Investigating Impact of Emission Inventories on PM2.5 Simulations over North China Plain by WRF-Chem

Xiaoyan Ma,Tong Sha,Jianying Wang,Hailing Jia,Rong Tian

DOI: https://doi.org/10.1016/j.atmosenv.2018.09.058

IF: 5

2018-01-01

Atmospheric Environment

Abstract:Anthropogenic air pollutant emissions have been increasing rapidly in China, leading to severe heavy haze events. In order to predict PM2.5 concentrations accurately, the emissions of precursor gases and primary particles are quite critical but remains large uncertainty. In this study we firstly compared the emissions of precursor gases and primary particles such as sulfur dioxide (SO2), nitrogen oxides (NOx), ammonia (NH3), carbon monoxide (CO), black carbon (BC), organic compounds (OC), non-methane volatile organic compounds (NMVOCs), PM(2.5 )and PM10, in three emission inventories (EIs) including INTEX-B (representing the year 2006), MEIC10 arid MEIC12 (representing the year 2010 and 2012) over North China Plain (NCP), one of the most polluted regions in China. It is found that annual total emissions of SO2 decreased 34% (26%) from INTEX-B to MEIC10 (MEIC12), possibly due to effective control and improvement of fuel combustion efficiency conducted in recent years. The NOx emissions increased 53% (58%) from INTEX-B to MEIC10 (MEIC12), and the major contributor to NOx was replaced from power in INTEX-B by industry in MEIC10/12. The VOCs emissions in MEIC10/12 are 21.1% lower than in INTEX-B possibly due to tightened emission standards and operating conditions. The simulations with three EIs, based on WRF-Chem, generally reproduced the observed seasonal variation of PM2.5 concentrations, i.e. high in winter and autumn but low in spring and summer. The spatial distribution from the simulations are overall in good agreement with each other, and the differences less than 20 mu g m(-3) in most regions. The maximums by the simulations with three EIs occur in autumn due to very low wind speed despite the highest emissions in winter. Comparisons in four representative cities indicate that the simulations are quite close to observations in April and July, but overestimate in October and December in all cities exception in Baoding, especially with INTEX-B. Large bias with INTEX-B is possibly because the simulations are compared with observations in 2014 while INTEX-B represents the emission of 2006 year which is more far away from 2014. A quantitative assessment by conducting statistical analysis gave evidence that the simulations with three EIs do not exhibit significant difference in temporal variations, in terms of correlation coefficients, standard deviation, and root-mean-square difference. Therefore, MEIC inventories are suitable for simulations of PM2.5 concentrations for recent years, while INTEX-B might be better for a climate study to simulate the condition decade ago. Temporal variations can be basically well reproduced by either INTEX-B or MEIC10/12 inventories.