Joint Inversion of Satellite-Based Isoprene and Formaldehyde Observations to Constrain Emissions of Nonmethane Volatile Organic Compounds

Abstract

Reducing uncertainties in nonmethane volatile organic compound (VOC) emissions is critical for understanding air quality and its impacts on human health and climate. In this study, we introduce an iterative joint mass balance inversion framework that synthesizes multiconstituent satellite observations, with a particular focus on CrIS isoprene and OMPS formaldehyde (CH₂O) column measurements. Our analysis targets East Asia during May–June 2016, using the GEOS-Chem model. Our joint inversion improves the model performance for both species compared to the satellite observations, as evident from increases in correlation coefficient (R) from 0.37 to 0.70 for isoprene and from 0.90 to 0.92 for CH₂O, and reductions of normalized mean bias from −79% to −64% for isoprene and from −23% to −16% for CH₂O. In contrast, single-species inversions of CH₂O- and isoprene-alone lead to minor reductions in R for the unobserved species. Cross-validation with airborne observations highlights the downside of single-species inversions, showing that the CH₂O-only inversion tends to overfit CH₂O and degrade the isoprene simulation, which is mitigated by simultaneously using satellite isoprene observations in the joint inversion. However, uncertainties in satellite isoprene observations limit the benefits of the joint inversion in regions with low concentrations, such as South Korea, due to the detection limits of CrIS isoprene retrievals. Despite this limitation, our joint inversion framework efficiently leverages information from multiple data sets to provide a more comprehensive understanding of VOC emissions, enabling future work to incorporate satellite measurements of additional species such as nitrogen oxides and potentially glyoxal.