Evaluation of the coupled high-resolution atmospheric chemistry model system MECO(n) using in situ and MAX-DOAS NO2 measurements

We present high spatial resolution (up to 2.2 x 2.2 km(2)) simulations focussed over south-west Germany using the online coupled regional atmospheric chemistry model system MECO(n) (MESSy-fied ECHAM and COSMO models nested n times). Numerical simulation of nitrogen dioxide (NO2) surface volume mixing ratios (VMRs) are compared to in situ measurements from a network with 193 locations including background, traffic-adjacent and industrial stations to investigate the model's performance in simulating the spatial and temporal variability of short-lived chemical species. We show that the use of a high-resolution and up-to-date emission inventory is crucial for reproducing the spatial variability and resulted in good agreement with the measured VMRs at the background and industrial locations with an overall bias of less than 10 %. We introduce a computationally efficient approach that simulates diurnal and daily variability in monthly-resolved anthropogenic emissions to resolve the temporal variability of NO2. MAX-DOAS (Multiple AXis Differential Optical Absorption Spectroscopy) measurements performed at Mainz (49.99 degrees N, 8.23 degrees E) were used to evaluate the simulated tropospheric vertical column densities (VCDs) of NO2. We propose a consistent and robust approach to evaluate the vertical distribution of NO2 in the boundary layer by comparing the individual differential slant column densities (dSCDs) at various elevation angles. This approach considers details of the spatial heterogeneity and sensitivity volume of the MAX-DOAS measurements while comparing the measured and simulated dSCDs. The effects of clouds on the agreement between MAX-DOAS measurements and simulations have also been investigated. For low elevation angles (<= 8 degrees), small biases in the range of -14% to C7% and Pearson correlation coefficients in the range of 0.5 to 0.8 were achieved for different azimuth directions in the cloud-free cases, indicating good model performance in the layers close to the surface. Accounting for diurnal and daily variability in the monthly-resolved anthropogenic emissions was found to be crucial for the accurate representation of time series of measured NO2 VMR and dSCDs and is particularly critical when vertical mixing is suppressed, and the atmospheric lifetime of NO2 is relatively long.