the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Amplified surface warming in the Southwest Pacific during the mid-Pliocene (3.3–3.0 Ma) and future implications
Jonny H. T. Williams
Sebastian Naeher
Osamu Seki
Erin L. McClymont
Molly Patterson
Alan M. Haywood
Erik Behrens
Masanobu Yamamoto
Katelyn Johnson
Abstract. Based on Nationally-Determined Contributions concurrent with Shared Socio-economic Pathway (SSP) 2–4.5, the IPCC predicts global warming between 2.1–3.5°C (very likely range 10th–90th percentile) by 2100 AD. However, global average temperature is a poor indicator of regional warming and Global Climate Models (GCMs) require validation with instrumental or proxy data from geological archives to assess their ability to simulate regional ocean and atmospheric circulation, and thus, to evaluate their performance for regional climate projections. The Southwest Pacific is a region that performs poorly when GCMs are evaluated against instrumental observations. The New Zealand Earth System Model (NZESM) was developed from the United Kingdom Earth System Model (UKESM) to better understand Southwest Pacific response to global change, by including a nested ocean grid in the Southwest Pacific with 80 % greater horizontal resolution than the global-scale host. Here, we reconstruct region Southwest Pacific sea surface temperature (SST) for the mid-Pliocene Warm Period (mPWP; 3.3–3.0), which has been widely considered a past analogue with an equilibrium surface temperature response of +3 °C to an atmospheric CO2 concentration of ~350–400 ppm, to assess the warming distribution in the Southwest Pacific. This study presents proxy SSTs from seven deep sea sediment cores distributed across the Southwest Pacific. Our reconstructed SSTs are derived from molecular biomarkers preserved in the sediment – alkenones (i.e., U37K' index) and isoprenoid glycerol dialkyl glycerol tetraethers (i.e., TEX86 index) and are compared with SSTs reconstructed from the Last Interglacial (125 ka), Pliocene Model Intercomparison Project (PlioMIP) outputs and transient climate model projections (NZESM and UKESM) of low to high range SSPs for 2090–2099 AD. Mean interglacial equilibrium SSTs during the mPWP for the Southwest pacific sites, were on average, 4.2 °C (1.8–6.1 °C likely range) above pre-industrial and show good agreement with model outputs from NZESM and UKESM under mid-range SSP 2–4.6 conditions. These results highlight that not only is the mPWP an appropriate analogue when considering future temperature change in the centuries to come, but also demonstrate that the Southwest Pacific region will experience warming that exceeds that of the global mean if atmospheric CO2 remains above 350 ppm.
- Preprint
(2344 KB) -
Supplement
(45 KB) - BibTeX
- EndNote
Georgia Rose Grant et al.
Status: open (until 05 May 2023)
-
RC1: 'Comment on egusphere-2023-108', Anonymous Referee #1, 09 Mar 2023
reply
In this study, the author synthesized the SST reconstructions from the Southwest Pacific and compared them with Pliocene simulation and future projections. They suggest that Southwest Pacific will experience more substantial regional warming than the global mean. The paper is well-written and can be accepted after minor revisions.
Line 103-104. The sentence is misleading. Anyway, the warm mPWP is an equilibrated climate. Whether the future climate (unequilibrated) can reach the mPWP condition remains unknown.
Figure 3. It is difficult to read Figure 3. Please consider replotting the figure with lines and shaded areas.
Figure 4. It is difficult to distinguish proxy in plot b.
Figure 5. The caption should be revised; consider using “Ensemble mean regional SST from core PlioMIP experiments”.
In Table 4, please add the simulated global mean. It will be helpful to show if the simulated regional warming exceeds the global mean.
Figure 6. The gradient plots look misleading. The mismatch between the simulated and the reconstructed gradient is mainly due to the reconstructed strong warming in DSDP590. Therefore, it is better to add some discussion about the gradient when the warming at DSDP590 is discussed.
Line 456, “recorded” is not a good word to describe simulations.
Line 571-575, it is better to remove this from the conclusion. In this study, the authors do not provide solid evidence to show the change in EAC. Even though the models underestimate the warming at DSDP590, they should show some signals for the EAC change. However, the authors do not investigate that.
Citation: https://doi.org/10.5194/egusphere-2023-108-RC1 -
RC2: 'Comment on egusphere-2023-108', Anonymous Referee #2, 21 Mar 2023
reply
The manuscript of Georgia Rose Grant et al. looks at regional warming in the Southwest Pacific between proxy and modelled Pliocene state and future predictions for the century. The manuscript essentially shows that regional warming can be more intense than global warming, which is not so surprising, in particular in regions outside of the tropics, but the study is innovative in showing the effect of model resolution on such warming. Resolution is usually lagging behind in paleo-simulations due to the complexity of changing boundary conditions and extensive spin-ups required. Here, the authors used a "trick" by taking pre-industrial conditions and SSP scenarios from the high resolution model NZESM, and compare it to the Pliocene state as some of these scenarios predict similar temperature anomalies as for the Pliocene. I would be slightly concerned here as the authors compare transient state to an equilibrium state; this is addressed by the authors from L540, but I am not sure they fully answer/acknowledge the issues.
Overall, I would say the study is a first step in looking at resolution effects in paleo-simulations, and particularly how high resolution models could better solve some ocean components which eventually would affect their modelled temperatures and proxy-model comparisons. Nevertheless, I find the study to be confusing in some places, mainly when dealing with climate sensitivity. UKESM and NZESM are the only 2 models used to argue that high ECS models provide a comparable warming signature as from the proxy record, but they are basically the same model with a difference of resolution, so I find it difficult to argue as such for ALL high ECS models. There are other high ECS models, even within PlioMIP2, and clearly the message here is not due to resolution, so it should be possible to reinforce it by using other models (CESM2 / HadGEM...). Otherwise it could be just a unique behaviour of UKESM/NZESM, and not something systematic for all high ECS models.
Moreover, I find the handling of temperatures anomalies odd. I do not understand why the authors calculated modelled anomalies relative to the observational dataset HadISST. If any of your model holds any kind of bias (such as non-zero radiative balance, numerical approximations due to the cluster used), this will not be taken into account by calculating the anomaly relative to HadISST. It would have been in some way if it was calculated relative to the pre-industrial state of each run. I get that the authors want to have the same "base" of comparison as for the geological proxy data. But I think there is a world where the conclusions of the authors do not hold because their temperature anomalies might be quite different.
Besides those two main points, I find the text well-written and relatively easy to follow. The figures are informative and easy to understand. I am not sure I would connect with plain lines sites which are separated by more than 10° of latitude, but this might be a more personal opinion.
L86: I find this sentence either wrong or vague. Many estimates show Last Glacial Maximum climate sensitivity to be weaker than modern/warmer paleoclimates due to state-dependency on climate feedbacks (see PALAEOSENS, 2012, for instance). Yes, warm-based estimates are usually more consistent with high ECS; however, you cannot summarize that the entire paleoclimate record relates to high ECS.
L119: Why cite Haywood et al. (2016) here, and not Haywood et al. (2011)?
L120: Connected to previous comment; I do not fully understand this sentence or even where those values come from. As I said in Technical comment, I cannot find your reference of Haywood et al. (2012), but I will assume you refer to Haywood et al. (2013). In that paper, the only number I could find for fast-feedback ECS would exceed 3 K. Also, if you wish to be thorough with ECS, you cannot ignore Hargreaves and Annan (2016), which effectively combined PRISM data and PlioMIP1 models and came up with 1.9 - 3.7 K, or more recently Renoult et al. (2020) with 0.5 - 5.0 K. I assume the numbers you provide are in range of median values, where in that case both median estimates of Hargreaves and Annan (2016) and Renoult et al. (2020) fall in the range of 2 - 3 K, and in that case you should actually say that you are showing median estimates. From a first read, I immediately thought you were presenting confidence intervals, and that your old ECS estimates are very well-constrained and low, in particular compared to PlioMIP2 estimates.
L192: I find weird and essentially a circular argument to say that the ECS of UKESM is higher than the IPCC range, since the high end of the IPCC range is itself calculated/constrained by models such as UKESM. This is similar as saying "UKESM lies on the high end of the IPCC range which itself is based on UKESM high ECS". I think this requires either a rephrasing, or maybe getting rid of the IPCC reference.
L201 - 205: I do not fully get this part. The way I understand it is that you calculated temperature anomalies of models between a perturbed (non-pre-industrial state) model run and an observational dataset, which implies that inherent biases in the models are kept in your perturbed model temperature. Most likely your pre-industrial climate is not at zero top-of-atmosphere radiative imbalance, which you could have corrected by comparing the perturbed to the modelled pre-industrial state. I understand you used HadISST for proxy data, and I completely agree on not using a model control here though. But here the issue is that you will obviously have consequences on your model temperatures which are not accounted for and which come from model physics and simplifications, or even numerical approximations made by using different clusters.
L338 - 340: Is there a better way of introducing those results? In one sentence we are given 4 ranges of temperature and 2 mean values for the same sites, and I quickly lost the thread.
L455: Here you report a PlioCore SST anomaly (3.2C), which is calculated as the anomaly between Pliocene ("perturbed") state and control. However, you compare it to the PlioCore SST average, which I guess is calculated from the difference between model and proxy (if I read Table 3 of the reference). However, I know that McClymont et al. (2020) have temperature anomalies of around 3.2C, so I would guess it is just a weird referencing. I would simply get rid of the Haywood et al. (2020) reference in that line.
L454 - 455: Overall I have a bigger issue because there is no uncertainty given on those values. You say the PlioCore simulations do not record the amplified warming signal in Southwest Pacific, but what is the full range of the simulated temperature there?
L521 - 523: I get the point but it is also speculative. You could have many other things which influence the SST at that site in a model and coincidentally make it close to the reconstructed SST at that site (e.g. clouds), and then you could conclude that changes due to paleogeography actually influence surface water masses distribution. A way of approaching this problem would be to run PlioCore with UKESM / NZESM, although I understand this would be an entire new study.
L547: What is "accepted" range? This is the likely range. The very likely would put max ECS at 5, and then I would not consider 5.4 to far exceed it. Also, again, this argument is circular. The range of the IPCC you are providing is dependant of UKESM. It works better when comparing to mPWP, since there UKESM was not used.
L552 - 554: In Fig.7b you show the average PlioMIP state. Do you also find a comparable warming signature in those regions for the other high ECS models (of e.g. the PlioMIP2 ensemble)? CESM2 has a comparable, even higher ECS than UKESM / NZESM (5.6 K, see Zhu et al. (2022)), or even COSMOS could show something similar. If it is not the case, what you are seeing might be only due to UKESM (and the similarities it has with NZESM), and not a systematic behaviour of high ECS models. It seems your argument mainly relies on the ECS of the models, so you would not need a high resolution one to test it on multiple models.
Technical comments:
L75: cryosphere
In bibliography: Where is Haywood et al., 2012? I could not find it.
L285: Weird sentence ("seasonality can be up to offset...")
L302 and elsewhere in this section: Details about using R are unnecessary, since you could have used any other language combined with statistical modules to perform those analyses (Python could have done that).
L350: "half the that", remove the.
L359: The STF? I forgot what it was and I could not search for it, then realised it is in your Fig.1... I would write it once more in the text.
L360: remove the "," before "the STF".
L363: Here you talk about Subtropical Front but do not use the acronym STF though...
Fig.7 : I would not plot lines, as there is such a limited amount of data, in particular it seems weird to connect with a straight line point at around -40 and -30.
L511: Similar as before, here STF is not used.
-------------
References:Haywood, A. M., Dowsett, H. J., Robinson, M. M., Stoll, D. K., Dolan, A. M., Lunt, D. J., Otto-Bliesner, B., and Chandler, M. A.: Pliocene Model Intercomparison Project (PlioMIP): experimental design and boundary conditions (Experiment 2), Geosci. Model Dev., 4, 571–577, https://doi.org/10.5194/gmd-4-571-2011, 2011.
Haywood, A. M., Hill, D. J., Dolan, A. M., Otto-Bliesner, B. L., Bragg, F., Chan, W.-L., Chandler, M. A., Contoux, C., Dowsett, H. J., Jost, A., Kamae, Y., Lohmann, G., Lunt, D. J., Abe-Ouchi, A., Pickering, S. J., Ramstein, G., Rosenbloom, N. A., Salzmann, U., Sohl, L., Stepanek, C., Ueda, H., Yan, Q., and Zhang, Z.: Large-scale features of Pliocene climate: results from the Pliocene Model Intercomparison Project, Clim. Past, 9, 191–209, https://doi.org/10.5194/cp-9-191-2013, 2013.
Hargreaves, J. C. and Annan, J. D.: Could the Pliocene constrain the equilibrium climate sensitivity?, Clim. Past, 12, 1591–1599, https://doi.org/10.5194/cp-12-1591-2016, 2016.
Renoult, M., Annan, J. D., Hargreaves, J. C., Sagoo, N., Flynn, C., Kapsch, M.-L., Li, Q., Lohmann, G., Mikolajewicz, U., Ohgaito, R., Shi, X., Zhang, Q., and Mauritsen, T.: A Bayesian framework for emergent constraints: case studies of climate sensitivity with PMIP, Clim. Past, 16, 1715–1735, https://doi.org/10.5194/cp-16-1715-2020, 2020.
Zhu, J., Otto-Bliesner, B. L., Brady, E. C., Gettelman, A., Bacmeister, J. T., Neale, R. B., et al. (2022). LGM paleoclimate constraints inform cloud parameterizations and equilibrium climate sensitivity in CESM2. Journal of Advances in Modeling Earth Systems, 14, e2021MS002776.
Citation: https://doi.org/10.5194/egusphere-2023-108-RC2
Georgia Rose Grant et al.
Georgia Rose Grant et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 218 | 75 | 8 | 301 | 19 | 1 | 2 |
- HTML: 218
- PDF: 75
- XML: 8
- Total: 301
- Supplement: 19
- BibTeX: 1
- EndNote: 2
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1