the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 06 Jul 2026
Permafrost thaw reshapes methane cycling across an interior Alaska peatland
Abstract. Permafrost thaw is expected to increase methane (CH4) emissions from northern peatlands, but it remains uncertain how thaw affects the microbial pathways driving CH4 production and oxidation. We measured CH4 and carbon dioxide (CO2) fluxes during July 2022 using static chambers and combined these measurements with δ13C-CH4 measurements, 16S rRNA and mcrA gene sequencing, as well as environmental data across a fine-scale thaw gradient in an interior Alaska peatland. Mean CH4 fluxes increased from near neutral (-0.1 ± 0.02 μmol m-2 s-1) in stable thaw stages to 40.2 ± 12.0 μmol m-2 s-1 in advanced thaw stages, while CO2 fluxes did not change. Thaw progression was associated with higher water tables and deeper seasonal thaw depths with an increase in methanogen relative abundance and shift in methanogen community composition. Hydrogenotrophic taxa increased in relative abundance with advanced thaw along the gradient, particularly Methanoregula, which explained 36 % of the variability in CH4 fluxes. Despite this shift, emitted δ13C-CH4 values fell within ranges commonly attributed to acetoclastic methanogenesis. Rayleigh fractionation modeling suggests that ~ 40 % of CH4 produced at depth was oxidized before reaching the atmosphere, enriching residual CH4 in 13C and altering the isotopic signature of emitted fluxes. These results highlight the need to integrate fluxes, isotopes, and microbial community data to fully resolve CH4 cycling processes in thawing permafrost peatlands.
- Preprint
(1369 KB) - Metadata XML
-
Supplement
(694 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-2809', Anonymous Referee #1, 31 Aug 2026
-
RC2: 'Comment on egusphere-2026-2809', Anonymous Referee #2, 22 Sep 2026
The authors simultaneously obtained static box CH 4/CO2 flux, box air and pore water δ 13C-CH 4, 16S rRNA and mcrA microbial communities, and water level/active layer/temperature environmental data along a stable early middle late freeze-thaw gradient in the Beta peatland of Alaska in July 2022. They found that CH 4 flux increased from near zero/weak absorption in the stable stage to about 40.2 μ mol m ⁻ ² min ⁻ ¹ in the late stage, while CO2 flux showed no significant change; As the water level rises and the active layer deepens, methane producing archaea overall increase and hydrogenotrophic Methanoegula significantly expands (explaining about 36% of CH 4 variation), but the δ 13C and apparent fractionation factors that emit CH 4 still fall within the range typically attributed to acetic acid cracking. The Rayleigh model estimates that about 40% of CH 4 is oxidized from deep to surface. This study has clear academic value for the identification of CH 4 pathways, model parameterization, and carbon feedback uncertainty in permafrost peatlands, and is worthy of publication. But before formal acceptance, the manuscript needs to be thoroughly revised. The specific revision suggestions are as follows:
- Inconsistent CH 4 flux units and questionable values (abstract, lines 15-25; methods, lines 130-135): The abstract states stable stage -0.1 ± 0.02 μ mol m ⁻² s ⁻¹, late stage 40.2 ± 12.0 μ mol m ⁻² s ⁻¹;; method, line 135, also states' CO2 reported in μ mol m ⁻² s ⁻¹ while CH 4... μ mol m ⁻² min ⁻¹ ', resulting in the appearance of' μ mol m ⁻ m ⁻ '. The layout is incorrect and the seconds/minutes are not standardized. 40 μ mol m ⁻² s ⁻¹ shows abnormally high CH 4 in peatlands, suspected to be due to μ mol m ⁻² min ⁻¹ or statistical errors in the original second values. Suggestion: Unify the flux unit throughout the manuscript, verify the original GasScouter output with the unit in formula (1), recalculate the mean/error, and synchronously correct the summary, Figure 3, and Table S2. If using the signed square root transformation (lines 280-285), explain the transformed unit and substitution method.
- Use the year 2021 instead of 2022 for soil moisture, and 2022 is relatively dry (lines 100-108): Due to instrument malfunction, the humidity of the two periods in 2021 was used to characterize the relative dryness of the sample plot. However, the author stated that 2022 was one of the driest in more than a decade, and humidity/hydrology is the core control factor for CH 4 generation. Using non synchronous humidity for correlation/environmental fitting will weaken the causal chain of "environment microorganism flux". Suggestion: Use 2022 water level, active layer depth, and temperature as the main environmental variables; Humidity is only used as a qualitative background and does not enter PERMANOVA/envfit/RDA, or supplementary data from nearby automatic meteorological/soil moisture sensors during the same period; Clearly declare that humidity results cannot be used for mechanism inference.
- The isotopic depth of pore water does not match the microbial stratification (microbial sampling lines 95-115; pore water lines 155-160): 16S/mcRA samples are taken at 0-15, 15-30, and 30-45 cm, while the δ ¹ ³ C of pore water is taken at 33 cm and 100 cm, respectively; 100 cm is usually deeper than the seasonal melting layer and also deeper than the microbial sampling depth. Using 'deep porewater' to perform Rayleigh analysis on the box gas and associating it with 30-45 cm mcrA results in spatial mismatch. Suggestion: Resample and analyze pore water at a depth of 0-15/15-30/30-45 cm; If 100 cm is retained, a separate core of ≥ 100 cm should be collected for 16S/mcRA, and the "deep pore water" and "deep community" should be strictly separated in the main text.
- The sample size for CH 4 oxidation quantification is too small and the hypothesis is too strong (lines 200-215; results lines 370-372): Only two early samples have both deep pore water and box gas δ, and the oxidation rate is only estimated for the middle and late stages; Using a single ε=-15 ‰, closed system Rayleigh, an "apparent oxidation rate" can be obtained in open diffusion soils. Suggestion: Conduct a sensitivity analysis of ε at -5-25 ‰ and provide a Bootstrap confidence interval; If possible, construct a mass balance/open transport model using multi-layer pore water δ ¹ ³ C; Clearly define all percentages as apparent oxidation to avoid confusion with absolute oxidation rate.
- Lack of direct evidence of methane oxidizing bacteria, but repeatedly attributed to oxidation (isotope 145-190 lines; discussion about 515-525 lines): The main text states that methane is present in all stages, and oxidation changes emit δ, but the full text does not provide the relative abundance, composition, or correlation coefficient with δ enrichment of methane at the 16S or functional gene level. Suggestion to supplement: Extract the abundance depth freeze-thaw stage distribution of methane oxidizing bacteria from 16S Guild or pmoA/mXO data as independent evidence of oxidation; If pmoA is supplemented unconditionally, at least the abundance of 16S oxidizing bacteria should be provided and combined with surface δ enrichment and water level oxygen content for analysis.
- The classification and process of mcrA need to be strengthened (lines 245-270): After amplification with ML_ccrA, only forward reads and BLAST were used to compare GTDB r226 and identify species with ≥ 90% amino acid consistency and ≥ 60 aa. For single copy functional genes, GTDB is not the optimal mcrA specific library, and there is no indication of nucleotide to amino acid translation or primer coverage bias. Suggestion: Parallel use of FunGene mcrA, NCBI mcrA/purified methane protein library, or MethHan database for comparison; Report on the coverage evaluation of primers for Methanorelula/Ethanobacterium/Ethanosarcina, etc; Provide the ASV species correspondence table and the proportion of unclassified species below the threshold.
7.16S and mcrA, relative abundance and activity are not consistent (lines 220-245, 395-450): 16S may have coverage bias for some archaea/methanogens using 515F/926R; Relative abundance after rareify does not equal activity or absolute quantity. The use of relative abundance in the text to explain CH 4 flux (Methanorelula explains 36%, total methane producers+stage explains 76%) can easily amplify the correlation. Suggestion: If feasible, supplement the absolute abundance or transcript (mcrA mRNA) of qPCR mcrA/pmoA; At least discuss relative abundance limitations and conduct dilution curves and coverage assessments; Provide cross validation or independent test sets for "interpretability" to avoid overfitting of small samples.
- Insufficient report on sample size and grouping details (lines 285-310, 370, 415-425): The final 71 samples for 16S and the final 80 samples for mcrA were not listed for each freeze-thaw stage x depth (0-15/15-30/30-45); In the advanced stage, there are very few detectable samples of box gas isotopes. The R ² given by PERMANOVA/RDA (such as MCrA stage R ²=0.30, seasonal melting depth R ² A=18.6%) requires sample size and post efficacy. Suggest adding a table for 'Number of sequencing/flux/isotope samples per stage and depth'; Adopt a more conservative permutation design for imbalanced data and report the effect size instead of just p in the main results.
- Insufficient treatment of bubble emissions (lines 120-143): Only 2 abnormally high absorption values of initial CH 4 were excluded and attributed to ebulling; However, in the late stage of high-throughput, there may be bubbles, and the static box+3 minutes will mix diffusion with bubbles on average. The δ ¹ ³ C fractionation of bubbles is different from diffusion, which will affect Rayleigh and path determination. Suggestion: Record the visible bubbles/flux nonlinear segments before and after sampling, and identify bubbles through segmented fitting; Discuss separately in isotope interpretation the light/heavy migration effects of ebulling on surface δ, or at least explain in the limitations.
- The apparent fractionation of α C for path discrimination is easily misled by oxidation (lines 175-195, results 345-350, Figure 5): using surface/box gas CH 4 and CO2 to calculate α C, if surface CH 4 has been oxidized and enriched, α will be smaller and disguised as acetic acid cracking; Although this point is discussed in the text, Figure 5 and the threshold are still presented based on the emission gas. Suggestion: Use "unoxidized/deep pore water" CO2 - CH 4 to calculate produced α and distinguish between production α and emission α; In the main text, the threshold method is limited to situations without strong oxidation or after corrected oxidation.
- Time matching between freeze-thaw stage division and July observation (lines 80-95, 315-320): The stages are divided based on the thickness of the active layer in September (Cox et al. 2025), while flux/isotope/core are in July; The 2022 drought may result in incomplete correspondence between the active layer in July and the classification in September. Suggest adding threshold values for each stage, number of years used, and consistency between ALT measured in July and classification; In the stable stage, there is no pore water, and there are only 2 samples in the early oxidation stage. The representative limitations for slope inference should be clarified within the limitations.
The processing of CO2 flux is too simplistic (lines 125-135, 330-335). CH 4 uses a dark box, CO2 uses a transparent box for 3 minutes to represent NEE; However, the coupling analysis of CH 4 oxidation, methane substrate turnover, and CO2/respiration is insufficient. Suggest adding black box CO2 or soil respiration decomposition (autotrophic/heterotrophic), and using CO2 isotopes and O2/redox as auxiliary evidence when discussing oxidation rates; If only NEE is used, it should be noted that it cannot represent soil respiration.
- Please add the following literature: Introduction/Research Blank Paragraph (in the middle of the Introduction section, if the line numbers of the entire text are around 120-180): Insert the literature after explaining "Permafrost that may increase peatland CH 4 but pathways remain uncertain"（ https://doi.org/10.3389/fmicb.2023.1181658 ； https://doi.org/10.1029/2026JG009919 ）
Citation: https://doi.org/10.5194/egusphere-2026-2809-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 202 | 130 | 29 | 361 | 28 | 33 | 33 |
- HTML: 202
- PDF: 130
- XML: 29
- Total: 361
- Supplement: 28
- BibTeX: 33
- EndNote: 33
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The authors investigated the effect of the thaw gradient of permafrost in a peatland in Alaska on the CO2 and CH4 fluxes, combining environmental variables like water table depth, seasonal thaw depth, soil temperature and soil moisture with changes in microbial community composition based on the 16S rRNA gene sequencing and analysis of methanogenic communities based on the mcrA gene and measures 𝛿13C-CH4 signatures released from thaw and from deeper porewater. This study addresses relevant scientific questions within the scope of BG, integrating “biological, chemical, and physical processes in terrestrial life with the geosphere, hydrosphere, and atmosphere”. The experiment is well-designed. The authors concluded that combining 𝛿13C-CH4 signatures with characterisation of the changes in methanogenic communities in response to thawing allowed a more reliable interpretation of methane cycling processes.
A taxonomic shift in methanogenic communities indicated an increasing contribution of hydrogenotrophic methanogen taxa, which may reveal a change towards hydrogenotrophic methanogens observed at sites in the advanced stages of thawing. In contrast, based solely on the isotope results, the false conclusion might be drawn that acetoclastic methanogenesis is dominant. Their study indicates that, based on the model used, approximately 40% of the methane produced is oxidised before leaving the soil profile, altering the interpretation of isotopic results and possibly explaining the discrepancy between the results obtained.
Overall, the results described are interesting; however, there is a lack of a more critical discussion of the methods used. (1) The estimated percentage of oxidised CH₄ is valid only for the variables assumed in the model. (2) The mcrA analysis was carried out using forward reads, and (3) the presence of a given physiological group, as determined by DNA, should be interpreted with greater caution as evidence of actual functional activity. (4) It is also necessary to describe in detail the potential impact of the spatial separation of the Permafrost Plateau, with stable and early thaw stages, from the Active Thaw Margin (intermediate, advanced) on the results obtained. It is understandable that in the field it is not possible to plan everything as one would in a laboratory, and that data collection is fraught with methodological problems, but these concerns should be discussed. (5) Raw NGS data should be deposited in an appropriate scientific repository, and the data access number should be included with the article before final acceptance of the manuscript. The journal Biogeosciences favours reliable, public repositories that comply with the FAIR principles and use persistent identifiers
In its current form, the manuscript requires a major revision.
Other comments:
Gene names, such as mcrA, should be written in italics throughout the text, whilst protein names should remain in normal typeface.
L.225: Check the sequence for 926R primer. It looks like you pasted the sequence for a different primer (806R). Additionally, in the supplementary materials, you used another primer 896R. Is it correct? The accuracy of the data in the supplementary materials should be checked.
L. 293 vs L. 375: what was the final number of 16S rRNA samples: 71 or 62? Methodological data referring to the initial number of samples and the numbers remaining after each stage of data preparation should be presented in a single place
The interpretation of Figures 6 and 8 could be enhanced by a brief description of the direction of the environmental-factor gradient at each melting stage. In particular, it is worth noting that the distinction between stable/early and intermediate/advanced communities essentially corresponds to changes in thaw depth, groundwater level, temperature, and CH₄ flux. Moreover, the figure caption should define all the environmental vectors shown in the graph.
L. 397-400: The authors list five types whose relative abundances change significantly. However, in the discussion, they refer to only one of them (Bacteroidota). Besides, in microbiology, classification at the phylum level is useful for the general characterisation of communities and for comparing communities of microorganisms in different environments. Therefore, it is good that fig. 7 is presented. However, such classifications are usually too general to allow for detailed functional interpretation, as presented in Discussion (L. 467-469), because individual phyla include taxa with highly diverse metabolic characteristics. To identify the groups of microorganisms most likely responsible for the observed changes in community composition, the authors should analyse changes at lower taxonomic levels, such as the family or genus level. It is also possible that a significant trend at the phylum level is largely due to a strong response from one or more dominant taxa at lower taxonomic levels.
L. 418-419 The non-linear trend in methanogen alpha diversity is potentially interesting, but has not been discussed. Both Shannon diversity and ASV richness decline from stable to intermediate thaw, then increase markedly during the advanced thaw phase. The authors should discuss potential ecological explanations for this phenomenon.