the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 24 Mar 2023
The Southern Ocean Freshwater release model experiments Initiative (SOFIA): Scientific objectives and experimental design
Neil Swart
Torge Martin
Rebecca Beadling
Jia-Jia Chen
Mathew H. England
Riccardo Farneti
Stephen M. Griffies
Tore Hatterman
F. Alexander Haumann
Qian Li
John Marshall
Morven Muilwijk
Andrew G. Pauling
Ariaan Purich
Inga J. Smith
Max Thomas
Abstract. As the climate warms, the grounded ice sheet and floating ice shelves surrounding Antarctica are losing mass at an increasing rate and injecting the resulting meltwater into the Southern Ocean. This freshwater input could feed back onto climate change, particularly since the Southern Ocean is a key contributor to global heat and carbon uptake. Nonetheless, almost all existing coupled climate models have fixed ice sheets, and lack the physics required to represent the dominant sources of Antarctic melt. These missing ice dynamics represent a key uncertainty that is unaccounted for in current global climate change projections. Previous studies have inserted additional Antarctic meltwater into models, demonstrating that it can alter Southern Ocean stratification, circulation, and sea ice, as well as influence remote atmospheric circulation, tropical precipitation, and global temperature. However, these previous studies have used widely varying rates of freshwater forcing, been conducted using different climate models and configurations, and have reached differing conclusions on the magnitude of meltwater-climate feedbacks. The Southern Ocean Freshwater release model experiments InitiAtive (SOFIA) brings together a team of scientists to quantify the climate system response to Antarctic meltwater input. In this paper, we summarize the state of knowledge on meltwater discharge from the Antarctic ice sheet and ice shelves to the Southern Ocean and explain the scientific objectives of our initiative. We propose a series of coupled and ocean/sea-ice model experiments, including idealized meltwater experiments, historical experiments with observationally consistent meltwater input, and future scenarios driven by meltwater inputs derived from stand-alone ice sheet models. Through coordinating a multi-model ensemble of simulations, with data housed in a common archive, SOFIA will be able to produce a consistent estimate of the climate system response to Antarctic meltwater, as well as the uncertainty of this response.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(1304 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1304 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Neil Swart et al.
Interactive discussion
Status: closed
-
CC1: 'Comment on egusphere-2023-198', Katherine Turner, 24 Mar 2023
Review of The Southern Ocean Freshwater release model experiments Initiative (SOFIA): Scientific objectives and experimental design by Swart et al.
The Southern Ocean Freshwater release model experiments InitiAtive (SOFIA) provides a set of climate model experimental protocols for quantifying the response to Antarctic meltwater. The protocols are general, which allows them to be applied to both high- and intermediate-complexity climate models.
The manuscript reviews the theory behind ice mass budgets, current representation of meltwater forcing in climate models, and historical trends and projections for ice sheet mass. Additionally, the manuscript covers how uncertainties in model architecture, internal variability, and the location of meltwater addition may impact the climate response. A thorough description of the experimental setup is located in the appendix.
The manuscript presents an interesting set of experiments to explore the role of freshwater forcing in the Southern Ocean, which is both highly uncertain and poorly represented in current climate model setups. The experimental setup includes three broad experiment types within tiers 1 and 2:
- (tier 1) Pre-industrial climate conditions forced with constant freshwater forcing,
- (tier 2) Historical climate conditions forced with linear freshwater forcing near the end of the historical runs, and
- (tier 2) Future projections with transient freshwater forcing calculated from ice models forced with RCP climate forcing (although I am unsure whether the ice models are forced with surface temperature changes or radiative forcing from GHGs + land use + aerosols)
There is a third tier of experiments that tests the sensitivity to the horizontal distribution of meltwater addition and to the heat fluxes involved in the phase change.
As I understand it, this paper aims to 1. provide experimental context for future papers which are already in the pipeline, and 2. advertise the SOFIA experiments to other climate modeling groups and encourage them to provide additional runs. Regarding aim 1, I think the paper is successful. Appendix A provides useful suggestions for creating ensembles, detailed descriptions for how forcings were calculated (particularly for the future projections), and a discussion of how ocean-ice models can contribute if SSS restoring is included. The standardization of experiments through SOFIA will help increase the transparency of future model-based studies and bring new, much-needed knowledge on Southern Ocean dynamics.
Regarding aim 2, I think the manuscript would benefit from some restructuring before submission. Signposted argumentation across the introduction, review, and ending discussion would be helpful in guiding the reader along the manuscript. I had difficulties understanding exactly how specific experimental setups would be used to reduce the various uncertainties and goals described in Sections 2 and 3. As a non-expert, I think the manuscript would benefit from a (sub)section in which the experiment aims are explicitly stated, rather than requiring the reader to infer some of these aims. I also think that connections to other model intercomparison projects (e.g., FAFMIP) should be emphasized, as these connections could attract modeling groups that already participate in these other projects.
Ultimately, I think the manuscript is a useful and timely work that merits submission. My suggestions below are to help readers from other fields within climate modeling appreciate the need to understand meltwater impacts and, in the long term, work towards creating coupled climate models that include interactive land and sea ice.
Signposting suggestions:
As it stands, each main section jumps directly into a subsection. A few sentences at the beginning of each section could provide extra motivation and emphasize the connections between the review material and the proposed experiments.
While the experimental descriptions are in the appendix, it would be useful to the reader to have a brief outline of the experiments and overarching themes of the project in the introduction. Perhaps in the introduction, the aims/motivating questions can even be listed with bullet points. That way, the reader can reference either the introduction or the appendix as they read the rest of the manuscript, depending on their level of involvement.
References to the SOFIA experiments are sparse and general (see line 128 for an example), which hinders readability. I recommend that the references to the experiments be more specific if the authors decide against adding an “experimental setup” section to the main text.
It would be nice to see more discussion, both within the main text and in the descriptions of the experiments in Appendix A, about why these specific experiments were chosen. For instance, regarding the preindustrial antwater experiment, how will the results be used in coordination with the other experiments (if at all), considering the differences in magnitudes? Will the authors test the linearity of the response using the idealized historical experiments? What are the advantages of using SSP-style forcings for future projections over the 1% CO2 experiments commonly used in other model intercomparison projects such as FAFMIP?
I would be interested to know if there are additional benefits to these runs outside quantifying uncertainty from the Southern Ocean. For instance, is there the potential to use some of these results to improve global or regional projections of sea level rise? The connection to biogeochemistry is mentioned briefly, but I would also be interested to read more about any (even speculative) impacts.
Minor points:
- Some sentences are difficult to read and have slight grammatical mistakes – for instance lines 19-21 has a misplaced modifier, and the clauses in 156-157 do not share the same subject. Overall, the manuscript has quite complicated sentence structures where the meaning can be unclear (e.g., line 163-164). I would recommend the longer sentences throughout the manuscript be broken up to improve clarity.
- Line 45: is there a contrasting study to Beadling et al., 2022 missing?
- Lines 97-98: What are the estimates from the Shepard et al. and Green et al. papers?
- Line 195: Behrens et al., 2015 (doi: 10.1002/2015JC011286) investigates this with the CMIP5 models – perhaps good to include
- Lines 288-289: A reference to FAFMIP in the main paper would also be a useful comparison as it is an established model intercomparison that includes freshwater forcing.
Citation: https://doi.org/10.5194/egusphere-2023-198-CC1 -
RC1: 'Comment on egusphere-2023-198', Anonymous Referee #1, 05 Apr 2023
The manuscript describes a new model intercomparison project devoted to the analysis of the impact of freshwater release in the Southern Ocean. Many studies have performed experiments imposing freshwater perturbations in this region but using different experimental designs, making an interpretation of the differences between the results of those experiments a complex task. This new initiative is thus timely. The protocol is clearly explained and several modelling groups have already agreed to participate in the exercise. It is thus very likely that the initiative will be a success.
1/ One of the stated main objectives is to have a protocol that is as simple as possible in order to ensure a wide participation. For instance, freshwater will be only added at surface, using simplified geographical distributions. The experimental design is consistent with this goal and this is well-adapted for an intercomparison exercise and to study why models have different responses to the same perturbations. However, it does not mean that the experiments will reproduce well the recent observations or the dynamics of the real system. It is perfectly fine to estimate the magnitude of the freshwater perturbation based on estimates of the mass balance of the Antarctic Ice sheet to have a reasonable order of magnitude but this can give the false impression that the intercomparison will give us a kind of measure of the uncertainty on the recent past and future impact of freshwater release in the Southern Ocean. The experiments will measure some elements of the uncertainty, related to model uncertainty, but not many others such as the impact of release at different depths or location, which could have a large impact too. This should be more clearly stated when the goals of SOFIA are described, starting for instance in the last sentence of the abstract which mentions ‘a consistent estimate of the climate system response to Antarctic meltwater, as well as the uncertainty of this response’ while it is only the uncertainty in the framework of the protocol related to the use of different models and different idealized scenario that will be estimated.
2/ In the same line, the manuscript mainly mentions CMIP6 type of models (see Table 2). This focus is consistent with the goals of the project but the introduction should describe more extensively the limitations of those models, in particular through a comparison with the results of higher resolution models. For instance, it is mentioned that there is a consensus on the expansion of the sea ice cover in response to the freshwater input but some studies have shown on the contrary a reduction of the ice thickness in response to ice shelf melting (e.g., Merino et al. 2018, Mackie et al. 2020, already cited in the manuscript). Changes in the dynamics of the slope current that cannot be resolved at the typical resolution of CMIP6 models also have an impact on the response to the perturbation and specifically on the exchanges with the continental shelves that can have a large impact on deep water formation, a topic that is part of the objectives of this study (Moorman et al. 2020; Beadling et al. 2022, already cited). This does not reduce the interest of the proposed intercomparison but should be discussed clearly to avoid an over-interpretation of the results of the intercomparison (see for instance lines 220-221).
3/ More and more ocean models include a representation of ice shelf cavities and we can expect that in the future several CMIP-type models will have such a representation. This is not discussed in the experimental design. Should these models be excluded from SOFIA and maybe be included in a future phase of the initiative as they will compute interactively some of the freshwater input that others will have to add at surface (for instance in the future scenario runs) or are they suggested adaptations to include them already at his stage?
4/ Minor points.
Line 96, I would replace ‘total’ by ‘additional’ as the total freshwater input includes also the climatological mean fluxes, not just the recent increase as more explicitly stated line 99-100
I would add in Table A1 that all the experiments are made with the Antarctic adjacent distribution, except 60Swater as currently the distribution is only clear for this experiment.
Citation: https://doi.org/10.5194/egusphere-2023-198-RC1 -
RC2: 'Comment on egusphere-2023-198', Nicholas Golledge, 19 Apr 2023
Swart et al: SOFIA
General comments
This paper outlines a new climate modelling initiative called SOFIA, that aims to assess the impact of including ice-sheet meltwater discharge on future projections of Southern Ocean and global climate evolution. This builds on the many studies that have done such experiments previously, but aims to formalise and standarise the procedures employed so that inter-model differences can be identified and quantified.The paper is succinct and sets out the intentions and procedures clearly. I don't think it does a particularly objective job of summarising the previous literature, however, with overt reliance on some of the same few citations throughout (Bronselaer et al 2018 & Fyke et al 2018 for example are both cited repeatedly, yet Fyke was simply a review paper and not a modelling study). I realise that the authors are more familiar with the climate model literature than the ice sheet modelling literature, but I think some further work could be useful in this area, to avoid the current bias.
The figures are fine and are useful, and with the modification suggested below I think Table 1 will be a very useful summary.
Specific comments
l2 - is 'injecting' the most appropriate verb to use here? The ice is melting slowly over a wide area, so maybe 'releasing' would be better?l2-3 - "This freshwater input could feed back onto
climate change, particularly since the Southern Ocean is a key contributor to global heat and carbon uptake" - I find this sentence a bit vague. Why not lean on previous research and be more definite: "Previous studies have shown that this freshwater input could lead to xxx and yyy, exacerbating climatic changes already underway" or something like that.l5-6 - "unaccounted for in current global
climate change projections" - I would say 'typically unaccounted for', because some GCMs with interactive ice sheets do exist, and not all 'climate change projections' are just about temperature, some relate to SLR for example, for which several scenario-based simulations incorporating meltwater feedback now exist. I would suggest taking a look at AR6 Ch9 for some info on that.l11 - is it the 'team of scientists' that is important here, or the fact that different models are being used with standardised methods?
l19 - citation to support the first statement perhaps?
l50-51 - "Since virtually
all existing Southern Ocean hosing experiments have involved only one model" - perhaps clarify that you mean 'have EACH USED only one model', because at the moment it reads as though all the experiments have used the same model, which is not the case, right?Table 1 - this is a very useful summary of the studies that have been undertaken. However, at the moment it only considers the climate model side of things. Since one of the key arguments being made to justify the SOFIA initiative is that few models incorporate interactive ice sheets, how about having a column that shows the full suite of model components being employed? E.g. 'OA' could be a standard ocean-atmosphere setup, whereas 'IOA' could be ice-sheet/ocean/atmosphere. I feel like this could be very useful for identifying which studies have 'closed the loop' as it were. For example, the Bronselaer paper is repeatedly cited throughout this paper but they didn't include a ice sheet model, whereas Golledge et al & Sadai et al did. The groupings in the table could then be adjusted to reflect this.
l113-115 - I think it might be worth mentioning that these kind of coupled experiments using interactive ice sheets are more common for Greenland - see for example Vizcaino et al 2015, Muntjewerf et al 2021 and various others. Again, much of this literature was assessed in AR6 Ch9 so could be easily discovered and included, perhaps more in the intro part of the paper than in this particular section though.
l133-4 - I think this statement "using these to force stand-alone ice sheet models" with the citation to the four papers cited is misleading, because Edwards et al presented only emulated results, not direct model outputs, and Golledge et al used a two-way (albeit offline) coupling specifically to capture the ice-ocean feedback. Perhaps this sentence can be rewritten to clarify what each of these studies did.
l154-6 - "Furthermore, feedbacks have been hypothesized, where increased meltwater input further enhances on-shore ocean heat transport through different processes in different regions (Hellmer et al., 2017; Hattermann, 2018; Bronselaer et al., 2018)." Yes, true, but why not also mention that some models have actually TESTED this feedback and quantified the response?
l176 - see note earlier regarding 'only one model'
Secion A4.3 - might be worth mentioning somewhere that Bakker et al 2017 used hosing experiments to explain some of the centennial-scale variability in climate during the Holocene - the timescale is different to the Historical experiments planned here, but the rationale and mechanics are the same.
Section A4.4 & 4.5, 4.6 - using ISMIP6 for basal melting is ok, but why then use an approximation for the calving flux? This assumes that the calving to basal melt ratio stays constant in time, which might not be true. It would be more physically robust to use calving fluxes directly from the ISMIP model outputs.
N R Golledge, 20th April 2023
Citation: https://doi.org/10.5194/egusphere-2023-198-RC2 -
CC2: 'Comment on egusphere-2023-198', Karen J. Heywood, 28 Apr 2023
This is a fantastic initiative and a most interesting paper. I wondered if you were aware of the paper by Richardson et al. In 2005, which was, I believe, the first to undertake an experiment using a coupled climate model (HadCM3) to investigate the impact of a meltwater addition around Antarctica, demonstrating an increase in Antarctic sea ice and subsurface ocean warming.
Short-term climate response to a freshwater pulse in the Southern Ocean
Glen Richardson, Martin R. Wadley, Karen J. Heywood, David P. Stevens, Helene T. BanksGeophysical Research LettersFirst published: 01 February 2005https://doi.org/10.1029/2004GL021586Of course, the models and experiments are much better nowadays, and longer model results are possible. I look forward to further results from Sofia!Citation: https://doi.org/10.5194/egusphere-2023-198-CC2 -
CC3: 'Comment on egusphere-2023-198', Paul Holland, 10 May 2023
Hi Folks
Many congratulations on this initiative, which is very timely and will be a very important addition to this field. The range of experiments out there at present does complicate matters, and SOFIA should clear that up very nicely. I have a few comments on the experimental design. Apologies for the unsolicited review!
Clearly the over-riding goal of SOFIA is to provide a consistent set of experiments that are practically achievable to the modelling centres, and that will necessitate some compromise on the physics. I am fully sympathetic to that. Therefore I am really expressing the below in order to raise a few possible caveats to the results that emerge, and to prompt future discussion, rather than to suggest any change in experimental design.
- The climate models will still have their ‘runoff’ (as shown in Figure 1b) from Antarctica during these experiments, but that will provide an uncontrolled additional Antarctic freshwater source to the ocean that may complicate the intercomparison of models. For example, there is a large increase in precipitation onto Antarctica during SSP5, which varies between models, and that will flow straight into the ocean. This precipitation increase can be the same size as changes in ice discharge during the 21st century so I think it is important (e.g. Seroussi et al 2020). The best solution could be to turn off runoff during these experiments and just implicitly accumulate the mass on Antarctica. Next best could be requesting further control experiments where needed (e.g. SSP585 without added FW) so that the effect of adding FW can be isolated from the model runoff changes. Simplest would be just analysing the runoff alongside the other results.
- It is a very uncertain choice taking the future parameterised melt from ISMIP6 models and multiplying by ~2 to account for calving. Calving could not change at all (larger, thinner ice shelves are implicit in the ISMIP6 melt rates) or could increase massively (ice shelves collapse as the grounding line retreats). Even if calving did increase, melting should be reduced accordingly, so it is not clear whether x2 is needed. One solution could be to take the change in ice sheet discharge across the grounding line from the ISMIP6 models, because that circumvents this point.
- Inputting freshwater into the surface, uniformly around Antarctica, with no latent heat extraction, produces a clear, tractable experiment but it is an unrealistic choice. In reality almost all of the additional mass flux has appeared in the Amundsen Sea, at depth, with a latent heat extraction. (Admittedly some ice shelf collapses have occurred on the peninsula - though those bergs were quickly exported - and the Ross and FRIS cavities could melt by 2100 under SSP5.) As with all of my points, I am not concerned about the model experimental design, just its interpretation. For me the stated design creates three issues: 1) Hosing the ocean with freshwater at the surface and no strong cooling produces a strong stabilising effect, which is not representative of real glacial melting. I assume this will cause excessive shutdown of dense water formation in ‘cold’ shelf regions, and excessive warm-water feedbacks in ‘warm’ shelf regions; 2) The uniform near-coastal distribution hoses the ocean directly in the ‘cold’ regions that form dense AABW, unlike in reality, thus over-stating the potential effects on AABW; 3) the 60S distribution hoses the ocean beneath the sea ice, over-stating the potential effects on sea ice and Southern Ocean SSTs. Issue 1 is compounded in issues 2 and 3, of course. So I think the current experiments are all extreme in over-stating the role of FW. For me the best solution would be to add the opposite extreme experiment with the all FW added in the Amundsen Sea only, distributed over depth, and with corresponding latent heat removed.
As I said above, all of these are just thoughts for discussion. SOFIA is fine as-is, and the authors should feel free to ignore these comments or take whatever action they feel sensible. I hope these thoughts help more than they hinder. Best of luck with SOFIA.
Cheers, Paul
Citation: https://doi.org/10.5194/egusphere-2023-198-CC3 - AC1: 'Response to reviewers and community comments', Neil Swart, 16 Jun 2023
Interactive discussion
Status: closed
-
CC1: 'Comment on egusphere-2023-198', Katherine Turner, 24 Mar 2023
Review of The Southern Ocean Freshwater release model experiments Initiative (SOFIA): Scientific objectives and experimental design by Swart et al.
The Southern Ocean Freshwater release model experiments InitiAtive (SOFIA) provides a set of climate model experimental protocols for quantifying the response to Antarctic meltwater. The protocols are general, which allows them to be applied to both high- and intermediate-complexity climate models.
The manuscript reviews the theory behind ice mass budgets, current representation of meltwater forcing in climate models, and historical trends and projections for ice sheet mass. Additionally, the manuscript covers how uncertainties in model architecture, internal variability, and the location of meltwater addition may impact the climate response. A thorough description of the experimental setup is located in the appendix.
The manuscript presents an interesting set of experiments to explore the role of freshwater forcing in the Southern Ocean, which is both highly uncertain and poorly represented in current climate model setups. The experimental setup includes three broad experiment types within tiers 1 and 2:
- (tier 1) Pre-industrial climate conditions forced with constant freshwater forcing,
- (tier 2) Historical climate conditions forced with linear freshwater forcing near the end of the historical runs, and
- (tier 2) Future projections with transient freshwater forcing calculated from ice models forced with RCP climate forcing (although I am unsure whether the ice models are forced with surface temperature changes or radiative forcing from GHGs + land use + aerosols)
There is a third tier of experiments that tests the sensitivity to the horizontal distribution of meltwater addition and to the heat fluxes involved in the phase change.
As I understand it, this paper aims to 1. provide experimental context for future papers which are already in the pipeline, and 2. advertise the SOFIA experiments to other climate modeling groups and encourage them to provide additional runs. Regarding aim 1, I think the paper is successful. Appendix A provides useful suggestions for creating ensembles, detailed descriptions for how forcings were calculated (particularly for the future projections), and a discussion of how ocean-ice models can contribute if SSS restoring is included. The standardization of experiments through SOFIA will help increase the transparency of future model-based studies and bring new, much-needed knowledge on Southern Ocean dynamics.
Regarding aim 2, I think the manuscript would benefit from some restructuring before submission. Signposted argumentation across the introduction, review, and ending discussion would be helpful in guiding the reader along the manuscript. I had difficulties understanding exactly how specific experimental setups would be used to reduce the various uncertainties and goals described in Sections 2 and 3. As a non-expert, I think the manuscript would benefit from a (sub)section in which the experiment aims are explicitly stated, rather than requiring the reader to infer some of these aims. I also think that connections to other model intercomparison projects (e.g., FAFMIP) should be emphasized, as these connections could attract modeling groups that already participate in these other projects.
Ultimately, I think the manuscript is a useful and timely work that merits submission. My suggestions below are to help readers from other fields within climate modeling appreciate the need to understand meltwater impacts and, in the long term, work towards creating coupled climate models that include interactive land and sea ice.
Signposting suggestions:
As it stands, each main section jumps directly into a subsection. A few sentences at the beginning of each section could provide extra motivation and emphasize the connections between the review material and the proposed experiments.
While the experimental descriptions are in the appendix, it would be useful to the reader to have a brief outline of the experiments and overarching themes of the project in the introduction. Perhaps in the introduction, the aims/motivating questions can even be listed with bullet points. That way, the reader can reference either the introduction or the appendix as they read the rest of the manuscript, depending on their level of involvement.
References to the SOFIA experiments are sparse and general (see line 128 for an example), which hinders readability. I recommend that the references to the experiments be more specific if the authors decide against adding an “experimental setup” section to the main text.
It would be nice to see more discussion, both within the main text and in the descriptions of the experiments in Appendix A, about why these specific experiments were chosen. For instance, regarding the preindustrial antwater experiment, how will the results be used in coordination with the other experiments (if at all), considering the differences in magnitudes? Will the authors test the linearity of the response using the idealized historical experiments? What are the advantages of using SSP-style forcings for future projections over the 1% CO2 experiments commonly used in other model intercomparison projects such as FAFMIP?
I would be interested to know if there are additional benefits to these runs outside quantifying uncertainty from the Southern Ocean. For instance, is there the potential to use some of these results to improve global or regional projections of sea level rise? The connection to biogeochemistry is mentioned briefly, but I would also be interested to read more about any (even speculative) impacts.
Minor points:
- Some sentences are difficult to read and have slight grammatical mistakes – for instance lines 19-21 has a misplaced modifier, and the clauses in 156-157 do not share the same subject. Overall, the manuscript has quite complicated sentence structures where the meaning can be unclear (e.g., line 163-164). I would recommend the longer sentences throughout the manuscript be broken up to improve clarity.
- Line 45: is there a contrasting study to Beadling et al., 2022 missing?
- Lines 97-98: What are the estimates from the Shepard et al. and Green et al. papers?
- Line 195: Behrens et al., 2015 (doi: 10.1002/2015JC011286) investigates this with the CMIP5 models – perhaps good to include
- Lines 288-289: A reference to FAFMIP in the main paper would also be a useful comparison as it is an established model intercomparison that includes freshwater forcing.
Citation: https://doi.org/10.5194/egusphere-2023-198-CC1 -
RC1: 'Comment on egusphere-2023-198', Anonymous Referee #1, 05 Apr 2023
The manuscript describes a new model intercomparison project devoted to the analysis of the impact of freshwater release in the Southern Ocean. Many studies have performed experiments imposing freshwater perturbations in this region but using different experimental designs, making an interpretation of the differences between the results of those experiments a complex task. This new initiative is thus timely. The protocol is clearly explained and several modelling groups have already agreed to participate in the exercise. It is thus very likely that the initiative will be a success.
1/ One of the stated main objectives is to have a protocol that is as simple as possible in order to ensure a wide participation. For instance, freshwater will be only added at surface, using simplified geographical distributions. The experimental design is consistent with this goal and this is well-adapted for an intercomparison exercise and to study why models have different responses to the same perturbations. However, it does not mean that the experiments will reproduce well the recent observations or the dynamics of the real system. It is perfectly fine to estimate the magnitude of the freshwater perturbation based on estimates of the mass balance of the Antarctic Ice sheet to have a reasonable order of magnitude but this can give the false impression that the intercomparison will give us a kind of measure of the uncertainty on the recent past and future impact of freshwater release in the Southern Ocean. The experiments will measure some elements of the uncertainty, related to model uncertainty, but not many others such as the impact of release at different depths or location, which could have a large impact too. This should be more clearly stated when the goals of SOFIA are described, starting for instance in the last sentence of the abstract which mentions ‘a consistent estimate of the climate system response to Antarctic meltwater, as well as the uncertainty of this response’ while it is only the uncertainty in the framework of the protocol related to the use of different models and different idealized scenario that will be estimated.
2/ In the same line, the manuscript mainly mentions CMIP6 type of models (see Table 2). This focus is consistent with the goals of the project but the introduction should describe more extensively the limitations of those models, in particular through a comparison with the results of higher resolution models. For instance, it is mentioned that there is a consensus on the expansion of the sea ice cover in response to the freshwater input but some studies have shown on the contrary a reduction of the ice thickness in response to ice shelf melting (e.g., Merino et al. 2018, Mackie et al. 2020, already cited in the manuscript). Changes in the dynamics of the slope current that cannot be resolved at the typical resolution of CMIP6 models also have an impact on the response to the perturbation and specifically on the exchanges with the continental shelves that can have a large impact on deep water formation, a topic that is part of the objectives of this study (Moorman et al. 2020; Beadling et al. 2022, already cited). This does not reduce the interest of the proposed intercomparison but should be discussed clearly to avoid an over-interpretation of the results of the intercomparison (see for instance lines 220-221).
3/ More and more ocean models include a representation of ice shelf cavities and we can expect that in the future several CMIP-type models will have such a representation. This is not discussed in the experimental design. Should these models be excluded from SOFIA and maybe be included in a future phase of the initiative as they will compute interactively some of the freshwater input that others will have to add at surface (for instance in the future scenario runs) or are they suggested adaptations to include them already at his stage?
4/ Minor points.
Line 96, I would replace ‘total’ by ‘additional’ as the total freshwater input includes also the climatological mean fluxes, not just the recent increase as more explicitly stated line 99-100
I would add in Table A1 that all the experiments are made with the Antarctic adjacent distribution, except 60Swater as currently the distribution is only clear for this experiment.
Citation: https://doi.org/10.5194/egusphere-2023-198-RC1 -
RC2: 'Comment on egusphere-2023-198', Nicholas Golledge, 19 Apr 2023
Swart et al: SOFIA
General comments
This paper outlines a new climate modelling initiative called SOFIA, that aims to assess the impact of including ice-sheet meltwater discharge on future projections of Southern Ocean and global climate evolution. This builds on the many studies that have done such experiments previously, but aims to formalise and standarise the procedures employed so that inter-model differences can be identified and quantified.The paper is succinct and sets out the intentions and procedures clearly. I don't think it does a particularly objective job of summarising the previous literature, however, with overt reliance on some of the same few citations throughout (Bronselaer et al 2018 & Fyke et al 2018 for example are both cited repeatedly, yet Fyke was simply a review paper and not a modelling study). I realise that the authors are more familiar with the climate model literature than the ice sheet modelling literature, but I think some further work could be useful in this area, to avoid the current bias.
The figures are fine and are useful, and with the modification suggested below I think Table 1 will be a very useful summary.
Specific comments
l2 - is 'injecting' the most appropriate verb to use here? The ice is melting slowly over a wide area, so maybe 'releasing' would be better?l2-3 - "This freshwater input could feed back onto
climate change, particularly since the Southern Ocean is a key contributor to global heat and carbon uptake" - I find this sentence a bit vague. Why not lean on previous research and be more definite: "Previous studies have shown that this freshwater input could lead to xxx and yyy, exacerbating climatic changes already underway" or something like that.l5-6 - "unaccounted for in current global
climate change projections" - I would say 'typically unaccounted for', because some GCMs with interactive ice sheets do exist, and not all 'climate change projections' are just about temperature, some relate to SLR for example, for which several scenario-based simulations incorporating meltwater feedback now exist. I would suggest taking a look at AR6 Ch9 for some info on that.l11 - is it the 'team of scientists' that is important here, or the fact that different models are being used with standardised methods?
l19 - citation to support the first statement perhaps?
l50-51 - "Since virtually
all existing Southern Ocean hosing experiments have involved only one model" - perhaps clarify that you mean 'have EACH USED only one model', because at the moment it reads as though all the experiments have used the same model, which is not the case, right?Table 1 - this is a very useful summary of the studies that have been undertaken. However, at the moment it only considers the climate model side of things. Since one of the key arguments being made to justify the SOFIA initiative is that few models incorporate interactive ice sheets, how about having a column that shows the full suite of model components being employed? E.g. 'OA' could be a standard ocean-atmosphere setup, whereas 'IOA' could be ice-sheet/ocean/atmosphere. I feel like this could be very useful for identifying which studies have 'closed the loop' as it were. For example, the Bronselaer paper is repeatedly cited throughout this paper but they didn't include a ice sheet model, whereas Golledge et al & Sadai et al did. The groupings in the table could then be adjusted to reflect this.
l113-115 - I think it might be worth mentioning that these kind of coupled experiments using interactive ice sheets are more common for Greenland - see for example Vizcaino et al 2015, Muntjewerf et al 2021 and various others. Again, much of this literature was assessed in AR6 Ch9 so could be easily discovered and included, perhaps more in the intro part of the paper than in this particular section though.
l133-4 - I think this statement "using these to force stand-alone ice sheet models" with the citation to the four papers cited is misleading, because Edwards et al presented only emulated results, not direct model outputs, and Golledge et al used a two-way (albeit offline) coupling specifically to capture the ice-ocean feedback. Perhaps this sentence can be rewritten to clarify what each of these studies did.
l154-6 - "Furthermore, feedbacks have been hypothesized, where increased meltwater input further enhances on-shore ocean heat transport through different processes in different regions (Hellmer et al., 2017; Hattermann, 2018; Bronselaer et al., 2018)." Yes, true, but why not also mention that some models have actually TESTED this feedback and quantified the response?
l176 - see note earlier regarding 'only one model'
Secion A4.3 - might be worth mentioning somewhere that Bakker et al 2017 used hosing experiments to explain some of the centennial-scale variability in climate during the Holocene - the timescale is different to the Historical experiments planned here, but the rationale and mechanics are the same.
Section A4.4 & 4.5, 4.6 - using ISMIP6 for basal melting is ok, but why then use an approximation for the calving flux? This assumes that the calving to basal melt ratio stays constant in time, which might not be true. It would be more physically robust to use calving fluxes directly from the ISMIP model outputs.
N R Golledge, 20th April 2023
Citation: https://doi.org/10.5194/egusphere-2023-198-RC2 -
CC2: 'Comment on egusphere-2023-198', Karen J. Heywood, 28 Apr 2023
This is a fantastic initiative and a most interesting paper. I wondered if you were aware of the paper by Richardson et al. In 2005, which was, I believe, the first to undertake an experiment using a coupled climate model (HadCM3) to investigate the impact of a meltwater addition around Antarctica, demonstrating an increase in Antarctic sea ice and subsurface ocean warming.
Short-term climate response to a freshwater pulse in the Southern Ocean
Glen Richardson, Martin R. Wadley, Karen J. Heywood, David P. Stevens, Helene T. BanksGeophysical Research LettersFirst published: 01 February 2005https://doi.org/10.1029/2004GL021586Of course, the models and experiments are much better nowadays, and longer model results are possible. I look forward to further results from Sofia!Citation: https://doi.org/10.5194/egusphere-2023-198-CC2 -
CC3: 'Comment on egusphere-2023-198', Paul Holland, 10 May 2023
Hi Folks
Many congratulations on this initiative, which is very timely and will be a very important addition to this field. The range of experiments out there at present does complicate matters, and SOFIA should clear that up very nicely. I have a few comments on the experimental design. Apologies for the unsolicited review!
Clearly the over-riding goal of SOFIA is to provide a consistent set of experiments that are practically achievable to the modelling centres, and that will necessitate some compromise on the physics. I am fully sympathetic to that. Therefore I am really expressing the below in order to raise a few possible caveats to the results that emerge, and to prompt future discussion, rather than to suggest any change in experimental design.
- The climate models will still have their ‘runoff’ (as shown in Figure 1b) from Antarctica during these experiments, but that will provide an uncontrolled additional Antarctic freshwater source to the ocean that may complicate the intercomparison of models. For example, there is a large increase in precipitation onto Antarctica during SSP5, which varies between models, and that will flow straight into the ocean. This precipitation increase can be the same size as changes in ice discharge during the 21st century so I think it is important (e.g. Seroussi et al 2020). The best solution could be to turn off runoff during these experiments and just implicitly accumulate the mass on Antarctica. Next best could be requesting further control experiments where needed (e.g. SSP585 without added FW) so that the effect of adding FW can be isolated from the model runoff changes. Simplest would be just analysing the runoff alongside the other results.
- It is a very uncertain choice taking the future parameterised melt from ISMIP6 models and multiplying by ~2 to account for calving. Calving could not change at all (larger, thinner ice shelves are implicit in the ISMIP6 melt rates) or could increase massively (ice shelves collapse as the grounding line retreats). Even if calving did increase, melting should be reduced accordingly, so it is not clear whether x2 is needed. One solution could be to take the change in ice sheet discharge across the grounding line from the ISMIP6 models, because that circumvents this point.
- Inputting freshwater into the surface, uniformly around Antarctica, with no latent heat extraction, produces a clear, tractable experiment but it is an unrealistic choice. In reality almost all of the additional mass flux has appeared in the Amundsen Sea, at depth, with a latent heat extraction. (Admittedly some ice shelf collapses have occurred on the peninsula - though those bergs were quickly exported - and the Ross and FRIS cavities could melt by 2100 under SSP5.) As with all of my points, I am not concerned about the model experimental design, just its interpretation. For me the stated design creates three issues: 1) Hosing the ocean with freshwater at the surface and no strong cooling produces a strong stabilising effect, which is not representative of real glacial melting. I assume this will cause excessive shutdown of dense water formation in ‘cold’ shelf regions, and excessive warm-water feedbacks in ‘warm’ shelf regions; 2) The uniform near-coastal distribution hoses the ocean directly in the ‘cold’ regions that form dense AABW, unlike in reality, thus over-stating the potential effects on AABW; 3) the 60S distribution hoses the ocean beneath the sea ice, over-stating the potential effects on sea ice and Southern Ocean SSTs. Issue 1 is compounded in issues 2 and 3, of course. So I think the current experiments are all extreme in over-stating the role of FW. For me the best solution would be to add the opposite extreme experiment with the all FW added in the Amundsen Sea only, distributed over depth, and with corresponding latent heat removed.
As I said above, all of these are just thoughts for discussion. SOFIA is fine as-is, and the authors should feel free to ignore these comments or take whatever action they feel sensible. I hope these thoughts help more than they hinder. Best of luck with SOFIA.
Cheers, Paul
Citation: https://doi.org/10.5194/egusphere-2023-198-CC3 - AC1: 'Response to reviewers and community comments', Neil Swart, 16 Jun 2023
Peer review completion
Journal article(s) based on this preprint
Neil Swart et al.
Neil Swart et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 1,175 | 628 | 27 | 1,830 | 23 | 11 |
- HTML: 1,175
- PDF: 628
- XML: 27
- Total: 1,830
- BibTeX: 23
- EndNote: 11
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Cited
1 citations as recorded by crossref.
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1304 KB) - Metadata XML