the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 24 May 2024
Soil and forest floor carbon balance in drained and undrained hemiboreal peatland forests
Abstract. Drainage of organic soil is associated with increasing soil carbon (C) efflux, which is typically linked to losses in C stock. In previous studies, soil in drained peatland forests has been reported as both a C sink and source depending on, e.g., soil nutrient and moisture regimes. However, most of the earlier research was done in boreal sites, and the impact of soil moisture regime on soil C stock is likely to vary across different climatic conditions and ecosystems, depending further on vegetation. In this study, we examined the soil and forest floor (including ground vegetation) C balance in drained and undrained hemiboreal forests to evaluate drainage impact on C balance. A two-year study was conducted in 26 drained and undrained forest stands with nutrient-rich organic soil in the Baltic states (Estonia, Latvia, Lithuania). To assess the C balance, measurements of soil heterotrophic and total respiration were carried out, along with the evaluation of C influx into the soil through litter, including fine foliar litterfall, herbaceous ground vegetation, and fine roots of trees. The CO2 emissions did not significantly differ between the study countries; therefore, one emission factor can be applied to characterize soil emissions in the Baltic States. It was observed that C influx into the soil through litter can compensate for the C losses caused by heterotrophic soil respiration, and neither drained nor undrained soils were proven to be losing their C stock. Comparing the C balances in drained and undrained sites, it was found that drainage of organic soils reduces their C sequestration by 0.43±2.69 t C ha−1 year−1.
- Preprint
(1527 KB) - Metadata XML
-
Supplement
(4614 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-1397', Jens-Arne Subke, 14 Jun 2024
The manuscript by Butlers et al presents a data set of soil CO2 flux measurements obtained across 26 drained and undrained forested peatland sites in the Baltic states. Fluxes are measured using static chamber methods, and aim to inform the carbon balance according to main tree cover or geographical location (by country) between drained and undrained sites. Additional flux measurements on trenched subplots were intended to provide heterotrophic flux estimates to balance with C inputs, but these data are not used in the analysis.
I found the approach interesting, especially as it enabled a simultaneous analysis of drainage and tree species, which have been found to be linked to C losses in organic soils recently. The comparison by country was less engaging, especially as the main motivation was to define the Baltic region as a distinctive geoclimatic region that contrasts with boreal conditions.
The manuscript is very long, and I struggled with the intended focus of it. Rather than building the study around testable hypotheses, it presents a wide range of measurements that are not all relevant to the main objectives. I think that a considerable evaluation of much sharper objectives and specific hypotheses is needed. This will guide the analysis which currently includes measurements that are not used for analysis of findings and results in a wide-ranging, similarly unfocussed discussion.
Unfortunately, there were significant methodological issues that seriously hamper the interpretation of findings. Static chamber methods for CO2 flux measurements can be problematic, or require careful consideration of the concentration gradient over time. It is well described in the literature that a build-up of CO2 in the chamber space reduced diffusion-driven soil CO2 efflux, and these chambers were deployed for extended periods. A non-linear correction is hence likely to be needed, and there are numerous studies that describe methods to do so (see detailed comments below).
The static chamber method contrasts with an infrared analyser-=based chamber approach for flux measurements intrenched areas. These resulted in higher flux values than static chamber measurements, and are discarded on this basis alone. Small differences in temperature and moisture can not explain this discrepancy, and it is not clear why the IRGA based fluxes were deemed erroneous, and the static flux estimates assumed to be correct.
I also missed a more critical engagement with flux results. Maximum values observed are very low for summer conditions, whilst minimum values observed in winter are very high. The constrained range of value across the season is unusual, and authors don’t offer any suggestions as to why soil biological activity was maintained at considerable levels in the depth of winter with negligible photosynthetic C supply and likely frozen surface soils where most metabolic activity originates. I suggest that are careful re-analysis of flux estimates is needed, and a careful interrogation of flux responses with more than just temperature.
The conclusions have to be much more carefully considered, given the considerable sources of uncertainty that the authors present quite openly. The data set may be suitable to derive total soil CO2 fluxes following a re-analysis, and estimating heterotrophic contributions using literature values may also be informative. But sweeping statements regarding source or sink functions based on highly uncertain flux estimates and selective use of either boreal or temperate comparison values is not helpful.
In conclusion, I can not recommend the manuscript for publication in its present form. A slimmed down version with a clearer focus on testable hypotheses and a careful re-evaluation of flux estimates maybe worth considering, but this requires a comprehensive revision.
35: Other studies exist that present the C balance between heterotrophic C loss from peat and inputs from litter in forested peatlands (Hermans et al 2022).
54-79: This paragraph gives a lengthy account of technical considerations for C accounting under the IPCC. It does not focus sufficiently on the scientific background and goals f the research. Whilst I appreciate the consideration for harmonised protocols and potential of bias from using contrasting schemes, this should be referenced or presented much more concisely to maintain a focus on the advancement of peatland drainage understanding of C balances per se, not technicalities in its reporting.
84-86: The hypothesis is not statistically testable. Of course, consistent emission factors can be used (tailored or not), but there is no statistical method to accept or refute such a statement. Please present an actual hypothesis 9that can be phrased as a null hypothesis, i.e. is statistically testable). As it is presented, the objective seems to be to collect and p[resent emissions data that can be used in future analyses – why is it not being analysed or synthesised here?
104 (Table 1; small detail): Where n = 2, please just state the respective values separated by comma, rather than “…”, which implies a wider range of values.
120-121: The second sentence of the paragraph seems to repeat the exact information given in the first sentence.
122; Hutchinson and Livingston (1993) describe opaque chamber methods, but you should provide specific detail of your chamber dimensions.
126: Heterotrophic decomposition of soil organic matter (peat) is surely also included!
132: 30 or 60-minute sampling intervals will lead to significant build-up of CO2 in chambers with likely non-linear diffusion flux. Using a simple linear regression is likely to under-estimate flux values. See e.g. Kutzbach et al. (2007; Biogeosciences, 4(6), pp.1005-1025). The degree of under-estimation is likely to be dependent on the degree of concentration build-up (i.e. flux magnitude).
135: This is not right. An ECD can not detect CO2.
148: Please clarify if the heterotrophic respiration measurements were taken over the same periods and on the same days as the main flux measurements. Why is a different system used for these measurements?
212: Not clear: Rhet measurements were made on litter-free soil with no root influence, so you should use different C input values compared to Rtot calculations.
219-223: Predicting soil C output using abiotic drivers is not trivial. And you have to provide significantly more information regarding the underlying regression used. From Table S6, I gather that you used a linear temperature response, which is unusual, as there is abundant literature to show that respiration follows an exponential temperature response. Soil moisture is an additional factor, and its influence should be investigated in combination with temperature. Potential underestimation bias from the concentration build-up in chambers is likely to affect higher fluxes during warmer conditions more strongly than colder/low flux conditions, resulting in a more liner response.
319: You appear to treat each chamber measurement as an independent observation, but as the locations were identical for each plot, you should account for this temporal pseudo-replication by applying repeated-measures statistics.
340-344: Did you attempt to use an exponential response, rather than linear regression?
370-377: It is unclear why there should be any difference in the correlation between Rtot or Rhet and different soil parameters, as one is derived directly from the other, so correlations should be equally as strong – or in any case, they are not independent from one another. Figure S6 shows correlation results for sil depths of “0 – 30 cm”, but the text references soil depth of 20-30 cm only.
544-546: This is also very unclear. What are “removals” by NEE observed in hemiboreal zones? Provide a reference and make it clear if this refers to a higher rate of NPP. Higher NPP relates to pretty much all ecosystem components, not just litter. And finally: are “Rhet C loss” and “Rhet rates” not the same flux?
597-601: Rather than speculating about whether the public finds results controversial (where is the evidence or motivation for this statement?), you should present robust interpretation of what can be concluded. The figures you cite show fluxes not significantly different from zero, so whilst they don’t support findings of soils being a c source, you can also not present them as a C sink.
608-610: This is unclear. What is the assumed C stock 100 years ago for this assertion? And why do you apply the temperate emissions factor when otherwise comparing to boreal or hemiboreal conditions in the manuscript?
833 (Figure S4): The units on they-axis seem wrong as values can not represent fluxes in mg C m-2 h-1.
Citation: https://doi.org/10.5194/egusphere-2024-1397-RC1 -
RC2: 'Comment on egusphere-2024-1397', Anonymous Referee #2, 11 Jul 2024
Title: Soil and forest floor carbon balance in drained and undrained hemiboreal peatland forests
Author(s): Aldis Butlers et al.
MS No.: egusphere-2024-1397
MS type: Research article
General comments:
The paper is targeting the important question of CO2-emissions from drained forested peatlands and their potentially higher emissions than undrained forested peatlands. This is important in the context of GHG reporting (UNFCCC). They establish extensive systems for measurements during a 2-year period in a total of 26 sites (n=19 are drained) distributed to all three Baltic countries. They underline the uncertainty of the current default IPCC emission factor for the region and that the transition between temperate and boreal zone (hemiboreal) may be poorly represented by the current IPCC default. Thus, I very much welcome and acknowledge the effort and recognize its importance. Particularly in light of future policy demands on the LULUCF sector emissions/sequestrations and the need to increase Tier levels and enhance documentation.
It is my opinion that the paper still needs some work to be ready for publication. This relates to the clarity in methods, thorough discussion of uncertainties including reference to magnitudes and drivers found in other studies (Rtot, Rhet), the use of extensive soil chemical data and on the text and priorities of both introduction and discussion. See specific comments below.
Specific comments:
The introduction largely refers to IPCC guidelines and national scale emission estimates based on UNFCCC (national submissions). I miss an effort to link the study to existing research literature on what the important drivers have been found to be, if emissions are mainly climate-driven or if they have also been found to be driven by vegetation, history, geology/landscape etc. It is stated that earlier results are “inconclusive” but you should at least mention what other studies have expected to find and what was concluded. A paragraph on how forested peatlands in the Baltic region may differ from those used in the IPCC default EF as well as the potential methodologies would be interesting in this context. The study focuses on nutrient rich sites. While the study has an apparent aim to contribute to higher certainty for the UNFCCC reporting for the Baltic countries, it is not shown to what extent the selected sites are area-representative for the drained peatlands for which these countries need to report.
Methods should include the history of the sites studied (I don’t find this), particularly when they were drained (and perhaps drainage channels were maintained over time as a typical management activity through time), their LU/area characteristics before planting (if they were planted). What does one know in terms of expecting that these forested peatlands were similar to the undrained forested peatland that are included in the study? The undrained sites in general show a higher tree basal area than the drained sites – do the undrained sites represent sites that would have been selected to be drained historically? In the paper you mention several places the “effect” of drainage. I claim that you are not measuring an effect of drainage, but you are comparing (contrast) two types of forest with apparent different management over time – and most likely the drainage happened long ago.
Also, sites likely do vary a little in ground vegetation composition – it would be timely to have a clear description of ground vegetation as you use only some of the vegetation components in the balance calculations.
Method description of respiration (total, heterotrophic), litter input fluxes and C balances (forest floor, soil) need to be supplemented by a figure with the fluxes that are measured and estimated and how they are combined to calculate the soil and forest floor balances. I believe this will make the methods much more clear as well as shorten the text.
You observe that measured Rhet results seem unreliable, very high compared to Rtot and with a poor correlation to temperature r2 < 0.3) relative to Rtot (r2 ca. 0.7-0.8). In some context (fx. line 462) you mention they are found to be in error. In other context you claim that they are unlikely to be subject to measurement errors but mention their likely influence from decomposing roots (line 499), or the lack of temperature and moisture measurements that reflect the actual measurement position (poor correlation to temperature, unknown potential effects of moisture fluctuations). As far as I can see you i) describe the field measurement methods for Rhet in detail, ii) discard the results (line 214) for use in the C balances, iii) do not present them in the Results but iv) refer to them with correlations with Corg%, C:N, porosity or BD (line 436 and onward…however, it is unclear to me if your reference to Rhet here is to the measured Rhet or the Rhet used in the C balance calculations (eq 2)). I realize you wish to present openly to the reader what you have done and which problems you encountered. I feel the balance in the paper of this challenge is wrong. I think what I would do would be: include the field method description of Rhet in the main text if your results from these measurements are still helping you in your research aims. If not – I would move most of it to the supplement. I would use more effort when selecting the empirical relationship by selecting (reviewing) more than one. The chosen one is from boreal forest (you claim in the intro that hemiboreal EF is likely not represented by temperate EF, back up your choice of a Rhet estimation regression from a boreal study). If I understand correctly that you refer to measured Rhet in the discussion (correlations with Corg%, C:N, porosity or BD) I would like to know why you believe this is relevant given the likely effects of decomposing roots on measured Rhet. Given the clear effect that your decision has on not using the measured Rhet in C-balances I would like to see in the main text a figure with magnitudes and correlations for Rhet and Rtot (clarifying to the reader) and an opportunity to clearly state what you use the measured Rhet for and what not. What are the correlations to temperature and the magnitudes one can expect? And please use different abbreviations for the measured Rhet and the estimated Rhet from eq. 2. An uncertainty discussion should include the uncertainties inherent in the choice of using the eq2 (alternative Rhet).
Both results and discussion use considerable space on describing observed effects/relationships between fluxes and soil nutrient characteristics. I miss a much more clear direction on these tests and on the discussion of their results and this direction should be set in the introduction, preferably as specific research questions and/or hypotheses.
Application of statistical tests (methods) are not clearly described in terms of testing expected biological relationships; every test should be used for a clear purpose. Given the few sites some of the statistical methods rely on very few observations per strata. I think it would be better to limit analyses to the mixed model analyses. I find that the PCA analyses are not utilized to their potential – fx. one could use the PC vectors as explanatory variables and – if they express environmental variability that is possible to interpret in a meaningful way – they may help to find a pattern in how the many measured variables influence/drive emissions (example: Callesen et al. 2006. Growth of Beech, Oak, and Four Conifer Species Along a Soil Fertility Gradient. Baltic Forestry Vol. 12, No. 1 (22)).
The discussion starts by targeting the errors observed in one of the flux-methods (unclear if Rhet is the measured one or the one from eq. 2). Rather, I believe the discussion should start by referring to the results actually obtained in the study on the outcomes you have targeted in your study aim (probably the balance rather than any specific flux?). And target uncertainties on specific fluxes (parts of the balance) in later sections. As an example, in the very last sentence of the section on soil heterotrophic respiration interpretation you state that roots cut in the process of installation was most likely the major reason for your errors…if this is the most important contribution then you should start the section on soil heterotrophic respiration error-discussions with this. Also, I would expect that you would find several studies who have found a similar challenge and it would be timely to refer to such.
Technical corrections:
Line numbers:
43ish: refer to IPCC EF factor uncertainties
47: transition rather than “halfway”
56: resulting rather than corresponding
57: absolute rather than direct
58: field inventories or GHG inventories?
61: revise wording, something is wrong here.
70: [16]?...wrong formatting of reference?
79: add an “the”….check this in general throught the manuscript (or get someone to do it)
80ish: I think we would like to know more about the nature of peatland forests in the Baltics.
87: you touch only very little on this subject which in your case is important as I assume stand history, time since drainage etc is highly relevant. To what extent new drainage or ditch network maintenance is relevant in the Baltics is relevant I do not know but what you are comparing here (old drainage) is not of relevance in conditions where ditches are new or maintained to actively manage water levels. At least, you are not providing sufficient information on this.
98: variations…in what?
101: site WTL should be included in the main text/tables…
Table1: the … should be replaced by – (the normal way to describe a range). Or alternatively [1;2]
106-ca 112: replace with/add to table. The text seems a bit messy, long. Given the lengthy references I would try to move them to the supplement.
131: unclear to me what it means, reword.
135: how was uncertainty estimated?
139: how much data discarded?
167: probably measured instead of “assessed”
187: it took a long time before I know if you dealt with ground vegetation abo- as well as bel-ground. Also unclear on shrubs. Please be clear on this earlier (hence, the need for a clarifying figure of measured and estimated fluxes and the resulting balances).
228: it was a bit unclear if eq 2 was used for all sites?, or if you only discard measured Rhet on some sites?
Table 2: do the number cover both drained and undrained sites? As an average or?
240: as total? (and total)
245: the use of the word “litter” here makes it unclear what you mean. Do you mean the litter production or flux?...be careful – litter does not tell the reader if it is a flux or stock or above or belowground.
249: you are NOT measuring the impact of drainage (see also comments above)
250-251: “two approaches” is an unclear formulation. Reword with more clarity.
266: what do you mean by “descriptive evaluation”?, visual?...
Figure 2: add test results to the figure. The chosen gray tone does not come across well.
272 +: I am lost in terms of figuring out which statistical tests/methods you are applying and how you prioritize/include the many different variables (see also comment above).
295-ca 298: where do these analyses relate to you stated research aim, hypotheses etc. + check legend/content of the referred table S5. NB, if you refer to many figures in the supplement to describe results then they should probably be in the main text.
304: be aware of your use of “soil water” and “ground water”
301: figure 4:44? -> check throughout your figures that the numbering is correctly formatted.
Figure 5: do you define the country acronums somewhere? When referring to figure 5 just use 5a , 5b….no need for “panel a”
339: I would just report annual results (these are the ones interesting for you research questions or?). in case the instantaneous results tell you more about drivers and processes they can be used for that purpose..fx possible to find the maximum flux rates or other to characterize the site fluxes in other ways. If interesting.
366: if the PCA results are important then include them in the main ms and explain you method in detail in the methods section including your aim when using it.
398: i don’t see these results are linked to you research aims. Express you research aims more specifically to fit to you analyses and the study.
Figure 8: illustrate the balance and name them as you do in your methods/discussion etc…”soil”…”forest floor”. Use fx effux (positive sing) and inlux (negative time)..if in GHG convention. You don’t measure autotrofic respiration….so you shouldn’t include in the figure?
Discussion: see general and specific comments above.
621: is it confirmed with any degree of certainty? (which analyses did you base your conclusion on?)
628: have you done thorough uncertainty assessments for both methods that makes you able to conclude that they are equally certain/uncertain?
631: do not use the “impact of drainage” – you are not (as least as far as info given) dealing with fresh drainage but contrasting two types of forest management in a short moment in time (2 yrs).
Citation: https://doi.org/10.5194/egusphere-2024-1397-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 284 | 95 | 19 | 398 | 30 | 14 | 13 |
- HTML: 284
- PDF: 95
- XML: 19
- Total: 398
- Supplement: 30
- BibTeX: 14
- EndNote: 13
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1