the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
A comparison of the last two glacial inceptions (MIS 7/5) via fully coupled transient ice and climate modeling
Abstract. Little is known about the evolution of continental ice sheets through the last two glacial inceptions (Marine isotope stages, MIS 7d and MIS 5d). Here, we present the results of a perturbed parameter ensemble of transient simulations of the last two glacial inceptions and subsequent interstadials (MIS 7e-7c, 240–215 ka and MIS 5e-5c, 122–98 ka) with the fully coupled ice/climate model LCice. LCice includes all critical direct feedbacks between climate and ice. As shown herein, it can capture the inferred sea level change (of up to 80 m) of the last two glacial inceptions within proxy uncertainty. One key underlying question of paleoclimate dynamics is the non-linear state dependence of the climate system. Concretely, in a model-centric context, to what extent does the capture of one climate interval in an Earth systems model guarantee capture of another interval? For LCice, the capture of present-day climate is insufficient to predict capture of glacial inception climate, as only a small fraction of ensemble members that performed "well" for present-day captured inception. Furthermore, the capture of inferred sea level change in one inception has weak correlation with the same outcome for the other.
After partial history matching against present-day and past sea level constraints, the resultant NROY (not ruled out yet) ensemble of simulations have a number of features of potential interest to various paleo communities, including the following. (i) In correspondence with the inferred last glacial maximum configuration, the simulated North American ice sheets are substantially larger than the Eurasian ice sheet throughout MIS 5d-MIS 5c and MIS 7d-MIS 7c. (ii) Hudson Bay can transition from an ice-free state to full ice cover (grounded ice) within 2000 years. (iii) The North American and Eurasian ice sheets advanced southward with rates well above 100 m/yr during the penultimate glacial inception and over 70 (Eurasia) and 90 (North America) m/kyr during last glacial inception. (iv) the Laurentide and Cordilleran ice sheets merge in their northern sectors in 13 out of 14 NROY simulations for MIS 7d, contrary to what is assumed from limited geological data. (v) larger ice sheets display a larger lag in the timing of stadial maximum ice volume compared to that of the insolation minimum; the North American ice sheet maximum lags 5.3 ± 0.5 kyrs behind the MIS 7d insolation minimum. Supplemental resources include a dynamic display of ice advance and subsequent retreat for a sub-ensemble of 14 NROY simulations from MIS 5d-5c and MIS 7d-7c.
Competing interests: One author (Lev Tarasov) is a member of the editorial board of "Climate of the Past". The other authors declare that they have no conflict of interest.
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
(3495 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 15 Apr 2025)
-
RC1: 'Comment on egusphere-2025-495', Anonymous Referee #1, 19 Mar 2025
reply
Geng et al. present numerical experiments of the last two glacial inceptions with a coupled ice sheet – climate model. The experiments are thoroughly performed but I found the manuscript uneasy, long and mostly descriptive, with not-so-pretty figures. I found no clear take-home messages from this impressive amount of simulations.
Major comments
1) My main comment is about the general purpose of the manuscript. At present the manuscript is mostly descriptive with almost inexistent discussions on the physical mechanisms at play. Given the multiple sources of uncertainty, a model is not expected to perfectly reproduce past changes and the authors seem fully aware of this. Instead, a model is based on physical assumptions which can be tested in multiple ways and confronted to palaeo data. Geng et al. manuscript mostly discuss simulated ice sheet changes and how it compares with the palaeo data. What can we learn from this? It is not particularly useful to know that some experiments work better than other, what is useful to know is “why some experiments work better and why others do not”. The main justification for using forward coupled ice sheet – climate model experiments instead of an inverse approach is to be able to discuss mechanisms. For example, from the title, I would expect to see some discussions on the difference in terms of mechanisms for the last two glacial inceptions. Differences in oceanic heat transport for the two inceptions and how they are linked (or not) with the Eurasian ice sheet growth (or else)? How can we link the differences in ice sheet geometry with the orbital forcing? Different patten of precipitations and/or temperature for the two inceptions or simply a shift driven by the insolation anomalies? Why the Antarctic ice sheet is so stable for the two time periods? What makes that one experiment is successful and not the other?...etc. I think there is material for many interesting discussions but the aim of the paper has first to be clarified.
2) A less important comment is that we have very few information of the coupling is done. Earlier work is referenced but the manuscript lacks some important clarifications in the methods. Please explain briefly how the downscaling works and how the surface mass balance is computed. Same for the sub-shelf melt (quadratic dependency? Near neighbour interpolation?…). What kind of ice sheet model is GSM? What is its spatial resolution? How friction and calving are computed? Also, it seems that a relative sea level solver is used but how does it work? Does it assume an eustatic homogeneous value at the start of the transient experiments or a perturbation of the present-day geoid is used? It seems that the bathymetry (ocean grid model) is unchanged, but then what happened for the air-sea fluxes and albedo in the Hudson Bay area when it becomes glaciated? This last point is highly relevant since Hudson Bay area results are highlighted in the manuscript.
3) A time-dependent bias correction is used to correct the climate data used to force the ice sheet model. This correction is maximal when the total ice volume is close to present-day and decrease with increasing ice volume. This is motivated by the fact that we cannot guarantee that the biases remain constant in time. While it sounds appealing I find this approach not really justified. The biases can also, in principle, increase with increasing ice volume. For example, a too wet mountainous area might become even wetter if the surface topography increases (same for temperature). Right now I think the reduction of the bias correction with increasing ice volume is unjustified and seems simply like tuning strategy. It would be nice to show simulations with constant bias correction and to properly discuss what are the implications of the chosen methodology on the results. It strikes as odd to claim to have explored thoroughly the parametric uncertainty with a large ensemble when ad-hoc tuning strategies are used (bias dependent correction and additional warming in certain locations).
4) I feel a bit frustrated to have no information on the spun-up ice sheets and climate at 122 ka and at 240 ka. What the initial ice sheets look like – in terms of ensemble mean and standard deviation? Have you evaluated these initial states against palaeo data? Is part of the transient response can be explained by the state of the ice sheet and the climate at the beginning of the simulation? I expect much larger ice sheet after the 7 ka spin-up at 240 ka compared to the ones at 122 ka since both CO2 and Northern Hemisphere summer insolation are lower.
5) I can find no coupled modern simulations in the manuscript. It seems that the calibration against the present-day climate has been run with LOVECLIM only, not including the interactive ice sheets component. This seems illogical since it is plausible that amongst the 90 NROY members some of them would produce unrealistic present-day ice sheets. Perhaps the members that produce the largest ice mass gain for the inceptions does not reproduce observed present-day ice sheets under modern boundary conditions. It would have been useful to run modern / Holocene ice sheet – climate simulations to rule out the members that do not manage to maintain present-day ice sheets.
6) Is albedo downscaled? If not, the temperature provided by ECBilt to GSM has probably a strong imprint of the original albedo. Also if the albedo is indeed computed on the native T21 grid, it can explains the stepwise ice sheet advances and/or retreat, which is then an artificial feature of the model. Please comment / discuss the albedo in the manuscript.
7) One improvement with respect to Bahadory and Tarasov (2018) is the representation of an interactive Antarctic ice sheet component. However there is virtually no information on Antarctica in the paper. Was it expected to have a stable Antarctic ice sheet for the time periods considered? Don’t we expect a retreated ice sheet at 122 ka? Do the coupling affect the large-scale climate through freshwater and/or salt rejection?
8) The quality of the figures can be largely improved. Most of them are blurry (Fig. 1 or 3 in particular) or contain too much information (Fig. 1 can be probably split into two sub-panels).
Minor comments
- P4L100: “4 x acceleration” what is meant here?
- P4L105: ECBilt contains ageostrophic correction terms.
- P4L108: NA is defined only P7. More generally, do we need these abbreviations?
- P5 Fig.2: It would be useful to have the frequency of exchanges (annual, monthly,...)
- P5L110: I am not sure how robust are the results of Lofverstrom and Liakka (2018). We cannot use a model at different resolutions without retuning some parameters. It is likely that the base model of Lofverstrom and Liakka (2018) – meaning the one that has been calibrated – was the one at high resolution.
- P5L123-124: If the ocean cannot be turned to land, what happens to the Hudson Bay area in terms of albedo, air-sea fluxes,...?
- P6L129-131: The daily variability is not taken into account if monthly means are used – or do you use some kind of a paremetrisation for daily variability?
- P6L134: Salinity is not used in the melt equation? For ice sheet points with no corresponding ocean points, do you use some kind of spatial interpolation?
- P6L136-137: If this is a novelty from earlier work it has to be a bit more described… How the ice shelves are handled? Does it make a difference to have the interactive Antarctic ice sheet? Does it impact the Northern Hemisphere simulated climate?
- P6L139: It is not enough to cite a paper in preparation. We need to know how the surface mass balance is computed since it is a critical point for glacial inception.
- P6L141-148: What formula is used to correct the climatic variables? A simple delta method? Also for evaporation and winds? Please provide the equations used (in supplement) to correct the different variables.
- P6L152: What can we learn from this? Event though a bias correction method is used, we still need to impose an ad-hoc correction. This additional correction is here to make the results “pretty” or does it really change climate and ice sheet dynamics? This point deserves further discussion.
- P6L155: The value 1 is not included in this range while it is what has been used as standard in earlier LOVECLIM works. How compatible are these values with previous model evaluations? Also, the climate sensitivity is low with standard LOVECLIM parameters but the ensemble used here is very large and it is possible that it does not need this artificial climate sensitivity increase.
- P7L172: Refer to appendix A2.
- Figure A1: Precipitation units?
- P7L186: How these dates have been selected?
- P7L187-189: With this methodology, the climate and ice sheets are supposed to be in equilibrium while the sea level record shows large changes, especially for MIS7. Assuming equilibrium is a practical choice but it would be useful to have a discussion on how alternative climate and ice sheet at the start of the transient experiments can impact the results.
- P7L190-191: 7 kyr is not enough to spin up ice sheet internal variables (temperature / viscosity).
- P7L191-192: What freshwater flux is given to the ocean in accelerated experiments? Conserving mass would lead to overestimated fluxes (x4 too high).
- P9 Fig. 3: The ensemble members seem to start with very similar ice volumes at 240 and 120 ka. This is very surprising given the fact that the climate is necessary very different (lower Northern Hemisphere insolation and GHG at 240 ka).
- P11 L253-257: This is the only section mentioning the Antarctic ice sheet. What is the geometry of the spun-up ice sheet? The largest shelves are simulated? I find it not convincing to invoke a direct correlation between Southern Hemisphere maximum in insolation and minimum ice volume. Surface melt is low in Antarctica and most of SMB changes are linked to precipitation and oceanic sub-shelf melt. Do you imply that the sub-surface oceanic temperatures are correlated with the local insolation? There is probably a better (positive) correlation between global mean temperature and Antarctic volume (wetter climate). In its current form the Antarctic results are not informative.
- P12 Fig. 4: After an initial ice growth the model melts the Eurasian and North American ice sheets, which is not supported by proxy data. Does it mean that the albedo – melt feedback is too strong?
- P12L266: Specify in the figure that this is the summer isotherm.
- P13 Fig. 6: why not adding the Antarctic ice sheet in this plot – a correlation with Northern Hemisphere insolation is not excluded.
- P13L270: How do you evaluate the sea-level temperature? Which lapse rate is used?
- P16L312-316: Do we have a control of sub-shelf melt? Please discuss.
- P18 End of section 3.3.2: The model simulates the “right” volume but underestimates the extent. Is it because the model is too wet despite the precipitation correction? Or sliding/deformation is underestimated?
- P19 Fig. 10: Dates for ice geometry?
- P20 Sec. 4.1: But then, what is the purpose of Sec. 3.3 where model results are confronted to uncertain geological data? Again for me the strength of a model is to understand mechanisms and processes. Uncertain models are not meant to reproduce uncertain data…
- P21L417 vs. P21L424: Full ice cover in 2 kyr or 1 kyr?
- P21L426: It is really possible to discuss results in this area given the fact that a fixed bathymetry is used and as such the surface mass balance model is evaluated using atmospheric fields above an ocean when it is land.
- P21L436: This is probably due to an artefact of the model since the ice mask is evaluated at the atmospheric model resolution. A whole grid cell can artificially switch from unglaciated to glaciated causing discontinuity in the albedo (hence near-surface air temperature). This result should not be discussed as if it is a real thing.
- P22L456-464: I am not sure that it is related to atmospheric circulation. Albedo is not downscaled which can cause an artificially homogeneous albedo over the whole ice sheet. A sub-grid parametrisation of albedo could also help to reproduce this feature.
- P22L466-472: It looks like a simple repetition of a previous section.
- P23L496: “model’s capability to simulate present-day climate” – it seems to me that this is not completely justified. The coupled ice sheet – climate model has not been used for present-day.
- P23L505-506: It is not really a new results. If some modellers have used deglacial ice sheets for pre-LGM it was for simplicity, but the effect of glacial isostasy is well known and accepted for a while now.
- P24L520: The code is not publicly available, meaning that it is not possible for other scientists to have a look of the coupling strategy in LCice.
Technical comments
- Abstract: problem with the units - 90 m/kyr
- Abstract: remove last sentence
- P2L45: “between the two” → 4 sea level reconstructions are discussed.
- Fig. 1: show x- and y-axis units. Figure is poor quality and really hard to read. To improve the readibility you could to separate CO2 and insolation from the sea level reconstructions. It is very blurry right now.
- Fig. 4 & 5: Capital “m”
- P15L308: Typo – inception
- P25 Fig. A1: Units of precip? Change the color palette so that it appears while when there is no change.
- P27 Tab. B1: Add a short description of these parameters and their units.
Citation: https://doi.org/10.5194/egusphere-2025-495-RC1
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
108 | 18 | 5 | 131 | 4 | 5 |
- HTML: 108
- PDF: 18
- XML: 5
- Total: 131
- BibTeX: 4
- EndNote: 5
Viewed (geographical distribution)
Country | # | Views | % |
---|---|---|---|
United States of America | 1 | 33 | 25 |
China | 2 | 13 | 10 |
Japan | 3 | 13 | 10 |
Canada | 4 | 11 | 8 |
Germany | 5 | 10 | 7 |
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1
- 33