the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Spatio-temporal dynamics of speleothem growth and glaciation in the British Isles
Abstract. Reconstructing the spatio-temporal dynamics of glaciations and permafrost largely relies on surface deposits, and is therefore a challenge for every glacial older than the last due to erosion. Consequently, glaciations and permafrost remain poorly constrained worldwide before c. 30 ka. Since speleothems (carbonate cave deposits) form from drip water and generally indicate the absence of an ice sheet and permafrost, we evaluate how speleothem growth phases defined by U-series dates align with past glacial-interglacial cycles. Further, we make the first systematic comparison of the spatial distribution of speleothem dates with independent reconstructions of the history of the British-Irish Ice Sheet (BIIS) to test how well geomorphologic ice reconstructions are replicated in the cave record.
The frequency distribution of 1,020 U-series dates based on three different dating methods between 300 and 5 ka shows statistically significant periods of speleothem growth during the last interglacial and several interstadials during the last glacial. A pronounced decline in speleothem growth coincides with the Last Glacial Maximum, before broad reactivation during deglaciation and into the Holocene.
Spatio-temporal patterns in speleothem growth between 31 and 15 ka agree well with the surface-deposit-based reconstruction of the last BIIS. In data-rich regions, such as northern England, ice dynamics are well-replicated in the cave record, which provide additional evidence about the spatio-temporal distribution of permafrost dynamics. Beyond the Last Glacial Maximum, the distribution of speleothem dates across the British Isles offers the opportunity to improve chronological constraints on past ice sheet variability, with evidence for a highly dynamic Scottish ice sheet during the last glacial. The results provide independent evidence of ice distribution complementary to studies of surface geomorphology and geology, and the potential to extend reconstructions into permafrost and earlier glacial cycles. Whilst undersampling is currently the main limitation for speleothem-based ice and permafrost reconstruction even in relatively well-sampled parts of the British Isles, we show that speleothem dates obtained using modern mass spectrometry techniques reveal a higher spatio-temporal resolution of glacial-interglacial cycles and glacial extent than previously possible. Further study of leads and lags in speleothem growth compared to surface deposition may provide new insights into landscape-scale dynamics during ice sheet growth and retreat.
- Preprint
(2133 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on cp-2024-48', Andy Baker, 10 Jul 2024
General Comment
The authors report an impressive new synthesis of speleothem U-Th ages from the British Isles, and use this to focus on a comparison between last glacial ice sheet location and the speleothem growth. It is a very valuable contribution to the literature.
Specific Comments
My questions to the authors largely relate to asking for more methodological detail so that the reader can reproduce the analyses and assess the certainty to which a speleothem U-Th age can be correctly attributed to the bins in section 4.2 ‘Spatial Distribution of Speleothem Growth’.
Line 98. Should the published sources of the compiled ages be cited somewhere e.g. a summary table by region? I see that the Figshare file does have complete or partial references, but that is separate from this document.
On lines 103-109. The authors state that they have chosen to use the published corrected U-Th ages, with justifications given. Some dates were rejected (line 106). So that future researchers could reproduce and expand on this work, please specify the criterion used for rejection of ages.
On line 107 it is stated that where uncertainties were not provided, the average was used (depending on analysis method). Was the age also taken into consideration e.g. as written, does this likely underestimate the uncertainty for older samples, and overestimate it for younger samples, and underestimate the uncertainty for samples that are detritally contaminated. Does the inclusion of these samples have an undue influence on the overall result?
Lines 123-150. Could the authors provide more detail on the pdf method, noting that it is ‘following Scroxton et al (2016)’ but that publication might not be accessible to all. In particular, could more information be provided on the calculation of the z-scores shown in Figure 2. Did the authors generate ‘10,000 synthetic ages …. calculated from the same exponential relationship to determine the predicted variation through time caused by chance’ (quoted from Scroxton et al 2016).
Section 4.2.1. Would the authors consider including some expert opinion on the quality of some of the critical sites that are considered in this section. For example, Crag Cave appears to be very important when comparing the timing of the ice sheet in the west of Ireland, but how reliable are those dates? I went to look at the supplemental datafile to see if one estimate of reliability, the 230Th/232Th in the speleothem, had been compiled, and that is not the case. However, I had access to the original publication (Fankhauser et al 2016, which might be paywalled for others) and could check that these speleothems had precise dates with very low detrital Th contamination. I would suggest that this would be useful information to convey in the text, and similarly for any other sites and samples that are critical to the interpretations made in section 4.
Lines 395-406. With apologies for the self-citation, a comment that Caseldine et al. (2008) report oak pollen in speleothems from the Yorkshire Dales well into this time period, which would agree with this interpretation.
References
Caseldine, C.J., McGarry, S.F., Baker, A., Hawkesworth, C. and Smart, P.L.: Late Quaternary speleothem pollen in the British Isles. Journal of Quaternary Science, 23, 193-200, https://doi.org/10.1002/jqs.1121, 2008.
Fankhauser, A., McDermott, F., and Fleitmann, D.: Episodic speleothem deposition tracks the terrestrial impact of millennial-scale last glacial climate variability in SW Ireland, Quat. Sci. Rev., 152, 104–117, https://doi.org/10.1016/j.quascirev.2016.09.019, 2016.
Scroxton, N., Gagan, M. K., Dunbar, G. B., Ayliffe, L. K., Hantoro, W. S., Shen, C. C., Hellstrom, J. C., Zhao, J. X., Cheng, H., Edwards, R. L., Sun, H., and Rifai, H.: Natural attrition and growth frequency variations of stalagmites in southwest Sulawesi over the past 530,000 years, Palaeogeogr. Palaeoclimatol. Palaeoecol., 441, 823–833, https://doi.org/10.1016/j.palaeo.2015.10.030, 2016.
Andy Baker
Citation: https://doi.org/10.5194/cp-2024-48-RC1 -
RC2: 'Comment on cp-2024-48', Anonymous Referee #2, 18 Sep 2024
My sincere apologies to the authors and editor for the time taken for this review.
This manuscript is an important contribution in its demonstration of the potential for constraining the extent and distribution of ice cover over time where cave records are available. It is somewhat hamstrung by the quality of a lot of the available data (Alpha spectrometric age determinations, predominantly conducted before the development of mass spectrometric dating in the late 1980s) but clearly illustrates the potential for a large project to address this in the UK. The study also clearly points to the potential for constraining timing and extent of older glacial advances for which the surface evidence no longer exists, although this would require a great deal of U-Th dating to locate older speleothem material.
The authors have chosen to use U-Th ages as originally published, which allows use of a larger dataset where the isotopic data required to recalculate the ages using current techniques were not always provided. I would argue that it would have been preferable to recalculate ages for which sufficient isotopic data were available and discard those where they were not (mostly ASU ages I would expect), giving a cleaner less noisy record at the cost of reduced data density - but I can accept this is a matter of opinion and going back to do that at this point would be an enormous task beyond the scope of this study
I recommend publication. I have some comments below for the authors consideration, they mostly affect wording with the exception of the error noted below in Fig 6 and the raw data table which should be corrected.
line 104: I understand the difficulty of compiling data from sources which did not always include the required information to recalculate published ages but it should be noted that Gaussian distribution of the calculated age is not the only consideration here. Our knowledge of the relevant constants has improved significantly since some of these were published, as have practices for calculating corrected ages (for instance many publications around the turn of the century incorrectly calculated the uncertainties of their corrected ages due to either a bug or common misconfiguration of widely used software at that time). Without this information, some of these dates might be best viewed as qualitative. I don’t think this is a show stopper, but it is perhaps an unnecessary source of noise affecting this study
line 108: Uncertainties vary a lot with age, and older ages will bias the average age uncertainty determined here. Might be better practice to use median although with low number of old ages might not make much difference. Is the average uncertainty determined using the age-filtered dataset?
115: I don’t follow the reasoning for excluding young ages on this basis, needs to be more clearly explained. I can understand they might flood the record otherwise but I wouldn’t call that a sampling artefact.
118: The question of how to address multiple dates on a single speleothem when compiling them into a histogram or PDF is always difficult. This approach intrinsically assumes random sampling, which is possible to do after a fashion if you set out into the field to sample randomly specifically for this purpose. But with a literature compilation approach (as most studies including this one have used) the difficulty is that the compiled data were mostly sampled anything but randomly, for completely different purposes. Where you have multiple dates available from a single speleothem the most appropriate approach might be to randomly choose one date and move on, but that is not really feasible, not least in that the reader must be convinced your choice was truely random. If you always take the oldest date your compilation is biased towards growth initiation times, if you take the youngest it is growth cessation times. I don’t think there is an easy answer to this, but I would be wary of biasing the compiled record by including more than two or three dates from any individual speleothem, unless it has clear major growth hiatuses then maybe treat each interval separately. Certainly, reducing the Crags Cave dataset at least as much as done here is justified.
Page 5: I agree, 5 kyr binning seems like a reasonable compromise here. The only alternative is bin width as a non-linear function of age and I don’t think anyone has gone there yet.
Page 5: The exponential normalisation is glossed over here. What was the exponential decay coefficient used in each case (for instance, expressed in terms of probability of speleothem removal per kyr) and how was this fitted to the data? Figure A1 shows this graphically but doesn’t explain how the fit is weighted to the data. The rate of decay used here would appear to have an impact on which of the normalised peaks are found to be significant at the next step.
Line 151: I certainly agree with this. Constant removal from consideration by erosion/burial might not hold up very well at thousands of years timescale but should be a reasonable model over hundreds of thousands of years, as seen in multiple previous studies from around the world.
Figs. 2, 3 and A2, and results section: These make a pretty convincing case for disregarding or at least minimising the use of the ASU data. It was a revolutionary breakthrough for its time and it is important to have compiled all that was available, but mixing these with mass spectrometric U-Th data degrades the overall quality of the dataset. A similar argument applies to figure 6.
196: I am always a little uneasy about hanging anything important from a single U-Th date. A cluster yes, even a pair, but there are just too many ways a single U-Th date can be wrong (due either to the sample itself, or potentially to the analytical process especially for ASU or TIMS).
In general the discussion section is very good, with adequate recognition of the limitations of the ASU data combined with advocacy for extending the mass spectrometric dating record.
207: The peak-trough smoothing observed in PDFs is predominantly controlled by some combination of age error and data density. It’s not clear to me that sampling from a wider area makes this any worse than from a single location, unless the relevant processes are operating at significantly different times in different regions under consideration.
209: “Statistically significant” has formal meaning difficult to associate with a decay-corrected PDF curve. I would suggest just “significant.”
Page 15, fig. 6: The authors appear to have incorrectly recorded uncorrected TIMS ages as being corrected TIMS ages for Dream Cave, shown here at 53.07N. The authors of the cited source paper did also supply corrected ages, of about 2 - 3 kyr younger than shown here for these two samples. Note they didn’t propagate any uncertainty associated with the correction (not doing so was common practice at the time).
524. Lechleitner et al 2021, appears to be an orphaned reference.
Citation: https://doi.org/10.5194/cp-2024-48-RC2
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
296 | 80 | 17 | 393 | 11 | 8 |
- HTML: 296
- PDF: 80
- XML: 17
- Total: 393
- BibTeX: 11
- EndNote: 8
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1