the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Using Icepack to reproduce Ice Mass Balance buoy observations in landfast ice: improvements from the mushy layer thermodynamics
Jean-François Lemieux
L. Bruno Tremblay
Adrienne Tivy
Joey Angnatok
François Roy
Gregory Smith
Frédéric Dupont
Abstract. The column thermodynamics package (Icepack v1.1.0) of the Community Ice Code (CICE) version 6 is used to reproduce observations from two Ice Mass Balance (IMB) buoys co-deployed in the landfast ice close to Nain (Labrador) in February 2017. A new automated surface retrieval algorithm is used to determine the ice thickness and snow depths from the measured vertical temperature profiles. The buoys recorded heavy snow precipitation over relatively thin ice, negative ice freeboards and delayed snow flooding. Icepack simulations are produced to evaluate the performance of the Bitz and Lipscomb (1999) thermodynamics used in the Environment and Climate Change Canada (ECCC) ice-ocean systems and to investigate the improvements associated with the use of mushy layer physics. Results show that the Bitz and Lipscomb (1999) scheme produces a smooth thermodynamics growth that fails to capture the observed variability in bottom sea ice congelation rates. The mushy layer physics produces similar temperature profiles but better captures the variability in congelation rates at the ice bottom interface, with periods of rapid ice growth that coincide with IMB observations. Large differences are also found associated with the snow-ice parameterization: the volume of snow-ice formed during flooding is largely underestimated when using a mass conserving snow-formation scheme, but largely improved when using the mushy layer parameterization in which sea-water is filling the porosity of the snow layer. Both schemes are however unable to reproduce the delayed snow-ice formation, as they rely on the hydrostatic balance and do not allow for negative freeboards. This calls for added brine fraction or ice porosity dependencies in the snow-ice parameterizations.
- Preprint
(4572 KB) - Metadata XML
- BibTeX
- EndNote
Mathieu Plante et al.
Status: final response (author comments only)
-
RC1: 'Comment on tc-2022-266', Anonymous Referee #1, 26 Jan 2023
Review on “Using Icepack to reproduce Ice Mass Balance buoy observations in land-fast ice: improvements from the mushy layer thermodynamics” by Plante et al.,
This manuscript needs a major revision or possible resubmission to the TC.
Undoubtedly, this research subject is important and is of potential interest to TC readers. This manuscript contains the following key elements: a) Icepack (v1.1.0); b) ice mass balance buoy (SAMS IMB), c) land-fast sea ice in Canadian Arctic Archipelago (CAA) and d) mushy layer~ slush layer (mixture of snow and ice).
By the way, please use “SIMBA (snow and ice mass balance apparatus)” in the revised manuscript to present SAMS IMB since this acronym has been used in many papers to name SAMS IMB.
The authors presented the Icepack model; processed the SIMBA data (observations) using a newly developed automatic SIMBA algorithm based on existing methods; simulated ice thickness (calculations) using the Icepack model; Summarized results (observations and calculations); Concluded that the modelled ice thickness is better when applying a mushy layer parameterization; pointed out the simulation errors and give suggestions on further actions. The storyline of this manuscript seems ok, but the presentation suffers various ambiguities and makes it difficult to follow and understand.
Several major comments:
1 What is the relationship between Icepack1.1.0 and Bitz and Lipscomb's (1999) thermodynamics model? To my understanding, CICE is a 2D dynamic-thermodynamic sea ice model developed by the Los Alamos National Laboratory. Icepack 1.1.0 is the one-dimensional module of the CICE model. Bitz and Lipscomb (1999) is an independent one-dimensional thermodynamic sea ice model. Please clarify those models and present clearly how they support each other.
2) Are you trying to develop Icepack or simply to validate Icepack using SIMBA observations? Why is Bitz and Lipscomb's (1999) scheme mentioned separately?
3a) The paper structure is not clear. The current chapters 2 and 3 mixture of many things and need to be reconstructed. One possibility could be
2 Data
Describe the data used in this study
2.1 Weather data
Describe weather conditions
2.2 SIMBA data
Describe SIMBA deployment and data
3 Method
Describe the model/algorithm used in the study
3.1 Icepack model
Surface energy budget
Heat conduction in snow and ice
Bottom heat and mass balance
Snow-ice interaction
3.2 SIMBA algorithm
I would like to see a sub-section dealing with the weather data.
3b) The result chapter needs significant updates too.
I would like to see a sub-section presenting analyses of weather data. This is very important for readers to understand your model performance and the snow-ice interactions. The weather part is missing entirely both in the data and result sections.
Do you have ice core samples to show how the snow ice was distributed vertically? It would be interesting to add some on-site photos.
4) Several figures can be improved.
- a) Figure 1 is not very representative. Please show a much larger domain so readers can better understand the region's geography. What is the distance between those two SIMBAs? What are the air temperatures and precipitation patterns of those two sites?
- b) Figure 7-12 need revisions. Can authors make those figures to be consistent with the SIMBA figures? The figure captions need improvement for better clarity. Some of the results lines need to be smoothed, e.g., 5-day running mean.
5) Surface retrieval algorithm validation: Could authors perform some statistical analyses to give a concrete assessment of your algorithm performance?
6) section 4.2 (In situ ice mass balance conditions) should be moved to the data section.
7) Icepack simulations section looks weak. I see a description of the results, but please carry out some in-depth analyses.
8) The discussion section looks weak too. I would like to see some tables and comparisons with other studies. I am sure there are a lot of land-fast sea ice modelling papers and snow-ice simulations. Please make some concrete discussions.
9) “Code and data availability. All codes (model and analysis) are available on github upon request. The buoy data are available upon request.” I think this statement is not acceptable to the TC. Please make your code and data available with doi link or weblink.
Citation: https://doi.org/10.5194/tc-2022-266-RC1 -
AC1: 'Reply on RC1', Mathieu Plante, 19 May 2023
The comment was uploaded in the form of a supplement: https://tc.copernicus.org/preprints/tc-2022-266/tc-2022-266-AC1-supplement.pdf
-
RC2: 'Comment on tc-2022-266', Anonymous Referee #2, 13 Mar 2023
In this paper model simulations using Icepack are compared with Ice Mass Balance (IMB) buoy data from land-fast ice close to Nain (Nunatsiavut), on the Labrador coast. A new algorithm is presented to determine ice thickness and snow depths from the measured vertical temperature profiles in IMB buoys. Model simulations were run with two different thermodynamic formulations: the Bitz & Lipscomb (1999) and the mushy layer. One of the purposes of this study was to evaluate the performance of the former which is used in Environment and Climate Change Canada (ECCC) ice-ocean systems, and the improvements that may be expected from the latter.
Whereas the mushy thermodynamics generally outperformed the Bits and Lipscomb approach, both were unable to reproduce delayed snow-ice formation as a result of relying on hydrostatic balance and not allowing for negative freeboards.
In the following paragraphs I present my general comments. Minor comments are incorporated directly on the manuscript.
General comments
The paper is very well written, and its contents are extremely clear. It is also well presented. Whilst the subject is not original, since comparisons between these two thermodynamic approaches were already carried out (e.g. Turner & Hunke, 2015; Bailey et al., 2020; DuVivier et al., 2021 – by the way, I suggest incorporating also the main achievements of DuVivier et al in a revised version of the paper), such comparisons were made using regional or global simulations, some with coupled models, introducing a number of factors that make it more difficult to disentangle the “pure” thermodynamic effects resulting from these two schemes. Here, 1D vertically resolved simulations are used, focused on thermodynamic processes alone. Moreover, results are compared with those of IBM buoys that provide a lot of spatial and temporal detail, regarding temperature and thickness of the snow, the snow-ice and the congelation ice.
The authors focus on possible thermodynamic reasons to explain the problems in reproducing delayed snow-ice formation. Whilst I am not criticizing this focus, I wonder if the problem here is mainly thermodynamic or mechanic, related for example with ice floe deformation. In fact, this possibility is mentioned in lines 374-375. On the other hand, it occurs to me (perhaps wrongly…) that when snow-ice if formed from the edges of an ice flow, this will change the porosity, making it more difficult for the water to penetrate further into the ice flow and continue snow-ice production. Such processes cannot be captured in 1D vertical simulations but may possibly be parameterized in 3D experiments. I guess some discussion about these aspects should be included. These problems of negative freeboard, flooding and snow-ice formation combining IBM and simulations with the CICE model were “touched” before by Duarte et al. (2020).
As far as I understood, the model was forced with re-analyses atmospheric data. Whilst I don’t think this is the ideal forcing for such an experiment, since it may introduce bias in the results that may confound a bit the effects, I understand that in situ measurements would be hardly available. In any case, the uncertainties in the forcing should be addressed in the paper, without the need to get into major details. Moreover, it is unclear to me how did the authors managed the ocean forcing. I guess that water temperatures were taken from the measurement arrays of the IBM buoys. However, Icepack expects data on current velocities and heat fluxes in/to the ocean layer in direct contact with the sea ice (by the way, what was the thickness assumed for the ocean slab layer?). I did not find information about these details in the paper, and I think they should be included in a revised version. In fact, it would help to have graphs showing the time series for all forcing functions, even if only in Supplementary info.
Comparisons between model results and observations are presented only for “thickness-related” variables. I think these should include the modeled and observed temperature profiles as well. Once again, such comparisons may be added to Supplementary info. Moreover, comparisons between model results and observations were addressed only visually, and I suggest using some metrics for an objective comparative evaluation of both thermodynamic approaches.
The model was run with a 3-h time step. I wonder if authors checked the results sensitivity to the time step. Duarte et al. (2020) found out that very small time steps may be necessary to avoid bias in the sea ice energy budget fluxes. Interestingly, some of these biases may cancel each other, not affecting model performance when it comes to the prediction of sea ice thickness. However, they may become relevant in coupled models by biasing the feedbacks between the sea ice and the atmosphere, for example. I understand that forcing frequency may limit such verification in this case, but this is something to keep in mind in a revised version.
As a final remark, I suggest transferring section 2 to Supplementary info, since most of its contents reproduce already published science (e.g. Hunke & Lipscomb, 2015).
References
Bailey, D. A., Holland, M. M., DuVivier, A. K., Hunke, E. C., and Turner, A. K.: Impact of a New Sea Ice Thermodynamic Formulation in the CESM2 Sea Ice Component, J Adv Model Earth Sy, 12, ARTN e2020MS002154, 10.1029/2020MS002154, 2020.
Bitz, C. M. and Lipscomb, W. H.: An energy-conserving thermodynamic model of sea ice, J Geophys Res-Oceans, 104, 15669-15677, Doi 10.1029/1999jc900100, 1999.
Duarte, P., Sundfjord, A., Meyer, A., Hudson, S. R., Spreen, G., & Smedsrud, L. H. (2020). Warm
Atlantic water explains observed sea ice melt rates north of Svalbard. Journal of Geophysical Research: Oceans, 125, e2019JC015662. https://doi.org/10.1029/2019JC015662
DuVivier, A. K., Holland, M. M., Landrum, L., Singh, H. A., Bailey, D. A., and Maroon, E. A.: Impacts of Sea Ice Mushy Thermodynamics in the Antarctic on the Coupled Earth System, Geophys Res Lett, 48, ARTN e2021GL094287, 10.1029/2021GL094287, 2021.
Turner, A. K. and Hunke, E. C.: Impacts of a mushy-layer thermodynamic approach in global sea-ice simulations using the CICE sea-ice model, J Geophys Res-Oceans, 120, 1253-1275, 10.1002/2014JC010358, 2015.
-
AC2: 'Reply on RC2', Mathieu Plante, 19 May 2023
The comment was uploaded in the form of a supplement: https://tc.copernicus.org/preprints/tc-2022-266/tc-2022-266-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Mathieu Plante, 19 May 2023
Mathieu Plante et al.
Mathieu Plante et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 333 | 133 | 20 | 486 | 8 | 7 |
- HTML: 333
- PDF: 133
- XML: 20
- Total: 486
- BibTeX: 8
- EndNote: 7
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1