the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 17 Oct 2023
Past Antarctic summer temperature revealed by total air content in ice cores
Dominique Raynaud
Qiuzhen Yin
Emilie Capron
Zhipeng Wu
Frédéric Parrenin
André Berger
Vladimir Lipenkov
Abstract. Seasonal temperature reconstructions from ice cores are missing over glacial-interglacial timescales, preventing a good understanding of the driving factors of Antarctic past climate changes. Here the total air content (TAC) record from an Antarctic ice core is analyzed over the last 440 thousand of years (ka). While the water isotopic record, tracer for annual mean surface temperature, exhibits a dominant ~100 ka cyclicity, the TAC record is associated with a dominant ~40 ka cyclicity. Our results show that the TAC record is highly correlated with the mean insolation over the local astronomical half-year summer. It also shows for the first time that it is highly correlated with local summer temperature simulated with an Earth system model of intermediate complexity. This suggests that the Antarctic TAC records could be used as a proxy for local summer temperature changes. Also, our simulations show that local summer insolation is the primary driver of Antarctic summer surface temperature variations while changes in atmospheric greenhouse gas concentrations and northern hemisphere ice sheet configurations play a more important role on Antarctic annual surface temperature changes.
Dominique Raynaud et al.
Status: open (until 12 Dec 2023)
-
RC1: 'Comment on egusphere-2023-2360', Anonymous Referee #1, 09 Nov 2023
reply
Review of Raynaud et al.: Past Antarctic summer temperature revealed by total air content in ice cores
The authors use existing ice core data and new climate model simulations to provide a new interpretation of the ice core total air content (TAC) proxy. Based on a model-data correlation analysis, they suggest that the EDC TAC record reflects local summer temperature. The idea is interesting if corroborated, and the paper is overall well-written.
Main comments:
(1 )After reading the paper, it was not exactly clear to me what the authors believe to be the mechanism for the TAC-summer temperature relationship. Which of these is it:
(a) Insolation controls summer temperature. Summer temperature controls TAC.
(b) Insolation controls summer temperature. Insolation controls TAC.
(c) We cannot distinguish between scenarios (a) and (b) based on the available evidence
(d) Something else altogether.Please indicate which of these options describes your understanding best, and please clarify this in the manuscript as other readers may have the same confusion as I do.
(2) The analysis relies on correlation analysis between the EDC TAC record, the mean insolation over the local astronomical half-year summer, and the modeled summer temperature. The correlation between TAC and modeled summer T is stronger than the correlation between TAC and insolation alone. Could this be because both the TAC and summer temperature have 100ka power, whereas the insolation metric does not? Presumably the summer T has 100ka power because the GHG forcing that was applied.
I would request the authors include additional figures to clarify these questions better.
In Figure 1 could you please add the time series and spectrum also for the (i) mean insolation over the local astronomical half-year summer, (ii) modeled summer temperature, (iii) modeled mean-annual temperature? This could also be a separate figure of course.
In Figure 2 could you please add a panel with the correlation between Half-year summer insolation and the modeled summer T
(3) Could you provide a figure that compares the accelerated simulations with the non-accelerated simulations for both summer and annual-mean temperatures? It seems important to assess how well these agree with one another.
(4) How well does the model succeed in simulation the modern-day EDC annual temperature cycle? That may be important in assessing how well it can simulate the seasonal contrast back in time. Does the model simulate the Antarctic inversion, or is this not resolved in an EMIC? My sense is that the success of the model in simulating the seasonal temperatures will be closely linked to how well it can simulate the strong Antarctic inversion. The inversion is likely to respond in different ways to ORB, GHG and ICE forcings, which may bias the model output. Please add some discussion and/or caveats to the paper on this aspect of the modeling.
Minor comments:
Line 20: “summer insolation” is vague in this context. Do you mean “mean insolation over the local astronomical half-year summer” again, or something else?
Line 25-26: “Mean annual temperature”: technically it would be the “precipitation-weighted condensation temperature”
Line 35: I think Southern Ocean sea ice extent is another key parameter here
Line 83: and “a” detailed description….
Line 91-98: why not change SH ice sheets? Those would be the most important for EDC, no?
Figure 1B: the magnitude of EDC glacial-interglacial temperature change is uncertain, as per the discussion in the paper later on. It may be better to just plot d2H or d18O instead.
Line 114 -115: I understand what is meant here, but the language is imprecise as there are not different “types” on insolation. Maybe rephrase to: “Comparing a proxy record with insolation is not necessarily straightforward because different metrics of insolation exist and their relationship with climate is not always clear”
Line 126: perhaps remind the reader that 75 degrees S is the latitude of EDC.
Figure 2: Please add a third row of panels in which you compare the half-year summer insolation to the modeled half-year summer temperature. That way the reader can evaluate the relationship between these two key parameters better
Line 172-175: is it possible to evaluate how well LOVECLIM simulates the cited reconstructed seasonal temperatures from West Antarctica to validate the model simulations?
Line 181-182: Please provide more justification for this statement. Is this based solely on the increased correlation coefficient?
Line 186: “This is information….” (information is singular)
Line 185-191: Can you clarify this discussion? Is TAC just a proxy for summer temperature, or is summer temperature actually driving TAC variations? The discussion here still assumes that TAC is driven by insolation directly.
Figure 5: For comparison, can you please also plot the mean annual temperatures from the accelerated simulations that span the full 440ka period?
Line 217: “Response of Antarctic climate” (remove the ‘a’ at the end of Antarctica)
Line 221-222: In light of the discussion that follows, perhaps rephrase in more neutral language? For example: “One may also note that the magnitude of the temperature change between glacial and interglacial is significantly smaller in the model as compared to the reconstruction”
Line 223: It is my understanding that Antarctic elevation changes would exacerbate the problem, as EDC likely had lower elevation during the LGM.
Line 239-251: This is really excellent, and very instructive. What is the role of winter temperature? Is it fully controlled by GHG, such that in the annual average both GHG (winter) and ORB (summer) show up?
Line 249: The GHG effect is 1 degree on both mean-annual and summer temperatures, so it is actually the same. Perhaps better to phrase it as the *relative* importance of GHG is bigger on annual-mean.
Fig, 6: The model appears to simulate a seasonal temperature cycle of around 16 degrees (twice the offset between summer and annual-mean). The observed seasonal cycle is closer to 40 degrees (from -25 in summer to -65 in winter). Please comment. How well does the model capture the seasonal variations in the modern day?
Section 6: can you me more clear in your usage of the words snow and firn? How do you define these? Except for the very fresh surface snow, anything older than 1 year I would call firn. But it appears you use the terms differently.
Line 273: “ice grains” instead of “snow grains”?
Line 274: “upper few meters of THE snow column”
Line 278: the critical density you refer to here is around 550 kg/m3, correct? Or do you mean the critical density for pore closure? Please specify for clarity. What depth is this at EDC?
Line 279-280: is it no correlation or a poor correlation? Has to be either, can’t be both.
Line 281: there is no surface temperature record. Do you mean a correlation between TAC and the modeled summer temperature?
Line 283: what is the number of pores per grain? Simply the ratio of the nr. of pores to the number of ice crystals?
Line 284: do you mean the critical density of firn?
Line 284: This transition between GBS and PLC occurs at around 550 kg/m3, correct? Please specify for clarity. What depth is this at EDC?
Line 304-307: See my main comment, I would appreciate more clarity on what the authors think the causal relationships between insolation, summer temperature, and TAC are.
Line 308: Ultimately this insight does not come from the TAC record, but from the climate model. Do you agree? Without the climate model, one would not have interpreted TAC as a summer temperature proxy.
Citation: https://doi.org/10.5194/egusphere-2023-2360-RC1 -
RC2: 'Comment on egusphere-2023-2360', Anonymous Referee #2, 17 Nov 2023
reply
General comments:
The manuscript explores the idea that there are no good reconstructions of summer temperature and suggests that TAC could be used as a proxy for summer temperature in Antarctica. The authors use a previously measured record of TAC from the EDC ice core, which spans the last 440 ka. They use the EDC TAC data to compare with a climate model simulation of summer temperature to make the case that TAC could be used as a summer temperature proxy.
Overall, the manuscript is good, and the idea for using TAC as a summer temperature proxy is important, especially if it could be shown to also be used on other ice cores (a subject for future research). I recommend its publications after the authors consider the below comments:
Specific comments:
Overall, it is difficult to follow if the authors main point is that the TAC is controlled by summer insolation or summer temperature, or both. The authors clearly describe that both summer insolation and summer temperature are well anti-correlated with TAC. To make this clearer, I would recommend reorganizing the conclusion, highlighting their main argument at the beginning of the first paragraph.
Line 190 the authors state that TAC can be used as a proxy for summer temperature, based on the strong anti-correlation with modeled summer temperature. Then in section 6 the authors highlight that summer temperature influencing firn metamorphism is only a low-accumulation site phenomenon. Does this mean that TAC could not be used as a temperature proxy in Greenland? If this is the case, maybe the language about using TAC as a proxy for summer temperature is too strong for their results. While their result is interesting, before saying that TAC is a proxy for summer temperature, the relationships should be verified at multiple ice core sites.
Line 49: What is the difference between V (air content) and TAC (total air content)? A differentiation of what the authors mean by V vs TAC is required.
Line 84: What is a 10x acceleration, and why is it important in this context? I’m guessing the 10x model was used to save resources, but I recommend that more information is presented about why the accelerated models were used instead of the non-accelerated simulations.
Line 111: The sentence beginning “this spectral characteristic…” suggests that orbital astronomical forcing drives changes in TAC. The studies correlations do show a link between TAC and orbital patterns, but I think this sentence is misleading, and should be changed to reflect that TAC is only correlated with astronomical forcing, not actually caused by it. Later, I believe the authors make the argument that temperature gradients are what is causing TAC to vary, not insolation alone.
Line 127: I like the description of the astronomical half-year, and it is intuitive as to why it would be used as opposed to ISI.
Line 151: Here the authors say that TAC can be considered a proxy driven by mean local summer insolation based on the correlation. I’m a bit unclear here, because the correlation is higher (0.58) between TAC and summer temperature than between TAC and insolation (0.39). Do the authors mean that insolation drives summer temperature drives TAC? Or are they referring to previous works results using integrated summer insolation? If so, what was the correlation coefficient in that instance?
Line 149, 150, 170: The authors use the term ‘correlated’ when I think they mean anti-correlated, or negatively correlated. This needs to be clear, as it can cause confusion. I recommend all instances are reviewed.
Line 177: Figure 2 does not show any correlation between summer temp and summer insolation. Recommend adding this to figure 2 or deleting this line.
Line 280: What does “no (poor) correlation” mean? Please give a correlation coefficient. You reference section 5, but maybe figure 1 would be a better reference?
Line 282: What does it mean to ‘affect negatively” the critical density? Does this mean the other factor decreases the critical density?
Line 308: I thought the comparison came from dD and modelled summer temperature, not TAC and dD?
Technical Comments:
Line 3 – insert “,a” before “..tracer”
Multiple places EDC is referred to as Dome C. Recommend using EDC throughout the paper.
Line 155 – Recommend starting new paragraph here, where you start to discuss orbital tuning.
Figure 3 units – Are your units on the left hand side correct? I think they should be the same order of magnitude as the right side.
Line 211 – Strange spacing issue here.
Line 290: Recommend deleting the second ‘here’ in the sentence.
Line 312: Missing a word. Perhaps “Our transient simulation which allows us…”
Citation: https://doi.org/10.5194/egusphere-2023-2360-RC2
Dominique Raynaud et al.
Dominique Raynaud et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 181 | 76 | 5 | 262 | 4 | 7 |
- HTML: 181
- PDF: 76
- XML: 5
- Total: 262
- BibTeX: 4
- EndNote: 7
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1