the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 Jul 2026
Fibre-optic strain sensing on an alpine glacier
Abstract. In-situ measurements of ice deformation are constrained by the spatial resolution of discrete sensors, leaving fundamental questions about small-scale ice rheology unresolved. Here we present the first application of fibre-optic Distributed Strain Sensing (DSS) in a glacier setting, deployed at Chessjengletscher, a small polythermal glacier in the Swiss Alps. We compare Brillouin and Rayleigh DSS techniques in a borehole drilled to 38.4 m depth, alongside Raman distributed temperature sensing. Both methods resolve highly variable strain rates at the sub-metre scale — detail unattainable with conventional tiltmeters or inclinometry. Fitting observed strain profiles to a plane-strain deformation model supports a near-linear ice rheology for temperatures within 1 °C of the pressure melting point, consistent with recent laboratory experiments on temperate ice. Corresponding rate factors align closely with accepted values – though likely tempered by incomplete borehole freeze-in, which decouples the cable from ice deformation over portions of the column and introduces uncertainty into our displacement estimates. Brillouin offers a simpler field workflow and kilometre-scale measurement range compatible with remote deployment while Rayleigh provides superior spatial resolution but requires a demanding calibration procedure ill-suited to autonomous field use, though it shows promise for laboratory-scale deformation experiments. DSS represents a step change in englacial deformation measurement with direct implications for the study of ice rheology, glacier dynamics, and hazard assessment at alpine and ice-sheet scales.
Competing interests: Matthias Buhler is employed by Marmota Engineering AG who were contracted by ETH Zurich to deploy and process strain data.
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.- Preprint
(11472 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-3329', Anonymous Referee #1, 05 Aug 2026
-
RC2: 'Comment on egusphere-2026-3329', Florent Gimbert, 21 Sep 2026
Review of « Fibre-optic strain sensing on an alpine glacier » by Robert Law et al., submitted to the Cryosphere
Overall assessment
This manuscript presents the first application of fibre-optic Distributed Strain Sensing (DSS), combined with the also innovative application of fibre-optic Distributed Temperature Sensing (DTS), to the measurement of deformation within a glacier borehole. The authors compare Brillouin- and Rayleigh-based measurements and combine these observations with distributed and discrete temperature measurements. The study is technically innovative, and the overall agreement between the different strain measurements suggests that DSS has considerable potential for resolving along-cable strain at spatial scales inaccessible to conventional borehole instruments.
However, the mansucript’s central message goes beyond the technical applicaiton of DSS/DTS, concluding that the measurements support a near-linear rheology for near-temperate ice (outlined in the abstract and throughout). This conclusion is of primary importance since such change in rheology compared to the classically used Glen’s flow law with n=3 has a strong impact on glacier modelling strategies and outcomes. In my view, however, this finding is not supported by the authors analysis. The three most prominent problems I have are the following ones. First, as acknowledged by the authors, the cable in the bottom section from 23 to 35 m deep expected to moslty constrain the rheological law is not coupled, causing the inferred deformation to be linear, consistant with a linear rheological law best fitting this artefact. Second, the interpretation of the DSS measurements using the plane-strain approximation is highly questionnable. One evidence of this assumption being violated here is that the displacement predicted at the surface from integrating the strain over depth is four times higher than the independent GNSS observation. Finally, the authors appear to be using the shallow ice approximation to estimate shear stress with depth, which has recently been demonstrated to not hold at least on an Alpine glacier configuration, such that borehole deformation measurements may provide little constraints on the Glen’s flow law exponent (Roldan Blasco et al., 2025).
I also have concerns on the overall presentation of the study and results. I don’t find the authors provide sufficient quantitative information on both the field setup and their theoretical considerations for the reader to judge the robustness of the assumptions and results. On the observational side, several aspects of the measurement strategy are insufficiently documented, particularly in Sections 2.3 and 2.4 (lines 160-224) : the GNSS surface velocity measurements over the period of acquisition are not shown, additional information about the local glacier geometry and dynamics including expectations of spatial velocity gradients, possible basal motion and bed geometry are not presented. On the theoretical side, the inferred glaciological quantities depend on several insufficiently justified assumptions. There are also numerous vague statements with lacking proper justification/quantification/references (e.g. l 56-58, l 187-189, l 221-224, l 231-234, l 253-254)
I expose my criticims more in details below. Addressing all these concerns would require more than a conventional major revision because they affect the central framing, analysis, and conclusions of the paper.
Major comments
1. Linear deformation in the bottom layer
The cable is known to be mechanically uncoupled from the surrounding ice at several locations along its depth. This uncoupling results in apparently linear deformation profiles, as is particularly evident within the lowermost ~10 m thick layer. Such linear deformation is a measurement artefact and therefore does not represent the local deformation of the ice.
To my knowledge, previous observations based on tiltmeters consistently show increasing non-linearity in ice deformation with depth (e.g. Maier et al., 2019; Roldan Blasco et al., 2025). The present study appears to be an exception, but this difference can be explained by an instrumental artefact that is both acknowledged by the authors and straightforward to understand. Consequently, the affected sections of the cable and, most critical for constraining the power law exponent, the deeper ice deformation, provide very limited information with which to constrain the rheological law.
I suspect that the resulting best-fitting rheological law being linear arises primarily from this measurement artefact. One way to reduce this bias would be to exclude the linear sections of the observed deformation profile from the power-law fit. Since these sections effectively represent interpolation across mechanically uncoupled portions of the cable rather than direct measurements of local ice deformation, allowing them to contribute to the fit artificially favours a more linear rheological response.
2. Assumption of negligible vertical extension or compression
A central assumption made by the authors is that all the deformation of the cable is induced by vertical shearing (the plane strain approximation, lines 187-194). The justification currently provided - the relatively simple geometry of the glacier and the use of similar assumptions in borehole inclinometry studies – first strongly lacks precision and references, but most importantly is not robust in the present case. Regarding the effect of geometry, the contribution of vertical extension/compression from longitudinal stress gradients should be evaluated using more elaborated numerical models. Regarding the justification that borehole inclinometry studies also do this hypothesis, I agree this is mostly true (references should be given), but I expect such assumptions to much more likely hold for inclinometry studies than for DSS studies. In the inclinometry case, tilt is by essence primarily sensitive to shear : a given shear deformation gives a large tilt change, while a vertical extension or compression of similar magnitude gives a small tilt change. For a fibre cable, it is the opposite, because the cable is vertical and by essence mesures elongation, such that vertical deformation has a disproportionate effect on the DSS measurements compared to the tilt measurements. The authors should acknowledge this difficulty and address the subsequent challenge of disentangling both contributions, which is a major objective if we want to use DSS as a robust technique to quantify ice rheology under shear.
An independant observation suggesting this source of bias is likely a major one is the finding that surface displacement obtained from integrating the DSS measurements assuming plane strain gives surface velocities four times larger than observed from GNSS. The statement at lines 284-285 that displacement integrated above 23 m matches the total GNSS displacement implies that the measured strain above 23 m already accounts for nearly all observed surface motion, leaving little or no contribution from deformation below 23 m or from basal sliding, where it is expected to actually be the largest.
The authors mention this bias in the discussion, among with others (lines 288-291). In my view, this is not a secondary caveat but one of the most plausible explanations for the unexpectedly large inferred displacement. I think this last point could be the main emphacis of the present paper.
3. Use of the shallow ice approximation to infer the shear stress profile
Although the authors do not state it explicitely, it appears to me that they use the shallow ice approximation to express shear stress as a function of depth, and use this expression to confront shear stress with the assumed plane-shear deformation. However, previous study shows based on numerical modelling (Roldan Blasco et al. 2025, https://doi.org/10.5194/tc-19-267-2025) that, due to 3D geometry complexity, internal ice stresses depart significantly from the shallow ice configuration, such that adjustement of the Glen’s power law exponent, which heavily relies on the stress description with depth, is underconstrained. A similar comment applies to the choice of f=1. The authors should evaluate the impact of their description of shear stress with depth on their finding of n, and evaluate the degree to which this exponent is actually constrained or not given all the uncertainties following my comments 1- to 3-.
4. Lacking information on the measurement strategy
Section 2.3 provides useful information about the installation and instruments (lines 160-185), but it does not describe the acquisition and processing strategy in sufficient detail for the reader to understand exactly how the reported strain profiles were obtained. For each DSS instrument, the authors should specify the gauge length, the integration time used for each measurement, the timescale over which differenciation is made, if there are any averaging performed over consecutive measurements and over the entire period, etc… This would allow assess the uncertainty associated with individual measurements and how these are mitigated when averaging or taking a long time period for differencing. There is also no information on any calibration procedure. It seems that the strain reported in the manuscript represents the difference between the measurements made on 19 August and 3 September 2025, is it ? If so, why the 8 August reference measurement was not used in the principal analysis and how this choice relates to the progression of cable freeze-in. We would also want to see a timeseries of GNSS velocities or some better constraints on how velocities may vary trhough the year at this site.
5. Imprecise/vague statements
At numerous instances the language is imprecise/vague. Section 2.1 states that glaciers with a similar geometry “typically do not have complicated flow fields” (lines 50-58), but based on which evidence this statement is supported, and what specifically the authors consider a complicated flow field ? For example, it is known that glaciers undergo vertical compression and extension from longitudinal stress changes, which would directly convert into strong along cable deformation measured by DSS, which the authors do not account for. A GNSS measurement of changes in vertical elevation through time would have been quite key to address this problem. Similarly, the description of the glacier geometry as “relatively planar” is used to justify a shape factor of 1 (lines 221-224), while the choice of representative slope and thickness is simply described as “reasonable” (lines 231-234), without sufficient geometric evidence or sensitivity analysis. In the Results, the authors state that the profiles “broadly conform” to several possible rheological behaviours (lines 247-249) and that the different measurement records are in “good agreement” (lines 244-246), but neither assertion is supported by quantitative criteria. Also surprising is the attribution of the DTS noise to the age of the interrogator and asserting that it has no effect on the results (lines 253-254), which appears as quite speculative. The authors also state that values of A fall very close to the standard ones but do not give numbers allowing the reader to judge (lines 260-261).
6. Interpretation of positive and negative strain
Lines 241-244 report both positive and slightly negative strain over different parts of the profile. Lines 249-251 then argue that the negative values do not result from temperature-change interference because the temperature had stabilised. Would be nice if the authors could explain further the physical meaning of these negative values. It should also be clarified how positive and negative strains can occur at similar depths or differ between the two branches of the looped cable
Minor comments
- Figure 2A: The schematic is not particularly informative in its present form. It could be improved by showing the borehole, cable geometry, coupled and potentially uncoupled sections, basal turnaround, the positions of the GNSS, the two boreholes, etc…
- Lines 59-60: The surface displacement of 1.43 m ashould be given in the results section, and accompanied by the measurement interval.
- Lines 170-172: Clarify why the GNSS displacement was measured in 2024 while the DSS experiment was conducted in 2025. Please provide the processing uncertainty and evidence that the 2024 velocity is representative of the 2025 DSS interval.
- Lines 181-182: Correct “The results presented here are between based on the two follow-up measurements.” State the analysed dates explicitly.
- Lines 183-185: Distinguish between sampling interval, spatial resolution, gauge length, and strain resolution for both instruments.
- Lines 247-249: Replace the statement that the profiles “conform” to anticipated displacement patterns with more cautious language reflecting the assumptions imposed in calculating those profiles.
- Lines 249-251: Explain the physical interpretation of negative strain rather than only excluding temperature change as its origin.
- Lines 253-254: The suggestion that DTS noise results from the age or unknown service history of the interrogator is speculative. Either provide supporting evidence or remove this explanation. The statement that the noise has “no bearing” on the results should also be supported quantitatively.
- Lines 258-260: Correct “plane-strain profile profiles.”
- Figure 5: Show uncertainty or confidence intervals for the fitted value of , not only the RMSE curves. The figure or caption should identify the depths included in the fit.
- Lines 282-285: The values of 6 and 1.43 m ashould be fully documented before they are interpreted in the Discussion.
- Lines 285-286: The two-week duration of the DSS measurement should be stated prominently in the Methods, Results, and Figure 4 caption.
- Lines 305-307: Correct “in line with with surface displacement.”
- Lines 307-309: The description of the site as a “surprisingly robust field laboratory” seems inconsistent with the unresolved coupling and displacement issues, and with the lack of a 3D model applied on the area which would largely helps interpreting these measurements. Please use more cautious wording.
- Line 338: Remove the additional closing parenthesis after “Law et al., 2021).”
- Lines 332-343: The broader claims that DSS can constrain ice rheology from grain to ice-sheet scales should be moderated or clearly presented as potential future applications. The present experiment highlights cable-strain measurement is likely high resolution in certain places of the borehole, but not an unambiguous recovery of local rheological properties.
Citation: https://doi.org/10.5194/egusphere-2026-3329-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 269 | 139 | 38 | 446 | 25 | 31 |
- HTML: 269
- PDF: 139
- XML: 38
- Total: 446
- BibTeX: 25
- EndNote: 31
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General Comments
This is an interesting and well written study demonstrating a variety of fiber-based measurements techniques in temperate ice. The authors have done a very good job of both reviewing the techniques but also making the most out of a challenging measuring environment (a weakly freezing borehole!). The paper is a sound contribution.
As a general comment, I would suggest that the authors expand their comparison of resolution of these techniques to other vertically resolvable measurements, such as APReS, etc.
Specific Comments
Line 74. While Raman backscatter can be used independently, the backscatter return can be significantly altered by local strain on the fiber, requiring more sophisticated fiber geometries (double ended) to account for non -uniform attenuation. I would suggest noting this here that even Raman sensing is not independent of strain and strain distribution. A small point but helps demonstrate the linkage between strain and temperature in all of these techniques.
Line 163: an overstuff of 1% appears large by some industry standards. Can you please double check this?
Line 185: I do not see a reference for the borehole diameter; can you give the reader some sense of the expected thermal equilibration time following freeze in? Some further discussion here on the borehole thermal equilibration would be helpful.While this is discussed further in the manuscript, I believe pointing out that freeze in in a temperate setting such as this is quite challenging to give the reader a foreshadowing of some of the issues that will arise later on.
Line 316: This is the first appearance (I may have missed it earlier) of an assumption of fiber failure at 1% strain. This may be a bit confusing to some readers when compared with the 1% overstuff reported in the loose buffered cable for temperature. I would suggest devoting a greater discussion to the maximum strain that can be measured by typical tight buffered (or bare fiber that matter). This would be helpful for other future uses. For example, does your tight buffered fiber contain any strength elements such as Kevlar? If so, how would be impact the potential for viscoelastic slip along the fiber. It may be just as easy to use bare fiber in this case, to ensure that there is not strain resistance due to other components of the tight buffering.
And while on the subject, readers may ask if there are radial strain issues on the fiber during freeze in that could impact longitudinal strain measurements. Tyler et al ( https://doi.org/10.3390/photonics11070630) demonstrated very negligible impacts to single mode fiber under high stress conditions suggesting that freeze in is unlikely to impact at least Raman backscatter.