the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 16 Jun 2026
Speleothem sulphate and trace elements constrain Alpine glacial inception across the Marine Isotope Stage 11/10 transition
Abstract. Alpine glacier histories beyond the Last Glacial Maximum are poorly resolved, impeding our ability to test how mountain glaciers respond to dynamic forcings on orbital to sub‑orbital timescales. This gap may be addressed through proxy records from alpine caves, in which subglacial speleothem growth is facilitated by recharge beneath temperate glaciers through sulphide-bearing epikarst. However, the use of conventional stable isotopes of carbon (δ13C) and oxygen (δ18O) in speleothem calcite yields ambiguities when interpreting environmental transitions and ice cover during glacial inception and retreat. Herein, we present a multiproxy speleothem record from Betten Cave (Melchsee Frutt, central Swiss Alps) spanning 415–360 ka that integrates sulphate stable isotopes (δ34SSO4, δ18OSO4) and trace‑element geochemistry with calcite δ13C–δ18O to diagnose redox state, sulphide‑oxidation pathways, and hydrological reorganization across the Marine Isotope Stage 11/10 glacial inception. Three environmental phases are identified from this dataset, marking the transition from (1) a vegetated and soil-covered montane valley to (2) a soil-limited periglacial setting, hydrologically influenced by glacier advance through the adjacent valley (~2,000 m a.s.l.), to (3) subglacial speleothem growth, in which a temperature glacier covered the lowest elevation of the cave system (~1,700 m a.s.l.). After 402 ± 4 ka, a long-term decrease in the isotopic offset between water and aqueous sulphate (Δδ18OSO4−H2O) records progressive oxygen limitation concomitant with glacier thickening and advance over the cave site by 372 ± 3 ka. Coeval peaks in cations and redox‑sensitive transition metals reflect the enhanced delivery of glacially comminuted detritus and coupled Fe–Mn redox cycling within the subglacial karst system. These results directly link sulphate oxygen- and sulphur-isotope systematics and trace‑element fingerprints to glacier dynamics, providing geochronologically precise benchmarks and a transferable framework for reconstructing mountain glacier behaviour where geomorphic records are incomplete.
- Preprint
(11744 KB) - Metadata XML
-
Supplement
(20492 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-2840', Anonymous Referee #1, 22 Jul 2026
-
AC1: 'Reply on RC1', Jonathan Baker, 09 Oct 2026
Reviewer comments are in bold text. Author replies are in unformatted text.
RC1: This manuscript presents a nice dataset obtained from the recently published stalagmites from Betten cave in the Swiss Alps (Honiat et al., 2026). The new proxies (isotopes in speleothem sulphate and trace elements) allow to interpret the environment associated with the speleothem formation in the studied time interval (transition MIS 11-MIS 10) and shed some light on the mechanisms related to subglacial speleothem precipitation. This is a very interesting study, well-written, original and that will help to further studies in this research line in other mountain ranges (eg. Pyrenees). I recommend publication, I just have some minor comments
Response: We thank you for your positive feedback and for the constructive suggestions below, which we have addressed and believe will improve the clarity of the manuscript.
- Line 37: temperature glacier – is it correct or a typo? I think it has to be changed to: temperate or warm-based glacier
Response: This was a typo and has been corrected to “temperate glacier”
- Line 64: in those depressions filled by lakes, is there any study on lake sediments that merits citation? It may be another type of continuous record to fill the gap!
Response: There are abundant studies from alpine lakes documenting Holocene environments, but the preservation potential of high-elevation depressions is small beyond the Late Glacial and extremely small prior to the Last Glacial Maximum (one rare exception is MIS 2 and Late Glacial deposition in the Central Alps at 2,900 m asl; Guglielmin et al., 2026, doi:10.1016/j.catena.2026.110032). Limited examples from MIS 3-4 include paleolake deposits in inner-alpine valleys (e.g., Barret et al., 2018, doi:10.1017/qua.2018.26; conference presentation by Spitaler et al., 2026, doi:10.5194/egusphere-egu26-20278). Therefore, we hope that our study would provide a benchmark, if future studies were to document lacustrine or other deposition corresponding to MIS 10–11 in a high-elevation setting sensitive to glaciation.
- Line 86: this paragraph about Paleoclimate records in alpine karst may be included in the following 1.3 about subglacial speleothems to simplify the introduction.
Response: We agree and have merged the two sections, retaining the Sect. 1.3 title.
- Line 107: change deposition by formation or by precipitation
Response: We have changed deposition to “precipitation”
- Line 172: I wonder if any cosmogenic date or OSL date are available in the Alps for this MIS 10 glaciation. If so, it would be important to reflect it in the text and in the map of Figure 1.
Response: Lateral moraines associated with the MIS 10 glaciations are identified pedostratigraphically in foreland deposits of the southern margin of the Western Alps (Gianotti et al., 2015; Monegato et al., 2023, doi:10.3929/ethz-b-000653791) but without radiometric constraints. Deposition in the Upper Rhine Graben of Northern Switzerland putatively records gravels associated with MIS 10, bounded by OSL dating of interglacial sands with ages of 351 ± 31 ka and 470 ± 41 ka (Preusser et al., 2021, doi:10.1007/s00531-021-02043-7). The latter work is cited already, but since the OSL ages correspond to deposits that stratigraphically bound MIS 10, we did not discuss them in detail. However, we would gladly consider your suggestion of including the locations of rare MIS 10 foreland deposits in Figure 1.
- Line 195: Since this methodology is quite novel (or at least it is not the usual one in speleothems) I suggest you to include a step by step procedure in the supplementary to help the readers in further studies.
Response: We agree, the methodology could be expanded to guide future studies. Since our Methods section is already compact, we suggest that a few additional paragraphs should not detrimentally impact the flow of the text. We can further cite Baker et al. (2026, doi:10.1038/s43247-026-03942-w), which has been published since our original submission and provides complementary details supporting the methodology.
- Results: to me, Figure 3 should appear earlier in the manuscript when trace element profiles are described. As it is now, the reader gets lost in the Result section since the figures with the trace elements are in the supplementary; this fact makes the reading uncomfortable and unclear. I suggest to include at least one figure with trace element data in the main text or place Figure 3 earlier in the text.
Response: We highly appreciate this feedback. In our opinion, especially in consideration of comments from Reviewer 2, it could be more efficient to split the datasets presented in Figure 3 into two figures, which may also incorporate concepts from Figure S4. This solution would allow us to present some of the trace-element profiles earlier in the main text and prevent readers from depending too highly on having to reference the Supplementary Materials.
- I like Figure 5, with the explanation of the three phases on the landscape, well-illustrated and easy to understand the different signals in the different stalagmites from the same cave. I am just not totally sure if this figure will be clearer if removing the satellite image and using an elevation model. Just a suggestion to check. Regarding Figure 6, I suggest increasing the size (of the figure and, specially, of the text, it is really small).
Response: Thank you for these positive words and suggestions. Since the elevation profile is presented in the righthand subpanels, we thought it more informative to provide the satellite imagery (depicting modern vegetation cover) on the lefthand subpanels. Although a DEM admittedly offers some advantages for interpretation of the topography and glacial landforms, the current form of the figure also allows readers to easily orient themselves in software like Google Earth to explore the landscape in more detail. We are therefore not opposed to testing a version without the satellite imagery but prefer to keep the presentation as is.
Finally, we agree that labels in Figure 6 should be larger and can easily address this detail in revision.
Citation: https://doi.org/10.5194/egusphere-2026-2840-AC1
-
AC1: 'Reply on RC1', Jonathan Baker, 09 Oct 2026
-
RC2: 'Comment on egusphere-2026-2840', Anonymous Referee #2, 07 Aug 2026
Comments on egusphere-2026-2840:
“Speleothem sulphate and trace elements constrain Alpine glacial inception across the Marine Isotope Stage 11/10 transition” by Baker et al.
The authors present new records of sulphate stable isotopes (d34S and d18O) and trace element data from four stalagmites from Betten Cave (Melchsee-Frutt, central Swiss Alps) covering the period from 415-360 ka. Using a multi-proxy approach including previously published calcite d18O and d13C data from the same stalagmites (Honiat et al. 2026), the authors present a reconstruction of the timing and extent of progressive glaciation across the MIS 11/10 glacial inception depicted in Fig. 5.
General comment:
The multitude of proxies and their interrelations make this paper quite complex and demanding, and I had a hard time to get through and really understand it. I think there is potential to improve the manuscript in a way to make it easier for the reader to follow.
Comments on content:
In the following I will focus on a couple of specific aspects that I believe are substantial weak points related to data analysis and interpretation:
Line 247:
The authors write: “Stable-isotope records are largely reproducible between intervals of synchronous growth” and refer to Honiat et al. 2026. This statement is fair for the d18Occ records but looking at the d13C records in the original paper, it is evident that the d13C record of sample 60 is in antiphase with the corresponding records of samples 13 and 15D in the period between 381-372 ka. Is this an issue of age uncertainties only?
Line 268:
The authors write that they converted calcite d18Occ into drip water d18O values. How was this conversion accomplished? The reconstruction of drip water d18O from d18Occ requires knowledge of the cave temperature and an appropriate calibration of the water-calcite oxygen isotope fractionation. None of this is mentioned in the text.
Since drip water d18O is subsequently employed to calculate Δd18OSO4-H2O that is then used as environmental proxy, a proper description of the d18Occ conversion into drip water d18O is indispensable.
Alternatively, fluid inclusion d18O could be measured directly. Based on the stalagmite images shown in Fig. S2, I guess there should be enough water in the samples.
Line 371: Section 4.2
The discussion provided in section 4.2 is very interesting and informative, but in my opinion, it is based on a misinterpretation of the underlying data.
Looking at the data records in Fig. 2a and Fig. 3, d13C is the only proxy that seems to provide evidence for an environmental change from interglacial to periglacial conditions, indicated by a distinctive, continuous increase from about 0 to +3 ‰ between 406 and 402 ka in sample 61D. Other proxy records like d34S, UAR(i), Na+K+Al, U, and Ni, in contrast, show clear discontinuities at 402 ka that are obviously related to the change from sample 61D to stalagmite 13. As the two samples originate from distant parts of the cave with 300 m altitude difference, it appears more likely that the observed “abrupt changes” are caused by site-specific differences of the karst bedrock rather than to the proposed transition to periglacial conditions. Although it is unfortunate that the growth periods of stalagmites 61D and 13 do not overlap, site-specific differences of the records become also obvious when comparing the overlapping parts of stalagmite 15D and 60. Moreover, the distinctive darker colour of stalagmites 61D and 60 compared to samples 13 and 15D (see Fig. S2) supports this view.
In this context, I was very surprised to see that only the oldest part of sample 13 between 402 and 393 ka was analysed for the present study although it grew between ca. 402 and 373 ka (Honiat et al. 2026) and thus overlaps with samples 15D (and 60). As samples 13 and 15D come from the same part of the cave at ca. 2000 m a.s.l., this looks to me like a missed opportunity to check how well the sulphate d34S and d18O records and the various trace element records reproduce across stalagmites from the same cave location.
Finally, for the overlapping records of samples 15D and 60 I suggest indicating the sample from which the individual d34S and UAR(i) data points in Fig. 2 were obtained. The same applies to sulphate d18O and Δd18OSO4-H2O. By the way, the data tables are missing in the Supplement.
Line 401:
Here the authors argue that the geographic setting of Betten cave is conducive to the proposed scenario of the MIS11/10 glacial inception, thereby implying that the paleo-topography of the area 400 kyr ago was similar to present. Is this plausible? Considering a 200 m tectonic uplift of the area over the last 400 kyr (see Honiat et al. 2026 and references therein), and substantial erosion by four other glaciations in this period, one might speculate that the landscape in the Melchsee-Frutt area looked different than today. In connection with the tectonic uplift, one could also argue that the present-day passages of Betten Cave were not only at lower altitude, but also significantly deeper below the paleo-surface.
This could be of relevance for the interpretation of the Mg and Sr data. If 400 kyr ago, a different limestone strata or marls with higher Sr content than the MF1, MF2 and MF3 bedrock references were in contact with the glacier, it could potentially explain the inconsistency between the CaveCalc modelling experiments and the measured data shown in Fig. S3e,f. During subsequent glaciations these strata were eroded, and other limestone and limestone-marl strata are now exposed to the surface.
Line 482:
The authors report “a distinct negative trend following the statistical break point near 400 ka, after which Δd18OSO4-H2O declines from +22‰ to +15‰, possibly tracking long-term oxygen limitation (Fig. 2b)”. Indeed, there is a declining trend in the data, however, the beginning of this decrease is uncertain due to the low temporal resolution of the record between 380 and 400 ka and the scatter of the data. To me it looks like the Δd18OSO4-H2O decrease starts at around 390 ka, rather than 400 ka and in this case, it is not concurrent with the “400 ka break point” and d13C increase between 406 and 402 ka.
Minor comments:
- I suggest adding a subsection 2.1 “sample description” at the beginning of “Material and Methods”.
- Figure 3: For graphical reasons, it doesn’t make much sense to invert the Ni concentration axis to illustrate the opposing shifts of Ni and U concentrations.
- Figure 4b shows significant overlap between interglacial and periglacial samples. Is there a chronological trend within the periglacial data? E.g. the older the samples the closer they plot to the interglacial data points?
- Note, in Fig. S4c the Sr data are plotted for samples 13 and 15D only but not for 61D and 60.
- Fig. S4a and b need more explanation in the caption to help the reader understanding the meaning of the plots. Only referring to Sinclair (2011) and Wassenburg et al. (2024) is not sufficient.
- Figure S7: Add some text in the figure caption to explain the meaning and relevance of this figure. It is not self-explanatory.
Final recommendation:
Based on the above comments on the manuscript I conclude that the authors are mistaken with critical parts of their data interpretation. The point is that the data do not really show what the authors expected them to show. I don’t think their geochemical model for subglacial calcite dissolution is wrong. It’s much simpler - it’s an unfortunate selection of samples from different parts of the cave in combination with a missing temporal overlap between samples 61D and 13 that caused the misinterpretation of the data.
I therefore recommend rejection of the manuscript.
Citation: https://doi.org/10.5194/egusphere-2026-2840-RC2 -
AC2: 'Reply on RC2', Jonathan Baker, 09 Oct 2026
Reviewer comments are in bold text. Author replies are in unformatted text.
The authors present new records of sulphate stable isotopes (d34S and d18O) and trace element data from four stalagmites from Betten Cave (Melchsee-Frutt, central Swiss Alps) covering the period from 415-360 ka. Using a multi-proxy approach including previously published calcite d18O and d13C data from the same stalagmites (Honiat et al. 2026), the authors present a reconstruction of the timing and extent of progressive glaciation across the MIS 11/10 glacial inception depicted in Fig. 5.
General comment:
The multitude of proxies and their interrelations make this paper quite complex and demanding, and I had a hard time to get through and really understand it. I think there is potential to improve the manuscript in a way to make it easier for the reader to follow.
Response: We recognize the complexities of this multiproxy approach and appreciate your feedback as we work to improve its clarity.
Line 247:
The authors write: “Stable-isotope records are largely reproducible between intervals of synchronous growth” and refer to Honiat et al. 2026. This statement is fair for the d18Occ records but looking at the d13C records in the original paper, it is evident that the d13C record of sample 60 is in antiphase with the corresponding records of samples 13 and 15D in the period between 381-372 ka. Is this an issue of age uncertainties only?
Response: We agree that this line should be revised to communicate that we refer principally to the reproducibility of calcite δ18O, which reflects a smoothed climatic signal from meteoric water inputs, whereas the reproducibility (albeit with some offsets) of δ13C requires some nuance.
It is plausible that age uncertainties alone can explain the offsets in calcite δ13C between the stalagmites, since 1) 230Th dating uncertainties increase substantially in this range, approaching secular equilibrium in the 230Th-234U-238U chain; and 2) each depositional model was calculated independently in OxCal, without any attempt to tune the proxy records, which could potentially align the millennial-scale variance.
However, our interpretation of calcite δ13C during subglacial growth is that kinetic 13C-enrichment leads to values exceeding that of the host rock, following previous work by Skiba et al. (2023) and Fohlmeister et al. (2023) on Betten Cave stalagmites. The extent of Prior Carbonate Precipitation (PCP) is not expected to be identical between drip sites in such environments, particularly because extensive sulphide oxidation increases the sensitivity of δ13C to PCP (Fig. S3) and may vary locally due to subglacial weathering and redox conditions. This behavior contrasts with temperate and warm-climate settings, in which hydroclimate can modulate PCP extent and thus produce common δ13C signals across the cave system (e.g. Patterson et al., 2026, doi:10.1016/j.epsl.2025.119758).
Despite the minor disagreements in δ13C across this interval, therefore, the fact that measured values remain consistently at or above δ13C of the host rock in all three stalagmites provides spatially reproducible evidence that the surface environment was soil-free and potentially subglacial after ~402 ka.
Line 268:
The authors write that they converted calcite d18Occ into drip water d18O values. How was this conversion accomplished? The reconstruction of drip water d18O from d18Occ requires knowledge of the cave temperature and an appropriate calibration of the water-calcite oxygen isotope fractionation. None of this is mentioned in the text.
Since drip water d18O is subsequently employed to calculate Δd18OSO4-H2O that is then used as environmental proxy, a proper description of the d18Occ conversion into drip water d18O is indispensable.
Alternatively, fluid inclusion d18O could be measured directly. Based on the stalagmite images shown in Fig. S2, I guess there should be enough water in the samples.
Response: The omission of this information was not intentional, and we appreciate you bringing it to our attention. However, the oversight can be rectified with minor updates to the Methods section and relevant figures.
The conversion of calcite δ18O to a drip-water equivalent value (δ18ODWE) was calculated using the T-dependent equation in Tremaine et al. (2011), assuming isotopic equilibrium and a cave temperature of 1°C. We believe that the assumption of isotopic equilibrium is reasonable for slow-growing calcite at low temperature, which tends to minimize kinetic fractionation (Daëron et al., 2019, doi:10.1038/s41467-019-08336-5). The choice of 1°C derives from the physical constraint that the cave must have remained unfrozen during speleothem growth but likely remained below the modern mean-annual surface temperature of 2–3°C. This information was reported in the Supplementary Data tables but regrettably not stated explicitly in Methods.
In revision, we would additionally cite Daëron et al. (2019) and use their equation in tandem to calculate the conversion to δ18ODWE, as well as propagate uncertainties from both equations and a range of parameters by assuming a uniform T distribution from 0–3°C. However, we emphasize that it is less vital to constrain the absolute δ18ODWE values, since the glacial inception is interpreted rather from the trend in Δδ18OSO4-H2O, whose magnitude far exceeds that of the uncertainties introduced by our choice of parameters and equations.
If δ18O values of fluid-inclusion water were available, we would rely on those data instead. However, preliminary tests failed to yield sufficient inclusion water, so the method must be reserved for future studies.
Line 371: Section 4.2
The discussion provided in section 4.2 is very interesting and informative, but in my opinion, it is based on a misinterpretation of the underlying data.
Looking at the data records in Fig. 2a and Fig. 3, d13C is the only proxy that seems to provide evidence for an environmental change from interglacial to periglacial conditions, indicated by a distinctive, continuous increase from about 0 to +3 ‰ between 406 and 402 ka in sample 61D. Other proxy records like d34S, UAR(i), Na+K+Al, U, and Ni, in contrast, show clear discontinuities at 402 ka that are obviously related to the change from sample 61D to stalagmite 13. As the two samples originate from distant parts of the cave with 300 m altitude difference, it appears more likely that the observed “abrupt changes” are caused by site-specific differences of the karst bedrock rather than to the proposed transition to periglacial conditions. Although it is unfortunate that the growth periods of stalagmites 61D and 13 do not overlap, site-specific differences of the records become also obvious when comparing the overlapping parts of stalagmite 15D and 60. Moreover, the distinctive darker colour of stalagmites 61D and 60 compared to samples 13 and 15D (see Fig. S2) supports this view.
Response: We appreciate your careful attention to the trends in minor elements and the contribution of site-specific influences on trace-element geochemistry. We agree that spatial heterogeneities in host-rock composition and flow path hydrology explain most of the large contrasts in the U systematics and composition of stalagmites 13 and 15D vs. 60 and 61D, consistent with the color contrast apparent in Fig. S2. Therefore, we emphasized throughout the manuscript that such heterogeneities contribute to the disparate signals (e.g. lines 296, 353, and 416).
However, these observations do not preclude their utility here, and we strongly disagree with the conclusion that δ13C is the only proxy that marks environmental changes at ~402 ka. Most importantly, the abrupt transition in δ34SSO4 cannot be explained by a local bedrock signal or the transition from stalagmite 61D to 13. Aqueous sulphate δ34S is a common and well-established proxy in surface and groundwater hydrology for sulphide oxidation at the aquifer/watershed scale, and the near homogenous signal between stalagmites 60, 13, and 15D (despite being the most spatially separated and quite obviously deriving minor elements from distinct lithologies in the epikarst) supports that interpretation here. A more parsimonious explanation of the available dataset is that δ34SSO4 reflects a shift in the predominant sulphur source in the broader cave system, which we attribute to the weathering and redox impacts of glacial advance. In particular, the higher δ34S variance in the glacial stalagmites (especially 60 and 13) can readily be explained by the varying intensity of sulphide oxidation in response to climatic instability (Honiat et al., 2026) but not by site-specific controls. Generally elevated sulphate concentrations after 402 ka support the inferred role of sulphide oxidation, and the covariant shifts in UAR, U, and Ni further signal that redox potential is driving both proxies, rather than lithology. Our recently published study from Klaus Cramer Cave, Austria provides an additional example of consistent and low δ34SSO4 values across glacial intervals, preserved in three stalagmites that also exhibit distinct trace-element chemistry and were collected from different parts of the cave (Baker et al., 2026, doi:10.1038/s43247-026-03942-w). Finally, the transition at 402 ka is marked by a statistical break toward declining Δδ18OSO4-H2O, discussed below.
That being said, we find many of your concerns valid and would like to address them by reframing the discussion and logical flow of the manuscript as follows:
In our opinion, the interpretation of conventional stable isotopes (δ13C/δ18O in calcite) presents a hypothesis of environmental transitions, which requires validation by independent proxies. Sulphate isotopes of S and O provide this test and elucidate the nature and timing of glacial advance. Our results provide the first example, in which the influence of subglacial hydrology on sulphate isotopes (especially δ18OSO4)—observed in modern settings—can be documented for a paleo-karst setting. Additionally, we can now cite the results of Baker et al. (2026), whose inferred timing of the MIS 10 glacial inception at Klaus Cramer Cave (370.8 ± 1.8 ka at 1950 m asl) independently corroborates the conclusion we draw from the combined stable-isotope datasets at Betten Cave (372 ± 3 ka at 1700–2000 m asl).
Given these constraints, we investigate further whether glacial advance and subglacial environments produce a consistent and diagnostic response in the trace-element geochemistry of stalagmites, which could potentially distinguish them from ice-free, interglacial phases. We conclude that although the intra-stalagmite behavior of key elements (including Ni vs. U, S, Na+K+Al, Fe+Mn, and the minor cations of Mg, Sr, and Ba that can only be modeled via incongruent calcite dissolution) appears consistent with subglacial environments per the processes outlined in our discussion, these signals cannot unambiguously demarcate subglacial from temperate environments without information from additional proxies.
Finally, we propose to clearly differentiate between ‘Upper Cave’ (13/15D) and ‘Lower Cave’ (60/61D) stalagmites in all time-series plots, so that the contribution of spatial heterogeneities is more clearly communicated. Rather than illustrate transitions in the absolute magnitude of trace elements, we intend to focus on how the intra-sample behavior of elements (e.g. via PCA) may differ between interglacial and glacial conditions.
In this context, I was very surprised to see that only the oldest part of sample 13 between 402 and 393 ka was analysed for the present study although it grew between ca. 402 and 373 ka (Honiat et al. 2026) and thus overlaps with samples 15D (and 60). As samples 13 and 15D come from the same part of the cave at ca. 2000 m a.s.l., this looks to me like a missed opportunity to check how well the sulphate d34S and d18O records and the various trace element records reproduce across stalagmites from the same cave location.
Response: It is inaccurate to suggest that we analyzed only the oldest part of sample 13 in this study, and we did not miss the opportunity to validate the reproducibility of δ18OSO4 and δ34SSO4 between overlapping segments of stalagmites 13 and 15D. In fact, these values replicate well between the two and both contribute to the documented trend in fO2 (see Table S3).
The choice not to analyze the younger segment of 13 via LA-ICP-MS derived purely from logistical and budgetary considerations, as well as our intent to preserve the slab against unnecessarily destructive subsampling. Unfortunately, we lacked the foresight to anticipate the complexities in trace-element signals, but we do not think that this data gap undermines the primary conclusions of this study. On the contrary, many elements (such as Mg, Sr, and S) exhibit surprisingly smooth continuity between temporally adjacent samples, and the generally low abundance of trace metals in 13 is consistent with the same features in stalagmite 15D.
Finally, for the overlapping records of samples 15D and 60 I suggest indicating the sample from which the individual d34S and UAR(i) data points in Fig. 2 were obtained. The same applies to sulphate d18O and Δd18OSO4-H2O. By the way, the data tables are missing in the Supplement.
Response: We appreciate the suggestion and will update the figures accordingly.
To access Supplementary Data tables, please see our data availability statement on line 648.
Line 401:
Here the authors argue that the geographic setting of Betten cave is conducive to the proposed scenario of the MIS11/10 glacial inception, thereby implying that the paleo-topography of the area 400 kyr ago was similar to present. Is this plausible?
Response: Yes, we believe it is plausible that the general structure of the valley and ridge were sufficiently close to support the outlined scenario, given the low rate of cirque floor incision and relatively small size of the glacier catchment. Undoubtedly, the valleys are deepened today, relative to the MIS 10 glacial, but this deepening would also have been partially offset by valley infilling between glacial cycles.
Considering a 200 m tectonic uplift of the area over the last 400 kyr (see Honiat et al. 2026 and references therein), and substantial erosion by four other glaciations in this period, one might speculate that the landscape in the Melchsee-Frutt area looked different than today. In connection with the tectonic uplift, one could also argue that the present-day passages of Betten Cave were not only at lower altitude, but also significantly deeper below the paleo-surface.
Response: We do not dispute that significant geomorphic evolution has significantly reshaped the glacial valleys of the Alps and therefore highlighted in Honiat et al (2026) and Baker et al. (2026) that uplift since MIS 11 should be considered when interpreting the timing of associated glacial advance and retreat during the Middle Pleistocene. However, rates of uplift/denudation are difficult to constrain and vary widely between catchments (Delunel et al., 2020, doi:10.1016/j.earscirev.2020.103407), which warrants caution against speculating about their relevance here. In the absence of local geomorphic evidence to the contrary, we have opted to present proxy data in the context of the modern topography, adding appropriate caveats.
This could be of relevance for the interpretation of the Mg and Sr data. If 400 kyr ago, a different limestone strata or marls with higher Sr content than the MF1, MF2 and MF3 bedrock references were in contact with the glacier, it could potentially explain the inconsistency between the CaveCalc modelling experiments and the measured data shown in Fig. S3e,f. During subsequent glaciations these strata were eroded, and other limestone and limestone-marl strata are now exposed to the surface.
Response: Agreed, and for this reason, we do not attempt to over-interpret the Sr/Mg mismatch of the CaveCalc experiment with our collected bedrock samples, which certainly under-represent the total geochemical heterogeneity of the complex carbonate stratigraphy. In particular, Sr concentration varies strongly across carbonate sequences in shallow marine platforms, primarily as a function of early meteoric diagenesis. We believe the best approach is to separate the stalagmites as we have done in Figure S4.
Line 482:
The authors report “a distinct negative trend following the statistical break point near 400 ka, after which Δd18OSO4-H2O declines from +22‰ to +15‰, possibly tracking long-term oxygen limitation (Fig. 2b)”. Indeed, there is a declining trend in the data, however, the beginning of this decrease is uncertain due to the low temporal resolution of the record between 380 and 400 ka and the scatter of the data. To me it looks like the Δd18OSO4-H2O decrease starts at around 390 ka, rather than 400 ka and in this case, it is not concurrent with the “400 ka break point” and d13C increase between 406 and 402 ka.
Response: Statistical breakpoints (with confidence intervals) can more objectively be defined by employing a non-linear regression function, accounting also for age-model uncertainties. This approach can easily be employed in revision to remove the subjective element.
Minor comments:
- I suggest adding a subsection 2.1 “sample description” at the beginning of “Material and Methods”.
Response: Agreed.
- Figure 3: For graphical reasons, it doesn’t make much sense to invert the Ni concentration axis to illustrate the opposing shifts of Ni and U concentrations.
Response: We would appreciate the opportunity to revise Figure 3 in light of previous discussions and will incorporate this suggestion.
- Figure 4b shows significant overlap between interglacial and periglacial samples. Is there a chronological trend within the periglacial data? E.g. the older the samples the closer they plot to the interglacial data points?
Response: Yes, and this trend is evident in the plot of fO2 in Figure 4c, notwithstanding the anomalously high value near 376 ka.
- Note, in Fig. S4c the Sr data are plotted for samples 13 and 15D only but not for 61D and 60.
Response: This omission was not intentional and has been fixed.
- Fig. S4a and b need more explanation in the caption to help the reader understanding the meaning of the plots. Only referring to Sinclair (2011) and Wassenburg et al. (2024) is not sufficient.
Response: We can provide this explanation in a few paragraphs accompanying the figure.
- Figure S7: Add some text in the figure caption to explain the meaning and relevance of this figure. It is not self-explanatory.
Response: We agree and will elucidate the design and significance of the figure in a paragraph of supplementary text.
Final recommendation:
Based on the above comments on the manuscript I conclude that the authors are mistaken with critical parts of their data interpretation. The point is that the data do not really show what the authors expected them to show. I don’t think their geochemical model for subglacial calcite dissolution is wrong. It’s much simpler - it’s an unfortunate selection of samples from different parts of the cave in combination with a missing temporal overlap between samples 61D and 13 that caused the misinterpretation of the data.
I therefore recommend rejection of the manuscript.
Response: We appreciate the critical feedback on the interpretation of elemental signals and believe these sections can be significantly improved per your suggestions, particularly in their graphical presentation. However, we dispute whether it’s been established that the signals are misinterpreted herein, and several criticisms clearly arose from a misunderstanding of the proxy system and data coverage. Speculating about the role of spatial heterogeneities (which we had already acknowledged in the original submission) does not rise to the level of evidence against the underlying interpretation.
The principal geochemical proxies utilized to document the glacial inception (stable isotopes of C, O, and S) clearly replicate where temporal overlap exists, and there is ample precedent to support their use at the scale of a karst aquifer / cave system, both from studies in hydrology and in speleothem paleoclimatology. Since the major conclusions of this manuscript rest on 1) the coeval shift in δ34SSO4 with δ13C and various redox indicators; and 2) the temporal trend in δ18OSO4 and Δδ18OSO4-H2O, which follows calcite δ18O and agrees with the chronological constraint from Klaus Cramer Cave, it seems unjustified to reject a manuscript over the minor conclusions that attempt to synthesize trace-element signals in multiple stalagmites.
The documented complexity of trace-element geochemistry over hundreds of meters in a spatially heterogeneous cave system with diverse epikarst lithologies warrants caution and caveat, such as presenting conclusions as tentative hypotheses to be tested in future studies—not rejection of the much stronger conclusions from other proxies. Given that many studies present elemental data from single stalagmite records (precluding any contradictory trends that might result from in-cave heterogeneity), we find it unreasonable that a more thorough analysis should be met with this recommendation. Therefore, we ask for the opportunity to address these concerns in a thorough revision of the manuscript.
Citation: https://doi.org/10.5194/egusphere-2026-2840-AC2
-
AC2: 'Reply on RC2', Jonathan Baker, 09 Oct 2026
Data sets
Speleothem sulphate and trace elements constrain Alpine glacial inception across the Marine Isotope Stage 11/10 transition Baker, J., Honiat, A., Wynn, P. M., Fohlmeister, J., Trüssel, M., Hollowood, S., Ruan, J., Mertz-Kraus, R., Edwards, R. L., & Spötl, C. https://doi.org/10.5281/zenodo.20259664
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 225 | 109 | 31 | 365 | 93 | 21 | 27 |
- HTML: 225
- PDF: 109
- XML: 31
- Total: 365
- Supplement: 93
- BibTeX: 21
- EndNote: 27
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Review of “Speleothem sulphate and trace elements constrain Alpine glacial inception across the Marine Isotope Stage 11/10 transition” by Baker et al.
This manuscript presents a nice dataset obtained from the recently published stalagmites from Betten cave in the Swiss Alps (Honiat et al., 2026). The new proxies (isotopes in speleothem sulphate and trace elements) allow to interpret the environment associated with the speleothem formation in the studied time interval (transition MIS 11-MIS 10) and shed some light on the mechanisms related to subglacial speleothem precipitation. This is a very interesting study, well-written, original and that will help to further studies in this research line in other mountain ranges (eg. Pyrenees). I recommend publication, I just have some minor comments
Line 37: temperature glacier – is it correct or a typo? I think it has to be changed to: temperate or warm-based glacier
Line 64: in those depressions filled by lakes, is there any study on lake sediments that merits citation? It may be another type of continuous record to fill the gap!
Line 86: this paragraph about Paleoclimate records in alpine karst may be included in the following 1.3 about subglacial speleothems to simplify the introduction.
Line 107: change deposition by formation or by precipitation
Line 172: I wonder if any cosmogenic date or OSL date are available in the Alps for this MIS 10 glaciation. If so, it would be important to reflect it in the text and in the map of Figure 1.
Line 195: Since this methodology is quite novel (or at least it is not the usual one in speleothems) I suggest you to include a step by step procedure in the supplementary to help the readers in further studies.
Results: to me, Figure 3 should appear earlier in the manuscript when trace element profiles are described. As it is now, the reader gets lost in the Result section since the figures with the trace elements are in the supplementary; this fact makes the reading uncomfortable and unclear. I suggest to include at least one figure with trace element data in the main text or place Figure 3 earlier in the text.
I like Figure 5, with the explanation of the three phases on the landscape, well-illustrated and easy to understand the different signals in the different stalagmites from the same cave. I am just not totally sure if this figure will be clearer if removing the satellite image and using an elevation model. Just a suggestion to check. Regarding Figure 6, I suggest increasing the size (of the figure and, specially, of the text, it is really small).