Exploiting OMI NO satellite observations to infer fossil-fuel CO emissions from U.S. megacities.

Fossil-fuel CO emissions and their trends in eight U.S. megacities during 2006-2017 are inferred by combining satellite-derived NO emissions with bottom-up city-specific NO-to-CO emission ratios. A statistical model is fit to a collection NO plumes observed from the Ozone Monitoring Instrument (OMI), and is used to calculate top-down NO emissions. Decreases in OMI-derived NO emissions are observed across the eight cities from 2006 to 2017 (-17% in Miami to -58% in Los Angeles), and are generally consistent with long-term trends of bottom-up inventories (-25% in Miami to -49% in Los Angeles), but there are some interannual discrepancies. City-specific NO-to-CO emission ratios, used to calculate inferred CO, are estimated through annual bottom-up inventories of NO and CO emissions disaggregated to 1 × 1 km resolution. Over the study period, NO-to-CO emission ratios have decreased by ~40% nationwide (-24% to -51% for our studied cities), which is attributed to a faster reduction in NO when compared to CO due to policy regulations and fuel type shifts. Combining top-down NO emissions and bottom-up NO-to-CO emission ratios, annual fossil-fuel CO emissions are derived. Inferred OMI-based top-down CO emissions trends vary between +7% in Dallas to -31% in Phoenix. For 2017, we report annual fossil-fuel CO emissions to be: Los Angeles 113 ± 49 Tg/yr; New York City 144 ± 62 Tg/yr; and Chicago 55 ± 24 Tg/yr. A study in the Los Angeles area, using independent methods, reported a 2013-2016 average CO emissions rate of 104 Tg/yr and 120 Tg/yr, which suggests that the CO emissions from our method are in good agreement with other studies' top-down estimates. We anticipate future remote sensing instruments - with better spatial and temporal resolution - will better constrain the NO-to-CO ratio and reduce the uncertainty in our method.