the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
The hydrological cycle and ocean circulation of the Maritime Continent in the mid-Pliocene: results from PlioMIP2
Daniel J. Lunt
Erica Hendy
Anna von der Heydt
Ayako Abe-Ouchi
Bette L. Otto-Bliesner
Charles J. R. Williams
Christian Stepanek
Chuncheng Guo
Deepak Chandan
Gerrit Lohmann
Julia C. Tindall
Linda E. Sohl
Mark A. Chandler
Masa Kageyama
Michiel L. J. Baatsen
Ning Tan
Qiong Zhang
Ran Feng
Wing-Le Chan
W. Richard Peltier
Xiangyu Li
Youichi Kamae
Zhongshi Zhang
Alan M. Haywood
Abstract. The Maritime Continent (MC) forms the western boundary of the tropical Pacific Ocean, and relatively small changes in this region can impact the climate locally and remotely. In the mid-Pliocene (from 3.264 to 3.025 million years before present), atmospheric CO2 concentrations were ~ 400 ppm, and the subaerial Sunda and Sahul shelves made the land-sea distribution of the MC different to today. Topographic changes and elevated levels of CO2, combined with other forcings, are therefore expected to have driven a substantial climate signal in the MC region at this time. By using the results from the Pliocene Model Intercomparison Project phase 2 (PlioMIP2) we study the mean climatic features of the MC in the mid-Pliocene and changes in Indonesian Throughflow (ITF) with respect to preindustrial. Results show a warmer and wetter mid-Pliocene climate of the MC and lower sea surface salinity in the surrounding ocean compared with preindustrial. Furthermore, we quantify the volume transfer through the ITF; although the ITF may be expected to be hindered by the subaerial shelves, 10 out of 15 models show an increased volume transport compared with preindustrial.
In order to avoid undue influence from closely-related models that are present in the PlioMIP2 ensemble, we introduce a new metric – the multi-cluster mean (MCM), based on cluster analysis of the individual models. We study the effect that the choice of MCM versus the more traditional analysis of multi-model mean (MMM) and individual models has on the discrepancy between model results and reconstructed proxy data. The clusters reveal spatial signals that are not captured by the MMM, so that the MCM provides us with a new way to explore the results from model ensemble that include similar models.
- Preprint
(3952 KB) -
Supplement
(6586 KB) - BibTeX
- EndNote
Xin Ren et al.
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2022-1281', Anonymous Referee #1, 16 Jan 2023
This manuscript examines simulations of the mid-Pliocene from the PlioMIP2 models. The authors focus on the mean climate features in the maritime continent. They find a warmer and wetter mid-Pliocene climate, with a lower sea surface salinity and in general a stronger Indonesian Throughflow (ITF) in the PlioMIP2 simulations. The author also explore the use of multi-cluster mean in summarizing multiple model results and noted the advantage of this method over the traditional multiple model mean.
The manuscript is very well-written and easy to follow. The manuscript fits well with the scope of Climate of the Past. I have a few comments that need to be addressed before publication. Please see details below.
- Results on ITF are not well connected with the rest of the manuscript. In other words, why should we care about the ITF in the Pliocene simulations (considering that we do not have proxy data to provide sufficient constraints on the model results)? In the current form of the manuscript, ITF is described separately from the SST and hydroclimate variables. Although, in the introduction, the author did cite literation on how the ITF is linked to coupled ocean-atmosphere variability and how the ITF may influence the monsoons. However, the authors results on ITF do not make any of the connection or mechanistic analysis. Given this disjoint, I am wondering whether the author should consider cutting the ITF results and focus on the regional SST and hydroclimate over the Maritime Continent instead.
- In the Discussion (Section 4.3), the authors stated that “but even models of the same model family may still produce different climatic signals depending on the analysis region or the studied climate characteristic.” Can you provide explanation for this interesting result? Is it because of the potentially different model resolution, or details of the boundary condition implemented by different authors, or internal variability?
- Are there available proxies on the hydroclimate (precipitation /evaporation and sea surface salinity) and ITF in the region? If yes, please include results and discussion on these comparisons. If no, please state it explicitly in the manuscript (that there is no available proxy for benchmarking models).
- Please consider adding a summary of model-proxy comparison of SST in the abstract.
Minor comments
- Lines 23–25: Rewrite and change into “A large amount of rainfall releases large quantities of latent heat into the atmosphere, which is an important driver of global atmospheric circulation”.
- Many of the multi-panel plots are not labeled with subplot label (such as (a) and (b). Please check and make sure all the subplots are properly labeled.
- Information should be provided on how the ocean salinity was initialized in the simulations. This information is needed because the authors examined the sea-surface salinity changes in the PlioMIP simulations (e.g., Figure 5d), and it is not clear whether the ice-volume effect has been accounted for in the simulations and has an imprint in Figure 5d.
- Line 266: “the relationship is not exactly linear.”
- Figure 10: cluster 5 (GISS) looks weird. The model resolution is ~2 degree (Table 1). It is hard to believe the precipitation anomaly has such a rich fine structure. Please double check and make sure calculation has been done correctly.
Citation: https://doi.org/10.5194/egusphere-2022-1281-RC1 - AC1: 'Reply on RC1', Xin Ren, 05 May 2023
-
CC1: 'Comment on egusphere-2022-1281', Tripti Bhattacharya, 13 Feb 2023
This paper takes a careful, nuanced analysis of the potential biases in PlioMIP2 models’ representation of mean climate state and whether this is related to bias in their simulation of Pliocene climate. This is important, since we normally justify the use of certain models to simulate past climates in a given region based on their ability to reproduce the pre-industrial climatology. However, if this is not the case, then we must be more careful of our choice of models.
The results presented required a huge amount of data compilation and analysis, and it is a nice demonstration of the utility of the multi-cluster mean approach to understanding model disagreement and consensus. I would recommend publication following consideration of some of the points below:
Figure 3: make the markers larger, more bold, it’s a little hard to see all of them directly. Would suggest using different colors rather than distinct symbols, or increase line weight or something
Make sure each panel is labeled (a) and b) do not appear on the figure
Figure 4: is it possible to propagate through the uncertainty (e.g. full error envelope of Pliocene proxy values) into the EOI400 discrepancy calculations? Or at least add some 95% error bounds.
Figure 5, and generally all figures with continental outlines: make the outlines of the land bold.
Figure 6 is very interesting. The relationship between ECS and SSTA over the maritime continent - I wonder if this can be formally connected to the ‘pattern effect’ literature. See for instance:
Dong, Y., Armour, K.C., Zelinka, M.D., Proistosescu, C., Battisti, D.S., Zhou, C. and Andrews, T., 2020. Intermodel spread in the pattern effect and its contribution to climate sensitivity in CMIP5 and CMIP6 models. Journal of Climate, 33(18), pp.7755-7775.
This literature points to the fact that long-term changes in climate feedbacks seems to depend on the relative warming in the western Pacific warm pool region.
While the text does a good job of articulating the differences between the MMM and the MCM, in the plots themselves they look quite similar for SSTA and precipitation. This might be a feature of the small size of the plots. Can the individual panels be made larger, and the dendrogram made much smaller, so that it is easier to compare and contrast the map panels?
Line 475: I would say more about the potential sources of proxy uncertainty for ODP 214. As of now the analysis of this potential source of model disagreement is quite cursory
Citation: https://doi.org/10.5194/egusphere-2022-1281-CC1 - AC2: 'Reply on CC1', Xin Ren, 05 May 2023
-
RC2: 'Comment on egusphere-2022-1281', Anonymous Referee #2, 02 Apr 2023
The study aims at evaluating and describing the main features of the Maritime Continent in mid-Pliocene through analysing model output of the second iteration of the Pliocene Model Intercomparison Project (PlioMIP2). To evaluate the model’s performance, the authors provide analyses on comparing the reconstructed and simulated SSTs around the Maritime Continent. In attempting to provide an overview of the MC climate the authors show results of sea surface temperature, surface water balance, zonal surface winds and sea surface salinity. To finalize, the authors have performed a cluster analysis to evaluate similarity among models and to reduce possible duplicated biases from models of the same family. I think providing a better understanding of the MC climate during the mid-Pliocene is of great significance, once the MC is very sensitive to changes in the both the Pacific and Indian Ocean as well as in large-scale atmospheric circulation. However, the study at its present form is still too preliminary and the manuscript needs substantial changes before publication. At the end of the reading, it is not clear how the MC climate was during the mid-Pliocene and the manuscript raises more questions than providing answers.
Major comments:
- The analysis and methods used by the authors are not sufficient to achieve their central objective. First, the hydroclimate of the MC is very complex and is under the influence of changes in mean state features (such as Hadley and Walker circulation, ITCZ position and the warming itself) and important internal modes of variability (such as ENSO, IOD and IOBW). Secondly, the study focuses on a very small region around the MC, from which is not possible to obtain a picture of the large-scale dynamics. Many results are described in terms of the small scales changes around the MC (i.e., North-eastern coast of New Guinea, central MC, gateway between MC and Australia, etc), which are not possible to be evaluated from the coarse resolution of the PlioMIP models. Additionally, the PlioMIP models are required to apply substantial changes in the land-sea mask in the MC and are of coarse resolution (some up to 4 degrees in the atmosphere), making its small-scale evaluation very difficult and uncertain. Furthermore, it is not possible to infer how the MC hydroclimate was during the Pliocene by simply evaluating the basic fields described in section 3.2. I recommend the authors to expand their study area to include the Pacific and Indian ocean changes, as well as southern Asia, which is a large land mass above the MC, where any small temperature changes may affect the atmospheric circulation and ITCZ position. Also, to evaluate the MC hydroclimate the authors must show results of Hadley and Walker circulations and the possible influence of changes in the main modes of variability (ENSO and IOD).
- It is not clear how three of the result’s sections (3.2, 3.3, 3.4) are linked to each other. These sections seem very independent from one another without a clear justification on why choosing these analyses to compose the manuscript. Section 3.2 must include more elements as mentioned above. In Section 3.3, the authors must show, through analyses, what the relative effect of an increased ITF in the mean MC hydroclimate is. After incorporating these new results, the authors will evaluate the utility of performing a cluster analysis (comment #7). In order to provide a comprehensive story of the MC climate, each analysis must be clearly justified.
- The authors have not performed any statistical significance analysis of the fields and processes evaluated. As such, it is not possible to know what the major changes simulated by most of the models are, and how these could be related to one another. Performing statistical significance analyses for each result is crucial before publication.
- In section 3.1, I suggest removing the cluster results for Figure 3. At this point of the manuscript the authors have not provided enough information of the cluster analysis and can confuse the readers. The analysis shown in Figure 4 is not very elucidative. Comparing discrepancies is very uncertain, especially for the mid-Pliocene where proxy-data show larger uncertainties. Is there any precipitation record in the MC that the authors could compare the PlioMIP results to? Borneo?
- In section 3.2, the authors analyse SST and P-E changes by averaging these variables over the study area. However, the study area encompasses the Indian and Pacific oceans as well as many artificial the occur due to a modified land-sea mask in the mid-Pliocene. As such, this is not a good metric to evaluate the MC climate.
- In section 3.3, the authors again try to address specific questions that are not possible to be addressed because either the study area is too restricted or beacuse of low resolution of the models. To address where the salt or heat anomalies originate from, it would be necessary to evaluate the large-scale heat and salt budget. It is not possible address whether the ITF anomalies originate from the surface or deep due to coarse resolution of the model and because of the land-sea masks, which will likely have different effects in each model due their different resolutions. I suggest the authors to focus on the possible role of the ITF transport on the MC hydroclimate. I am also very confused on what is being shown on Figure 8. The legend says ‘ocean current’ but the unit is Sv, which is a unit of transport.
- In my understanding a cluster analysis would be more appropriate if the PlioMIP models did not show a clear agreement on the changes for the MC climate, in which some models show very distinct results that could be masking important simulated features. However, without statistical significance analyses it is not possible to evaluate the usefulness of the cluster analysis. Also, the authors argue that the cluster analysis would reduce the influence of models of the same family in the MMM results, but it is not shown that models of the same family produce similar results. In fact, I found some results very different. SSTs from CESM2 are very different from all other CESM models. CCSM4-UoT is quite distant from CCSM4 and CESM1.2. NorESM models are several steps apart. Finally, the authors say that ‘the MCM can avoid signals being overweighted from the same family of models, but one could argue that the MCM could also vanish changes that are simulated by most models. For example, cluster 3 of SST includes nearly half of the models analysed.
- Discussion, conclusions and abstract must be rewritten on the light of the comments above.
Minor comments
- 34: ITF must have units of transport and not temperature.
L 104: Review grammar of Q2 or rephrase it.
L 107: Q4. It must be first demonstrated that there is a duplication of biases in the PlioMIP ensemble. Models’ results from the same family can change quite substantially.
- 135: The title of section 2.2 must be modified to not mislead the readers, once it is not performed any simulations specific to this study.
Figure 3: The location of the sites used to construct this figure must be plotted in figure 1, otherwise it is not conductive for a good reading.
Figure 4: Do you use the mean SST around the MC to plot this figure? If so, this is not appropriate for the proxy-data because there are only a few sites around the MC.
L 245: You cite fig. 6, but fig. 5 has not been cited yet.
L 266-268: What do you mean by ‘linearity is not exactly linear’? This sentence seems to be confusing what is shown on the plot.
L 284-285: this sentence needs a better theoretical explanation.
Figure 6 needs a statistical significance analysis with correlation coefficient and p-value.
L 295: I suspect the SOS decrease in the North Indian Ocean maybe related to a northward shift of the ITCZ, which is an important feature of the MC climate.
L 301-304: This is a too simplicity view of the drivers of the ITF. The ITF is an important feature of the large-scale ocean circulation and is not driven by density gradients between the Pacific and Indian oceans.
L 329: Results must be described with the assistance of statistical methods to quantify significance. Means and standard deviations? Medians and inter-quartile ranges?
L 351: Why it could be expected a reverse in the direction of the ITF?
L 388: Why is it necessary to remove the regional mean SSTa in figure 9b?
Figure 11: It is not possible to clearly see the colour corresponding to the proxy-data.
Figures: All figures need statistical significance analyses in order to be more confident of the results described in the text.
Citation: https://doi.org/10.5194/egusphere-2022-1281-RC2 - AC3: 'Reply on RC2', Xin Ren, 05 May 2023
Xin Ren et al.
Xin Ren et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 397 | 114 | 19 | 530 | 38 | 5 | 7 |
- HTML: 397
- PDF: 114
- XML: 19
- Total: 530
- Supplement: 38
- BibTeX: 5
- EndNote: 7
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1