Numerical study of air pollution over a typical basin topography: Source appointment of fine particulate matter during one severe haze in the megacity Xi'an

Highlights

• The hybrid modelling system is designed to study the sources of PM2.5 in Xi’an.

• The FAC2 and IOA of the system reaches to 84% and 0.81, better than the original.

• More than 60% of the PM2.5 at the urban center is contribute by Xi’an sources.

• The rural areas have higher transport contribution of PM2.5 pollutions in Xi’an.

Abstract

Many cities are located in lands with typical basin topographies, which are not conducive to the spread of air pollutants. In the winter of 2016/2017, a severe haze happened in Xi’an, the main city in the Guanzhong Basin in central China. When the peak daily concentration of fine particulate matter (PM_2.5) reaches 499 μg/m³, the source of the atmospheric pollution needs to be found urgently in order to take countermeasures. The comprehensive air quality model with extensions, coupled with the tracer tagging particulate source apportionment technology (PSAT) module, and an improved emission inventory, higher grid resolution, and bigger inner domain area, have been applied to quantify the contributions of local and regional emissions to the PM_2.5 pollutions. The model performed well in time period considered in this study. The correlation of the simulated daily PM_2.5 concentration data reaches 0.82, and the fraction of predictions within a factor of two of observations approaches 84%. With the PSAT module, the PM_2.5 contributions from local and regional sources to the urban centre and rural areas during the severe winter haze event are analysed in detail. The PM_2.5 concentrations in the urban centre in Xi’an is mainly originating from local emissions (60%), and Xianyang City is the largest contributor among the surrounding source regions (11.6%), while the transportation sector outside the Shaanxi Province (5.1%) also contributes significantly. Comparatively, the rural areas have lower local contributions and higher transport contributions. In particular, in the northern rural area Yanliang, the contribution from surrounding source regions approaches 82%. The results of this study suggest that to improve the air quality in a typical basin city, a regional-scale coordinated emissions control should be used, focusing on the emissions from both local and surrounding areas.

Introduction

Owing to the significant threat posed by air pollution, researchers have used various methods to quantify the impact of regional emissions on nearby receptor regions. The most direct approach involves using observational data to identify the pollution source regions. For example, researchers have employed aerosol observational data combined with a backward trajectories method (Dan et al., 1999, Polissar et al., 2001, Wang et al., 2015, Zhao et al., 2015, Liu et al., 2016, Ojha et al., 2017, Mengdi et al., 2019) or positive matrix factorization (PMF) (Kim and Hopke, 2005). Other approaches employ atmospheric models, including the Lagrangian transport and dispersion model (Hov et al., 1994, Trickl et al., 2003) and Eulerian chemical transport models (Park et al., 2003, An et al., 2007, Chin et al., 2007). Eulerian chemical transport models combined with a sensitivity analysis have great potential to determine the source-receptor relationship, even on an intercontinental scale (Liu and Mauzerall, 2007, Koo et al., 2009, Ling et al., 2017, Luecken et al., 2018, Uranishi et al., 2019). Three sensitivity analysis methods with Eulerian chemical transport models have been proposed. The most straightforward and simple approach is the brute-force method (BFM, or zero-out analysis). The BFM runs with the emissions from the source region of interest set to zero, repeats running with perturbed emissions, and compares the two simulation results (Koo et al., 2009). The BFM has obvious drawbacks in that the number of tests needed linearly increases the computational cost and can only study primary pollutants. Therefore, two additional methods, which are more efficient and accurate, have been proposed: the adjoint sensitivity method (Sandu et al., 2005) and the decoupled-direct method (DDM) (Dunker, 1981, Dunker, 2002). The adjoint approach provides information on influence areas by a backward analysis of the meteorological fields, while the DDM integrates the corresponding sensitivity coefficient equations forward. Both methods have been extensively applied to quantify the source areas that influence the pollutant concentrations (Martien and Harley, 2006, Hakami et al., 2007, Napelenok et al., 2007, Sandu and Zhang, 2008, Uranishi et al., 2017, Shixian et al., 2018).

Tracer-based approaches, such as the particulate source apportionment technology (PSAT) in comprehensive air quality model with extensions (CAMx) (Wagstrom et al., 2008), the tagged species source apportionment (TSSA) (Tonnesen and Wang, 2004), and the nested air quality prediction model system (NAQPMS) (Li et al., 2008, Wu et al., 2011) have also been gaining popularity. The PSAT is a tagged species method that can track source contributions of both primary and secondary particulate matter in the chemical transport model.

Many studies have applied tracer-based approaches to analyse the source regions of air pollution. Using the online air pollution tagged module of the NAQMPS, Wu et al. (2011) found that more than 50% of the sulfur dioxide (SO₂) and inhalable particulate matter (PM₁₀) in Beijing originated from the surrounding regions within 1.1 km. The CAMx model with the PSAT module was also applied to calculate the contribution to the total PM_2.5 mass of the regions around Shanghai, and the results showed that over 50% comes from outside sources during haze episodes in autumn (Wang et al. 2014). Wu et al. (2013) applied the PSAT to analyse the influence of different emission activities on the PM concentrations in the Pearl River Delta and found that 69.6% of the fine particulate matter in Guangzhou was transported from the surrounding regions in December 2011.

Previous studies have mainly focused on the economically developed regions in China, such as the Beijing–Tianjin–Hebei region, the Yangtze River Delta, and the Pearl River Delta, where extensive research has been conducted on air pollution and a large amount of emission data is available. The Guanzhong Basin in Midwest China also faces serious regional air pollution problems because of its topography and economic development (Zhang et al., 2002, Cao et al., 2012). Many experts have addressed this, focusing on observational data analyses (Han et al., 2006, Tao et al., 2012, Zhu et al., 2015). It is also necessary to analyse the influence of air pollutants inside and outside the Guanzhong Basin on the PM concentrations in Xi’an. Li et al. (2016) employed the weather research and forecasting model coupled to chemistry (WRF-Chem) and the BFM to estimate the source contributions to elemental carbon (EC) pollution in the Guanzhong Basin; their results showed that the local emissions contributed by approximately 60% to the annual mean EC concentration. Xue et al. (2014) used the CAMx with PSAT module to study the pattern of the regional transport of PM_2.5 over China and noted that the PM_2.5 transport contribution to the Shaanxi province was approximately 31%. However, previous studies have not identified the specific source regions and contributions to PM_2.5 in Xi’an City.

In this study, the CAMx with the PSAT module was employed to study the contributions of different emission regions, including Xi’an and the other four cities in the Guanzhong Basin (i.e. Baoji, Weinan, Xianyang, and Tongchun), and the long-range (i.e. south, north, and outside Shaanxi) regional transportation of fine particulate matter during a heavy haze episode in the winter from the end of 2016 to the beginning of 2017 in Xi’an, the main city in the Guanzhong basin.