Evaluation of Particulate Matter Source Apportionment Forecasts during the MAPS-Seoul Field Campaign

Abstract:

We report forecasting model performance analysis results of Comprehensive Air quality Model with extensions (CAMx) simulation evaluated with flight measurements during Megacity Air Pollution Studies–Seoul (MAPS–Seoul) field campaign. The primary focus of this study is two-fold: (1) the air quality forecasting model performance for O3, PM10/2.5 and their precursors over the Yellow Sea to measure the model’s ability to account for the transport process and (2) the utilization of modeled source-receptor relationship to understand the root of systematic model under-prediction for PM10 and PM2.5 forecasts. MAPS–Seoul, conducted in the Seoul Metropolitan Area (SMA) in the summer of 2015, was an integrated research program covering ground monitoring and aloft measurement with aircrafts. To support this field campaign, air quality forecasting was performed with Weather Research and Forecasting (WRF) - Sparse Matrix Operator Kernel Emissions (SMOKE) - CAMx modeling framework. WRF model simulations initialized with National Centers for Environmental Prediction Global Forecasting System (NOAA/NCEP-GFS) were prepared for daily meteorological forecasts. Emission inventories used in this study are Model Inter-Comparison Study-Asia (MICS-Asia) 2010 for Asia and Clean Air Policy Support System (CAPSS) 2010 for South Korea. Simulated PM10 concentrations were evaluated with observed PM10 concentrations at ground monitoring sites of the AirKorea network in SMA. During the campaign period, average simulated PM10 concentrations showed significant underprediction, over 30% (~35 ㎍/㎥) lower than those observed at sites. To examine source-receptor relationship as a way to identify the cause of underprediction, we ran CAMx with Particulate matter Source Apportionment Technology (PSAT). The air quality forecasting model is based on the with 27-km horizontal grid resolution over Northeast Asia.