Evaluating TROPOMI CO Column Observations Using CU AirSOF During BB-FLUX

Jake Rowe, ANYL MS Defense, Volkamer group

"The Biomass Burning Fluxes of Trace Gases and Aerosols (BB-FLUX) field campaign was carried out during the summer of 2018 with the primary goal of quantifying emission fluxes of trace gases by mass balance of actual wildfires. To characterize these fluxes, the University of Colorado Airborne Solar Occultation Flux (CU AirSOF) instrument was flown below biomass burning plumes to measure vertical trace gas columns, such as carbon monoxide (CO, with the first overtone fit from 4214 - 4254 cm-1), along the direct solar beam. The TROPOspheric Monitoring Instrument (TROPOMI) provides measurements of trace-gas maps in the Ultraviolet-Visible (UV-Vis.) and shortwave-IR (SWIR) spectral regimes (e.g. CO from 4277 - 4303 cm-1). A subset of flights during BB-FLUX were dedicated to position the aircraft below smoke plumes while the satellite measures the same scene aloft.

Radiative transfer simulations were conducted to estimate the effect of aerosol multiple scattering in smoke plumes, which is not routinely accounted for in TROPOMI CO retrievals at short-wave infrared wavelengths. Aerosol multiple scattering recovers sensitivity losses in the absence of surface albedo (by up to 80%), but the loss is greatly reduced (5-10%) if surface albedo exceeds 10%. The difference in the temporal and spatial scales, and measurement geometries sampled from the aircraft and satellite are actively addressed by 1) focusing on near coincident case studies, 2) comparing spatial integrals of CO differential vertical column densities (dVCDs) across plumes, and 3) using the FLEXible PARTicle (FLEXPART) dispersion model to correct for different sampling times. CU AirSOF revealed variations in CO dVCD integrals on the order of 24% (up to 37%) over 30 mins (consecutive transect measurements) that characterize changes in emissions on the satellite sub-pixel scale. TROPOMI CO is found to be systematically higher relative to the aircraft by +26% for the operational product (+23% scientific product) without FLEXPART correction; such high-bias is reduced to +8.7% (+5.8%) upon adding the FLEXPART correction. The small but systematic high-bias in TROPOMI CO is most likely due to unaccounted aerosol multiple scattering, atmospheric variability, or a combination thereof."