This paper examines the creation of fine resolution maps at 100 m x 100 m resolution using statistical downscaling for the area of Prague, as a case study. This Czech city was selected due to the fine resolution proxy data available for this city. The reference downscaling methodology used is the linear regression and the interpolation of its residuals by the area-to-point kriging. Next to this, several other methods of statistical downscaling have been also executed. The results of different downscaling methods have been compared mutually and against the data from the monitoring stations of Prague, separately for urban background and traffic areas. The downscaled maps in 100 m x 100 m resolution have been constructed for the area of Prague for three pollutants, namely for NO2, PM10 and PM2.5. Several methods of the statistical downscaling have been compared mutually and against the data from the monitoring stations. In general, the best results are given by the linear regression and the interpolation of its residuals, either by the area-to-point kriging or the bilinear interpolation. In the maps, one can see overall realistic spatial patterns, the main roads in Prague are visible through higher air pollution levels. This is distinct especially for NO2, while for PM10 and PM2.5 the differences between road increments and urban background are smaller as would be expected. The results of the case study for Prague have proven the usefulness of the statistical downscaling for the air quality mapping, especially for NO2. In addition, the population exposure estimates based on the downscaled mapping results have been also calculated.

Framework Partnership Agreement No. OCP/EEA/CAS/21/005-ETC/HE Work programme 2022