Methane concentration
Dissolved gases in meltwaters and in periglacial creeks, rivers, and lakes consist primarily of methane and carbon dioxide; no heavier hydrocarbons were detected. Methane concentrations in meltwaters range from 12–20 nM, similar to those in periglacial rivers and creeks (9–24.7 nM). Periglacial lakes exhibit a broader range and higher values (10.8–93.5 nM; Fig. 2; Supplementary Table 1). All analysed waters are supersaturated with methane—by up to a factor of 20—relative to atmospheric equilibrium (4.9 nM)7. Despite this supersaturation, the measured methane concentrations are 1–3 orders of magnitude lower than those reported for meltwaters in southwestern Greenland (e.g., Russell Glacier, 110 nM; Leverett Glacier, 270 nM; Isunnguata Sermia, 1,575 nM)3,5,7,8. However, they are similar to some glaciers in the Northwest Greenland7 and exceed concentrations measured in southern Greenland, where values are near atmospheric equilibrium (e.g., Kiattut Sermiat, 9 nM)5,7. Concentrations are also lower than those reported for several glacial meltwater systems elsewhere, such as in Canada (136–899 nM)22, but fall within the range observed in Alaska (2.8–120 nM)23 and China (9–16 nM; Fig. 2)24.
Fig. 2: Methane concentrations in meltwaters, periglacial lakes, rivers, and creeks.
The diagram shows meltwater (yellow hexagons; avg. 15.2 nM, max. 19.7 nM, min. 11.8 nM), rivers and creeks (orange triangles; avg. 17.1 nM, max. 24.7 nM, min. 9 nM), lakes (purple circles; avg. 38.2 nM, max. 93.6 nM, min. 10.8 nM) data for North Greenland (blue area) but also methane concentrations for south and southwest (avg. 1071 nM, max. 14,026 nM, min. 1.4 nM), central west (avg. 70 nM, max. 347 nM, min. 9 nM), and northwest Greenland meltwaters (avg. 65 nM, max. 184 nM, min. 0.4 nM), and proglacial waters in Canada, Alaska, China, Eastern Alps, and Svalbard (empty squares) for comparison. Notably, the lowest methane concentrations are associated with cold-based glaciers in the southern (e.g., Kiattuut Sermiat), northern (this study), and northwestern Greenland Ice Sheet, whereas the highest concentrations occur in the southwestern sector (e.g., Isunnguata Sermia), where glacier beds are predominantly thawed31. Comparison data were obtained from Christiansen et al.8, Pain et al.5, Lamarche-Gagnon et al.3, Sapper et al.22, Konya et al.23, Du et al.24, Ragnoli et al.48, Kleber et al.14, and Hatton et al.7. Selected glaciers mentioned in the text are shown. The dashed line indicates the atmospheric equilibrium concentration of methane (4.9 nM)7.
The relatively low methane concentrations observed in our study area are unexpected, given that the northern margin of the GrIS overlies an actively degassing petroleum system. In analogous geological settings such as Svalbard, methane concentrations in proglacial waters fed by groundwater springs are several orders of magnitude higher (up to 106 nM)14. We did not observe such features in the study area. The concentrations measured here are also lower than those reported from regions where glaciers rest directly on crystalline bedrock, including southwestern Greenland3. The processes responsible for the low methane concentrations potentially include: (1) specific characteristics of the underlying petroleum system, such as low gas content or extensive prior degassing; (2) the spatial extent of the petroleum system beneath the northern GrIS25 (Supplementary Fig. 3), (3) low rates of subglacial microbial methanogenesis, potentially due to limited availability of organic substrates, (4) the absence of long-term microbial methanogenesis within bedrock fracture networks26, (5) methane oxidation within the subglacial environment3, (6) early hydrocarbon migration during the Devonian, combined with effective degassing associated with major basin uplift during the Late Devonian to Early Carboniferous (370–350 Ma)27,28, (7) possible shorter water residence time in the subglacial environment of northern Greenland relative to southern GrIS basins29,30, (8) the existence of a thick permafrost acting as a seal for deep gas migration11, and (9) the existence of cold-based glacier patches in North Greenland that limit basal meltwater production31. Cold-based glaciers are also present in northwest and south Greenland31, where methane concentrations are similarly low7. The low methane concentrations measured in North Greenland can also be attributed to sampling at the end of the melt season, which in this region typically occurs in September32. Consistent with this seasonal effect, late‑season meltwater measurements from the southwestern GrIS show an approximately ten‑fold decrease in methane concentrations relative to peak melt‑season values, reaching levels comparable to those observed in this study (ca. 18 nM)3. Despite lower concentrations, meltwaters in northern Greenland remain consistently supersaturated with methane and thus act as a net atmospheric source.
Stable carbon isotopes
Stable carbon isotopes of methane and associated carbon dioxide were used to elucidate the origin of methane. Meltwaters exhibit δ13CCH4 values ranging from − 41.9‰ to −35.1‰ and δ13CCO2 values between −13.6‰ and −7‰ (Supplementary Table 1). These values reflect methane that is markedly enriched in ¹³C relative to that dissolved in glacial meltwaters elsewhere across the GrIS, where values are typically < −50‰3,5,33,34. Periglacial rivers and creeks display comparable isotopic compositions, with δ13CCH4 values from −42.1‰ to −38.2‰ and δ13CCO2 values from −13.4‰ to −9.6‰. In contrast, periglacial lakes show a broader range, particularly for methane, with δ13CCH4 values spanning −58.6‰ to −38.8‰ and δ13CCO2 values from −11.3‰ to −6.7‰ (Supplementary Table 1). For comparison, we also analysed gas trapped in a bitumen-bearing limestone sample from Polaris Foreland (Hauge Bjerge Formation; Fig. 1 and Supplementary Figs. 1 and 2). This gas contains heavier hydrocarbons (ethane, propane, butane, and pentane) with a methane-to-heavier‑hydrocarbon ratio of approximately seven and exhibits δ13CCH4 of −48.6‰ and δ13CCO2 of −14.2‰. These signatures clearly indicate a thermogenic, oil-associated origin (Fig. 3). Despite the absence of measurable heavier hydrocarbons in meltwater and periglacial stream samples, their isotopic signatures indicate a predominantly thermogenic methane source, and/or potential influence of oxidation processes. (Fig. 3). However, several additional lines of evidence support a thermogenic origin: (1) the combined δ13CCH4 and δ13CCO2 values plot within the field characteristics of oil-associated thermogenic methane; (2) isotopic signatures resulting from oxidation of microbial methane typically exhibit much wider δ13CCH4 and δ13CCO2 ranges22,35,36,37, (3) active degassing of underlying petroleum systems is documented11, (4) the close match between the isotopic composition of methane and carbon dioxide in meltwaters and in the associated bitumen-rich bedrock, and (5) shorter water residence time in the subglacial environment29,30, which limit the development of strongly reducing conditions required for methanogenesis and interaction times in cold subglacial environments where metabolic rates are inherently slow1,38,39. The δ13CCH4 and δ13CCO2 values from periglacial lakes similarly point to a predominantly thermogenic, oil‑associated source, except in lakes that are not clearly connected to glacial runoff. These isolated lakes show distinctly microbial signatures36, with δ13CCH4 < −56‰ (i.e., sites 5LAN and 8LAN; Fig. 1; Supplementary Table 1).
Fig. 3: Genetic diagram for methane.
Carbon stable isotopes of methane (δ13CCH4) and carbon dioxide (δ13CCO2) in waters sampled in North Greenland: meltwater (yellow hexagons), rivers and creeks (orange triangles), and lakes (purple circles). Subglacial meltwaters from southwestern Greenland3 (empty squares) and degassed limestone with oil inclusions of the Hauge Bjerge Formation, Franklinian Basin (red star) were plotted for comparison. The sample from this study, located within the “Primary microbial” field, originates from a lake that shows no clear connection to glacial runoff (site 8LAN). The arrow indicates the methane oxidation trend. Diagram fields from Milkov and Etiope36.
Implications
The thermogenic origin of methane beneath the North GrIS contrasts sharply with observations elsewhere in Greenland, where high dissolved methane concentrations are typically attributed to microbial methanogenesis within subglacial sediments3,4,7,34. The comparatively low concentrations observed in northern Greenland demonstrate that methane storage beneath ice sheets cannot be inferred from bedrock type alone (e.g., degassing sedimentary units), and that additional factors such as sampling location and timing relative to the melt season may also exert a strong influence. In addition, the presence or absence of organic‑rich sediments plays a critical role in sustaining microbial methanogenesis and generating elevated methane concentrations in meltwaters—sometimes exceeding 3 × 10⁴ nM⁴— even where glaciers overlie crystalline bedrock1,4,6,7,40. The expression of this microbial potential is modulated by hydrological residence times, which limit microbial processing under cold subglacial conditions, as well as by deglaciation history and permafrost evolution, which regulate the quantity and reactivity of organic carbon. Meltwaters from cold-based glaciers in the North, Northwest, and South Greenland31 contain, in general, lower methane concentrations than the ones in Southwest and West Greenland7, which have warmer, thawed beds (Supplementary Fig. 3)31. In addition, large permafrost‑affected areas of the pan‑Arctic shelf, for instance, contain organic carbon that is poorly reactive, thereby limiting microbial methanogenesis and increasing the relative contribution of deeper methane sources at the surface41. This framework is consistent with our observations, in which methane is predominantly derived from geological sources independently of subglacial biogeochemical processes. Further permafrost aggradation following glacier retreat may, however, form effective barriers to methane migration from depth11, thereby reducing its expression at the surface.
Our results identify a previously under-recognised methane source beneath the GrIS—degassing from petroleum‑bearing formations—and highlight the need for geological methane contributions to be incorporated into methane emission budgets. The magnitude and persistence of this geological source warrant further investigation. The extent to which these findings can be extrapolated across the northern sector of the GrIS remains poorly constrained. However, our observations from major drainage systems, including Peterman and C.H. Ostenfeld glaciers, together with the southward extent of petroleum systems within the Franklinian Basin (Supplementary Fig. 3), suggest that thermogenic methane may be distributed over a broad area beneath the northern GrIS. The western limit of thermogenic influence aligns with the margin of the Franklinian Basin, beyond which methane in the northwestern GrIS is primarily microbial (Supplementary Fig. 3)7. We further hypothesise that continued GrIS retreat could enhance petroleum‑system degassing by reactivating underlying tectonic structures12,13, potentially increasing methane delivery to subglacial drainage networks and the atmosphere. In addition, our findings suggest that higher methane concentrations are more likely beneath the GrIS in regions hosting younger petroleum systems, such as the Nuussuaq Basin with Cretaceous and Palaeogene source rocks in West Greenland42.
Our measurements of methane in meltwaters and periglacial rivers, creeks, and lakes across North Greenland expand the limited dataset of greenhouse‑gas observations from the High Arctic. These measurements reveal pronounced spatial variability in both the concentration and isotopic composition of subglacial methane associated with the GrIS. By providing constraints from North Greenland, this dataset improves the empirical basis for estimating methane emissions from the rapidly retreating ice sheet. These observations also offer essential benchmarks for evaluating and refining existing and future modelling efforts2,43, and underscore the need to incorporate potential geological methane sources when assessing the contributions of large ice masses to Arctic and global carbon budgets.