;In the last three years (2013-2015), the most serious air pollution county in Taiwan is Yunlin. Douliu city, the capital of Yunlin, has many emission sources of particulate matter indicating complicated aerosol environment. In this study, we use 10 years (2005-2015) PM2.5 data of Douliu aera in autumn to analysis its temporal variation, spatial distribution and correlation with meteorology conditions. Aerosol data obtaining from an experiment in 2015 autumn at Douliu city has been used to further analyze aerosol vertical distribution and aerosol optical properties in both surface and vertical column. We try to use meteorology data, aerosol vertical distribution and the aerosol optics both at surface and vertical column to understand what the reason of aerosol concentration and the deterioration of air quality in Douliu city in autumn.
The daily mean PM2.5 concentrations for Taixi, Lunbei, and Douliu from 11 years (2005-2015) fall season show the highest value for Douliu and the lowest value for Taixi. However, during polluted events, Taixi PM2.5 concentration growth rates is highest and Douliu is lowest. All three sites show decreasing trend of PM2.5 concentrations in the past ten years, especially for Douliu site. Hourly PM2.5 data reveal Douliu concentration increase at daytime and decrease at nighttime, whereas an opposite day-night trend for Taixi, suggesting it may in relation to the local land-sea breeze circulation. Correlation coefficients (R) between four meteorology conditions (relative humidity, wind speed, temperature, and vertical stability) and PM2.5 concentrations at Douliu are -0.30, -0.29, -0.18, 0.19, respectively. If we consider all of above meteorological parameters together with PM2.5 concentration, R-squared can reach 0.22. It suggests that 22 percent of PM2.5 concentration variation is associated with meteorology conditions.
Results from 2015 field experiment showed that three meteorology parameters (relative humidity, wind speed and temperature) have better correlation coefficient with higher absorption aerosols (i.e. low single-scattering albedo). Correlation coefficient between temperature and absorption coefficient is 0.63 durning polluted event period. Higher correlation coefficient between meteological prameters and aerosol size are found to be -0.60, 0.50, 0.56 for RH, wind speed, and temperature, respectively. PM2.5 concentration shows positive correlation with AOD in the experimental period in general but shows a negative correlation durning polluted event period. This result implies the main sources of air pollution are local emissions and short-term near ground transport during the normal days. As contrast, lidar observation reveals high altitude aerosols downward transport to the ground durning polluted days. Positive correlation between PM2.5 concentration and PBL height also suggests that PM2.5 concentration will increase when atmospheric mixing is stronger. Results from case studies show that the increasing of surface air pollution in Douliu are due to local emission, aerosol transport by land-sea breeze circulation, nighttime residual layer downward to the surface by atmospheric vertical convection, and poor diffusion by high vertical stability. This study has implications on air quality diagnostic, forecast, as well as control policy making for the high PM2.5 area such as Douliu.