Aerosol Chemical Component: Simulations with WRF-Chem and Comparison with Observations in Nanjing

Tong Sha,Xiaoyan Ma,Hailing Jia,Rong Tian,Yunhua Chang,Fang Cao,Yanlin Zhang
DOI: https://doi.org/10.1016/j.atmosenv.2019.116982
2019-01-01
Abstract:Secondary inorganic aerosols, including sulfate, nitrate, and ammonium (SNA), are the predominant components of fine particles (PM2.5). Reasonable representations of SNA formation in numerical models can largely improve the predictions of PM2.5 concentrations and effectively help implement emission control strategies. Despite the Atmospheric Pollution Prevention and Control Action Plan has been implemented since 2013, PM2.5 concentration during 2017 in Nanjing, one of the megacities in China, still exceeded the World Health Organization-recommended safe level (35 m(-3)). In this study, WRF-Chem model was applied to simulate aerosol chemical components in PM2.5 during April 2016 and January 2017 in Nanjing, and the simulations are evaluated with in-situ observations. Our results show that the model can reasonably reproduce the temporal variability of PM2.5 in two seasons, but significantly underestimate the sulfate concentrations by 71% (84%) in April (January), and overestimate the nitrate concentrations by 67% (45%) in April (January). The simulated ammonium is overall consistent with the observations. Meanwhile, the model tends to overestimate SO2 concentrations by 20% (74%) in April (January). Several sensitivity studies are conducted to explore the mechanisms for underestimation of sulfate and overestimation of nitrate, and found that the conversion rate of SO2 to sulfate is significantly underestimated in the model. Tripling the gas-phase oxidation rate of SO2 by OH only enhances sulfate by 67% (72%) in April (January), indicating gas-phase oxidation is not the main causes for the underestimations in the model. However, inclusion of SO2 heterogeneous oxidation in aerosol water can largely increase the simulated sulfate by 84% (196%) in April (January), and also better reproduce the diurnal variations of sulfate compared to the reference run. It should be noted that the simulated sulfate is still 47% (53%) lower than the observations in April (January), though inclusion of heterogeneous reaction can substantially improve the simulation performance of SNA.
What problem does this paper attempt to address?