the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 02 Jul 2026
Evaluation of MARv3.14 over the Greenland Ice Sheet
Abstract. Accurately estimating the surface mass balance (SMB) of the Greenland Ice Sheet (GrIS) is essential to quantify its contribution to sea-level rise. The polar regional atmospheric climate model MAR is widely used to simulate GrIS SMB and to force ice sheet dynamics models, highlighting the need for a thorough evaluation. Here, we evaluate the latest MAR version (MARv3.14) over Greenland at 5 km spatial resolution and examine the impact of coarser resolutions (10–30 km) on the simulated SMB and its components. MAR outputs are compared to a range of independent observations, including in situ SMB measurements, automatic weather station (AWS) records of near-surface meteorological variables, satellite-derived melt extent, and albedo products. At 5 km, MAR reproduces the observed SMB with a root-mean-square error (RMSE) of 0.51 m and a correlation of 0.93. For near-surface meteorological variables and surface energy budget fluxes, the model RMSE is smaller than the corresponding observed natural variability (i.e., standard deviation), indicating non-significant model error. Prescribing bare-ice albedo improves the model performance in the ablation zone, while biases remain in the accumulation zone, suggesting that further improvements are required in the snow albedo scheme. In addition, simulated melt timing is consistent with satellite-based melt extent products. Sensitivity experiments reveal that discrepancies between simulations at different spatial resolutions are mostly limited to the ice sheet margins where strong SMB and topographic gradients occur, notably in the southeast of Greenland where precipitation peaks. Differences in integrated SMB generally remain small and mostly non-significant, while individual components, i.e., precipitation and runoff, exhibit larger resolution-dependent variations. We find that a resolution of at least 10 km is required to accurately capture the GrIS climate and SMB. Reducing computation time about 10-fold relative to 5 km simulation, a 10 km grid makes a good compromise for long-term climate projections.
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 paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(24247 KB) - Metadata XML
-
Supplement
(4100 KB) - BibTeX
- EndNote
Status: open (until 13 Aug 2026)
- RC1: 'Comment on egusphere-2026-2341', Kristiina Verro, 03 Aug 2026 reply
-
RC2: 'Comment on egusphere-2026-2341', Anonymous Referee #2, 05 Aug 2026
reply
This works evaluation a new version of MAR against a range of ground and satellite observations across Greenland. There is thorough statistical analysis spatially from both satellite and station measurements. Overall, this manuscript is well-written and provides a clear benchmark of the MAR model. It is particularly useful to compare benefits from higher resolution versus computation time. The evaluation is fairly comprehensive for mean values, and it does acknowledge areas for model improvement, e.g., cloud representation.
I believe a couple relatively minor improvements are needed. First, in lines 113-4 the authors state that they only compare MAR to observed radiative fluxes. This is true through the results, in which case I think the language in the abstract and elsewhere would be more precise to say that comparisons are to surface radiative fluxes rather than mentioning "SEB fluxes", e.g., line 9. However, Table 2 and section 4.1 actually include turbulent heat fluxes. This inclusion means that the turbulent fluxes should be described in the methods, including if they’re bulk or eddy covariance, uncertainty, and which direction is positive, and the language that only radiative fluxes are included should be removed.
Second, the authors note unphysical longwave fluxes on lines 118-120. There are unphysical upward longwave fluxes because the downward looking radiometer absorbed shortwave radiation reflected from the surface (e.g., Kuipers Munneke et al. 2009). While the biases are only obviously wrong when upwelling longwave is greater than ~315.7 Wm-2, the values are biased for all times with incoming shortwave radiation. I doubt this will significantly impact the results, but the upwelling longwave fluxes should be corrected (e.g., Sledd et al. 2026) instead of using biased values.
Some additional small language changes would improve readability, also noted below.
L 94: What are the snowfox records?
L 170-171: How can a version both be an improvement and statistically the same?
L 201: A nearly 10 Wm-2 LWD bias seems notable compared to the magnitudes of upwelling shortwave and longwave fluxes, no?
249-50: I found this sentence difficult to understand the first time I read it. Maybe revise for clarity?
L 262: Given this concern, is there are better threshold/metric to use for melt to compare the model and satellite observations? Is the threshold for melt from MAR only considered at the surface or does it include melt in the full column?
L 265-6: Would this still be included as melt from MAR? Is there a way to define surface melt in MAR for comparison with the observations?
L329: At the end of this sentence, it would have been helpful to mention that the following paragraphs are referencing Table 2.
L 433: You’re using the density of ice, but if you’re considering surface melt wouldn’t that include some fresh snow with lower density for some locations (e.g., Fausto et al. 2018)?
Kuipers Munneke, P., Van Den Broeke, M. R., Reijmer, C. H., Helsen, M. M., Boot, W., Schneebeli, M., and Steffen, K.: The role of radiation penetration in the energy budget of the snowpack at Summit, Greenland, The Cryosphere, 3, 155–165, https://doi.org/10.5194/tc-3-155-2009, 2009.
Sledd, A., Gallagher, M. R., Shupe, M. D., Cox, C. J., Hawley, R., Town, M. S., Guy, H., Marshall, H.-P., Neely III, R. R., Pettersen, C., Walden, V. P., Hebson, C., Martin, A., Olson, E., and Pickell, D.: Summer subsurface temperature variability in the percolation zone of southwest Greenland: high resolution observations of the top meter of firn, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2026-1842, 2026.
Fausto, Robert S., et al. "A snow density dataset for improving surface boundary conditions in Greenland ice sheet firn modeling." Frontiers in Earth Science 6 (2018): 51.
Citation: https://doi.org/10.5194/egusphere-2026-2341-RC2
Data sets
MARv3.14.3-ERA5 outputs at 5 km spatial resolution Xavier Fettweis and Guillaume Timmermans https://doi.org/10.5281/zenodo.19691263
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 92 | 53 | 13 | 158 | 18 | 15 | 12 |
- HTML: 92
- PDF: 53
- XML: 13
- Total: 158
- Supplement: 18
- BibTeX: 15
- EndNote: 12
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The authors are presenting a new MAR version, v3.14, and have gone through extensive and thorough model evaluation in the context of simulated surface mass balance (SMB) over the domain of Greenland Ice Sheet. MAR outputs are compared to a range of observations, including in situ measurements, AWS records, satellite-derived melt extent, and albedo products. For example, the discussion on ablation underestimate (Sect. 4.1) describes, in detail, comparison to SMB observations from station to station. They have also conducted a model resolution-dependency analysis.
The paper is generally well structured, but needs some technical editing. I would like to see a structured Conclusions section. There are a lot of comparisons made in this paper, and the Conclusions section is the perfect place to summarize the main points found in the analysis that are of interest for a MARv3.14 (data) user. Please provide concrete, specific outcomes rather than vague, generalized summary. The Results section needs language polishing / stylistic editing. Namely, the over use of parentheses and over-referencing should be reviewed, as it severely disrupts the reading. This is also true in the Discussion section. At times, the Results section paragraphs are hard to read, as there is no clear structure (e.g 3.4 Albedo representation). I suggest starting with a strict lead sentence, which is the opening sentence of a paragraph that introduces its central point, and then following with supporting sentences with data. As this is a paper using a lot of different datasets to evaluate MAR, the results need to be presented clearly.
I have only some minor scientific comments, mostly asking to expand and elaborate. Overall, the evaluation is done properly and the results make sense. MAR is an important tool in climatological studies, and such a thorough and rigorous improvement and evaluation is appreciated.
Suggestions for improving the content are provided in the supplement pdf. I recommend the paper to be published at the TC, after these minor, mostly technical, improvements.