the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 16 Jan 2025
Seasonal and interannual variability of freshwater sources for Greenland's fjords
Abstract. The magnitude, source, release location, and timing of freshwater fluxes that end up in the numerous Greenland fjords is of special interest for ice-ocean interactions and ecosystems. In this study, we investigate intra- and interannual variability of the various freshwater sources for Greenland’s fjords in seven climatologically distinct regions. For this, we use direct and statistically downscaled output from regional climate models for the mass fluxes, process-based estimates of basal melt and observational data for solid ice discharge. For the period 1940/1958 to 2023, we separately quantify runoff from the Greenland ice sheet, peripheral ice caps and tundra regions, and precipitation directly falling in the fjords. From 2009 onwards, the available data allows us to resolve the full seasonal cycle of freshwater fluxes. The results indicate a diverse range of relative contributions from freshwater sources between seasons and regions. Freshwater input in fjords in the wet southeast and northwest is dominated by solid ice discharge (55 % and 67 %, respectively) with a small contribution of tundra runoff, whereas in the relatively drier north, northeast, and southwest the contribution of tundra runoff is more important (20 %, 25 % and 30 %, respectively). Precipitation in fjords and tundra runoff can represent a large fraction of the monthly total, i.e. up to 11 % and 35 %, respectively, for winter and spring. However, the relative contribution of tundra runoff has been decreasing in time, the result of rapid increases in ice sheet and ice cap runoff over the past decades following atmospheric and oceanic warming. We show that the regional glacier-integrated melt-over-accumulation ratio (MOA) is a good predictor for the relative contributions of solid ice discharge, tundra runoff, and ice sheet runoff. These findings have implications for the use of freshwater fluxes forcing in regional ocean models and fjord studies, and enhance our understanding of their impact on ocean and fjord circulation and biogeochemistry.
Competing interests: At least one of the (co-)authors is a member of the editorial board of The Cryosphere.
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 preprint. The responsibility to include appropriate place names lies with the authors.- Preprint
(3785 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-3735', Ken Mankoff, 19 Jan 2025
* OverviewThis manuscript quantifies total freshwater fluxes into fjords at regional scale for Greenland. It is similar to but improves on some existing products. Notably this is the first product I'm aware of that addresses freshwater input over fjords, although it does so in a fairly simplistic manner. This also adds temporal resolution to the Karlsson (2023) http://doi.org/10.34194/geusb.v53.8338 product which is monthly but no annual time series.
Overall I'm supportive of this paper, but think it needs a few major updates that I detail below.
* Issues
** Custom regions
L129-131 states "We made one adjustment by moving the boundary between SE and CE northward to better follow hydrological catchments, as well as to make the regions more comparable in size."
Must you do this? Please don't do this. I assume you are not trying to make your product less useful and usable and add confusion and extra effort for others, but without good justification for adding yet another new and different basin and region product, I cannot let this comment go.
See https://x.com/glasheologist/status/1557208390853505024 There are entire community projects (e.g., https://iacs-cryo.github.io/Delineation-WG/deliverable1 ) and many papers showing that comparing products A and B is hard or impossible or incorrect because they used different boundaries.
In addition to using standard regions, I would strongly encourage you to provide data (even if not results in the paper) at fjord scale. Or in map (spatial) form at GIS scale. Your work is only possible because of all of the data products you were able to ingest and process. You should strive to make your products even more usable than what you ingested - raise the bar if possible. While some of those products were specifically data products (and published in ESSD), many others were science-focused papers such as this manuscript but that also shared usable data. Data shared at fjord scale can be easily combined up to region or GIS scale. Data shared spatially at GIS scale (at a reasonably high resolution, e.g < 25 km grid cell) can be processed down to region or fjord resolution. The only way to have data less useful than providing it at regional scale is to provide a single value at GIS scale.
I know Mankoff products are available at fjord scale, Karlsson is GIS wide spatial format (can be accumulated at fjord scale), and as can your RCM inputs. I'm not sure if any of your inputs are provided at only regional scale.
In summary, use standard regions, and provide data everywhere in spatial format, or at glacier scale in vector format.
** MOA
Sect. 2.3 Annoying to see a new term and have to jump around to learn about it. Add a brief intro to MOA here? Anyway, 4.2 just refers straight back to 2.3. I still don't know what MOA is beyond a mathematical definition. You're introducing a new term and concept here (I think? I've never seen it and you have no citation for it). Add some text? What's the point of MOA? What do we learn from it? What are its weaknesses and limitations? What's a reasonable range for it? Min, max, average, median? How does it change spatially or temporally and why? Some of these are addressed in the text, but very few.
I've now spent some time thinking about MOA and trying to understand it, looking at Fig. 6, etc.
MOA = (melt + rain) / (snowfall − sublimation)
=> low MOA = low melt or high snowfall
=> high MOA = high melt or low snowfallThis seems to mostly be an atmospheric phenomenon? But perhaps controlled by elevation too? What about providing a spatial map of average MOA. Does something interesting pop out? I assume nothing interesting in the interior where MOA goes to 0 because of no melt. I'd expect regions with high winter snowfall to also have high summer melt (e.g., Southern Greenland) and regions with low accumulation to also have low melt (N. Greenland). The ablation area is where it gets interesting. Is it just a proxy for width of ablation zone? Something else?
I also assume it's only useful as an annual average, not monthly. How it changes in time is also unlikely to be interesting, as we know there's an increase in melt that is larger than the increases in snowfall. But maybe I'm missing something.
I think you need to introduce MOA as a stand-alone product before you start correlating it with external things like Discharge (Fig. 6a) at the beginning of Sect. 4.2.
What is the point of 6a? I think it's a complicated way of saying that when there is high melt, the proportion of discharge goes down.
* Minor comments
+ L24: How does something enter a fjord above a grounding line?
+ Paragraph starting L42: Seems like the text here should make reference to Mankoff (2020) http://doi.org/10.5194/essd-12-2811-2020 which uses your RCM (and MAR) to distribute both ice sheet and tundra runoff at stream resolution. L51 in particular, tundra runoff is not excluded by Mankoff (2020) http://doi.org/10.5194/essd-12-2811-2020. That paper does not "combine different datasets" (L50), but is trivial to combine with other Mankoff products to get solid discharge + RCM runoff terms. I also note that the Mankoff discharge product provides an estimate of discharge depth for every stream outlet, addressing the surface vs subsurface discharge issues raised here.
+ L143: There is no explicit uncertainty section in Sect. 2, but it sounds like there should be? Is it 1 or 2 sigma?+ L164: Direct inputs in January? Seems unlikely w/ ice cover? I realize you address this at the end of the document, but I think it should be addressed more explicitly throughout.
+ L164: Basal amount is mostly steady state. What is the goal or reporting this small amount on this month? What is the significance of this sentence?
+ L240: Mankoff 2020 (freshwater) shows RCMs and obs disagree by >> 100%, not 30 - 50 %.
+ Table A1 and elsewhere: Do you know things to 0.1 %?
+ What version of Mankoff (2019) data did you use?
+ Maybe change Mankoff (2019) to Mankoff (2020) http://doi.org/10.5194/essd-12-1367-2020
+ Your Zenodo dataset contains folders and files like ".DS_Store" and "__MACOSX/data/temp/._MOA_plot_input_mean_per_region_1990-2023.csv". Consider adding the DOI/URL of "all versions" that Zenodo provides and in the text mentioning which version you used. This way if you update the data in the future, readers can find the version in the paper or the latest version.
+ "Flux": This is a technical term with specific units of mass or length^3 (volume) per unit time per unit area. You are never using per unit area, and should probably never use the word flux (except perhaps when referring to ice discharge flux gates). I believe the correct term is either "mass flow rate" if Gt/yr or "volume flow rate" if km^3/yr.
+ What is sublimation? Is it the net term (deposition + condensation - evaporation - true sublimation), or is it true sublimation?
+ Sect 4.6 It's great that this product addresses freshwater input to fjord surfaces. But I encourage you to be more up-front about the limitations of this new approach, the primary limitation being sea ice. It's the last section before conclusions, but is more important than that. Going back to highly precise uncertainty (fractions of a percent!), what exactly is your uncertainty measuring? Some arbitrary mathematical function of the RCM, or is it telling us something useful about what we do and do not know about freshwater fluxes at regional and temporal scale? If the former, that's easy but not terribly useful. If the latter, that may be more useful for downstream users, but your uncertainty in winter months must then increase because of the neglected sea ice processes. That is, perhaps you can address sea ice in this work in some manner, without adding a full model of sea ice growth, winter RCM accumulation onto the sea ice, and then sea ice advect & decay which allows the accumulated mass to then enter the fjord. That's out of scope for this paper, but I hope some treatment may not be out of scope.
+ Table 1: MAR "source" is bold but nothing else on that line is bold.
+ Fig 2: y-axis units might be "Mass flow rate" or "volume flow rate" not "loss"?
+ Fig 3: y-axis units are not Flux but Mass flow rate.
+ Fig 6a: Add units to y-axes. Replace symbols with letters (i.e "NO, "NW" in the plot, no legend needed).
+ Table A3, A4: If MOA is interesting and you keep it, should it be a column in these tables (region average MOA)?
+ Competing interests: This is weirdly phrased. Why not state who has which specific competing interests?
+ Ack: Cite LanguageTool (URL if no scientific paper)
+ Cite software used.
+ Cite all data products DOIs, not just scientific papers. Mention versions of data products.
+ When you say "From 1990 to 2023" I don't know if this includes 2023 or not. I recommend "A through B" if you went to the end of B.
Citation: https://doi.org/10.5194/egusphere-2024-3735-RC1 -
AC1: 'Reply on RC1', Anneke Vries, 05 Apr 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2024-3735/egusphere-2024-3735-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Anneke Vries, 05 Apr 2025
-
RC2: 'Comment on egusphere-2024-3735', Anonymous Referee #2, 26 Feb 2025
This study investigates the variability of freshwater sources in Greenland’s fjords across different regions and seasons, using data from climate models and observational records. It highlights how contributions from solid ice discharge, tundra runoff, and precipitation have evolved over time, with a significant increase in ice sheet runoff due to warming trends.
The manuscript is well-written, and the results improve upon some existing products. My suggestions, outlined below, mostly pertain to clarifications of the results.
Specific comments:
L52-54: This sentence is quite confusing. I suggest instead writing what the tundra contribution is for the whole model time period for all these studies (Bamber et al, 2018, Mankoff et al, 2020, Igneczi and Bamber, 2024). Something along the lines of:
“Bamber et al, 2018, estimated that tundra runoff added on average ?? % from 1958-2016, while Mankoff et al, 2020 found a larger number of.. Igneczi and Bamber (2024) estimated an even higher contribution of [..].”
L68: Consider whether it is really necessary to abbreviate freshwater as “FW” in the text. The less abbreviations, the easier it will be to read.
L80: Delete “MAR henceforth”, the abbreviation has already been given on the previous line “(MAR)”
L93-98: If the RACMO simulations exist back to 1940 at a 5.5 km resolution, could you not use the whole time series by conducting the statistical downscaling of Noel et al (2016)? Or what is the reason for only using 1958-present? Also, why did you choose to not statistically downscale the tundra?
L98: Is there no dataset from MAR of the tundra runoff which you can compare to? This is the only variable where only one dataset is used
L102-103: Why don’t you use the RACMO product from 1940 here?
Table 1: Isn’t the tundra runoff also interpolated onto a 1 km resolution? There is no arrow in the table like for the runoff.
Section 2.2: consider removing the adjustment you added to Slater et al. (2020), as it will be easier to compare to other studies if the same borders are used.
L150-51: I’d remove this line: “the Total average FW flux since 2010 is 1239 (± 180) Gt yr −1 , calculated to compare to other studies in Sect. 4.1”. The number can just be provided in Sec 4.1:
L164-65: change to “Basal melt accounts for a maximum of 3± 0.6 % (March)”, to clarify that march is the month with the highest contribution from basal melt
Section 3.2: Information on the trend of the freshwater components are mostly given for NO. Can you write this for the other areas too?
L173-97: Since the discussion is mostly about how much each component contributes to the total runoff for each area, I would suggest changing the numbers in this section to percentages
L186: change “in line with earlier studies” to “in line with an earlier study” (or alternatively, have more than one reference)
L188: delete “now”
L196: what does “(O (10 km))” mean?
L232-235: this is a results, I suggest moving to section 3.2
L235-36: “The average annual rate is slightly smaller than the total Arctic FW flux found by Bamber et al. (2018) (1300 Gt yr −1 since 2010), which included non-Greenland ice caps.” - I would delete this sentence, since it is not calculated for the same region.
Section 4.1: The results could also be compared to Mankoff et al (2020)
L236-37: “GrIS runoff between 2010-2016 is higher in Bamber et al. (2018) and Igneczi and Bamber (2024) than estimated in this study.” → this sentence is too vague, please provide numbers. And why is it only compared for 2010-16, aren’t both studies for longer periods?
L239: which period is the comparison with bamber et al (2018) from?
L240-41: “.. similar to values found in this study using RACMO output.” - this is too vague, please provide the numbers
Section 4.2: I am not sure what the added value of the MOA is here. Since you are discussing the fraction of the total runoff, it seems obvious that regions with less melt will have a higher contribution from the solid ice discharge and regions with high melt with have a smaller contribution from the solid ice discharge. And regions with high melt will have a higher MOA, while regions will less melt will have a smaller one. So MOA does not seem to provide any additional information on what is actually happening beyond the obvious – or do I misunderstand what the meaning of the parameter is? Please clarify
L251-52 / Figure 6(c-d): I am not sure what the point is of Figure 6c, d. Since runoff and ice discharge are the main contributors, won’t these always have a linear relationship? if there is a high contribution from ice discharge (e.g. 70%), of course the contribution from runoff has to be low. And when there is a high contribution from runoff, the contribution of ice discharge cannot be high. I suggest deleting this sentence/these figures, unless there is something important I am missing?
L262: add the relative distribution in Bamber et al (2018)
L264: what is a “relatively high contribution”? Can you provide percentages?
L337-341: please cite the datasets in addition to the papers
Figure 3: why does (a) and (b) show 1990-2004 and 2005-2023 instead of the whole period? This is not discussed in the text. I suggest either only showing the values for the whole time period (1990-2023), or alternatively add some text about the difference between the two time periods.
Figure 6: I think the “R” on the figures should be “r”, like in the caption?
Citation: https://doi.org/10.5194/egusphere-2024-3735-RC2 -
AC2: 'Reply on RC2', Anneke Vries, 05 Apr 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2024-3735/egusphere-2024-3735-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Anneke Vries, 05 Apr 2025
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 186 | 59 | 10 | 255 | 6 | 8 |
- HTML: 186
- PDF: 59
- XML: 10
- Total: 255
- BibTeX: 6
- EndNote: 8
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 79 | 29 |
| Netherlands | 2 | 39 | 14 |
| Germany | 3 | 36 | 13 |
| China | 4 | 10 | 3 |
| Greenland | 5 | 9 | 3 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 79