the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 20 Nov 2023
Greenland's firn responds more to warming than to cooling
Abstract. The porous layer of snow and firn on the Greenland Ice Sheet stores meltwater and limits sea level rise. This buffer is threatened in a warming climate. To better understand the nature and timescales of firn’s response to air temperature change, we use a physics-based model to assess the effects of atmospheric warming and cooling on firn air content. We identify an asymmetric response of Greenland’s firn to air temperature: firn loses more air content due to warming compared to the amount gained from commensurate cooling. In dry firn, this asymmetry is driven by the highly nonlinear relationship between temperature and firn compaction, as well as the dependency of thermal conductivity on the composition of the firn. The influence of liquid water accentuates this asymmetry. In wet firn areas, melt increases nonlinearly with atmospheric warming, thus enhancing firn refreezing and further warming the snowpack through increased latent heat release. Our results highlight the vulnerability of firn to temperature change and demonstrate that firn air content is more efficiently depleted than generated. This asymmetry in the temperature–firn relationship may contribute to the overall asymmetric mass change of the Greenland Ice Sheet in a changing climate across many timescales.
- Preprint
(5867 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2023-2629', Anonymous Referee #1, 04 Jan 2024
1 Overview
==========This is a modeling study to investigate the response of firn to
changes in air temperature. The investigators run a simple
step-forcing experiment to detrermine the spatial and temporal
responses of a simulated firn column over Greenland, to step changes
of 1 degree C in both warming and cooling. The experiment and
procedures are well-described and the paper is generally well written.
Many interesting insights are illustrated in the paper, and it looks
to be an interesting contribution to the literature on firm in a
warming world. While the main conclusions may be unsurprising to
whose who think about firn a great deal, the conclusions are
well-founded and are well worth publishing. I have mainly a few
relatively minor issues with the existing study and they should be
straightforward to address.
2 Minor issues
==============- Timescale of the experiment: As the authors note (lines 151-152 and
165), the model in the perturbed state (both warming and cooling)
comes _close_ to equilibrium (as shown in figures 2 and 7 for
example) but doesn't quite get there. It would definitely be a more
satisfying experiment to run these models to equilibrium. That
would also allow an additional result- the quantification of the
total change in FAC for the changes, and the total response time to
equilibrium (bringing home the point the authors make about the
dependence of response time on mean state climate- in lines
165-167).
- Additinal experiments: It seems that once the forcing framework is
in place and the perturbations are set for 1 degree, it would not be
difficult to 'turn the crank' on other perturbation experiments;
most simple would be doubling the warming or cooling, but another
more interesting question would be what is the magnitude of cooling
required to match the change in FAC from the 1 degree warming
experiment? There's an extent to which the answer to the existing
experiment is self-evident to folks who study firn... additional
experiments like this might make this a more widely-cited paper.
- Appropriateness of title, given Figure 9c: at and near line 250, the
authors state that the majority (1282/1724 = 74%) of the grid cells
fall into the category labelled "Equal response" in figure 9. This
is graphically shown in figure 9c. Given this statement, does the
title (claiming unequal response) make sense? I certainly concede
the point made in lines 252 and 262 that within this 5% bound the
warming is usually greater (pinkish tint in Figure 9d). Figure 2c
makes the point (of the title) well but then it's a bit undermined
by the later discussion of figure 9c. Maybe instead of "Greenland's
firn" in the title you might insert "Greenland's marginal firn"? Or
change some of your binning here in this paragraph to make the point
in the title more clearly made? Perhaps this is pedantic but seemed
worth discussing.
3 Mainly small issues, line-by-line
===================================- Line 32: I would argue that firn air content and density are not
really separate properties in this context; just different ways of
describing the same property. Since air content is the key term
used in the title and throughout, I'd either suggest a sentence
relating the two, or leaving density out of it entirely.
- Line 33 Not sure MacFerrin et al 2022 is the right reference for the
concept of compaction (they didn't come up with it!). Not clear a
reference is needed here, but if so, please go back further in the
literature- MacFerrin et al 2022 cite many other studies like Herron
and Langway 1980, Morris and Wingham 2014, among others. If you
want to stick with references already existing in the manuscript I'd
suggest Herron and Langway 1980 and Arthern and Wingham 1998.
- Line 80: non-modelers might not know what a "bucket scheme" is;
suggest a short description.
- Line 122: By "modern Greenland Ice Sheet" I think you mean
'holocene', right? It should really reflect something like the 1991
Greenland Ice Sheet, which is definitely changed in the past three
decades (thus "modern" does not mean 2023).
- Line 139: presumably the reason the big changes are slightluy inland
from the edge is that at the edge FAC is zero to begin with
(ablation zone)? If so, this sentence may either not be needed or
could include this point...
- Figure 4: It took me a while to understand the observations made in
lines 181-182, because there's a lot going on in this figure.
Perhaps in lines 181-182 refer to the colorbar for 4c,d to remind
readers that the initial FAC in the control is indeed shown in those
panels...
- Figure 6: if the editors want to save space, this would be a good
figure to cut, as it illustrates essentially a null result which was
expected.
- Line 224 and Figure 7f: Also a more general comment about this
figure- the annual/seasonal cycles of changes in FAC are apparent
particularly in 7f but the differnces from the max in 7f and a min
in 7e are pretty stark; this is worth discussing at least a little
bit.
- Line 239: If we are genuinely in the ablation zone, the snowfall
cannot "almost completely" melt away in summer, it must completely
melt away (and thus there would be by definition 0 FAC). Call this
"here near the equlibirium line" instead of "here in the ablation
zone".
- Line 250: The fact that the large majority of grid cells fall into
the "almost equal" category sort of implies that the title of the
paper might be an overgeneralization. See my 'minor point' above on
this.
- Line 272: instead of the equivocal "impacts the bulk thermal
conductivity", state the generalized relationship (ie increasing air
content reduces conductivity)- it may be obvious to some readers but
not all; it's implied in lines 272-3 but could be made more clear to
the reader.
- Figure 10: This is a pretty complicated figure, which may be the
point, but it's challenging to follow the description in the text
along with the figure. One option would be to have letters or
numerals on each box, so that insetad of referencing 'Figure 10' in
the text, you can refer to a specific box or set of boxes in the
figure.
- All of section 3.4: It's not clear that these are actually results,
and they read like discussion. Consider putting them in section 4.
- Line 299: here you state "most regions" have a greater change in FAC
from warming than from cooling- this needs to be reconciled with the
statements around Line 250 and in figure 9 where most of the area of
the ice sheet falls into the "equal response" category. Also see my
'minor point' about the title of the paper above.
- Line 345: "While a lot can be learned" a but colloquial. Suggest
changing to "much can be learned".
- Line 547: reference to Kuipers Munneke et al 2015: Van den Brooke's
name is broken up in an incorrect way, it should be Van den Brooke,
M. R. (note I didn't check all of these references I just noticed
this one, suggest double-checking the typesetting- probably a simple
BiBTeX issue).Citation: https://doi.org/10.5194/egusphere-2023-2629-RC1 -
AC1: 'Reply on RC1', Megan Thompson-Munson, 13 Mar 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-2629/egusphere-2023-2629-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Megan Thompson-Munson, 13 Mar 2024
-
RC2: 'Comment on egusphere-2023-2629', Anonymous Referee #2, 11 Feb 2024
Review of
Greenland's firn responds more to warming than to cooling
by Megan Thompson-Munson et al.
GeneralThis is an interesting paper that investigates the response of the Greenland ice sheet firn layer to idealized positive and negative temperature perturbations on century time scales. The paper is well and concisely written, and the figures are of excellent quality. My comment are therefore relatively minor and should be fairly easy to address.
Major commentsl. 71: "SNOWPACK uses near-surface atmospheric conditions as input". This is confusing; I assume more input parameters are required to calculate the surface mass and energy balance at the upper boundary of the firn model? Such as surface radiation and mass fluxes? Upon further reading this is specified in l. 92, but please correct here to avoid confusion.
l.77: "Unlike semi-empirical firn models used in similar studies (e.g., Kuipers Munneke et al., 2015), SNOWPACK does not use a positive degree day parameterization for calculating melt." As it is written here, this is incorrect. The "semi-empirical" refers to the way snow densification is parameterised and does not refer to the way these models are forced. The study of Kuipers Munneke et al (2015) cited here used an idealized/simplified melt forcing simply to facilitate interpretation of the development of firn aquifers. Other studies by that same author use realistic surface mass and energy balance to drive the firn model, i.e. equivalent to what has been done here. There remains an important difference though. In the model setting used here, SNOWPACK calculates its own surface energy balance, whereas in most other studies the firn models are directly forced by surface skin temperature from regional climate models, being the result of the closure of the surface energy balance. Although not critical for this study, it is good to clearly separate these different approaches.
Section 2.3. The model coverage of Greenland is at relatively low resolution, and points that do not meet the spin up criteria are removed. As a result, only 1724 locations remain, an order of magnitude less than when the ice sheet would have been resolved at e.g. 10 x 10 km grid cells. This low resolution is also apparent from e.g. Fig. 1. Could you comment at how that potentially influences the (ice sheet integrated) results, and also regionally especially in regions with strong climate gradients? And can you indicate in Fig. 1 the outline of the contemporary ice sheet, so that it becomes clear which parts of the ablation zone have been removed from the analysis?
Same section and results section: you select a single year as a baseline for all your experiments. Am I correct that Figs. 1a, b, c then represent the climate of 1991? Apart from the absence of a trend in firn air content for that year (is this averaged over the ice sheet, how does it hold regionally?), can you also provide information on the temperature/accumulation/melt of that particular year relative to climatology? It appears from Fig. 1 that melt is rather low along the western marginal ice sheet (although I realise that most of the ablation zone is not part of the model domain).
Forcing experiments: if I understand correctly, only 2 m air temperature is varied, but other variables that normally are closely connected to air temperature (incoming longwave radiation, specific humidity, rainfall fraction) are left unchanged? If so, please mention this specifically to avoid confusion.
Fig. 7: What causes the fluctuations with frequency of about 1/decade in panels a and c?
Minor and/or textual commentsl. 20: asymmetric -> temporally asymmetric (to distinguish it from spatially asymmetric, which I thought was meant when first reading this sentence)
l. 25: drives meltwater runoff -> drives enhanced meltwater runoff
l. 31: limit -> limit and/or delay (in the case of aquifers)
l. 32 and following discussion: Firn air content and density are presented here as two separate characteristics, but it would seem to me that they are one-to-one coupled? Or am I overlooking something?
l. 139: "slightly inland from the ice edge". As large parts of ablation zone are not considered, the ice edge can be >100 km away. Perhaps better use "slightly inland from the equilibrium line".
Citation: https://doi.org/10.5194/egusphere-2023-2629-RC2 -
AC2: 'Reply on RC2', Megan Thompson-Munson, 13 Mar 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-2629/egusphere-2023-2629-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Megan Thompson-Munson, 13 Mar 2024
-
RC3: 'Comment on egusphere-2023-2629', Erin Pettit, 13 Feb 2024
Review of Thompson Munson - Greenland’s firn responds more to warming than to cooling.
This work presents and overview of how Greenland’s firn porosity might change under slight warming or cooling scenarios.
The modeling and results are insightful and interesting - I really just want more! But in a shorter, more concise structure, and figures that really draw the reader to the main points. I agree with several of the points of the other reviewer - showing the model reaching equilibrium would be great. And testing another magnitude of change. But I realize that is effort. Also, as a physicist, I really want to see the results tied back to the physics of firn, since this is what would allow us to take these results and transfer them to other places…
General comments:
- the abstract suggests that the work will focus on the physics of the nonlinear relationships, then discuss this in the context of Greenland, but the paper jumps immediately into the large scale integrated Greenland response without showing much of the physical reasoning. For example, how much does a steady trend in air temperature affect the firn compared to an individual warm event (one hot summer, for example)?
- at the end of the abstract/introduction, I was still trying to figure out where this paper was headed. It was unclear to me what the goals of the paper were and what methods would be used or what the findings were (other than big general statements). It would be nice to be able to extract from the abstract and introduction what is really novel about what this paper is offering and how it is getting there. I think this could be helped by just making each sentence, each paragraph a bit more specific (not longer, just more specific).
- The methods section is a bit limiting from my perspective. Perhaps this is because I like to know what physics is going on! More information is provided about the surface boundary condition than to the internal physics of the firn model - a summary of the key assumptions made by the model would be really helpful. The model also uses the same parameters as a previous paper (but I don’t have time to go read that paper) and then is compared to the results from that paper to suggest that the model is working well. I’m a bit confused as it seems to me that the statement implying validation (Line 133) of the model is just comparing the model to itself.
- Also related to methods - the goal of the paper implied trying to figure out how firn responds. But given that the results are only provided in the context of large scale changes on Greenland, I don’t know how to take the results and apply them to a mountain glacier or to Antarctica. I would suggest an approach of sweeping through a range of parameters and looking at 1D results for each, to show the changes in the snowpack that are possible. Or at a minimum, spending much more time on results such as Fig 7/8 - where a few modeled points with different characteristics are compared.
- this paper is long and quite wordy in ways that I don’t think are necessary to get the message across. Many general statements that could be deleted or condensed. Figures made more focused on the
- There are a lot of figures that to me seems to re-inforce the same conclusions and I got a bit overwhelmed trying to figure out what new I was learning from each figure. Each figure should have a clear message that is easy to grasp even if the reader is skimming the papers.
- I really like the idea that this asymmetry might affect the way the ice sheet grows or shrinks, but I feel like that message got a little lost along the way.
Specific comments:
Abstract:
- In the first sentence I’d suggest saying that the porous layer of snow stores meltwater and limits **the rate at which Greenland contributes to sea level rise**.
- the abstract states that the the paper demonstrates that fir air content is more efficiently depleted… but doesn’t provide any information to the reader as to the specific findings.
Introduction
Line 24 - check wording, should be contribute, instead of contribution
Lines 32-40 - I understand the authors are trying to be succinct, but much of this paragraph is written in fairly general statements and the citations do not seem appropriate. For example the idea that meltwater changes firn was discovered much before 2015. If the authors wish to cite the more recent papers, then please be specific about what new knowledge those newer papers contributed.
Methods
The model - In general I would like to seem more details. The reader should be able to read this paper and get at least the basic elements of the model the parameters used and any necessary details without having to go read another paper or papers first.
The experiments - Line 100 - “we require” - why? I see why you want no trends, but our real world does have noise at all timescale, so even in a steady climate there will be noise. I would suggest repeating the experiments with constant climate but added noise. Perhaps that is a different paper (if the authors want to contact me for more information on this please do, I’ve done some not-yet- published work on firn thermal structure in noisy versus “pure” steady climate conditions).
Line 104-105 - We verified - how? What do you consider negligible? I am assuming you ran the same model using a different year’s climate and the results on warming/cooling were less than 1% different or something?
Finally some of the limitations of the model and the experimental design are not noted until the very end of the discussion section - and it seems like it would be nice to put those right up in the methods because they are choices that have been made in designing the experiment and the analysis.
Results
Lines 120-135 - this entire paragraph is presented as results, but to me it seems like a confirmation that the model mostly does what we expect. Maybe just rewording this paragraph? statements like “melt occurs where the temperatures are highest” seems so obvious that I wonder what the authors want me to learn from this paragraph.
Line 136-139 - these statements could use specifics. “Idealized warming causes depletion of firn air content” - by how much, how measured. The first specifics offered are for the total spatially integrated air content.
I think overall this section could be condensed down and focused on comparing what was learned to what was expected.
Line 151 - timescales of response to a perturbation are typically defined by an e-folding time, I’m not quite sure I understand the need to use the percentages.
Line 167-169 - again very general statements.
Figures
- most of the figures really could use contours instead of just colors. I can’t tell dark from darker colors and the authors point to various areas where they expect us to see color variations (such as SE Greenland coastal areas). Also - many of the figures contain the full spinup period - this seems like a lot of wasted space, perhaps modify it to zoom in on the results that are the most useful for telling the story.
- is there a reason not to use an e-folding time to express the timescales, Figure 3, for example, could be just one plot instead of 6. Or maybe 2 if the authors want to emphasize how much change happens in the first 2.5% of time.
- Among figures 7 an 8, a few things could make the data more digestible - In fig 7 for example, All 6 could be plotted on the same curve as relative change in FAC. Because of the different scales, it is hard to compare them. Also Fig8 take a while to digest - similarly the “control climate” takes up a lot of ink, and then what my mind really wants to see is a diagram what a firn core might look like in each situation. Also Fig 8 only shows two locations - it would be helpful to find a way to express this information for all of them.
Again - lots to like in this paper, most of my comments are towards focusing the message and making it easier to read and digest.
Erin Pettit
Citation: https://doi.org/10.5194/egusphere-2023-2629-RC3 -
AC3: 'Reply on RC3', Megan Thompson-Munson, 13 Mar 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-2629/egusphere-2023-2629-AC3-supplement.pdf
-
AC3: 'Reply on RC3', Megan Thompson-Munson, 13 Mar 2024
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 282 | 90 | 23 | 395 | 13 | 9 |
- HTML: 282
- PDF: 90
- XML: 23
- Total: 395
- BibTeX: 13
- EndNote: 9
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1