1 Introduction
Nitrate (NO) and its precursors (NO NO NO2) play important roles in the atmosphere. Tropospheric NOx cycling produces ozone (O3), a key component of the atmospheric oxidative capacity (Finlayson-Pitts and Pitts, 2000). NOx emitted from various sources undergoes oxidation to form HNO3, which contributes to acid rain (Shammas et al., 2020) and particulate matter (as NO; Zhai et al., 2021); moreover, upon deposition, nitrate alters the nutrient balance in ecosystems (Duce et al., 2008). Owing to increasing fossil fuel and chemical fertilizer use since the beginning of the industrial revolution, NOx levels in the atmosphere have increased, which is reflected in the elevated NO concentrations in ice cores, including those collected from Greenland (Neftel et al., 1985; Mayewski et al., 1986). Despite efforts to curb NOx emissions through pollution mitigation measures and NOx removal techniques, including three-way catalytic converters, the decline in the ice core NO concentrations has been relatively gradual, as observed in Greenland (Iizuka et al., 2017) and in an Alpine ice core (Eichler et al., 2023). This underscores the necessity to understand atmospheric NO dynamics beyond just precursor NOx emissions, for which ice core NO can offer invaluable historical perspectives.
In addition to NO concentrations, NO stable isotopic compositions provide valuable information. Nitrogen isotopes (N) differ among NOx sources and can be used to identify the origin of the NO (Hastings, 2010; Hastings et al., 2013). Previous studies of Greenland ice cores have identified decreases in the N values of NO (N(NO) as early as 1850 CE, which subsequently accelerated after 1950 CE (Hastings et al., 2009; Geng et al., 2014). These decreases in N have been interpreted as a change in source: increased anthropogenic emissions of NOx from fossil fuel combustion (Hastings et al., 2009) and/or NOx derived from soil amended with fertilizer (Felix and Elliott, 2013). In addition, complex factors control N(NO). Previous studies have shown that isotopic fractionation can occur during gas–particle partitioning and washout (Freyer, 1991), kinetic NO2 oxidation (Walters and Michalski, 2015a), and NOx cycle equilibrium in the atmosphere (Walters et al., 2015b; Walters and Michalski, 2016). Another interpretation of the decrease in N in ice cores is related to changes in isotopic fractionation between gaseous HNO3 and particulate NO that resulted from acidity changes (Geng et al., 2014). The mass-independent oxygen isotope fractionation signals (O O O) of NOx and NO can reflect the oxygen source during oxidation (Michalski et al., 2003; Alexander et al., 2009). Positive O values of NO (O(NO) occur as a result of excess 17O (i.e., deviation from mass-dependent fractionation) transferred from O3 to NO during photochemical cycling of NO and NO2 and the oxidation of NO2 into HNO3. Thus, middle to high latitudes have O(NO) values of 22 ‰–34 ‰ for atmospheric deposition (Michalski et al., 2003, 2012). The O(NO) measurements in ice cores have also been used to investigate historical atmospheric processes (Geng et al., 2017).
However, postdepositional NO loss in snow or ice can reduce NO concentrations and change its isotopic compositions (Röthlisberger et al., 2000, 2002; Frey et al., 2009; Akers et al., 2022). Indeed, NO in snow can undergo photolysis by ultraviolet (UV) light (–350 nm; Berhanu et al., 2014), which produces NO2 that is released into the atmosphere via diffusion or wind pumping. Although NO2 can partially reoxidize into NO in the atmosphere, postdepositional processes can lead to decreases in the NO concentrations in ice cores (Meusinger et al., 2014; Erbland et al., 2015). In addition, postdepositional processes can also cause significant isotopic fractionation (from ‰ to ‰), and the remaining NO becomes enriched in 15N (Berhanu et al., 2015). In contrast, O(NO) is not directly affected by photolysis but, rather, by the cage effect, in which the intermediate photoproducts (NO2, NO, or ONOO−) undergo recombination reactions within snow grains to reform nitrate with exchange with water oxygen or react with radicals (e.g., OH) to regenerate NO after being emitted to the atmosphere (McCabe et al., 2005; Jiang et al., 2021). The alteration of O(NO) can also occur through the reoxidation of NO2 sourced from snow, which leads to nitrate formation in the overlying atmosphere (Erbland et al., 2013). Because NO photolysis in snow only occurs in the photic zone, the degree of postdepositional alteration is mostly controlled by the snow accumulation rate, as demonstrated in Antarctica (Akers et al., 2022). Even at the Greenland Summit ice core site, the N(NO) values of the snowpack are higher than those of the surface snow and the overlying atmosphere (Jarvis et al., 2009; Geng et al., 2014; Fibiger et al., 2016). The Greenland Ice Sheet Project 2 (GISP2) ice core exhibits decreasing N(NO) values from glacial to interglacial periods (Hastings et al., 2005; Geng et al., 2015), which has been interpreted as the result of two potential causes: (1) changes in the NOx source or (2) postdepositional effects related to the snow accumulation rate and dust concentrations. Overall, given the impacts of postdepositional processes, ice core records of NO and its isotopes require careful consideration.
Accurate understanding of atmospheric NO and its related nitrogen cycles can be more reliably obtained from ice core records with minimal postdepositional alterations. In this context, the southeastern Dome (SE-Dome) site in Greenland has distinct characteristics, including a snow accumulation rate of m water equivalent per year (m w.e. a−1; for the period from 1960 to 2014; Iizuka et al., 2017), which is approximately 4 times greater than that at the Summit site ( m w.e. a−1; Fig. 1c; Geng et al., 2014). The relatively high accumulation rate is expected to reduce the impact of postdepositional loss, preserving a more representative record of atmospheric NO deposition. Furthermore, statistical analysis using the Mann–Kendall test for monotonic trends (Kendall, 1975; Mann, 1945) revealed no significant decadal trends in snow accumulation at either the SE-Dome (Kawakami et al., 2023) or the Summit (; data from Geng et al., 2014) site, suggesting that within-site accumulation variability does not strongly influence nitrate trends at these locations. The Summit and SE-Dome sites receive atmospheric inputs from similar source regions in North America and western Europe, based on Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) 7 d backward-trajectory modeling (Fig. 1b and c). However, the extent to which postdepositional processes affect NO isotopic compositions at these sites has not been systematically assessed. In this study, we present the N(NO) and O(NO) records obtained from a 90.45 m ice core drilled at the SE-Dome site. By applying the TRansfer of Atmospheric Nitrate Stable Isotopes To the Snow (TRANSITS) model (Erbland et al., 2015; Jiang et al., 2021) with site-specific modifications, we evaluate the extent to which postdepositional processes influence NO at SE-Dome and assess its suitability for reconstructing past atmospheric NOx emissions. Our results demonstrate that SE-Dome preserves a robust NO isotopic signal with minimal postdepositional modification, making it a valuable archive for investigating anthropogenic changes in atmospheric NO over the Northern Hemisphere, particularly from eastern North America and western Europe.
Figure 1
(a) Map of the southeastern Dome (SE-Dome, star) and Summit (circle) sites in Greenland (annual precipitation from ERA5 reanalysis climate data; Hersbach et al., 2020). Probability distributions for the air masses overlying the (b) SE-Dome and (c) Summit sites from 7 d three-dimensional back-trajectory analysis based on HYSPLIT modeling (1960–2019). Detailed back-trajectory analysis procedures were the same as those in Iizuka et al. (2018). (d) Annual accumulation rate at the SE-Dome (Furukawa et al., 2017) and Summit (Geng et al., 2014) sites.
[Figure omitted. See PDF]
2 Materials and methods2.1 Samples
This study was based on a 90.45 m ice core drilled at the SE-Dome site (67.18° N, 36.37° W; 3170 m a.s.l., meters above sea level) in 2015 (Iizuka et al., 2016). The age–depth scale was determined using the oxygen isotope matching method, which matches the O variations between ice core records and isotope-enabled climate model estimates, and indicated that this ice core covers the period from 1959 to 2014 (Furukawa et al., 2017). The reliability of this dating method generally falls within the 95 % confidence interval (typically around an average of months). The greatest uncertainty was reported at 2 months in some years (Furukawa et al., 2017). We divided the ice core samples into four seasons: spring (21 March–20 June), summer (21 June–20 September), fall (21 September–20 December), and winter (21 December–20 March). For samples analyzed at a two-season resolution (1959–1980 and 1995–2014), spring and summer were combined into summer, whereas fall and winter were combined into winter. This approach was adopted to ensure a consistent sample size and minimize analytical uncertainty for periods with lower temporal resolution, while still capturing seasonal variations relevant to NO deposition and atmospheric conditions.
All SE-Dome ice core samples used in this study were stored in a refrigerated room ( °C) at the Institute of Low Temperature Science (Hokkaido University, Sapporo, Japan). Each ice sample (3 cm 4 cm cross-dimension) was cut using a band saw in a refrigerated room ( °C) and decontaminated by removing the outermost mm of ice with a ceramic knife in a class 10 000 clean booth, resulting in a loss of approximately 30 % of the original sample weight. The remaining 70 % of the cleaned samples was shipped frozen ( °C) to the Tokyo Institute of Technology (Tokyo Tech, Yokohama, Japan). The samples were then stored in a freezer at °C until analysis.
2.2 Sample analysis
The NO in each sample () was separated from other ions using ion chromatography (IC; Dionex Integrion, Thermo Fisher Scientific) according to the methods described by Noro et al. (2018). Changes in the isotopic compositions of NO during ion chromatographic separation were ‰ for N values and within the analytical error range for O values (Noro et al., 2018). After ion separation, NO in solution was converted and neutralized to the Na+ form by passing it through an ion exchange column. The isotopic compositions of NO were measured using a bacterial method that converts NO to N2O (Sigman et al., 2001; McIlvin et al., 2011), followed by N2O decomposition via microwave-induced plasma (MIP), a technique developed at Tokyo Tech (Hattori et al., 2016). Isotopic reference materials, as well as United States Geological Survey (USGS) standards 32, 34, 35, and their mixtures (prepared in 18.2 M cm water), were also analyzed with the samples using the same analytical processes.
Stable isotopic compositions are reported as , where denotes 15N, 17O, or 18O, while denotes the isotope ratios, such as , , and , determined for both sample and standard materials. The N, O, and O values are reported in per mil (‰) notation. The N values are relative to atmospheric N2 (air), while the O and O values are relative to Vienna Standard Mean Ocean Water. By propagating the analytical uncertainties for the IC separation and replicating isotopic measurements of USGS standards 34, 35, and 32, the estimated combined uncertainties were ‰ for both N(NO) and O(NO).
2.3 TRANSITS modeling
The TRANSITS model (Erbland et al., 2015), a multilayer one-dimensional isotopic model, was used to simulate NO recycling across the air–snow interface (i.e., UV photolysis of NO, NOx emission, local NO2 oxidation, and NO deposition) and its associated isotopic effects. The model is operated at a weekly resolution (52 time steps per year), and the default snow depth resolution is 1 mm. In each step, NO photolysis is calculated according to the depth-dependent photochemical flux and NO concentration. All generated NO2 enters the atmosphere and is reoxidized to NO, which is deposited on the surface snow with the primary NO from long-range transport in the next step. The original snow moves downward as snowfall continues, and newly deposited snow is divided into 1 mm layers. Once the NO is buried beneath the light transmission band, the layer is regarded as an archive. We adapted the parameters of the TRANSITS model, originally developed for the Summit site by Jiang et al. (2021). This model reproduces the seasonal variation pattern of N(NO) in the surface snow at the Summit site reported by Jarvis et al. (2009), highlighting the importance of postdepositional processes at site. In addition, this model estimated the net loss of NO (4.1 %) and associated changes in N(NO) ( ‰) and O(NO) ( ‰) between primary deposition and NO archived in the ice, under an estimate of the horizontal export fraction of locally reoxidized NO () of 35 % (Jiang et al., 2021). In this study, we applied the same model under the SE-Dome condition by adjusting the parameters to examine the effects of snow NO photolysis on the NO concentration and its isotopes.
The snow accumulation rate was set at 1.01 m w.e. a−1, based on the 1960–2014 average from SE-Dome ice core data (Iizuka et al., 2017), with additional tests conducted at rates of 0.25, 0.6, and 1.4 m w.e. a−1. The mass balance of NO between the snow and atmosphere depends on NO influxes and outfluxes. We expressed the NO flux as , which includes the primary NO flux from long-range transport (), NO flux from NO photolysis (), atmospheric NO deposition flux (), and ice core NO flux buried beneath the light band (). These fluxes reflect changes in NO and its isotopic compositions in the snow and atmosphere. The TRANSITS model considers that two processes can change N(NO) as a result of isotope fractionation from UV photolysis and NO deposition (i.e., co-condensation and dry deposition). The nitrogen isotope fractionation constant during photolysis () was calculated using a ratio of the 14NO and 15NO photolysis rates (J , where represents the photolysis rate constant) in each snow layer (Erbland et al., 2013). The at different depths () () was calculated according to Eq. (1):
1 where is the actinic flux, while and are the quantum yield and absorption cross-section of NO photolysis, respectively. The quantum yield of NO photolysis has significant uncertainties (Meusinger et al., 2014). However, it is unlikely that the quantum yield of NO photolysis would differ substantially between the SE-Dome and Summit sites. As this study compares the differences between these sites, which have notably different snow accumulation rates, the quantum yield of NO photolysis at the SE-Dome site was set to the same value (0.002) as that estimated for the Greenland Summit site by Jiang et al. (2021). The absorption cross-sections of 14NO () and 15NO () were derived from Berhanu et al. (2014). The nitrogen isotope fractionation constant during deposition () was set to ‰ (Erbland et al., 2015). For the oxygen isotopes, only the mass-independent fractionation signal (O) was modeled. The cage effect (i.e., decrease in O of the snow NO owing to secondary chemistry during NO photolysis; McCabe et al., 2005) was set to 15 % according to Erbland et al. (2015), and the TRANSITS model calculated the exchange of oxygen atoms with water during UV photolysis and atmospheric NO–NO2 cycling, both of which alter O. The O(NO) fractionation mechanisms in the TRANSITS model during these processes are explained in detail in Jiang et al. (2021).
The atmospheric boundary layer at the SE-Dome site was assumed to be a zero-dimensional well-mixed box, and the snowpack was assumed to be a stack of snow layers deposited at different times. Weekly air temperatures (), pressures (), and average boundary layer heights () from 1950 to 2020 were obtained from the second-generation European Centre for Medium-Range Weather Forecasts atmospheric analysis of global climate (ERA5) (Hersbach et al., 2020; Khalzan et al., 2022). O3, OH, peroxyl radical (RO2 and HO2), and BrO concentrations were used to calculate the rates of NO–NO2 cycling and NO2 oxidation to HNO3. However, because these records were not available for the SE-Dome, they were extracted from the outputs of the v.12.9.3 (10.5281/zenodo.3974569, The International GEOS-Chem User Community, 2020) GEOS-Chem atmospheric chemical transport model (
An -folding depth, which is the depth to which light enters the snow layer and attenuates to an initial intensity of (owing to absorption and scattering), for the SE-Dome site was calculated using the snow density (), the calculated specific surface area (SSA), and fixed light-absorbing impurity concentrations (Jiang et al., 2021). The of 400 kg m−3 for the SE-Dome site was obtained from an observation at the SE-Dome (Oyabu et al., 2016). The SSA for the SE-Dome site was determined to be 47.0 m2 kg−1 using the relationship between the SSA and , according to a previous study (Domine et al., 2007): 2 where SSA is in units of square centimeters per gram (cm2 g−1) and the units for were changed to grams per cubic centimeter (g cm−3). For the light-absorbing impurity concentrations, we established constant concentrations of the three main light-absorbing impurities in snow: dust, soot (BC), and organic humic-like substances (HULISs). The dust concentration was set to 33.94 ng g−1 according to an average concentration for the SE-Dome ice core from 1960 to 2014 (Amino et al., 2021). Owing to a lack of direct observations for BC and HULISs at the SE-Dome, we assumed these concentrations based on the Ca2+ concentration ratio between the Summit and SE-Dome sites. Here, [BC]Summit and [HULIS]Summit were set to 1.4 and 31 ng g−1, respectively, according to Jiang et al. (2021). Furthermore, [Ca2+] at the Summit site was set to 6.5 ng g−1 according to an average of the 2 m shallow snowpack observation (Geng et al., 2014); [Ca2+] at the SE-Dome site was set to 11.6 ng g−1 according to an average from the SE-Dome ice core from 1960 to 2014 (Iizuka et al., 2018). Thus, [BC]SE-Dome and [HULIS]SE-Dome were calculated as 2.2 and 47.6 ng g−1, respectively, and used for the model calculation. An -folding depth of 10 cm was obtained based on the above inputs. The calculated -folding depth for the SE-Dome site was consistent with previous estimates from the GEOS-Chem model investigating the impact of postdepositional effect in snow (Zatko et al., 2016).
The horizontal export fraction of locally reoxidized NO () under the SE-Dome condition was calculated with the same scheme described for the Antarctic Plateau (Erbland et al., 2015) and Greenland Summit (Jiang et al., 2021), as outlined in the following equations: In these equations, , , and represent the lifetimes of horizontal transport, oxidation of NO2 by OH radicals, and vertical deposition, respectively. and denote the respective summer boundary layer height and horizontal characteristic, while is the mean horizontal wind speed, is the rate constant for the NO OH reaction, and is the dry deposition velocity of HNO3 (Jiang et al., 2021). The values used for the calculation are summarized in Table S2, and the same physicochemical values as those of Summit were used, while parameters such as temperature and boundary layer height were incorporated from ERA5 data. We obtained to be 47 % under the SE-Dome condition, but the calculated may oversimplify the processes governing NO deposition and the chemical loss pathways of NOx, as discussed previously (Jiang et al., 2021). Therefore, we considered the sensitivity of postdepositional alteration to variations in between initial deposition and the point at which NO becomes archived in the ice.
At the initial time () in the TRANSITS model, the NO concentration was set to 71.12 ng g−1 based on the average NO concentration in the SE-Dome I ice core (Iizuka et al., 2018), while the N(NO) and O(NO) values in the snowpack were set to 0 ‰ and 30 ‰, respectively, according to previous TRANSITS settings (Erbland et al., 2015; Jiang et al., 2021). This initial isotopic parameter does not affect the model interpretation of changes in N(NO) and O(NO) due to postdepositional processing. The 3-year distributions of NO and its isotopes were simulated with and without NO photolysis scenarios under the SE-Dome condition, from which profiles of NO concentrations, N(NO), and O(NO) were output. The parameters used in the TRANSITS model are summarized in Table 1 and Supplement data file 1.
Table 1Parameters used in the TRansfer of Atmospheric Nitrate Stable Isotopes To the Snow (TRANSITS) model for the SE-Dome ice core.
Parameters | Description | Value | Unit* | Data origin |
---|---|---|---|---|
Snow density | 400 | kg m−3 | Oyabu et al. (2016) | |
SSA | Snow-specific surface area | 47.0 | m2 kg−1 | Domine et al. (2007) |
[BC]SE-Dome | BC concentration | 2.2 | ng g−1 | See text |
[Dust] | Dust concentration | 33.94 | ng g−1 | Amino et al. (2021) |
[HULIS]SE-Dome | HULIS concentration | 47.6 | ng g−1 | See text |
TCO | Total column ozone | See Supplement data file 1 | DU | Jiang et al. (2021) |
Boundary layer height | See Supplement data file 1 | m | ERA5 (Hersbach et al., 2020; Khalzan et al., 2022) | |
Temperature | °C | |||
Pressure | hPa | |||
[O3] | O3 concentration | 26.66–32.30 | ppb | GEOS-Chem v12.9.3 ( |
[BrO] | BrO concentration | 0.06–0.76 | ppt | |
and | and concentrations | See Supplement data file 1 | molec. cm−3 | |
Snow accumulation | 101 | cm a−1 | Iizuka et al. (2017, 2018) | |
Primary nitrate flux | 16.28 | kg N m−2 a−1 | ||
Export fraction | 47 % | – | See Table S2 | |
N isotope fractionation constant during photolysis | ‰ | Erbland et al. (2013) | ||
N isotope fractionation constant during deposition | 10 | ‰ | ||
Cage effect | 15 | % | Erbland et al. (2015) |
* Note that ppt denotes parts per trillion.
2.4 Statistical analysisXLSTAT 2023 (Addinsoft, Paris, France) was used for the Mann–Kendall trend analysis over 1959–2014. SPSS Statistics 25 (IBM SPSS, Armonk, NY, USA) was used to perform the tests of annual changes in the N(NO) and NO concentrations. Statistical significance was set at .
3 Results
3.1 Nitrate isotope records from the SE-Dome ice core
The NO isotope data and fluxes obtained from the SE-Dome ice core are shown in Figs. 2 and S1. The seasonal variations were larger in the samples analyzed at a four-season resolution (1981–1994), which may have been caused by age errors (Furukawa et al., 2017). Accordingly, mass-weighted averages were calculated for the summer and winter fractions from the seasonal samples during 1981–1994. From 1959 to 2014, the N(NO) values were generally higher in summer ( ‰) than in winter ( ‰) (Figs. 2a and S1a). To assess annual changes in N(NO) over this period, we calculated the annual mass-weighted average N(NO) values and found that they decreased from 1959 to 1974 and exhibited no significant () trends after 1975 (mean value of ‰) (Fig. 2a). No clear relationship was observed between the annual variations in the N(NO) and NO concentrations ().
Using a similar method to that employed for N(NO), we also calculated the mass-weighted average and annual mass-weighted average for O(NO). The O(NO) also exhibited a seasonal pattern, with lower values in the summer ( ‰) than those in the winter ( ‰), yielding a mass-weighted average of ‰ over the entire period (Figs. 2b and S1). The average annual O(NO) values were relatively high ( ‰) in 1988, whereas low values were observed in 2013 and 2014 (Fig. 2b). Excluding these particular years, no significant annual increases or decreases () were observed in the O(NO) values.
Figure 2
The NO isotope data and fluxes from the SE-Dome ice core: (a) N(NO), (b) O(NO), and (c) NO flux (mmol m−2 a−1).
[Figure omitted. See PDF]
3.2 TRANSITS model resultsWe aimed to know the changes in NO from primary deposition to the ice core archive. However, these changes primarily depend on the value – the fraction of NO exported from the site of photolysis. Therefore, we calculated the dependency of the N(NO) and O(NO) values on at the SE-Dome using the same approach as Jiang et al. (2021) (Fig. 3). The postdepositional alterations in N(NO) and O(NO) between initial deposition and the ice core NO concentration at the SE-Dome were dependent on . As shown in Fig. 3, an inverse relationship was observed between and N(NO) and O(NO). In the case of N, when is high, a larger proportion of isotopically light NO is removed from the snowpack due to strong isotopic fractionation during photolysis. As a result, the remaining NO in the snow becomes isotopically enriched in 15N. In contrast, for O(NO), when is high, the impact of photolysis-induced isotopic fractionation on O(NO) is minimal; thus, its value remains largely unchanged. Conversely, when is low (i.e., a significant portion of the nitrogen species is emitted and subsequently redeposited), the NO with lower O(NO), originating from NO OH reaction, dominates the signal, leading to a decrease in O(NO). These distinct mechanisms explain why N(NO) and O(NO) exhibit opposite trends.
However, the degree to which the changes in N(NO) and O(NO) were dependent on was less evident at the SE-Dome site than at the Summit site (Fig. 3). Notably, even when the snow accumulation rate for SE-Dome is adjusted from the minimum (0.6 m w.e. a−1) to the maximum (1.4 m w.e. a−1) values, as shown in Fig. 1d, the results indicate that changes in N(NO) and O(NO) are less sensitive to compared to Summit. Furthermore, when a snow accumulation rate of 0.25 m w.e. a−1, equivalent to that used in the Summit study (Jiang et al., 2021), was applied, the variations were nearly identical to those observed at Summit (Fig. S2). This suggests that the differences in postdepositional alterations for N(NO) and O(NO) are primarily caused by differences in accumulation rates.
Using the method of Erbland et al. (2015), was calculated as 47 % at SE-Dome, reflecting an estimated net loss of 1.4 % due to postdepositional alteration in the NO concentration at the site, with corresponding changes in N(NO) and O(NO) of ‰ and ‰, respectively (Fig. 3). In contrast, the estimated net loss of NO under the Summit condition showed higher values (4.1 %) and associated changes in N(NO) ( ‰) and O(NO) ( ‰) when the value was 35 %, as estimated previously (Jiang et al., 2021). Thus, even when using the same evaluation criteria, the postdepositional alteration in NO and its isotopic compositions at the SE-Dome site were smaller than those at the Summit site. We note that, under an extreme condition of %, the estimated changes in N(NO) from initial deposition due to postdepositional processing under the SE-Dome condition were ‰, which is significantly lower than that under the Summit condition of ‰ (Fig. 3).
Figure 3
Sensitivity of the changes in N(NO) and O(NO) of the ice core nitrate to . Positive or negative values indicate deviations from initial deposition. The shaded area in the SE-Dome calculations represents results obtained using snow accumulation rates of 0.6 and 1.4 m w.e. a−1.
[Figure omitted. See PDF]
Figure 4 shows the results obtained from the TRANSITS model for NO and its isotopic compositions for the SE-Dome site when considering an value of 47 %. The model considering photolysis showed a maximum 6 % decrease in the annual NO concentrations during spring and early summer compared to the scenario without photolysis (Fig. 4a). The postdepositional effects (primarily due to photolytic isotopic fractionation) caused a fluctuation in N(NO) of ‰ to ‰, with higher values in summer ( ‰) and lower values in winter ( ‰) (Fig. 4b). The variation in the O(NO) value, which was initially set at 30 ‰, is attributed to a slight decrease in the atmospheric NO concentration owing to its reoxidization during spring to summer. Thus, when photolysis is minimal in the fall and winter, the O(NO) values remained close to the initial value (30 ‰) (Fig. 4c). Conversely, during spring and summer, when N(NO) values increase, a decreasing O(NO) trend was observed (Fig. 4c). However, the extent of this change is minimal, with values reaching a minimum of ‰. The seasonality in the NO concentration and postdepositional alteration in its isotopic composition were less obvious at SE-Dome than that at Summit, where N(NO) varied by ‰ and O(NO) by ‰ (Jiang et al., 2021).
Figure 4
TRANSITS model results for the SE-Dome site: (a) NO concentration, (b) N(NO), and (c) O(NO). The black and red lines represent the variations calculated with and without NO photolysis, respectively.
[Figure omitted. See PDF]
4 Discussion4.1 The N(NO) values from the SE-Dome and Summit sites
Figure 5 shows the annual average N(NO) values obtained from the SE-Dome ice core and the previously published N(NO) values from the Summit site (Hastings et al., 2009; Geng et al., 2014). Decreasing trends in N(NO) were observed in both the Summit and SE-Dome ice cores until approximately 1974, after which no clear changes occurred (Fig. 5). Notably, based on the overlapping analysis period from 1959 to 2006 (), the annual N(NO) values in the SE-Dome ice core were found to be ‰ lower than those in the Summit ice core (Fig. 5). The observed differences in N(NO) values between the SE-Dome and Summit cores may be attributed to (1) variations in the N values of NO deposited at the two sites and/or (2) variations in the degree of postdepositional alterations between the two sites. These two points are discussed in detail below.
Figure 5
The N(NO) values obtained from the SE-Dome (this study) and Summit (Hastings et al., 2009; Geng et al., 2014) ice cores. The shaded area for N(NO) values in SE-Dome represents the propagated errors of the annual average, based on the seasonal concentration and N variations.
[Figure omitted. See PDF]
First, regarding the differences in the N values of NO deposited at the two sites, there are two main sources of the NO deposited in the Greenland ice core. One long-range source is derived primarily from anthropogenic sources outside of Greenland. The other source involves NOx released from NO photolysis within the snowpack, which is then reoxidized and redeposited. Although the air masses at the SE-Dome and Summit sites have similar source regions – North America and western Europe (Fig. 1b and c) – the degree of influence from reactive nitrogen sources differed between sites (Fig. S3). At the SE-Dome site, the contributions from outside Greenland were relatively high, with nearly equal influence from North America and European Union (EU) countries (Fig. S3b). In contrast, the contributions from EU countries were relatively low at the Summit site, while the North American region (mostly Eastern Canada) and inner Greenland had greater contributions (Fig. S3c). The NOx sources from western Europe and North America are not necessarily similar; for example, differences in the relative contributions of various NOx sources are reflected in their N values, with NOx from coal and biomass tending to have higher N values, while NOx emissions from oil, natural gas, and soil tend to have lower N values (e.g., Elliott et al., 2019). To date, there have been no studies comprehensively comparing the isotopic composition of atmospheric NO between Europe and North America. The limited available data show that N values in total atmospheric NO (sum of gaseous HNO3 and particulate NO) in the northeastern USA range from ‰ to ‰ (Bekker et al., 2023), while those in rainwater NO (including both gaseous HNO3 and particulate NO) in Switzerland range from to ‰ (Freyer, 1991), which are indistinguishable. A recent study (Song et al., 2021) compiled the N values of precipitated NO between urban and non-urban areas in Europe ( and , respectively) and North America ( and , respectively), showing no clear distinction between the two regions, although Europe exhibited slightly higher values. Hence, there is no clear evidence that the long-term N trends in European countries consistently remain lower than those in the USA or Canada; thus, the ‰ lower N(NO) values observed at SE-Dome, which is relatively more influenced by air masses from Europe, cannot be explained solely by air mass origin differences.
Considering the potential impact of snow-sourced NOx and reoxidized NO, it is important to note that the extent of recycled NOx from NO photolysis in the Greenland ice sheet differed between these two sites. As modeled by Zatko et al. (2016), recycled NOx is typically more important at inland sites such as the Summit than coastal sites such as the SE-Dome. Additionally, the contribution of air masses from inside Greenland was higher at the Summit site than at the SE-Dome (Fig. S3). Nevertheless, the N values of NOx and reoxidized NO are typically low due to isotopic fractionation during snow NO photolysis in the snow and ice, with N(NO) in high-latitude air masses attributed to photochemical NOx production in snow, resulting in N(NO) values of ‰ to ‰ in polar regions (e.g., Savarino et al., 2007; Morin et al., 2009; Shi et al., 2021). Thus, the contribution of locally recycled NO, which was greater at the Summit, cannot explain why N(NO) values were lower at the SE-Dome than at the Summit site.
Finally, regarding the differences in postdepositional alterations between two sites, we applied parameters specific to the SE-Dome in the TRANSITS model (Figs. 3 and 4). The model results for the SE-Dome, accounting for postdepositional NO photolysis, showed a net NO loss of 1.3 % and a N(NO) increase of ‰ (see Sect. 3.2). In comparison, the Summit condition resulted in a % net NO loss and a ‰ N(NO) increase (Jiang et al., 2021). Although the lower N(NO) values at the SE-Dome can be partially explained by the model, they cannot be fully accounted for quantitatively. However, the estimated ‰ increase in N(NO) at the Summit site may be underestimated due to an underestimation of (Jiang et al., 2021). Indeed, an observational study (Honrath et al., 2002) indicates that most of the NOx and/or HNO3 emitted from the snow at Summit is largely exported from the local boundary layer if no wet deposition occurs, suggesting that the value can reach under Summit conditions. Therefore, the actual net NO loss and N(NO) variation at the Summit site may have been larger than the 4 % estimated by Jiang et al. (2021). Thus, when considering higher values, the difference due to postdepositional alterations could be higher than the modeled difference between the SE-Dome ( ‰, this study) and Summit ( ‰, Jiang et al., 2021) sites. Indeed, when considering an extreme condition of %, the difference between the Summit and SE-Dome sites becomes ‰ (Fig. 3). Overall, although there is some uncertainty in the model, it is likely that the SE-Dome experienced less postdepositional alteration, thus preserving the atmospheric N(NO) values more effectively than at the Summit site.
Although the contribution of different NO sources cannot be entirely ruled out, our analysis shows that the observed N(NO) value at SE-Dome, which is ‰ lower than that at Summit, can largely be attributed to differences in postdepositional alterations.
4.2 Seasonal variations in NO isotopesThe postdepositional effect, as estimated from the TRANSITS model, yielded a summer–winter difference in N(NO) of ‰ and a difference smaller than 0.5 ‰ in O(NO), as described in Sect. 3.2 (Fig. 4b and c). In contrast, except for the anomalous years (1959–1961, 1972, 1995, 2005, and 2013), the observed summer–winter differences for respective years in the SE-Dome ice core were ‰ (0.4 ‰–9.8 ‰) for N(NO) and ‰ (from ‰ to ‰) for O(NO), respectively (Fig. 2a and b), which were larger than the differences estimated by the TRANSITS model. Thus, the observed differences between the summer and winter NO isotopes were not solely explained by postdepositional alteration. Consequently, seasonal differences in N(NO) and O(NO) likely reflect atmospheric changes.
The observed seasonal N(NO) trend (high in summer and low in winter) at the SE-Dome site was consistent with observations made at two coastal Arctic sites (Morin et al., 2008, 2012) but inconsistent with typical seasonal N(NO) values of aerosols in midlatitude regions that are high in winter and low in summer (Freyer, 1991; Freyer et al., 1996; Lim et al., 2022). Although the specific process has yet to be identified, the factors controlling high N(NO) values in the summer have been comprehensively reviewed (Jiang et al., 2024, and references therein). One possibility is that physicochemical transformations of NO related to temperature influence N(NO) values, as suggested by a strong correlation between high N(NO) values and summer air temperatures (Morin et al., 2008). Another possibility is the incursion of anthropogenic sources, as proposed by Morin et al. (2009), which is supported by indications that air parcels originating from regions with greater anthropogenic influence carry higher N(NO) values. This is supported by observational studies on atmospheric N(NO) (e.g., Vicars and Savarino, 2014) and the increased frequency of air masses originating from North America during summer compared to winter (Kahl et al., 1997). While definitive conclusions regarding these observations have not yet been determined, it is hypothesized that the observations may be influenced by a combination of factors, including NOx sources, gas–particle partitioning variability influenced by temperature (Freyer, 1991) and acidity (Geng et al., 2014), oxidation pathways (Walters et al., 2015a), and differences in transport efficiency and removal processes (Heaton, 1987; Beyn et al., 2014). Future studies should examine the differences in N(NO) variations between both the source and remote regions. Such comparative analyses could enhance the current understanding of the underlying processes that influence isotopic compositions in different geographical contexts.
The observed seasonal changes in O(NO) (high in winter and low in summer; Fig. 2b) were consistent with typical seasonal variations in O(NO) (e.g., Michalski et al., 2003). In summer, increased sunlight promotes the formation of HNO3 via NO OH reactions, leading to lower O(NO) values. Conversely, in winter, N2O5 hydrolytic or NO3 radical pathways forming HNO3 in the presence of O3 predominate and result in increased O(NO) levels. Although this kind of seasonal variation in O(NO) is well known, we confirmed the historic occurrence of similar seasonal variations in the atmosphere. Although the scope of the current study limits further discussion in this regard, future research should explore the differences in O(NO) between summer and winter during the preindustrial period, when anthropogenic contributions of NO were significantly lower.
4.3 Decadal variations in NO isotopes
As discussed in Sect. 4.1, the SE-Dome ice core recorded atmospheric NO deposition with minimal postdepositional effects. The decadal N(NO) trend obtained from ice cores in Greenland has been interpreted to indicate changes in the NOx source (Hastings et al., 2009) and/or atmospheric acidity from the beginning of the industrial revolution to the present (Geng et al., 2014). In response to emission controls since 1975, reasonable changes in dominant NOx emissions and the adoption of NOx removal technology (such as three-way catalytic converters) are expected, which can affect the N(NO) values (e.g., Walters et al., 2015b). As for acidity, if N(NO) is primarily controlled by atmospheric acidity, as proposed by Geng et al. (2014), N(NO) should have increased after approximately 1975 when the atmospheric acidity decreased (owing to SO2 emission controls; Hattori et al., 2021). However, the N(NO) values obtained herein did not increase until 2014, indicating that there are multiple factors controlling ice core N(NO) values. As this study only covers a relatively limited period (60 years), future studies should address and compare longer ice core records from different regions. Such comparisons would be beneficial for understanding the factors behind isotopic variations, thereby enabling more accurate interpretations of isotopic records reconstructed from ice cores.
The unusually high NO fluxes observed in the summer of 1987 (1.97 mmol m−2 a−1) and the spring of 1992 (1.38 mmol m−2 a−1) were also notable (Fig. 2c). The extent of forest fires in North America could be the primary driver of this phenomenon, based on coincident high NH fluxes during these periods (Iizuka et al., 2018). In 1992, the Mount Pinatubo eruption may have influenced the observations, considering the high SO concentration (13.7 ; Iizuka et al., 2018). The N(NO) values during the summer of 1987 (1.3 ‰) and the spring of 1992 (4.8 ‰) were relatively high compared with other years during which N(NO) was less than 0 ‰. These higher N(NO) values may be related to biomass burning associated with forest fires ( ‰ to ‰; Chai et al., 2019). Stratospheric NO inputs may also have high N values, as observed in Antarctic aerosols (Savarino et al., 2007). However, the O(NO) values in the summer of 1987 (28.2 ‰) and the spring of 1992 (29.5 ‰) were not clearly different from other years, which is not consistent with the high O(NO) trend during preindustrial biomass burning (i.e., forest fires) reported in the previous study (Alexander et al., 2004). We also note that no biomass-burning tracers were detected in 1987 or 1992 (Parvin et al., 2019). Further research is therefore required to link nitrate isotopes with specific events such as biomass burning.
The O(NO) record from the SE-Dome core did not exhibit clear trends over the past 60 years. During this period, changes in atmospheric oxidants have occurred, such as increases in tropospheric O3 over Arctic regions (Law et al., 2023). It is reasonable to estimate that higher O3 can induce increases in O(NO) by (1) promoting NO2 formation from NO O3 reactions and (2) promoting NO2 oxidation to NO3 (and subsequently to HNO3) by O3. However, such changes were not recorded in the O(NO) data from the SE-Dome ice core. During this period, atmospheric sulfate formation was changed by the promotion of in-cloud S(IV) O3 reactions, based on increases in O(SO) from the same SE-Dome ice core (Hattori et al., 2021). Thus, further research is required to determine the mechanism(s) behind the observed constant O(NO) values in ice cores after emission controls by comparing O(NO) values estimated using chemical transport models such as GEOS-Chem (Alexander et al., 2009, 2020). Based on the O(NO) values recorded in the GISP2 ice core, the variations have been attributed to the intricate BrONO2 hydrolysis mechanism, which extends beyond the small fluctuations in the ) ratio in the relatively colder climate of a glacial period (Geng et al., 2017). Thus, reactive-halogen chemistry may also be a factor that impacts changes in the atmospheric oxidation capacity, specifically in high-latitude regions in the Northern Hemisphere.
4.4 Comparison with other ice core data
The N(NO) value of ice cores collected in Lomonosovfonna, Svalbard, was ‰ after 1950 (Vega et al., 2015), which is lower than that at the SE-Dome and Summit sites. Given that snow accumulation at Lomonosovfonna ( m w.e. a−1; Vega et al., 2015) was higher than at Summit, these low N(NO) values may reflect less postdepositional alteration. Notably, the relatively low N(NO) values at Lomonosovfonna and the SE-Dome site were consistent with the low N(NO) values in aerosols observed at two Arctic stations (Morin et al., 2008, 2012). Notably, the SE-Dome site and Svalbard both had lower N(NO) values and higher accumulation rates than the Summit site. Additionally, Svalbard is closer to Europe than Greenland, which may indicate a regional source difference. Ice core N(NO) records reported from Lomonosovfonna also exhibited decreasing trends until the 1970s, whereas an increase in N(NO) was only observed at Lomonosovfonna after the 1990s (Vega et al., 2015). Such differences may be attributed to differences in NOx sources and spatial chemistries in the Arctic, although it is unclear whether this difference was caused by anthropogenic sources, natural sources, transport, or a combination of these factors.
The N(NO) records in ice cores from the Tibetan Plateau also exhibit decreasing trends from 1955 to 2011 (Li et al., 2020). The N(NO) of this Tibetan Plateau ice core ( ‰ in 1951–2011) is also substantially higher than those of Arctic ice cores, indicating a different regional context. Comparing N(NO) records from different locations would be beneficial for determining the regional physical or chemical behaviors of NO from emission to deposition. This would allow us to better assess the impacts of human activity on nitrogen cycling and take corresponding measures to reduce the adverse effects of NO on climate and biogeochemical cycles. However, in Antarctica (Akers et al., 2022), the ice core N(NO) values varied significantly depending on the snow accumulation rate. It is therefore important to estimate postdepositional alteration for each ice core, refine models with recent information (Shi et al., 2023), and perform reverse calculations for atmospheric N(NO) (Jiang et al., 2024). The TRANSITS model used in this study is effective for assessing the sensitivity of nitrate preservation and isotopic compositions to snow accumulation within the same environment. However, as discussed, the archived isotopic composition depends on the value, which varies spatiotemporally. Therefore, efforts should be made to update models that account for the photolysis of snow NO and the recycling and redistribution of reactive nitrogen in boundary layer chemistry within a global chemical transport model (Zatko et al., 2016) by incorporating N information. Additionally, a recent study emphasized the potential impact of microbial alterations on both NO concentrations and its isotopic compositions in an Asian glacier (Hattori et al., 2023); thus, interpretations of NO concentrations and N(NO) records in ice cores should proceed with caution. We recommend that interpretations of NO concentrations and N(NO) records in ice cores should be accompanied by O(NO) or O(NO) records whenever possible to verify atmospheric NO preservation without postdepositional biological alteration.
5 Conclusions
In this study, we reported -year (1959–2014) records of NO isotopic compositions from the SE-Dome ice core in Greenland. The observed N(NO) values in the SE-Dome ice core were consistently ‰ lower than those in the Summit ice core record. The high snow accumulation rate at the SE-Dome site reduces the sensitivity of NO to postdepositional processes, which was supported by outputs from the TRANSITS model. Therefore, we concluded that the SE-Dome ice core, which exhibits superior NO preservation, is a promising tool for reconstructing changes in atmospheric nitrogen cycling driven by anthropogenic activity. This study was based on results from the SE-Dome I ice core ( m), which covers the past 60 years. The SE-Dome II core (drilled in 2021) preserves records that extend back to 1800 CE (Iizuka et al., 2021; Kawakami et al., 2023). Thus, there is considerable potential for future research aimed at reconstructing NO aerosol dynamics from the beginning of the industrial revolution to the present. Additionally, while regional comparisons of ice core N(NO) records are beneficial for describing the regional physicochemical behaviors of NO, it is necessary to account for regional differences in postdepositional alteration when analyzing the spatiotemporal variations in atmospheric NO isotopes.
Data availability
The data used in this study are available from
The supplement related to this article is available online at
Author contributions
SH conceptualized the study; ZW, SH, AT, SI, ZJ, SM, and YI curated the data; ZW, SH, AT, ZJ, SI, and YI performed the formal analysis; SH and YI acquired funding; SH, AT, NY, KF, SI, SM, and YI conducted the investigation; SH, ZJ, LG, and JS developed the methodology; SH managed the project; ZW and SH validated the results; ZW, KF, SH, and ZJ visualized the data; ZW and SH wrote the original draft; and KF, SI, ZJ, LG, JS, RU, AL, and YI reviewed and edited the manuscript.
Competing interests
The contact author has declared that none of the authors has any competing interests.
Disclaimer
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors.
Acknowledgements
The authors gratefully acknowledge the NOAA Air Resources Laboratory (ARL) for providing the HYSPLIT transport and dispersion model (Stein et al., 2015). We also wish to acknowledge Jihong Cole-Dai for providing data on the annual snow accumulation rate at the Summit site.
Financial support
This study was supported by the National Natural Science Foundation of China (grant nos. 42484851, W2441015, and 42473011); the Fundamental Research Funds for the Central Universities, including the General Program (grant no. 0206/14380918), the International Collaboration Program (grant no. 0206/14380225), the Cemac “GeoX” Interdisciplinary Program (grant no. 0206/14380205), and additional grants (grant nos. 0206/14380232, 0206/14380204, 0206/14380150, 0206/14380185, and 0206/14380174); and start-up funding from Nanjing University. This study was also supported by MEXT/JSPS KAKENHI (grant nos. 18H05292, 23H00511, and 23K18516) and the Arctic Challenge for Sustainability (ArCS II) project (grant no. JPMXD1420318865). This study was performed under a cooperative research program of the Institute of Nature and Environmental Technology, Kanazawa University (project no. 22042, 23039). Shohei Hattori was supported by MEXT/JSPS KAKENHI (grant no. 20H04305), Lei Geng was supported by the National Natural Science Foundation of China (grant no. W2411030), and Zhuang Jiang received support from the Fundamental Research Funds for the Central Universities.
Review statement
This paper was edited by Eliza Harris and reviewed by Matthew Johnson and one anonymous referee.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
© 2025. This work is published under https://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Abstract
Stable isotopes of atmospheric nitrate (NO
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
Details









1 International Center for Isotope Effects Research, State Key Laboratory of Critical Earth Material Cycling and Mineral Deposits, Nanjing University, Nanjing 210023, China
2 International Center for Isotope Effects Research, State Key Laboratory of Critical Earth Material Cycling and Mineral Deposits, Nanjing University, Nanjing 210023, China; Frontiers Science Center for Critical Earth Material Cycling, School of Earth Sciences and Engineering, Nanjing University, Nanjing 210023, China; Department of Chemical Science and Engineering, School of Materials and Chemical Technology, Tokyo Institute of Technology, Kanagawa 226-8502, Japan
3 Department of Chemical Science and Engineering, School of Materials and Chemical Technology, Tokyo Institute of Technology, Kanagawa 226-8502, Japan
4 National Key Laboratory of Deep Space Exploration, School of Earth and Space Sciences, University of Science and Technology of China, Hefei, 230026, Anhui, China
5 Institute of Nature and Environmental Technology, Kanazawa University, Kakuma-machi, Kanazawa, Ishikawa 920-1192, Japan
6 Graduate School of Environmental Studies, Nagoya University, Nagoya 464-8601, Japan
7 Institute of Low Temperature Science, Hokkaido University, Sapporo 060-0819, Japan
8 University Grenoble Alpes, CNRS, IRD, Grenoble INP, INRAE, IGE, 38000 Grenoble, France
9 Department of Chemical Science and Engineering, School of Materials and Chemical Technology, Tokyo Institute of Technology, Kanagawa 226-8502, Japan; National Institute of Information and Communications Technology, Koganei, Tokyo 184-8795, Japan; Earth-Life Science Institute, Institute of Science Tokyo, Tokyo 152-8550, Japan
10 University Grenoble Alpes, CNRS, IRD, Grenoble INP, INRAE, IGE, 38000 Grenoble, France; Aix Marseille University, CNRS, IRD, INRAE, CEREGE, Aix-en-Provence, France