the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 04 Jan 2026
Bedrock topography and ice thickness distribution of three major Patagonian outlet glaciers unveiled by helicopter-borne ground penetrating radar
Abstract. We present the first helicopter-borne ground-penetrating radar dataset over Glaciar Viedma, Upsala, and Perito Moreno, three of the largest outlet glaciers of the Southern Patagonian Icefield in South America. The dataset comprises 116.021 individual ice-thickness measurements along 232 km of flight tracks. Data were acquired during two campaigns in March and April 2022 as well as in October 2024 using a 25 MHz shielded broadband antenna deployed as a helicopter sling load. For the first time, we reveal the complex subglacial topography of these glaciers in their lower regions and measured bed reflections at Glaciar Upsala in depths of up to 800 m. The newly obtained measurements were incorporated into an ice-thickness reconstruction method to derive glacier-wide ice-thickness distribution maps and the corresponding bedrock topography. The latter exerts primary control on the response of water-terminating glaciers to a changing climate. Our 100-m gridded ice-thickness maps indicate that the three glaciers had a combined ice volume of 831.14 km³ in the year 2000. The dataset and the well-constrained glacier-wide grids provide a valuable basis for future studies aiming to better understand the mechanisms driving glacier retreat and the susceptibility of these large outlet glaciers to climate change. All data are publicly available at Zenodo (https://doi.org/10.5281/zenodo.17464164; Koch et al., 2025a).
- Preprint
(2689 KB) - Metadata XML
-
Supplement
(496 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on essd-2025-678', Shin Sugiyama, 06 Mar 2026
-
RC2: 'Comment on essd-2025-678', Anonymous Referee #2, 28 Mar 2026
Bedrock topography and ice thickness distribution of three major Patagonian outlet glaciers unveiled by helicopter-borne ground penetrating radar
General comment: the manuscript presents a valuable investigation of the ice thickness and the bed topography of three Patagonian outlet glaciers, revealing ice thickness up to 800 m. Such a record of penetration depth combined with the dense coverage of GPR profiles is a valuable adding to the existing knowledge. Therefore, I think a comparison with previous studies is essential to frame the results of the manuscript in the context of Patagonian glaciers ice thickness.
The structure of the whole manuscript needs to be improved, as some results are described in the section methods. Figures 3 and 4 show two GPR sections which should require a dedicated section in the results, not in the methodology (data interpretation is 3.2, while it should be the first subsection of the results, so 4.1). Also, as the maximum depth of investigation is exceptionally high in Upsala Glacier, I suggest to show the profile where it can be seen, so it may be a solution to merge figures 3 and 4 into one figure showing a GPR profile for each glacier. The coherency between figures and text should be revised, as in paragraph 4.1 about ice thickness, the description of channelized bedrock topography is associated to figure 5, which shows ice thickness. For the data processing and modelling, it would be better to have more detail about the parameters’ choice.
A note about the data release: it would be more useful for the public to have also the processed data in a format that is usable from most of the community. The standard format for geophysical data is a SEG-Y (.sgy). I encourage the authors to think about releasing also the processed sections, otherwise consider to release the raw file in .sgy format.
Considering the comment above, I recommend the publication of the manuscript after revisions, mainly involving the comparison with previous dataset and focusing on improving the quality of the manuscript.
I thank the authors for the very interesting reading, and I am very looking forward to see this paper published.
Specific comments
L14: I suggest to present the dataset in terms of kilometer of profiles interpreted, instead of points measurements. So, the sentence would be “The dataset comprises X km of ice-thickness measurements along 232 km of flight tracks”.
L18: Reporting the highest ice thickness is perfect, as it is really extraordinary, but reaching 800 m may be due to a combination of factors, not a standard. So, treat this value with care. Have you calculated the distribution of the ice thicknesses and defined a reasonable depth of investigation?
L38: the abbreviation NPI is never used again in the manuscript, so remove it
L61: have you checked if there are some measurements of ice thickness from satellite or airborne radar?
L65: you have already defined SPI, so there is no need to do it again
L69: same comment as L14. You can report the % or interpreted bedrock over the whole survey.
Figure 1: I think that a zoom in is necessary here to see the survey lines better.
L98: if you decide to use “Glaciar” is fine but be sure to be consistent through all the manuscript. Anyway, the English version “Glacier” should be used.
L116: replace “seismic measurement” with “seismic survey”
L134-138: a reference to support the hypothesis of temperate glaciers is needed here, as well as for the scattering due to water pockets
L142: is it a commercial GPR antenna or it was built by Blindow et al., 2011?
L154: the references should appear in chronological order
L156: Was the stacking set to 256 in the acquisition settings? Just for clarity
L175: you have already named the software, just remove the sentence and start straight with processing steps
L183: why did you choose 20 traces for background removal? I guess 20 traces gave the best result, right?
L190-198: I am a bit confused about the steps here, what is the manual layer determination? Was it associating the first arrival to the topographic surface? If so, what are the ice layers then? From the description I am a bit confused with the interpretation step, which is the picking phase, right? The air-layer correction is the topography correction where you considered the elevation of the platform, the elevation of the helicopter and the topography? I suggest to be more precise in the description of the processing steps. Have you tried other migration algorithm? Why did you choose diffraction stacking?
L202-211: as said in the general comment, these lines should be moved in the first subsection of the results. The figures should be described more carefully, for instance the “smiley features” at the bottom of figure 3 are not comment. Please move somewhere else the (a) because it covers a part that is described in the text. Since you comment on the surface morphology between km 4 and km 5 it is hard to see what you are referring to. For the detailed features maybe consider to use the appendix for more figures. In the text there is a reference to figure 3(c) which does not have a correspondence in the figure.
I suggest to merge figure 3 and 4 into one figure with three panels plot, one for each glacier. For Upsala, plot the GPR sections where the ice thickness reaches 800m. I also recommend to plot the section with the topography correction applied.
L228: give value for picking uncertainty, how was it calculated? Referring to velocity uncertainty, you are using 0.168 m/ns which is the velocity for pure ice. It is totally fine to use it, but as you said that the glaciers are temperate, have you tried to quantify the difference in terms of ice thickness if you use a lower velocity, as is expected for warm ice?
Also, Perito Moreno glacier has some snow layer, how thick is it? Does it affect the time to depth conversion if you take this layer into account? Maybe you should discuss also this source of uncertainty.
L276: give value for the plausible bounds. I think you should give more details for the whole modelling procedure
L286: as above for L14 and L69
L290: I guess we are talking about Perito Moreno Glacier
L296: You are describing figure 5 showing ice thickness, so the bedrock morphology description is not coherent with the figure. You should mention the channelized bed topography referring to figure 7. Please take care of the coherency between figures and text.
L300: please show previous measurement in fig 5a
L303: same as L69
L309: same as above
L325: REFs, which ones?
L323-327: I think a comparison with previous dataset is highly valuable at this point, as I said before. I agree with Reviewer 1 that this is one of the values of your measurements.
L330: in line 290 you say that the maximum ice thickness for Perito Moreno Glacier is 706m, not 681m.
L333: as said before, discussion about bed morphology should refer to figure 7
Figure 6: I suggest to put the flight path on the glaciers or at least the lines where the bedrock was interpreted. Also, contouring lines might help with the visualization.
L349: I am not sure what “constriction” means here?
L375: the system is optimized for operation over temperate ice, but at this point one would be curious about details. This info should be in the method section where you describe the system. The scattering due to water has a higher impact than the lateral reflection. You never mentioned the lateral reflections due to steep rocky walls; it is a common problem with airborne survey. Did you have any? Did you correct them with migration? If so, this should be said in the methods when describe processing steps.
L385: these issues during acquisition should be first said in the methods when you describe the survey. Then, they can be mentioned again here in the discussion when you discuss the attenuation of the signal.
Figure 8: from figure 5 to 9, they look almost identical in the layout. It makes the manuscript a bit linear. I would put figure 8 in the appendix, you can describe GPR uncertainty in the dedicated section referring to figure 8 in the appendix. I would leave figure 9 in the main text.
Citation: https://doi.org/10.5194/essd-2025-678-RC2 - RC3: 'Comment on essd-2025-678', Andres Rivera, 01 Apr 2026
Data sets
Ground Penetrating Radar Data, Ice thickness fields and bedrock topography maps of Perito Moreno, Viedma and Upsala Glacier Moritz Koch et al. https://doi.org/10.5281/zenodo.17802904
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 383 | 265 | 26 | 674 | 88 | 26 | 24 |
- HTML: 383
- PDF: 265
- XML: 26
- Total: 674
- Supplement: 88
- BibTeX: 26
- EndNote: 24
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Review comment on “Bedrock topography and ice thickness distribution of three major Patagonian outlet glaciers unveiled by helicopter-borne ground penetrating radar” by Moritz Koch et al.
- General comments
This manuscript reports ice thickness and bed elevation data for three lake-terminating glaciers in the Southern Patagonia Icefield. The data sets are based on helicopter-borne ice-radar measurements and numerical modeling of ice thickness distributions. Observational ice thickness data are sparse in Patagonia due to environmental challenges (severe weather conditions, heavily crevassed glaciers, large amounts of meltwater) and logistical constraints (glacier access, helicopter availability, material transport). The authors conducted helicopter observations over three large, rapidly changing glaciers on the eastern side of the Southern Patagonia Icefield. The obtained field data were used to estimate glacier ice thickness distributions using a numerical model. Also presented are error analyses of the ice thickness from the radar measurement and modeling.
The radar-measured ice thickness data are novel and important to the glaciological community. The presented field data cover critical areas of three large glaciers that have been the focus of glacier studies in the region. Our understanding of the rapid changes in these glaciers will be improved by ice-thickness and bed-geometry data. The modeled ice thickness and bed elevation maps are very useful for studies in the future. For example, they create a possibility of reliable ice flow modeling, which will contribute to more accurate projections of glacier mass loss and sea level rise. For these reasons, I congratulate the authors on obtaining the valuable data and strongly support the publication of the manuscript and the data in Earth System Science Data.
I list some suggestions to improve the presentation of text and data, from the perspective of a glaciologist engaged in the region. I hope the authors receive more specific guidance from other reviewers on the technical aspects of radar data processing and ice thickness reconstructions.
- Specific comments
The measured and modeled ice thickness and bed elevation are presented on maps with color scales. They show nice overviews, but I have a problem with the maps when detailed features are discussed in the text. For example, the subglacial ridge near the front of Glaciar Perito Moreno is mentioned several times in the text (Line 292, Line 333), but it is not clear which part of the map is referred to. Additional enlarged maps focused on the regions of interest would be very useful for readers. Please also consider using a discrete color scale and/or contour lines, because continuous color scales are less useful for quantitative discussion (Figures 6 and 9).
The bedrock elevation maps (Figure 7 and Supplementary Figure 1) are very useful. It would be more useful for glacier modeling, if lake bottom elevations near the glacier fronts are included. I am proposing this idea because we have published the lake bathymetry near the glaciers studied in this paper (Sugiyama et al., 2016). The bathymetry data are available online.
Sugiyama, S. et al. (2016), Thermal structure of proglacial lakes in Patagonia, J. Geophys. Res. Earth Surf., 121, 2270–2286, https://doi.org/10.1002/2016JF004084.
Sugiyama, S. et al. Bathymetry data from glacial lakes in Patagonia, Mendeley Data, V2, https://doi.org/10.17632/47rhmznjs6.1
Details of the ice thickness modeling are missing. I assume the modeling requires various input data and parameters, including ice velocity, mass balance and sliding rate. However, nothing is clear from the manuscript. Please consider describing sufficient details, including how the ice thickness from the radar survey was incorporated into the model.
I think the text is not very smooth in some places. Please get help to improve your English.
- Specific comments
Line 13: "three of the largest outlet glaciers" >> According to (DeAngelis, 2014), Glaciar Viedma, Upsala and Perito Moreno are 2nd, 4th and 12th largest glaciers in terms of area. I agree that they are well known and important glaciers, but I think "three of the largest" is misleading.
Line 14: "116.021" >> I understand "." is used as a thousands separator, but it is confusing because it's used as a decimal point elsewhere.
Line 22-23: "the well-constrained glacier-wide grids" >> What do you mean?
Line 32: "(Bojinski et al., 2014)" >> I think you need some other citations to state that glacier changes are clear indicators of "anthropogenic" climate change. Something like these?
Ben Marzeion et al. 2014. Attribution of global glacier mass loss to anthropogenic and natural causes. Science, 345, 919-921. https://doi.org/10.1126/science.1254702
Roe, G. H., Christian, J. E., and Marzeion, B. 2021. On the attribution of industrial-era glacier mass loss to anthropogenic climate change, The Cryosphere, 15, 1889–1905, https://doi.org/10.5194/tc-15-1889-2021
Line 40: " Patagonian Andes" >> Because you have just introduced, you'd better discuss "SPI and NPI".
Line 47-48: "Bathymetric surveys of recently retreated outlets" >> Sounds odd to me because bathymetric surveys are performed in lakes and the ocean, but recently retreated outlets mean outlet glaciers.
Line 49: "Minowa et al., 2023b" >> This appears before Minowa et al., 2023a. Please check the citations of Minowa's papers. It looks 2021a and 2021b, and 2023b and 2023c are duplicated (Line 561-574).
Line 60: "outlets" >> outlet glaciers?
Line 64: GPR is already introduced in Line 61.
Line 65: SPI is already introduced in Line 39.
Line 68: "116.021" >> 116,021 or 116021
Line 79: "expanding" >> extending?
Figure 1, 1st line in the caption: "rotated by 90 degree" >> Isn't it (b) which is rotated from north up?
Line 97: "Three of the largest": I think this is OK because they are 1st, 2nd and 4th largest glaciers on the eastern side of SPI.
Line 101: "approximately 896 km^2 and 779 km^2": Please revise to make it clear that these areas are for entire glaciers but not for accumulation zones. "approximately" sounds odd because 896 and 779 are quite precise numbers. The numbers are not consistent with previous studies (De Angelis, 2014, 974 and 647 km^2) (Lo Vecchio et al., 2024, ~645 km^2 for Upsala). How do you measure the areas?
Line 114: Please consider referencing Minowa et al. (2026), as well.
Minowa, M. et al. 2026. Triggering mechanisms of dynamic mass loss at a freshwater-calving glacier in southern Patagonia, Earth and Planetary Science Letters, 681, 119930. https://doi.org/10.1016/j.epsl.2026.119930
Line 116: "a single borehole" >> Actually, it was a single location, but we drilled two boreholes (Sugiyama et al., 2011).
Line 119: Move (Millan et al., 2019) to Line 119 after "… airborne gravity survey."
Line 123-125: This is already mentioned in Line 66-68.
Line 127-133: What about setting up a table to present the numbers associated with the three glaciers?
Line 143: "6 x 4 x 1 m">>"6 x 4 x 1 m3" or "6 m x 4 m x 1 m".
Line 144: "comparatively" >> What are you comparing with?
Line 159: Define "GNSS".
Line 163: "multi-frequency" >> " dual-frequency"?
Line 164: Please indicate the locations of the base stations in Figure 5.
Line "164-165": "a rover antenna was …" >> Already mentioned in Line 159.
Line 166: Abbreviations (TBC and PPP) are not necessary because they are not used.
Line 167: "rover trajectory were" >> rover trajectories were?
Line 170: "GPST–UTC offset " >> What do you mean by this?
Line 206: " Between kilometer 4 and 5 " >> Sounds odd to me.
Line 215: (2016)
Line 217: Isn't the radical sign too short? Don't you need numbers for the equations?
Line 220: "acquisition velocity" >> What do you mean?
Line 223: Please define the variables, x, y, H, ipsilon.
Line 224: What do you mean by "picks"?
Line 229: "picking uncertainty" >> Do you mean something like "uncertainty in time of measurement"?
Line 228: Please use italic for variables, c and t. You do not need to define "TWTT".
Line 237: "WGMS)" >> WGMS
Line 246-247: Omit "with the air layer removed prior to calculation".
Line 257: I suggest "thickness and bed elevation maps for entire glacial basins (for the year 2000)".
Line 266: I suggest " Comparing the modelled and observed ice thickness values (Appendix Fig. ...".
Line 267: How do you compute "misfit"? What do you mean by "The triangular model mesh"?
Line 276: " we perturbed the surface mass-balance gradient within plausible bounds" >> Please give values for the "plausible bounds". Anyway, I encourage the authors to describe more details about the modeling procedures.
Line 287: Do you mean that bed elevation was not available for 41% of the survey routes over the glacier?
Line 289: What about setting up a single table to present these numbers associated with each glacier?
Line 290: Please indicate in the plot the location of the measured maximum depth.
Line 292: Please indicate in the plot the location of the ridge. I encourage the authors to provide a larger plot of key areas, including the region near the front of Glaciar Perito Moreno. The following sentences describing ice thicknesses (180, 220, 130, 140 m) are difficult to follow without such an additional plot.
Figure 5, 2nd line in the caption: What do you mean by "densest survey grid was collected"?
Line 304: "The mean ice thickness" >> Is this the mean along the flight route or the mean of available data points?
Line 306: "stepped glacier bed" >> Can you indicate in the plot which region you are referring to? An enlarged plot helps readers.
Line 323: "small-scale" >> How small?
Line 325: Why don't you compare your measurements with existing data? Because such data sets are widely used, it is essential to validate them with observational data. That's one of the values of your measurements, I think. For example, Minowa et al. (2021) used inverted ice thickness for ice flux calculation, and Lannutti et al. (2024) studied the thickness distributions of the three glaciers that you studied. By the way, what do you mean by "REFs"?
Line 327: "2D reconstructions" >> the reconstruction of bed elevation?
Line 333: Please show "subglacial ridge" on the plot.
Line 335: "1.292 m" >> 1292 m Where is the deepest part of the glacier?
Line 337: "well constrained" >> What do you mean by this? Based on which data?
Line 340: "1.400 m" >> 1400 m
Line 346: "in the Supplement" >> Refer to the figure number.
Line 350: "After" >> Above?
Line 347-Line 359: Please consider providing additional maps focused on the regions near the terminus.
Figure 7, 3-4th line in the caption: Omit "areas at or below".
Line 367: ... cross points of survey routes?
Line 388: Please clarify that Figure 9 shows the uncertainty of the ice thickness obtained by modeling.
Figure 9: Please indicate the regions covered by the radar survey because the uncertainty should be directly affected by the existence of field data.
Line 409: Omit "densely gridded".
Reviewed by Shin Sugiyama