PNNL

Abstract

Fossil fuel carbon dioxide (CO2ff), the main driver of global warming and climate change, is often co-emitted with nitrogen oxides (NOx) and precursors to ground-level ozone from anthropogenic sources like power plants or vehicles. In urban and suburban areas, satellite-based NO2 can be used as a proxy to track the emissions of CO2ff. Because of NO2’s shorter lifetime, urban NO2 plumes are more distinguishable from backgrounds and more sensitive to variations in emissions. However, the combination of these two gases is limited by the asynchrony among NO2 and CO2 monitoring satellites. We used CO2ff simulated by the Weather Research and Forecasting model coupled with Chemistry (WRF-Chem) model to reconcile the tropospheric NO2 vertical column density (VCD) from the Tropospheric Monitoring Instrument (TROPOMI) and column-averaged dry-air mole fractions of carbon dioxide enhancements (?XCO2) from Orbiting Carbon Observatory 3 (OCO-3) Snapshot Area Maps (SAMs) over a multicity area, Washington D.C.-Baltimore (DC-Balt), and a basin city, Mexico City. NO2/CO2ff ratios over DC-Balt are smaller than Mexico City, indicative of stricter emission restrictions, a more combustion-efficient vehicle fleet, and higher combustion efficiency due to lower altitude in DC-Balt. For single-track cases, the spatial correlations between NO2 and ?XCO2 over Mexico City are stronger than DC-Balt because the NO2 and CO2 are mostly trapped in the valley of Mexico City, while DC-Balt is severely affected by distant sources (i.e., US East Coast cities). Using multi-track averaging, spatial correlation coefficients increase with the number of days used for averaging. The correlations reached a maximum when averaging more than 12 continuous images for DC-Balt and more than 10 continuous images for Mexico City. This finding indicates that multi-track averaging using modeled CO2ff as a proxy is helpful to filter the noise in single-track images, to cancel the interference from distant sources, and to magnify correlations between NO2 and CO2ff. Mexico City showed stronger spatial correlations but weaker temporal correlations than DC-Balt due to biomass burning hot spots and large transport errors caused by the trapping effects of the surrounding mountains. Tracking the 20-day moving average of CO2ff emissions using TROPOMI NO2 seems technically feasible, considering the relationship between correlation coefficients and the number of available satellite images.