Characteristics of Water-soluble Inorganic Ions in Aerosol and Precipitation and their Scavenging Ratios in an Urban Environment in Southwest China

Received: August 9, 2020
Revised: December 8, 2020
Accepted: December 8, 2020

 Copyright The Author(s). This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are cited.

Cite this article:

Lin, C., Huo, T., Yang, F., Wang, B., Chen, Y., Wang, H. (2021). Characteristics of Water-soluble Inorganic Ions in Aerosol and Precipitation and their Scavenging Ratios in an Urban Environment in Southwest China. Aerosol Air Qual. Res. 21, 200513. https://doi.org/10.4209/aaqr.200513

HIGHLIGHTS

• Aerosol and precipitation samples were collected simultaneously over one-year period.
• Dynamic characteristics of particulate matter during rainfall processes were investigated.
• Scavenging ratios of major inorganic ions and gas pollutants were estimated.
• The relative wet scavenging contributions of gases and particles were calculated.

Keywords: Wet scavenging, Water-soluble inorganic ions, PM2.5, Precipitation chemistry

1 INTRODUCTION

Wet scavenging efficiently removes particulate matter and their gaseous precursors from the atmosphere (Wang et al., 2018a). In-cloud (rainout) and below-cloud (washout) scavenging are the two components of wet scavenging processes for air pollutants. For particulate matter, rainout and washout mechanisms are different, with rainout involving mostly nucleation of aerosols (acting as cloud condensation nuclei) and to a much less extent impaction scavenging, whereas washout involving mainly impaction scavenging (collected by falling raindrops) (Zhang and Vet, 2006). For gases, dissolution in liquid water droplets is the mechanism for both rainout and washout processes. The relative contributions of rainout and washout to the total wet deposition could be estimated based on the sequential precipitation sampling measurements (Aikawa and Hiraki, 2009). For example, using this method, Xu et al. (2017) estimated that washout contributed about 56%, 61%, and 47% to total wet scavenging of SO₄^2–, NO₃^–, and NH₄⁺ respectively, in Beijing, and Ge et al. (2016) obtained 88%, 92%, and 84% in Dalian and 56%, 50%, and 46% in Dandong in North China. Furthermore, washout accounted for about two-thirds of total wet scavenging for NO₃^– and half for SO₄^2– in Japan using the same method (Aikawa et al., 2014; Kajino and Aikawa, 2015).

Scavenging ratio (W), defined as the ratio of a pollutant concentration in precipitation to that in air, can be a useful indicator for wet scavenging efficiency (Engelmann, 1971). This parameter depends on the physical and chemical characteristics of both particles and precipitation, and can vary by up to two orders of magnitude with different chemical species. Budhavant et al. (2020) estimated the scavenging ratios of ten major chemical species in PM₁₀ over the Northern Indian Ocean, and obtained values of less than 50 for black carbon and up to 3000 for Cl^– and NO₃^–. Cheng and Zhang (2017) investigated the scavenging ratios of eight inorganic ions at 13 monitoring sites across Canada and obtained values varying by a factor of up to 6, with larger values for chemical species with higher fractions in coarse particles, e.g., Ca²⁺ and NO₃^–. The important role that below-cloud scavenging plays on ambient aerosol concentration under different rain conditions has also been demonstrated by field experiments and numerical studies (Zhang et al., 2004; Gao et al., 2019; Lu et al., 2019; Luan et al., 2019).

Due to the strong dependence of scavenging ratio on the characteristics of particles and precipitation, existing database are subject of large uncertainties. Scavenging ratio obtained from parallel measurements of ambient particulate matter and precipitation chemistry over long sampling periods may reduce such uncertainties. Furthermore, it is even more complicated for wet-scavenged chemical species involving both particulate- and gaseous-phase pollutants in air, such as SO₄^2–, NO₃^–, and NH₄⁺. Kasper-Giebl et al. (1999) estimated that 89–96%, 4–12%, and 49–79% of SO₄^2–, NO₃^–, and NH₄⁺ in precipitation were from the scavenging of their ambient particulate-phase ions, respectively, with the rest from the gaseous pollutants (SO₂, HNO₃, and NH₃). Similarly, Cheng and Zhang (2017) suggested that HNO₃ dominated the particulate NO₃^– in wet scavenging with a percentage contribution of about 72%, whereas the contributions from particulate SO₄^2– and NH₄⁺ were greater than the corresponding gas SO₂ and NH₃, with an average percentage contribution of 63% and 70%, respectively. Comprehensive investigation on the relative contributions from gases and particles to the total wet deposition may improve the wet scavenging models of SO₄^2–, NO₃^–, and NH₄⁺, and consequently reduce the uncertainties in prediction of particle concentrations during precipitation periods.

Wanzhou is a small urban city, situated in southwest China. The climate in Wanzhou is characterized by high relative humidity, extremely low wind speeds all year round, and abundant rainfall with annual precipitation of about 1200 mm. Thus, wet scavenging is expected to play an important role in removing pollutants from the atmosphere. Previous studies have focused on the dry and wet deposition fluxes of nitrogen or other pollutants in southwest China (Wang et al., 2016; Wang et al., 2018a), but provided no information on the wet scavenging processes or scavenging ratios of water-soluble inorganic ions. The present study aims to fill this knowledge gap by simultaneously measuring PM_2.5 and precipitation chemistry in different seasons of a year. The main objectives are to (1) investigate the dynamic characteristics of PM_2.5 and PM₁₀ during rainfall processes, (2) characterize major water-soluble inorganic ions in PM_2.5 and precipitation, and (3) estimate the scavenging ratios of major inorganic ions and gaseous precursors as well as their relative contributions.

 
2 METHODS

 
2.1 Sampling and Chemical Analysis

PM_2.5 and precipitation samples were collected in parallel in Wanzhou. The sampling site is located on the roof of a building inside the Chongqing Three Gorges University (108°13′E, 30°48′N), about 28 m above the ground, which is influenced by local vehicular emission and point sources (Wang et al., 2016). 23-h PM_2.5 samples were collected by an air pollutant sampler (URG Corp., URG-3000K, North Carolina, USA) from 11:00 a.m. to 10:00 a.m. the next day at a flow rate of 15 L min^–1. Two channels of the air pollutant sampler were used to collect PM_2.5 samples in parallel. The left channel was preceded by two annual denuders coated with glycerol/Na₂CO₃ and glycerol/citric acid solution for trapping HNO₃ and NH₃ sequentially, and then a Teflon filter was equipped for mass weighting. The right channel was loaded with quartz filter for water-soluble inorganic ions analysis. Daily precipitation samples were collected using an automatic atmospheric deposition sampler equipped with a rain gauge sensor (APS-3A, Xianglan Scientific Instruments Co., Ltd., China). Wanzhou has a subtropical monsoon climate with temperature above 0°C all year round, hence, precipitation at this location is almost always in rain. Although rainfall samples were collected on a daily basis in order to match PM_2.5 sampling campaign, the rain gauge sensor could automatically record the detailed information of each rainfall event, including the beginning and ending time, besides the rainfall amount. Sampling campaigns were conducted in four months in 2015: from 2^nd to 29^th in April, 2^th to 30^th in July, 16^th October to 13^th November, 16^th December to 14^th January in 2016, representing spring, summer, autumn, and winter, respectively. Note that PM_2.5 samples were not available from 2^nd to 7^th July due to the sampler maintenance. PM_2.5 samples were stored in the dark and at –18°C until analysis to prevent the evaporation of volatile compounds, while rainfall samples were filtered through a 0.45 µm filter after collection and refrigerated at 4°C.

Five cations (Na⁺, NH₄⁺, K⁺, Mg²⁺, Ca²⁺) and three anions (Cl^–, SO₄^2–, NO₃^–) were determined by an ion chromatography (DX-600, Dionex, Sunnyvale, CA) for both PM_2.5 and rainfall samples. Cations were measured using a CS12A column with 20 mM methanesulfonic acid (MSA) as an eluent while anions were analyzed using an AS11-HC column with 30 mM KOH as an eluent. Ions in rainfall samples were measured directly, while PM_2.5 samples should be extracted using ultrapure water before ion chromatography analysis. HNO₃ and NH₃ captured by denuders were extracted using ultrapure water first, and then measured by an ion chromatography. A detailed description of PM_2.5 and rainfall sampling as well as water-soluble inorganic ions analysis were presented in Wang et al. (2016, 2018b). For Na⁺ in PM_2.5, the field blanks varied largely and might result in large uncertainties. For Mg²⁺ and Ca²⁺, most of PM_2.5 samples were below the detection limits due to their relatively low concentrations. Hence, only five water-soluble inorganic ions in PM_2.5 including SO₄^2–, NO₃^–, NH₄⁺, Cl^–, and K⁺ were discussed in the following section.

Hourly concentrations of PM_2.5 and PM₁₀ were measured using online PM_2.5/PM₁₀ analyzer (FH62C14, Thermo Fisher, U.S.). Hourly concentrations of SO₂ were not available at the sampling site and obtained from an air quality monitoring station about 6 km away.

 
2.2 Data Analysis

The rate of change in particle concentration (ΔC) for each rainfall event was defined as follows:

where C_b is the average concentration of PM_2.5 or PM₁₀ within 3 h before the rain, C_d is the average concentration during the rain. If the value of ΔC is positive, it means that PM_2.5 or PM₁₀ decreases when rainfall occurs, otherwise, it indicates that PM_2.5 or PM₁₀ increases during rain. Calculation of ΔC requires that the free-rainfall period between two events is higher than 3 h, otherwise, consecutive rainfall events are combined (Aikawa et al., 2014). As a consequence, a total of 7, 8, 14, and 13 rainfall events were obtained in spring, summer, autumn, and winter, respectively (Table S1).

Scavenging ratio of major inorganic ions in each month was calculated based on monthly mean concentration as follows (Kasper-Giebl et al., 1999; Cheng and Zhang, 2017):

where C_prec is the volume weighted mean (VWM) concentration of inorganic ions in precipitation (mg L^–1), C_air is the mean concentration of water-soluble inorganic ions in the air (µg m^–3), and ρ_a and ρ_w are the density of air (1200 g m^–3) and water, respectively.

SO₄^2–, NO₃^–, and NH₄⁺ in rainfall originated from both particles and gaseous precursors. Hence, it is assumed that the difference between the total and particulate wet scavenging was from the gaseous precursor. Scavenging ratio of gaseous precursor was calculated followed by Cheng and Zhang (2017). The procedure for estimating the scavenging ratio of HNO₃ was described in detail, and the calculation for the scavenging ratio of SO₂ and NH₃ was done similarly.

The particulate NO₃^– concentration in rainfall was first estimated as follows:

where [pNO₃^–]_prec is the part of NO₃^– wet scavenging from particulate NO₃^–. W_fPM and W_cPM are monthly scavenging ratio of fine and coarse particles, respectively. W_fPM could be calculated from the scavenging ratio of K⁺, while W_cPM could be calculated based on the average scavenging ratio of Ca²⁺, Mg²⁺, and Na⁺. In this study, the coarse particles data were not available, thus the contributions from coarse particles were neglected. [pNO₃^–]_fPM and [pNO₃^–]_cPM are monthly mean concentration of NO₃^– in fine and coarse particles, respectively. Note that the concentrations of major inorganic ions in coarse particles were not available, alternative data assumptions were made, which were discussed in the following section in detail. Then, the scavenging ratio of HNO₃ and its contribution were calculated as follows:

where [total NO₃^–]_prec means the monthly VWM concentrations of NO₃^– in rainfall, [HNO₃]_prec is the concentration of NO₃^– in rainfall from gas contribution, and [HNO₃]_air is ambient concentration of gas HNO₃. %HNO₃ and %NO₃^– are the contributions of HNO₃ and particulate NO₃^– to the total NO₃^– wet scavenging, respectively. In some cases, the calculated [HNO₃]_prec value was negative, then we assumed that only particulate NO₃^– contributed to the wet scavenging and the contribution from gas HNO₃ was negligible.

 
3 RESULTS AND DISCUSSION

 
3.1 Variations of Hourly Particulate Matter Concentrations during Rain Events

During the observation periods, the accumulated rainfall amounts were 79.2, 152.8, 85.8, and 31.7 mm in spring, summer, autumn, and winter, respectively, while the corresponding total rainfall hours were 57.1, 61.7, 38.2, and 53.3 h. Note that the rainfall data were obtained in a typical month rather than three months in each season. In Wanzhou, rainfall was the most abundant in summer and least in winter. Although the total rainfall hours were similar in spring, summer, and winter, the accumulated amounts varied about five times due to the different rainfall intensities. Rain could be classified into three categories based on the rainfall intensity (Luan et al., 2019): light rain (0.1–2.5 mm h^–1), moderate rain (2.6–7.6 mm h^–1), and heavy rain (> 7.6 mm h^–1). Regarding the rain events recorded by rain gauge sensor, the predominant rain category was light rain in all seasons, in particular in spring and winter where the numbers of light rain events accounted for about 96% and 100% of the total events, respectively. Moderate and heavy rain events contributed to about 15.8% and 17.9% of the total events in summer, and 5.3% and 3.6% in autumn, respectively.

Particulate matter concentrations were affected by many factors, such as source emissions and wet scavenging effects. Fig. 1 shows the hourly concentrations of PM_2.5, PM₁₀, and rain intensities during the observation periods. For those rainfall events with relatively long durations and/or high rainfall intensities, PM_2.5 and PM₁₀ concentrations decreased greatly during rain, e.g., on 19^th April, 15^th July, and 25^th October in 2015, and 7^th January 2016, indicating that wet scavenging played an important role on the decreases of the PM_2.5 and PM₁₀ concentrations. However, for those rainfall events with short durations or low intensities, the scavenging effects exhibited large variations, implying that other factors would primarily contribute to the variations of particle concentrations. For example, on 21^st July with the rain duration of 0.8 h and rain intensity of 0.13 mm h^–1, PM_2.5 and PM₁₀ concentrations decreased by 39% and 48% during rain, respectively. We found that this phenomenon was caused by the increased PM_2.5 and PM₁₀ concentrations during the rush hours in the evening, and when the rush hours ended after 21:00 local time, the air pollutants concentrations began to decrease. This can be supported by the increased NO₂ concentrations during the rush hours and the decreased concentrations one or two hours before the rain. That means the decreases of PM_2.5 and PM₁₀ concentrations on this day were ascribed to the emission reduction after rush hours rather than the wet scavenging effects. On 24^th December, PM_2.5 and PM₁₀ concentrations increased by about 27% during rain period although it lasted about 8 hours from 7:00 to 17:00 local time. The reason for this phenomenon was similar to the rain event on 21^st July, which was caused by the high emissions during the rush hours when rain occurred, resulting in relatively high average PM_2.5 and PM₁₀ concentrations. Furthermore, no evident variations of PM_2.5 and PM₁₀ concentrations were observed before and during rain on 7^th April, in which the duration was about 5 h and rainfall intensity was greater than 1.0 mm h^–1. In total, those distinct scavenging effects for different rainfall events indicated that other factors also affected the overall scavenging effects besides rainfall duration and intensities.

Fig. 1. Hourly concentrations of PM_2.5, PM₁₀, and rainfall intensity in spring, summer, autumn, and winter.

The variations of PM_2.5 and PM₁₀ concentrations were not uniform for each event, showing increasing or decreasing trends when rainfall occurred (Fig. 2). For PM₁₀, almost all events showed decreasing trends in spring. In other seasons, the number of events with decreasing trends accounted for about 62.5–71.4% of the total rain events. For PM_2.5, the number of events presenting decreasing trends accounted for 50–64.3% of the total rainfall events, implying the important role of wet scavenging effects on decreasing particle concentrations.

Fig. 2. The variations of PM₁₀ (a) and PM_2.5 (b) concentrations for each rainfall event.

Generally, the decreasing rates of PM₁₀ by rainfall were higher than PM_2.5 for every event (Fig. 2), which can be explained by the size-dependent below-cloud impaction scavenging efficiency (Wang et al., 2010). Besides particle size distribution, removal rates are also related to the concentrations of particles before rain. The rates of change in particle concentrations were grouped with PM₁₀ or PM_2.5 concentrations before rain at 20 µg m^–3 interval bins. As shown in Fig. 3(a), ΔC typically increased with the increases of PM₁₀ and PM_2.5 concentrations, and the highest change rate in PM₁₀ and PM_2.5 concentrations appeared in the range of 80–100 µg m^–3 and 60–80 µg m^–3, respectively. This trend was also observed by Olszowski (2016), in which the scavenging effect was well correlated with higher particle concentrations before rain. Note that ΔC showed decreasing trends when PM₁₀ and PM_2.5 concentrations were higher than 100 µg m^–3. This phenomenon indicated that other meteorological conditions might govern the severe PM₁₀ and PM_2.5 pollution and then the wet scavenging effect was overwhelmed.

Fig. 3. The variations of PM₁₀ and PM_2.5 grouped by particle concentrations before rain (a), and rain intensity bins (b).

It has been reported that the rainfall intensity also influenced the variations of particle concentrations (He and Balasubramanian, 2008; Luan et al., 2019). As shown in Fig. 3(b), ΔC was highest (higher than 50% for both PM₁₀ and PM_2.5) when rainfall intensities were higher than 2 mm h^–1, indicating that moderate or heavy rain exhibited relatively high scavenging efficiency on particulate matter. This can also be explained by the precipitation-intensity dependent below-cloud scavenging efficiency (Wang et al., 2010). Scavenging effects were positive for each moderate or heavy rainfall event with the value of ΔC in the range of 8.3–90.9% for PM₁₀ and 2.6–93.5% for PM_2.5, respectively. However, no obvious trends of ΔC were observed for light rains, likely because other factors overwhelmed the below-cloud scavenging effect.

 
3.2 Gaseous Precursors and Water-soluble Inorganic Ions in PM_2.5 and Rainfall

Among the three gaseous precursors discussed here, the annual mean concentrations were the lowest for HNO₃, followed by SO₂, and highest for NH₃ (Table 1). From a seasonal perspective, SO₂ showed the highest concentration in spring and lowest in summer. SO₂ is primarily emitted from coal combustion for residential heating, power plants and industrial use in China. Wanzhou is located in southwest China with a temperature of above 0°C all year round, and residential heating using coal may not be extensive in this region. Thus, the seasonal patterns of SO₂ were mainly influenced by the meteorological conditions and chemical transformations rather than source emissions. High temperature and strong solar radiation in summer could enhance photochemical reactions, leading to the decrease of gaseous precursor concentration such as SO₂ and the increase of the secondary aerosol such as SO₄^2–. In addition, the high planetary boundary layer height in summer favors the pollutants dispersion and then leads to the low SO₂ concentration. Contrary to the case of SO₂, HNO₃ and NH₃ exhibited the maximum concentrations in summer and lowest values in winter and autumn, which were likely associated more with the source emission and thermodynamic behavior of NH₄NO₃ than other conditions (Wang et al., 2018b).

The annual mean concentration of PM_2.5 was 41.6 ± 24.5 µg m^–3, with the highest concentration in winter and lowest in summer. The total concentration of the five major water-soluble inorganic ions (WSIIs) was 20.3 µg m^–3, accounting for about 48.7% of PM_2.5. All five water-soluble inorganic ions displayed the highest concentrations in winter (Fig. 4 and Table 1), which were partly related to the unfavorable diffusion conditions, i.e., extremely weak wind (< 1 m s^–1) and low precipitation amount. In addition, the relative humidity was generally higher than 80% in winter, which was conducive to secondary inorganic aerosol formation through heterogeneous reactions and thus increased the concentrations of SO₄^2–, NO₃^–, and NH₄⁺. Biomass burning for residential heating in winter probably contributed to the higher concentrations of Cl^– and K⁺. The lowest concentrations of the five inorganic ions appeared in summer due to the favorable diffusion conditions and abundant precipitation with the exception of SO₄^2–, which was in autumn. Furthermore, the pronounced high concentrations of NO₃^– in winter and extremely low concentrations in summer were governed by the thermodynamic equilibrium of NH₄NO₃, since the low temperature and high relative humidity in winter were conducive to the formation of particulate NO₃^–, whereas high temperature in summer enhanced the volatilization of NH₄NO₃ (Wang et al., 2018b). Compared with the other inorganic ions, the relatively higher concentration of SO₄^2– in summer was mainly ascribed to the intense photochemical reactions as mentioned above. In PM_2.5, the winter/summer ratios were 1.7, 6.7, 2.3, 25, and 2.5 for SO₄^2–, NO₃^–, NH₄⁺, Cl^–, and K⁺, respectively, indicating that NO₃^– and Cl^– presented the largest seasonal variations.

Fig. 4. Seasonal patterns of (a) PM_2.5, (b) Cl^–, (c) K⁺, (d) SO₄^2–, (e) NO₃^–, and (f) NH₄⁺ in PM_2.5 as well as the seasonal VWM concentrations of inorganic ions in rainfall.

Among the five major water-soluble inorganic ions in PM_2.5, SO₄^2– is the most abundant ions, showing seasonal mean contributions of 46–64% to WSIIs during the sampling periods, implying the important contributions of coal combustion to PM_2.5 pollution. Besides SO₄^2–, NO₃^– and NH₄⁺ were also predominant components of WSIIs. In autumn and winter, NO₃^– and NH₄⁺ showed comparable contributions, each of which was about half of that of SO₄^2–. However, SO₄^2–, NO₃^–, and NH₄⁺ presented distinct contributions to WSIIs in summer, where the percentage of SO₄^2– was about 3 times that of NH₄⁺ and even 7 times higher than that of NO₃^–. The relatively high contribution of SO₄^2– and quite low contribution of NO₃^– were consistent with the patterns of their concentrations. The fractions of Cl^– and K⁺ in WSIIs were minor, ranging from 0.37% in summer to 3.8% in winter for Cl^– and from 1.6 to 2.6% for K⁺.

As shown in Fig. 4 and Table 1, the seasonal variations of eight inorganic ions in rainfall followed the patterns of those in PM_2.5, showing the highest VWM concentrations in winter and lowest in summer. Besides the relatively low concentrations of pollutants in the air, dilution effect due to the abundant rainfall was also responsible for the lower concentrations of those inorganic ions in summer. Compared the winter/summer ratios of the inorganic ions in PM_2.5 and rainfall, it could be found that the seasonal variations were more pronounced in PM_2.5 than in rainfall except for SO₄^2–, implying the sensitivity of the air pollutants to the meteorological conditions. Although Ca²⁺ concentrations were very low in PM_2.5, they were almost comparable to NH₄⁺ in rainfall in spring and autumn, highlighting a high contribution from coarse particles to rainfall. On an equivalent amount basis, rainfall was dominated by NH₄⁺ and SO₄^2–, followed by Ca²⁺ and NO₃^– (Table 1). The distribution of inorganic ions abundance in rainfall at Wanzhou was similar to Hangzhou (Han et al., 2019), but slightly different from the mean value across China which showed SO₄^2– and Ca²⁺ as the most abundant species (Li et al., 2019b).

 
3.3 Scavenging Ratios of Major Inorganic Ions

Calculating scavenging ratios of inorganic ions and gaseous pollutants required more parameters, including the concentrations of major inorganic ions in fine particles, coarse particles and rainfall, and the concentrations of SO₂, HNO₃, and NH₃. Fine/coarse fractions of these ions were not measured in the present study, but were previously reported for urban sites in Chongqing (Li et al., 2018a, b), which can be considered to be representative of the location of the present study. Briefly, Mg²⁺ and Ca²⁺ were mainly distributed in coarse particles, SO₄^2–, NH₄⁺, and K⁺ were primarily in fine particles, and Na⁺, NO₃^–, and Cl^– were distributed in both fine and coarse particles (Table S2). In this study, the concentrations of inorganic ions in coarse particles were not available. Considering the relatively high fractions of coarse particles for Mg²⁺, Ca²⁺, and Na⁺, the scavenging ratios of these ions will be highly uncertain if their concentrations in coarse particles were ignored. Meanwhile, the scavenging ratio of coarse particles (W_cPM) cannot be obtained due to the lack of coarse particles data, since W_cPM was determined by averaging scavenging ratios of Na⁺, Mg²⁺, and Ca²⁺ (Cheng and Zhang, 2017). Nevertheless, it is assumed that the influences due to the lack of coarse particles data would be minor on those inorganic ions dominated in fine particles. Thus, the scavenging ratios were only analyzed for SO₄^2–, NO₃^–, NH₄⁺, Cl^–, and K⁺ using their concentrations in PM_2.5. Taking into account the fractions of K⁺ in fine particles around 80%, the scavenging ratio of K⁺ multiplied by 0.8 was assumed to be the scavenging ratio of fine particle (W_fPM).

Seasonal and annual mean scavenging ratios of five inorganic ions and three gases are summarized in Table 2. On an annual basis, SO₄^2–, NO₃^–, and Cl^– showed comparable scavenging ratios at around 450, which was about 13–25% higher than the values for NH₄⁺ and K⁺. As shown in Fig. S1, positive correlations of equivalent concentrations between [NH₄⁺] and [SO₄^2– + NO₃^–] were observed in the four seasons, with the correlation coefficients higher than 0.90 and the slopes of linear regressions being around 1.0, suggesting that SO₄^2– and NO₃^– were fully neutralized by NH₄⁺, and (NH₄)₂SO₄ and NH₄NO₃ were the two major chemical forms in PM_2.5. Consequently, it seems that SO₄^2–, NO₃^–, and NH₄⁺ should have consistent scavenging ratios due to their coexistence in fine particles. However, slightly higher values of SO₄^2– and NO₃^– than NH₄⁺ were observed. As mentioned above, about 30%–40% of Cl^– and NO₃^– were found in coarse particles, thus the relatively higher scavenging ratio of NO₃^– was in part ascribed to the contributions from coarse particles. The slightly higher scavenging ratio of SO₄^2– was probably attributed to the in-cloud rainout process, since SO₄^2– could be formed efficiently through in-cloud oxidation of SO₂ by H₂O₂ or O₃, and then the cloud droplets including SO₄^2– could be removed as the raindrop falls. Based on Eqs. (2)–(5), the annual mean scavenging ratio was 116, 353, and 18 for SO₂, HNO₃, and NH₃, respectively. The different scavenging ratios of the three gases could be partly explained by their different solubilities. HNO₃ is the most soluble species, followed by NH₃, and SO₂ is the least soluble gas, with a Henry’s law constant of around 2.1 × 10³, 6.1 × 10^–1, and 1.2 × 10^–2 mol m^–3 pa^–1, respectively (Sander, 2015). NH₃ solubility is higher than SO₂, their relative low scavenging ratio might be responsible for the short lifetime of NH₃. The air pollutants incorporated into precipitation occurs at cloud level, however, the ambient concentrations were measured at the ground. Over this vertical distance, NH₃ may transform into particulate NH₄⁺ and consequently lead to the wet scavenging mainly from particulate NH₄⁺ rather than gas NH₃.

Different seasonal patterns of the scavenging ratios were observed for different inorganic ions and gases. Seasonal variations in scavenging ratios are caused by many factors, such as ambient concentrations of the ions, rain intensities and associated precipitation chemistry and droplet spectra. These factors would affect the dissociate rate of related pollutants between ambient air and rain droplets, below-cloud evaporation of droplets, etc., all of which would contribute to the overall calculated scavenging ratio. The maximum scavenging ratio was observed in autumn for both SO₄^2– and NH₄⁺, and the minimum value appeared in summer and winter, respectively. In contrast, the highest scavenging ratio was found in summer for Cl^– and NO₃^– and lowest in autumn and winter, respectively. The highest scavenging ratios of SO₂, HNO₃, and NH₃ were observed in autumn, spring, and summer, respectively.

The scavenging ratio values of SO₄^2–, NH₄⁺, and K⁺ obtained in the present study generally agreed well with existing data reported in literature (Table 2); however, those of NO₃^– were slightly lower than reported in the other regions, which was probably related to the different size distributions among different regions, among other factors. Contrary to the predominant coarse mode distribution of NO₃^– influenced by natural sources or sea salt in other regions (Zhang et al., 2008), about 70%–80% of NO₃^– was found in fine mode in the study region, which might result in lower scavenging ratio since rainfall scavenges coarse particles more efficiently. Additionally, the sampling site in this study is an urban site, which is severely affected by the vehicle emissions and hence results in relatively high ambient concentrations of NO₂ and NO₃^–. For the three gases, limited NH₃ scavenging ratios were reported in the previous studies. The scavenging ratio of SO₂ in this study was within the range of those values in the literatures, while the value for HNO3 was somewhat lower in this study (Table 2).

As expected, particulate SO₄^2– was the dominant contributor to rainfall SO₄^2–, accounting for 69.4%, while gas SO₂ contributed about 30.6% (Fig. 5), which were consistent with those conducted at rural sites in Canada (Cheng and Zhang, 2017). Similar to SO₄^2–, particulate NO₃^– accounted for about 70% of the rainfall NO₃^–, which was much higher than those reported by Cheng and Zhang (2017) (around 28%) and Kasper-Giebl et al. (1999) (about 4–12%). The discrepancies between our results and other literatures were probably attributed to the different characteristic of sampling location, where the sampling site with higher ambient NO₃^– concentrations exhibited high relative contributions of particulate NO₃^–. The relative contribution of particulate NH₄⁺ to total wet scavenging was 86.6%, which was higher than those in Cheng and Zhang (2017) (70%) and Kasper-Giebl et al. (1999) (48–79%).

Fig. 5. Relative contributions of particulate and gas-phase to the total wet scavenging of SO₄^2–, NO₃^–, and NH₄⁺ during the sampling periods.

 
4 CONCLUSIONS

Water-soluble inorganic ions in ambient fine particles and in precipitation were characterized, and relative contributions from particulate and gaseous pollutants to inorganic ions in precipitation were quantified. PM_2.5 and PM₁₀ concentrations were generally decreased after occurrence of moderate or heavy rain event, but this was not always the case after light rain event due to too many other factors affecting the ambient concentration. The scavenging ratio values derived for the five inorganic ions that were mostly distributed in fine particles varied by 25% on average, e.g., about 450 for SO₄^2–, NO₃^–, and Cl^–, 364 for NH₄⁺, and 395 for K⁺, likely due to their slightly different fine/coarse fractions. Scavenging ratios for gaseous species investigated here differed greatly, e.g., 353 for HNO₃, 116 for SO₂, and 18 for NH₃, likely related to their different solubility or lifetime in air. Besides, major sources and chemical processes might have contributed to the ambient concentrations of some chemical species and thus substantially lowered the scavenging ratio, e.g., for the case of NH₃. Seasonal variations of the scavenging ratios were a factor of 1.7–8 for different air pollutants. Wet scavenging of SO₄^2–, NO₃^–, and NH₄⁺ were primarily from particulate forms (> 69%), especially considering that only fine particles were considered in this estimation due to the lack of the coarse particles data. The scavenging ratio data obtained in this study may be used for improving the prediction of particle concentrations during rain periods using air quality models.

 
ACKNOWLEDGEMENTS

This work was supported by the National Natural Science Foundation of China (grant numbers 41405027, 41831285), Key Research and Development Projects of Sichuan Province (grant number 2020YFS0056), and the National Key R&D Program of China (grant numbers 2016YFC0200400, 2018YFC0214002).