the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Where can rewetting of forested peatland reduce extreme flows?
Abstract. Historical drainage to improve forestry practices has resulted in 0.6–0.7 million hectares drained forested peatland in Sweden. This has reduced the storage of water in the landscape and may impact greenhouse gas emissions, biodiversity and the damping of extreme water flows. National restoration actions therefore aim at rewetting 0.1 million hectares of forested peatland in Sweden, despite the limited and sometimes contradictory evidence in the impacts of rewetting. To clarify the potential impact on extreme flows and their cause-effects relationships from rewetting, we simulated flow under various conditions of the climate, local hydrology and rewetting practices (ditch blocking alone or combined with reduced tree cover). For this, we used the HYPE model setup across Sweden (450 000 km2) with improved calculations of runoff in drained forest and routines for inflow and outflow regions. National evaluation of changes in discharge extremes was combined with a detailed study in south-east Sweden, with the aim to understand rewetting impacts at various scales. We found that the change in discharge extremes from catchments of 10 km2 is small, because there is considerable mixing with runoff from various landcover. Hence, at the larger scale, rewetting is not an efficient measure to combat droughts or floods. However, for ecosystems in the streams only draining peatlands, rewetting can have an impact if appropriate sites for restoration are selected. The results show that groundwater level prior to rewetting and reduced tree cover are governing the effect on water runoff. Wetland allocation and management practices are thus crucial if the purpose is to reduce flow extremes in peatland streams.
- Preprint
(2490 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on hess-2024-271', Anonymous Referee #1, 30 Oct 2024
General comments
Rewetting of peatland does impact greenhouse gas emissions, biodiversity and extreme low and high flows. The authors investigated the effect of rewetting on high and low flows in Sweden. They employ an hydrological model (S-HYPE), which has been calibrated across Sweden. In addition to the national scale, they performed a sensitivity analysis with S-HYPE in a part of Sweden to determine which factors determine the hydrologic effect of rewetting. They show that at a larger scale (> 10km2) rewetting has hardly any impact on extremely high and low flows. At a smaller scale (< 10km2) the effects are mainly determined by the groundwater levels before rewetting and the reduction of tree cover after rewetting.
The current insecurities regarding rewetting policies, extensive distributed modelling of rewetting and the systematic sensitivity analysis make the study very relevant to readers of Hydrology and Earth System Sciences. The modeling study is novel and allows them to answer their research questions. I especially appreciate Figures 7 and 9, which explain clearly why rewetting does not always result in higher low flows or larger high flows.
In general, the paper reads fluently and is well structured. Below I list several specific comments which would strengthen the paper.
Specific comments
Line 39: The term “air-water holding capacity” is not common among hydrologists. Please explain what you mean.
Line 94: “whereas peat soils cover 17 % of the entire surface and 7 % of forests”. Probably you mean that 7% of the entire surface is forested peat soil. This is not clear in your sentence.
Line 115: “following 5 years of initialization”. Do you mean that the initialization occurred in the years 2007-2011? Clarify in the text.
Line 119: “these HRUs are described using three soil layers extending to 0.25 m, 0.7 m and 1.5 or 2.25 m below the soil surface”. In the next paragraph, you describe the land cover data. Mention also the source of your soil profile data.
Line 142: Table 2 should be Table 1.
Line 146: In the baseline scenario you assumed a ditch depth of 0.7 m. As ditch depth is important in the sensitivity analysis, what is the base of 0.7 m?
Lines 163-167: You show that the change of tree density after rewetting has a significant impact on interception and evapotranspiration and thus on high and low flows. As your results entirely depend on modeling, we should be sure that sound modeling concepts are used for interception and evapotranspiration. Describe the used concepts for interception and evapotranspiration and show how reliable the concepts are.
Line 163 and Table 2: You discuss the reduction in tree cover and its effect on model parameters. How large is the reduction in tree cover that you have in mind?
Figure 2: The amount of information is relatively small in relation to the space it takes. You can explain the main result in the text and, if needed, move the Figure to Supplementary Material.
Figure 9: The difference is unclear between the sub-figures right-center, left-bottom and right-bottom. Do you need these 3 subfigures?
Table 3 and Figure 3: You apply S-HYPE with (Scen. A and B) and without (Scen. G and H) regional calibration. Figure 3 shows that the impact of regional calibration is relatively limited. Mention this when you discuss Figure 3.
Lines 364-368: This sentence is too long. Split the sentence.
Line 389: This is the conclusion section. Before you discuss the impact for policymakers and field research, answer the main research questions of your study:
- What are the main drivers behind the heterogenous impacts of rewetting on discharge extremes?
- Where can rewetting of forested peatland reduce extreme flows?
Line 399: Be more specific on implications for other ecosystem services and risks.
Citation: https://doi.org/10.5194/hess-2024-271-RC1 -
RC2: 'Comment on hess-2024-271', Anonymous Referee #2, 06 Feb 2025
This paper carries out a modelling experiment (using the SMHI HYPE model set up for Sweden: S-HYPE) to assess the effectiveness of forest peatland restoration in buffering low flows and high flows. This is achieved by various semi-distributed modelling scenarios based on new data on the distribution of ditches in Sweden and various assumptions about ditch depth and effectiveness to establish baseline simulations. Adjustments to model parameterisation to mimic the various hydrological process impacts of peatland restoration in relevant spatial domains. The modelling concluded that peatland restoration will not have any significant benefits in terms of mitigating high flows or enhancing low flows. This partly reflects the influence of non-drained areas in larger catchments and the fact that many of the effects of peatland drainage are irreversible, at least on normal policy decision making time scales.
This is important work that is likely to have a large readership. Currently large sums of money are being spent globally on peatland restoration with – often overly optimistic - expectations that benefits will accrue in terms of mitigating hydrological extremes and enhancing carbon sequestration. As the authors correctly argue, modelling experiments like that presented in this paper have an important role here, where much experimental work is very small-scale and can yield contradictory findings from often short-term monitoring. So, I find the work both novel and interesting.
That said, in its current form, it impossible to evaluate the results presented. Consequently, I could not yet support publication in HESS for the reasons presented below.
- Lack of information about HYPE and its use in this application: The paper provides very little information on the HYPE model. I am aware that it is a well-known model, but even as someone with extensive experience in hydrological modelling, I had to read the original Lindstrom et al. (2010) paper to understand the model sufficiently well to have a basic appreciation of how it functions and how it has been adapted for this application. Many interested readers of the paper will likely be experimentalists who have very limited modelling knowledge. Clearly the paper should be sufficiently “stand-alone” that other papers don’t have to be read to have a basic understanding of the work. I strongly recommend as a minimum a basic description of the model work flow and a conceptual diagram of the model and its basic parameterisation (in terms of stores and fluxes) showing how it was adapted to mimic the effects of drainage/peatland restoration. Otherwise, it is impossible for anyone other than an experienced HYPE user to know what has been done.
- No information on model calibration: Virtually no information is given in the text on how the model was calibrated or what the associated uncertainties in predictions are. We are told that the S-HYPE model was calibrated (though the reader is not told how – just referred to another paper pre-dating the calibration period) that for 2006-2020 with a mean NSE for Q of 0.79 and a small volume error. No information is given in the text on efficiency of low flow predictions (despite them being central to the objectives of the study). Many will see this as being a critical flaw in the work as we have no means of evaluating whether the model is fit for purpose or not.
- No evaluation of model uncertainty: Given the inevitable uncertainties in the modelling: a mean NSE for Q of 0.79 sounds quite good but is suggesting high uncertainties at least for some sites for high flow predictions alone. Consequently, many of the very small percentage changes simulated for the various scenarios in Table 3 are likely within the uncertainty of the model and may not be significant. The authors need to provide more of the evidence they have on why they think the modelling overcomes these issues. Certainly statements like “Rewetting…..cannot help improve water security… …in catchments >10km2” (L364) are strong and unambiguous, yet the reader has no means of assessing their validity on the basis of the information presented in the paper. This also seems potentially dangerous given the policy relevance of such findings.
Setting aside these more fundamental criticisms about the presentation of the modelling itself, the paper is quite well-written and the quality of the graphics is good. I am also supportive of the general approach; though I think the title should make clear that this is a modelling experiment. As it stands, many are likely to expect a more empirical data-driven study. Related to this, the authors should remain aware throughout that these are modelling results – not data! Despite the preceding comments, I would be inclined to agree with the study’s conclusions. However, because it is such an important near-policy topic, the modelling really needs to be presented in a much more robust way and show more critical thinking in terms of interpreting the results with model uncertainties in mind.
I therefore hope in revision that the authors provide the additional information requested and present their important work in a more robust way that will increase its impact.
Citation: https://doi.org/10.5194/hess-2024-271-RC2
Data sets
Research data for the article Elenius et al, 'Where can rewetting of forested peatland reduce extreme flows?' Maria Elenius et al. https://doi.org/10.5281/zenodo.13472209
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
303 | 53 | 179 | 535 | 9 | 8 |
- HTML: 303
- PDF: 53
- XML: 179
- Total: 535
- BibTeX: 9
- EndNote: 8
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1