Towards monitoring localized CO<sub>2</sub> emissions from space: co-located regional CO<sub>2</sub> and NO<sub>2</sub> enhancements observed by the OCO-2 and S5P satellites

Abstract. Despite its key role for climate change, large uncertainties persist in our knowledge of the anthropogenic emissions of carbon dioxide (CO2) and no global observing system exists allowing to monitor emissions from localized CO2 sources with sufficient accuracy. The Orbiting Carbon Observatory-2 (OCO-2) satellite can retrieve the column-average dry-air mole fractions of CO2 (XCO2). However, regional column-average enhancements of individual point sources are usually small compared to the background concentration and its natural variability. This makes the unambiguous identification and quantification of anthropogenic emission plume signals challenging. NO2 is co-emitted with CO2 when fossil fuels are combusted at high temperatures. It has a short lifetime of the order of hours so that NO2 columns often exceed background levels by orders of magnitude near sources making it a suitable tracer of recently emitted CO2. Based on six case studies (Moscow, Russia; Lipetsk, Russia; Baghdad, Iraq; Medupi and Matimba power plants, South Africa; Australian wildfires; and Nanjing, China), we demonstrate the usefulness of simultaneous satellite observations of NO2 and the column-average dry-air mole fraction of CO2 (XCO2). For this purpose, we analyze co-located regional enhancements of XCO2 observed by OCO-2 and NO2 observed by the Sentinel-5 Precursor (S5P) satellite and estimate the CO2 plume's cross-sectional fluxes. We take advantage of the nearly simultaneous NO2 measurements with S5P's wide swath by identifying the source of the observed XCO2 enhancements, excluding interference with remote upwind sources, allowing to adjust the wind direction, and by constraining the shape of the CO2 plumes. We compare the inferred cross-sectional fluxes with the Emissions Database for Global Atmospheric Research (EDGAR), the Open-Data Inventory for Anthropogenic Carbon dioxide (ODIAC), and, in the case of the Australian wildfires, with the Global Fire Emissions Database (GFED). The inferred cross-sectional fluxes range from 32 Mt CO2/a to 158 Mt CO2/a with uncertainties (1σ) between 23 % and 72 %. For the majority of analyzed emission sources, the estimated cross-sectional fluxes agree within their uncertainty with either EDGAR or ODIAC or lie in between them. We assess the contribution of multiple sources of uncertainty and find that the dominating contributions are related to the computation of the effective wind speed normal to the plume's cross-section. The planned European Copernicus anthropogenic CO2 monitoring mission (CO2M) will not only provide precise measurements with high spatial resolution but also imaging capabilities with a wider swath of simultaneous XCO2 and NO2 observations. Such a mission, in particular as a constellation of satellites, will deliver CO2 emission estimates from localized sources at an unprecedented frequency and level of accuracy.