Ship-Borne Observations of Atmospheric CH 4 and δ 13 C Isotope Signature in Methane over Arctic Seas in Summer and Autumn 2021

: Determining the sources of methane emissions in the Arctic remains a complex problem, due to their heterogeneity and diversity. Information on the amount of emissions has signiﬁcant uncertainties and may differ by an order of magnitude in various literature sources. Measurements made in the immediate vicinity of emission sources help to clarify emissions and reduce these uncertainties. This paper analyzes the data of three expeditions, carried out in the western Arctic seas during Arctic spring, summer, and early autumn in 2021, which obtained continuous data on the concentration of methane and its isotope signature δ 13 C. CH 4 concentrations and δ 13 C displayed temporal and spatial variations ranging from 1.952 to 2.694 ppm and from − 54.7‰ to − 40.9‰, respectively. A clear correlation was revealed between the surface methane concentration and the direction of air ﬂow during the measurement period. At the same time, even with advection from areas with a signiﬁcant anthropogenic burden or from locations of natural gas mining and transportation, we cannot identify particular source of emissions; there is a dilution or mixing of gas from different sources. Our results indicate footprints of methane sources from wetlands, freshwater sources, shelf sediments, and even hydrates.


Introduction
Significant reserves of methane are concentrated in the Arctic region, including the water column and seabed; however, some of the marine sources are poorly investigated. For example, various estimates of methane emissions from the sea surface at high latitudes differ by more than three orders of magnitude: from 0.013 to 20 Tg yr −1 [1][2][3][4][5][6][7][8]. Methane can enter the ocean from several sources, such as biogenic production in oxic surface waters, the formation of methane in bottom sediments, thermogenic (fossil) methane, and lateral transport by rivers. As a rule, it is difficult to separate the contribution from any of the listed sources due to the complexity of measurements. Even with the involvement of data on the isotopic signature of methane δ 13 C in the ambient air, the results show the presence of mixed sources [9][10][11].
In addition, there is no unified conception of which region makes the greatest contribution to Arctic marine emissions. Thus, in [4,5,12], it was concluded that the East Siberian Arctic Shelf (ESAS) gives from 10 to 17 Tg yr −1 on the basis of data from a number of marine expeditions. ESAS is the widest and shallowest shelf of the World Ocean and contains considerable reserves of submarine permafrost [13]. Methane hydrate stability is maintained in the special thermobaric conditions; the source hydrates, according to some estimates, may exceed by two orders of magnitude the total amount of methane in the of studying regional transport of air masses and its influence on the dynamics of methane in the Arctic region. This paper presents the results of ship-borne measurements of methane and its isotope signature performed in the Norwegian, Barents, Kara, and White Seas in 2021. One cruise was made for Arctic spring; the Kara Sea was partly covered by ice. Two other cruises were made during the navigation period when the seas were free of ice, i.e., according to [6], emissions in the ocean-atmosphere system were insignificant. Thus, the large-scale advection of air masses can be considered as the main factor of the variability of the surface methane concentration over the sea. The results of this research may be of interest in terms of studying regional transport of air masses and its influence on the dynamics of methane in the Arctic region.

Materials and Methods
An automated analysis system [9,10] based on the G2132-i instrument manufactured by Picarro Inc. (USA, Santa Clara, California) was used for obtaining the data analyzed in this paper. This instrument is designed to measure methane mole fraction in the range of 1800 to 12,000 ppb with a precision of less than 5 ppb and to measure the δ 13 CСН4 value with a precision of less than 0.8‰. The time resolution of the instrument is about 30 s, according to the manufacturer's data sheet. Such a property will not affect the average concentration picture, with averaging of 10 min or more in any case [9].
The instrument was calibrated as follows. First, it was previously calibrated according to a secondary standard, which was a 1 L cylinder with compressed air provided by the NILU (Norwegian Institute for Air Research, Kjeller, Norway) with definite methane, CO2, and δ 13 C values, calibrated, in turn, to known international standards [8]. Then, the instrument calibrated thus was used to measure the true concentrations in cylinders with calibration gas mixtures (CGMs) provided by Linde Gas Rus company (Moscow region, Russia) with a significantly better accuracy than the CGMs themselves. Subsequently, this allowed the use of these CGMs as secondary standards as well as the checking of the long-term stability of the instrument readings. Calibrations were carried out before and after each expedition.
The analysis system was located in a cabin at the deck of the research vessel steering bridge. The air inlet was a downward funnel fixed at open air (see Figure 2) at 18.5 m height above sea level with PTFE tubing 1/8" OD and 10 m length. Air flow was maintained at 1.5 L/min with a special pump of the complex, and the transit time from the inlet to the instrument turned out to be about 3 s.

Materials and Methods
An automated analysis system [9,10] based on the G2132-i instrument manufactured by Picarro Inc. (USA, Santa Clara, California) was used for obtaining the data analyzed in this paper. This instrument is designed to measure methane mole fraction in the range of 1800 to 12,000 ppb with a precision of less than 5 ppb and to measure the δ 13 C CH4 value with a precision of less than 0.8‰. The time resolution of the instrument is about 30 s, according to the manufacturer's data sheet. Such a property will not affect the average concentration picture, with averaging of 10 min or more in any case [9].
The instrument was calibrated as follows. First, it was previously calibrated according to a secondary standard, which was a 1 L cylinder with compressed air provided by the NILU (Norwegian Institute for Air Research, Kjeller, Norway) with definite methane, CO 2 , and δ 13 C values, calibrated, in turn, to known international standards [8]. Then, the instrument calibrated thus was used to measure the true concentrations in cylinders with calibration gas mixtures (CGMs) provided by Linde Gas Rus company (Moscow region, Russia) with a significantly better accuracy than the CGMs themselves. Subsequently, this allowed the use of these CGMs as secondary standards as well as the checking of the long-term stability of the instrument readings. Calibrations were carried out before and after each expedition.
The analysis system was located in a cabin at the deck of the research vessel steering bridge. The air inlet was a downward funnel fixed at open air (see Figure 2) at 18.5 m height above sea level with PTFE tubing 1/8 OD and 10 m length. Air flow was maintained at 1.5 L/min with a special pump of the complex, and the transit time from the inlet to the instrument turned out to be about 3 s.
Continuous observations of CH 4 mole fraction, carbon dioxide, and water vapor concentrations, as well as the values of the isotopic signature δ 13 C in methane, were carried out in the ambient air of the seas of the Arctic region in the periods of 19 June to 5 July, 29 July to 25 August, and 27 August to 29 September 2021 from the board of the research vessel (R/V) Akademik Mstislav Keldysh (cruisers AMK-83, AMK-84, and AMK-85). The path of the AMK-83 ran from the port of Arkhangelsk, Russia through the White and Barents Seas to the Kara Sea, in which observations were made in the areas of the St. Anna Trough and the Novaya Zemlya archipelago; the final destination of the cruise was the port of Arkhangelsk. The path of the AMK-84 cruise ran from the port of Arkhangelsk to the Barents Sea and the Norwegian-Greenland basin to the Spitsbergen archipelago, then through the Fram Strait to the southwestern periphery of the Nansen Basin and then through the Barents Sea to the port of Arkhangelsk. The route of the AMK-85 cruise ran from the port of Arkhangelsk through the White and Barents Seas to the Kara Sea, where observations were made in the areas of the Novaya Zemlya islands and the Novaya Zemlya depression and then to the port of Arkhangelsk. Route maps of the AMK-83, AMK-84, and AMK-85 cruises and the distribution of methane mole fraction in the surface layer along the expedition route are shown in Figure 3.  Data filtering. Data from a meteorological station (AIRMAR 150WX, (AIRMAR Technology Corp., Milford, NH, USA) was used to exclude the influence of the ship's stack emissions. We used the relative wind direction, where a tailwind could bring smoke from the ship's chimney back onto the system and thus contaminate the measurements. We also used CO2 data; potentially contaminated data corresponded to an increase in concentration of 10 ppm or more from the mean for each analyzed region. In total, about 7% of the data were flagged as potentially contaminated by the ship and not considered in the following analysis.

Results
Separate regions were allocated for three expeditions: White Sea, Barents Sea, Kara Sea, Norwegian Sea, and for the port of Arkhangelsk. The statistical characteristics were calculated for these areas and for the concentration of CH4 and isotopic signature δ 13 C ( Table 1). The obtained dataset allows us to identify some regularities in the spatial and temporal variability of the methane concentration in the lower atmosphere; first of all, there is a good correspondence between the СН4 concentration values and the direction of air mass removal in the periods of analyzed cruises. Data filtering. Data from a meteorological station (AIRMAR 150WX, (AIRMAR Technology Corp., Milford, NH, USA) was used to exclude the influence of the ship's stack emissions. We used the relative wind direction, where a tailwind could bring smoke from the ship's chimney back onto the system and thus contaminate the measurements. We also used CO 2 data; potentially contaminated data corresponded to an increase in concentration of 10 ppm or more from the mean for each analyzed region. In total, about 7% of the data were flagged as potentially contaminated by the ship and not considered in the following analysis.

Results
Separate regions were allocated for three expeditions: White Sea, Barents Sea, Kara Sea, Norwegian Sea, and for the port of Arkhangelsk. The statistical characteristics were calculated for these areas and for the concentration of CH 4 and isotopic signature δ 13 C ( Table 1). The obtained dataset allows us to identify some regularities in the spatial and temporal variability of the methane concentration in the lower atmosphere; first of all, there is a good correspondence between the CH 4 concentration values and the direction of air mass removal in the periods of analyzed cruises. During the AMK-83 cruise, the average CH 4 concentrations are recorded at the level of 1.962-1.968 ppm (Figure 4) for all seas, from the White to the Kara. The most intense emissions from wetlands are also noted during this period [42]. The lowest concentration of methane in the ambient air is observed in the Norwegian Sea (average 1.969 ppm) at the end of summer, in AMK-84. A gradual accumulation of methane in the ambient air begins in the Barents and White Seas, with average values reaching 1.980 and 1.989 ppm, respectively. The average methane concentration in all seas exceeded 2 ppm during the AMK-85 expedition, a significant part of which took place in September. During this period, there is already an active accumulation of methane in the Arctic and a weakening of the intensity of its sinks, while such a source as wetlands is no longer as significant as in the middle of summer [42]. Emissions become less significant, and methane sinks also become weaker as the temperature decreases [42]. This voyage is also interesting because of the greatest variety of directions of air masses advection. Figure 4 shows the distribution functions of methane concentration for different seas, covered in the three analyzed expeditions. A rather small variation of CH 4 concentrations is typical for summer cruise, with a standard deviation of 0.007-0.01 for AMK-83 cruise and 0.006-0.01 for AMK-84 (Figure 3a-f, Table 1). A bimodal structure is distinguished in some cases, as, for example, for the White Sea in AMK-83 and AMK-85 (Figure 3a,g), which is a consequence of two dominant directions of air flow transport-from Scandinavia and from the northwest, without crossing land areas.
The concentration of methane is always higher in the port of Arkhangelsk, where many anthropogenic sources are permanently active, than those observed over the open seas. The maximum was noted during the September expedition; the average value reaches 2.155 ppm. Because of the significant anthropogenic burden, data of port of Arkhangelsk were not analyzed in the paper.
Trajectory analysis. The backward five-day trajectories (every 6 h for 00, 06, 12, 18 UTC) of air flows, built on the basis of the NOAA Hysplit Lagrangian transport model [43,44], were analyzed for all expeditions. To build trajectories, we used GDAS meteorology 1 × 1, model vertical velocity; the end point is 500 m above sea level. Series of methane concentration with generalized transport direction were constructed ( Figure 5).   Table 1). A bimodal structure is distinguished in some cases, as, for example, for the White Sea in AMK-83 and AMK-85 ( Figure  3a,g), which is a consequence of two dominant directions of air flow transport-from Scandinavia and from the northwest, without crossing land areas.
The concentration of methane is always higher in the port of Arkhangelsk, where many anthropogenic sources are permanently active, than those observed over the open seas. The maximum was noted during the September expedition; the average value reaches 2.155 ppm. Because of the significant anthropogenic burden, data of port of Arkhangelsk were not analyzed in the paper.
Trajectory analysis. The backward five-day trajectories (every 6 h for 00, 06, 12, 18 UTC) of air flows, built on the basis of the NOAA Hysplit Lagrangian transport model [43,44], were analyzed for all expeditions. To build trajectories, we used GDAS meteorology 1 × 1, model vertical velocity; the end point is 500 m above sea level. Series of methane concentration with generalized transport direction were constructed ( Figure 5). AMK-83 expedition was carried out during the period of low methane content in the surface air in late June-early July. The lower methane concentration corresponds to the advection from the southwest, from the North Atlantic, according to the trajectory analysis. On the AMK-83 cruise, the minimum methane concentration corresponds to the air flow from the Baltic, which was recorded once for the entire period on 19 June from 19:30 to 20:00 UTC and gives the averaged methane mole fraction as 1.954 ppm. Decreased values also correspond to the advection from the west and northwest, in cases where the flow did not cross the Scandinavian Peninsula (on average 1.956 ppm, in some cases 1.966 ppm). Correspondingly, an average of 1.977 ppm was noted when the air flow crossed Scandinavia, and the trajectories from Taimyr corresponded to a CH 4 content equal to an average of 1.970 ppm, which is a relatively high value.
During this period, no natural fires were recorded on the peninsula; in addition, a rather low methane abundance is noted at the Dikson ground station (Dikson island, Krasnoyarsk Region, Russia) during this time [45]. During the northern advection, the methane content above the sea surface was at the level of 1.966 ppm. The highest values of methane concentration in AMK-83 correspond to the removal from the east and southeast, from the Gydan Peninsula and Yamal, where some of the largest gas production and natural gas transportation regions in Russia are concentrated; the average CH 4 concentration was 1.988 ppm. AMK-83 expedition was carried out during the period of low methane content in the surface air in late June-early July. The lower methane concentration corresponds to the advection from the southwest, from the North Atlantic, according to the trajectory analysis. On the AMK-83 cruise, the minimum methane concentration corresponds to the air flow from the Baltic, which was recorded once for the entire period on 19 June from 19:30 to 20:00 UTC and gives the averaged methane mole fraction as 1.954 ppm. Decreased values also correspond to the advection from the west and northwest, in cases where the The same patterns are observed on the AMK-84 as on the previous one. The minimum methane concentration was noted in the Norwegian Sea on 30 and 31 July 2021. The average methane concentration was 1.952, with a minimum of 1.946 ppm for the northern and north-western transport (Table 1, Figure 5). The average methane concentration slightly increased to 1.959 ppm for a northern wind in the Norwegian Sea. Increased values of CH 4 concentration are noted during the eastern transport, on average 1.982 ppm. The average methane mole fraction reaches 1.969 ppm with the southwestern transport. The CH 4 content slightly increases in the Kara and White Seas, where, with north and northeastern winds, the average value reaches 1.980 ppm, and, with north/northwestern outflow, the concentration increased to 1.981 ppm.
The longest cruise was AMK-85, the main part of which took place in September 2021, significantly repeating the route of AMK-83. The White and Barents Seas were covered for a significant part of the time (from 31.08 to 14.09) that the vessel spent in the Kara Sea, near the southwestern part of the Novaya Zemlya. This voyage exhibits the highest variability in methane concentration as well as the greatest variation in the direction of air flows. By September, a seasonal increase in the CH 4 content is also noticeable in the surface layer of the atmosphere. So, the minimum concentration of CH 4 is noted at the beginning of the expedition in the White Sea during the western air transfer from the Scandinavian Peninsula (average 1.972 ppm, minimum 1.958 ppm). In the Barents Sea, advection from the north and northeast prevailed. The average methane concentration was 1.978 ppm; the methane concentration slightly increased, up to 1.995 ppm, with air capture from Taimyr.
The ship spent a long time in the roadstead southeast of Novaya Zemlya in the Kara Sea. The increased methane concentration during this period corresponds to the advection from the mainland, especially from the gas production areas, where the CH 4 concentration significantly exceeds 2 ppm. A minimum of methane of 1.966 ppm is noted for the advection from the west in the Kara Sea. CH 4 is always above 2 ppm for the eastern directions of transport. The concentration rises sharply, up to 2.060 ppm, in the case of air flow crossing the mainland, especially regions of possible natural gas production. The increased methane concentration corresponds to the transport from the southern regions of Russian European territory and the north of Kazakhstan for the AMK-85 cruise (Figure 4, 24-27 September 2021, the average methane concentration is 2.050 ppm). Anthropogenic sources and natural fires are considered as possible reasons for the increased concentration of methane, which still took place in the European Territory of Russia and the Urals.
Isotopic composition. Figure 6 shows the δ 13 C variability. In all likelihood, the ships almost all of the time were distant from methane sources.  The Keeling plot method was used for more accurate analysis of potential methane contributors [46][47][48]. The largest possible range of mixing ratios of СН4 in the air mass is required to ensure maximum accuracy in determining the isotopic characterization of excess methane in the air mass. The difficulty in identifying sources during ship-borne measurements is that the air flow is not uniform and accumulates methane from several sources while passing over different areas. Figure 7 shows examples of Keeling plots. The Keeling plot method was used for more accurate analysis of potential methane contributors [46][47][48]. The largest possible range of mixing ratios of CH 4 in the air mass is required to ensure maximum accuracy in determining the isotopic characterization of excess methane in the air mass. The difficulty in identifying sources during ship-borne measurements is that the air flow is not uniform and accumulates methane from several sources while passing over different areas. Figure 7 shows examples of Keeling plots.  Despite a significant anthropogenic burden around the port of Arkhangelsk, the Keeling plot shows dominance of emissions from wetlands (Figure 7a). Figure 7b,c show that values are outside the range (more depleted) than expected for Finland wetlands [32]. But this range of values was typical for this investigation. Figure 7b,c are well out- Despite a significant anthropogenic burden around the port of Arkhangelsk, the Keeling plot shows dominance of emissions from wetlands (Figure 7a). Figure 7b,c show that values are outside the range (more depleted) than expected for Finland wetlands [32]. But this range of values was typical for this investigation. Figure 7b,c are well outside the range (more depleted) than expected for northern wetlands. Figure 7d is much more depleted than expected wetland sources. It is within the range of d 13 C reported for ESAS sediment sources in [34]. So, a seafloor source of CH 4 is possibly represented here. The results are consistent with data published, for example, in [40], where it was shown that at the Nordic Seas the share of methane emissions from hydrates dominates.

Discussion and Conclusions
Observations of both methane concentration and its isotopic ratio δ 13 C were carried out in summer and autumn of 2021 in the Arctic aboard the R/V Akademik Mstislav Keldysh. As expected, during the period covered by the research, a significant influence on the methane concentration variability in the lower atmosphere is determined mainly by large-scale processes, namely by the direction of air masses advection. The minimum methane concentration during the expeditions occurred in Arctic spring and summer, observed due to air mass removal from the west direction. A similar result was obtained, for example, during a marine expedition [40] where the minimum was obtained in the western seas of the Arctic. The maximum methane concentration in the ambient air is observed during advection from the northwest of Siberia (areas of potential natural gas production and transportation). In the autumn period, the highest methane concentration was recorded during advection from natural gas production areas (northwest of Siberia) and invasion from the south of the European territory of Russia and north of Kazakhstan, where the anthropogenic load was high and wildfires were noted (https://firms.modaps. eosdis.nasa.gov/map/, last accessed on 20 December 2021). However, the ship, almost all of the time, was at a considerable distance from methane sources. The Keeling plot analysis did not reveal a prevailing source of methane during all three campaigns. According to the analysis, results emissions from freshwater sources, wetlands, seafloor sources, and hydrates contribute to the measured variations.