Inferring and evaluating satellite-based constraints on NOx emissions estimates in air quality simulations

Abstract

Abstract. Satellite observations of tropospheric NO2 columns can provide top-down observational constraints on emissions estimates of nitrogen oxides (NOx). Mass-balance based methods are often applied for this purpose, but do not isolate near-surface emissions from those aloft, such as lightning emissions. Here, we introduce an inverse modeling framework that couples satellite chemical data assimilation to a chemical transport model and infers satellite-constrained emissions totals using the iterative finite-difference mass-balance method. The approach improves the finite-difference mass-balance inversion by isolating the near-surface emissions increment. We apply the framework to estimate lightning and anthropogenic NOx emissions over the Northern Hemisphere. Using overlapping observations from the Ozone Monitoring Instrument (OMI) and the Tropospheric Monitoring Instrument (TROPOMI), we compare NOx emissions inferences from these satellite instruments, as well as the impacts of emissions changes on modeled NO2 and O3. OMI inferences of anthropogenic emissions consistently lead to larger emissions than TROPOMI inferences, attributed to a low bias in TROPOMI NO2 retrievals. Updated lightning NOx emissions from either satellite improve the chemical transport model’s low tropospheric O3 bias. Combined lightning and anthropogenic updates inferred from satellite observations can improve the model’s ability to represent background and ground-level O3 concentrations, an ongoing policy consideration in the U.S. as domestic and international emissions control strategies evolve.

Inferring and evaluating satellite-based constraints on NOx emissions estimates in air quality simulations

“…Increases in background areas with small NO 2 column totals and subsequent LNO x increases in these areas suggest a low bias in modeled background NO 2 relative to observations from both satellites. A low bias agrees with the findings reported by other model and satellite NO 2 comparisons (Silvern et al, 2019;Qu et al, 2021;Goldberg et al, 2017). The LNO x emissions adjustments inferred here decrease the differences between modeled and satellitederived NO 2 in the upper troposphere and decrease the bias that differences in the upper troposphere can introduce to the subsequent ANO x inversion.…”

“…LNO x satellite-inferred emissions from both satellites increase O 3 and subsequently improve the model's low O 3 bias across all seasons, with the strongest effect in the summer. This suggests a low background NO 2 in our prior simulation, consistent with several studies demonstrating that models underestimate background NO 2 (Goldberg et al, 2017;Qu et al, 2021;Silvern et al, 2019).…”

“…Over background areas, the analysis increments that result from assimilation of observations from both satellites generally agree. The consistency suggests a low bias in modeled background NO 2 concentrations and also agrees with the low bias in CMAQ-modeled free-tropospheric NO 2 reported by Goldberg et al (2017). Such a bias can contribute to the positive analysis increment over background areas.…”