Combustion-related activities dominate atmospheric ammonia in the Pearl River Delta region, China

Abstract

Ammonia (NH₃) is a key precursor of fine particulate matter (PM_2.5) in the air; however, its emission sources at different heights remain poorly understood in the Pearl River Delta (PRD) region of China. In this study, we simultaneously collected PM_2.5 samples at three atmospheric heights (ground, 118 m, and 488 m) based on the atmospheric observatories of Canton Tower, the tallest structure in the PRD region. Our results showed that the average NH₄⁺ concentrations were 2.7 ± 1.4, 3.0 ± 1.8, and 2.6 ± 1.7 μg/m³ at the ground site, 118 m, and 488 m during the sampling campaign, with no significant difference (p > 0.05) among the three heights. However, the stable nitrogen isotope composition values in NH₄⁺ (δ¹⁵N-NH₄⁺) displayed a significant correlation with height (p < 0.05). We further calculated the initial δ¹⁵N-NH₃ values and performed source apportionments using the Bayesian Isotope Mixture Model. The results indicated that the mean contributions of agriculture, waste, vehicle, biomass burning, NH₃ slip, and coal combustion were 9.9% ± 4.4%, 8.3% ± 5.5%, 29% ± 8.0%, 16% ± 2.2%, 25% ± 6.0%, and 12% ± 3.4%, respectively, at the ground site during the sampling campaign. By contrast, the contributions of sources at 488 m remained relatively stable due to the limited influence of local activities. Overall, our study highlights the dominant role of combustion sources in NH₃ emissions in the PRD region, with their contribution being highly dependent on atmospheric height.

Graphical abstract

Keywords

Ammonia / Stable nitrogen isotope / Source apportionment / Canton Tower / PRD region

Highlight

	● The NH4+ concentrations showed no significant differences at different altitudes.
	● The δ15N-NH4+ values decreased with increasing altitude in urban areas.
	● Combustion source was the largest emitter for NH3 in the PRD region.
	● The contributions of different sources to NH3 vary with height.

1 Introduction

As the most abundant alkaline gas in the atmosphere, ammonia (NH₃) can effectively react with acidic substances (e.g., SO₂ and NO_x) and transfer to ammonium aerosol (NH₄⁺), which is a significant component of fine particulate matter (PM_2.5) with an aerodynamic diameter of less than 2.5 μm (Wei et al., 2023). On average, the mass contribution of NH₄⁺ in PM_2.5 is approximately 5% globally (Weagle et al., 2018). Yet, this contribution would increase to 10%–20% during severe haze pollution events in regions with intensive anthropogenic activities, such as East Asia and South Asia (Nirmalkar et al., 2023; Wang et al., 2023a). To date, both field observations and simulation model studies have confirmed that reducing NH₃ emissions is a cost-effective strategy for the mitigation of PM_2.5 pollution in the world (Gu et al., 2021; Xu et al., 2022; Zheng et al., 2022; Vo and Christiansen, 2024).

Technology-based bottom-up emission inventories (EI) indicate that atmospheric NH₃ primarily originates from fertilizer application, livestock, waste, vehicle, coal combustion, and biomass burning. Agricultural sources, namely fertilizer application and livestock, are generally considered the largest emitter globally, contributing 83% (U.S.EPA, 2023), 93% (EEA, 2023), 83% (Li et al., 2021), and 81% (Sahoo et al., 2024) in the USA, Europe, China, and India, respectively. However, a large uncertainty exists in estimating NH₃ emissions based on EI methodology due to the highly variable emission factors of key sources. For instance, Farren et al. (2020) measured NH₃ emissions from 230000 passenger cars and found that current EI-derived NH₃ emissions are underestimated by a factor of 17. In addition, long-term global satellite observations have revealed that industry is a substantial source of atmospheric NH₃ and is underestimated by at least one order of magnitude by EI (Van Damme et al., 2018). Since the adopted emission factors and the associated activities of sources depend on the information obtained by researchers, EI-based NH₃ estimates for specific regions (e.g., China) often lack sufficient convergence (Li et al., 2017a; Li et al., 2021). This may lead to significant errors when formulating NH₃ reduction policies and assessing the impacts of NH₃ reductions on air quality using atmospheric chemical transportation models.

The stable nitrogen isotope composition (δ¹⁵N) is a reliable tool for source apportionment of atmospheric NH₃, as it provides distinct signatures of δ¹⁵N-NH₃ in key emission sources. For example, Pan et al. (2016) quantified the relative contributions of agricultural emission, NH₃ slip, and fossil fuel in atmospheric NH₃ in Beijing, China, finding that agricultural contributions accounted for 84% on clean days and 10% on hazy days. In Colorado’s Front Range urban corridor, Felix et al. (2023) found that vehicle was the largest emitter (37%) of atmospheric NH₃, followed by biomass burning (34%), livestock (18%), and fertilizer (12%). Most δ¹⁵N-related studies focusing on the source apportionment of NH₃ or NH₄⁺ have primarily examined the atmosphere near ground level (Xiao et al., 2020; Chang et al., 2021; Feng et al., 2022; Walters et al., 2022; Feng et al., 2023; Kawashima et al., 2023; Li et al., 2023; Zhang et al., 2023). However, knowledge of its vertical distribution remains poorly constrained due to the difficulties associated with sampling (Li et al., 2017b; Wu et al., 2019; Wu et al., 2024). For example, Wu et al. (2019) measured the concentration and δ¹⁵N signatures of NH₄⁺ in PM_2.5 samples collected at three atmospheric heights (8 m, 120 m, and 260 m) on a 325 m tower in Beijing, North China. Their results indicated that the relative contributions of emission sources to NH₄⁺ were not uniform, with combustion sources playing a more significant role at the ground level compared to higher atmospheric altitudes. This finding was further corroborated by a recent study based on tower observations conducted at the same site (Wu et al., 2024). However, these studies were limited to North China and specific seasons, with the maximum sampling height restricted to 260 m. Such limitations hinder a comprehensive understanding of the sources of atmospheric NH₃ across different seasons, regions, and higher altitudes. Compared to the ground site, the higher atmosphere is less affected by local sources, thus better representing the regional characteristics of the research area. In this study, year-long PM_2.5 samples were simultaneously collected from three different heights (ground, 118 m, 488 m) using atmospheric observatories at Canton Tower in Guangzhou, the largest city in South China. We also measured the δ¹⁵N signatures of NH₄⁺ and systematically investigated the vertical variations in the relative contributions of key NH₃ sources, which will significantly enhance our understanding of NH₃ sources and provide valuable insights for formulating effective emission reduction policies in this megacity.

2 Materials and methods

2.1 Aerosol sampling

The sampling campaign was conducted as part of the PRD region Aerosol Isotope Observation (PRDAIO) project (Chen et al., 2022; Jiang et al., 2022; 2023; Wang et al., 2023b). Three high-volume PM_2.5 samplers (PM_2.5-ASM-1, Guangzhou Mingye Environmental Protection Technology Co., Ltd., China) with a flow rate of 1 m³/min were installed at three atmospheric observatories (ground, 118 m, and 488 m) at the Canton Tower in Guangzhou (113.3°E, 23.1°N), China. Guangzhou, the largest city in South China, is situated in the Pearl River Delta region (PRD) and is a highly urbanized metropolis with a population of approximately 18 million. The Canton Tower, the tallest structure in the PRD region, is located in downtown Guangzhou and is surrounded by dense traffic networks and residential areas. A total of more than 100 48-h PM_2.5 samples were simultaneously collected using quartz microfiber filters (0.052 m², Whatman, UK) at the three heights from October 2018 to August 2019.

2.2 Measurements of NH₄⁺ concentration and stable nitrogen isotope

Before sampling, all quartz filters were heated in a muffle furnace at 450 °C for 6 h to eliminate potential impurities. After sampling, all filter samples were stored in a refrigerator at −20 °C until analysis. Quartz filters were treated with 20 mL of ultrapure water and sonicated for 40 min. Subsequently, the extracted solutions were filtered through 0.22 μm membranes, and the concentrations of NH₄⁺ in the solutions were determined using an ion chromatograph (Dionex ICS-5000, Thermo Fisher Scientific Inc., USA). The mean instrumental detection limit (LOD) and measured precision were < 5 μg/L and ± 5%, respectively (Chen et al., 2022).

For the δ¹⁵N measurements, the NH₄⁺ was initially converted to nitrite (NO₂⁻) using hypobromite (BrO⁻) (AR, Sinopharm Chemical Reagent Co., Ltd, China), which was subsequently transformed into nitrous oxide (N₂O) using hydroxylamine (NH₂OH) (AR, Sinopharm Chemical Reagent Co., Ltd, China) under strongly acidic conditions. Finally, the isotopic analysis was conducted using an isotope ratio mass spectrometer (MAT253, Thermo Fisher Scientific, USA) (Liu et al., 2018). The δ¹⁵N-NH₄⁺ values are reported relative to the standard (atmospheric N₂) and expressed in parts per thousand (‰) according to the following equation (Eq. (1)):

(1)\begin{array}{rl}& {\delta }^{15}{\mathrm{N}(\text{‰})=[(}^{15}{\mathrm{N}/}^{14}{\mathrm{N})}_{\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}}-{(}^{15}{\mathrm{N}/}^{14}{\mathrm{N})}_{\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{a}\mathrm{r}\mathrm{d}}]/\\ & {(}^{15}{\mathrm{N}/}^{14}{\mathrm{N})}_{\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{a}\mathrm{r}\mathrm{d}}\times 1000.\end{array}

Each sample was calibrated using three international reference materials (IAEA-N1, USGS25, and USGS26), which have δ¹⁵N values of 0.4‰, −30.4‰, and 53.7‰, respectively. The analytical precision for the δ¹⁵N-NH₄⁺ measurements was better than 0.3‰ (Liu et al., 2014).

2.3 Calculation of the initial δ¹⁵N-NH₃

The initial δ¹⁵N-NH₃ can be calculated using the isotope mass balance model (Pan et al., 2016) as follows (Eq. (2)):

(2){\delta }^{15} {\text{N-NH}}_{3\text{(gas)}}= {\delta }^{15}{\text{N-NH}}_{4(\text{aerosol})}^{+}-{\epsilon }_{({\text{NH}}_4^{+}{\text{-NH}}_3)}\times (1-f ).

In the above equation, δ¹⁵N-NH_3(gas) and δ¹⁵N-NH₄⁺_(aerosol) denote the δ¹⁵N value of initial NH₃ and the measured δ¹⁵N value of aerosol NH₄⁺, respectively. The parameter f means the fraction of NH₃ converted into aerosol NH₄⁺. Since NH₃ was not collected during the sampling campaign in this study, the f value was derived from a previous study that simultaneously measured the concentrations of NH₃ and NH₄⁺ in the PRD region (Yue et al., 2015). ε is the nitrogen isotopic enrichment factor (Eq. (3)), which is a function of temperature (T, Kelvin) (Urey, 1947; Li et al., 2012; Kawashima and Ono, 2019; Walters et al., 2019):

(3){\epsilon }_{({\mathrm{N}{\mathrm{H}}_4}^{+}-\mathrm{N}{\mathrm{H}}_3)} =12.4678\times 1000 /\mathit{T}-\mathrm{7.6694.}

2.4 Isotope mixing model

The Stable Isotope Analysis in R (SIAR) package was used to quantify the relative contributions of major sources to atmospheric NH₃ (Parnell et al., 2010). SIAR integrates the isotopic signatures (δ¹⁵N) and associated uncertainties of potential sources, providing probabilistic estimates of their contributions. The initial δ¹⁵N-NH₃ in the atmosphere was treated as a mixture of multiple sources. The isotopic endmembers of key sources were parameterized using a prior distribution based on a newly established NH₃ source profile, incorporating both the mean and variance. The model applies Monte Carlo simulations to reconcile the initial δ¹⁵N-NH₃ value of the mixture with the distributions of potential source endmembers. Each stimulation iteration generated percentage data for source contribution estimates. Final results, derived from 10000 potential solutions, account for variability in isotopic measurements, fractionation processes, and uncertainty in source proportion estimation. This methodological approach enables detailed quantification of source apportionment and provides critical insights into dominant contributors under varying atmospheric conditions.

3 Results and discussion

3.1 Vertical profile of meteorological parameters and NH₄⁺ concentration level

The meteorological parameters observed at the Canton Tower exhibit strong associations with altitude (Fig.1). For example, the average temperature during the sampling campaign was 24.6 ± 5.1 °C at the ground site, while it decreased to 23.1 ± 4.8 °C at 118 m and 20.3 ± 4.5 °C at 488 m (Table S1). By contrast, wind speed exhibited a significant increasing trend with increasing height. During the entire observation period, the wind speeds at the ground site, 118 m, and 488 m were 0.63 ± 0.2, 2.2 ± 0.9, and 3.1 ± 1.2 m/s, respectively. Obviously, this vertical variation in meteorological parameters is beneficial for the transport of air pollutants from the ground to higher altitudes. Furthermore, the atmospheric environment at 488 m is more susceptible to the impacts of regional transportation due to its strongest wind speed among the three heights.

The average NH₄⁺ concentrations at the ground site, 118 m, and 488 m were 2.7 ± 1.4, 3.0 ± 1.8, and 2.6 ± 1.7 μg/m³, respectively. However, the differences in NH₄⁺ concentrations among the three heights were not statistically significant (p > 0.05) (Fig.1), suggesting that the NH₄⁺ was uniformly mixed across the heights during the 48-h sampling period. We further investigated the correlation between NH₄⁺ concentrations and meteorological parameters. The results showed that all NH₄⁺ concentrations at all three heights were not significantly correlated with temperature or humidity (p > 0.05), with low R values (Figs. S1–S3). However, significant and negative correlations were observed between NH₄⁺ concentrations and wind speed at 118 m and 488 m (p < 0.05) (Figs. S2 and S3), suggesting that NH₄⁺ levels at higher altitudes are more influenced by regional transportation. By contrast, no significant correlation was found between NH₄⁺ concentration and wind speed at the ground site (p > 0.05) (Fig. S1), reflecting the strong impact of local NH₃ emission on the ground NH₄⁺ concentration. Overall, the average mass percentage of NH₄⁺ in PM_2.5 in this study was approximately 6% and PM_2.5 concentrations at all heights showed a strong increase alongside rising NH₄⁺ concentrations (p < 0.001), indicating that NH₄⁺ is a key component regulating the PM_2.5 pollution in Guangzhou area.

3.2 Vertical profile of δ¹⁵N-NH₄⁺ signature

Although NH₄⁺ concentrations did not vary significantly across altitudes, we found that δ¹⁵N-NH₄⁺ values at Canton Tower exhibited significant variation with heights (p < 0.05) (Fig.1 and Tab.1). The average δ¹⁵N-NH₄⁺ values at the ground site, 118 m, and 488 m were 16‰ ± 5.2‰, 12‰ ± 3.8‰, and 7.4‰ ± 7.1‰, respectively. The observed δ¹⁵N-NH₄⁺ values in this study displayed a decreasing trend as NH₄⁺ concentrations increased, with a more pronounced negative correlation at higher altitudes (Fig. S1–S3), indicating that the increase in NH₄⁺ concentration at high heights is associated with greater ¹⁵N depletion. The δ¹⁵N-NH₄⁺ values in this study were approximately 2 to 3 times higher than those observed in Beijing, North China, using a 325 m meteorological tower (Table S2). Given the substantial difference in δ¹⁵N-NH₄⁺ values between the two cities, we inferred that NH₃ in Guangzhou may be more impacted by anthropogenic sources characterized by δ¹⁵N enrichment, resulting in higher δ¹⁵N compared to Beijing. Additionally, we found that the δ¹⁵N-NH₄⁺ values in both Guangzhou and Beijing decreased with increasing height (Table S2) (Wu et al., 2019; Wu et al., 2024), suggesting that the upper atmosphere is more susceptible to low δ¹⁵N-NH₃ sources such as livestock and fertilizer application in the adjacent areas.

3.3 δ¹⁵N-NH₃ endmembers of sources

The current bottom-up emission inventory indicates that nearly all NH₃ in the PRD region comes from the emissions of vehicle exhaust, biomass burning, coal combustion, NH₃ slip, livestock, fertilizer application, and waste (Huang et al., 2021). Therefore, it is essential to revisit the δ¹⁵N-NH₃ endmembers of these sources to improve the accuracy of NH₃ source apportionment based on ¹⁵N measurements. Since a large ¹⁵N diffusion fractionation occurs during the process of passive sampling for NH₃ (Pan et al., 2020), only δ¹⁵N-NH₃ values of the sources using the active sampling technique were considered in this study.

In total, more than 300 δ¹⁵N-NH₃ values for 7 sources were obtained based on 13 studies (Table S3). Obviously, the δ¹⁵N-NH₃ values in the combustion-related sources were significantly higher than those in volatilization sources. Also, all δ¹⁵N-NH₃ values among sources were significantly different (p < 0.05), except for the values between fertilizer application and livestock (p > 0.05) (Fig.2), suggesting that it is preferable to combine fertilizer application and livestock into a single agricultural source. Therefore, the δ¹⁵N-NH₃ endmembers were categorized as follows (Table S4): agriculture (−25‰ ± 6.1‰), waste (−33‰ ± 5.0‰), coal combustion (−18‰ ± 6.5‰), biomass burning (−11‰ ± 6.4‰), NH₃ slip (−2.8‰ ± 7.8‰), and vehicle exhaust (3.4‰ ± 5.0‰). Furthermore, the six primary ammonia sources were divided into two distinct types: combustion-related sources (−7.1‰ ± 3.2‰) and volatilization sources (−28‰ ± 3.2‰). Combustion-related sources include vehicle exhaust, NH₃ slip, and coal combustion, while volatilization sources comprise agriculture and waste.

3.4 Source apportionment of NH₃ in different heights in the PRD region

In this study, the average of the initial δ¹⁵N-NH₃ values at the ground site, 118 m, and 488 m were −2.3‰ ± 5.2‰, −5.8‰ ± 3.9‰, and −11‰ ± 7.3‰, respectively. The average value of the initial δ¹⁵N-NH₃ at the ground level was 2.5 and 4.7 times higher than those at 118 m and 488 m, indicating that NH₃ at the ground level is more influenced by δ¹⁵N-rich emission sources than by those in the higher atmosphere. Fig.3 illustrates the source apportionment of atmospheric NH₃ using the estimated initial δ¹⁵N-NH₃ values based on two assumptions. Whether NH₃ emissions are assumed to originate from six sources (vehicle, NH₃ slip, biomass burning, coal combustion, agriculture, waste) or two sources (combustion and volatilization), the results consistently indicate that combustion sources are significant contributors to NH₃ in this megacity of the PRD region. In the case of six sources, the largest source of NH₃ at the ground site was vehicle (29%), followed by NH₃ slip (25%), biomass burning (16%), coal combustion (12%), agriculture (10%), and waste (8%), meaning that the contributions of combustion sources and volatilization sources were 82% and 18%, respectively. This result is very close to the scenario where only two sources are considered in the SIAR model, which showed that the contribution of combustion sources to NH₃ at the ground site was 87% with the remaining 13% from volatilization sources (Fig.3). This strong agreement between the two scenarios suggests that the dataset of the δ¹⁵N-NH₃ endmembers in sources is reliable for source apportionment of NH₃. It is important to note that the influence of volatilization sources on NH₃ becomes more pronounced with increasing height, as confirmed by observations in North China (Zhang et al., 2020), mainly due to the limited influence of local combustion-related emissions at higher altitudes.

Different seasonal characteristics of NH₃ source apportionments were found at different heights (Fig.4). At the ground site, which is easily influenced by the local sources, the largest NH₃ source is vehicle with a contribution of 32% in autumn, followed by NH₃ slip (28%), biomass burning (15%), coal combustion (11%), agriculture (8%), and waste (6%). This meant that the volatilization sources accounted for only 14% during this season. Similar source contribution profiles were also observed during winter and spring at the ground site. However, in summer, the volatilization sources contributed as much as 40% due to the high temperatures during this season, which may strongly enhance the emission of NH₃ from these types of sources. The seasonal variation of the NH₃ source contribution profile at 488 m is notably different from that at the ground site. The contributions of all sources exhibit a relatively stable level throughout the year at an altitude of 488 m, indicating that NH₄⁺ particles at this height are less influenced by direct NH₃ emissions from various sources and can serve as a more reliable receptor site for the PRD region. Nevertheless, the contribution of volatilization sources (27%–35%) to NH₃ at 488 m remains considerably lower than that of combustion sources (65%–73%) throughout all seasons. Taken together, all the seasonal and vertical variations in the source apportionment of NH₃ in this study strongly suggest that combustion activity is the dominant source of NH₃ in the PRD region.

3.5 Uncertainty analysis of f on the results of the source apportionment using SIAR

In this study, the f value was obtained from the previous research conducted in the PRD region (Yue et al., 2015), which is comparable to those in other cities in China (Table S5). A prior study found that the NH₃ concentrations did not vary significantly across different altitudes in urban areas of China (Zhang et al., 2018). The NH₄⁺ concentrations in this study also showed no significant variation with altitude and aligned with the vertical distribution patterns of NH₃ in urban areas of China. In addition, the NH₄⁺ concentrations were insignificantly correlated with temperature or humidity (p > 0.05). These findings indicate that the gas-to-particle conversion of atmospheric NH₃ did not vary significantly across different altitudes in Guangzhou during the sampling period. Therefore, we used the f value from the near-ground levels at all altitudes, consistent with previous studies conducted at the 325 m meteorological tower in Beijing (Wu et al., 2019; Zhang et al., 2020; Wu et al., 2024).

To assess the impact of the f value on the results of the source apportionment of NH₃, a detailed sensitivity analysis of the f value in the SIAR model was performed. When the f values were adjusted by decreasing or increasing them by 10%, 20%, 30%,40%, and 50%, the average uncertainties for the NH₃ source apportionments from agricultural, waste, vehicle exhaust, biomass burning, NH₃ slip, and coal combustion were 18%, 26%, 18%, 3.6%, 18%, and 10%, respectively. The average uncertainties in the NH₃ source apportionment at ground, 118 m, and 488 m were 20%, 14%, and 13%, respectively. Fig.5 demonstrates that combustion sources were the dominant contributors to NH₃ at all heights, and their contribution decreased with increasing altitude regardless of how the f values varied within the range of sensitivity analyses.

4 Conclusions

We measured NH₄⁺ concentrations and their δ¹⁵N signatures at three atmospheric observatories (ground, 118 m, 488 m) at the Canton Tower in Guangzhou, China. The average concentrations of NH₄⁺ were 2.7 ± 1.4 μg/m³ at the ground site, 3.0 ± 1.8 μg/m³ at 118 m, and 2.6 ± 1.7 μg/m³ at 488 m. A strong positive correlation was observed between NH₄⁺ and PM_2.5 concentrations at all heights (p < 0.001), suggesting that NH₄⁺ is a key component driving PM_2.5 formation in the PRD region. Significant differences were found in δ¹⁵N-NH₄⁺ signatures among the heights (p < 0.01), with values of 16‰ ± 5.2‰ at the ground site, 12‰ ± 3.8‰ at 118 m, and 7.4‰ ± 7.1‰ at 488 m. Additionally, we calculated the initial δ¹⁵N-NH₃ values, and the SIAR model revealed that the contributions from different sources varied with height. On average, the contributions from agriculture, waste, vehicle, biomass burning, NH₃ slip, and coal combustion at the ground site were 9.9% ± 4.4%, 8.3% ± 5.5%, 29% ± 8.0%, 16% ± 2.2%, 25% ± 6.0%, and 12% ± 3.4%, respectively. By contrast, the corresponding contributions were 14% ± 3.6%, 13% ± 5.7%, 20% ± 5.9%, 17% ± 0.84%, 19% ± 4.0%, and 16% ± 1.4% at 118 m, and 16% ± 3.5%, 17% ± 8.1%, 17% ± 5.7%, 16% ± 2.2%, 17% ± 4.0%, and 16% ± 1.6% at 488 m, respectively. Overall, our findings suggest that combustion sources are the predominant contributors to the atmospheric NH₄⁺ loading in this highly urbanized region of China, rather than volatilization sources. These insights should be seriously considered in future NH₃ emission reduction strategies and PM_2.5 pollution mitigation efforts.

References

[1] Berner A H, David Felix J. (2020). Investigating ammonia emissions in a coastal urban airshed using stable isotope techniques. Science of the Total Environment, 707: 134952

[2] Bhattarai N, Wang S, Xu Q, Dong Z, Chang X, Jiang Y, Zheng H. (2020). Sources of gaseous NH₃ in urban Beijing from parallel sampling of NH₃ and NH₄⁺, their nitrogen isotope measurement and modeling. Science of the Total Environment, 747: 141361

[3] Bokhoven C, Theeuwen H J. (1966). Determination of the abundance of carbon and nitrogen isotopes in Dutch coals and natural gas. Nature, 211(5052): 927–929

[4] Chang Y, Cheng K, Kuang Y, Hu Q, Gao Y, Huang R J, Huang C, Walters W W, Lehmann M F. (2022). Isotopic variability of ammonia (δ¹⁵N-NH₃) slipped from heavy-duty vehicles under real-world conditions. Environmental Science & Technology Letters, 9(9): 726–732

[5] Chang Y, Zhang Y L, Kawichai S, Wang Q, Van Damme M, Clarisse L, Prapamontol T, Lehmann M F. (2021). Convergent evidence for the pervasive but limited contribution of biomass burning to atmospheric ammonia in peninsular Southeast Asia. Atmospheric Chemistry and Physics, 21(9): 7187–7198

[6] Chen Z, Pei C, Liu J, Zhang X, Ding P, Dang L, Zong Z, Jiang F, Wu L, Sun X, Zhou S, Zhang Y, Zhang Z, Zheng J, Tian C, Li J, Zhang G. (2022). Non-agricultural source dominates the ammonium aerosol in the largest city of South China based on the vertical δ¹⁵N measurements. Science of the Total Environment, 848: 157750

[7] David Felix J, Elliott E M, Gish T J, Mcconnell L L, Shaw S L. (2013). Characterizing the isotopic composition of atmospheric ammonia emission sources using passive samplers and a combined oxidation-bacterial denitrifier approach. Rapid Communications in Mass Spectrometry, 27(20): 2239–2246

[8] EEA(2023). European Union Emission Inventory Report 1990–2022—Under the UNECE Convention on Long-range Transboundary Air Pollution (Air Convention). Copenhagen, Denmark: EEA

[9] Farren N J, Davison J, Rose R A, Wagner R L, Carslaw D C. (2020). Underestimated ammonia emissions from road vehicles. Environmental Science & Technology, 54(24): 15689–15697

[10] Felix J D, Berner A, Wetherbee G A, Murphy S F, Heindel R C. (2023). Nitrogen isotopes indicate vehicle emissions and biomass burning dominate ambient ammonia across Colorado’s Front Range urban corridor. Environmental Pollution, 316: 120537

[11] Feng S, Xu W, Cheng M, Ma Y, Wu L, Kang J, Wang K, Tang A, Collett J L Jr, Fang Y. . (2022). Overlooked nonagricultural and wintertime agricultural NH₃ emissions in Quzhou County, North China Plain: evidence from ¹⁵N-stable isotopes. Environmental Science & Technology Letters, 9(2): 127–133

[12] Feng X, Chen Y, Du H, Feng Y, Mu Y, Chen J. (2023). Biomass burning is a non-negligible source for ammonium during winter haze episodes in rural North China: evidence from high time resolution ¹⁵N-stable isotope. Journal of Geophysical Research: Atmospheres, 128(3): e2022JD038012

[13] Feryer H D. (1978). Seasonal trends of NH₄⁺ and NO₃⁻ nitrogen isotope composition in rain collected at Jiilich, Germany. Tellus, 30(1): 83–92

[14] Gu B, Zhang L, Dingenen R V, Vieno M, Grinsven H J V, Zhang X, Zhang S, Chen Y, Wang S, Ren C. . (2021). Abating ammonia is more cost-effective than nitrogen oxides for mitigating PM_2.5 air pollution. Science, 374(6568): 758–762

[15] Heaton T H E. (1987). ¹⁵N/¹⁴N ratios of nitrate and ammonium in rain at Pretoria, South Africa. Atmospheric Environment, 21(4): 843–852

[16] Huang Z, Zhong Z, Sha Q, Xu Y, Zhang Z, Wu L, Wang Y, Zhang L, Cui X, Tang M. . (2021). An updated model-ready emission inventory for Guangdong Province by incorporating big data and mapping onto multiple chemical mechanisms. Science of the Total Environment, 769: 144535

[17] Jiang F, Liu J, Cheng Z, Ding P, Xu Y, Zong Z, Zhu S, Zhou S, Yan C, Zhang Z. . (2022). Dual-carbon isotope constraints on source apportionment of black carbon in the megacity Guangzhou of the Pearl River Delta region, China for 2018 autumn season. Environmental Pollution, 294: 118638

[18] Jiang F, Liu J, Cheng Z, Ding P, Zhu S, Yuan X, Chen W, Zhang Z, Zong Z, Tian C. . (2023). Quantitative evaluation for the sources and aging processes of organic aerosols in urban Guangzhou: insights from a comprehensive method of dual-carbon isotopes and macro tracers. Science of the Total Environment, 888: 164182

[19] Kawashima H, Ono S. (2019). Nitrogen isotope fractionation from ammonia gas to ammonium in particulate ammonium chloride. Environmental Science & Technology, 53(18): 10629–10635

[20] Kawashima H, Yoshida O, Suto N. (2023). Long-term source apportionment of ammonium in PM_2.5 at a suburban and a rural site using stable nitrogen isotopes. Environmental Science & Technology, 57(3): 1268–1277

[21] Li B, Chen L, Shen W, Jin J, Wang T, Wang P, Yang Y, Liao H. (2021). Improved gridded ammonia emission inventory in China. Atmospheric Chemistry and Physics, 21(20): 15883–15900

[22] Li L, Lollar B S, Li H, Wortmann U G, Lacrampe-Couloume G. (2012). Ammonium stability and nitrogen isotope fractionations for NH₄⁺-NH₃(aq)-NH₃(gas) systems at 20–70 °C and pH of 2–13: applications to habitability and nitrogen cycling in low-temperature hydrothermal systems. Geochimica et Cosmochimica Acta, 84: 280–296

[23] Li M, Liu H, Geng G, Hong C, Liu F, Song Y, Tong D, Zheng B, Cui H, Man H. . (2017a). Anthropogenic emission inventories in China: a review. National Science Review, 4(6): 834–866

[24] Li Y, Liu J, George C, Herrmann H, Gu M, Yang M, Wang Y, Mellouki A, Pan Y, Felix J D. . (2023). Apportioning atmospheric ammonia sources across spatial and seasonal scales by their isotopic fingerprint. Environmental Science & Technology, 57(43): 16424–16434

[25] Li Y, Thompson T M, Van Damme M, Chen X, Benedict K B, Shao Y, Day D, Boris A, Sullivan A P, Ham J. . (2017b). Temporal and spatial variability of ammonia in urban and agricultural regions of northern Colorado, United States. Atmospheric Chemistry and Physics, 17(10): 6197–6213

[26] Liu D, Fang Y, Tu Y, Pan Y. (2014). Chemical method for nitrogen isotopic analysis of ammonium at natural abundance. Analytical Chemistry, 86(8): 3787–3792

[27] Liu J, Ding P, Zong Z, Li J, Tian C, Chen W, Chang M, Salazar G, Shen C, Cheng Z. . (2018). Evidence of rural and suburban sources of urban haze formation in China: a case study from the Pearl River Delta region. Journal of Geophysical Research. Atmospheres, 123(9): 4712–4726

[28] Nirmalkar J, Jung J, Han S, Dong Z, Xu Z, Fu P, Pavulari C M. (2023). Chemistry of PM_2.5 in haze events in two East Asian cities during winter–spring 2019. Atmospheric Environment, 293: 119457

[29] Pan Y, Gu M, Song L, Tian S, Wu D, Walters W W, Yu X, Lü X, Ni X, Wang Y. . (2020). Systematic low bias of passive samplers in characterizing nitrogen isotopic composition of atmospheric ammonia. Atmospheric Research, 243: 105018

[30] Pan Y, Tian S, Liu D, Fang Y, Zhu X, Zhang Q, Zheng B, Michalski G, Wang Y. (2016). Fossil fuel combustion-related emissions dominate atmospheric ammonia sources during severe haze episodes: evidence from ¹⁵N-stable isotope in size-resolved aerosol ammonium. Environmental Science & Technology, 50(15): 8049–8056

[31] Parnell A C, Inger R, Bearhop S, Jackson A L. (2010). Source partitioning using stable isotopes: coping with too much variation. PLoS One, 5(3): e9672

[32] Sahoo P, Sahu S K, Mangaraj P, Mishra A, Beig G, Gunthe S S. (2024). Reporting of gridded ammonia emission and assessment of hotspots across India: a comprehensive study of 24 anthropogenic sources. Journal of Hazardous Materials, 479: 135557

[33] Shao S, Zhang Y, Chang Y, Cao F, Lin Y, Mozaffar A, Hong Y. (2020). Online characterization of a large but overlooked human excreta source of ammonia in China’s urban atmosphere. Atmospheric Environment, 230: 117459

[34] Smirnoff A, Savard M M, Vet R, Simard M C. (2012). Nitrogen and triple oxygen isotopes in near‐road air samples using chemical conversion and thermal decomposition. Rapid Communications in Mass Spectrometry, 26(23): 2791–2804

[35] Song L, Walters W W, Pan Y, Li Z, Gu M, Duan Y, Lü X, Fang Y. (2021). ¹⁵N natural abundance of vehicular exhaust ammonia, quantified by active sampling techniques. Atmospheric Environment, 255: 118430

[36] U.S.EPA . (2023). . ,

[37] Urey H C. (1947). The thermodynamic properties of isotopic substances. Journal of the Chemical Society (Resumed), 1947(5): 562–581

[38] Van Damme M, Clarisse L, Whitburn S, Hadji-Lazaro J, Hurtmans D, Clerbaux C, Coheur P F. (2018). Industrial and agricultural ammonia point sources exposed. Nature, 564(7734): 99–103

[39] Vo T, Christiansen A E. (2024). Impact of recent agricultural ammonia increases on fine particulate matter burden over the Midwestern United States. ACS Earth & Space Chemistry, 8(11): 2209–2217

[40] Walters W W, Chai J, Hastings M G. (2019). Theoretical phase resolved ammonia–ammonium nitrogen equilibrium isotope exchange fractionations: applications for tracking atmospheric ammonia gas-to-particle conversion. ACS Earth & Space Chemistry, 3(1): 79–89

[41] Walters W W, Karod M, Willcocks E, Baek B H, Blum D E, Hastings M G. (2022). Quantifying the importance of vehicle ammonia emissions in an urban area of northeastern USA utilizing nitrogen isotopes. Atmospheric Chemistry and Physics, 22(20): 13431–13448

[42] Walters W W, Song L, Chai J, Fang Y, Colombi N, Hastings M G. (2020). Characterizing the spatiotemporal nitrogen stable isotopic composition of ammonia in vehicle plumes. Atmospheric Chemistry and Physics Discussion, 20(19): 11551–11567

[43] Wang W, Gao Y, Shao L, Fan C, Li X, Li Y, Liu M, Zhou X. (2023a). Chemical compositions and possible transportation of PM_2.5 during two haze periods in a coastal city of the North China Plain. Geological Journal, 58(12): 4417–4427

[44] Wang Y, Liu J, Jiang F, Chen Z, Wu L, Zhou S, Pei C, Kuang Y, Cao F, Zhang Y. . (2023b). Vertical measurements of stable nitrogen and oxygen isotope composition of fine particulate nitrate aerosol in Guangzhou city: source apportionment and oxidation pathway. Science of the Total Environment, 865: 161239

[45] Weagle C L, Snider G, Li C, Van Donkelaar A, Philip S, Bissonnette P, Burke J, Jackson J, Latimer R, Stone E. . (2018). Global sources of fine particulate matter: interpretation of PM_2.5 chemical composition observed by SPARTAN using a global chemical transport model. Environmental Science & Technology, 52(20): 11670–11681

[46] Wei Y, Tian X, Huang J, Wang Z, Huang B, Liu J, Gao J, Liang D, Yu H, Feng Y. . (2023). New insights into the formation of ammonium nitrate from a physical and chemical level perspective. Frontiers of Environmental Science & Engineering, 17(11): 137

[47] Wu L, Ren H, Wang P, Chen J, Fang Y, Hu W, Ren L, Deng J, Song Y, Li J. . (2019). Aerosol ammonium in the urban boundary layer in Beijing: insights from nitrogen isotope ratios and simulations in summer 2015. Environmental Science & Technology Letters, 6(7): 389–395

[48] Wu L, Wang P, Zhang Q, Ren H, Shi Z, Hu W, Chen J, Xie Q, Li L, Yue S. . (2024). Dominant contribution of combustion-related ammonium during haze pollution in Beijing. Science Bulletin, 69(7): 978–987

[49] Xiao H W, Wu J F, Luo L, Liu C, Xie Y J, Xiao H. (2020). Enhanced biomass burning as a source of aerosol ammonium over cities in central China in autumn. Environmental Pollution, 266: 115278

[50] Xu W, Zhao Y, Wen Z, Chang Y, Pan Y, Sun Y, Ma X, Sha Z, Li Z, Kang J J S B. . (2022). Increasing importance of ammonia emission abatement in PM_2.5 pollution control. Science Bulletin, 67(17): 1745–1749

[51] Yue D, Zhong L, Zhang T, Shen J, Zhou Y, Zeng L, Dong H, Ye S. (2015). Pollution properties of water-soluble secondary inorganic ions in atmospheric PM_2.5 in the Pearl River Delta Region. Aerosol and Air Quality Research, 15(5): 1737–1747

[52] Zhang Y, Benedict K B, Tang A, Sun Y, Fang Y, Liu X. (2020). Persistent nonagricultural and periodic agricultural emissions dominate sources of ammonia in urban Beijing: evidence from ¹⁵N stable isotope in vertical profiles. Environmental Science & Technology, 54(1): 102–109

[53] Zhang Y, Ma X, Tang A, Fang Y, Misselbrook T, Liu X. (2023). Source apportionment of atmospheric ammonia at 16 sites in China using a Bayesian isotope mixing model based on δ¹⁵N-NH_x signatures. Environmental Science & Technology, 57(16): 6599–6608

[54] Zhang Y, Tang A, Wang D, Wang Q, Benedict K, Zhang L, Liu D, Li Y, Collett J L Jr, Sun Y. . (2018). The vertical variability of ammonia in urban Beijing, China. Atmospheric Chemistry and Physics, 18(22): 16385–16398

[55] Zheng M, Wang Y, Yuan L, Chen N, Kong S. (2022). Ambient observations indicating an increasing effectiveness of ammonia control in wintertime PM_2.5 reduction in Central China. Science of the Total Environment, 824: 153708

RIGHTS & PERMISSIONS

Higher Education Press 2025