The critical role of volatile organic compound emissions in nitrate formation in Lhasa, Tibetan Plateau: insights from oxygen isotope anomaly measurements

Atmospheric particulate nitrate aerosol (${\mathrm{NO}}_{\mathrm{3}}^{-}$), produced via the oxidation of nitrogen oxides (${\mathrm{NO}}_{\mathit{\text{x}}}=\mathrm{NO}+{\mathrm{NO}}_{\mathrm{2}}$), plays an important role in atmospheric chemistry and air quality, yet its formation mechanism remains poorly constrained in the plateau region. In this study, we report for the first time the yearly variation in the signatures of the stable oxygen isotope anomaly ($\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}=\mathit{\delta }{}^{\mathrm{17}}\mathrm{O}-\mathrm{0.52}\times {\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$) in ${\mathrm{NO}}_{\mathrm{3}}^{-}$ collected in the urban Lhasa (3650 meters above sea level (m a.s.l.)), on the Tibetan Plateau, China. Our results show that NO₂+OH was the largest contributor to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation (46±26 %), followed by NO₃+volatile organic compound (VOC) (26±18 %) and N₂O₅+H₂O (28±11 %) using the Bayesian Isotope Mixture Model. Notably, there were significant differences in the NO₂+OH, NO₃+VOC, and N₂O₅+H₂O pathways between spring and the other three seasons (T test, p<0.05). Using the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) dispersion model, we highlighted the influence of VOC emissions from regions such as Afghanistan and northern India, which enhanced ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations in Lhasa during spring. Furthermore, the diurnal distribution of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ oxidation pathways varied distinctly across seasons, suggesting that these differences in ${\mathrm{NO}}_{\mathrm{3}}^{-}$ pathways were attributed to aerosol liquid water content (ALWC), VOC concentrations, and the atmospheric lifetime of ${\mathrm{NO}}_{\mathrm{3}}^{-}$.

1 Introduction

Nitrate aerosol (${\mathrm{NO}}_{\mathrm{3}}^{-}$) is a key component regulating the mass concentration of atmospheric fine particulate matter (PM_2.5), which is highly related to air quality (Colmer et al., 2020), public health (Zhang et al., 2019, 2017; Geng et al., 2021), and the climate system (Clark and Tilman, 2008). Globally, the mass contribution of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ in PM_2.5 is in the range of 5 %–30 % (Huang et al., 2014; Xu et al., 2019; Salameh et al., 2015; Espina-Martin et al., 2024; Bell et al., 2007; Sun et al., 2022), depending on the locations and the severities of air pollution. For example, it was reported that ${\mathrm{NO}}_{\mathrm{3}}^{-}$ accounted for 22 %, 27 %, and 26 % of PM_2.5 in megacities in China (Zong et al., 2020), Europe (Espina-Martin et al., 2024), and the US (Sun et al., 2022), respectively. In addition, some studies found that the contribution of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ would increase by 3–8 times with the occurrence of the particle-derived haze pollution (Ge et al., 2024; Song et al., 2019; Yin et al., 2022; Walters et al., 2024).

It is well known that atmospheric ${\mathrm{NO}}_{\mathrm{3}}^{-}$ is formed by the oxidation of nitrogen oxides (${\mathrm{NO}}_{\mathit{\text{x}}}=\mathrm{NO}+{\mathrm{NO}}_{\mathrm{2}}$) with different oxidants such as O₃, OH, and RO₂ (Text S1 in the Supplement). In general, atmospheric chemical transportation models are employed to depict the detailed oxidation pathways of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation. However, there remains considerable uncertainty in modeling the contribution of individual oxidation pathways to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation, particularly the N₂O₅+H₂O pathway, due to the wide variability of key parameters such as the N₂O₅ uptake coefficient, which has been shown to vary significantly depending on aerosol composition, relative humidity, and temperature. For example, it was reported that the predicted N₂O₅ uptake to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation in Beijing, as estimated using WRF-Chem, ranged from 5 % to 21 % (Su et al., 2017). Higher contributions between 66 % and 85 % have been observed when applying the CMAQ model in Beijing (Qiu et al., 2019). Therefore, the application of alternative techniques is crucial for providing more reliable estimates and enhancing our understanding of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation mechanisms, in addition to the insights gained from atmospheric chemical transportation models.

Stable oxygen isotope anomaly ($\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}={\mathit{\delta }}^{\mathrm{17}}\mathrm{O}-\mathrm{0.52}\times {\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$) is recognized as a powerful tool to track the formation pathways of atmospheric ${\mathrm{NO}}_{\mathrm{3}}^{-}$ (Zhang et al., 2024; Feng et al., 2023). This is because the oxygen atoms in the terminal positions of O₃ exhibit an elevated Δ¹⁷O ($\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}=\mathrm{39}\pm \mathrm{2}\mathrm{‰}$) (Vicars and Savarino, 2014), whereas the Δ¹⁷O values of other atmospheric oxidants (e.g., H₂O, OH, and RO₂) that can be incorporated into ${\mathrm{NO}}_{\mathrm{3}}^{-}$ are very close to 0 ‰. (Dubey et al., 1997; Barkan and Luz, 2003; Alexander et al., 2020) Therefore, $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}\left({\mathrm{NO}}_{\mathrm{3}}^{-}\right)$ serves as a unique tracer of O₃ involvement in its formation pathways, offering valuable insights into the relative contributions of individual reactions. In recent years, the use of $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}\left({\mathrm{NO}}_{\mathrm{3}}^{-}\right)$ to elucidate ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation has garnered considerable attention. Walters et al. (2024) reported that the major formation pathways of annual HNO₃ production in the northeastern US were NO₂+OH (46 %), N₂O₅ uptake (34 %), and organic nitrate hydrolysis (12 %), with notable seasonal variability. Additionally, Zhang et al. (2022) observed that the contribution of nocturnal chemistry to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation increased at night, peaking at 72 % around midnight. In contrast, the contribution of NO₂+OH rose with sunrise, reaching its highest fraction (48 %) around noon. However, nearly all current Δ¹⁷O-related observations have been conducted in the plain cities, with little attention given to plateau cities, where atmospheric conditions generally suffer from distinct energy consumption patterns and unique climatic factors (e.g., intense solar radiation). In this study, we present detailed results from comprehensive field observations conducted in Lhasa (3650 meters above sea level, m a.s.l.), one of the highest cities in the world, located on the Tibetan Plateau, China. For the first time, we quantify the relative contribution of three oxidation pathways to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation in Lhasa on the basis of ambient measurements for Δ¹⁷O signatures in ${\mathrm{NO}}_{\mathrm{3}}^{-}$.

2 Materials and methods

2.1 Sampling campaign

PM_2.5 samples were collected on the roof of a building (∼ 15 m above ground) at the Meteorological Bureau of Lhasa (91.08° E, 29.40° N; Fig. 1) in China. Lhasa, the capital of the Tibet Autonomous Region, is a rapidly developing city with a population of ∼950 000 and an urban area of ∼30 000 km² (Office of the Leading Group, 2025). The sampling site is surrounded by mixed land use, including residential areas, government offices, religious temples, and commercial zones, with minimal heavy industry. The strong solar radiation and large diurnal temperature variations in this sampling site can lead to pronounced changes in the boundary layer height, which in turn significantly influence vertical mixing and the transport of air pollutants.

The sampling campaign was conducted from June 2022 to July 2023 using a high-volume PM_2.5 sampler, which operated at a flow rate of 1.0 m³ min⁻¹. Samples were collected once a week, with each sampling session lasting 48 h, except during intensive sampling periods in the summer (30 June–14 July 2022) and winter (28 January–7 February 2023). During these intensive periods, each sample was collected for 12 h, from 8:00 to 20:00 LT and from 20:00 to 8:00 LT on the following day. During the autumn of 2022, Lhasa experienced intermittent COVID-19 control measures, including restricted movement, reduced traffic activity, and temporary lockdowns in urban areas (Lhasa Municipal People's Government, 2025). Before sampling, all quartz filters ($\mathrm{8}\mathrm{in}.\times \mathrm{10}\mathrm{in}.$, i.e., 20.3 cm×25.4 cm, Pallflex) were calcined in a muffle furnace at 450 °C for 6 h to prevent impurities from contaminating the collected PM_2.5 samples. After sampling, the samples were collected and stored in a freezer at −20 °C.

Figure 1 Geographic position of the sampling site in Lhasa, China.

2.2 Measurements of water-soluble ions and isotopes

Water-soluble ions were measured by ion chromatography (Dionex ICS-5000, Thermo Scientific Inc.) (Chen et al., 2022). In brief, a part of the filter membranes (4.54 cm²) was cut using a 17 mm diameter punch and placed in a 15 mL centrifuge tube with 10 mL of 18.2 MΩ ultrapure water. The tube was then subjected to ultrasonic treatment in an ice-water bath for 30 min to prevent ion volatilization. The extract was filtered through a 0.22 µm filter into a 30 mL sample bottle. This process was repeated with an additional 10 mL of water to ensure full extraction. The final extract was analyzed by ion chromatography. The method detection limits (MDLs) for Cl⁻, ${\mathrm{NO}}_{\mathrm{3}}^{-}$, ${\mathrm{SO}}_{\mathrm{4}}^{\mathrm{2}-}$, Na⁺, ${\mathrm{NH}}_{\mathrm{4}}^{+}$, K⁺, Mg²⁺, and Ca²⁺ were 0.001, 0.001, 0.003, 0.02, 0.01, 0.02, 0.006, and 0.02 mg L⁻¹, respectively.

Stable oxygen isotopes (δ¹⁷O, δ¹⁸O, Δ¹⁷O, and $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}=\mathit{\delta }{}^{\mathrm{17}}\mathrm{O}-\mathrm{0.52}\times {\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$) of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ were determined using an isotope ratio mass spectrometer (MAT253, Thermo Fisher Scientific, USA) at Nanjing University of Information Science and Technology (Fan et al., 2021; Zhang et al., 2022). Briefly, ${\mathrm{NO}}_{\mathrm{3}}^{-}$ from filter extractions (containing at least 0.8 µg N) was converted into gaseous N₂O using the bacterial denitrifier method. N₂O was then further thermally decomposed into O₂ and N₂ in a gold tube heated to 800 °C. The produced O₂ was analyzed for oxygen isotopes by an isotope ratio mass spectrometer. The duplicate analysis showed that the errors were within 1.32 % for $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$.

2.3 Primary data sources

Meteorological parameters, including ambient temperature (T), relative humidity (RH), rainfall, radiation, wind direction (WD), and wind speed (WS) during the sampling campaign, were obtained from the Meteorological Bureau of Lhasa. Additionally, NO₂ and O₃ during the sampling campaign were downloaded from the National Meteorological Information Center (https://air.cnemc.cn:18007/, last access: 21 December 2023).

2.4 Evaluation of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ oxidation pathways

In our study, we aimed to quantify the relative contribution of different oxidation pathways to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ production based on $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$. Due to the low Cl⁻ concentrations observed in Lhasa, the ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation pathways considered in this study were limited to NO₂+OH, NO₃+volatile organic compound (VOC), and N₂O₅+H₂O. Although NO₃+VOC is generally considered a minor pathway in continental regions (Alexander et al., 2009), we included it because elevated VOC concentrations were observed at our sampling site in Lhasa, influenced by both biomass burning (e.g., incense burning) and anthropogenic sources (e.g., vehicle emissions) (Tang et al., 2022). The relative contributions of the three pathways were determined using a Δ¹⁷O-based mass balance approach (Michalski et al., 2003), as shown in Eqs. (1) and (2):

$$\begin{array}{}\text{(1)}& {\displaystyle }& {\displaystyle }\begin{array}{rl}{\displaystyle }& {\displaystyle }\mathrm{\Delta }{{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}}^{-}=(\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}{)}_{{\mathrm{NO}}_{\mathrm{2}}+\mathrm{OH}}\\ {\displaystyle }& {\displaystyle }\times f_{{\mathrm{NO}}_{\mathrm{2}}+\mathrm{OH}}+(\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}{)}_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}\times f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}\\ {\displaystyle }& {\displaystyle }+(\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}{)}_{{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}}\times f_{{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}}\end{array}\text{(2)}& {\displaystyle }& {\displaystyle }{f}_{{\mathrm{NO}}_{\mathrm{2}}+\mathrm{OH}}+f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}+f_{{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}}=\mathrm{1}\end{array}$$

where $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ value is the Δ¹⁷O value of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ in PM_2.5. The ($\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}{)}_{{\mathrm{NO}}_{\mathrm{2}}+\mathrm{OH}}$, $(\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}{)}_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$, and $(\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}{)}_{{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}}$ correspond to the Δ¹⁷O values from NO₂+OH, NO₃+VOC, and N₂O₅+H₂O, respectively. The Δ¹⁷O values for each pathway were calculated using Eqs. (3)–(5) (Savarino et al., 2016; Alexander et al., 2009):

$$\begin{array}{}\text{(3)}& {\displaystyle }& {\displaystyle }\left(\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}{)}_{{\mathrm{NO}}_{\mathrm{2}}+\mathrm{OH}}\right(\mathrm{‰})={\displaystyle \frac{\mathrm{2}}{\mathrm{3}}}\mathit{\alpha }\times \mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }\text{(4)}& {\displaystyle }& {\displaystyle }\begin{array}{rl}{\displaystyle }& {\displaystyle }\left(\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}{)}_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}\right(\mathrm{‰})={\displaystyle \frac{\mathrm{2}}{\mathrm{3}}}\mathit{\alpha }\times \mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }\\ {\displaystyle }& {\displaystyle }+{\displaystyle \frac{\mathrm{1}}{\mathrm{3}}}\times \mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }\end{array}\text{(5)}& {\displaystyle }& {\displaystyle }\begin{array}{rl}{\displaystyle }& {\displaystyle }(\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}{)}_{{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}(\mathrm{‰})={\displaystyle \frac{\mathrm{1}}{\mathrm{3}}}\mathit{\alpha }\times \mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }\\ {\displaystyle }& {\displaystyle }+{\displaystyle \frac{\mathrm{1}}{\mathrm{2}}}\left({\displaystyle \frac{\mathrm{2}}{\mathrm{3}}}\mathit{\alpha }\times \mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }+{\displaystyle \frac{\mathrm{1}}{\mathrm{3}}}\times \mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }\right)\end{array}\end{array}$$

Previous studies have demonstrated a linear correlation between Δ¹⁷O−O₃ and $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }$ with Δ¹⁷O−O₃ values ranging from 20 % to 40 % in tropospheric O₃ (Vicars and Savarino, 2014; Ishino et al., 2017). The equations are shown as follows (Vicars et al., 2012):

$$\begin{array}{}\text{(6)}& \mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }=\mathrm{1.5}\times \mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}\end{array}$$

Based on previous observations of tropospheric O₃, the $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }$ average value was approximately 39 ‰. The α value represents the proportional contribution of O₃ to the NO oxidation pathway and can be estimated using the following Eqs. (7) (Alexander et al., 2009). When NO_x is in photochemical steady state, Δ¹⁷O−NO₂ can be represented using the following Eq. (10):

$$\begin{array}{}\text{(7)}& {\displaystyle }\mathit{\alpha }={\displaystyle \frac{K_{p\mathrm{1}}\left[{\mathrm{O}}_{\mathrm{3}}\right]\times \left[\mathrm{NO}\right]}{K_{p\mathrm{1}}\left[{\mathrm{O}}_{\mathrm{3}}\right]\times \left[\mathrm{NO}\right]+K_{p\mathrm{2}}\left[{\mathrm{HO}}_{\mathrm{2}}\right]\times \left[\mathrm{NO}\right]+K_{p\mathrm{3}}\left[{\mathrm{RO}}_{\mathrm{2}}\right]\times \left[\mathrm{NO}\right]}}\text{(8)}& {\displaystyle }{K}_{p\mathrm{1}}=\mathrm{3.0}\times {\mathrm{10}}^{-\mathrm{2}}\times {\mathrm{e}}^{\left(-\frac{\mathrm{1500}}{T}\right)}\text{(9)}& {\displaystyle }{K}_{p\mathrm{2}}=K_{p\mathrm{3}}=\mathrm{3.5}\times {\mathrm{10}}^{-\mathrm{12}}\times {\mathrm{e}}^{\frac{\mathrm{270}}{T}}\text{(10)}& {\displaystyle }\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{2}}=\mathit{\alpha }\times \mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }\end{array}$$

where T represents the ambient temperature (K) (Kunasek et al., 2008). The HO₂ mixing ratios were estimated using empirical equations in the absence of direct HO₂ observations (Kanaya et al., 2007). Due to the lower temperatures in Lhasa during non-summer seasons, HO₂ concentrations were assessed using a formula derived from winter conditions.

Winter

$$\begin{array}{}\text{(11)}& {\displaystyle }& {\displaystyle }\begin{array}{rl}{\displaystyle }& {\displaystyle }\left[{\mathrm{HO}}_{\mathrm{2}}\right]/\mathrm{ppt}=\exp (\mathrm{5.7747}\times {\mathrm{10}}^{-\mathrm{2}}[{\mathrm{O}}_{\mathrm{3}}\left]\right(\mathrm{ppb})-\mathrm{1.7227})\\ {\displaystyle }& {\displaystyle }\text{for daytime}\end{array}\text{(12)}& {\displaystyle }& {\displaystyle }\begin{array}{rl}{\displaystyle }& {\displaystyle }\left[{\mathrm{HO}}_{\mathrm{2}}\right]/\mathrm{ppt}=\exp (\mathrm{7.7234}\times {\mathrm{10}}^{-\mathrm{2}}[{\mathrm{O}}_{\mathrm{3}}\left]\right(\mathrm{ppb})-\mathrm{1.6363})\\ {\displaystyle }& {\displaystyle }\text{for nighttime}\end{array}\end{array}$$

Summer

$$\begin{array}{}\text{(13)}& {\displaystyle }& {\displaystyle }\begin{array}{rl}{\displaystyle }& {\displaystyle }\left[{\mathrm{HO}}_{\mathrm{2}}\right]/\mathrm{ppt}=\exp (\mathrm{2.0706}\times {\mathrm{10}}^{-\mathrm{2}}[{\mathrm{O}}_{\mathrm{3}}\left]\right(\mathrm{ppb})+\mathrm{1.0625})\\ {\displaystyle }& {\displaystyle }\text{for daytime}\end{array}\text{(14)}& {\displaystyle }& {\displaystyle }\begin{array}{rl}{\displaystyle }& {\displaystyle }\left[{\mathrm{HO}}_{\mathrm{2}}\right]/\mathrm{ppt}=\mathrm{0.2456}+\mathrm{0.1841}\left[{\mathrm{O}}_{\mathrm{3}}\right]\left(\mathrm{ppb}\right)\\ {\displaystyle }& {\displaystyle }\text{for nighttime}\end{array}\end{array}$$

2.5 Stable isotope analysis in the R (SIAR) model

In this study, stable isotope analysis in the R (SIAR) model was employed to estimate the relative contributions of three main pathways to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ (Parnell et al., 2010). The SIAR model is well-suited for analyzing multiple formation pathways, as it effectively incorporates uncertainties and parameter variability, leading to more reliable estimates. Specifically, this model allows for a detailed analysis of oxygen isotope (Δ¹⁷O), enabling accurate modeling of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation pathways based on oxygen isotope measurements. The SIAR model is a Bayesian mixture model, mathematically formulated as follows:

$$\begin{array}{rl}{\displaystyle }& {\displaystyle }{X}_i={\sum }_{j=\mathrm{1}}^K{p}_j\times f_{ij}\\ {\displaystyle }& {\displaystyle }{p}_{\mathrm{1}}+p_{\mathrm{2}}+\mathrm{\dots }+p_k=\mathrm{1}\\ {\displaystyle }& {\displaystyle }{f}_{ij}\sim N({\mathit{\mu }}_j,{\mathit{\omega }}_j^{\mathrm{2}})\end{array}$$

where X_i is the observed Δ¹⁷O value for sample i ($i=\mathrm{1},\mathrm{2},\mathrm{3},\mathrm{\dots },N$), and p_j is the proportional contribution of each ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation pathway j to the sample i. f_ij is the Δ¹⁷O value of formation pathway j for sample i and follows a normal distribution with mean (μ_j) and variance (${\mathit{\omega }}_j^{\mathrm{2}}$). Within the Bayesian framework, prior distributions are assigned to each p_j, and these are updated with the observed data X_i to obtain posterior distributions, allowing for inference of the proportional contributions p_j of each pathway.

2.6 Aerosol liquid water content (ALWC) and the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT)

To evaluate the influence of ALWC on ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation, ALWC was calculated using the ISORROPIA II model developed by Fountoukis and Nenes (Fountoukis and Nenes, 2007). The ISORROPIA II model includes two modes: the forward mode, which requires the concentrations of both particulate and gaseous pollutants as inputs, and the reverse mode, which only requires the concentrations of particulate pollutants. The model computes the ALWC in both modes based on particulate pollutant concentrations (e.g., ${\mathrm{NH}}_{\mathrm{4}}^{+}$, Na⁺, Ca²⁺, K⁺, and Mg²⁺), as well as ambient RH and T. In this study, the reverse mode was employed due to the lack of gaseous pollutant concentration observations.

Additionally, the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) model was utilized to compute 72 h back trajectories during the sampling campaign. HYSPLIT, developed by the National Oceanic and Atmospheric Administration Air Resources Laboratory, is available on their website (https://www.ready.noaa.gov/HYSPLIT.php, last access: 1 December 2024). This model has been widely used for simulating the transport and dispersion trajectories of pollutants such as PM_2.5, VOC, O₃, and NO_x, among others (He et al., 2022; Zhao et al., 2015; Cao et al., 2023). Backward trajectories for each sampling day were calculated at an altitude of 3650 m using meteorological data from the Global Data Assimilation System (GDAS), available through the US Air Resources Laboratory (https://www.ready.noaa.gov/data/archives/gdas1/, last access: 15 December 2023).

3 Results

3.1 Overview of the meteorological parameters in Lhasa during the sampling campaign

Figure 2a presents the daily variations in meteorological parameters, including T, RH, rainfall, and solar radiation. During the sampling campaign, the annual average T was 11.5 °C, ranging from −2.83 to 24.2 °C. The highest average T was observed in summer (19.7 °C), while the lowest (3.11 °C) was recorded in winter. RH varied between 6.67 % and 66.8 %, with the lowest average RH occurring in winter (17.1 %) and the highest in summer (35.6 %). The near-surface layer of Lhasa is influenced by a thermal low-pressure system, and the southwest monsoon, active between June and September, transports moisture-laden air from the Indian Ocean, resulting in increased rainfall during summer. Solar radiation intensity exhibited a seasonal trend consistent with those of T and RH, peaking in summer (394 W m⁻²) and reaching its lowest levels in winter (220 W m⁻²). The dominant WD was southeast in spring and southwest in the other three seasons (Fig. 3). WS was highest in spring and lowest in autumn.

Figure 2 (a) shows the time series of temperature (°C), relative humidity (%), rainfall (mm), radiation (W m⁻²), ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentration (µg m⁻³), and $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ (‰) from June 2022 to July 2023. (b) shows the average values at different seasons with their statistical significance.

Figure 3 The bivariate polar plot illustrates the seasonal variation in the mass concentration of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ in relation to wind speed (WS, m s⁻¹) and wind direction (WD, degrees).

3.2 ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentration

${\mathrm{NO}}_{\mathrm{3}}^{-}$ mass concentrations ranged from 0.10 to 1.72 µg m⁻³, with an average value of $\mathrm{0.62}\pm \mathrm{0.31}\mathrm{\mu }\mathrm{g}{\mathrm{m}}^{-\mathrm{3}}$. ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations exhibited distinct seasonal patterns. As shown in Fig. S1 in the Supplement, the equivalent concentrations of [${\mathrm{SO}}_{\mathrm{4}}^{\mathrm{2}-}+{\mathrm{NO}}_{\mathrm{3}}^{-}$] were considerably higher than those of [${\mathrm{NH}}_{\mathrm{4}}^{+}$], indicating that ${\mathrm{NH}}_{\mathrm{4}}^{+}$ was insufficient to fully neutralize ${\mathrm{NO}}_{\mathrm{3}}^{-}$. This suggests that a portion of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ may have existed in other forms, such as KNO₃ and Ca(NO₃)₂. This inference is supported by the strong positive correlations between ${\mathrm{NO}}_{\mathrm{3}}^{-}$ and K⁺ (r=0.64, p<0.1) and Ca²⁺ (r=0.43, p<0.01), especially in spring, as shown in Fig. S2. In contrast, ${\mathrm{NO}}_{\mathrm{3}}^{-}$ showed relatively weak negative correlations with T ($r=-\mathrm{0.27}$, p<0.01) and RH ($r=-\mathrm{0.22}$, p<0.1), indicating that under the specific atmospheric conditions in Lhasa, meteorological parameters might not be the dominant factors controlling the gas-particle partitioning of ${\mathrm{NO}}_{\mathrm{3}}^{-}$. The maximum monthly average values of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentration occurred in spring ($\mathrm{0.83}\pm \mathrm{0.35}\mathrm{\mu }\mathrm{g}{\mathrm{m}}^{-\mathrm{3}}$) with the instantaneous maximum reaching 1.72 µg m⁻³, whereas the lowest was recorded in autumn ($\mathrm{0.23}\pm \mathrm{0.13}\mathrm{\mu }\mathrm{g}{\mathrm{m}}^{-\mathrm{3}}$) with an instantaneous minimum of only 0.09 µg m⁻³ (Table 1). The elevated ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations in spring could be attributed to biomass burning from South and Southeast Asia (Figs. S3 and S4). The strong correlation between ${\mathrm{NO}}_{\mathrm{3}}^{-}$ and K⁺ in spring further supports this explanation.

In spring, high ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations were associated with weak southeasterly winds ($<\mathrm{3}\mathrm{m}{\mathrm{s}}^{-\mathrm{1}}$) in the bivariate polar plot, suggesting probable impacts from local emissions (Fig. 3). The southeasterly sector of the sampling site includes residential areas, agricultural land, and major transportation routes, which are potential NO_x sources. In spring, intensified agricultural activities (e.g., fertilization, biomass burning) might increase NO_x emissions. Meanwhile, low wind speeds likely limit atmospheric dispersion, promoting the local accumulation of precursors and enhancing ${\mathrm{NO}}_{\mathrm{3}}^{-}$ production. During the rainy summer, shorter ${\mathrm{NO}}_{\mathrm{3}}^{-}$ lifetimes indicated a weak influence from regional transport, with a more pronounced contribution from local emissions. In autumn, ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations were relatively low, which coincided with strict local COVID-19 restrictions in Lhasa. These measures significantly reduced human activity and traffic, leading to suppressed local emissions. Despite low wind speeds typically favoring pollutant accumulation, ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations remained low, suggesting that both reduced local sources and seasonal meteorological conditions constrained ${\mathrm{NO}}_{\mathrm{3}}^{-}$ production. Nevertheless, the persistence of measurable ${\mathrm{NO}}_{\mathrm{3}}^{-}$ under such stagnant conditions also implied a potential contribution from regional transport during this period. In winter, elevated ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations under low wind speeds ($<\mathrm{3}\mathrm{m}{\mathrm{s}}^{-\mathrm{1}}$) emphasized the significant contribution of local emissions. These findings underscored that both regional transport and local emissions were important contributors to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations in Lhasa. Furthermore, based on our day-night sampling scheme, no nychthemeral (day–night) differences in ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations were detected (Table S1 in the Supplement). A similar day-night pattern of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations has also been observed in Beijing (Luo et al., 2020).

Table 1 Average values of water-soluble ions and $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ during the sampling campaign.

3.3 Oxygen isotopes of ${\mathrm{NO}}_{\mathrm{3}}^{-}$

To explore the three major oxidation pathways of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation, 53 samples representing varying ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations across different seasons were selected for oxygen isotope measurements (Fig. 2b). The $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ values ranged from 18.3‰ to 34.1‰, with an average of 26.3±3.13 ‰, which is slightly lower than the global average of 28.6±4.5 ‰ simulated by the Global Chemical Transport Model (Alexander et al., 2020). As shown in Table S2, the observed $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ values in this study were similar to most mid- and low-latitude regions, but lower than those in polar regions (∼32 ‰). As listed in Table S1, the average $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ values in spring, summer, autumn, and winter were 28.8±8.0 ‰, 25.5±2.20 ‰, 25.6±1.35 ‰, and 25.9±3.56 ‰, respectively. The differences in $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ values between spring and summer, as well as between spring and winter, were statistically significant (p<0.05). The elevated $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ values in spring could be attributed to a higher proportion of nocturnal pathways that enrich $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ values, such as NO₃+VOC and N₂O₅+H₂O pathways. In contrast, the lower $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ values in the other three seasons suggested a greater production of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation via the NO₂+OH pathway, leading to more negative $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ values. Diurnal variation in $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ values also differed across seasons (Fig. S5). In summer, the average of $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ values during the day (25.3±2.39 ‰) was lower than at night (26.7±1.03 ‰). Conversely, in winter, the average of $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ values during the day (28.0±3.79 ‰) was significantly higher than at night (24.4±3.85 ‰). Similar diurnal patterns, with higher daytime $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ values and lower nighttime values, have also been observed in winter in the US (Vicars et al., 2013) and other cities in China (He et al., 2018).

4 Discussion

4.1 A comparison of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ oxidation pathways in Lhasa with other megacities in plain regions

Typically, observations of $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ and estimated α (the proportion of O₃ oxidation in NO₂ production rate) values are employed to quantify the contributions of major ${\mathrm{NO}}_{\mathrm{3}}^{-}$ oxidation pathways in conjunction with a Bayesian model. The α value ranged from 0.63 to 0.93, with an average of 0.83±0.06, suggesting the significance of O₃ participation in NO oxidation during the sampling campaign. On the other hand, our α values were lower than those (0.85–1) for other midlatitude regions (Alexander et al., 2009). The α values are influenced by the relative amount of O₃, HO₂, and RO₂ in NO_x cycling. Due to the generally high O₃ concentrations (O₃>50 ppb) observed in Lhasa, nearly all α values exceeded 0.8 (Fig. S6). To evaluate the impact of key parameters on the estimated contributions of different ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation pathways, we conducted a sensitivity analysis by assuming the α values and Δ¹⁷O value of the terminal oxygen atoms of O₃ ($\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }$). As listed in Table S3, the assumptions of α and $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }$ had an impact on the ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation mechanisms. When $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }$ was fixed at 39 ‰, increasing α from 0.7 to 0.9 led to a notable increase in the relative contribution of the NO₂+OH pathway from 25 % to 46 %, while that of the NO₃+VOC pathway decreased from 46 % to 25 %. The N₂O₅+H₂O pathway remained nearly constant, with contributions ranging between 28 % and 29 %, indicating that this pathway is relatively insensitive to changes in α values. Similarly, when α was varied within a reasonable range (0.68–0.93), increasing the $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }$ value from 37 ‰ to 39 ‰ led to an increase in the NO₂+OH contribution from 37 % to 46 % and a corresponding decrease in the NO₃+VOC contribution from 35 % to 26 %. Again, the N₂O₅+H₂O contribution remained stable at ∼28 %. These results suggest that the estimated contributions of NO₂+OH and NO₃+VOC pathways are sensitive to assumptions about α and $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }$, whereas the contribution of the N₂O₅+H₂O pathway is relatively robust under the tested conditions. Because Lhasa is characterized by relatively high VOC concentrations and $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{O}}_{\mathrm{3}}^{\ast }$ is generally close to 39 ‰, we consider our parameter assumptions reasonable for further estimating ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation pathways for each sample.

On average, the relative contributions of NO₂+OH ($f_{{\mathrm{NO}}_{\mathrm{2}}+\mathrm{OH}}$), NO₃+VOC ($f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$), and N₂O₅+H₂O ($f_{{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}}$) to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation in Lhasa during the sampling campaign were 46±26 %, 26±19 %, and 28±11 %, respectively. To better understand the characteristics of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation mechanisms in Lhasa, we performed a detailed comparison across China for the relative contributions of key oxidation pathways using the Δ¹⁷O methodology (Fig. 4). Overall, similar to most Chinese cities, ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation in Lhasa was predominantly driven by the NO₂+OH pathway, exhibiting distinct seasonal and regional variations. In particular, the average $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$ values were generally several times higher in spring in Lhasa than in other urban cities. Compared to rural/remote areas, the average $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$ values showed higher fractions in Lhasa, revealing the influence of anthropogenic emissions, i.e., vehicle exhaust and heating, on ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation. In Lhasa, the capital of Tibet, field measurements over different years showed a substantial increase in VOC concentrations in urban areas of the Tibetan Plateau, comparable to those in North China (Tang et al., 2022), revealing the importance of the active NO₃+VOC pathway for ${\mathrm{NO}}_{\mathrm{3}}^{-}$ pollution formation in Lhasa. In fact, recent studies have recognized NO₃+VOC as a major formation mechanism for ${\mathrm{NO}}_{\mathrm{3}}^{-}$ production. For instance, Fan et al. (2021) found that the $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$ values in Beijing increased from 17 % in summer to 32 % in winter based on $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ measurements. He et al. (2018) estimated the $f_{{\mathrm{NO}}_{\mathrm{2}}+\text{OH}}$ and $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$ values and found that they were in the range of 16 %–56 %, underscoring the significant roles of these pathways during haze events in Beijing. Similarly, Feng et al. (2023) also reported that the $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$ values were up to 49.6 % in winter in northern China. In Guangzhou, Wang et al. (2023) noted that the average $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$ value was at 488 m (25 %), which was higher than that at the ground (12 %). Furthermore, Li et al. (2022) reported that the $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$ values increased from 5 % in urban to 13.5 % in rural regions in Northeast China. Although the specific nighttime RO₂ production mechanism in Lhasa remains unclear, studies in other cities have demonstrated that the NO₃+VOC pathway was the dominant channel for nighttime RO₂ production (Fisher et al., 2016), which in turn led to the formation of alkyl and multifunctional nitrates (RONO₂) and eventually ${\mathrm{NO}}_{\mathrm{3}}^{-}$. In such cases, the RO₂ concentration is expected to be correlated with NO₃ radical production, which depends on the reaction rate of O₃ and NO₂ (Brown and Stutz, 2012). Given the relatively high nighttime O₃ concentrations in Lhasa, it is plausible that O₃-driven nighttime NO₃ chemistry plays an important role, thereby enhancing NO₃+VOC derived from RO₂ production and ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation. Global modeling studies also support the significance of this pathway. For instance, Alexander et al. (2020) reported that the NO₃+VOC pathway via the RONO₂ mechanism accounted for 3 % of global ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation on average. The relatively high $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$ values observed in Lhasa are broadly consistent with these findings, especially under conditions of high VOC concentrations and strong nighttime oxidant levels.

Figure 4 Summary of the relative contributions of key oxidation pathways using the Δ¹⁷O methodology across China (data given in Table S4). Colors for the study labels indicate the type of sampling location: urban areas (red) and rural/remote areas (black). The pie charts show the relative contribution of different pathways to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation: $f_{{\mathrm{NO}}_{\mathrm{2}}+\mathrm{OH}}$ ( deep yellow), $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$ (deep green), and $f_{{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}}$ (light blue).

4.2 Seasonal and diurnal variations of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ oxidation pathways

Figure S7 illustrates the seasonal variations in the relative contributions of the three main oxidation pathways to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation. When comparing different seasons, the $f_{{\mathrm{NO}}_{\mathrm{2}}+\mathrm{OH}}$ values were lower (p<0.01) in spring (22.6 %) than in winter (50.8 %), summer (52.9 %), and autumn (73.2 %). The dominance of the NO₂+OH pathway in autumn is consistent with observations at Mount Everest during the autumn season of 2017 and 2018, suggesting that ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation on the Tibetan Plateau in autumn may be mainly driven by the NO₂+OH pathway (Lin et al., 2021; Wang et al., 2020b).

A significant increase in the $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$ values was observed in spring (p<0.05). First, O₃ and NO₂ are precursors of NO₃. In this work, the highest concentrations of O₃ were found in spring ($\mathrm{114.9}\pm \mathrm{18.1}\mathrm{\mu }\mathrm{g}{\mathrm{m}}^{-\mathrm{3}}$), likely leading to elevated NO₃ concentrations. Additionally, the low T and reduced OH radical concentrations in spring facilitate the reaction of NO₂ and O₃ to synthesize NO₃. This might be an appropriate reason for the $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$ values in spring. High-altitude locations such as Nepal (5079 m a.s.l.) and Qomolangma Station (4300 m a.s.l.) have experienced stratospheric ozone intrusions, especially in spring and winter, as reported in previous studies (Zhang et al., 2025, 2022; Cristofanelli et al., 2010; Morin et al., 2007; Lin et al., 2016; Yin et al., 2017; Wang et al., 2020b). Notably, such intrusions in spring may elevate tropospheric O₃ levels in Lhasa, resulting in a mixture of tropospheric and stratospheric O₃ that enhances ${\mathrm{NO}}_{\mathrm{3}}^{-}$ production. Second, previous studies have indicated that the Afghanistan–Pakistan–Tajikistan region, the Indo-Gangetic Plain, and the Meghalaya–Myanmar region could transport industrial VOC emissions to various zones in Tibet from west to east. Additionally, agricultural areas in northern India could contribute biomass burning-related VOC emissions to the middle-northern and eastern regions of Tibet (Li et al., 2017). During our sampling campaign, South and Southeast Asia air clusters were notably prevalent in the springtime, coinciding with intensive fire spots observed in Afghanistan, Pakistan, India, Nepal, and Bhutan (Figs. S3 and S4). These observations, combined with the prevailing South and Southeast Asia air mass trajectories in spring, strongly suggest that long-range transported VOCs from South Asia were delivered to Lhasa and likely participated in local ${\mathrm{NO}}_{\mathrm{3}}^{-}$ production via the NO₃+VOC pathway. Moreover, recent studies have shown that ambient VOC concentrations in the urban areas on the Qinghai–Tibet Plateau were comparable to those in the North China Plain (Tang et al., 2022). The input of VOC emissions through long-range transport might further elevate VOC concentrations, thereby promoting ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation via the NO₃+VOC pathway and contributing to the enhanced $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$ values observed in spring. While VOCs appear to play a dominant role in the process, it should be noted that other nitrogen species (e.g., NO, NO₂) associated with biomass burning emissions may also be transported over long distances and influence ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation in Lhasa. These co-transported nitrogen compounds, although not directly quantified in this study, could further contribute to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ production in spring. Taken together, these findings provide strong evidence that long-range transport of biomass burning emissions, particularly from South Asia, can substantially influence springtime ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation in Lhasa.

Figure 5 The relative contributions (mean ± SD values) of NO₂+OH, NO₃+VOC, and N₂O₅+H₂O to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation during the day and night in (a) summer and (b) winter in Lhasa during the sampling campaign.

Similarly, the $f_{{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}}$ values exhibited their highest contributions during spring, with significant seasonal differences (p<0.05) except when compared to summer (p>0.05). Typically, high RH enhances ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation via the N₂O₅+H₂O pathway. However, studies have revealed that during sandstorm events, a significantly large N₂O₅ uptake coefficient was observed on urban aerosols in spring (Xia et al., 2019). In this study, the mean Ca²⁺ concentration in PM_2.5 was found to be the highest in spring, suggesting a possible role of dust in facilitating N₂O₅ uptake. Additionally, the N₂O₅+H₂O pathway has been reported to be promoted by elevated ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations (Lin et al., 2021), which were also highest in spring. Therefore, the increased $f_{{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}}$ values during spring might be attributed to the combined effects of lower RH, elevated Ca²⁺ levels, and high ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations.

Interestingly, distinct diurnal patterns of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ oxidation pathways were observed during the sampling campaign (Fig. 5). In summer, the NO₂+OH pathway showed a significantly higher contribution during the daytime (55.1 %) compared to nighttime (44.9 %), which is attributed to increased OH radical synthesis during longer days and higher temperatures in Lhasa (Rohrer and Berresheim, 2006). A previous study indicated that lower NO₂ and higher O₃ concentrations enhance the relative contribution of the OH pathway to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation (Wang et al., 2019). Additionally, the concentration of ALWC (the detailed information is given in Text S3) was higher at night than during the day in summer, favoring ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation through nocturnal formation. In winter, $f_{{\mathrm{NO}}_{\mathrm{2}}+\mathrm{OH}}$, $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$, and $f_{{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}}$ were similar during both day and night. Typically, photolytic destruction and chemical reactions with NO are rapid sinks during the daytime, with lifetimes generally less than 5 s, resulting in extremely low concentrations. Similarly, the atmospheric lifetime of N₂O₅ under sunlight is also very short (Wang et al., 2018). Thus, daytime NO₃ and N₂O₅ chemistry is often considered negligible. However, a recent study revealed that a non-negligible amount of NO₃ radicals can persist during the daytime in cold months, owing to the limited solar radiation (Hellén et al., 2018). Wang et al. (2020a) found that the daytime production rate of NO₃ can be substantial due to elevated concentrations of O₃ and NO₂, suggesting that the mixing ratios of NO₃ and N₂O₅ during the day may not be negligible. Furthermore, in winter, lower temperatures and elevated NO₂ concentrations facilitate a quasi-steady-state equilibrium between NO₃ and N₂O₅, slowing the overall reactivity of the ${\mathrm{NO}}_{\mathrm{3}}^{-}$ precursors (Brown et al., 2003). This equilibrium condition minimizes diurnal fluctuations in precursor concentrations, resulting in relatively stable nocturnal and daytime ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation pathways, including NO₃+VOC and N₂O₅+H₂O. Nevertheless, we acknowledge that the exact role of daytime NO₃$/$N₂O₅ chemistry remains uncertain in Lhasa and should be further assessed using concurrent field observations or chemical transport models. Moreover, when interpreting the diurnal differences in $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ values, the atmospheric lifetime of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ must be considered. Given that the atmospheric lifetime of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ is generally more than 12 h, each sample might reflect both daytime and nighttime ${\mathrm{NO}}_{\mathrm{3}}^{-}$ production impacting $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ values (Park et al., 2004; Vicars et al., 2013).

Figure 6 Influence of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ (µg m⁻³), temperature (°C), RH (%), ALWC (µg m⁻³), NO₂ (µg m⁻³), O₃, and O_x (µg m⁻³) on ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation pathways (%).

4.3 Integrated analysis of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ oxidation pathways in Lhasa

As shown in Fig. S8, the NO₃+VOC pathway emerged as the major contributor to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation during periods of high ${\mathrm{NO}}_{\mathrm{3}}^{-}$ spikes. To elucidate the ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation pathways under different ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations, ${\mathrm{NO}}_{\mathrm{3}}^{-}$ samples were categorized into different concentration ranges (Fig. 6). We found that the $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$ values increased and $f_{{\mathrm{NO}}_{\mathrm{2}}+\mathrm{OH}}$ values decreased with the ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations increasing. Although recent field radical measurements in urban sites in China found that OH and HO₂ radicals during haze periods were comparable to those on clean days (Slater et al., 2020; Yang et al., 2021), our results suggested that the NO₃+VOC pathway still played an important role in ${\mathrm{NO}}_{\mathrm{3}}^{-}$ production under high ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations in Lhasa, possibly due to enhanced VOC emissions. In addition to concentration effects, meteorological factors typically also regulate the ${\mathrm{NO}}_{\mathrm{3}}^{-}$ oxidation pathways. Typically, high T promotes the ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation via $f_{{\mathrm{NO}}_{\mathrm{2}}+\mathrm{OH}}$ values (Han et al., 2015). However, our study revealed that the relationship between T and $f_{{\mathrm{NO}}_{\mathrm{2}}+\mathrm{OH}}$ values did not consistently show a positive trend. Further analysis indicated that NO₂ and O₃ concentrations were negatively correlated, with lower NO₂ concentrations paired with elevated O₃ levels (Fig. S9). $f_{{\mathrm{NO}}_{\mathrm{2}}+\mathrm{OH}}$ values reached their minimum when NO₂ was between 15 and 20 µg m⁻³ and O₃ was within 100–120 µg m⁻³. Although OH radicals exhibit a higher oxidation potential (2.8 V) than O₃ (2.07 V), their atmospheric availability is much lower than that of O₃ (Carslaw et al., 1999; Dubey et al., 1997). Therefore, NO₂ at lower concentrations is more likely to be oxidized by OH than by O₃, even though O₃ concentrations were high. With increasing NO₂ concentrations, the availability of OH radicals for oxidizing NO₂ became lower, resulting in a relatively higher proportion of NO₂ being oxidized by O₃ although O₃ concentrations were low. However, when the concentration of O₃ was below 20 µg m⁻³, O₃ concentrations were not sufficient to oxidize NO₂ due to the higher NO₂ concentrations and OH radicals for oxidizing NO₂ would redominate. These observations underscore that in high-altitude urban environments like Lhasa, OH effectiveness is more important for ${\mathrm{NO}}_{\mathrm{3}}^{-}$ oxidation pathways than that of O₃. Additionally, we identified an intriguing positive correlation between the atmospheric oxidizing capacity (${\mathrm{O}}_{\mathit{\text{x}}}={\mathrm{NO}}_{\mathrm{2}}+{\mathrm{O}}_{\mathrm{3}}$) and $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$ values. $f_{{\mathrm{NO}}_{\mathrm{3}}+\text{VOC}}$ values were lowest when O_x was less than 90 µg m⁻³, corresponding to a maximum contribution from the NO₂+OH pathway. This suggests that O_x is more indicative of the pathways of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation in the atmosphere compared to either NO₂ or O₃ alone. Typically, high RH and ALWC were also positively correlated with $f_{{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}}$. However, RH was associated with variable contributions from the N₂O₅+H₂O pathway in our study, while increasing ALWC significantly enhanced this pathway, indicating ALWC as a more reliable indicator of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation.

4.4 Implications

The oxidation pathways of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ in Lhasa, China, were constrained using a full year of $\mathrm{\Delta }{}^{\mathrm{17}}\mathrm{O}-{\mathrm{NO}}_{\mathrm{3}}^{-}$ measurements from 2022 to 2023. Based on seasonal data, we observed a significant increase in the relative contribution of the NO₃+VOC to ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation during spring. Furthermore, the diurnal distribution of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ oxidation pathways varied distinctly across seasons. To better understand the factors influencing these pathways, we integrated meteorological conditions, NO_x precursors, and ALWC for a more comprehensive analysis of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation. The results revealed that O_x and ALWC were more reliable indicators of ${\mathrm{NO}}_{\mathrm{3}}^{-}$ oxidation pathways than meteorological factors. Notably, Lhasa's unique high-altitude environment, such as strong solar radiation, persistently high O₃, and elevated VOC, promotes active NO₃+VOC chemistry, especially in spring. Atmospheric ALWC is primarily produced by hygroscopic aerosols such as ${\mathrm{SO}}_{\mathrm{4}}^{\mathrm{2}-}$, ${\mathrm{NH}}_{\mathrm{4}}^{+}$, and Cl⁻. Therefore, in addition to controlling NO₂, O₃, and VOC, reducing these hygroscopic aerosols is crucial for effective PM_2.5 pollution control.

Although this study provides valuable insights into ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation mechanisms in Lhasa, we must acknowledge the associated uncertainties due to the lack of comprehensive observational constraints in Lhasa. Specifically, the limited understanding of local RO₂ concentrations led us to adopt empirical parameterizations and refer to measurements from other regions, which inevitably introduced uncertainty into the pathway apportionment. In addition, the absence of direct observations of nighttime NO emissions and the NO₂–NO isotope exchange processes in this region further complicates the interpretation of diurnal variations in ${\mathrm{NO}}_{\mathrm{3}}^{-}$ formation pathways. To improve the robustness of Δ¹⁷O-based pathway analysis, future studies should consider synchronous measurements of both NO₂ and ${\mathrm{NO}}_{\mathrm{3}}^{-}$ isotopes.