the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 24 Oct 2023
Review article: A systematic review of terrestrial dissolved organic carbon in northern permafrost
Dolly N. Kothawala
Lars J. Tranvik
Abstract. As the permafrost region warms and permafrost soils thaw, vast pools of soil organic carbon (C) become vulnerable to enhanced microbial decomposition and lateral transport into aquatic ecosystems as dissolved organic carbon (DOC). The mobilization of permafrost soil C can drastically alter the net northern permafrost C budget. DOC entering aquatic ecosystems becomes biological available for degradation as well as other types of aquatic processing. However, it currently remains unclear which landscape characteristics are most relevant to consider in terms of predicting DOC concentrations entering aquatic systems from permafrost regions. Here, we conducted a systematic review of 111 studies relating to, or including, concentrations of DOC in terrestrial permafrost ecosystems in the northern circumpolar region published between 2000–2022. We present a new permafrost DOC dataset consisting of 2,276 DOC concentrations, collected from the top 3 m in permafrost soils across the northern circumpolar region. Concentrations of DOC ranged from 0.1–500 mg L-1 (median = 41 mg L-1) across all permafrost zones, ecoregions, soil types, and thermal horizons. DOC concentrations were greatest in the sporadic permafrost zone (101 mg L-1) while lower concentrations were found in the discontinuous (60 mg L-1) and continuous (59 mg L-1) permafrost zones. The highest median DOC concentrations of 66 mg L-1 and 63 mg L-1 were found in coastal tundra and permafrost bog ecosystems, respectively. Coastal tundra (130 mg L-1), permafrost bogs (78 mg L-1), and permafrost wetlands (57 mg L-1) had the highest DOC concentrations in the permafrost lens, representing a potentially long-term store of DOC. Other than in Yedoma ecosystems, DOC concentrations were found to increase following permafrost thaw and were highly constrained by total dissolved nitrogen concentrations. This systematic review highlights how DOC concentrations differ between organic- or mineral-rich deposits across the circumpolar permafrost region and identifies coastal tundra regions as areas of potentially important DOC mobilization. The quantity of permafrost-derived DOC exported laterally to aquatic ecosystems is an important step for predicting its vulnerability to decomposition.
- Preprint
(1751 KB) - Metadata XML
-
Supplement
(818 KB) - BibTeX
- EndNote
Liam Heffernan et al.
Status: open (until 05 Dec 2023)
-
RC1: 'Comment on tc-2023-152', Anonymous Referee #1, 20 Nov 2023
reply
- General comments: My recommendation for this manuscript is to resubmit after major revisions. The results are of significant interest to The Cryosphere, but the current manuscript requires considerable modifications before considering immediate publication. Specifically, the introduction and discussion sections are currently written in a structure that is difficult for readers to understand. Additionally, numerous English grammar issues have been identified in the sentences. Focused improvements in these areas could greatly enhance the quality of the manuscript.
- Detailed Comments
- [L55-56] I think you need references to back this up.
- [L67-71] The sentence seems overly lengthy. It may be beneficial to break it down into shorter sentences for clarity.
- [L71-73] The sentence is grammatically incomplete and requires revision for clarity.
- [L74-76] I'm uncertain whether the current sentence is necessary as the opening line of this paragraph, considering the section primarily discusses the lateral movement of DOC.
- [L79-81] The sentence seems more fitting for the previous paragraph. I'm unsure if it's essential for the narrative progression of the current section.
- [L74-99] The critical concepts in the second paragraph of the introduction might be 'the lateral transport of DOC,' 'permafrost landscape dynamics,' and their impacts on the 'Arctic freshwater ecosystems.' However, due to a lack of connectivity between these concepts in the paragraph, it is challenging to grasp the narrative and logic the author intends to convey. I think addressing this issue to clarify how these concepts interrelate is necessary.
- [L100-117] This paragraph should clearly articulate how the research was conducted based on the research status and limitations identified in the preceding sections, including any particular methodology used, the objectives of the study, and the hypothesis. However, as it stands, this information is difficult to discern.
- [L112-114] This sentence seems misplaced in the introduction and appears more appropriate for the methods section.
- [L122] Explaining the research question design in connection with the hypothesis presented in the introduction would be beneficial.
- [L598-614] Finding a significant distinction between this paragraph and the results section is challenging. I recommend integrating this paragraph into the results section. Otherwise, it might be better to use this section to outline the overall direction of the discussion. For example, you could describe how the results will be used to substantiate each hypothesis.
- [L615-629] This section aims to prove the hypothesis that soils rich in organic matter have higher DOC concentrations. It explains this by linking to the differences in soil classification. However, a more detailed explanation of the unique characteristics of each soil class that could influence DOC content variations is needed.
- [L630-651] The content in this section is not being communicated clearly. It should be more concise and put the topic sentence in the head of the paragraph. Moreover, it might be better to establish discussion points that connect to the hypothesis and develop the logic accordingly.
- [L652-733] For clarity in this section, I recommend organizing the content in the following order: findings/claims based on the results of this study, supporting evidence from previous research, and the implications of this particular discussion point.
- There are additional detailed revision and improvement requests, but I will consider them after the major manuscript revisions have been addressed.
Citation: https://doi.org/10.5194/tc-2023-152-RC1 -
RC2: 'Comment on tc-2023-152', Anonymous Referee #2, 24 Nov 2023
reply
- General comments: I would like to thank the authors for this impressive effort in synthesizing and collating this nice database. The permafrost community needs a better understanding of the magnitude of lateral organic carbon export in the Arctic region and its significance for the climate feedback of thawing permafrost. In that regard, this study is timely and relevant. However, I have a few major concerns below before I can properly assess the quality of the manuscript.
My main concern is regarding the validity of the interpretation of DOC variability across ecosystem types because the authors, despite mentioning it, ignore the important effect of the various soil porewater extraction methods on the results. It is known that different extraction methods yield different soil porewater recovery rates with large effect on solute concentrations. These effects can’t be ignored. Otherwise, the authors would be overinterpreting the data. I suggest some possible additional analyses below to potentially sort this out.
In addition, the introduction is lacking a clear narrative which brings more confusion to the reader throughout the manuscript. While the authors try to explain why it is important to consider DOC and its export during permafrost thaw, they don't mention previous efforts in reporting soil DOC and current understanding of the variability in soil DOC, current state of the art etc. The aim of the paper only comes at l. 327, after mentioning the non-negligible effect of extraction methods, this is too late and should be included in the introduction. And again, the “assessment of methods” is a prerequisite to being able to “assess the concentration and mobilization of DOC in terrestrial permafrost ecosystems”.
In general, I also wonder whether this study falls under the “Review” or “Original research paper” type. I am missing a critical analysis of the available research and practices on terrestrial DOC in permafrost to date for it to be a proper review paper. The review is only partial and includes only the data collation aspect. On the other hand, there quite a lot of data analysis which could well fit under an original paper. This has implications for the title of the paper, where currently “A systematic review of…” should be removed from the title.
- Detailed Comments
- The novelty of the study is not just limited to what is stated l. 108-112, but to be able to better identify this novelty, a properly constructed introduction is required with reference to previous studies, e.g., Guo et al. 2020, Langeveldt et al 2020.
Previous syntheses on Permafrost carbon cycling mentioned DOC export as an important fluxes but it is still overlooked e.g.,:
https://www.annualreviews.org/doi/10.1146/annurev-environ-012220-011847
https://link.springer.com/article/10.1007/s13280-016-0872-8 - L. 94-99, it is true that the different sources of DOC will impact its biodegradability, however, the current dataset doesn’t allow to distinguish the different sources, this is, I believe, outside the scope of this study. First step is to describe and understand factors driving the natural variability in DOC, whether the current set of data is able to represent real variability or whether methodological artefacts are the main driver behind DOC variability.
- L. 114-117 where do the hypotheses come from? This is also where the introduction is lacking a clear narrative, the hypotheses are not introduced based on previous research findings, not anchored in current understanding. The reader needs some background.
- L. 122-127 a similar point to the previous point, why asking those questions? The introduction should be built around the various hypotheses and research questions so that the questions do not appear out of nowhere here.
- L. 215-226 and Fig. 3. In contrast to the other categorical variables displayed in Fig. 2 and introduced earlier, (i.e., ecoregions, soil classes, permafrost zones, thermal horizons), the ecosystem types are never introduced properly despite their central role in the data analysis later on, also in the PLS. To avoid future confusion and improve clarity, the different ecosystem types should be described, or the authors should point to relevant references where the reader can learn about the different ecosytems. Reference to Table S1 would be useful here and please add some description and relevant references for each ecosystem type.
- L. 210-214 I count only 10 extraction methods while l. 313 and onwards, 11 extraction methods are mentioned. Here again, some background information on the extraction methods is required, such as expected recovery, passive versus active, destructive or not, references to studies describing the methods. Include a table with references.
- Since the extraction methods can yield very different porewater recovery, e.g.,
https://www.sciencedirect.com/science/article/pii/S0883292706002289
https://www.sciencedirect.com/science/article/pii/S0038071705003111
https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/j.1365-2389.1996.tb01413.x
the effect of the extraction method cannot be ignored before analyzing for differences across ecosystem types etc. L. 312-325 is a good starting point and highlight the importance of “water-based versus solid (soil)” extraction methods which showed the highest ANOVA F score, among all reported ANOVA results. To me, this shows that extraction methods are more important than ecosystem types for explaining DOC variability. Am I wrong? This section must show whether the extraction methods are responsible for a larger share of the observed variability in DOC than other categorical variables, such as ecosystem types. If this is the case, then the remaining of the analysis must be adapted and e.g., performed by “extraction method”.
What happens to the findings in section 3.2 and 3.3 if the analysis is performed by “extraction method”. - L. 831-837 This information is coming way too late, this should start your results section so that you can probably rule out some of the issues I raise above. Need to be expanded and to describe how the extraction methods would impact observed DOC variability.
- L. 310 to 312, the median of the DOC grouped by filter size are quite different and to declare “We consider the effects of filter size to be minor” just after that, is a bit odd. The large variation in the medians here might in fact reflect some differences caused by extraction methods. Same comment for carbon measurement methods further below, if you cannot control for the extraction method, the values you report don’t have a proper sense. What if you group the data by extraction method?
- The reason to use the PLS is somewhat confused. What is the actual aim of the PLS? L. 241-243 it is mentioned that PLS is used to “assess the performance of continuous and categorical variables in predicting DOC concentrations”. But later, l. 464-480, the results of the PLS are used to better understand the variability in DOC, make links between processes and DOC variability. If the aim was to predict DOC, the authors should only use variables that are easily available at larger spatial scales, I wonder what the value is of predicting DOC with TDN and C:N ratio. In fact, the point is not to predict, but to investigate relationships to improve our process understanding.
In general, this section 2.4 is somewhat too dense and is a mixture of various statistical analyses whose aims are not completely clear. The authors should divide this section in sub-sections where, for each analysis, the objective is clearly stated. Here I am missing a relevant method to distinguish between the effect of DOC extraction method versus ecosystem types and other environmental variables in the DOC variability. - Fig. 5 How were the predictor variables selected? Figure S. 5 shows a complete PLS with all variables considered but is not referenced in the main text. The authors mentioned the “Variable Importance in Projections (VIP)” l. 244 but those values are not reported. Please add a table (eventually in the supplementary) where all variables have this “VIP” score and a description of how the variables were selected, was it only based on this VIP >1?
L. 248-249, why did the authors decide to split the ecosystem classes in 2? The answer might be along the lines 754-780 but it comes too late and it not linked. - L. 735-753 this section seems weak, where the authors end up saying that they cannot conclude based on too little data.
- BDOC data. L. 303-304 the authors mentioned that they consider later the smaller sample size during result interpretation, but this is not completely clear. Reading through l. 530-541, which is based on all BDOC data, including 3-days to 90 days incubation data, VERSUS l. 542-554, which is based on the highest BDOC values available for each ecosystem type, it is still not completely clear whether the same conclusions are found.
- Since there is a systematic increase with time, by ecosystem type in the BDOC data isn’t it possible to correct for the length of the incubation experiment in Fig. 6?
- Other comments
- Supplementary Table and Figures are named S1, S2, etc in the SI while they are referenced as A1, A2 in the main text. Some of the figures are not even referenced (this is also pointed out in some comments above and below).
- L. 112-114 this belongs to the methods.
- L. 238 how did you control for the month? Maybe this is what Fig. S2 is about, would need to be referenced.
- L. 282, why did you keep this reference then?
- L. 379-382 the description of the different classes would be better in the text (Methods section)
- L. 383-386 same here.
- L. 475-477 Have you tested that? “no clear or obvious trends in SoilC, TDN, C:N ratios, and SUVA across ecosystem types”
Citation: https://doi.org/10.5194/tc-2023-152-RC2 -
CC1: 'Comment on tc-2023-152', Tatiana Raudina, 01 Dec 2023
reply
General comments: I would like to thank the authors for attempting to systematize these databases. A lot of effort and time was spent collecting and processing them, and it is wonderful that they are considered in such kind of study (review article). However, I would like to leave some comments for possible improvement of accessibility of the data and their interpretation.
Detailed Comments
1. Line 215 - Sites were classified according to ecosystem type, and these included coastal tundra, forest, peatland, permafrost bog, permafrost wetland, retrogressive thaw slump, upland tundra, and Yedoma. Ecosystem classification is based on the general site description in the article, the provided ecosystem classification within the article, and site data including vegetation composition, permafrost conditions, and ecoregion.
I am wondering how peatland, permafrost bog, permafrost wetland were identified? A wetland is an area saturated with water and includes various aquatic ecosystems. In your research, you focused on terrestrial ecosystems. Was there some narrower sampling within the wetlands or were all heterogeneities and water bodies taken into account, including lakes, hollows, streams, etc.? Because this could also be the reason that permafrost wetland had the lowest DOC concentrations (7-10 mg L-1). Another question regarding peatlands. Were they permafrost-affected and where was the main location of these ecosystems? The authors that "Our goal was to assess the concentration and mobilization of DOC in terrestrial permafrost ecosystems", however, in Fig.1 there is only one peatland sampling location. For a better understanding, it would be helpful to provide clear definitions/descriptions of these ecosystems and what specific landscapes you included within them.
Accordingly, have you somehow considered the microtopography/spatial heterogeneity of your ecosystems (for example, mounds/fens/hollows/etc.)? Please note that even within the same type of ecosystems, DOC concentrations will differ by more than two times (e.g.: https://www.mdpi.com/2076-3263/9/7/291 or https://doi.org/10.5194/bg-14-3561-2017). It is not clear whether you averaged (or provided median values) for all the data within one ecosystem?
2. Line 209 - We also included the soil class found at the site (Histel, Histosol, Orthel, and Turbel; USDA, 1999) and whether the DOC was from the organic or mineral soil. You also mentioned that organic layer depth was included and that the highest DOC concentrations are found within organic-rich Histosol and Histel soils.
Have these variables (samples taken from mineral and organic horizons) been considered in the PLS? Since there is no information about these variables in the description of the results. Aren't they significant drivers of DOC concentrations? In general, the result of the PLS is not clear (Line 464). That is, the DOC concentrations are largely determined by the TDN and C:N ratio?
3. Line 312 - DOC concentrations were found to be significantly different between samples subject to the 11 different extraction methods used (ANOVA: F(10, 2515) = 21.8, p < 0.001), and between water based and soil (solid) based extraction methods (ANOVA: F(1, 2524) = 182.1, p < 0.001).
Indeed, the extraction method, sample subject location, as well as the size of the filter, have a large impact on the DOC concentration. Why then did you decide not to focus on extraction methods and filter sizes? After all, this is a review article and must take such important parameters into account. For a better understanding, it is worth explaining what specific extraction methods were used?
4. Line 403 - According to the text it follows that “The majority of permafrost wetland sample locations were found in Russia. However, Fig. 1 shows different information and most ecosystem types are permafrost bog. At the same time, most of the studies done in Russia, which you included to generate database deal with permafrost bog or permafrost peatlands (e.g.:
https://doi.org/https://doi.org/10.1016/j.chemgeo.2017.10.002
https://doi.org/10.1016/j.scitotenv.2018.04.059
https://doi.org/10.1016/j.chemosphere.2020.128953
https://doi.org/https://doi.org/10.1016/j.chemgeo.2013.07.016 and others.
It is also worth noting that there is not a single reference to studies from Western Siberia in the text (although Fig. 1 shows that quite a lot of DOC concentrations were taken from sites in Western Siberia).
5. Line 808 - Our results suggest that the high concentrations of DOC in permafrost bogs remains relatively stable upon thermokarst formation, although individual studies do indicate that thawing peat may provide a reactive source of DOC (Panneer Selvam et al., 2017).
How do you explain your results? Because the thawing of frozen peat can actually lead to a large release of DOC and macro- and microelements (e.g.:
https://www.sciencedirect.com/science/article/abs/pii/S0045653520331507
Citation: https://doi.org/10.5194/tc-2023-152-CC1 -
RC3: 'Comment on tc-2023-152', Anonymous Referee #3, 04 Dec 2023
reply
General Comments
I agree the study provides ‘unique and valuable insights’ into C dynamics in circumpolar north ecoregions. It pulls together an enormous database of DOC concentrations and fluxes from hundreds of sites to compare export and biodegradability of DOC from different ecosystems. The authors under-sell the scope and breadth of their review.
The assessment of sampling approach and DOC analyser method is also interesting. This, alongside the recommendations for study design in the discussion, will be a valuable resource for future studies
However, there are some issues to be addressed. The main aim of the study not clear, and the introduction doesn’t map on to the hypotheses or results, making it hard to determine the reasons behind the hypotheses. I am concerned that significant differences between filter size, extraction and measurement method were not incorporated into the analysis of DOC concentration, as there were clear impacts of these on the DOC concentrations that could explain differences between studies that are currently being attributed to ecosystem or another factor. The figure captions are so long it makes the figures feel really complex, whereas they are actually relatively simple figures showing clear results.
Specific Comments
Abstract
Line 24 – do you mean vast pools of water containing SOC (like peat pools/ponds) or vast stores of SOC?
Introduction
Line 56 – Mentioning RCP 8.5 here makes it seem like this is the only scenario in which there will be permafrost melt. Could you make it clear that there will be increased melt across a range of scenarios?
Line 61 – I think ‘additionally’ would be more appropriate than ‘alternatively’ here, as surely C will be lost via both pathways?
Line 83 – Do you mean ‘deepening’ here?
Line 101 – ‘top 3 m of terrestrial ecosystems’ – it isn’t clear what you mean here.
Line 116 – hypothesis 3 –there is nothing in the introduction that explains this hypothesis – why would DOC be most biodegradable from those ecosystems?
Methods
Line 138 – why was this additional search carried out on Google Scholar, when it was included in the original search?
Table 2 – can you include numbers of studies that were included after each screening stage? Or is this information included in the results?
Table 2 – several of these criteria seem quite similar to those in Table 1 (e.g. language, type of study), can you make it clear what was screened at each stage?
Line 185 – why did you remove DOC concentrations over 500 mg L-1?
Line 202 – why did you choose 20% as the cut-off point?
Line 215 – there are a lot of classification types and categories listed here – could you include more detail or definitions, or an example of a site classification to show how they work together? e.g. DOC sample from site X from continuous permafrost in Arctic tundra, on histosol soil, from permafrost-free horizon.
Line 239 – this definition of permafrost lens and active layer samples in amongst the data analysis is strange – could it be moved to the section above where the categories and classifications are introduced?
Results
Line 282 – why was the Olefeldt et al study included if it did not report DOC concentrations?
Line 291 – if the number of DOC mobilisation measurements is so low and therefore the results are not considered, why is a lot of the introduction about mobilisation? It made it seem like mobilisation would be a focus of the study.
Line 310 – this shows a significant effect of filter size on DOC concentration, yet you state “we consider the effects of filter size to be minor” without any justification. Please explain why you assessed the impact of filter size and then discount the result.
Line 312 – I understand that these results were not the focus of the study, however they are interesting and worth reporting. Could you include a table to show the number of DOC concentrations or studies by each filter size, extraction method and measurement methods? It could go in the supplementary information. I think it will help future studies choose which method to use, and be highly citable.
Line 344 – linking ecoregion to latitude – could you include an average (mean, median, whichever is appropriate) of the latitude of each ecoregion if you are linking that to DOC concentration?
Line 346 – could you standardise/shorten the way you report median, LQ and UQ? The current way leads to long chunks of text/numbers in brackets that break up the sentences and make it hard to follow. The n values for each permafrost zone, ecoregion, soil class and thermal horizon are in Figure 2, so you don’t need to put those in the brackets. Something like (58 (20-107) mg L-1) would be shorter than (n = 442; 58 mg L-1; LQ = 20 mg L-1; UQ = 107 mg L-1) as it is currently.
Figure 2 – this is good, clear way to represent database findings
Line 406 – Table S1 states there were 145 DOC concentrations from 9 studies in the Yeodoma ecosystem, whereas the text here says 118 DOC concentrations from 9 studies.
Line 427 – this paragraph is so difficult to read as there are so many numbers in the text.
Figure 4 caption – there are seven categories shown in the plot, out of eight total. In the caption, you state that you have not shown results for peatland or permafrost-free sites, which would make nine categories in total. Can you make sure you are consistent with how the ecosystem categories are reported?
Line 451 – could this go into the methods data analysis section? It is describing the analysis rather than results.
Line 468 – this sentence doesn’t make sense “The positive relationship between DOC and total dissolved nitrogen soil carbon content (SoilC)…”
Line 470 – Some of this is discussion rather than results. The sentence about aromatic content makes it sound like you had data of aromatic content to compare with SUVA values.
Line 476 – you mention figure A3 (which is called S3 in the SI) but not figure A2 anywhere in the text.
Line 489 – the figure captions are so long. I didn’t think the figures were that complicated, but the caption makes it seem much more complex.
Line 511 – mobilisation of DOC – I thought there wasn’t enough data to assess mobilisation?
Line 512 – I am confused by this section – it seems to be repeating results from section 3.2. Can you make it clear which characteristics/properties are assessed in each section?
Line 530 – these BDOC results are interesting, yet it feels like they are being buried in this section.
Discussion
Line 598 – again, referring to the top 3 m is not clear what you mean. Do you mean the top 3 m of soil? Were plants included in height?
Line 618 – were these studies included in your review?
Line 631 – were the thermal layers consistently deep across all sites?
Technical Corrections
Line 203 – brackets aren’t quite right here
Line 343 – “violin plots of both he discontinuous…” he should be “the”
Line 402 – should Table S1 be referenced here?
Figure 4 caption – “and (b) (b) the number…” too many (b).
Line 811 – “Whereas the database…” this sentence does not make sense.
Citation: https://doi.org/10.5194/tc-2023-152-RC3
Liam Heffernan et al.
Liam Heffernan et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 116 | 54 | 10 | 180 | 21 | 3 | 5 |
- HTML: 116
- PDF: 54
- XML: 10
- Total: 180
- Supplement: 21
- BibTeX: 3
- EndNote: 5
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
