Estimation of PM2.5 concentration over East Asia using GOCI-II aerosol optical properties and machine learning model

Airborne particulate matter of diameter≤2.5 µm (PM2.5) is associated with detrimental effects on human health. The paucity of in-situ surface PM2.5 observations is a major obstacle to elucidation of the temporal and spatial variability of PM2.5 concentration. We used hourly aerosol optical products from the 2nd Geostationary Ocean Color Imager (GOCI-II) to estimate hourly PM2.5 concentrations over East Asia during January 2021. Additional input variables included meteorological and atmospheric trace gas data. Temporospatially weighted PM2.5 concentrations were used in assessing variations in ground-level PM2.5 levels. An oversampling technique and the Random Forest machine-learning method were applied. Results indicate stable model performance over the region with dynamic fluctuations in PM2.5 concentration, with 10-fold cross-validation results yielding R2 and RMSE values of 0.909 and 8.434 µg m−3, respectively.