Top-down downscaling of a global emission inventory in the area of Northern China

<p>This study develops top-down methods to downscale a global emissions inventory provided by the Copernicus Atmosphere Monitoring Service (CAMS) from a resolution of 10 kilometers to a resolution of 1 kilometer in the area of Northern China. Information extracted from various high-resolution proxies is used as weight factors to distribute the original emissions in 10x10 km grids to emissions in 1x1 km grids. Among five lumped emissions sectors, three of them (transportation, residential and agriculture) are area sources while two of them (industrial and energy) are point sources. For emissions from area sources, the original emissions are first bilinearly interpolated on defined 1x1 km grids in order to smooth spatial distribution of emissions. Correspondingly, the point source emissions are conservatively converted into 1x1 km grids. To downscale the emissions of the transportation sector, road maps including motorways and railways from OpenStreetMap are processed to derive total road length in every 1x1 km grid, which is summed up by weights of different road types. Cropland fraction is used to weight agricultural emissions of 1x1 km grids. Population is taken as a proxy for downscaling of residential emissions. Emissions of energy sector (mainly from power plants) are downscaled based on annual emissions of nitrogen dioxide from individual power plants due to their good correlation with annual emissions of other pollutants. Accuracy of energy sector downscaling is depending on how many power plants are taken into account. Downscaling of industrial sector emissions takes population in industrial area as proxy (weight factor); several proxies are needed here to generate 1x1 km grids covered by industrial area with defined properties. Weighted values of 1x1 km grids are rescaled to conserve the total emission in area of 10x10km grids. Two comparative simulations driven by the CAMS emissions (10 km) and the downscaled emissions (1 km) are performed to test the sensitivity of simulated pollutants to emission resolutions using model of Weather Research and Forecast coupled with Chemistry (WRF-Chem).</p>