2 reveals possible biases in the emission location and modeled wind field that can guide CO2 inversions. By further zooming into an urban area, we find that the combustion efficiency related to heavy industrial activities in Shanghai is distinguishable from that in Los Angeles, as inferred from observed CO:CO2 enhancement ratios. We also attempt to constrain the spatially-resolved ERs of NOx:CO2 for a few cities and investigate their spatiotemporal variability for sector-specific signals. Our analyses may shed light on emission characteristics over more cities around the globe, especially when data from future satellite missions (e.g., CO2M, GeoCarb) become available.