Abstract

The Korea-United States Air Quality (KORUS-AQ) field study was conducted during May–June 2016 to understand the factors controlling air quality in South Korea. Extensive aircraft and ground network observations from the campaign offer an opportunity to address issues in current air quality models... Show moreThe Korea-United States Air Quality (KORUS-AQ) field study was conducted during May–June 2016 to understand the factors controlling air quality in South Korea. Extensive aircraft and ground network observations from the campaign offer an opportunity to address issues in current air quality models and reduce model-observation disagreements. This study examines these issues using model evaluation against the KORUS-AQ observations and intercomparisons between models. Six regional and two global chemistry transport models using identical anthropogenic emissions participated in the model intercomparison study and were used to conduct air quality simulations focusing on ozone (O3), aerosols, and their precursors for the campaign. Using the KORUSv5 emissions inventory, which has been updated from KORUSv1, the models successfully reproduced observed nitrogen oxides (NOx) and volatile organic compounds mixing ratios in surface air, especially in the Seoul Metropolitan Area, but showed systematic low biases for carbon monoxide (CO), implying possible missing CO sources in the inventory in East Asia. Although the DC-8 aircraft-observed O3 precursor mixing ratios were well captured by the models, simulated O3 levels were lower than the observations in the free troposphere in part due to too low stratospheric O3 influxes, especially in regional models. During the campaign, the synoptic meteorology played an important role in determining the observed variability of PM2.5 (PM diameter ≤ 2.5 μm) concentrations in South Korea. The models successfully simulated the observed PM2.5 variability with significant inorganic sulfate-nitrate-ammonium aerosols contribution, but failed to reproduce that of organic aerosols, causing a large inter-model variability. From the model evaluation, we find that an ensemble of model results, incorporating individual models with differing strengths and weaknesses, performs better than most individual models at representing observed atmospheric compositions for the campaign. Ongoing model development and evaluation, in close collaboration with emissions inventory development, are needed to improve air quality forecasting. Show less