Statistical predictability of wintertime PM2.5 concentrations over East Asia using simple linear regression

• DOI: 10.1016/j.scitotenv.2021.146059
• Publication Date: 2021-02-25T07:36:08Z
• Authors: Jaein I. Jeong; Rokjin J. Park; Sang-Wook Yeh; Joon-Woo Roh
• Keywords: 13. Climate action; 01 natural sciences; 0105 earth and related environmental sciences

Abstract The interannual meteorological variability plays an important role in wintertime air quality in East Asia. In particular, monsoons and the El Nino Southern Oscillation (ENSO) are known as important mechanisms for determining wintertime PM2.5 concentrations. In addition, Arctic Oscillation, North Atlantic Oscillation, and Pacific Decadal Oscillation are also known to affect meteorological conditions and thus PM2.5 concentrations in East Asia. Here, we used a global 3-D chemical transport model (GEOS-Chem) with assimilated meteorological fields to investigate the long-term (1980–2014) relationship between 16 different climate indices and wintertime PM2.5 concentrations in this region. We show that wintertime PM2.5 concentrations in Northeast Asia (33–41°N, 118–141°E) are highly correlated with ENSO indices and the Siberian high-pressure system. Furthermore, we develop a simple linear regression (SLR) model for the prediction of wintertime PM2.5 concentrations. Despite the use of a single predictor, the SLR model shows good performance with r > 0.72 in reproducing targeted PM2.5 concentrations. The hit and false alarm rates are 77% and 11%, respectively, indicating the high predictive accuracy of the model. In particular, the model shows excellent performance for capturing the abnormal variability of wintertime PM2.5 concentrations in Northeast Asia.