Comparison of regional chemistry-modelled NO2 tropospheric columns and profiles with TROPOMI observations and 4-azimuth MAX-DOAS measurements

Regional chemistry-transport models typically simulate the physical and chemical state of the atmosphere at a high spatial resolution, e.g. of less than 7 km. At this relatively high spatial resolution, air quality and relevant processes within cities can be assessed to facilitate strategic mitigation planning. Comparison of regional models with satellite and ground-based observations helps validate the models and evaluate emission inventories, as well as satellite retrieval algorithms. For example, an underestimation of atmospheric trace gases (like often found for NO₂) by satellite observations can be improved by providing high-resolution input fields from regional models.

MECO(n), a global-to-regional chemistry climate modeling system, in which the finer resolved domains receive their initial and boundary conditions on-line from the next coarser model instance, was set-up with Germany as focus. 1-way nested MECO(3)  simulations were performed for May 2018 with spatial resolution up to ~2.2 km × 2.2 km in the finest domain. Model simulations accounting separately for both TNO MACC III and EDGAR 4.3.2 anthropogenic emissions are evaluated against TROPOMI observations. A diurnal factor was applied to road transport emissions to account for their temporal variation. For the comparison with TROPOMI data, we applied a novel method of online sampling of model fields along the satellite overpass by also accounting for the difference in local solar time across the swath width, which can be up to 90 minutes. Modified airmass factors in the TROPOMI data product, using the model calculated NO₂ a priori profiles and taking into account averaging kernels, resulted in an improved agreement of the spatial pattern of NO₂ vertical column density (VCD) between model and satellite.

NO₂ VCDs over Mainz, calculated using model output at the finest model resolution, were compared against MAX-DOAS observations for the simulation period. Vertical profiles of NO₂ were also retrieved in 4 azimuth directions around Mainz by profile inversion of MAX-DOAS measurements. The temporal (e.g. day-to-day and diurnal) variation of the 3-D NO₂ field derived from the model was evaluated against the MAX-DOAS observations. For the cloud-free days, the model is able to reproduce the temporal development with satisfactory temporal correlation (slope=0.7, r=0.5) of the NO₂ VCDs. For a direct comparison of measured slant column densities of NO₂, height-resolved 2-D box airmass factors were calculated using McArtim (Monte Carlo Atmospheric radiative transfer model) and applied to the modelled trace gas profiles along individual elevation angles of the measurements. This comparison procedure accounts for the complex dependency of the MAX-DOAS column densities on the 3D (vertical and horizontal) trace gas distribution in the measurement direction.