the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 28 Aug 2026
Quantifying 3D topographic effects on geothermal heat flow beneath the Northeast Greenland Ice Stream
Abstract. The North East Greenland Ice Stream (NEGIS) is the largest ice stream in Greenland, and previous studies debate the reasons for its initiation so far into the interior of the ice sheet. Here, we explore how important 3D geothermal heat flow (GHF) is to effective heat transport as a contributing factor. We estimate the effect of high-resolution bed topography on GHF using the finite-element framework pyGIMLi to perform 3D thermal simulations. The bed topography is derived from recent airborne ice-penetrating radar data and covers an area of approximately 40 x 60 km² surrounding the EastGRIP ice core site at the onset of the ice stream. As large-scale GHF models currently show substantial discrepancies in north-east Greenland, we evaluate the 3D effects for two contrasting background GHF models. Our results show that incorporation of high-resolution bed topography leads to local GHF deviations of up to 20 % relative to background models, closely following topographic patterns. Therefore, incorporating high-resolution GHF, especially in regions with a large ice flux, could reduce uncertainties in our understanding of the origin and stability of ice streams. However, for the NEGIS itself, our preferred background model of low GHF implies that topographic effects channeling GHF do not significantly alter the behavior of the ice stream.
- Preprint
(4789 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 09 Oct 2026)
- RC1: 'Comment on egusphere-2026-4746', Magnus Tumi Gudmundsson, 01 Oct 2026 reply
-
RC2: 'Comment on egusphere-2026-4746', Mareen Lösing, 09 Oct 2026
reply
This manuscript investigates how high-resolution subglacial bed topography modifies geothermal heat flow (GHF) beneath the onset region of NEGIS. The authors compare three-dimensional conductive thermal models with one-dimensional calculations for two contrasting background GHF models, and evaluate the results against an empirical topographic correction. The study addresses a relevant question for basal thermal boundary conditions and makes use of the recently available swath-radar bed elevation dataset.
The comparisons demonstrate that topography can introduce substantial local variability while the background GHF model largely controls the regional mean. The manuscript is generally well structured, and the comparison between modelling approaches is valuable. However, some aspects of the model construction and numerical implementation need clearer explanation, and some figures and terminology require revision.
Main comments
Lines 79–83
Please explain explicitly how the laterally variable radiogenic heat production is calculated from each background GHF model. An equation or a description of the calculation would make this step reproducible. In particular, clarify the surface elevation or reference geometry used when bed topography is excluded, and how Moho and LAB depths enter this calculation.
Section 2.1
The manuscript gives the analytical 1D expression but does not explicitly state the governing equation solved in 3D. Please include it, together with the relevant assumptions and boundary conditions. The lateral resolutions at the bed, Moho and LAB are provided; please also describe the vertical discretization or element sizes, particularly near the bed where temperature gradients are evaluated. Also showing the swath-radar survey coverage in Figure 1 would help readers understand the domain selection. If the model domain is given directly by the survey outline, then it would be good to mention this somewhere.
Line 75
Please clarify that the prescribed 0°C is an approximation to the basal pressure melting point and neglects its dependence on ice thickness.
Figure 1C
Here it would help to specify the interpolation method used to obtain the 500 m grid and to mention which topographic dataset was used. A comparison of the 1 km and 500 m topography grids would help demonstrate the extent to which the interpolation affects this experiment.
Section 3 and Figures 4–6
Please make explicit where you are referring to the 3D results relative to the corresponding 1D calculations. Figures 4 and 5 appear to repeat the same 3D GHF maps; please explain their distinct purposes or consider consolidating them. Adding the corresponding GHF–bed elevation correlation plots for the 1D cases would help show how the relationship changes in 3D. Please also add the profile locations to Figure 4 and/or Figure 5, so readers can connect the maps with Figure 6.
The manuscript discusses the flexibility of the 3D framework and notes that finer resolution substantially increases computational cost. It would be very interesting if you could provide indicative runtimes for the 1D and 3D approaches and for the 1 km and 500 m calculations, together with basic hardware information.
Minor comments/corrections
Lines 26–30
Clarify that the studies suggesting more moderate GHF values don’t constrain or reproduce observed ice-flow velocities. This would make the comparison with Smith-Johnsen et al. clearer.
Lines 31–37
It would be nice to give the magnitudes and uncertainties of the 3 different NorthGRIP estimates.
Lines 56–58
Check the Colgan citation: the background GHF model is described elsewhere as Colgan et al. (2022), whereas Colgan et al. (2021) provides the topographic correction.
Section 2
Consider “Methods and Data,” matching the order of the subsections.
Figure 1
The geology is difficult to distinguish against the coloured bed elevation. Consider muting the topographic background or using greyscale.
Line 107
Check the panel references: the Moho and LAB depth maps are panels (b) and (c).
Lines 111–116
Delete " without NGRIP" from the sentence: “We use the GHF model of Colgan et al. (2022) without NGRIP as the observed high ice borehole measurement at NGRIP,..."
Specify that the five ice-core locations used by Greve include NGRIP.
Line 113
Consider “lower-GHF” and “higher-GHF,” or “relatively cold” and “warmer,” to make clear that the model labels are comparative.
Line 121
Could merge “results for the cold and warm models”
Line 131
Add the missing space before “(Fig. 5(b), (c))”.
Line 131
Remove brackets here “across (Fig. 6(a)-(d)) along (Fig.6(e)-(h))”.
Line 141
Should be Figure 6 instead of Figure 5.
Figure 5 caption
Replace the second “(d)” with “(e).”
Equation (2)
I would place the melt-rate unit in the surrounding text rather than attaching them to the variable within the equation.
Figure 6 caption
Check whether the line described as “thin dashed” is actually solid, and identify explicitly whether the temperature sections are from the 3D model.
Lines 153–155
Specify that “small-scale variations” refers to variations in GHF.
Line 164
Replace the semicolon between the Bullard and Jeffreys references with “and”.
Code and data availability
Correct “Morlighem and et al.” and check the formatting of the pyGIMLi citation.
Appendix A, Lines 263–265
Cite Figure A2 explicitly when discussing the correlation plots.
References
Update the ISMIP7 community recommendations reference to the published version:
Lösing, M., Colgan, W., Stål, T., Ebbing, J., Busck, A. G., Zhang, T., Seroussi, H. L., McCormack, F., Fahrner, D., Stearns, L., Svendsen, S. H., & Reading, A. (2026). Community heat flow recommendations: suitable basal boundary conditions for Greenland and Antarctica in ISMIP7. GEUS Bulletin, 62. https://doi.org/10.34194/r0w9rf81
Citation: https://doi.org/10.5194/egusphere-2026-4746-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 229 | 121 | 49 | 399 | 50 | 47 |
- HTML: 229
- PDF: 121
- XML: 49
- Total: 399
- BibTeX: 50
- EndNote: 47
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Quantifying 3D topographic effects on geothermal heat flow beneath the Northeast Greenland Ice Stream
Frienstein et al.
Review comments by:
Magnús T. Gudmundsson, 28 September 2026
This is a most useful study where detailed modeling is used to analyse the possible/likely causes for an ice stream in the northern part of the Greenland Ice Sheet. Overall, the work is sound and I recommend acceptance after minor revisions. I have no major comments on this work but several minor comments/suggestions to improve the clarity and correct some very minor typos/etc.
The only comment that may not qualify as (very) minor, is a suggestion on figure 1 – nothing wrong with it, but it is very difficult to find the geological units listed. I therefore suggest that the panels will not be two, but three, where the first panel shows the bed topography, and a panel is inserted (the same size) in the middle, where the geological units are shown. This will make the figure much easier to read. Also, it would do no harm to mention in the caption the age span of these formations – how old are the rocks in Greenland? Would make the figure/paper easier to read.
Minor comments:
Line 2-3:
„Here, we explore how important 3D geothermal heat flow (GHF) is to effective heat transport as a contributing factor.“
Suggest rewording slightly for clarity:
Here we explore how important 3D subglacial topography may be in modyfing goethermal heat flow (GFH) in this area. (or something similar)
Line 45-53: Consider adding a sentence here stating how topography affects heat flow – not everybody may be familiar with this mechanism (e.g. explained in Colgan et al. 2022). This is also addressed in Appendix A – should be referred to here.
Line 78: For clarity: add k1 and k2 in brackets after crust and mantle:
The thermal conductivities for the crust (k1) and the mantle (k2) are constant with k1 = 2.7Wm−1K−1 and k2 = 3.0Wm−1K−1.
Line 142: ….pattern is visible at greater depths and….
Figure 6: Something not right here. The temperature profiles are in Kelvin, but that cannot be correct. Should this be °C?
Line 164-165: Were these observations (empirical data) obtained by Bullard and Jeffries? Or where these results of calculations? If the latter is true change wording to make clear.
Also, should this not be …..Bullard (1938) and Jeffries (1938).
198-199: The statement that friction is a significantly greater contributing factor is very important. How much larger would this be? Also, friction leading to higher melting, would the higher melting/supply of meltwater lead to reduced friction? It would be good to add a sentence or two explaining this better, possibly adding what the studies (by others?) indicate on melting/heating in this situation? This is not in any way invalidating your results, but would be helpful to give a better picture.
A general comments on the figures (minor but relevant). The extensive use of abbrevation in figures and caption, makes them more difficult to understand. The majority of people who go through papers read the title and abstract and scan through the figures. It is therefore very helpful to make the figures and caption as clear as possible and is also more likely to keep the attention of the readers. Consider adding to the captions:
Fig. 2: Here LAB depth is shown. Add to the caption …..(c) the depth to the lithosphere-astenosphere boundary (LAB) in the area.
Fig.3: (a) The geothermal heat flow (GHF) from Colgan….. (d) the calculated radiogenic heat production (RHP) for the warm model….
Fig. 4: ….(a) the calculated 1D and (b) 3D cold geothermal heat flow (GHF) model…..
Fig. 5: Geothermal heat flow (GHF) from the 3D….
Fig. 6: Perhaps not strictly necessary – but helpful nevertheless: …..and (orange, thick dashed) geothermal heat flow (GHF) models.
Fig. 7: …and (b) the 3D cold geothermal heat flow (GHF) models…….
Fig. 8: ….(a) the cold 3D geothermal heat flow (GHF) model…….
Same comment on Fig. A1 – GHF – geothermal heat flow (GHF)…..
Appendix A – what about using this title:
Appendix A: Relative contribution of 3D topography to geothermal heat flow
Line 248: consider phrasing as: We calculate the relative contribution (DG) of 3D effects……
Appendix B:
Figure A2 (b) – horizontal axis label: should this be Topographic correction (DG)?
Appendix C:
Line 276: Rephrase: The results indicate that by doubling the final resolution….. (or similar)