Emissions of atmospherically reactive gases nitrous acid and nitric oxide from Arctic permafrost peatlands

Soils are important sources of nitric oxide (NO) and nitrous acid (HONO) in the atmosphere. These nitrogen (N)-containing gases play a crucial role in atmospheric chemistry and climate at different scales because of reactions modulated by NO and hydroxyl radicals (OH), which are formed via HONO photolysis. Northern permafrost soils have so far remained unexplored for HONO and NO emissions despite their high N stocks, capacity to emit nitrous oxide (N2O), and enhancing mineral N turnover due to warming and permafrost thawing. Here, we report the first HONO and NO emissions from high-latitude soils based on measurements of permafrost-affected subarctic peatlands. We show large HONO (0.1–2.4 µg N m−2h−1) and NO (0.4–59.3 µg N m−2h−1) emissions from unvegetated peat surfaces, rich with mineral N, compared to low emissions (⩽0.2 µg N m−2h−1 for both gases) from adjacent vegetated surfaces (experiments with intact peat cores). We observed HONO production under highly variable soil moisture conditions from dry to wet. However, based on complementary slurry experiments, HONO production was strongly favored by high soil moisture and anoxic conditions. We suggest urgent examination of other Arctic landscapes for HONO and NO emissions to better constrain the role of these reactive N gases in Arctic atmospheric chemistry.


Introduction
Traditionally, high-latitude soils have been considered negligible sources of nitrogenous gases to the atmosphere. Cold and often wet soil conditions slow down the decomposition and mineralization processes (Robinson 2002), thus limiting the availability of mineral nitrogen (N) for microbial pathways producing N-containing gases. However, a recent synthesis shows that nitrous oxide (N 2 O) emissions commonly occur in permafrost-affected soils and can reach high rates in soils with high N availability, for example, as a consequence of disturbed or completely lacking vegetation cover or input of additional N from permafrost thaw (Voigt et al 2020). These observations suggest that, in addition to N 2 O release, other N-containing gases might occur from permafrostaffected soils, but they have never been studied.
Particularly important in this context are nitrous acid (HONO) and nitric oxide (NO), atmospherically reactive gases which are emitted from soils (Kubota and Asami 1985, Galbally et al 1987, Su et al 2011, Oswald et al 2013 (figure 1). The photolysis of HONO generates the hydroxyl radical (OH), a key atmospheric oxidant that initiates the removal of pollutants, such as carbon monoxide (CO) and the potent greenhouse gas methane (CH 4 ) (Lelieveld et al 2004). OH participates in the formation of secondary organic aerosols by oxidizing volatile organic compounds (VOCs), which further trigger cloud formation (Claeys 2004). Nitric oxide, on the other hand, regulates the ozone (O 3 ) cycle in the atmosphere as Figure 1. Nitrogen transforming processes release nitrous acid (HONO), nitric oxide (NO), nitrous oxide (N2O), and dinitrogen (N2) from soil to the atmosphere. HONO, NO, and N2O is produced in aerobic ammonia (NH4 + ) oxidation (blue arrows) by chemolithotrophic bacteria, archaea, and comammox organisms (complete ammonia oxidation to nitrate) (Oswald et al 2013, Ermel et al 2018, Han et al 2021. Ammonia oxidation can also support HONO formation via abiotic reactions between hydroxylamine (NH2OH) and water (Ermel et al 2018). The denitrification process (turquoise arrows) contributes to soil NO and N2O emissions during the sequential reduction of nitrate (NO3 − ) under anaerobic conditions (Knowles 1982). Soil nitrite (NO2 − ) produced via nitrification, denitrification, and dissimilatory nitrate reduction to ammonium (DNRA, red arrows) can be converted chemically in the presence of reduced metals (e.g. Fe +2 ) and organic substances (e.g. humic acid) into gaseous NO and N2O in the process called chemo-denitrification (green arrow) (Coby and Picardal 2005). Nitrite can also react chemically with proton (H + ) to form HONO and NO (Cleemput and Samater 1995, Su et al 2011, Bhattarai et al 2021. The italic terms above the colored arrows indicating nitrogen transforming processes denote the genes encoding enzymes carrying out the respective processes. The amo is ammonia monooxygenase oxidizing NH4 + to NH2OH, hao is hydroxylamine oxidoreductase oxidizing NH2OH to NO2 − , nxr is nitrite oxidoreductase oxidizing NO2 − to NO3 − , nap and nar are periplasmic and cytoplasmic enzymes respectively reducing NO3 − to NO2 − , nrf is formate-dependent dissimilatory nitrite reductase converting NO2 − to NH4 + , nirK, and nirS are copper and cytochrome-cd1 based enzymes respectively reducing NO2 − to NO, nor is heme or heme-copper based enzyme reducing NO to N2O, and nosZ is nitrous oxide reductase reducing N2O to atmospheric N2. well as the formation of another atmospheric oxidant, the nitrate (NO 3 − ) radical (Atkinson 2000). Therefore, HONO and NO have a crucial role in atmospheric chemistry and climate, and understanding their emission sources is paramount.
Numerous studies have reported substantial HONO and NO emissions from agricultural soils (Davidson and Kingerlee 1997, Stehfest and Bouwman 2006, Su et al 2011, Maljanen et al 2013, Oswald et al 2013, Scharko et al 2015, Wu et al 2019, Bhattarai et al 2021, where fertilization and other management practices lead to the high availability of mineral N. The global emissions of NO x (NO + NO 2 ) from agricultural soils (3.7 Tg N yr −1 ) are dominated by NO emissions (Ciais et al 2013). Agricultural soils are also globally significant sources of the potent greenhouse gas N 2 O (3.8 Tg N yr −1 ) (Tian et al 2020), suggesting that the production mechanisms of HONO, NO, and N 2 O are interlinked and related to high N turnover rates in soil (Maljanen et al 2013, Wu et al 2019. Indeed, studies have clearly shown the role of microbial N-transformation pathways, most importantly nitrification and denitrification, behind N 2 O, HONO, and NO production (Oswald et al 2013, Scharko et al 2015, Ermel et al 2018, Wu et al 2019, Bhattarai et al 2021, Han et al 2021 (figure 1). In addition to reports from agricultural soils, a few studies have reported HONO and NO emissions from soils under natural vegetation, such as biocrust (Weber et al 2015) and forest soils (Mushinski et al 2019). However, pristine Arctic soils have so far remained entirely unexplored concerning HONO (Bhattarai et al 2021) and NO (Davidson and Kingerlee 1997) emissions, despite their well-documented capability to emit N 2 O and extensive soil N stocks, which are becoming increasingly vulnerable to mobilization (Voigt et al 2020) as a consequence of warming and permafrost thaw (Meredith et al 2019).
Some of the highest N 2 O emissions from permafrost-affected soils have been found in peatlands (Voigt et al 2020). Permafrost-affected peatlands cover up to 46% (1.7 × 10 6 km 2 ) of the Northern Hemisphere peatland and are important storehouses of soil organic carbon (185 Pg C) and N (10 Pg N) (Hugelius et al 2020). Permafrost-affected peatlands in the Arctic region have surfaces uplifted by permafrost, the so-called palsas and peat plateaus (Seppälä 2011, figure 2). These are common geographical features occurring in discontinuous and sporadic permafrost zone (Sannel andKuhry 2011, Borge et al 2017). They are formed by permafrost aggradation, which lifts the peat surface, leading to drier conditions than the surrounding unfrozen peatland surface (Seppälä 2003). As a result of wind abrasion, parts of the palsas and peat plateaus lack A detailed explanation about sampling sites is presented in the material and methods, and a close view of the studied peat surfaces is provided in figure S1. Photo credit: 1. First author, 2. Richard E Lamprecht, and 3. Maija E Marushchak. vegetation (Seppälä 2003). These unvegetated (bare) peat surfaces, due to well-drained conditions and lack of competition for N between plant and soil microbes, possess high total N content, high mineral N content, and low C/N ratio (Marushchak et al 2011, Kaverin et al 2016. Due to the high availability of mineral N, unvegetated peat surfaces on permafrost peatlands are known to emit substantial amounts of N 2 O in-situ (Repo et al 2009, Marushchak et al 2011, Siljanen et al 2019, and N 2 O emissions have been shown to increase with permafrost thaw and climate warming (Voigt et al 2017a(Voigt et al , 2017b. Such high mineral N content and high N 2 O emission rates from permafrost-affected peatlands indicate a high potential for HONO and NO emissions, as these N-containing gases are an integral part of the soil N cycle (figure 1). Here, we hypothesized that Arctic permafrost peatlands are sources of HONO and NO, as they are of N 2 O, with higher emissions from unvegetated compared to vegetated surfaces due to better mineral N availability for the microbial N cycle because of lack of competition for mineral N between plant and soil microbes. To address our hypothesis, we assessed HONO and NO fluxes from intact peat cores from permafrost peatlands collected from Seida (NW-Russia), Kevo, and Kilpisjärvi (N-Finland) under laboratory conditions and N 2 O fluxes in-situ from all sampling sites except that in Kilpisjärvi. Furthermore, we conducted complementary slurry experiments to explore the effect of soil moisture on HONO and N 2 O production. We also assessed the changes in HONO and NO emissions induced by permafrost thawing by measuring intact cores sampled from collapsing palsa edges.

Site description
To test our hypothesis, we collected intact peat cores (0-0.1 m depth) from unvegetated and vegetated palsa and peat plateau surfaces (figures 2 and S1 available online at stacks.iop.org/ERL/17/024034/mmedia) in three subarctic peatlands in northern Finland and northwestern Russia (figure 2). We collected only the surface peat layer for our study because these layers connect the peatlands with the atmosphere, thus playing a vital role in the exchange (emissions or uptake) of gases between the atmosphere and peatland. Of the three studied subarctic permafrost peatlands, two were palsa mires in the sporadic permafrost zone in northern Finland: Peera palsa located in Kilpisjärvi (68.88 • N, 21.05 • E, 496.5 m a.s.l.) and Vaisjäeggi palsa mire located in Kevo (69.82 • N,27.17 • E, 295 m a.s.l.); and the third was the Seida peat plateau (67 • 030 N, 62 • 550 E, 100 m a.s.l.) with discontinuous permafrost located in northwestern Russia. The mean annual temperature and precipitation in Kilpisjärvi, Kevo, and Seida are −1.9 • C and 487 mm, −1.3 • C and 433 mm, and −5.6 • C and 501 mm, respectively (Repo et al 2009, Pirinen 2012. Additional details are provided in the supplementary material.

Peat core sampling and transportation
Intact peat cores were sampled from unvegetated (bare) peatland surfaces and adjacent vegetated surfaces (figure S1). We collected 14 (five vegetated and nine unvegetated), 19 (12 vegetated and seven unvegetated), and six (three vegetated and three unvegetated) intact peat cores in Kilpisjärvi, Kevo, and Seida, respectively. The vegetated cores included ground vegetation and small vascular plants. The dominant vegetation species were Rubus chamaemorus L. and Empetrum nigrum subsp. hermaphroditum in Kilpisjärvi, E. nigrum subsp. hermaphroditum and R. chamaemorus L. in Kevo, and mosses, Sphagnum Dicranum sp. and lichen, Cladina sp. in Seida. Six had only lichens but no vascular plants among the 12 vegetated peat cores from Kevo (figure S1). Among the nine unvegetated peat cores from Kilpisjärvi, four were collected from collapse scars, which are common microtopographic features at this site (figures 2 and S1). In Kilpisjärvi and Kevo, intact peat cores were sampled using polyvinylchloride (PVC) rings (h = 0.12 m, Ø = 0.1 m). A circular boundary (exact size as PVC area) was cut to the peat surface to a depth slightly deeper than the PVC ring by a sterilized knife (70 vol.% alcohol) during the peat core sampling. The cylindrical peat core was then extracted and inserted into the PVC core, after which extra depth (>0.1 m) was removed. This step was done to avoid peat compaction and to maintain the in-situ bulk density. The peat cores from Kilpisjärvi and Kevo were transported in cold conditions (+7 • C-8 • C) to the laboratory at the University of Eastern Finland (UEF) and further stored at +4 • C before the measurements. In Seida, surface peat samples with dimensions of 0.1 × 0.05 m (width × depth) were sampled with a sterilized knife, frozen within 24 h after sampling, and transported in zipper bags to UEF, Finland, where they were stored at −18 • C ± 2 • C until the analyses. The samples from the remote Seida study site were transported in frozen condition since that was the only way to ensure that the sample stayed cold during the shipment, which took several days.

Flux measurement of HONO, NO, and N 2 O and soil analyses
Before the HONO, NO, and N 2 O flux measurements, we pre-incubated the peat cores at +13 • C for 4 d to acclimatize the soils to near in-situ summer conditions. After unfreezing and inserting into the PVC rings described above, the Seida peat cores were stored at +4 • C for 9 d before the pre-incubation at +13 • C. An incubation temperature of +13 • C was selected because it was close to the ambient air temperatures during the sampling (Kilpisjärvi = 12.0 • C, Kevo = 15.4 • C, Seida = 13.3 • C). Moisture loss from the peat cores during the incubation period was compensated daily by adding Milli-Q water, thus maintaining the moisture content (table S1) close to the in-situ condition.
The HONO and NO fluxes from the intact peat cores were measured in the laboratory at room temperature (21 • C) in the dark using a dynamic chamber (made from Teflon, Polytetrafluoroethylene, (PTFE)), V = 0.034 m 3 ) system . This method represents the current state-of-the-art HONO and NO measurements due to the major technical challenges associated with in-situ HONO and NO measurements, particularly in remote areas such as the Arctic. The N 2 O fluxes were measured using a static chamber system in the laboratory  and in-situ (Repo et al 2009, Siljanen et al 2019 conditions. After the gas flux measurements, intact peat cores were homogenized and processed for peat physiochemical properties (table S1). More details are provided in the supplementary material.

HONO production under oxic and anoxic slurry conditions
In addition to in-situ moisture conditions, we further explored the role of high moisture (∼100% water holding capacity (WHC)) on HONO emissions in a soil incubation experiment under oxic and anoxic conditions. The effect of moisture on HONO production was tested with unvegetated Seida peat because it had the highest HONO and NO emissions under the in-situ moisture content, indicating an efficient mineral N cycling associated with HONO and NO production. Here, we assessed only HONO (not NO) as influenced by moisture because gaseous N loss via HONO was 25 times less (µg HONO-N vs µg NO-N) compared to NO from the intact cores with low moisture content (gravimetric water content (GWC), 1.7 g H 2 O g −1 , ∼37% water filled pore space (WFPS), table S1) and we wanted to assess if an increase in moisture would increase HONO emissions or not. Additionally, we investigated the role of abiotic (chemical) HONO production by repeating the slurry experiment with Seida peat sterilized by autoclaving to stop any microbial activities. To relate the HONO production with N 2 O production in peat slurries, we also followed the N 2 O production over the same measurement period.
More details are provided in the supplementary material.

Statistics
Unless otherwise specified, statistical analysis and graphical presentations were done using R statistical software (R version 4.0.5), except for figure 1 and the conceptual model in figure 4, which were created with PowerPoint and a graphic software Inkscape (version 1.0.1), respectively. Before statistical testing, data distribution was assessed using the Shapiro-Wilk test, Q-Q plots, and histograms. Non-normally distributed data were transformed (log 10 or square root) when required to achieve a normal distribution. We used one-way analysis of variance followed by Bonferroni's post hoc test on normally distributed data, whereas a non-parametric Kruskal-Wallis test followed by Dunn's post hoc test on data showing a nonnormal distribution. More details are provided in the supplementary material.

Magnitude and drivers of HONO and NO emissions
We found HONO and NO emissions from the studied permafrost peatlands ( figure 3(a)). Emission rates varied considerably between (Seida > Kevo > Kilpisjärvi) and within (unvegetated > vegetated) the study sites ( figure 3(a)), which is typical for N cycling processes and N gas emissions from soils in general (Butterbach-Bahl et al 2013). As hypothesized, the unvegetated peat surfaces had a higher mineral N content ( figure 3(b)), and they emitted HONO (between 0.1 and 2.4 µg N m −2 h −1 ) and NO (between 0.4 and 59.3 µg N m −2 h −1 ) at higher rates compared to the vegetated surfaces (HONO, χ 2 (1) = 7.391, P = 0.0065 and NO, χ 2 (1) = 10.542, P = 0.0011, figure 3(a), table S1). The low HONO and NO emissions in the vegetated surfaces (⩽0.2 µg N m −2 h −1 ; figure 3(a), table S1) were associated with lower mineral N content than in the unvegetated surfaces (F(1,37) = 5.47, P = 0.0248). The HONO emissions measured from permafrost peatlands are smaller than the global median (18 µg N m −2 h −1 ) (Bhattarai et al 2021), but within the range reported for boreal peatlands drained for agriculture (1.7-6 µg HONO-N µg m −2 h −1 ) (Maljanen et al 2013). Notably, the mean NO emission from the unvegetated peat surfaces of the highest emitting Seida site (59.3 µg N m −2 h −1 ; table S1) is almost two-fold the global mean NO emission (30.8 µg N m −2 h −1 ) (Davidson and Kingerlee 1997) and more than two-fold the mean N 2 O emission (24.8 µg N m −2 h −1 ) from peatlands across the permafrost region (Voigt et al 2020), highlighting the importance of this newly revealed N gas emissions from Arctic permafrost peatlands.
The HONO and NO emissions in the studied palsa and peat plateau surfaces were strongly interlinked (r s = 0.84, P < 0.05, figure S2), implying common production pathways with similar environmental regulatory factors like N availability (figure 1). Both HONO and NO emissions correlated negatively with the C:N ratio (r s = −0.41 for HONO and r s = −0.47 for NO, P < 0.05, figure S2) and positively with the total mineral N content (NH 4 + + NO 3 − , r s = 0.74 for HONO and r s = 0.84 for NO, P < 0.05, figure S2). These correlations reflect the tight linkage between N gas production and N mineralization in these pristine soils where, in the absence of external N inputs (low N deposition (Bobbink et al 2010) and biological N 2 fixation (Stewart et al 2014)), mineralization of organic matter is a dominant source of mineral N needed in microbial HONO and NO production. The association of low C:N ratios with high N mineralization rates (Liu et al 2017)

and N 2 O emissions (Klemedtsson et al 2005) has been demonstrated in previous studies.
From the two mineral N species determined, NO 3 − exhibited particularly high positive correlations with HONO and NO emissions (r s = 0.83 and r s = 0.97, respectively, P < 0.05, figure S2), while the correlations were weaker with NH 4 + (r s = 0.35 for both HONO and NO, P < 0.05, figure S2). The strong correlations between HONO and NO emissions and NO 3 − suggest that nitrification is of great importance for HONO and NO production, either as a direct pathway of HONO and NO production (Oswald et  Based on the measurements from the peat cores, it is not possible to draw solid conclusions about the contribution of nitrification and denitrification to HONO and NO production or the effect of soil moisture on the production pathways. Soil moisture content in the unvegetated peat surfaces was highly variable (GWC 1.7-2.7 g H 2 O g −1 , table S1). The HONO and NO emissions tended to be higher in drier soils (r s = −0.42 for HONO and r s = −0.46 for NO, P < 0.05, figures 3(a), S2 and table S1). This implies the high importance of nitrification (see above), which is an aerobic process requiring welldrained oxic soil. However, in our slurry incubation experiment with unvegetated Seida peat, we observed a significantly higher HONO (χ 2 (1) = 12.403, P = 0.0004; figures 3(c), S4 and table S2) and N 2 O (χ 2 (1) = 4.23, P = 0.0376, figure S4) production under anoxic than oxic conditions. This points towards an important role of the reductive microbial Kilpisjärvi-n = 5 biologically independent samples for vegetated, n = 9 for unvegetated; Kevo-n = 12 for vegetated, n = 7 for unvegetated; Seida-n = 3. In (c), n = 3 laboratory replicates. Note: in (a), the scale on the y-axis is different for HONO and NO, and in (c), the y-axis has a log10 scale. Asterisks show statistical significances between the peat surfaces in (a) and (b). The levels of significance are: * * * P < 0.001, * * P < 0.01, * P < 0.05, and n.s. denotes statistically non-significant difference. The lowercase letters in (c) represent statistically significant differences (P < 0.05) in HONO emissions between the sampling days in oxic and anoxic conditions. pathways (Wu et al 2019), such as denitrification, in the HONO production (Bhattarai et al 2021).
The results of the slurry experiment (⩾100% WHC), where anoxic conditions promoted HONO production (mean emission rates 66-254 µg N m −2 h −1 and 3-10 µg N m −2 h −1 in anoxic and oxic conditions, respectively, figure 3(c) and table S2) may seem to conflict with the results from the peat core measurements where the highest HONO emissions (2.4 µg N m −2 h −1 , figure 3(a)) were observed under low soil moisture content (table S1). However, this kind of non-linear moisture dependence can be easily explained by the multiple pathways contributing to HONO production in soils and their distinct moisture dependencies (figure 1). The double peak of HONO and NO production in soils as a factor of soil moisture, one at a low soil moisture content (20%-30% WHC) and another at a high soil moisture content (∼100% WHC), has been reported previously (Wu et al 2019). The important implication of the wide range of soil moisture contents supporting HONO and NO production in permafrost peatlands is that these gases can be produced in variable environmental conditions, with high mineral N availability as a key prerequisite (figures 3 and S2).

Lack of association with N 2 O emissions and importance of the whole peat profile for N gas production
The HONO and NO emitting unvegetated Seida cores were also emitting N 2 O (0.36 µg N m −2 h −1 , table S1) under laboratory conditions, as was expected based on previous studies which have reported a close link between these nitrogenous gases (Maljanen et al 2013. Interestingly, the same was not true for the unvegetated peat cores from Kevo: these cores showed net uptake of atmospheric N 2 O despite their high HONO and NO emissions. The N 2 O uptake was even higher in the vegetated cores ( figure S3). This N 2 O uptake is in contrast with the substantial N 2 O emissions from unvegetated permafrost peatland surfaces measured in-situ in this study (Kevo, figure S3), as well as in previous studies (Marushchak et al 2011, Gil et al 2017, Voigt et al 2017a, 2017b). This discrepancy between N 2 O emissions in the field and the laboratory may be because the latter measurements included only the top layer (0.1 m) of the peat core. The moisture content in the cores (median WFPS 53%) was mostly below the optimum (WFPS 65%-100%) for the N 2 O emissions (Voigt et al 2017b). Indeed, in previous studies on such permafrost peatlands, the highest soil pore gas N 2 O concentrations have been observed in the middle of the seasonally thawing active layer at ∼30-55 cm, indicating maximum N 2 O production in the deeper and moist peat layers (Gil et al 2017, Voigt et al 2017a, 2017b. Also, the lower N 2 O emissions in laboratory conditions compared to the in-situ from Seida (figure S3) can be explained by the importance of the deeper peat layers for N 2 O production. Although the consumption of atmospheric N 2 O usually occurs in water-logged soils (Hallin et al 2018), this process has also been found in dry soils under oxic conditions (Wu et al 2013, Siljanen et al 2020, including high-arctic soils (Brummell et al 2014, Wagner et al 2019. Furthermore, the nosZ gene encoding the N 2 O reductase enzyme, which carries out N 2 O reduction to N 2 , has been detected in the surface layer of uplifted permafrost peatlands, similar to those studied here (Palmer andHorn 2012, Hetz andHorn 2021). Usually, N 2 O reduction occurs when the energetically more favorable electron acceptor NO 3 − is not available. In agreement with this, the unvegetated Kevo peat with net N 2 O uptake had a much lower NO 3 − content (9.9 µg N g −1 , 16.8% of all mineral N) than the N 2 O emitting Seida peat (178.9 µg N g −1 , 74.1% of all mineral N) ( figure 3(b)). Under in-situ conditions, a part of N 2 O diffusing from the deeper layers may be consumed in the surface layer by organisms capable of N 2 O reduction. The same microbes may switch to using atmospheric N 2 O when disconnected from the deeper peat profile.
So, how can we explain continuous HONO and NO emissions despite the observed N 2 O uptake in the unvegetated peat cores from Kevo (figures 3(a) and S3)? From the possible HONO and NO precursors (figure 1), NO 2 − is the first product to be formed in both oxidative (i.e. nitrification) and reductive (i.e. denitrification or DNRA) microbial pathways of the soil N cycle. This microbially produced NO 2 − can be lost as gaseous HONO and NO (a primary precursor for N 2 O) via various biotic and abiotic reactions (figure 1) before it is used in, for example, the reduction steps towards N 2 O production in denitrification. Thus, substantial HONO and NO production can consume the mineral N species (NO 3 − and NO 2 − ), which serve as N 2 O precursors, thereby effectively reducing the potential for N 2 O production (figure 1). Indeed, continuous NO emissions were observed while N 2 O flux ranged from zero to N 2 O uptake in a grassland soil (Werner et al 2014), which is a phenomenon that could also be occurring for HONO.

Emissions of HONO and NO from Arctic soils at present and in future
Our results suggest that HONO and NO emissions can have importance for the total loss of nitrogenous gases from permafrost-affected peatlands, adding significantly to the previously discovered gaseous N release in the form of N 2 O emissions (Voigt et al 2020). Because of the tight linkage between mineral N availability (figure S2), we have good reasons to believe that the HONO and NO emissions from Arctic peatlands will increase with warming and the associated increase in N mineralization (figure 4). This will have several consequences to atmospheric chemistry (figure 4), including the cooling effects associated with oxidation of the potent greenhouse gas CH 4 and secondary aerosol formation and further cloud formation, also associated with a cooling effect (Emmons et al 1997, Atkinson 2000, Schmale et al 2021. Indeed, in an in-situ warming experiment at the Seida study site, warmer temperatures led to enhanced N 2 O emissions from unvegetated peat surfaces and triggered N 2 O emissions from vegetated peat that had negligible N 2 O emissions under ambient temperatures (Voigt et al 2017a). The increase in N 2 O emissions reflects enhanced N turnover and might well be associated with increased HONO and NO emissions, but this is yet to be proven in future studies.
Furthermore, permafrost thaw may also promote HONO and NO release in Arctic permafrost peatlands through the input of other mineral N together with moisture changes induced by ground collapse associated with the thaw. We, indeed, observed higher NO emissions (figure S5) associated with higher moisture content from collapsing palsa surfaces than from stable ones without signs of permafrost degradation. It has to be noted that since HONO (figures 3(c), S4 and table S2) and NO  figure 1 for detail) of mineral nitrogen (N), such as ammonia (NH4 + ) oxidation (dark red color) and nitrate (NO3 − ) reduction (turquoise color). The emissions of HONO and NO from both unvegetated and vegetated surfaces may increase with climate warming, causing enhanced N mineralization from soil organic matter (SOM), as well as with climate change-induced disturbances such as permafrost-thaw, thermokarst, extreme weather events, and wildfires (Voigt et al 2020). Upon photolysis (dotted yellow arrow), hydroxyl radicals (OH) and nitric oxide (NO) are produced from HONO, which together with NO participate in various atmospheric reactions (Atkinson 2000) (light blue oval), thus influencing the chemical balance of the Arctic atmosphere. CO-carbon monoxide, VOCs-volatile organic compounds, SOA-secondary organic aerosols, O3-ozone, and PAN-peroxyacetyl nitrate in the light blue oval. The longer arrows for HONO and NO compared to those for N2O are used to highlight their relevance with regards to atmospheric reactions and do not represent the magnitude of the emissions of the different gases. (Wu et al 2019) production is enhanced by wetter soil conditions, occurring in the deeper peat layers, the emissions observed in our peat core measurements are likely lower than those occurring in-situ. This should be tested by future studies with deeper peat profiles or, preferably, by in-situ measurements when it will become technically feasible in places like the Arctic.

Conclusions
We have shown that permafrost peatlands in remote Arctic regions emit atmospherically important nitrogenous gases, HONO and NO. In urban areas, the production of NO x (NO 2 + NO) (Delmas et al 1997) and HONO (NO 2 + organic soot particles) (Stemmler et al 2006) are primarily related to fuel combustion. Still, this emission source is very limited in remote Arctic regions with minimal anthropogenic influence. Under these circumstances, emissions of OH precursors, such as HONO and NO, from Arctic soils can be crucial for atmospheric chemistry, especially when the same ecosystems are known to emit substantial amounts of VOCs (Kramshøj et al 2016, Seco et al 2020. Because OH, NO and VOCs jointly participate in atmospheric reactions (Atkinson 2000) (figure 4), their frequent availability in the Arctic atmosphere likely plays a crucial role in the O 3 budget and for aerosol-climate interactions, which can be of global significance (Schmale et al 2021). However, our current knowledge about OH, NO, and VOCs sources and sinks in the Arctic region remains poorly understood (Emmons et al 1997, Schmale et al 2021. In addition, our observations of substantial NO emissions from permafrost peatlands suggest that the current estimate of the contribution of natural soils to global NO x emissions (15%), where Arctic soils are not considered yet, might be an underestimate (Ciais et al 2013). To better understand the Arctic aerosol-climate interactions and the regional O 3 and the global NO x budget, we stress the importance of studying the HONO and NO emissions more thoroughly from Arctic peatlands and mineral soils not studied here. Only when a solid understanding of the magnitude of HONO and NO emissions and their regulatory factors (e.g. dry-wet cycles) have been established, it will become feasible to make predictions of how their emissions from the Arctic will change with warming and how they will affect the atmospheric chemistry and climate (figure 4). Therefore, we conclude that a better assessment of diverse Arctic landscapes for HONO and NO emissions are urgently needed.

Data availability statement
The data that support the findings of this study are available upon reasonable request from the authors and also from the link, https://doi.org/10.5281/zenodo.5864117.