From sedimentary production to water-column accumulation: temperature and spatial controls on littoral methane

Kiefel, Katharina J.; Peeters, Frank; Sepulveda-Jauregui, Armando; Ropella, L. Loraine; Martinez-Cruz, Karla

doi:10.5194/egusphere-2026-5425

Preprints

https://doi.org/10.5194/egusphere-2026-5425

Preprints

14 Sep 2026

| 14 Sep 2026

From sedimentary production to water-column accumulation: temperature and spatial controls on littoral methane

Katharina J. Kiefel, Frank Peeters, Armando Sepulveda-Jauregui, L. Loraine Ropella, and Karla Martinez-Cruz

Abstract. Littoral sediments are important sources of methane (CH₄) to lake water, but it remains unclear how strongly sedimentary CH₄ production controls sediment-water fluxes and dissolved CH₄ concentrations across seasons and spatially heterogeneous littoral zones. We measured sediment organic carbon, potential CH₄ production, sediment-water CH₄ fluxes, dissolved CH₄ concentrations, and stable carbon isotopes of CH₄ and CO₂ at seven littoral sites in Upper Lake Constance (ULC) and Lower Lake Constance (LLC) over four seasons. Potential CH₄ production, sediment-water CH₄ fluxes, and dissolved CH₄ concentrations were significantly related and showed similar temperature dependencies. Their apparent activation energies ranged from 0.81 to 1.11 eV and did not differ significantly. These results are consistent with a process chain from sedimentary production via sediment-water fluxes to dissolved CH₄, although the shared temperature dependence of the three variables means that their relationships do not, by themselves, demonstrate causal propagation along this chain. Spatial patterns, however, became weaker along this process chain. Within ULC, differences in potential CH₄ production were reflected in sediment-water fluxes but were no longer significant in dissolved CH₄ concentrations. Between the two lakes, dissolved CH₄ concentrations were significantly higher in LLC than in ULC, whereas potential CH₄ production and sediment-water CH₄ fluxes did not differ significantly. Potential CH₄ production increased with sediment organic carbon, suggesting that substrate availability may have contributed to spatial differences in CH₄ production. Carbon isotope patterns were consistent with a sedimentary, mainly acetoclastic CH₄source and indicated modification of part of the dissolved CH₄ pool after release from the sediment. However, oxidation alone could not explain the isotopic contrast between the lakes, suggesting an additional influence of lateral mixing and hydrodynamic transport. Our results suggest that sedimentary CH₄ production contributes to controlling littoral CH₄ supply, whereas oxidation and physical transport can weaken or redistribute this signal in the water column.

Received: 09 Sep 2026 – Discussion started: 14 Sep 2026

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Preprint (PDF, 1017 KB)

Supplement (583 KB)

Download & links

Katharina J. Kiefel, Frank Peeters, Armando Sepulveda-Jauregui, L. Loraine Ropella, and Karla Martinez-Cruz

Status: final response (author comments only)

RC1:
'Comment on egusphere-2026-5425', Anonymous Referee #1, 07 Oct 2026
Comments to “From sedimentary production to water-column accumulation: temperature and spatial controls on littoral methane” by Kiefer et al.
General comments
This study presents an extensive dataset on the potential production, transport, and concentration of methane (CH₄) in a connected two-lake system, with both spatial and seasonal resolution. The authors propose a plausible process chain linking sediment methane production, sediment-water CH₄ fluxes, and methane concentrations in the water column. This interpretation is supported by similarities in temperature dependencies and correlations among the measured variables. The statistical analyses are clearly presented and appropriately account for uncertainties in the observations. However, the interpretation remains somewhat inconclusive because three potentially important processes, namely the oxidation of CH₄, the lateral exchange with pelagic waters and oxic methane production (OMP), were not quantified. Consequently, it is difficult to assess to what extent the observed methane concentrations and their spatial and temporal variability can be attributed to the proposed sediment-based process chain.

Specific comments
The authors acknowledge the possibility that OMP influences CH₄ concentrations in the water column (l. 37). However, OMP is not considered further when discussing the drivers of CH₄ concentrations (e.g., ll. 28, 51, 68, 431, 451, and 508). Given that OMP could potentially contribute to the observed concentrations, I suggest that this process is considered consistently throughout the discussion.

Technical corrections
Please clearly distinguish between natural logarithms (ln) and logarithms to base 10 (log₁₀). This would improve the reproducibility of the analysis. For example, in Equation 4, log appears to be used to denote the natural logarithm. Is this convention used consistently throughout the manuscript? If so, I would recommend to use "ln" instead for clarity. (In the figures of the supplementary material, "ln" is already used for the natural logarithm.)

Figure 3 and 5: R²g and p values could be included in the facets to easily see the strength of the correlations.

l. 149: Please check the grammar of this sentence.

l. 278: Should “A” be replaced by “One”?

l. 359: The beginning of this sentence could be rephrased. In the context of the preceding sentence, the consecutive use of “By contrast” and “In contrast” is somewhat confusing.
Citation: https://doi.org/10.5194/egusphere-2026-5425-RC1
RC2:
'Comment on egusphere-2026-5425', Anonymous Referee #2, 12 Oct 2026
General Comment
This study combines potential CH4 production, sediment-water CH4 fluxes, dissolved CH4 concentrations and stable carbon isotopes across seven littoral sites and several seasons. Few studies bring these measurements together at this scale, and the dataset is a valuable contribution to understanding how sedimentary CH4 production relates to CH4 in lake water. Their central question, where the sediment signal fades on its way into the water column, is a useful way to look at littoral CH4 dynamics. I also appreciate the authors’ care in acknowledging the limits of their data.
However, some of the main conclusions are not yet fully supported by the analyses. In particular: (1) the regressions linking production, flux and dissolved CH4 cannot separate a causal chain from the shared response to temperature; (2) the claim that spatial differences weaken along the chain rests on comparing significant with non-significant results; and (3) some methodological details, such as the treatment of inorganic carbon in SOC and corrections for gas removed during incubations, could affect key results. Most of these points can be addressed with the existing data, and I believe doing so would strengthen the paper considerably. Detailed comments follow.
Major Comments
In Fig. 3, potential CH4 production was measured using sediment collected only in spring. The Arrhenius model was then used to predict production rates at temperatures observed during other sampling periods. This means that, for a given site, changes in predicted CH4 production across seasons are entirely determined by temperature.

Therefore, the strong relationship between predicted CH4 production and sediment-water CH4 flux in Fig. 3a may just reflect their shared response to temperature, rather than showing that higher production leads to higher flux.

The reported slopes also suggest this possibility. Based on the activation energies, the ratio of the fitted activation energies is approximately 0.94/0.81 = 1.16, very close to the reported slope of 1.14. Similarly, for dissolved CH4 against flux (Fig. 3b), the expected slope is 1.18, compared with the reported slope of 1.04.

So, Fig. 3 does not clearly provide an independent test of hypothesis (i) beyond the temperature relationships already shown by the Arrhenius models.

Authors could compare the site-specific values standardized to 20 °C (Fig. 4) across the seven sites. This would show whether sites with higher CH4 production also have higher fluxes and dissolved CH4 concentrations independent of seasonal temperature variation. The results in Section 3.3 already suggest that this relationship weakens at the dissolved CH4 stage.

For each site, sediment from five cores was pooled into one sample, and one slurry was divided into three incubation vials. These vials are technical replicates rather than independent field replicates. This means that the natural spatial variability in CH4 production within each site is not captured by the variation among vials.

The significant site differences in potential CH4 production (p ≤ 0.002, Section 3.3) may therefore overstate the statistical evidence for spatial differences. The authors should clearly distinguish technical from biological replication, explain how this affects the uncertainty estimates, and interpret the site comparisons more cautiously.

The conclusion that spatial differences weaken along the CH4 process chain is mainly based on significant site differences in production and flux, but non-significant differences in dissolved CH4 (Section 3.3). However, significant differences in one variable and non-significant differences in another do not necessarily mean that spatial variability has decreased.

Dissolved CH4 also has fewer replicate measurements than flux (3 vs. 8), and the Holm correction across 21 site comparisons may reduce the ability to detect differences. The authors can compare effect sizes and their confidence intervals or directly test whether between-site variability decreases along the process chain. This could support their interpretation of weakening of spatial variability.

Potential CH4 production was measured only using spring sediments and then extrapolated to other seasons using temperature. But substrate availability, microbial communities, and methanogenic capacity also changes seasonally. So the predicted production rates assume that the relationship measured in spring remains valid throughout the year. This assumption needs to be justified or explicitly acknowledged as a limitation, particularly when interpreting seasonal relationships between production, flux, and dissolved CH₄.

The Methods state that SOC was measured using a Shimadzu TOC analyser with a solid sample combustion unit, but do not explain how inorganic carbon was excluded. This is particularly relevant for Lake Constance, where carbonate-containing sediments may contribute substantially to total carbon. Could the authors clarify whether samples were acid-treated or inorganic carbon was measured separately and subtracted? This distinction is important because the SOC values are used to investigate the relationship between organic substrate availability and potential CH4 production (Fig. 5).

Section 2.4.1: 0.5 mL of gas is removed every 2–5 days over 30–120 days. Given each vial contained ~ 20 mL of headspace, repeated sampling could remove a substantial amount of gas over the incubation period. It is unclear whether the withdrawn gas was replaced (e.g., with nitrogen) or whether the CH4 removed during sampling was included in the production calculations. If neither was done, repeated sampling could affect the estimated production rates, particularly for longer incubations at lower temperatures, and potentially influence the estimated activation energy. The authors may have already accounted for this, but a brief clarification in the Methods would be helpful.

All three LLC sampling sites are located in Gnadensee, the most enclosed sub-basin of Lower Lake Constance. Therefore, the comparison between ULC and LLC is effectively a comparison between four ULC sites and three Gnadensee sites, rather than a comparison representative of both lakes.

With only 3-4 sites per lake, between-site variation estimated using REML may be uncertain. This can be acknowledged as a limitation and qualify that the observed differences, and the proposed effects of hydrodynamic mixing, may be specific to Gnadensee rather than LLC as a whole.

Additional Scientific and Methodological Comments
Abstract
Statements in the abstract that are stronger than the results support:
Potential CH4 production increased with sediment organic carbon, but this relationship was not statistically significant (p = 0.118, Section 3.3). It should be described as a positive trend rather than an established relationship.

A minor point: Abstract states that oxidation alone could not explain the isotopic differences between lakes. However, the Discussion acknowledges that the contributions of oxidation, bubble dissolution, and mixing cannot be clearly separated. It may be more accurate to say that other processes could also contribute, rather than ruling out oxidation alone.

The abstract is difficult to follow because it introduces several interpretations and then immediately qualifies them. Rather than adding further caveats, I suggest simplifying the abstract around the main findings and distinguishing observations from proposed mechanisms. For example, the shared temperature dependence of production, flux, and dissolved CH4 is an important result, but it does not need to be presented as evidence for a process chain and then immediately qualified. Similarly, the roles of oxidation and hydrodynamic transport could be presented as possible explanations without trying to establish which process is responsible. This would make the main findings clearer and avoid overstating what the analyses can demonstrate.

Introduction
Hypothesis (i) says dissolved CH4 is "primarily controlled" by the chain from sediment. Since the study is not measuring the relative importance of the other controls, “primarily” is an overpromise. Consider a more cautious wording that would match what the data can show.

The Introduction frames the study partly around the importance of large lakes (>100 km²), but only ULC meets this definition (472 km²). LLC is smaller (62 km²), and all LLC sites are in Gnadensee (~13 km²). The authors already note that the two lakes differ in size, morphology and hydrodynamics (lines 61–62), and they later use these differences to explain the higher dissolved CH4 in LLC (Section 4.4). I suggest making this contrast the main framing of the Introduction. This would turn the lake comparison into a question the study sets out to answer and would connect the Introduction more directly to the Discussion.

Minor note: The Introduction discusses several mechanisms influencing CH4 transport like wave-driven release, ebullition, plant-mediated transport, and macrophyte coverage. However, not all of these mechanisms are addressed by the study design. For example, sampling was deliberately conducted during calm conditions, and the enclosed-core measurements do not directly distinguish diffusion from ebullition. Macrophyte coverage is also introduced as a possible control but is not examined further.

Including these mechanisms is useful, but the Introduction could focus more on the processes the study actually investigates, while briefly acknowledging which transport pathways are not directly studied. This would create a clearer connection between the background literature, research questions, and methods. Similarly, CH4 production in oxygenated water is mentioned once but not discussed further. Since this could contribute to dissolved CH4 independently of sedimentary inputs, it would be helpful to briefly discuss its potential relevance to the study and how it might affect the interpretation of the proposed sediment-to-water CH4 process chain.

“Lakes cover 2.2% of global land area and emit approximately 53 Tg CH4 annually.” Saunois et al. (2025) reports emission estimates from both lakes and small natural ponds. Are there numbers reported specifically for Lakes?

Methods
For water samples, the headspace was nitrogen. For sediment, the 120 mL vials only had 5 mL of sediment, and the manuscript does not say whether the remaining space was flushed. If it was air, the CH4 and the CO2 in air could affect the isotope values. This matters for the comparison between sediment and water isotopes and will help make the method become reproducible.

Section 2.2 describes how headspace CH4 was collected and measured, but it is unclear how these measurements were converted to dissolved CH4 concentrations in the original water samples. The authors provide the sample volumes and initial pressure, but not the calculation or solubility correction used. Since the bottles contained a concentrated NaCl solution, it would be helpful to clarify how the effect of salt on CH4 solubility was accounted for. A brief description or equation would make the method easier to reproduce.

Minor: The Arrhenius models assume a common activation energy across all 7 sites, allowing only the intercepts to change. This assumption could influence the site-specific estimates at 20 °C (Fig. 4) and their uncertainty. It would be helpful to check whether temperature sensitivities differ among sites, for example, by testing a site × temperature interaction, or to briefly justify the use of a common activation energy.

Results
Although Figures 2 and 4 show the measured and estimated CH4 variables in physical units, a brief numerical summary of their ranges or typical values for each lake would make comparisons with other studies easier.

Potential CH4 production is reported per unit volume of sediment, while sediment-water CH4 flux is reported per unit area. Multiplying production by the assumed active sediment depth (5 cm) would give an approximate comparison of the two rates. Although potential production measured in lab does not represent actual field observation, this comparison could help check if the estimated production capacity is consistent with the measured fluxes. The authors already have the data needed for this comparison, which could provide additional context for interpreting the proposed process chain.

In ULC, dissolved CH4 decreased significantly from autumn to winter, while sediment-water flux did not show a significant change. In LLC, flux decreased significantly from summer to autumn, while dissolved CH4 did not change significantly. These differences in seasonal patterns suggest that dissolved CH4 concentrations may not always closely follow sediment-water fluxes, and that other processes, such as oxidation or mixing, may influence them. This observation is relevant to hypothesis (i) and could be highlighted in the Results and discussed further in the Discussion.

Each Arrhenius model includes seven site-specific intercepts alongside temperature. It would be useful to clarify how much of the reported variance explained (93%, 53%, and 80%) is attributable to temperature versus differences among sites.

Sediment α_app values range from 1.034 to 1.075, spanning values associated with both acetoclastic and hydrogenotrophic methanogenesis. Given this range, it would be helpful to explain more clearly what supports the interpretation that acetoclastic methanogenesis dominates. The authors could also discuss whether the variability in α_app suggests differences in methanogenic pathways among sites or seasons.

To improve readability: The paragraph on site differences in sediment isotopes lists many pairwise results, and the pattern changes each season. With three cores per site, some of these may be noise. A summary would be clearer, and it would help to say whether the isotope patterns line up with the production pattern (SIP and MUN high).

Discussion
Sections 4.1 and 4.4 offer different explanations for the observations in LLC. Section 4.1 suggests that greater oxidation could contribute to the more enriched water-column CH4 isotopes, while Section 4.4 suggests that weaker lateral mixing could contribute to higher dissolved CH4 concentrations. These processes could interact, but they are discussed separately. It might be useful to consider them together and discuss how oxidation and physical transport together might affect CH4 concentrations and isotopic composition in the two lake settings.

Section 4.1 describes the “clearest oxidation signal” as occurring in winter. Although the Results examine seasonal differences in isotopic composition, it is unclear what specifically supports the interpretation that oxidation was most evident in winter. This is interesting because lower temperatures can reduce microbial oxidation rates, although other environmental factors also influence oxidation. The authors could clarify the evidence supporting this interpretation and discuss possible explanations for the winter pattern.

In the Introduction the importance of littoral methane cycling for whole-lake CH4 budgets and the potential effects of warming are the motivation for the study. But the Discussion does not clearly return to these broader questions. For example, what do the observed differences between sedimentary production, sediment-water flux, and dissolved CH4 mean for estimating lake methane emissions? What can the measured temperature sensitivities tell us about potential responses to warming, and where are their limitations?

It would strengthen the Discussion to connect the main findings back to these broader questions, while considering how the sampling design (calm-weather conditions, one year of observations, and enclosed-core flux measurements) affects their interpretation at larger spatial and temporal scales. A brief discussion of these implications would help connect the study's motivation to its conclusions.

It is interesting that the lab and field activation energies are similar since in the lab, only temperature changes but in the field, temperature moves together with season, so light, plant growth, fresh organic matter and mixing are all bundled into the field values for flux and dissolved CH4 . The match in activation energies could partly be coincidence. If possible, the authors should also compare their values directly with published ones.

Section 4.2 discusses the role of pore-water CH4 concentration gradients in driving sediment-water fluxes. Pore-water samples from 0–5 cm and 5–20 cm were analysed for stable isotopes using an instrument that also measures CH4 concentrations. If reliable concentration data are available from these measurements, they could provide additional insight into the relationship between pore-water CH4 and sediment-water fluxes across sites and seasons. This could help strengthen the proposed link between sedimentary CH4 production and flux.

To improve readability: The Discussion starts with isotopes, which come last in the results and are not part of the hypotheses. Following the hypothesis order (chain, temperature, space, then isotopes) would be easier to read.

Technical Corrections
Line 108: The text says wind data came from "the DWD weather station in Constance", but the reference is HOSTRADA, a gridded product, not station data. One of them should be corrected.

Lines 118, 128: "−0.9 ± 0.1 bar" should be stated as relative to ambient pressure.

Line 123: "(1s)" should be "(1σ)".

Lines 125 and L207: "Matlab" and "MATLAB" should be spelled the same way.

Line 149: "from randomly five sediment cores" should be "from five randomly selected cores".

Lines 159–160: "except that the vials were shaken for 1 minute": the water samples were also shaken for one minute, so the "except" does not make sense.

Line 195: "Belenmite" and "VDPB" should be "Belemnite" and "VPDB".

Lines 218 and L226: "kB" and "kb" should match.

Eq. 4: S Is the "log" the natural logarithm?

“Section 2.6.1.” ends with a period

Fig. 6 caption: "The hulls illustrate the observed … fractionation pathways": the hulls are fields from the literature (Whiticar, 1999), not observations. "Typical" would be more accurate than "observed".

"Site" and "station" are used interchangeably, choosing one will improve readability.

Citation order in brackets: sometimes alphabetical (Line 459), sometimes by date (Lines 47–48), sometimes neither (Line 490).

"&" and "and" in citations: "Peeters & Hofmann" in some places and "Peeters and Hofmann" or "D'Ambrosio and Harrison" in others.

References
Jansen et al. (2022) and McGinnis et al. (2006) use a different style than the rest.

Capitalisation of titles is inconsistent, for example Bertolet et al. (title case) against most others (sentence case), and "Journal of chemical and engineering data" (Wiesenburg).
Citation: https://doi.org/10.5194/egusphere-2026-5425-RC2

Katharina J. Kiefel, Frank Peeters, Armando Sepulveda-Jauregui, L. Loraine Ropella, and Karla Martinez-Cruz

Supplement

https://doi.org/10.5194/egusphere-2026-5425-supplement

Katharina J. Kiefel, Frank Peeters, Armando Sepulveda-Jauregui, L. Loraine Ropella, and Karla Martinez-Cruz

Viewed

Total article views: 531 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
339	133	59	531	59	71	57

HTML: 339
PDF: 133
XML: 59
Total: 531
Supplement: 59
BibTeX: 71
EndNote: 57

Views and downloads (calculated since 14 Sep 2026)

Month	HTML	PDF	XML	Total
Sep 2026	218	76	21	315
Oct 2026	121	57	38	216

Cumulative views and downloads (calculated since 14 Sep 2026)

Month	HTML	PDF	XML	Total
Sep 2026	218	76	21	315
Oct 2026	121	57	38	216

Viewed (geographical distribution)

Total article views: 531 (including HTML, PDF, and XML) Thereof 531 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 29 Oct 2026

Short summary

Methane is a powerful greenhouse gas, yet its movement from lake sediments into the water is not fully understood. We measured methane production, release from sediments, and accumulation in shallow waters at seven sites across four seasons in Lake Constance. All three increased strongly with temperature and were closely linked. Local differences weakened after methane entered the water, suggesting that mixing and methane removal help shape where it accumulates.


Total:	0
HTML:	0
PDF:	0
XML:	0