We evaluated the nation-wide anthropogenic NOX and SO2 emissions in the Korean National Emissions Inventory, named the Clean Air Policy Support System 2016, based on three-dimensional air quality model simulation. The air quality modeling system comprised of WRF (Weather Research and Forecasting), SMOKE (Sparse Matrix Operator Kernel Emissions), CMAQ (Community Multiscale Air Quality) predicted NO2 and SO2 levels at a 9-km horizontal grid resolution during 2016. Normalized mean biases of simulated provincial annual mean NO2 and SO2 concentrations ranged from −41.2∼11.4%, and −46.7%∼ 133.1%, respectively. Over-prediction of the annual mean NO2 concentrations were observed mostly from metropolitan areas such as Seoul (10%), Incheon (11%), and Busan (11%). The NO2 level in Gangwon was significantly under-estimated by −41%. Annual mean SO2 concentrations were over-predicted in Ulsan, Chungnam, and Jeonnam by 118%, 133%, and 46% in which point source emissions amount to ∼70% of the provincial totals. It is noted that temporal allocation of the bottom-up emissions is associated with the model biases of monthly mean NO2 and SO2 concentrations. When the bottom-up emissions are revised based on the ratios between the observed and modeled concentrations for each province, the national total emissions have increased by 14% for NOX and decreased by 18% for SO2 for the year.
