Hot spots of methane emission in West Siberian middle taiga wetlands disturbed by petroleum extraction activities

Introduction. The concentration of methane in the Earth's atmosphere, the second most potent greenhouse gas, continues to rise since 2007 [Canadell et al., 2021]. The need to significantly reduce the anthropogenic emission of methane into the atmosphere in order to limit the increase in global temperature by 2100 within 2°C relative to the period from 1850 to 1900 is recognized by both the scientific community [IPCC, 2021] and the leadership of most countries of the world, including Russia, who signed and ratified the Paris Agreement, adopted following the results of the 21st Conference of the UN Framework Convention on Climate Change [Climate Agenda of Russia, 2021]. Reduction of methane emissions and control over it throughout the territory of managed ecosystems will require huge resources and investments, development of new climate-smart technologies. A reasonable compromise may be to identify the most important sources of methane within managed ecosystems (also called “hot spots”) and to introduce changes in their land-use in accordance with the principles of sustainable development and science-based environmental management.

The major type of economic activity in the taiga natural zone of West Siberia is oil production [Koleva, 2007; Volkova, 2010]. Since 35-40% of the West Siberian middle taiga area is covered with waterlogged ecosystems - wetlands and floodplains [Peregon et al., 2009; Terentieva et al., 2016], a significant part of this infrastructure is located in wetland ecosystems and has a strong impact on them. In this paper, we made the first attempt to understand, how the most common types of disturbances by oil production (road, pipeline and electric power transmission line construction) can affect methane emissions from the most common disturbed waterlogged ecosystems in the region (oligotrophic raised bogs on a terrace or watershed) and eutrophic lowland swamps in the floodplain). We measured methane emission from the surface of disturbed wetland ecosystems, physicochemical and biological factors influencing it, to identify which ecosystems are hot spots of methane emission.

Objects. The study area was located 50 km southeast of the city of Khanty-Mansiysk, on the right bank of the Irtysh River, in the natural zone of the middle taiga. The climate of this region is subarctic (Dfc according to Köppen). In the floodplain of the Irtysh the most common types of wetlands are sedge-grass open swamps and sogras (treed sedge-grass wetlands), on terraces and the watershed - pine-shrub-sphagnum ecosystems (ryams) and ridge-hollow complexes [Liss et al., 2001]. The thickness of the peat layer in raised bogs on the terrace and watershed varied from 2 to 3 m; in sogra – from 3.5 to 4 m; in open floodplain swamps thickness of organic-rich horizon never exceeded 0.4 m. For floodplain ecosystems we investigated influence of a four-lane access road on changing the hydrological functioning of open swamps (points OO and OK), as well as the effect of cross-cut in a sogra (SP) compared to an undisturbed sogra (SE). For raised bogs on the terrace and watershed, we study the influence of asphalt two-lane roads which act as dams, preventing the flow of water from one side of the road to the other resulting in flooding to upstream areas (GMKO1 and GMKO2) and drying in downstream areas (GMKS) in ridge-hollow complexes. In ryams and ridge-hollow complexes The effect of cross-cutting on methane emission in ryams (RP1 and RP2) as well as pipeline installation in ryam (RTO1) and ridge-hollow complex (RTO2) were also studied. During a cross-cut tree layer was destroyed, the vegetation and moss cover was compacted (RP1) or mostly destroyed (RP2 and SP). Access roads were constructed 3 (four-lane) and 10-15 (asphalt two-lane) years ago. Pipelines were installed 2-3 years ago.

Methods. Methane flux was measured using the static chamber method [Hutchinson and Mosier, 1981]. In the course of one flux measurement four syringes were taken from the chamber on the interval of 10 min. Total duration of one flux measurement was 30 minutes. Three consecutive replicates of the flux measurements were carried out on each of the three collars per each investigated ecosystem. Interval between two consecutive flux measurements was 10 min. Water were sampled from the depth of 20 cm below water table level (WTL) in two replicates to determine dissolved organic carbon (DOC) content at the points GMKO2, GMKS, RTO1, RTO2, RP2, as well as in an undisturbed ryam ecosystem 50 m away from the points RTO1 and RP2. The concentration of DOC was measured by a Flash 2000 elemental analyzer using an AS1310 automatic liquid sampler (both Thermo Fisher Scientific, USA). In each studied ecosystem for each collar the values of WTL (cm, positive water is below the level of the moss surface), pH and electrical conductivity (μS·cm^-1) of water were measured. All calculations were carried out in the MATLAB software environment R2022a (MathWorks, USA).

Results and discussion. Methane emission varied from 0.005 to 41.7 mg·m^-2·h^-1 with a median of 2.1 mg·m^‑2·h^‑1. Fluxes were not distributed normally (p < 0.0001, N = 33), but could be described by the lognormal distribution (p = 0.15) and the Weibull distribution (p = 0.22). Such a significant distribution asymmetry indicates that changes of land-use practice in several ecosystems with the highest methane emission could help to reduce methane emission significantly without substantial modifications of the whole landscape. The dependence of the methane flux on WTL differs depending on both disturbance and ecosystem types. Within one ecosystem, the maximum emission values can be observed both in most flooded sites (RP2, GMKS), in sites with intermediate WTL values (GMKO1, RTO2, OK), and in sites with the highest WTL (RTO1). One of the markers of methane emission hot spots is the appearance of ruderal plants Eriophorum vaginatum and Trichophorum cespitosum in different ecosystems and on disturbances of different types. Eriophorum vaginatum is one of the first species to settle on bare peat in cross-cuts (RTO1 and RTO2) and footprints after heavy equipment (RP2) in raised bogs, as well as on seismic survey lines in sogra (SP). Trichophorum cespitosum was found in the upstream area of the road, where a zone of excessive moisture has formed resulting in degradation of the moss and vegetation cover and peat decomposition (GMKO1). In all these five ecosystems, methane flux from sites covered with Eriophorum vaginatum and Trichophorum cespitosum was 2 or more times higher compared to the surrounding sites where these species were absent.

The maximum values of methane emission among all studied ecosystems are in the WTL range from -2 to 8 cm (see Fig. 1). In studied raised bogs, the emission from the flooded upstream areas (GMKO1 and GMKO2) was significantly lower (p = 0.0082, N = 8) than from the dried downstream areas (GMKS), if we exclude the point with Trichophorum cespitosum, where high methane emission is attributed, presumably, to the influence of the plant community and not with to the different WTL, as described in the section above. In contrast, for floodplain wetlands, emission from the open sedge bog in the drying area (OO) was significantly lower (p = 0.02, N = 6) than from the flooded open swamp with Phalaris arundinacea (OC). This difference could be explained by changes in local ecohydrology and hydrochemistry after the road construction. Methane emission from ridges in GMKO1 and GMKO2 ecosystems (median 1.5 mg·m^-2·h^-1) exceeds by an order of magnitude the median of methane emission from middle taiga ridges Western Siberia (0.13 mg·m^-2·h^-1 according ^‑to [Kleptsova et al., 2010]). Due to flooding in the upstream area of the roads, WTL in ridges decreased compared to values typical for these ecosystems (mean ± standard deviation is 35 ± 14 cm according to [Kleptsova et al., 2010]). However, the grass-moss layer of the ridges did not degrade, and the methane emission from them turned out to be comparable with the emission from undisturbed ridges with the same WTL values (Fig. 2).

Methane emission from temperate and subarctic swamps is typically characterized by a lower optimal WTL value (ranging from -20 cm to -5 cm) compared to bogs [Bao et al., 2021]. Therefore, flooding of the Phalaris arundinacea swamp (OK) resulted in optimal conditions for methanogenesis in all three studied sites of this ecosystem with WTL ranging from -12 to 3 cm. The methane emission in each site of the Phalaris arundinacea swamp was higher than the third quartile for the entire sample obtained in this study. The open sedge bog (OO) separated from the rest of the floodplain by the road was characterized by a higher WTL (from -5 to 12 cm), far from optimal. In addition, the soil temperature in these ecosystems, located at a distance of 600 meters from each other, differed by 9-11°C in a peat layer from 0 to 20 cm. The same pattern was observed in sogra wetland, where temperature of the upper 20 cm in cross-cut bare peat was 6-8°C higher than in undisturbed site, separated from floodplain by access road. Thus, both the temperature and hydrological regimes contribute to the fact that the methane emission from the flooded floodplain open swamp (OK) is significantly higher than from the floodplain bog in the drying area (OO point). A similar pattern was observed for the treed floodplain swamp (SP and SE points, respectively).

The concentration of DOC in the water of natural and disturbed ecosystems of the low ryam was significantly higher than in the hollow of the ridge-hollow complex (p < 0.01, N = 5). The same pattern was observed for Canadian wetlands and was explained by the fact that DOC production occurs mainly in the aeration zone above the WTL. Since in ryams and ridges WTL it is higher than in hollows, the rate of plant litter decomposition is twice as high as in hollows (Moore, 2009). The higher rate of decomposition can explain both the higher EC (faster mineralization) and the lower pH (higher acidogenesis) in the low ryam. It is noteworthy that during the disturbance and subsequent recovery of the vegetation in the ryam, the concentration of DOC in the peat pore water increased by almost one and a half times, while in the hollow of the ridge-hollow complex it did not change considerably compared to the value in undisturbed wetland ecosystem.

Conclusion. Measurements of methane emission from wetlands of the West Siberian middle taiga disturbed during oil production and its physicochemical and biological factors showed that several of these ecosystems are intensive sources of this greenhouse gas. Although this is only a snapshot taken at the end of June 2021, and it is necessary to study the seasonal dynamics of the methane flux for more reliable conclusions, several indicators of methane emission hot spots could be suggested. Presence of ruderal plants such as Eriophorum vaginatum and Trichophorum cespitosum marks such a hot spots throughout different ecosystems. Ecosystem-specific range of WTL optimal for methane emission could also be a reliable indicator of these hot spots. Response of methane emission to the construction of roads depends on type of wetland ecosystems. In raised bogs, hollows in the upstream area emit less methane than undisturbed ecosystems, while in the downstream area emission is higher. Emission from ridges in flooded ridge-hollow complexes increases with the decrease of the WTL in them, similarly to natural undisturbed ridges. Nutrient-rich floodplain swamps response differently to changes in the hydrological regime. The emission of methane from open and forested swamps in the drying area is lower than from flooding area. This is explained not only by different WTL optimums for methane emission between bogs and swamps but also differences in temperature (6-11°С) of the surface organic-rich layers of floodplain wetlands in the flooding area compared to drying area. The methane emission from heavy vehicle tracks in low ryam is driven by the change in WTL relative to its optimum for methane emission from raised bogs.