Figures and Tables

top

Introduction

Across the northern permafrost region, 1,662 Tg of carbon per year is released to the atmosphere as carbon dioxide (CO₂) and methane (CH₄) during the winter season (October-April; Natali et al., 2019

Natali, S.M., Watts, J.D., Rogers, B.M., Potter, S., Ludwig, S.M., et al. (2019) Large loss of CO₂ in winter observed across the northern permafrost region. Nature Climate Change 9, 852–857. https://doi.org/10.1038/s41558-019-0592-8

; Arndt et al., 2023

Arndt, K.A., Hashemi, J., Natali, S.M., Schiferl, L.D., Virkkala, A.-M. (2023) Recent Advances and Challenges in Monitoring and Modeling Non-Growing Season Carbon Dioxide Fluxes from the Arctic Boreal Zone. Current Climate Change Reports 9, 27–40. https://doi.org/10.1007/s40641-023-00190-4

). Winter carbon emissions may be preceded by subsurface mobilisation and transfer of dissolved organic carbon (DOC) within taliks (Parazoo et al., 2018

Parazoo, N.C., Koven, C.D., Lawrence, D.M., Romanovsky, V E., Miller, C E. (2018) Detecting the permafrost carbon feedback: Talik formation and increased cold-season respiration as precursors to sink-to-source transitions. The Cryosphere 12, 123–144. https://doi.org/10.5194/tc-12-123-2018

; Walter Anthony et al., 2024

Walter Anthony, K.M., Anthony, P., Hasson, N., Edgar, C., Sivan, O., et al. (2024) Upland Yedoma taliks are an unpredicted source of atmospheric methane. Nature Communications 15, 6056. https://doi.org/10.1038/s41467-024-50346-5

; Bergman et al., 2025

Bergman, O., Walter Anthony, K., Eliani-Russak, E., Sivan, O. (2025) Nitrogen Redox Controls on Greenhouse Gas Production in Yedoma Taliks. Global Change Biology 31, e70356. https://doi.org/10.1111/gcb.70356

), processes that remain poorly constrained. A talik is defined as “a part of the ground at a site with permafrost that remains unfrozen year round” (Lewkowicz et al., 2024

Lewkowicz, A.G., Wolfe, S.A., Roujanski, V.E., Hoeve, E., O’Neill, H.B., et al. (2024) An Illustrated Permafrost Dictionary Canadian Permafrost Association. https://ostrnrcan-dostrncan.canada.ca/handle/1845/347162

) and may be either between the base of the active layer and permafrost (supra-permafrost talik), or within a permafrost profile (intra-permafrost talik; Devoie et al., 2024

Devoie, É.G., Connon, R. F., Beddoe, R., Goordial, J., Quinton, W.L., Craig, J.R. (2024) Disconnected active layers and unfrozen permafrost: A discussion of permafrost-related terms and definitions. Science of The Total Environment 912, 169017. https://doi.org/10.1016/j.scitotenv.2023.169017

). Supra-permafrost taliks can be either closed or lateral (O’Neill et al., 2020

O’Neill, H.B., Roy-Leveillee, P., Lebedeva, L., Ling, F. (2020) Recent advances (2010–2019) in the study of taliks. Permafrost and Periglacial Processes 31, 346–357. https://doi.org/10.1002/ppp.2050

): closed taliks are surrounded by frozen ground with talik water relatively stagnant, having only a localised sphere of influence and limited flow potential; and lateral taliks can form lateral water flow pathways, leading to an increase in DOC transfer from soil to streams in winter (Connon et al., 2018

Connon, R., Devoie, É., Hayashi, M., Veness, T., Quinton, W. (2018) The Influence of Shallow Taliks on Permafrost Thaw and Active Layer Dynamics in Subarctic Canada. Journal of Geophysical Research: Earth Surface 123, 281–297. https://doi.org/10.1002/2017JF004469

; Devoie et al., 2024

Devoie, É.G., Connon, R. F., Beddoe, R., Goordial, J., Quinton, W.L., Craig, J.R. (2024) Disconnected active layers and unfrozen permafrost: A discussion of permafrost-related terms and definitions. Science of The Total Environment 912, 169017. https://doi.org/10.1016/j.scitotenv.2023.169017

). Thus, distinguishing between closed and lateral taliks is crucial for determining how supra-permafrost taliks regulate winter carbon transfer from soils to streams (Kurylyk and Walvoord, 2021

Kurylyk, B.L., Walvoord, M.A. (2021) Permafrost Hydrogeology. In: Yang, D., Kane, D.L. (Éds.) Arctic Hydrology, Permafrost and Ecosystems, 493–523. Springer International Publishing. https://doi.org/10.1007/978-3-030-50930-9_17

; Zhao et al., 2026

Zhao, Y., Zheng, C., Gelfan, A., Watanabe, K., Liu, H., et al. (2026) Frozen Soil Hydrological Processes and Their Effects: A Review and Synthesis. Reviews of Geophysics 64, e2024RG000839. https://doi.org/10.1029/2024RG000839

).

Water tracks are unchanneled zones overlying permafrost that concentrate water flow downslope, and exhibit higher soil moisture and snow thickness than the surrounding areas (Del Vecchio and Evans, 2025

Del Vecchio, J., Evans, S.G. (2025) Climate and Hydrogeological Controls on Water Tracks in Permafrost Landscapes. Reviews of Geophysics 63, e2024RG000854. https://doi.org/10.1029/2024RG000854

). Moreover, soil moisture increases along the water tracks, leading to wetter areas downstream of the water tracks, mainly fed by water from upstream during the growing season (Tananaev et al., 2021

Tananaev, N., Isaev, V., Sergeev, D., Kotov, P., Komarov, O. (2021) Hydrological Connectivity in a Permafrost Tundra Landscape near Vorkuta, North-European Arctic Russia. Hydrology 8, 3. https://doi.org/10.3390/hydrology8030106

). Increasing soil moisture associated with water flows enhances advective heat transfer, promoting the development of taliks, potentially favouring closed taliks where the soil moisture is lower upstream, and lateral taliks where the soil moisture is higher downstream of the water track (Devoie et al., 2021

Devoie, É.G., Craig, J.R., Dominico, M., Carpino, O., Connon, R.F., Rudy, A.C.A., Quinton, W.L. (2021) Mechanisms of Discontinuous Permafrost Thaw in Peatlands. Journal of Geophysical Research: Earth Surface 126, e2021JF006204. https://doi.org/10.1029/2021JF006204

). Existing methods for talik detection are based on mechanical, geophysical and thermal approaches (O’Neill et al., 2020

O’Neill, H.B., Roy-Leveillee, P., Lebedeva, L., Ling, F. (2020) Recent advances (2010–2019) in the study of taliks. Permafrost and Periglacial Processes 31, 346–357. https://doi.org/10.1002/ppp.2050

). The gap in biogeochemical methods reflects the exceptional logistical challenges of collecting soil pore water from taliks during deep winter, when frozen surface conditions and extreme cold make in situ sampling difficult.

Talik connectivity may be assessed using the stable isotope composition of silicon (δ³⁰Si), based on Hirst et al. (2023)

Hirst, C., Monhonval, A., Mauclet, E., Thomas, M., Villani, M., Ledman, J., Schuur, E.A.G., Opfergelt, S. (2023) Evidence for late winter biogeochemical connectivity in permafrost soils. Communications Earth and Environment 4, Article 1. https://doi.org/10.1038/s43247-023-00740-6

, who measured δ³⁰Si in soil pore waters to distinguish isolated from connected soil pore water during the late winter to snow melt transition. We apply it here to a colder, earlier stage of winter, prior to any onset of thaw. Upon freezing of soil pore water under a closed system, the decrease in liquid water volume drives an increase in the silicic acid concentration in soil pore waters over the amorphous silica solubility (Dietzel, 2005

Dietzel, M. (2005) Impact of cyclic freezing on precipitation of silica in Me–SiO2–H2O systems and geochemical implications for cryosoils and -sediments. Chemical Geology 216, 79–88. https://doi.org/10.1016/j.chemgeo.2004.11.003

). Amorphous silica precipitation induces silicon isotope fractionation by preferentially incorporating lighter isotopes into solid phases (from nanosized solid phases termed colloids to particulate sized phases), leaving the dissolved soil pore water fraction isotopically heavier (Oelze et al., 2015

Oelze, M., von Blanckenburg, F., Bouchez, J., Hoellen, D., Dietzel, M. (2015) The effect of Al on Si isotope fractionation investigated by silica precipitation experiments. Chemical Geology 397, 94–105. https://doi.org/10.1016/j.chemgeo.2015.01.002

). This process may happen in closed taliks (Hirst et al., 2023

Hirst, C., Monhonval, A., Mauclet, E., Thomas, M., Villani, M., Ledman, J., Schuur, E.A.G., Opfergelt, S. (2023) Evidence for late winter biogeochemical connectivity in permafrost soils. Communications Earth and Environment 4, Article 1. https://doi.org/10.1038/s43247-023-00740-6

), leaving the soil pore water from taliks isotopically heavier than soil pore water from the surrounding frozen soil (soil pore water δ³⁰_talikSi > δ³⁰_frozenSi). In lateral taliks, water flow prevents silicic acid concentration. In addition, the freezing front facilitates colloid mobilisation through preferential flow paths (Mohanty et al., 2014

Mohanty, S.K., Saiers, J.E., Ryan, J.N. (2014) Colloid-Facilitated Mobilization of Metals by Freeze–Thaw Cycles. Environmental Science and Technology 48, 997–984. https://doi.org/10.1021/es403698u

) from the frozen ground to the lateral talik. This leads to a soil pore water isotopically lighter in lateral taliks with the accumulation of light δ³⁰Si bearing-colloids compared to the surrounding frozen soil (soil pore water δ³⁰_talikSi < δ³⁰_frozenSi).

This study aims to biogeochemically distinguish supra-permafrost taliks located upstream, downstream, and outside a water track from Eight Mile Lake (Healy, Alaska; Fig. 1), using soil pore water chemistry (pH, conductivity, DOC concentrations, and δ³⁰Si) from the taliks and the frozen portions of the soil within the active layer. The chemistry of local first order (inlet) and second order stream waters (Panguingue Creek River) is investigated to assess the implications of talik connectivity for DOC transfer to streams (pH, conductivity, δ¹⁸O, DOC concentrations, radiocarbon of DOC (Δ¹⁴C-DOC), dissolved CO₂ and CH₄ concentrations, Si concentrations, and δ³⁰Si). As winters become increasingly warmer in permafrost regions (Graham et al., 2017

Graham, R.M., Cohen, L., Petty, A.A., Boisvert, L.N., Rinke, A., Hudson, S.R., Nicolaus, M., Granskog, M.A. (2017) Increasing frequency and duration of Arctic winter warming events. Geophysical Research Letters 44, 6974–6983. https://doi.org/10.1002/2017GL073395

), we compare two contrasted winters to examine how interannual differences in air temperature (winter 2025 warmer than winter 2024) influence carbon transfer to streams (see SI).

Full size image

top

Contrasting Interannual Winter Conditions

At Eight Mile Lake, winter air temperatures in 2023–2024 and 2024–2025 were significantly higher compared to the previous ten winters (−11.3 ± 9.1 °C; p < 0.001; Wilcoxon test; Kennedy et al., 2026

Kennedy, E., Celis, G., See, C., Schaedel, C., Mauritz, M., Taylor, M., Ledman, J., Natali, S., Schuur, E.A.G. (2026) Eight Mile Lake Research Watershed: Hourly meteorological data, 2004-2025, Bonanza Creek LTER - University of Alaska Fairbanks. BNZ:453. https://doi.org/doi:10.6073/pasta/c1b76570a7f3d4feaef428e98970b2be

), and the winter freezing degree day was higher in 2024–2025 (−1528 °C) than in 2023–2024 (−1911°C; calculated by summing all negative values between 1 November and 14 May; Farquharson et al., 2022

Farquharson, L.M., Romanovsky, V.E., Kholodov, A., Nicolsky, D. (2022) Sub-aerial talik formation observed across the discontinuous permafrost zone of Alaska. Nature Geoscience 15, 475–481. https://doi.org/10.1038/s41561-022-00952-z

). Winter 2024–2025 was then warmer than winter 2023–2024 overall, and this was also observed more specifically during the sampling period (February-March): air temperatures were significantly higher (p value < 0.001) of 1.9 °C in 2025 (−7.4 ± 6.9 °C (mean ± s.d.) with a minimum of −25 °C) than in 2024 (−9.3 ± 10.6 °C (mean ± s.d.) with a minimum of −34 °C; Fig. 1d). As a result, February-March soil temperatures were significantly higher (p value < 0.001) in 2025 than in 2024 for each depth at each site, with increases ranging from 1.5 to 2.4 °C (Fig. 1e). Snow depth at the site was thinner both upstream and outside of the water track in 2025 (15 ± 8 cm, n = 6) than in 2024 (23 ± 13 cm, n = 6; Fig. 1b). Despite a large variability in snow depth, this trend is consistent with regional observations of thinner snow depth in 2025 (42 ± 23 cm) than in 2024 (50 ± 23 cm) in Healy (Thompson et al., 2025

Thompson, A., Ledman, J., Celis, G., Mauritz, M., Taylor, M., Schuur, E.A.G. (2025) Eight Mile Lake Research Watershed, Carbon in Permafrost Experimental Heating Research (CiPEHR): CiPEHR snow depth manual data 2009-2025, Bonanza Creek LTER - University of Alaska Fairbanks. BNZ:629. https://doi.org/doi:10.6073/pasta/79eafbe49cff540115cd6869e382f28c

). These contrasting air and soil temperatures and snow layer thicknesses between the two winters may then influence talik biogeochemical connectivity.

top

Chemical Composition of the Soil Pore Waters and Streams

In soil pore waters from the taliks and the frozen portions of the soil, pH, conductivity, and DOC concentrations varied from 2.9 to 5.1 (3.9 ± 0.6), 10.3 to 105 μS cm^-1 (53.5 ± 28.5 μS cm^-1), and 25.4 to 299 mg L^-1 (100 ± 59.0 mg L^-1) respectively (Fig. 2a-c). In soil pore waters from the taliks sampled both in 2024 and 2025 (n = 6; upstream and downstream of the water track), DOC concentrations are 45 (U2), 26 (D1) and 75 % (D3) lower in 2025 compared to 2024, respectively, suggesting DOC dilution within a wetter water track under warmer conditions in 2025 (Del Vecchio and Evans, 2025

Del Vecchio, J., Evans, S.G. (2025) Climate and Hydrogeological Controls on Water Tracks in Permafrost Landscapes. Reviews of Geophysics 63, e2024RG000854. https://doi.org/10.1029/2024RG000854

).

Full size image

At the inlet (first order stream), water pH and conductivity varied respectively between 6.3 and 4.3, and between 136 and 65 μS cm^-1 (Fig. 2, Table 1). Water δ¹⁸O values and DOC concentrations were respectively 4 % and 24 % higher in 2025 compared to 2024. Values of Δ¹⁴C-DOC were slightly higher in 2025 (−1.7 ‰) than in 2024 (−19.4 ‰), corresponding to young DOC radiocarbon ages of ∼0 and 86 years BP, respectively, potentially derived from recently fixed plant litter in the downslope area close to the inlet (Benner et al., 2004

Benner, R., Benitez-Nelson, B., Kaiser, K., Amon, R.M.W. (2004) Export of young terrigenous dissolved organic carbon from rivers to the Arctic Ocean. Geophysical Research Letters 31. https://doi.org/10.1029/2003GL019251

). Dissolved CO₂ and CH₄ concentrations in inlet waters were considered stable between 2024 (420 μmol L^-1 CO₂ and 19.4 μmol L^-1 CH₄) and 2025 (443 μmol L^-1 CO₂ and 11.9 μmol L^-1 CH₄) given the typical variability of these quantities in streams in permafrost areas (Street et al., 2016

Street, L.E., Dean, J.F., Billett, M.F., Baxter, R., Dinsmore, K.J., et al. (2016) Redox dynamics in the active layer of an Arctic headwater catchment; examining the potential for transfer of dissolved methane from soils to stream water. Journal of Geophysical Research: Biogeosciences 121, 2776–2792. https://doi.org/10.1002/2016JG003387

).

Parameters (units)	Inlet (first order stream)		Panguingue Creek River (second order stream)
	2024	2025	2024	2025
*pH     (-)	6.3	4.3	7.4	6.0
*Conductivity (μS cm-1)	136	65	274	317
δ18O        (‰)	−17.2	−16.4	−18.5	−19.1
δD      (‰)	−145.6	−137.2	−148.2	−146.1
*DOC    (mg L-1)	29.4	36.6	1.7	2.7
Δ14C      (‰)	−19.4	−1.7	−521.0	−508.2
CO2     (μmol L-1)	420	443	139	91
CH4     (μmol L-1)	19.4	11.9	0.030	0.045
*Si      (mg L-1)	2.5	1.3	7.5	7.5
*δ30Si     (‰)	2.04	1.33	1.71	1.14

At the Panguingue Creek River (second order stream), water pH and conductivity varied respectively between 7.4 and 6.0, and between 274 and 317 μS cm^-1 (Fig. 2, Table 1). Water δ¹⁸O values were 3 % lower, and DOC concentrations were 57 % higher in 2025 compared to 2024. Values of Δ¹⁴C-DOC were higher (i.e. less depleted) in 2025 (−508.2 ‰, 5628 years BP) compared to 2024 (−521 ‰, 5841 years BP), indicating a younger DOC source in 2025. Dissolved CO₂ and CH₄ concentrations remained similar in 2024 (139 μmol L^-1 CO₂ and 0.030 μmol L^-1 CH₄) and 2025 (91 μmol L^-1 CO₂ and 0.045 μmol L^-1 CH₄) in the Panguingue Creek River water, in agreement with the absence of changes observed at the inlet.

top

Silicon Isotopes Reveal Talik Connectivity (Closed versus Lateral)

In soil pore waters from the taliks and from the frozen portions of the soil, Si concentrations varied from 4.0 to 18.1 mg L^-1, and the silicon isotope compositions (δ³⁰Si) varied from −0.02 to 2.88 ‰ (Fig. 2d-e). To characterise talik connectivity (closed versus lateral), we consider the difference in δ³⁰Si values between soil pore waters from the taliks and from the frozen portions of the soil, i.e. the Δ_Frozen-Talikδ³⁰Si (Fig. 2f).

In 2024, the Δ_Frozen-Talikδ³⁰Si (Fig. 2f) had two distinct patterns depending on the locations of the cores along the slope: upstream of the water track, δ³⁰Si was higher in the taliks than in the frozen portions of the soil (Δ_Frozen-Talik δ³⁰Si < 0) with a mean difference of −1.08 ‰; whereas downstream of the water track, δ³⁰Si was lower in the taliks than in the frozen portions of the soil (Δ_Frozen-Talik δ³⁰Si > 0) with a mean difference of +0.86 ‰, consistent with Si isotope fractionation during freezing processes (Hirst et al., 2023

Hirst, C., Monhonval, A., Mauclet, E., Thomas, M., Villani, M., Ledman, J., Schuur, E.A.G., Opfergelt, S. (2023) Evidence for late winter biogeochemical connectivity in permafrost soils. Communications Earth and Environment 4, Article 1. https://doi.org/10.1038/s43247-023-00740-6

). Upstream of the water track (δ³⁰Si_talik > δ³⁰Si_frozen), the heavier signature is consistent with closed system Si isotopic fractionation (Hatton et al., 2019

Hatton, J.E., Hendry, K.R., Hawkings, J.R., Wadham, J.L., Opfergelt, S., Kohler, T.J., Yde, J.C., Stibal, M., Žárský, J.D. (2019) Silicon isotopes in Arctic and sub-Arctic glacial meltwaters: The role of subglacial weathering in the silicon cycle. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences 475, 20190098. https://doi.org/10.1098/rspa.2019.0098

): in closed taliks. Downstream of the water track (δ³⁰Si_talik < δ³⁰Si_frozen), the lighter signature indicates an open system without Si isotopic fractionation in lateral taliks. Downstream taliks were also shallower (40 ± 17 cm) than upstream (64 ± 5 cm; Fig. 1b, Table S-1), consistent with a wetter, more connected zone that may transfer carbon to streams in deep winter.

In 2025, Δ_Frozen-Talik δ³⁰Si was negative upstream of the water track and positive downstream of the water track, which confirms the results obtained in 2024 (Fig. 2f): closed taliks are upstream and lateral taliks are downstream of the water track. Moreover, δ³⁰Si values from soil pore waters outside of the water track exhibit patterns similar to those measured upstream of the water track: Δ_Frozen-Talik δ³⁰Si was negative with δ³⁰Si in the taliks and 0.70 ‰ higher than in the frozen portions of the soil. As a result, taliks from outside of the water track are closed, like upstream of the water track.

top

Tracing Lateral Talik Contributions to Streams Using Silicon Isotopes

At the inlet (first order stream), dissolved Si concentrations and δ³⁰Si values were 47 % and 35 % lower, respectively, in 2025 compared to 2024 (Fig. 2d,e, Table 1). Inlet δ³⁰Si was lower in 2025 (1.33 ‰) than in 2024 (2.04 ‰), a shift toward lighter values, consistent with an increased contribution of soil pore waters from lateral taliks (0.59 ± 0.58 ‰) in 2025. The increase in δ¹⁸O values from the inlet waters in 2025 compared to 2024 also supports this higher contribution of soil pore waters under warmer conditions (Table 1; Hirst et al., 2022

Hirst, C., Mauclet, E., Monhonval, A., Tihon, E., Ledman, J., Schuur, Edward. A.G., Opfergelt, S. (2022) Seasonal changes in hydrology and permafrost degradation control mineral element-bound DOC transport from permafrost soils to streams. Global Biogeochemical Cycles 36, e2021GB007105. https://doi.org/10.1029/2021GB007105

). As a result, our data suggest that, under warmer conditions (2025), winter inlet waters reflect an increased input from lateral talik pore waters, in addition to a persistent contribution from deep sub-permafrost groundwater (O’Donnell et al., 2012

O’Donnell, J.A., Aiken, G.R., Walvoord, M.A., Butler, K.D. (2012) Dissolved organic matter composition of winter flow in the Yukon River basin: Implications of permafrost thaw and increased groundwater discharge. Global Biogeochemical Cycles 26, GB0E06. https://doi.org/10.1029/2012GB004341

; Roux et al., 2026

Roux, P., Hirst, C., Villani, M., du Bois d’Aische, E., Osy, C., et al. (2026) Radiogenic Sr isotopes reveal extended seasonal windows of soil-river dissolved organic carbon transfer in an Arctic permafrost catchment. Chemical Geology 712, 123427. https://doi.org/10.1016/j.chemgeo.2026.123427

).

At the Panguingue Creek River (second order stream), Si concentrations remained similar in 2024 and 2025 (7.5 mg L^-1), and δ³⁰Si values were 33 % lower in 2025 compared to 2024 (Fig. 2d,e, Table 1). Panguingue Creek δ³⁰Si was lower in 2025 (1.14 ‰) than in 2024 (1.71 ‰), a shift toward lighter values, consistent with an increased contribution of soil pore waters from lateral taliks (0.59 ± 0.58 ‰) in 2025, similar to the pattern observed at the inlet. The decrease in δ¹⁸O in 2025 compared to 2024 (Table 1) likely reflects intra-river processes (Gibson and Prowse, 2002

Gibson, J.J., Prowse, T.D. (2002) Stable isotopes in river ice : Identifying primary over-winter streamflow signals and their hydrological significance. Hydrological Processes 16, 873–890. https://doi.org/10.1002/hyp.366

), including reduced river ice thickness during the warmer winter (90 cm in 2024 and 68 cm in 2025; see SI), rather than a direct soil signal. This indicates that Panguingue Creek, as a second order stream, reflects an increased input from talik pore waters under warmer conditions (2025), in addition to a persistent contribution from deep sub-permafrost groundwater (O’Donnell et al., 2012

O’Donnell, J.A., Aiken, G.R., Walvoord, M.A., Butler, K.D. (2012) Dissolved organic matter composition of winter flow in the Yukon River basin: Implications of permafrost thaw and increased groundwater discharge. Global Biogeochemical Cycles 26, GB0E06. https://doi.org/10.1029/2012GB004341

; Roux et al., 2026

Roux, P., Hirst, C., Villani, M., du Bois d’Aische, E., Osy, C., et al. (2026) Radiogenic Sr isotopes reveal extended seasonal windows of soil-river dissolved organic carbon transfer in an Arctic permafrost catchment. Chemical Geology 712, 123427. https://doi.org/10.1016/j.chemgeo.2026.123427

), and intra-river processes.

top

Implications for Winter Carbon Mobilisation and Transfer into Streams

Lateral taliks being hydrologically connected during winter (Zhao et al., 2026

Zhao, Y., Zheng, C., Gelfan, A., Watanabe, K., Liu, H., et al. (2026) Frozen Soil Hydrological Processes and Their Effects: A Review and Synthesis. Reviews of Geophysics 64, e2024RG000839. https://doi.org/10.1029/2024RG000839

), they likely enhance the transfer of DOC to streams. As the sampling period in 2025 was warmer than in 2024, the increase in DOC concentrations in the first and the second order stream waters in 2025 compared to 2024 suggests an increase in this mobilisation and the transfer of DOC from soils to streams (Fig. 3). The parallel increase in Δ¹⁴C values at the second order stream (from −521.0 in 2024 to −508.2 ‰ in 2025; Table 1) under warming conditions also supports a greater contribution of younger DOC from talik pore waters (Δ¹⁴C-DOC = −123.5 ± 120.2 ‰; Fig. S-1a). This increase in transfer of young DOC from lateral taliks in 2025 is in line with δ³⁰Si and δ¹⁸O values at the first order stream, and δ³⁰Si at the second order stream (Fig. 2, Table 1). Stream waters were sampled once per year at each site, precluding formal statistical testing of interannual differences. However, the consistent direction of change across several independent parameters (DOC concentration, Δ¹⁴C-DOC, δ³⁰Si, and δ¹⁸O) demonstrates that these shifts reflect an increase in lateral talik contribution to streams under warmer conditions, rather than sampling variability alone.

Full size image

At both the first and second order streams, dissolved CO₂ and CH₄ concentrations were stable between 2024 and 2025, (Table 1), suggesting that the mobilisation of DOC from lateral taliks in 2025 compared to 2024 does not translate into a change in the production of CO₂ and CH₄ and transfer to streams. A similar decoupling between DOC mobilisation and CO₂/CH₄ production was reported in Yedoma thaw affected lakes, where highly labile permafrost DOC nonetheless accumulated rather than being rapidly mineralised (Ollivier et al., 2026

Ollivier, S., Séjourné, A., Hatté, C., Bouchard, F., Noret, A., Hughes-Allen, L., Costard, F., Gandois, L. (2026) Massive concentrations of old dissolved organic carbon from Yedoma thaw in lakes in Siberia. Communications Earth and Environment 7, 200. https://doi.org/10.1038/s43247-026-03229-0

). This decoupling contrasts with conditions in the soil itself: in free standing water from a shallow unconfined saturated zone at the most downstream site (Fig. S-1b; D4), dissolved CO₂ and CH₄ concentrations increased ∼2 and ∼17 fold, respectively, between 2024 (CO₂ = 303 μmol L^-1, CH₄ = 2.7 μmol L^-1) and 2025 (CO₂ = 590 μmol L^-1, CH₄ = 45.0 μmol L^-1), pointing to stronger CO₂ and CH₄ production in saturated soils under warmer conditions. This enhanced soil production, however, provides no evidence of a corresponding transfer of dissolved CO₂ and CH₄ to streams. Alternatively, we cannot exclude that part of the transferred DOC is mineralised along the flow path between the first and second order streams, without being captured by our fixed point sampling. Enhanced winter DOC transfer from lateral taliks and active microbial CO₂ and CH₄ production in talik soils (Fig. S-1b) may also have delayed effects on riverine greenhouse gas dynamics during subsequent high flow period after deep winter (Castro-Morales et al., 2022

Castro-Morales, K., Canning, A., Körtzinger, A., Göckede, M., Küsel, K., et al. (2022) Effects of Reversal of Water Flow in an Arctic Floodplain River on Fluvial Emissions of CO₂ and CH₄. Journal of Geophysical Research: Biogeosciences 127, e2021JG006485. https://doi.org/10.1029/2021JG006485

), contributing to the permafrost driven increase in annual stream CO₂ emissions (Mu et al., 2025

Mu, C., Li, K., Liu, S., Wei, Y., Mu, M., et al. (2025) Recent intensified riverine CO₂ emission across the Northern Hemisphere permafrost region. Nature Communications 16, 3616. https://doi.org/10.1038/s41467-025-58716-3

).

top

Conclusion

This study provides geochemical evidence of lateral talik connectivity using silicon isotopes. Contrasting δ³⁰Si signatures between taliks and frozen soil pore waters (heavier upstream and lighter downstream of the water track) distinguish closed from lateral taliks. The higher DOC concentrations in stream waters in 2025 than in 2024 suggest a greater lateral talik contribution to streams under warmer conditions. Together, these results indicate that lateral taliks act as critical subsurface pathways linking winter soil pore waters DOC mobilisation to aquatic carbon cycling. As winters continue to warm, the expansion of lateral taliks may enhance DOC transfer to streams, but do not necessarily lead to immediate increases in riverine dissolved CO₂ and CH₄ concentrations during winter. This study represents a first step toward understanding talik biogeochemical connectivity, given the logistical challenges of winter sampling in remote permafrost environments. Expanding observations across multiple sites and combining geochemical tracers with existing geophysical and thermal characterisation of taliks will be essential to fully assess the biogeochemical significance of lateral taliks in permafrost ecosystems.

top

Acknowledgments

We thank the analytical platform MOCA at UCLouvain, L. Smeyers for his contribution to field sampling, and the valuable input from LandSense team K. Van Oost, V. Vanacker, F. Jonard and S. Lambot (Action de Recherche Concertée (ARC), no. 21/26–119, funded by the Fédération Wallonie-Bruxelles (FWB)). We acknowledge funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (ERC Starting Grant, WeThaw, n°714617) to SO, from the Fonds National de la Recherche Scientifique to SO, MV and E d BA (FNRS: FC69480, FC49507, FC54613, and CDR MOIST), from BELSPO (RT/23/LIFTHAW), from CircleU (Seed funding CEDRIC), and from VOCATIO grant (Belgium) to MV. We thank the editor, Andreas Kappler, and the two reviewers, Aaron Bufe and one anonymous reviewer, for their valuable comments and suggestions.

Editor: Andreas Kappler

top

Data Availability

The data related to this article is available online at: https://doi.org/10.14428/DVN/LPFBWO

top

References

top

Supplementary Information

The Supplementary Information includes:

• Materials and Methods
• Supplementary Figures S-1 and S-2
• Supplementary Information References

Download the Supplementary Information (PDF)