the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 16 Jul 2025
Kinetic Grain Growth in Firn Induced by Meltwater Infiltration on the Greenland Ice Sheet
Abstract. The microstructure of polar firn governs its porosity, permeability, and compaction rate, and is thus critical to understanding surface elevation change, heat and gas exchange, and meltwater infiltration on ice sheets. Previous studies in high-elevation dry firn have identified two atmospheric drivers of kinetic grain growth, though both produce only millimetre-scale layers near the surface. Here, we demonstrate that meltwater infiltration in the percolation zone of the Greenland Ice Sheet (GrIS) produces centimetre- to decimetre-scale layers of kinetic grain forms, ranging from faceted crystals to depth hoar, persisting to depths of up to 16 m. We analysed subsurface temperature time series from a transect on the western GrIS to resolve thermal regimes associated with infiltration-driven kinetic grain growth. Two distinct mechanisms responsible for faceting were identified: one associated with the onset of the wet layer, the other with preferential meltwater flow events. For both mechanisms, elevated vapor fluxes were calculated and diminished grain sphericity was observed in SNOWPACK model simulations, implying each can facilitate kinetic grain growth. Wet layer onset was the dominant mechanism, producing pronounced reductions in sphericity and the most enduring faceted layers. Additionally, the rate of wetting front propagation influenced the longevity of faceted layers, with rapid infiltration preferentially producing lasting, lower-sphericity firn grains. As surface melt expands across the GrIS, constraining the influence of these faceted layers on meltwater storage, surface elevation change, and chemical transport will become increasingly important.
- Preprint
(2548 KB) - Metadata XML
-
Supplement
(370 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Review of Gehl et al. on meltwater-induced kinetic grain growth in firn', Peter Kuipers Munneke, 13 Aug 2025
-
RC2: 'Comment on egusphere-2025-3002', Mahdi Jafari, 21 Aug 2025
- General Overview
This manuscript presents a study on kinetic grain growth in firn induced by meltwater infiltration on the Greenland Ice Sheet. The authors provide new and interesting observations that extend previous understanding of kinetic grain growth, which has typically been confined to thin, near-surface layers of millimeter-scale thickness. Here, the authors report the presence of faceted layers up to 1 m thick and preserved at depths as great as 16 m, which is a novel and noteworthy finding.
The work is timely and relevant, as it sheds light on the interplay between meltwater infiltration, temperature gradients, and firn microstructure evolution in the percolation zone. The proposed mechanism—that meltwater infiltration advects heat into cold firn and establishes strong local temperature gradients sufficient to drive kinetic grain growth—is plausible and well-motivated. The integration of field observations with SNOWPACK modeling provides a valuable perspective, although some aspects of the methodology and model assumptions need further clarification.
Overall, the manuscript addresses an important gap in the literature by linking microstructural evolution of firn to surface processes, meltwater dynamics, and densification models. With additional clarification and refinement in key sections, this study could make a significant contribution to the understanding of snow and firn processes on the Greenland Ice Sheet.
- Major/minor Comments
- It would be helpful if the authors could provide a short description of “kinetic grain growth” in the Methods section. This would give readers a clearer understanding of the physical process that governs kinetic grain growth.
-
Lines 30–31: Can you clarify if you are referring to the wind-compaction process observed in Arctic snow covers, which leads to a hard slab forming on top of the depth hoar layer? What is the timescale of forming a sub-millimeter hard slab compared to faceted layers that result from kinetic grain growth? This paragraph states that diffusive vapor transport drives the rounding process, leading to hard slab burial several meters deep with rounded grains of ~2 mm radius. Could the authors also check whether other types of water vapor transport, such as convection, may occur in this context?
-
Lines 42–44: Can you please clarify the significance of convection-driven water vapor transport in the Greenland Ice Sheet? Jafari et al. (2020, 2022, 2023) have shown that if convection persists over long timescales, it can substantially decrease density at the bottom of the snowpack and increase density at the top. Could the authors be more explicit about which types of water vapor transport drive the kinetic grain growth rate and discuss the relevance of convective vapor transport in this context?
References:
Jafari, M., Lehning, M., and Sharma, V. (2022), Convection of water vapor in snowpacks. J. Fluid Mech.
Jafari, M., Gouttevin, I., Couttet, M., Wever, N., Michel, A., Sharma, V., and Lehning, M. (2020), The impact of diffusive water vapor transport on snow profiles in deep and shallow snow covers and on sea ice. Frontiers in Earth Science.
Jafari, M., Lehning, M. (2023), Convection of snow: when and why does it happen? Frontiers in Earth Science, 11, 1167760. - Lines 80–83: For the lower elevation, the temperature measurements show that the summer heatwave penetrates more deeply. The manuscript explains this as (1) due to increased latent heat — is this because of warmer air at lower elevations? — and (2) due to higher thermal conductivity from greater firn density. Could the authors clarify which of these is the main driver? Also, what is the reason for the larger density at lower elevation — could this be related to stronger katabatic winds favoring wind compaction?
-
Section 2.4, SNOWPACK simulations: Can you please clarify whether you used a SNOWPACK version that includes the vapor transport scheme? If yes, could you provide a comparison between the vapor flux calculated from the measured temperature gradients and the flux simulated by SNOWPACK?
-
Line 158: Please correct the citation “with mass transport by vapor flow (Lehning et al., 2002a)”. The more recent reference for SNOWPACK with water vapor transport is:
Jafari, M., Gouttevin, I., Couttet, M., Wever, N., Michel, A., Sharma, V., and Lehning, M. (2020), The impact of diffusive water vapor transport on snow profiles in deep and shallow snow covers and on sea ice. Frontiers in Earth Science.
-
Lines 162–164: Regarding the idealized simulations, it is not clear how using meteorological forcing with an annual wavelength is justified. Please elaborate on this assumption, especially since it is explicitly stated that a mean annual value is set for the firn temperature.
-
Line 167, Table S2: Is LWR a net value (i.e., LWR = ILWR − OLWR, where ILWR = incoming longwave radiation and OLWR = outgoing longwave radiation)? It is not clear what is meant by “heat fluxes during the spinup to retain a vertical deep-firn temperature profile at −16.5°C.” Could the authors clarify this statement?
-
Lines 169–171: Can you please confirm that the data were averaged from 1995 to 2020 (25 years) at each time step (hourly resolution)? Does this averaging then provide the time series for air temperature, shortwave radiation, and longwave radiation used as forcing data?
-
Lines 254–256: This appears to contradict what I observe in Figure 8, where the sphericity increases after reaching its maximum reduction. Could the authors clarify and be more specific about the sphericity reduction shown in Figure 8?
-
Line 281: It seems that the authors are referring to a different plot, not Figure 7, since Figure 7 is about wet layer onset as the faceting mechanism. Could the authors clarify this reference?
-
Caption of Figure 7: I think both columns should belong to the same simulation setup. Please specify which faceting mechanism is shown — it should be the wet layer onset, as this plot corresponds to Section 3.3. What causes the transient change seen in the black box in Figure 7(c)? Is it a numerical artifact or a physical process? Could the authors add another row showing the time series of grain types, which would be very informative for readers? SNOWPACK simulations include grain type as an output that can be used for post-processing.
-
Line 287: Please indicate that subplots (E) and (F) of Figure 8 correspond to the “Preferential flow/piping” scenario. Change the reference from “(Fig. 8)” to “(Fig. 8 E and F)”.
-
Lines 316–319: This is very important regarding densification models. Faceted grains have a very different densification pattern compared to dry firn conditions. Do the SNOWPACK simulations include densification specifically elaborated for faceted grains? If not, please mention this limitation for snow models such as SNOWPACK.
Citation: https://doi.org/10.5194/egusphere-2025-3002-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 803 | 29 | 14 | 846 | 18 | 33 | 36 |
- HTML: 803
- PDF: 29
- XML: 14
- Total: 846
- Supplement: 18
- BibTeX: 33
- EndNote: 36
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Review of Kinetic grain growth in firn induced by meltwater infiltration on the Greenland Ice Sheet, submitted for publication in The Cryosphere by Kirsten L. Gehl, Joel T. Harper, and Neil F. Humphrey
===
In this manuscript, an explanation for the occurrence of faceted grains in Greenland firn up to a depth of 16 metres is sought. Traditional mechanisms of faceted grain formation focus on processes close to the surface and in cold locations, but none are able to explain the formation and occurrence of faceted crystals deeper down in the firn, and in locations where frequent surface melt occurs. As far as I can judge, this is the first attempt to explain faceted grain formation in the Greenland percolation zone. It's an interesting and original subject, fit for The Cryosphere and its readership.
The central hypothesis is that meltwater infiltration (either as a wetting front or by means of preferential flow in pipes) creates very strong temperature gradients within the firn, between the temperate wet snow and the cold firn below the wetting front or surrounding pipes of preferential flow. Such temperature gradients act as hotspots for kinetic growth of faceted crystals.
This is an interesting hypothesis, made plausible by SNOWPACK modelling and by an inventory of in-situ observations of faceted crystals spanning a large range of elevations and time.
I do have some concerns about (1) the presentation and analysis of the observational data; (2) the modelling setup; and (3) the effectiveness of the figures. It is unlikely that these concerns will wipe out the central hypothesis of the paper. But if care is taken to clarify and work out a few points, it will make for a better and even more convincing paper.
===
(1) analysis of observational data
A critical temperature gradient (CTG) is introduced, defined as a temperature gradient whose absolute value equals 10 deg C m^-1. My feeling is that the statistics about the exceedance of this CTG in the observational record (presented for example in figures 4 and 6, and section 3.2) depend quite a bit on the vertical resolution of the temperature observations. Firn temperature is recorded with a vertical resolution between 0.125 and 1 m, limiting the ability to observe the highest temperature gradients. How does vertical resolution affect TG observations? How are the statistics in figure 6 affected by it? If you downsample a record at 0.125m resolution to 1m,
The statistics in figure 6 are a bit unclear to me. Is the cumulative CTG duration (panels 6a and 6b) specified per year? Or in any other way normalized? If the duration depends on the total length of the data set and/or the completeness of the data, then how to compare apples to apples?
In general, I feel that the occurrence of faceted grains is nicely presented in figures 2 and 3, and the statistics of large TGs in the subsequent figures. But is there a way to connect these two types of observations more precisely, if only for a case study or two? I.e., tying together observations of faceted grains in a particular firn core to the TGs of the preceding year at that location? Such an illustration would strengthen the observational link between TG and faceted grains.
(2) the SNOWPACK modelling setup
When introducing SNOWPACK in section 2.4, it remains unclear what physics are implemented in the model that ensure that faceted grain growth is simulated reliably. How is sphericity evolution modelled?
Regarding the atmospheric forcing of the model: is 2-m air temperature or surface temperature used from ERA5? How well are strong near-surface air temperature gradients resolved within the atmospheric boundary layer? How reliable is ERA5 surface temperature?
In the synthetic wetting front simulation, two superimposed sinusoidal curves represent annual and daily cycles in surface temperature. But nowhere does this temperature exceed the melting point. How, then, is surface melt produced? Have you checked the surface melt volume in the model with observations? Does the surface energy budget in SNOWPACK make sense? Would it be possible to plot a time series of surface melt along with figure 7?
(3) Effectiveness of figures
I find figure 5 quite uninteresting and irrelevant. These temperature profiles are monthly averaged, likely eliminating most TGs exceeding the CTG. In fact, the highest TGs are near the surface, but this is no support for the hypothesis of meltwater-induced TGs. Indeed, only in May is there a temperature gradient exceeding the CTG, but that is before the melt season and not in the location where meltwater-related faceted grain growth occurs. So I suggest replacing this figure by a figure with a few examples of instantaneous observed temperature profiles that illustrate TGs around an observed wetting front, a piping event, both positive and negative. It would be far more elucidating what kind of TGs are typically encountered in a melting firn pack than the current figure 5.
As mentioned, are the statistics in figure 6 independent of sample size and record length?
Figure 7 is a great figure! I suggest to explicitly mark the wetting front and piping episodes in the figure, just with a stylized line and word labels in the figure.
Figure 8 nicely shows how the wetting front is wiping out faceted grains. This process, which is casually mentioned in line 227, is crucial in understanding that the actual occurrence of faceted crystals depends on TG around infiltrating meltwater, but also on the presence of meltwater itself. Perhaps interesting to expand this figure with a infographic, or like suggested for figure 7, to annotate panel b with a line of text that explains what is happening.
I'm not sure if the division into a main text and a supplement is really necessary for this paper. The paper is not really long, and the normal methods section seems a fine place to put all the information of the supplement into.
===
Minor remarks
L. 9: Be specific here. Instead of "two atmospheric drivers", write something like: "Kinetic grain growth is known to occur when very large temperature gradients cause vapor transport within the firn. In previous work, warm wintertime winds and summertime absorption of sunlight have been identified to cause such large gradients. Here we demonstrate that meltwater infiltration can also cause such gradients, explaining observations of faceted grains in the percolation zone, and at greater depth. "
L. 25: The best way of citing Amory et al. is "The Firn Symposium Team, 2024". Charles Amory is lucky to have a name starting with A :-), but no specific author order is implied in the author list of that paper.
L. 39: Here is a good opportunity to describe in one or two lines what kinetic grain growth precisely is, and how it is physically different from rounding.
L. 51: remove "later"
L. 57: In a separate paragraph, here is a good place to put forward more explicitly your central hypothesis, namely that meltwater is also a cause for enhanced temperature gradient driven kinetic grain growth.
L. 71. The last sentence is more for a conclusions section (and is already mentioned in the abstract too). Better remove here.
L. 290. This deserves some more attention. Is kinetic grain growth by wetting front advance more important / more frequent than by preferential flow? Is there any firn core evidence that the former mechanism explains the majority of faceted crystals?