Nitrous acid budgets in the coastal atmosphere: potential daytime marine sources

Nitrous acid (HONO), a vital precursor of atmospheric hydroxyl radicals (OH), has been extensively investigated to understand its characteristics and formation mechanisms. However, discerning fundamental mechanisms across diverse environments remains challenging. This study utilizes measurements from Mount Lao, a coastal mountain in eastern China, and an observation-based chemical box model (OBM) to examine HONO budgets and their subsequent impacts on atmospheric oxidizing capacity. The model incorporates additional HONO sources, including direct emissions, heterogeneous conversions of NO₂ on aerosol and ground surfaces, and particulate nitrate photolysis. The observed mean HONO concentration was 0.46 ± 0.37 ppbv. The updated model reproduced daytime HONO concentrations well during dust and photochemical pollution events. During dust events, daytime HONO formation was dominated by photo-enhanced heterogeneous reactions of NO₂ on aerosol surfaces (> 50 %), whereas particulate nitrate photolysis (34 %) prevailed during photochemical pollution events. Nevertheless, the model uncovers a significant unidentified marine HONO source in a “sea case”, with its HONO production rate reaching up to 0.70 ppbv h⁻¹ at noon. Without considering this unidentified source, an extraordinarily high photolysis coefficient of nitrate and/or a heterogeneous uptake coefficient of NO₂ would be required to match observed HONO concentrations. This missing marine HONO source affected the peak O₃ production rate and OH radical concentration by 36 % and 28 %, respectively, at the observation site. Given the limited HONO observation data in coastal and marine settings, the unidentified HONO source may cause an underestimation of the atmosphere's oxidizing capacity. This study highlights the necessity for further investigation of the role of HONO in atmospheric chemistry in coastal and marine environments.

1 Introduction

Atmospheric nitrous acid (HONO) serves as a pivotal precursor of hydroxyl radicals (OH) (Alicke et al., 2003), accounting for up to 60 % of daytime OH radicals (Kleffmann et al., 2005; Czader et al., 2012). Thus, HONO establishes itself as a critical source of OH in both urban and rural environments, surpassing the contribution from ozone (O₃) photolysis (Acker et al., 2006; Elshorbany et al., 2012; Gu et al., 2022a). Consequently, HONO substantially influences the formation of secondary pollutants, including secondary aerosols and O₃, exerting a considerable effect on air quality and climate change (Xing et al., 2019; Y. Yang et al., 2021).

Recent studies pinpoint four primary sources of atmospheric HONO: (a) direct emissions from traffic (Liao et al., 2021), biomass burning (Nie et al., 2015; Theys et al., 2020), soil (Su et al., 2011), and livestock farming (Zhang et al., 2023); (b) homogeneous reaction of NO + OH, which is generally regarded as a significant process in polluted urban areas during daytime when NO and OH concentrations are relatively high (Gu et al., 2022a); and (c) heterogeneous reactions of NO₂ on various surfaces, such as mineral dust (Underwood et al., 2001), soil (Kebede et al., 2016), and aqueous surfaces (Wojtal et al., 2011). The uptake coefficient of NO₂, γ(NO₂), on these surfaces remains uncertain and is subject to varying factors, sparking debates regarding the importance of the heterogeneous conversion of NO₂ (Broske et al., 2003; Xue et al., 2022). (d) There is photolysis of adsorbed nitric acid (HNO₃) and particulate nitrate (pNO${}_{\mathrm{3}}^{-}$), crucial contributors to daytime HONO formation (Ye et al., 2017; Gen et al., 2022), particularly in clean environments (Zhou et al., 2011; Ye et al., 2016a). However, HONO formation mechanisms in different environments remain contentious and require more detailed model evaluations (Jiang et al., 2022).

Despite its short daytime atmospheric lifetime, HONO is frequently observed at high concentrations at noon (Ye et al., 2016a; J. Yang et al., 2021; Jiang et al., 2023). Traditional mechanisms cannot fully explain these observed daytime HONO peaks, indicating the presence of additional daytime HONO sources (missing sources of daytime HONO). Over recent decades, researchers have extensively investigated the missing sources of daytime HONO in various environments (Kleffmann, 2007; Lee et al., 2016; Jiang et al., 2022; Zhang et al., 2022). However, our limited understanding of these unidentified HONO sources has hindered accurate assessments of atmospheric free radicals and oxidizing capacity (Tang et al., 2015). In areas with high concentrations of NO₂ and particulate matter, missing sources are often ascribed to the photolytic enhancement of heterogeneous NO₂ reactions (Su et al., 2008; Czader et al., 2012; Lee et al., 2016; Tong et al., 2016), which can also be intensified by the presence of other substances (e.g., ammonia or sulfur dioxide) (Ma et al., 2017; L. Li et al., 2018; Ge et al., 2019). Conversely, in remote areas, nitrate photolysis or soil emissions are perceived as significant contributors to daytime HONO sources (Su et al., 2011; Ye et al., 2016a; Cui et al., 2019). In polluted mountainous areas, the vertical transport of air masses may also contribute to observed daytime HONO concentrations (Jiang et al., 2020; Xue et al., 2022). During dust storms, the particle surface area increases sharply, potentially enhancing the heterogeneous reaction of NO₂, yet the evaluations of dust impacts on daytime HONO are scarce (Wang, 2003). Overall, most existing HONO source studies lack quantitative assessments based on models and fail to provide comparative analyses across different environmental scenarios.

The marine boundary layer (MBL) with a large air–water interface, where the ocean and atmosphere exchange trace gases, heat, and aerosol particles (Wurl et al., 2016) and the interfacial photochemistry processes often occur (Bruggemann et al., 2018), is utterly different from inland environments. The opposite diurnal variations of HONO with a peak concentration at noon at marine sites implied different predominant HONO processes compared with polluted inland areas (Jiang et al., 2022). Furthermore, recent observations of HONO in coastal and marine regions indicate the existence of marine HONO sources (Ye et al., 2016a; Crilley et al., 2021; J. Yang et al., 2021; Jiang et al., 2022). The observed accelerated NO₂-to-HONO conversion in marine air masses suggests that air–marine interactions enhance HONO production (Zha et al., 2014; J. Yang et al., 2021). However, the heterogeneous conversion of NO₂ on the vast air–water interface, a potential source of marine HONO, remains uncertain (Wojtal et al., 2011; Yu et al., 2021; Zhu et al., 2022). Crilley et al. (2021) only obtained a factor of 5 lower ocean-surface NO₂-to-HONO conversion than previous studies; there was still a debate on the importance of ocean-surface-mediated conversion of NO₂ into HONO. Nitrate photolysis is believed to contribute to marine HONO sources (Ye et al., 2016a; Andersen et al., 2023), but significant controversy persists (Romer et al., 2018; Shi et al., 2021). The specific influencing factors remain unclear (Zhang et al., 2020; Andersen et al., 2023), with some studies suggesting other factors may be responsible (Wojtal et al., 2011; J. Yang et al., 2021). Accordingly, Jiang et al. (2023) highlighted the contribution of the dust-surface photocatalytic conversion of reactive nitrogen compounds to HONO formation and the important role of halogen chemistry in HONO simulation in CVAO (Cabo Verde). However, most existing studies still rely on steady-state analysis, and there is a lack of quantitative research determining whether current HONO mechanisms can adequately explain observed marine daytime HONO concentrations.

Mount Lao, located on the eastern coast of Qingdao, China, experiences influence from various air masses from the continent and the ocean. During the spring of 2021 (27 April–19 May), when dust and O₃ pollution occurred frequently, we conducted measurements on Mount Lao to explore the daytime HONO budgets in the coastal atmosphere. Utilizing the latest HONO formation mechanisms in the box model, we found that the existing parameters adequately accounted for the HONO sources during both dust and photochemical pollution periods. However, we identified a significant discrepancy between the simulated and observed HONO in a “sea case”. This discrepancy suggests that a substantial daytime source of marine-derived HONO is absent from the current chemical mechanisms. To compensate for this missing source, either an unprecedentedly large enhancement factor (EF) of nitrate photolysis or a heterogeneous uptake coefficient of NO₂ would be necessary if attributed solely to these known HONO sources.

2 Methods

2.1 Field measurements

Field measurements were conducted on the southeastern coast of Mount Lao (36.15^∘ N, 120.68^∘ E; 166 m a.s.l.) in Qingdao (Fig. 1), approximately 1 km away from the Yellow Sea. The geographical location and elevation of Mount Lao make it an optimal location for examining the contrasts between marine and continental air masses and the chemical processes within the marine boundary layer. The relatively pristine condition of the area, coupled with minimal levels of anthropogenic activities such as industrial emissions, establish Mount Lao as a representative of a clean environment. The field campaign was carried out during the spring of 2021 (27 April–19 May 2021), a period when the air quality of Qingdao is often affected by dust storms from Mongolia and northwestern China as well as by O₃ pollution. Consequently, the site at Mount Lao provides an opportune platform for investigating the fundamental formation mechanisms of HONO under diverse environmental conditions.

Figure 1 Maps showing the location of the monitoring site. Panel (a) is colored by the tropospheric NO₂ column density in May 2021 from the Ozone Monitoring Instrument (OMI, https://www.earthdata.nasa.gov/, last access: 27 November 2023), and panels (b) and (c) are colored by the geographical height from the Geospatial Data Cloud (http://www.gscloud.cn/, last access: 27 November 2023).

HONO was quantified using a water-based long-path absorption photometer (WLPAP, Beijing Zhichen Technology Co., Ltd, China). Ambient HONO was absorbed by deionized water alone, after which it reacted with a reagent comprising 3.44 g of sulfanilamide and 0.2 g of N-(1-naphthyl)-ethylenediamine-dihydrochloride (NED) in 10 L of deionized water, leading to the formation of an azo dye. Two channels were employed to extract HONO and interfering gases, respectively. The absorbance of the azo dye was measured using a fiber optic spectrometer (USB 4000, Ocean Optics, USA) at both the measurement wavelength (550 nm) and the reference wavelength (580 nm). Regular automatic zero measurements using ultrapure nitrogen were conducted every 2 d to correct for baseline drift. The detection limit and detection ranges were 2 and 5 pptv–2 ppmv, respectively.

A suite of commercial online analyzers monitored the concentrations of NO_x, O₃, SO₂, and CO (42i, 49i, 43i, and 48i, respectively, Thermo Fisher Scientific Inc, USA). PM_2.5 was measured using a hybrid nephelometric–radiometric particulate mass monitor (SHARP-5030i, Thermo Fisher Scientific Inc, USA), while PM₁₀ mass data were obtained from the China National Environmental Monitoring Center (https://quotsoft.net/air/, last access: 27 November 2023). During the field campaign, 57 VOC (volatile organic compound) canister samples were collected at 2 h intervals from 09:00 to 19:00 local time on pollution episode days and at 6 h intervals from 09:00 to 21:00 on non-episode days. These VOC samples were subsequently analyzed using gas chromatography and mass spectrometry (TT24xr, Markes, UK; GC–MS, Thermo Fisher Scientific Inc, USA) (Liu et al., 2021). A wide-range particle spectrometer (WPS, Model 1000XP, MSP, USA) was employed to determine the atmospheric particle number size distributions from 10 nm to 10 µm. Taking into account the hygroscopic growth, the relative humidity-adjusted aerosol surface area concentration (Sa) was calculated based on the determined particle number size distributions; 95 offline particulate samples were collected every 3 h from 07:00 to 19:00 and every 12 h from 19:00 to 07:00 utilizing a high-volume air-sampling system (TE-5170, Tisch Environmental Inc, USA). The inorganic compositions of the samples, including Cl⁻, NO${}_{\mathrm{3}}^{-}$, SO${}_{\mathrm{4}}^{-}$, NH${}_{\mathrm{4}}^{+}$, Na⁺, K⁺, Mg²⁺, and Ca²⁺, were determined via ion chromatography (Dionex ICS-600, Thermo Fisher Scientific Inc, USA). Meteorological data, including temperature, relative humidity (RH), pressure, wind speed, and wind direction, were monitored by an ultrasonic integrated weather station (RS-FSXCS-N01-1).

This study distinguishes between the sea case and the “land case” by analyzing the backward trajectories of the air masses. Specifically, considering the short lifetime of HONO, the MeteoInfo model (Wang, 2012) was used to calculate 6 h air mass backward trajectories starting at a height of 200 m above ground level using meteorological parameters from the Global Data Assimilation System (GDAS, https://www.ready.noaa.gov/data/archives/gdas1, last access: 27 November 2023. The criteria for differentiating between the sea case and land case are based on the time spent over land or sea during the 6 h backward air mass, with cases that spent less than 1 h over land designated as a sea case (J. Yang et al., 2021). Following this criterion, we selected a total of 18 sea cases and 13 land cases (Table S1). The observation data for the sea case and the land case were averaged for subsequent analysis.

2.2 Model setup

An observation-based chemical box model (OBM) was employed to explore the HONO budgets and atmospheric oxidizing capacity. The chemical mechanism used for the modeling was obtained from the Master Chemical Mechanism (MCM) website (https://mcm.york.ac.uk/MCM, last access: 27 November 2023) and was based on the MCM v3.3.1 as proposed by Jenkin et al. (2015). The observed data of HONO, O₃, NO, NO₂, SO₂, CO, VOCs, pNO${}_{\mathrm{3}}^{-}$, Sa, temperature, RH, and pressure were averaged or interpolated to a time resolution of 5 min, except for VOCs and pNO${}_{\mathrm{3}}^{-}$, which were linearly interpolated to a time resolution of 1 h to constrain the model (Yang et al., 2018). The calculation of the photolysis rate of NO₂, JNO₂, was determined using Eq. (1):

$$\begin{array}{}\text{(1)}& J{\mathrm{NO}}_{\mathrm{2}}=J{\mathrm{NO}}_{\mathrm{2}\left(\mathrm{TUV}\right)}\times {\displaystyle \frac{{\mathrm{UV}}_{\mathrm{observed}}}{{\mathrm{UV}}_{\mathrm{TUV}}}},\end{array}$$

where JNO_2(TUV) and UV_TUV are obtained from the Tropospheric Ultraviolet and Visible (TUV) radiation model (https://www.acom.ucar.edu/Models/TUV/Interactive_TUV, last access: 27 November 2023). The UV_observed was obtained from the NASA GES DISC (https://disc.gsfc.nasa.gov/, last access: 27 November 2023). Other photolysis frequencies were calculated in the OBM and scaled by JNO₂. The time series of JNO₂ is presented in Fig. S1. The model was pre-run for 1 d to stabilize the simulation of unconstrained species, e.g., atmospheric radicals and some secondary VOCs.

In the MCM v3.3.1, the formation of HONO is originally attributed to a homogeneous reaction, specifically NO + OH → HONO. This study extends the existing mechanism by incorporating additional sources of HONO into the chemical model. A description of these sources and their associated mechanisms is provided in the following, and the corresponding parameters are listed in Table 1.

Table 1 Summary of the HONO sources and sinks included in the box model.

2.2.1 Description of HONO sources and sinks adopted in the OBM

Direct emission

In the atmosphere, HONO can be directly released through the exhaust emissions of various sources. The HONO $/$ NO_x emission ratio, which typically averages around 0.8 %, is a common parameter used to gauge the impacts of these vehicular emissions on HONO concentration (Kleffmann et al., 2003; Czader et al., 2012; Lee et al., 2016; Xue et al., 2020a). However, the ratio can fluctuate between 0.3 % and 1.6 %, depending on engine and fuel types (Kurtenbach et al., 2001). Prior research indicates that direct emissions contribute significantly to HONO concentrations in urban settings (Zhang et al., 2019; Kramer et al., 2020). However, in rural and background areas, the vehicular contribution is comparatively insignificant (Y. Liu et al., 2019; Xue et al., 2022). Consequently, the contribution of vehicle emissions to HONO is not constant and varies based on the environment and traffic density. In this study, we employed the widely used ratio of 0.8 % for modeling scenarios and sensitivity simulations using ratios of 0.4 % and 1.6 %.

Homogeneous reaction of OH + NO → HONO

$$\begin{array}{}\text{(R1)}& \mathrm{NO}+\mathrm{OH}+\mathrm{M}\to \mathrm{HONO}+\mathrm{M}\end{array}$$

The reaction of NO + OH is considered an important gas-phase reaction for HONO formation, particularly during pollution periods when concentrations of NO and OH are high (Gu et al., 2022a). We employed the box model to calculate the reaction rate using complex rate coefficients from the MCM website (https://mcm.york.ac.uk/MCM/rates/complex, last access: 27 November 2023).

Heterogeneous reaction of NO₂ on aerosol surfaces

$$\begin{array}{}\text{(R2)}& {\displaystyle }{\mathrm{NO}}_{\mathrm{2}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}\stackrel{\text{aerosol\hspace{0.17em} surface}}{⟶}\mathrm{HONO}+{\mathrm{HNO}}_{\mathrm{3}}\text{(R3)}& {\displaystyle }{\mathrm{NO}}_{\mathrm{2}}+hv\stackrel{\text{aerosol\hspace{0.17em} surface}}{⟶}\mathrm{HONO}\text{(2)}& {\displaystyle }{k}_{\mathrm{aerosol}}=\mathrm{0.25}\times v_{{\mathrm{NO}}_{\mathrm{2}}}\times \mathrm{Sa}\times {\mathit{\gamma }}_{\mathrm{a}}{\mathit{\gamma }}_{\mathrm{a}}=\mathrm{8}\times {\mathrm{10}}^{-\mathrm{6}}\text{(3)}& {\displaystyle }\begin{array}{rl}{k}_{\mathrm{aerosol},hv}& =\mathrm{0.25}\times v_{{\mathrm{NO}}_{\mathrm{2}}}\\ & \times \mathrm{Sa}\times {\mathit{\gamma }}_{\mathrm{a},hv}\times {\displaystyle \frac{J{\mathrm{NO}}_{\mathrm{2}}}{J{\mathrm{NO}}_{\mathrm{2}},\mathrm{noon}}}{\mathit{\gamma }}_{\mathrm{a},hv}=\mathrm{4}\times {\mathrm{10}}^{-\mathrm{5}}\end{array}\text{(4)}& {\displaystyle }{v}_{{\mathrm{NO}}_{\mathrm{2}}}=\sqrt{{\displaystyle \frac{\mathrm{8}RT}{\mathit{\pi }M}}}\end{array}$$

The heterogeneous conversion of NO₂ on surfaces is a significant source of HONO in the atmosphere. As illustrated by Reactions (R2) and (R3), NO₂ reacts with water and light on aerosol surfaces to produce HONO. The HONO formation rate from heterogeneous reactions is typically first-order with respect to NO₂ concentration (Aumont et al., 2003), and the reactivity of NO₂ is known to be significantly enhanced under irradiated conditions compared to darkness (C. Yu et al., 2022). In this study, the uptake coefficients of NO₂ on the aerosol surface under dark and irradiated conditions, γ_a and γ_a,hv, were set to 8 × 10⁻⁶ and 4 × 10⁻⁵ (Lelièvre et al., 2004; Vandenboer et al., 2013), respectively. The molecular speed of NO₂ ($v_{{\mathrm{NO}}_{\mathrm{2}}}$, m s⁻¹) was calculated using Eq. (4), where R represents the ideal gas constant, 8.314 J mol⁻¹ k⁻¹; T is the absolute temperature (K); and M is the relative molecular weight of NO₂ (g mol⁻¹). Sa is the surface area concentration (m² m⁻³) estimated from particle number concentrations measured by the WPS.

Heterogeneous reaction of NO₂ on ground surfaces

$$\begin{array}{}\text{(R4)}& {\displaystyle }{\mathrm{NO}}_{\mathrm{2}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}\stackrel{\text{ground\hspace{0.17em} surface}}{⟶}\mathrm{HONO}+{\mathrm{HNO}}_{\mathrm{3}}\text{(R5)}& {\displaystyle }{\mathrm{NO}}_{\mathrm{2}}+hv\stackrel{\text{ground\hspace{0.17em} surface}}{⟶}\mathrm{HONO}\text{(5)}& {\displaystyle }{k}_{\mathrm{ground}}=\mathrm{0.25}\times v_{{\mathrm{NO}}_{\mathrm{2}}}\times {\mathit{\gamma }}_{\mathrm{g},hv}\times {\displaystyle \frac{S}{V}}{\mathit{\gamma }}_{\mathrm{g}}=\mathrm{1}\times {\mathrm{10}}^{-\mathrm{6}}\text{(6)}& {\displaystyle }\begin{array}{rl}{k}_{\mathrm{ground},hv}& =\mathrm{0.25}\times v_{{\mathrm{NO}}_{\mathrm{2}}}\times {\mathit{\gamma }}_{\mathrm{g},hv}\\ & \times {\displaystyle \frac{S}{V}}\times {\displaystyle \frac{J{\mathrm{NO}}_{\mathrm{2}}}{J{\mathrm{NO}}_{\mathrm{2}},\mathrm{noon}}}{\mathit{\gamma }}_{\mathrm{g},hv}=\mathrm{2}\times {\mathrm{10}}^{-\mathrm{5}}\end{array}\text{(7)}& {\displaystyle \frac{S}{V}}={\displaystyle \frac{\mathrm{1.7}}{\mathrm{BLH}}}\end{array}$$

Equations (5) and (6) delineate the parameterizations for the heterogeneous reaction of NO₂ on the ground surfaces in both the absence and presence of light. The uptake coefficients of NO₂ on the ground surface under dark and irradiated conditions, γ_g and γ_g,hv, respectively, were set to 1 × 10⁻⁶ and 2 × 10⁻⁵ (Kleffmann et al., 1998; Stemmler et al., 2006), respectively. Under ambient conditions, the relative importance of gas uptake on ground and aerosol surfaces is uncertain, with the influence of land use categories and chemical compositions (Li et al., 2019). The surface-to-volume ratio, $\frac{S}{V}$, is calculated by an effective surface of 1.7 m² per geometric surface in Eq. (7) (Vogel et al., 2003). Within the model, the boundary layer height, BLH, is projected to increase from 300 m at dawn to 1500 m at 14:00 and then decrease back to 300 m at dusk (Xue et al., 2014).

Photolysis of particulate nitrate

$$\begin{array}{}\text{(R6)}& {\displaystyle }p{{\mathrm{NO}}_{\mathrm{3}}}^{-}+hv\to \mathrm{HONO}\text{(8)}& {\displaystyle }k={\displaystyle \frac{\mathrm{J}\left({\mathrm{pNO}}_{\mathrm{3}}^{-}\right)}{J{\mathrm{HNO}}_{\mathrm{3}},\mathrm{noon}}}\times J{\mathrm{HNO}}_{\mathrm{3}}\left(\mathrm{MCM}\right)\end{array}$$

In Eq. (8), the photolysis rate constant of gaseous HNO₃ at noon, JHNO₃,noon, is chosen to be ∼ 7 × 10⁻⁷ s⁻¹ based on a previous study (Ye et al., 2016b). JHNO₃(MCM) is calculated by the box model. Recent research has shown that the photolysis rate of particulate nitrate is significantly faster than that in the gas and aqueous phases (Zhou et al., 2003; Ye et al., 2016a). We adopt a median value of 8.3 × 10⁻⁵ s⁻¹ in our simulation based on a range provided by Ye et al. (2017). Considering the uncertainty of the parameter values of the abovementioned HONO formation mechanisms, we conducted the sensitivity tests with lower and upper values in Sect. 3.2 and 3.3.

Photolysis of HONO

$$\begin{array}{}\text{(R7)}& \mathrm{HONO}+h\mathit{\nu }\to \mathrm{OH}+\mathrm{NO}\end{array}$$

The primary loss pathway of HONO is through photolysis following sunrise, which significantly contributes to the atmospheric OH budget. The photolysis rate of HONO, J(HONO), in the OBM is constrained by JNO₂.

Homogeneous reaction between HONO and OH

$$\begin{array}{}\text{(R8)}& \mathrm{HONO}+\mathrm{OH}\to {\mathrm{H}}_{\mathrm{2}}\mathrm{O}+{\mathrm{NO}}_{\mathrm{2}}\end{array}$$

The relevant kinetic parameter of the reaction between HONO and OH is available from the MCM mechanism, and its reaction rate coefficient is dependent solely on temperature.

Dry deposition of HONO

$$\begin{array}{}\text{(9)}& k={\displaystyle \frac{\mathrm{vHONO}}{\mathrm{BLH}}}\end{array}$$

Here, vHONO is the dry deposition velocity of HONO (cm s⁻¹). Harrison and Kitto (1994) suggested that the range of vHONO was 0.2–1.7 cm s⁻¹, and a value of 1.0 cm s⁻¹ was employed in this study. The dry depositions of HONO, O₃, and other species, including peroxides, carbonyls, and organic acids, are also considered in the OBM (Xue et al., 2014).

3 Results and discussion

3.1 Concentration levels and temporal variations

Figure 2 displays the time series of HONO, HONO $/$ NO₂, NO_x, O₃, CO, SO₂, PM_2.5, and pNO${}_{\mathrm{3}}^{-}$ along with meteorological parameters (i.e., temperature, RH, and wind) measured throughout the field campaign. The presence of missing data in the time series resulted from instrument maintenance and calibration. Instrument maintenance and calibration resulted in gaps in the time series data. The observation site underwent dust periods on 27–28 April and 7–8 May as well as periods of photochemical pollution on 5–6, 13, and 17–18 May. In this study, a photochemical pollution period is classified as a day when the maximum daily 8 h average O₃ concentration (MDA8O₃) exceeds 75 ppbv (the Grade II National Ambient Air Quality Standard). A dust period is recognized when the peak PM₁₀ concentration surpasses 150 µg m⁻³, and the PM_2.5 $/$ PM₁₀ ratio falls below 0.4 based on previous research (Liu et al., 2006; Wu et al., 2020). Section 3.2 provides a comprehensive explanation of the differences in pollutant concentrations and HONO budgets during dust and photochemical pollution periods.

Figure 2 Time series of HONO, meteorological parameters, and related species measured during the campaign. The yellow-shaded areas correspond to the period of dust, while the pink-shaded areas represent the period of photochemical pollution.

Table 2 summarizes the descriptive statistics of the species and meteorological parameters measured during the observation period. The average (± standard deviation, SD) temperature and RH were 15.1 ± 3.4^∘ and 68.7 ± 26.1 %, respectively, indicating a moderate spring temperature and relatively high RH influenced by marine air masses. The primary pollutant concentrations were relatively low, as indicated by the mean mixing ratios of 0.9 ± 1.7, 5.9 ± 4.8, 284.0 ± 118.8, and 1.0 ± 0.8 ppbv for NO, NO₂, CO, and SO₂, respectively. These low levels suggest that Mount Lao is a relatively clean site with minimal impact from nearby anthropogenic sources. The high O₃ concentration (60.4 ± 15.8 ppbv) implies that photochemical reactions were relatively strong during observation. The mean concentrations of PM_2.5 and pNO${}_{\mathrm{3}}^{-}$ were 21.2 ± 21.0 µg m⁻³ and 4.6 ± 5.0 µg m⁻³, respectively.

Table 2 Statistics of measured species and meteorological parameters during the campaign.

DL: detection limit.

During the campaign, the mean concentration of HONO was 0.46 ± 0.37 ppbv, with a maximum mixing ratio of 3.14 ppbv recorded at 17:00 on 4 May. The concentration level of HONO at Mount Lao is lower than at urban sites with higher NO₂ concentrations (D. Li et al., 2018; Hao et al., 2020; Y. Yu et al., 2022). However, it is notably higher than other clean coastal and remote marine sites, as Table S2 illustrates (Villena et al., 2011; Zha et al., 2014; Meusel et al., 2016; Crilley et al., 2021; Zhu et al., 2022). Past research conducted in urban and rural areas found that the HONO $/$ NO₂ ratio, which indicates the extent of NO₂ conversion to HONO, typically ranges from 0.02 to 0.08 (Jiang et al., 2022). The higher HONO $/$ NO₂ value (0.13) measured at Mount Lao highlights the potentially significant role of non-NO_x-related HONO sources or higher heterogeneous conversion of NO₂ efficiency at this site.

Figure 3 Average diurnal variations of (a) HONO, (b) HONO $/$ NO₂, (c) particle nitrate, (d) O₃, (e) CO, (f) SO₂, (g) NO₂, (h) NO, (i) PM_2.5, (j) temperature, (k) RH, and (l) pressure during the observation period. The shaded area indicates the range of half of the standard deviation.

Table 3 Comparison of the statistics for the measured species and meteorological parameters during dust, photochemical pollution, and non-polluted periods in the daytime (07:00–17:00).

Figure 3 illustrates the average diurnal patterns of HONO and related species. The diurnal cycle of CO and SO₂ is similar, with peak concentrations observed during midday and relatively stable concentrations during nighttime. The concentration of O₃ increases with the accumulation of photochemical generation during the afternoon and decreases steadily after sunset. In contrast to most urban or rural locations, the concentration of HONO at Mount Lao peaks at noon, similarly to remote areas (Ye et al., 2016a; Jiang et al., 2022). It is important to note that both HONO and NO_x exhibit a second daytime peak in the late afternoon primarily caused by specific days with relatively high emissions or transport events. After excluding these days, the diurnal variation of HONO aligns more closely with observations in clean regions (Jiang et al., 2022). We have included the diurnal variation of HONO after removing days with afternoon peaks in the supporting information. Since their impact is largely confined to specific days, it does not significantly affect our subsequent analysis, particularly the sea case analysis. NO_x, comprising NO and NO₂, shows a similar temporal variation trend to HONO, suggesting potential photolytic sources for them (Reed et al., 2017). During the daytime (07:00–17:00), the average concentration of HONO was 1.56 times higher than at night (17:00–7:00), with concentrations of 0.54 ppbv during the day and 0.35 ppbv at night. Given the short lifetime of HONO during the day – only a matter of minutes – the noon HONO peak concentration suggests an in situ photochemical source of HONO (Kasibhatla et al., 2018). The ratio of HONO to NO₂ shows an increasing trend until sunrise, suggesting heterogeneous conversion from NO₂ to HONO during nighttime. However, unlike urban areas where the ratio of HONO to NO₂ decreases during the daytime (Zhang et al., 2019; Gu et al., 2022a), the ratio even increases during the midday period at Mount Lao, implying that HONO from sources other than NO₂ conversion also significantly contributes to HONO concentration (J. Yang et al., 2021). Considering that the influence of HONO on the OH radical and O₃ is primarily observed during the daytime, the higher concentration of HONO during the daytime at the Mount Lao site suggests the presence of strong daytime HONO sources. The primary objective of our following study is to analyze the daytime budgets of HONO in the coastal atmosphere of Mount Lao.

3.2 Daytime HONO budgets in dust and photochemical pollution periods

The daytime HONO budgets were examined during periods of dust and photochemical pollution using an updated OBM, with the aim of assessing whether our current understanding of HONO sources is sufficient to explain observed concentrations. Table 3 presents the mean daytime concentrations of HONO and other species during the dust, photochemical pollution, and non-polluted periods (i.e., days devoid of dust and photochemical pollution). On average, the daytime HONO concentrations during the dust and photochemical pollution periods are 0.57 ± 0.39 ppbv and 0.44 ± 0.29 ppbv, respectively. The dust period exhibits significantly higher concentrations of NO₂, PM_2.5, pNO${}_{\mathrm{3}}^{-}$, and Sa, with increased factors of 1.4, 2.6, 2.3, and 2.3, respectively, compared to the non-polluted period. During the photochemical pollution period, the daytime mean values of O₃, CO, SO₂, PM_2.5, pNO${}_{\mathrm{3}}^{-}$, and JNO₂ are 78.8, 353.8, 1.7 ppbv, 25.0, 6.2 µg m⁻³, and 7.0 × 10⁻³ s⁻¹, respectively. These values are approximately 1.4, 1.3, 2.4, 1.5, 2.1, and 1.6 times higher than those during the non-polluted period.

Figure 4 compares the observed and modeled HONO concentrations during dust and photochemical pollution periods and illustrates the contribution of various sources and sinks to the HONO budget. The study examines two scenarios: the base case, which only considers the homogeneous reaction NO + OH, and the model case, which considers all seven HONO sources outlined in Table 1. The results indicate that the base case significantly underestimated the HONO concentration, consistent with previous studies (Y. Liu et al., 2019; Zhu et al., 2022). However, the model case effectively replicates the observed HONO concentrations for both periods, even the high noon concentrations. The index of agreement (IOA) values for HONO during the dust and photochemical pollution periods are 0.96 and 0.88, respectively. This suggests that the updated parameterization scheme employed in the model can adequately account for the observed HONO concentration at Mount Lao.

Figure 4 Daytime HONO budgets in the dust (a, c) and (b, d) photochemical periods at Mount Lao. The base case only considers the homogeneous reaction of NO + OH, and the model case considers the updated HONO sources described in this study.

During the daytime, the average modeled production rates of HONO are 1.66 and 0.90 ppbv h⁻¹ for the dust and photochemical pollution periods, respectively. The maximum HONO production rate is significantly higher during the dust period (3.50 ppbv h⁻¹) compared to the photochemical pollution period (1.69 ppbv h⁻¹) and is even comparable to levels observed during haze periods at polluted urban or rural sites (Zhang et al., 2019; Xue et al., 2020b; Gu et al., 2022a). Based on the model results of detailed HONO budgets, the dominant pathway for daytime HONO production during the dust period is photo-enhanced heterogeneous conversion of NO₂ on the aerosol surface, accounting for 53 % (0.87 ± 0.66 ppbv h⁻¹) of the simulated daytime HONO production rate. Wang (2003) reported sudden increases in HONO concentration during nocturnal dust storm events and observed a higher ratio of HONO to NO₂ (0.18). The enhanced efficiency of NO₂ to HONO on mineral dust particles suggests a potentially significant impact of dust aerosol on nitrogen compound distribution. Further research is needed to understand the contribution of dust to HONO formation and nitrogen cycling during the daytime as well as its global impact. For the photochemical pollution period, the major sources of HONO include the photolysis of particulate nitrate, the photo-enhanced heterogeneous conversion of NO₂ on the aerosol surface, and the homogeneous reaction of OH and NO, which contribute 34 %, 27 %, and 27 %, respectively, of the daytime HONO production rate. This points to the significant role of photochemical processes under intense solar radiation. Direct emissions have a negligible contribution during both the dust and photochemical pollution periods, accounting for less than 2 %. The photolysis of HONO is the dominant loss pathway throughout the day for all the measurement periods, accounting for more than 90 % of the HONO sinks.

The model is subjected to sensitivity tests by increasing or decreasing selected parameters by factors of 2 and 5 (Table S4, Figs. S3 and S4). Even with such a broad range of parameter variation, the heterogeneous reaction of NO₂ on the aerosol surface and the photolysis of nitrate to form atmospheric HONO remain significant sources of HONO in both dust and photochemical pollution periods. This suggests that our current understanding of HONO sources, based on existing mechanisms, can generally explain the observed concentrations of HONO. However, it is important to note that differences in parameter selection can significantly affect the relative contributions of each pathway. Given the considerable uncertainties in the uptake coefficient of NO₂ and the enhancement factors of photolysis of nitrate, further experimental studies are necessary to evaluate their effects on HONO under different environmental conditions.

3.3 Missing daytime HONO source in the sea case

Recent field studies suggest potential unidentified daytime sources of nitrous acid in the marine atmosphere, with high daytime HONO levels recorded (Ye et al., 2016a; J. Yang et al., 2021). Figure 5 shows the diurnal variations of the selected sea case and land case, with the corresponding statistical results in Table S3. In the sea case, daytime concentrations of typical primary pollutants, such as CO and SO₂, are significantly lower than those in the land case (251 ± 59 ppbv vs. 335 ± 115 ppbv and 0.7 ± 0.4 ppbv vs. 1.4 ± 0.8 ppbv for CO and SO₂, respectively). Concurrently, the sea case shows a lower daytime temperature (15.2 ± 3.0^∘C vs. 18.7 ± 3.8^∘C) and a higher RH (76.3 ± 25.9 % vs. 47.3 ± 20.3 %) compared to the land case. This is consistent with our understanding of marine air masses, which tend to be cleaner and more humid. These findings validate our classification method of the land case and the sea case based on the backward air mass trajectory.

Figure 5 Average diurnal variations of HONO and related parameters in the “sea case” and the “land case” during the campaign at Mount Lao. The shaded area indicates half of the standard deviation.

Secondary pollutants in marine air masses, such as O₃, PM_2.5, and pNO${}_{\mathrm{3}}^{-}$, also register lower daytime concentrations than in the land case (59.4 ± 10.3 ppbv vs. 63.4 ± 13.3 ppbv, 13.2 ± 5.8 µg m⁻³ vs. 29.9 ± 22.8 µg m⁻³, and 1.3 ± 0.5 µg m⁻³ vs. 10.0 ± 3.3 µg m⁻³, respectively). Though the daytime HONO concentration in marine air masses is less than that in the land case (0.42 ± 0.25 ppbv vs. 0.51 ± 0.22 ppbv), it maintains a relatively high level, particularly during intense photolysis periods around noon when the HONO concentration in the sea case marginally increases. NO_x concentrations in the sea case are also lower than in the land case, but the difference is less substantial than primary pollutants, with both NO and NO₂ showing concentration peaks around noon. Nighttime observations in the sea case show a higher HONO $/$ NO₂ ratio (0.12), which has been noted in earlier studies (Zha et al., 2014), suggesting strong nocturnal HONO formation in marine air masses. Here, we focus on the sources of HONO during the day under the influence of marine air masses.

Utilizing the updated chemical model, we examine the HONO budgets in both the sea and land cases. In the land case, the simulated HONO concentration aligns well with the observed HONO concentration, with a high IOA value of 0.94 (Fig. 6a). The peak HONO production rate observed at Mount Lao (2.69 ppbv h⁻¹) surpasses that calculated in continental air masses at Hok Tsui, Hong Kong (less than 1.5 ppbv h⁻¹) (Gu et al., 2022b). The contributions of photo-enhanced heterogeneous reactions of NO₂ on the aerosol surface (22 %) and photolysis of pNO${}_{\mathrm{3}}^{-}$ (20 %) are comparable (Fig. 6c). Model results reveal that the homogeneous reaction between NO and OH is the predominant HONO formation pathway, contributing an average of 44 % (0.52 ± 0.38 ppbv h⁻¹). Despite a lower absolute rate than in urban areas, the relative contribution is significant (Gu et al., 2022a; Y. Yu et al., 2022). This result suggests that, similarly to the findings in the cases of dust and photochemical pollution, the current model's parameterization reasonably accounts for the observed HONO concentration in the land case.

Figure 6 Comparison of the observed and modeled daytime (07:00–17:00) HONO concentrations and modeled HONO budgets in the land case (a, c) and the sea case (b, d).

However, in the sea case, while the updated model has improved in simulating HONO concentrations, with an average concentration increasing from 0.05 to 0.11 ppbv, it falls short of the observed concentration (0.42 ppbv), indicating a substantial unidentified HONO source. At noon, the missing HONO production rate (P_missing) can reach up to 0.70 ppbv h⁻¹. This value is slightly higher than the result calculated by Meusel et al. (2016) on Cyprus (about 0.5 ppbv h⁻¹) but lower than that reported by J. Yang et al. (2021) in coastal Qingdao (up to 1.83 ppbv h⁻¹, including all non-NO + OH pathways). Sensitivity tests are conducted to assess the impact of parameter selection on simulation results, but even with much larger parameters (Table S4), the model fails to explain the observed HONO concentrations (Fig. S5).

Figure 7 Comparison of simulated net O₃ and OH radical production rates (solid lines) and concentrations (dashed lines) with and without HONO measurement data constraints and relative diurnal contributions of different OH radical sources with HONO constrained in the overall case (a, c, e) and the sea case (b, d, f). The shaded area indicates the standard deviation.

The correlation analysis reveals that the missing HONO production rate correlates strongly with JNO₂ and JNO₂ × pNO${}_{\mathrm{3}}^{-}$ (Fig. S7), with correlation coefficients (r) of 0.90 and 0.73, respectively. This indicates that the missing HONO sources are closely related to photochemical processes. This concurs with recent multi-site HONO analysis results, which propose a significant role of photochemical processes in observed HONO concentrations in remote areas (Jiang et al., 2022). We postulate that all missing HONO originates from photochemical processes and have calculated the required EFs for nitrate photolysis rates (Sect. S2) and the uptake coefficient required for NO₂ on aerosol and ground surfaces (Fig. S6). To account for the observed HONO concentrations, the required EF is approximately 4000. While Andersen et al. (2023) noted that the EF increases with decreasing nitrate concentration, a 4000-fold difference exceeds all laboratory and field observations to date (Ye et al., 2017; Andersen et al., 2023). The required uptake coefficient of NO₂ on aerosol and ground surfaces reaches 4 × 10⁻⁴, exceeding previous laboratory studies (Stemmler et al., 2007; J. Liu et al., 2019). Although NO₂ uptake coefficients on the order of 10⁻⁴ have been measured in laboratory experiments under conditions of high SO₂ concentration (280 ppbv) and moderate acidity (pH = 5) (Liu and Abbatt, 2021), our observational site features a lower SO₂ concentration (∼ 1 ppbv) and slightly acidic aerosols (pH = 3.1). These conditions suggest that the uptake coefficients should be considerably lower than the laboratory-measured 2 × 10⁻⁴. It is worth noting that previous research has indicated that the presence of halogens can enhance NO₂ uptake, which could potentially explain the higher NO₂-to-HONO conversion ratios in marine environments (Yabushita et al., 2009). However, further research is needed to explore this possibility. Overall, the observed missing HONO source in the sea case cannot be explained by the current photochemical processes. This deviates from the findings of Zhu et al. (2022), who discovered that nitrate photolysis could explain the observed HONO concentrations in clean marine air masses using a moderate EF of 29. In recent years, many observations have noted distinct HONO characteristics under the influence of marine air masses, differing from those in continental air masses, but the specific mechanisms are still lacking (Meusel et al., 2016; Crilley et al., 2021; J. Yang et al., 2021; Zhu et al., 2022). The ocean surface contains abundant nitrogen-containing substances (e.g., dissolved nitrate, ammonia, aliphatic amine, dissolved free amino acids) (Donaldson and George, 2012; Altieri et al., 2016) encompassing both organic and inorganic nitrogen. This is particularly true in polluted coastal areas where the surface nitrogen content is rich. This merits investigation into whether these nitrogen-containing substances in the alkaline sea surface microlayer can directly affect HONO production or enhance HONO formation by photolysis on the formed sea salt aerosols. Additionally, the presence of halogens in oceanic air masses might promote nitrate photolysis (Zhang et al., 2020).

3.4 Impacts of HONO on O₃ and OH production

To quantify the impact of HONO, especially in the marine atmosphere, on O₃ and OH radicals, we conduct further scenario simulations using a chemical box model. In the “with HONO” scenario, we input the observed HONO concentrations to constrain the model. In contrast, in the “without HONO” scenario, we turn off seven HONO production pathways summarized in Table 1 and set the input HONO concentrations to zero. The differences between these scenarios illustrate the impact of HONO chemistry on O₃ and OH radicals in the atmosphere. To further investigate the effect of the missing HONO sources on marine air masses, we establish a third simulation scenario, “without missing HONO”, in the marine air mass simulation. In this scenario, the model includes the latest HONO formation mechanisms discussed earlier but without the constraint of observed HONO.

HONO significantly influences the production of O₃ and OH radicals, regardless of whether the overall situation during the observation period (“overall case”) or the situation within the marine air masses (sea case) is considered (Fig. 7). The net O₃ production rate is determined by the difference between the O₃ production rate (P(O₃)) and loss rate (L(O₃)) (Xue et al., 2014). Specifically, the absence of HONO results in a decrease in the net O₃ production rate and the OH radical primary production rate in the overall case from 7.39 ppbv h⁻¹ and 1.44 × 10⁷ molec. cm⁻³ s⁻¹ to 3.41 ppbv h⁻¹ (a 54 % reduction) and 2.81 × 10⁶ molec. cm⁻³ s⁻¹ (an 81 % reduction), respectively. Regarding concentration, the absence of HONO chemistry results in reductions in the average OH radical concentration from 3.6 × 10⁶ to 1.9 × 10⁶ molec. cm⁻³ and the peak OH concentration from 5.2 × 10⁶ to 2.7 × 10⁶ molec. cm⁻³. Similarly, in the marine air masses, the production rates of O₃ and OH decrease from 6.22 ppbv h⁻¹ and 7.69 × 10⁶ molec. cm⁻³ s⁻¹ to 3.33 ppbv h⁻¹ (a 46 % reduction) and 2.14 × 10⁶ molec. cm⁻³ s⁻¹ (a 72 % reduction), respectively, without HONO chemistry. These findings are consistent with previous observational studies (J. Yang et al., 2021; Gu et al., 2022a), highlighting the significant impact of HONO on O₃ and OH radicals in the atmosphere.

Notably, in the sea case, if the observed values are not input as constraints and only the updated mechanisms are used (without missing HONO), the model still significantly underestimates the impact of HONO on O₃ and OH radicals. Specifically, missing marine HONO sources contribute 36 % to the peak net O₃ production rate from 9.24 to 5.90 ppbv h⁻¹ and 28 % to the peak OH concentration (from 3.4 × 10⁶ to 2.4 × 10⁶ molec. cm⁻³). Regarding the relative contribution to OH radicals, HONO accounts for 79 % and 55 % in the overall case and sea case, respectively, both exceeding the combined contribution of other pathways (photolysis of O₃ contributes 14 % and 25 %, respectively). Given the relatively limited observational data on HONO in coastal or marine areas and the unclear understanding of the missing HONO sources in the ocean, the impact of marine emissions on atmospheric oxidizing capacity may be significantly underestimated. This underscores the importance of further research in this area to enhance our understanding of the role of HONO in atmospheric chemistry, especially in marine environments.

4 Conclusions

This study presents a comprehensive investigation of the characteristics and sources of nitrous acid (HONO) in the coastal environment of Qingdao. The analysis utilizes observational data from the Mount Lao site in Qingdao during spring 2021 and an updated chemical box model that integrates HONO mechanisms. The focus lies on discerning the unidentified HONO sources in marine air masses and comprehending their effects on atmospheric chemistry, emphasizing O₃ and OH radical production.

Despite a relatively pristine coastal atmosphere, HONO concentrations are considerably higher than previously thought (0.46 ± 0.37 ppbv), notably during daytime. This observation persists in lower primary pollutant concentrations such as CO and SO₂ within marine air masses, suggesting missing HONO sources tied to photochemical processes. An updated chemical model's simulation reveals that the mechanisms behind HONO formation can satisfactorily account for the observed HONO concentrations during the dust and photochemical pollution periods. However, in marine scenarios, the model falls short of matching observed concentrations, pointing to a strong unidentified HONO source within the marine atmosphere. Sensitivity tests and correlation analyses emphasize the importance of photochemical processes in unidentified HONO sources. Nevertheless, if these unknown sources are attributed to either nitrate photolysis or heterogeneous NO₂ reactions, the necessary nitrate photolysis rate and the heterogeneous uptake coefficient of NO₂ would surpass the upper thresholds established by current laboratory studies. In light of these findings, future research must target uncovering the mechanisms behind the missing HONO sources in marine air masses. Specifically, the role of nitrogen-containing substances at the ocean's surface and the potential influence of halogens in promoting nitrate photolysis warrant further examination.