Predictability of PM2.5 in Seoul based on atmospheric blocking forecasts using the NCEP global forecast system

Abstract

We investigate the predictability of PM2.5 (particulate matter with diameters <= 2.5 mu m) concentration in Seoul using atmospheric blocking forecasts during four winter seasons (December, January, and February) of 2015-2019. Atmospheric blocking is predicted using the National Centers for Environmental Prediction (NCEP) Global Forecast System (GFS), and PM2.5 concentration in Seoul of the same period are obtained from the surface measurements of the National Institute of Environmental Research (NIER). We define blocking indices using two methods: the anomaly method and the gradient method. Correlation analysis establishes the relationship between the blocking and PM2.5 concentrations in Seoul. The result based on the anomaly method exhibits strong positive and negative correlation coefficients respectively over the East Sea and Japan and the vicinity of Lake Baikal. The correlation coefficient from the blocking index of the gradient method reveals a similar pattern, and the maximum values are found within the longitude range of 130-150 degrees E Building on the relationship between PM2.5 in Seoul and the blocking index, we perform PM2.5 prediction using the GFS forecast of the atmospheric blocking index. The prediction shows reliable performance at least up to a five-day forecast. This atmospheric blocking based dynamical and statistical PM2.5 prediction can be a useful tool for mid-range PM2.5 forecast.