Investigations of comparison uncertainties for airborne validation of air quality satellite products

When validating atmospheric satellite observations, several error sources must be taken into account: the uncertainties of the satellite products, the uncertainties of the reference measurements, and the representativity of the latter with respect to the investigated satellite pixels. Compared to static ground-based reference measurements, airborne observations reduce the spatial component of the representativity error. Recent airborne campaigns indicate a remaining low-bias for TROPOMI tropospheric NO2 VCDs above polluted areas. This bias has been attributed in particular to wrong assumptions on the NO2 profiles in the satellite products. 

In the context of the RAMOS and SVANTE projects, we started regular continuous mapping of the NO2 tropospheric VCDs above Bucharest and Berlin, respectively. Both activities make use of compact whiskbroom imagers, namely SWING. In Bucharest, we also measure the profiles of NO2 and of aerosols from the aircraft and perform car-based DOAS measurements of tropospheric NO2 underneath the aircraft. We study the error budgets of the validation of the TROPOMI tropospheric NO2 VCD product in these two situations. We quantify the added values of the ancillary observations in Bucharest and assess the temporal component of the representativity error. Given the time duration of a scientific flight, several configurations are possible for our whiskbroom observations, and it may be useful to undersample satellite pixels to cover a large area. This work is therefore also useful to optimize the flight patterns and information content of future validation flights.