In their manuscript "The late Pliocene jet stream: Changes and drivers of the mean state and variability" Abigail E. C. Buchan and coauthors investigate differences in jet stream strength, position and variability between recent climate and a warm period within the Pliocene about 3 Million years ago that has been investigated as a potential analogue to future warm climates. A novelty is to perform the analysis based on a multi-model simulation ensemble from PlioMIP2 that is supported with detailed study of individual drivers based on a detailed factorization analysis of the HadCM3 model and of similar results based on two more models from the PlioMIP2 ensemble. The authors find that the majority of models support a weaker, more poleward jet with larger interannual variability, but also note substantial variability of simulated jet stream dynamics across the model ensemble. If considering the different drivers of Pliocene climate, the authors find strongest impact is from paleogeography, with carbon dioxide being a minor contributor, thereby limiting the value of the Pliocene as a direct analog to near-future changes in jet stream variability. Nevertheless, the authors highlight the value of the Pliocene as a laboratory to better understand the different contributors to jet stream strength and variability. They specifically highlight the value of this information in re-evaluating comparison of models and proxies in the context of different mean climate patterns in the Pliocene.

I have read the manuscript with great interest. I think that analysis and overall manuscript generally are already in good shape. I find the study generally suitable for publication in Climate of the Past. Nevertheless, I also think that there is still potential to improve the presentation and interpretation of study results and to increase clarity of text and figures to the readership. I suggest to return the manuscript to the authors for major revisions. Below I provide comments that I hope the authors will find helpful in revising the manuscript.

Overarching comments:

1. Relevance of the research and link to proxy data

From my personal point of view the discord between mid-Pliocene jet stream position being more northward in an environment with reduced meridional temperature gradient, and understanding based on modern observations suggesting an equatorward shift in the jet stream (line 55ff of the manuscript) is relevant. Could the authors shed more light on this problem based on the simulations that they analyse? Furthermore, the manuscript refers at many locations the link to proxy records, illustrating that combination of proxy records and models could improve our understanding of past extreme events. Nevertheless, the major take home message that I gain from these statements are problems with recording / reproducing extremes in both methods. In my opinion this proxy-model topic stays in general rather unclear. More examples would be helpful to better understand the relevant points that the authors would like to make here. I provide some example below.

2. Consideration of jet stream variability across two different regions (North Pacific and North Atlantic)

The authors separate the analysis across two geographic regions, the North Pacific and the North Atlantic region. In fact, when studying the figures, it is not always clear which region is shown in analyses. This should be clarified. Furthermore, at least to me it remained a bit unclear from the manuscript what is gained by studying both different regions, and to which extent the findings of each region have different relevance for our understanding of climate dynamics. The respective text in lines 170ff remains a bit vague with regard to this motivation. Since the authors focus a lot on AMOC, and AMOC having its major impact on the heat distribution in the North Atlantic region, I could imagine that the North Atlantic region's jet stream may be more relevant for various dynamics in Europe, for example. Nevertheless, I get the impression that more analyses are presented for the North Pacific region (although, as noted above, I might be wrong as it is not always clear from figure captions which analysis region is actually shown). Please consider during your revisions to make the motivation for focusing on each of these two regions clearer, also highlighting the specific relevance that findings derived for each region's jet stream have for our understanding of the (Pliocene) climate system.

3. Terminology

I am not sure whether readers outside the PlioMIP modelling community find it easy to link simulation names to specific time periods or to the employed forcing factorization. Consequently, I suggest to the authors to reconsider the chosen terminology along the following lines:

3.1 Definition of the main time period Late Pliocene: The authors chose the term Late Pliocene that comes with the update from PlioMIP2 to PlioMIP3 and define this in their lines 20-21. For clarity to the reader, here it would make sense to also refer to Haywood et al. (2024) regarding the renaming.

3.2: While the authors provide Table 2 as a service to the readers towards linking PlioMIP2 and PlioMIP3 terminology, my feeling is that non-experts of PlioMIP terminology could find comprehending section 2.1.1 a bit challenging; in particular, you describe results from PlioMIP2 simulations and refer to them with the newer PlioMIP3 term Late Pliocene. I think that the link to simulation Eoi400 (and its overlap with simulation LP) should be made more clear here. Towards improved clarity, I suggest to move text currently written in lines 115-117 to the beginning of the section.

3.3: I note that the link of boundary conditions to PlioMIP3 names is not always obvious from the PlioMIP3 terminology. For example, in simulation PI_lp-orog the term PI identifies PI icesheets, while in simulation PI_lp-ice it refers to PI orography (your Table 2). I think it may be helpful if you provided in your forcing factorization analysis clear indications as to the affiliation of all relevant boundary conditions to PI or LP background states. For example, in Figure 5 you could add the text "(except ice sheets)" to the heading of panel d. Comprehension of Figures 6, 10, and 11 could likewise profit from illustrating which parts of the boundary conditions are LP vs. PI by adding respective information in brackets after simulation names. This would help in linking results to the state of specific boundary conditions.

3.4: Terminology of the term "orography" in the context of boundary condition changes. In my personal experience the term orography is most often used in referring to elevations (bathymetry or land height surface). In PlioMIP terminology, on the other hand, the term refers more generally to combined changes of elevations, land sea mask, and other characteristics outside of ice sheet regions (maybe one could call this combination of changes called "orography" as modifications of geography outside ice sheet regions). Since in the context of this study the term orography is often used by you to implicitly refer to differences in land-sea mask that influence the expression of the AMOC (e.g. line 243 ff), I think it may make sense to explicitly spell out those characteristics of orogography that are of relevance in a specific context. In line 245 it would be particularly important, towards clarity, to replace the term "orography changes" by "gateway changes", and in line 266 by "land-sea mask changes". Also in lines 279ff it could be considered which occurrences of the term orography rather refer to land-sea mask changes specifically.

3.5: interannual variability vs. variability vs. differences. I note that I sometimes got lost in the descriptions of variability of jet stream vs. differences in the variability over time and between different models. Please carefully review text passages where you describe and discuss related findings and make sure that the meaning is always clear. Example: You speak of interannual variability heavily in the abstract and once in the outlook (line 334), but that term actually never appears in any other part of the main manuscript. What is the difference between interannual variability addressed in outlook and abstract vs. variability described in other sections? Another example, line 293-294: Here you speak of variability in the jet stream across the 8 simulations for these models. Do you rever here to interannual variability of the jet stream in each simulation, or to differences between jet stream expression between different simulations? Also see other locations.

3.6: The authors focus on boreal winter when analyzing the jet stream. Yet, it is not always clear from figure captions which seasons the results present (see specific comments below). Furthermore, sometimes the analysis uses DJF means, sometimes just velocities of January. Please provide reasoning for this choice in the revised manuscript.

3.7: the terms "mean" vs. "multi-model mean (MMM)" seem not to be used consistently. For example, line 200 ff. I think one appearance of "mean" rather refers to the MMM than to the mean state of the jet stream, which is also a topic of interest addressed in the study.

4. Abbreviations

Abbreviations are not always consistently defined and used. Late Pliocene is sometimes abbreviated, sometimes it is not. Please use a consistent strategy across the manuscript and supplement. In figures, please make sure that abbreviations in figure elements are always defined in the figure caption - this is relevant in particular also for the supplement, where selected figures may be studied more outside the context of manuscript-defined terminology.

5. Clarity of formulation, argumentation

Below I will highlight some text passages where the reasoning of the authors stayed unclear to me or where text elements seem to me disruptive to the overall tread of argumentation of the manuscript. Please reconsider your argumentation aims in such cases and clarify text accordingly.

6. selection of indices in the comparison of large-scale climate metrics to jet stream changes and variability

In line 226ff and Fig. S5 authors analyze the relationship between change in jet stream latitude vs. ESS/ECS. The authors do not find a strong relationship. To me the motivation to choose ESS/ECS as a metric did not quite become clear. Maybe the authors could clarify their reasoning in the text, and generally expand the related text section. Furthermore, ESS and ECS describe the response of >global< climate to changes in forcing and boundary conditions. Yet, in their motivation and also in their analysis, the authors highlight the relevance of the meridional temperature gradient for the expression of the jet stream. Taking this fact into account, wouldn't a comparison between jet stream and polar amplification of a model appear more promising when aiming to understand the link of jet stream changes to large-scale climate metrics of climate models? A stronger polar amplification will relate to a larger reduction of the meridional temperature gradient, and therefore one could also expect to find a clearer relationship between jet stream and polar amplification.

7. typography

Across the manuscript spacing between scalars and physical units should be observed in both text and figure elements (pressure levels are consistently referred to without spacing).

Specific comments:

line 11-13: "This is important as these differences might suggest a shift in the distributions of climate variables, such as temperature and precipitation, which could have implications for how proxy data and model simulations are compared." Towards increased clarity, please be more specific regarding the reference states that you compare here with the term "shift". Do you mean Pliocene vs. future or a different quality?

line 45-49: Here the reasoning is unclear to me. The authors refer to difficulties in deriving proxy records towards reconstructing extreme events. As a remedy, authors highlight the value of climate models towards increasing our understanding of drivers of extreme events and of their change to then motivate hypothesization on the impact of such events in the proxy record. Thereafter, the authors again highlight the value of studying a climate, where ample proxy data is available, towards understanding the skill of climate models in simulating extreme events. Yet, the authors just highlighted problems with reconstructing extreme events from the proxy record. How is this problem being overcome in the model-data comparison? I hypothesize that the authors would like to hint here at gaining a better understanding of changes in large-scale climate patterns and variability based on climate models towards subsequently enabling the study of the transfer of such changes into proxy archives via proxy system models. Yet, at least to me such an argumentation seems to remain unclear from the manuscript. Please clarify.

line 54: add the term "meridional" to "temperature gradients"

line 61/62: "linked to" used twice in close proximity

line 92: abbreviation LP undefined. It may be sensible to also here once more provide the link to the MPWP for readers not yet familiar with the PlioMIP3 terminology

line 92-96: this reference of proxy-based findings to jet stream variability seems a bit out of place. One could shorten this to the first two sentences of that paragraph. The remainder of the text could be better linked to the remainder of the manuscript, or maybe just moved to a later section where the focus is again on proxy-based evidence (currently this is in lines 208ff, but I note that also this text section seems a bit out of place in the results section and could be better placed in the discussion)

line 99: Do the authors mean to say here: "This will also provide a perspective on the usefulness of the Late Pliocene as a >past< analogue for >future< jet stream variability"?

line 100: LP?

line 102: Text of the introduction seems to end abruptly here. Also refer to the conclusions section, and maybe provide a final sentence as a transition to the methods section, maybe regarding the motivation of your study?

line 111: Should also Hudson Bay and Western Antarctic land sea mask changes be explicitly mentioned here due to their importance for AMOC dynamics (refer to Otto-Bliesner et al. (2017) wrt. Hudson Bay)

line 113-114: the argument that "CESM2 and CCSM4_Utrecht are not examined here due to a difference in coordinate systems" remains unclear. For clarity, please explain here what exactly differs so that those models' simulations are ignored in your study.

line 127: authors could be more explicit regarding their grading of a "good climate". Does mean model climate agree well with proxy records or with the overall PlioMIP2 simulation ensemble? Is location of a model well within multi-model ensemble metrics "good"? Please clarify text accordingly.

line 137-138: key here is the use of the LP land-sea mask in comparison to the originally proposed PI land sea mask, am I right? Please explicitly spell this out for clarity for readers not familiar with differences between enhanced and standard PlioMIP boundary conditions.

line 143: remove comma after "sheets"

line 143: For clarity, the sentence could be expanded as follows: "leads to a change in the orography and vegetation since PlioMIP boundary conditions do not discriminate between ice sheet extent and height."

line 146ff: you speak here of >new< HadCM3 simulations. Since HadCM3 was already part of PlioMIP2 (Haywood et al., 2020), it would be important to explicitly describe in which regard these simulations are comparable to the "classical" HadCM3 PlioMIP2 simulations to which they are compared. Were there any changes in HPC, model version, model parameterizations, or spinup procedure between the older and newer simulations that would be relevant to be highlighted in this section?

line 152: replace "variable" with "climate variable of interest"?

line 157-158: should reference to Hersbach et al. (2023) be moved to after "ERA5 reanalysis data" for clarity?

line 168-169: Please already name here the different pressure levels that you will be analyzing, and please also provide a motiviation why you use exactly these in contrast to other levels. In particular the link between, and conclusions drawn from, the analysis of jet stream dynamics in higher and lower pressure levels is in my opinion not always clear from your manuscript.

Line 170ff: You refer here to the method by Lie et al. (2015). Since that publication is behind a paywall it would be important to provide as many relevant methodological details as possible also in your manuscript. Outlining selection of pressure levels, and reflecting on the definition of the height to find the maximum of the jet stream, could be particularly relevant towards comprehension of your analysis method. Since you focus your analysis on three discrete pressure levels, can you be sure to fully capture changes in jet stream position and strength that may occur at intermediate pressure levels and that may be likewise relevant towards the understanding of the dynamics that your study aims at?

line 177: add comma after "December"

line 179: "but is not done here" - I suggest to provide a (reference to) the motivation why this is not done here.

line 180ff: you speak here about variability of the jet stream. Maybe more precisely speak of variability in the jet stream's wind speed maximum? There are also other aspects of variability wrt. the jet stream (position, height, width, etc.)

line 187: add closing bracket

line 190: weakening of, plus fix brackets

line 191-195: One could reflect on the difference in agreement of models and observations between North Pacific (better) and North Atlantic (worse)

line 198: "being faster" -> "wind speed being higher"

line 202: to better identify the models that you are singling out here, change text to "exhibiting a large equatorward shift >4°"? I also suggest to add a comma after "The mean"

line 208: please check formulation "in to the", add comma after "layer", and check the formulation "indicating that the zonal wind at each layer are linked" - for the latter, beyond grammar, the implications are not fully clear to me. Do the authors want to suggest here that changes in winds at each layer are linked to each other, so that proxy derived evidence could be employed to verify modelled jet stream changes across the atmosphere column? Please clarify.

line 208ff: This whole section appears to me more a discussion than a result, move it to the respective section? I think it would fit into outlook and further directions since the proxy based information is only touched by the authors, rather than fully explored (in my humble opinion).

line 219: Please provide reference for the statement that links spatial resolution to jet stream behavior.

line 224-225: please be more specific - are they better, could they be better?

line 226: "To examine possible", delete plural s of jets

line 242: Please check whether also Zhang et al. (2021) should be cited here

line 244: capitalize Northern Hemisphere

line 245: I think here it is particularly important to replace the term "orography changes" with "gateway changes";  Weiffenbach et al. (2023) do not mention the terms orography or elevation even once, bathymetry is used only once, but then in relation to the description of the HadGEM3 setup rather than to specify details in ocean gateway shape or other bathymetric features

line 253: extend the sentence to "are inherently linked per boundary condition design in PlioMIP"? If, instead of a combined ice sheet / elevation data set, ice sheet reconstruction was provided via both ice sheet extent and ice sheet thickness, then one could test the impact of albedo vs. elevation effects by means of a "white-blanket" experiment

line 257: looked at -> investigated?; for clarity, consider adding an "only" after "CCSM4-Utrecht"

line 260: "the the" -> "of the"

line 263: not sure whether the term "varied" is the best choice here since you talk of factorization with separated variation of boundary conditions; maybe replace "varied" by "diverse"? Furthermore, the sentence reads as if you talked of a variability of a variability - please check and clarify formulation where necessary. I think you may speak here of "The model-to-model discord is larger than the difference caused by the change from PI to LP boundary conditions", but I am not sure.

Line 264: consider reformulating the text "no conclusions can be drawn from this data set" (you actually draw conclusions directly thereafter)

line 265: when speaking of the change in the jet stream, do you mean here a change in jet stream variability?

line 266: also here orography refers more to land sea mask?

line 268: Formulation unclear. Do you mean to say here that "The variability is higher in LP than in PI in most models"?

line 271: change formulation to "considering jet stream dynamics"?

line 274-275: change formulation to "and higher spatial resolution, indicate that the jet stream is weaker and more variable in the LP"?

line 276: add "(7 of 11)" after "with the majority". This would, in my humble opinion, be a more precise statement that does not conceal that a quarter of the models actually show a different signal.

line 281: orography -> land sea mask?

line 284: fix typo "to c achange"; I do not unerstand the meaning of the formulation "The results of the forcing factorisation hold, independent of the 100 year period chosen to perform this analysis". Please clarify formulation accordingly.

line 286: I think a reference should be provided

line 289: I do not understand the formulations "split in sea ice" (do you mean difference in sensitivity to sea ice?) and "reduction in Arctic sea surface area" (do you refer here to ice covered Arctic Ocean area?)

line 290: feedbacks positively -> reinforces?

ine 291-292: Here an explict analysis of polar amplfication vs. jet stream metrics may be useful to support or refute this statement.

line 294-295: Indeed, COSMOS is an outlier. Yet, the formulation that you use could suggest that it was the only model for which the jet stream was stronger and less variable in the LP, while in reality there are two more models (so in total 3 of 11 models) showing a similar effect (at smaller amplitude), see Fig. 9. I suggest to clarify the formulation accordingly.

line 304: than in -> than?

line 304-306: I think here a reference to AABW from the AMOC plots (Fig. 6) my be relevant.

line 308: add comma after "set up"? Here I do not fully understand the reference to internal variability of climate models. Do you mean here time-variability (which would be the classical term that comes to my mind when speaking of internal variability), or are you aiming at structural uncertainty as in the PMIP triangle (Haywood et al. 2016), referring to uncertainties in forcings, boundary conditions, their implementation, and model initialization?

line 310: exhibits -> exhibited?

line 311:: "This" -> "The effect"?

line 312: change to "factors, potentially including model resolution" - if I did not misunderstand that you do not explicitly separate model resolution as a contributor, but you find evidence that suggests that it may play a role.

line 315: here you explicitly list vegetation as a cause for changes in the meridional temperature gradient; yet, if I understand correctly, its contribution is not explicitly separated from other boundary conditions in your analysis, so its quantitative contribution cannot be related to those by other drivers based on your work alone. Yet, there are previous studies that could be cited. O'ishi and Abe-Ouchi (2013) show that vegetation dynamics reinforce LGM cooling with a particular impact in the northern high latitudes, while O'ishi et al. (2021) and Arima and Yoshimori et al. (2025), and references therein, show a similar reinforcing effect of vegetation changes for warmer climates at orbital time-scales.

line 316-317: My understanding is that this rule could be extended - COSMOS has implemented a land sea mask change while nevertheless not showing a weaker jet in LP

line 318: Since you talk about other studies here, please clarify whether with Late Pliocene you refer here to LP as at other locations of the manuscript, or to MPWP, or to the Pliocene more generally.

line 329: change "the distribution of key climate variables" to "the key climate patterns"? Regarding the reference to the impact on the model comparison to proxy data, could one or two clear examples be provided for clarity?

line 333: add "climate" in front of variability, and maybe provide a few examples (NAO?). This is towards clarity that you speak here of internal modes of variability (ENSO, NAO, AMO, ...) rather than of other aspects of variability

Tables:

Table 1: For clarity one could adapt the last sentence of the caption to: "Models with an asterisk (*) have used for simulation LP a simplified paleogeography with a land-sea mask that is unchanged from modern conditions in the pre-industrial control simulation (E280)." (please check details of implementation in each model once more and adapt formulation accordingly if necessary). Also define abbreviations LP and PlioMIP2 here.

Table 2: Terminology is a bit unclear. Table 1 lists the various PlioMIP2 models, while you here define PlioMIP3 terminology as the standard. For a study that is largely focusing on PlioMIP2 simulations, and that appears outside the PlioMIP3 context, this may be a bit confusing to readers. For example, when reading the text first I was wondering whether the HadCM3 simulations presented by you already refer to the PlioMIP3 phase. Yet, this does not seem to be the case since you only focus on the overlap of simulations between PlioMIP2 and PlioMIP3. For clarity, I suggest to reformulate the Table header, e.g. as follows: "Boundary conditions from a selection of Pliocene Model Intercomparison (PlioMIP) simulations. Simulations have been prepared in the framework of PlioMIP2 (Haywood et al., 2016, 2020) but are overlapping with the PlioMIP3 simulation ensemble (Haywood et al., 2024). We provide both PlioMIP2 and PlioMIP3 nomenclature for ease of comparison to studies from both frameworks."

Figures:

Please check font sizes of all text elements in all figures. Some labels and annotations are difficult to read

Figure 1: fix spacing of scalars and units in panel subheadings, define any appreviations PI and LP in the caption

Figure 2: fix spacing of scalars and units in legend; on the right hand side of the x-axis you list Mean, but in the text you speak of Multi-Model-Mean (MMM)

Figure 3: spacing of hPa, definition of time slice abbreviations

Figure 4: spacing of hPa, definition of time slice abbreviations - consider to add "except ice sheets" to the subheader of subpanel d for clarity?

Figure 5: as per my comment to Fig. 4.

Figure 6: Since the AMOC effects are so much dependent on gateway changes, I think towards better understanding the results it would be helpful if the description of each subpanel would be accompanied by listing those settings that differ from modern (gateways, elevation outside ice sheet regions, ice sheets). Furthermore, it seems to me that simulation PI, PI_lp-orog, and LP_pi-ice are favorable to pronounced Antarctic Bottom Water circulation (i.e. the lower limb of the AMOC). These simulations feature modern ice sheets, and, maybe more importantly, modern land sea mask in the Southern Hemisphere, if I am not mistaken. This may be a point to reflect on in your manuscript (and if only as an outlook to the importance of geography in the Southern Hemisphere to AMOC and jet stream dynamics). Potentially reformulate the caption to "Maxima of each experiment annotated." (you use the term "value" twice in one sentence).

Figure 7: indicate region (North Pacific vs. North Atlantic) shown here in the caption; clarify formulation: "Latitude of maximum zonal wind speed and standard deviation of this latitude (Jet latitude variability)"?

Figure 8: spacing for hPa, definition of time period abbrevations; fix "for in"; clarify choice of January vs. DJF (this potentially in the main text); when trying to understand differences in variability of the jet stream I found this to be difficult based on the metrics provided in the figure. Would it be sensible to also quantify time-variability, as it is present in each panel, as a number, so that different levels of variability can be more easily compared with each other? Potentially you could provide standard deviation over time as an annotation above each subpanel (similarly to done wrt. the AMOC maximum in Fig. 6). Furthermore, but I am not sure due to difficulties with interpretation as outlined above: Is there a link between spatial resolution of a model and jet stream variability evident also from this figure? The models that seem to have the largest variability / most smeared out patterns in Fig. 8 (COSMOS, MRI-CGCM2.3, NorESM-L) also seem to be among the models with lowest spatial resolution (3.75x3.75, 2.8x2.8, 3.75x3.75) - exceptions being IPSLCM5A/2 and MIROC4m that likewise have low spatial resolution but seem to behave a bit different regarding changes in jet stream variability than the other three models.

Figure 9: Does this figure show results for North Pacific? Clarify caption. Fix typo in "jet stream"; consider adding a comma after "quadrants" and consider deleting the comma after "decreased variability".

Figure 10: spacing in hPa, 100 Januaries (?); also here consider providing a metric of variability that is more easily comparable between simulations (see my comment to Fig. 8). I am asking myself whether it would be sensible to further explore the change in variability due to CO2 in the Pliocene (lowermost two panels). You speak in section 3.3.1 about a small overall contribution of CO2, but it is not clear whether you refer here to a comparison between simulations with modern (uppermost panels) or LP (lowermost panels) geography. I get the impression that there is an impact, but a simple scalar metric like standard deviation (see comments above) could further clarify this.

Figure 11: capitalize Northern Hemisphere; add information identifying the region and the pressure level for which results are shown here.

Supplementary Material:

Please observe font size in figures (in particular S1, S2, S7, S8, S9, but also other Figures are in some cases at the limit of readability)

Please observe units (no latitude unit provided in S1 and S2 y-axes, no space used in pressure level skalars/units)

Please always make sure that information in figures is self-contained, i.e. that it is clearly explained what is shown and why specific models have been excluded from the analysis / computations:

Fig S1/S2 show three groups of curves without indicating the attribution of each group to a specific quantity. I assume curve groups show results for the three different pressure levels as in the main text. For clarity please indicate them at least in the caption, if not in a legend or as an annotation.

Make clear in captions which season/month is shown, and for clarity use same terminology as in the manuscript (winter->DJF in Fig. S1,2?; Fig. S3,4,6 do not provide any respective information)

Figure S3/4 show >changes< in zonal wind speed, could be indicated for clarity (e.g. "LP-PI zonal wind speed anomaly")

Figure S5 caption is incomplete, I assume it should read: "Change in jet stream latitue vs. ratio of Earth system sensitivity (ESS) and Equilibrium climate sensitivity (ECS). ...". Please also explicitly add a note regarding the reason for excluding HadGEM3 and MRI-CGCM2.3, since the motivation is not obvious from looking at the figure itself. Further, reference to CCSM4-UoT appears twice in the legend, and I assume this is by mistake.

Figure S9 caption is incomplete, should it read: "Relationship between January sea ice area and jet stream speed (left), and jet latitude variability (right)". Please explicitly state pressure level(s) for which data is shown here.

Make sure that abbreviations (LP, PI) used in figure elements are defined in the caption for self-consistency (S6, S7, S8). Also define such abbreviations if only used in captions (S1, S2, S3, S4).

Although you provide a lookup table regarding PlioMIP2/3 experiments in the main text, I would avoid mixing terminology (S9 uses old PlioMIP2 naming, while other Figures use PlioMIP3 naming)

When interpreting Fig. S5 in addition to Figs. 7 and 9 of the main text, it would be helpful if the same symbols were used for identical models.

Typos: the blue box groups experiments (S9).

References:

Haywood, A. M., Dowsett, H. J., Dolan, A. M., Rowley, D., Abe-Ouchi, A., Otto-Bliesner, B., Chandler, M. A., Hunter, S. J., Lunt, D. J., Pound, M., and Salzmann, U.: The Pliocene Model Intercomparison Project (PlioMIP) Phase 2: scientific objectives and experimental design, Clim. Past, 12, 663–675, https://doi.org/10.5194/cp-12-663-2016, 2016.

Haywood, A. M., Tindall, J. C., Dowsett, H. J., Dolan, A. M., Foley, K. M., Hunter, S. J., Hill, D. J., Chan, W.-L., Abe-Ouchi, A., Stepanek, C., Lohmann, G., Chandan, D., Peltier, W. R., Tan, N., Contoux, C., Ramstein, G., Li, X., Zhang, Z., Guo, C., Nisancioglu, K. H., Zhang, Q., Li, Q., Kamae, Y., Chandler, M. A., Sohl, L. E., Otto-Bliesner, B. L., Feng, R., Brady, E. C., von der Heydt, A. S., Baatsen, M. L. J., and Lunt, D. J.: The Pliocene Model Intercomparison Project Phase 2: large-scale climate features and climate sensitivity, Clim. Past, 16, 2095–2123, https://doi.org/10.5194/cp-16-2095-2020, 2020.

Haywood, A. M., Tindall, J. C., Burton, L. E., Chandler, M. A., Dolan, A. M., Dowsett, H. J., Feng, R., Fletcher, T. L., Foley, K. M., Hill, D. J., Hunter, S. J., Otto-Bliesner, B. L., Lunt, D. J., Robinson, M. M., and Salzmann, U.: Pliocene Model Intercomparison Project Phase 3 (PlioMIP3) – Science Plan and Experimental Design, Global Planet. Change, https://doi.org/10.1016/j.gloplacha.2023.104316, in press, 2024.

Li, X., Jiang, D., Zhang, Z. et al. Mid-Pliocene westerlies from PlioMIP simulations. Adv. Atmos. Sci. 32, 909–923 (2015). https://doi.org/10.1007/s00376-014-4171-7

Arima, N., Yoshimori, M., Abe-Ouchi, A., O’ishi, R., Chan, W.-L., Sherriff-Tadano, S., and Ogura, T.: Impact of the temperature-cloud phase relationship on the simulated Arctic warming during the last interglacial, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2025-4109, 2025.

O'ishi, R. and Abe-Ouchi, A.: Influence of dynamic vegetation on climate change and terrestrial carbon storage in the Last Glacial Maximum, Clim. Past, 9, 1571–1587, https://doi.org/10.5194/cp-9-1571-2013, 2013.

O'ishi, R., Chan, W.-L., Abe-Ouchi, A., Sherriff-Tadano, S., Ohgaito, R., and Yoshimori, M.: PMIP4/CMIP6 last interglacial simulations using three different versions of MIROC: importance of vegetation, Clim. Past, 17, 21–36, https://doi.org/10.5194/cp-17-21-2021, 2021.

Otto-Bliesner, B. L., A. Jahn, R. Feng, E. C. Brady, A. Hu, and M. Löfverström (2017), Amplified North Atlantic warming in the late Pliocene by changes in Arctic gateways, Geophys. Res. Lett., 44, 957–964, doi:10.1002/2016GL071805.

Weiffenbach, J. E., Baatsen, M. L. J., Dijkstra, H. A., von der Heydt, A. S., Abe-Ouchi, A., Brady, E. C., Chan, W.-L., Chandan, D., Chandler, M. A., Contoux, C., Feng, R., Guo, C., Han, Z., Haywood, A. M., Li, Q., Li, X., Lohmann, G., Lunt, D. J., Nisancioglu, K. H., Otto-Bliesner, B. L., Peltier, W. R., Ramstein, G., Sohl, L. E., Stepanek, C., Tan, N., Tindall, J. C., Williams, C. J. R., Zhang, Q., and Zhang, Z.: Unraveling the mechanisms and implications of a stronger mid-Pliocene Atlantic Meridional Overturning Circulation (AMOC) in PlioMIP2, Clim. Past, 19, 61–85, https://doi.org/10.5194/cp-19-61-2023, 2023.

Zhang, Z., Li, X., Guo, C., Otterå, O. H., Nisancioglu, K. H., Tan, N., Contoux, C., Ramstein, G., Feng, R., Otto-Bliesner, B. L., Brady, E., Chandan, D., Peltier, W. R., Baatsen, M. L. J., von der Heydt, A. S., Weiffenbach, J. E., Stepanek, C., Lohmann, G., Zhang, Q., Li, Q., Chandler, M. A., Sohl, L. E., Haywood, A. M., Hunter, S. J., Tindall, J. C., Williams, C., Lunt, D. J., Chan, W.-L., and Abe-Ouchi, A.: Mid-Pliocene Atlantic Meridional Overturning Circulation simulated in PlioMIP2, Climate of the Past, 17, 529–543, https://doi.org/10.5194/cp-17-529-2021, 2021.