Evaluating the impact of spatial resolution on tropospheric NO2 column comparisons within urban areas using high-resolution airborne data

NASA deployed an airborne UV/Visible spectrometer, GeoTASO, in May-June 2017 to produce high resolution (approximately 250 × 250 m), gapless NO2 datasets over the western shore of Lake Michigan and over the Los Angeles Basin. Results show that the airborne tropospheric vertical column retrievals compare well with ground-based Pandora spectrometer column NO2 observations (r2=0.91 and slope of 1.03). Apparent disagreements between the two measurements can be sensitive to the coincidence criteria and are often associated with large local variability, including rapid temporal changes and also spatial heterogeneity that may be observed differently by the sunward viewing Pandora observations. The gapless mapping strategy executed during the 2017 GeoTASO flights provides data suitable for averaging to coarser areal resolutions to simulate satellite retrievals. As simulated satellite pixel area increases to values typical of TEMPO, TROPOMI, and OMI, the agreement with Pandora measurements is degraded as localized polluted plumes observed by Pandora are spatially averaged over larger areas (aircraft-to-Pandora slope: TEMPO scale=0.88; TROPOMI scale=0.77; OMI scale=0.57). This behavior suggests that satellite products are representative of individual Pandora observations up to a certain pollution scale that depends on satellite spatial resolution. In these two regions, Pandora and TEMPO or TROPOMI have the potential to compare well up to pollution scales of 30x1015 molecules cm-2. Two publicly available OMI tropospheric NO2 retrievals are both found to be biased low with respect to Pandora observations (NASA V3 Standard Product slope = 0.18 and Berkeley High Resolution Product slope=0.30). However, the agreement improves when higher resolution a priori inputs are used for the tropospheric air mass factor calculation. Overall, this work explores best practices for satellite validation strategies by showing the sensitivity to product spatial resolution and demonstrates how the high spatial resolution NO2 data retrieved from airborne spectrometers, such as GeoTASO, can be used with high temporal resolution surface observations to evaluate the influence of spatial heterogeneity on validation results.