New method for evaluating winter air quality: PM2.5 assessment using Community Multi-Scale Air Quality Modeling (CMAQ) in Xi'an

Particulate matter is the main air pollutant in China, especially in Xi'an in recent years. Since 2013, the WRF-SMOKE-CMAQ model system has been used to build an air quality model system for daily air quality forecasting in Xi'an. The emission inventory was built based on several anthropogenic emission inventories and open access emission datasets, and the model evaluation is presented to verify the emission inventory for particulate matter in Xi'an. Comparing the daily observed and simulated fine particulate (PM2.5) concentrations for four winters in different years (from 2014 to 2017), the model performs well in all studied time periods. The correlation coefficient of the simulated daily PM2.5 concentration data are all larger than 0.58, reaches 0.80 in 2016, and the fraction of predictions within a factor of two of observations (FAC2) are all above 66%. The differences of simulated results based on emission-unchanged system between 2014 and 2015 indicate that the slightly deteriorating air quality of 2015 is affected by the unfavorable air diffusion condition. The PM10 concentration increases from 95.9 mu g/m(3) to 110.3 mu g/m(3), and the PM2.5 from 82.4 mu g/m(3) to 95.4 mu g/m(3). According to the error analysis in model performance, the serious polluted situation of 2016 is mostly because of the sharp increased dust emissions. The emission-unchanged simulated particulate matter concentrations have little variation from 2015 to 2016, but the observation data increase obviously, that results in dramatically change of Mean Bias (MB). The absolute MB of PM10 increase from 70.3 mu g/m(3) to 135.2 mu g/m(3), and PM2.5 from 0.46 mu g/m(3) to 69.9 mu g/m(3). While the improved air quality in 2017 is attributed to both the better weather condition and the emission-reductions. The emission-unchanged simulated results decrease, and the absolute MB even have bigger decrease, that of the PM10 concentration reduce by 47 mu g/m(3), and PM2.5 by 37 mu g/m(3).