There has been some deliberation over this paper among the three editors handling the PMIP experiment description paper (one of whom has a self-declared conflict of interest, which was taken into account). Fundamentally none of us consider the paper satisfactory at present, but James Annan and I (who do not have a conflict of interest) have come up with a potential solution, which should allow this PMIP4 “bookmark” to appear in the tome of the CMIP6 special issue.

We think that the paper should be reformulated as a review or summary paper. One framing could be to think in terms of being of interest and use to people currently of outside of the PMIP mainstream, who would benefit from an introduction to PMIP and a rough guide to which experiments they might be able to do. They probably wouldn't be so willing to wade through 4 different detailed experimental protocol papers to see if they would work out which are relevant to them and what the important differences are. Such people exist! We have met them! I imagine that the resulting summary would also be useful to new PMIP researchers as well as those who so far have only been involved in one of the PMIP sub-MIPs, but are interested in getting a wider appreciation of the whole project.

In order to fulfil these requirements the paper needs a major overhaul. The manuscript is too long and does not need to cover so much detail, I don’t think it needs so much background, and many of the figures seem unnecessary, and they are anyway rather incomplete (eg 3 and 4 are two sets of partial forcings for a couple of experiments, 2 and 5 are bits of analysis that people have done for past PMIPs, but are by no means comprehensive). The tables are better, although they will need updating with the information from the now accepted/published sub-MIP papers. The text also needs to be made consistent with the other three papers, which we went to some lengths to coordinate, in terms of both the language and the actual requirements of the different MIPs.

The reviewer recommends rejection. If this manuscript were merely a review of the PMIP experiments as presented in the GMD/CP special issue, then I would reject, and encourage the authors to simply write a non-peer-reviewed preface to the special issue. However, this manuscript is also part of the CMIP6 special issue, and has value in that context, as an overview paper that summarises PMIP activities to non-PMIP modellers (and to PMIP modellers that are involved in only part of the project).

My first impression of this revision was that the manuscript is still much too long to serve its purpose as an enlightening overview of PMIP activities. However, other scientists are usually more patient than I am with overly long papers, so I decided to send the manuscript back out for review. I chose the reviewer quite carefully to be someone who I think is precisely the kind of scientist to whom the manuscript should be of interest (a paleo-curious climate modeller). It is therefore thoroughly disappointing that they find the manuscript "neither particularly useful nor exciting to read". The reviewer's general suggestions sound very attractive, "As an introduction to the other four papers, it could have placed the scientific questions first and spend more energy on explaining how the new experiments are exciting and can be analysed in ways that were not possible before, e.g. multi-state constraints, but this part (section 4) appears more to be an afterthought and not particularly concrete. This could have been done in a shorter format. ... In the worst case, however, the redundancy of protocol is confusing and can lead to errors in implementation. I therefore recommend stripping the manuscript of experimental protocol information, i.e. practically all of section 3 and parts of section 4."

A reduction in length of 50% would not be inappropriate in my (possibly still a bit extreme!) opinion.

I suspect that the other comments from the reviewer (about the out of sample-ness of the experiments and correctness of forcing) really refer to a slight imprecision in language rather than any fundamental disagreement, and so it is worth amending the text to make sure that such misunderstandings are minimised.

The delightful news is that, for the first time on reading a version of this manuscript, I see something that is an appropriate contribution to GMD. Well done!

I think that a review paper at GMD should read well. On that score, the manuscript needs improvement. It is littered with grammatical errors, and there are many cases where the word choice is at least weird and sometimes inappropriate, leading to what I perceive as errors in the manuscript. Lots of little bits of weird grammar add up to a manuscript that does not read at all well well. Experience tells me that a few improvements will make a huge and positive difference.

I know that some of your co-authors are much better than me at English, so I am wondering how much input they have had to the manuscript. I think it would be appropriate, therefore, to include an author contribution statement. Please can you also ask one person to go through the whole manuscript with the goal of improving the language?

In the notes below I include a non-exhaustive (I just carried on until I became exhausted!) selection of pointers to what I perceive as inaccuracies and inconsistencies, both in content and language.

Abstract

1.“The goal of the Paleoclimate Modelling Intercomparison Project (PMIP) is to understand the response of the climate system to different climate forcings and feedbacks for documented climatic states very different from the present and historical climates.”

I think this is incorrect. Feedbacks _are_ the response!

2.“ Climate models are usually developed using the present and historical climates as references, but climate projections show that future climates could lie well outside these conditions.”

“Could lie well outside”?? Surely with high confidence, future climate will lie well outside the historical period?

3.“therefore provide stringent tests for the state-of-the-art models and a way to assess whether their sensitivity to forcings and feedbacks is compatible”

As above, feedbacks are the response.

Introduction

4.“We have to look back ca. 3 million years to find a period of Earth’s history when atmospheric CO2 concentrations were similar to the present day (the mid-Pliocene Warm Period) and several tens of million years (e.g. the early Eocene, ~55 to 50 million years ago) to find concentrations similar to those projected for the end of this century.”

It took me about 5 attempts to parse the above sentence.

How about, “We have to look back three million years to find a period of Earth’s history when atmospheric CO2 concentrations were similar to the present day (the mid-Pliocene Warm Perios), and several tens of million years further to find concentrations similar to those projected for the end of this century (e.g. the early Eocene, ~55 to 50 million years ago).”

Note: the sense sentence is consistent to the change “2.” that I suggested for the abstract.

5. “the main external factors driving climatic changes are the astronomical parameters, which determine the seasonal and latitudinal distribution of incoming solar energy.”

Astronomical parameters cannot “drive” climate change. It is the variations in the pattern of incoming solar energy that drives the climate changes. I am guessing that “astronomical parameters” is the term for the parameters in models that are used to describe the shape of the orbit of the earth round the sun in climate models…? If so, then I suppose that changing these parameters does change _modelled_ climate. But I think that here you mean to talk about real climate rather than modelled climate.

2.1 PMIP4-CMIP6 entry cards: the mid-Holocene (midHolocene) and last glacial maximum (lgm)

6. “The LGM is particularly relevant because”

What is the LGM “relevant” to?

7. “However, the PMIP3-CMIP5 simulations of MH and LGM climates show only moderate skill in predicting reconstructed patterns of climate change overall (Hargreaves et al., 2013; Hargreaves and Annan, 2014; Harrison et al., 2014; Harrison et al., 2015).”

This needs rephrasing. Suggest referring to something less precisely defined than skill. I think that only the first two of these references discuss model skill, and that there was no skill at all for the mid-Holocene, but moderate skill for the LGM.

8. “This arises because of persistent problems in simulating regional climates.”

I am sure there us truth here, as evidenced in the differing patterns of change produced by climate models run for past and future scenarios. However, it is not the whole story. There is evidence that patterns in paleodata are also not well established. I am thinking here of the zero correlation between the earlier CLIMAP and later MARGO SST fields at the LGM.

9. “For example, state-of-the-art models cannot fully properly reproduce”

For grammar, only “fully” or “properly” not both. For climate science “adequately” might be better. All models are wrong so it will never be fully properly reproduced. The question is whether the reproduction is sufficient from the models to be useful.

10. “The protocol for the PMIP4-CMIP6 lgm simulations accounts for this uncertainty and includes a choice between..”

This is not very clear without reading the LGM paper.

How about, “The protocol for the PMIP4-CMIP6 lgm simulations accounts for this uncertainty by permitting modellers to choose between..” ?

11. “Altogether, the lgm experiments will allow testing the impact of these different ice sheet reconstructions (Figure 2) and of the dust forcing, which was not included in the PMIP3 set-up.”

I think this may be a reference to the sensitivity tests in the lgm paper? If so, say so, and mention which “tier” of experiments these are.

2.3 The LIG

12. “The PMIP4-CMIP6 lig127k experiment will help to determine the interplay of [things] on ice sheet thermodynamics 20 and dynamics.”

I have no idea how to determine an interplay. Please use words that make sense.

13. “ Vegetation reconstructions (Salzmann et al., 2008) indicate that the area of deserts was smaller than today and boreal forests were present in high northern latitude regions covered by tundra today.”

The sentence above only took about 3 attempts to understand.

Maybe, “Vegetation reconstructions indicate that the area of deserts was smaller than today, and boreal forests were present in high northern latitude regions that are today covered by tundra (Salzmann et al., 2008) .”

14. “The [] experiment (Haywood et al., 2016) is designed to understand the long- term response of the climate system..”

I am not sure that what you have written is wrong, but it suggests to me that you are expecting the experiment to understand something. The experiment is not sentient so it cannot understand. The scientists are the sentient ones.

14. “It [The experiment] will also be used to address the response of ocean circulation, Arctic sea-ice and modes of climate variability (e.g. El Niño Southern Oscillation)…”

address -> assess ?? I think we address questions rather than responses.

2.5 Paleoclimatic and paleoenvironmental data for the PMIP4-CMIP6 periods

15. “ The evaluation of climate simulations such as the PMIP4-
CMIP6 ones …”

Suggest, “The evaluation of climate simulations such as those conducted in PMIP4-CMIP6…”

16. “We expect new data sets to become available, which will increase the
number of possible model-data comparisons for the PMIP4-CMIP6 periods.”

I don’t like this. If these data sets are going to appear within the timeframe of the project they need to be close to completion now, and therefore will be known to you, and so you can here indicate what they are. Otherwise, delete.

17. “Much of the information on paleoclimates stems from the impact of climatic changes on the environment,…”

“Stems from” just doesn’t seem quite right. Suggest “arises from”…

18. “Some paleoclimatic indicators such as meteoric water isotopes have to be computed as the climate model is running, but are also examples of this forward modelling approach.”

Took several takes to understand.

Perhaps, “Paleoclimatic indicators such as meteoric water isotopes, that have to be computed as the climate model is running, are also examples of this forward modelling approach.”

19. “Modelling the impacts of past climate changes on the environment is key to understand how climatic signals..”

“is key to understand” -> “is key for understanding”…

3.1. Comparisons with paleoclimate and paleoenvironemental reconstructions, benchmarking and beyond

20. “Model-data comparisons for each period will be one of the first tasks conducted after completion of these simulations. One new feature common to all periods is that we will make full use of the fact that modelling groups must also run the piControl and historical experiments.”

“full use of”?! No idea what that might mean. Suggest deleting “full”.

21. “Indeed, existing paleoclimate reconstructions have used different modern reference states, and this has been shown to have an impact on the magnitude of reconstructed changes (e.g. Hessler et al., 2014).”

Confusing. Are you suddenly discussing observations here rather than models? I do not recall that models have had problems with using inappropriate control climates. Reconstruction is not unique to either realm and reconstruction are often a combination of models and observations. Why is it pertinent here to be discussing observations?

22. “Running both the piControl and historical simulations was not systematic in previous phases of PMIP, which prevented investigation of the impact of these reference states for model-data comparisons.”

Hasn’t the control has been a requirement since the start of PMIP? Maybe just limit this discussion to how you are going to use the historical runs.

23. “The new set of simulations will provide a way of quantifying this source of reconstruction uncertainty, as will comparisons with present-day observations and reanalysis data sets (Obs4MIPS, Ferraro et al, 2015).”

OK - even more confused now! Are we in the realms of model, observation, or both here?

In short, I don’t understand the message of the first paragraph of this section!

24. “Comparisons of PMIP4-CMIP6 simulations with available paleoclimatic reconstructions is a unique contribution to CMIP6 in terms of evaluating model biases for climates different from the historical one.”

Very ugly sentence. I am not sure how much I can improve it, however.

Here’s my attempt: “The paleoclimatic reconstructions from paleoclimate measurements are a unique resource within CMIP6 as they make it possible for model biases to be evaluated for climates that are very different to those over the historical period.”

25. “Benchmarking the ensemble of the PMIP4-CMIP6 simulations, for all the periods, will therefore allow quantifying the..”

->? “Benchmarking the ensemble of the PMIP4-CMIP6 simulations for all the periods will therefore allow quantification of the..”

26. “Another promising activity will consist in analysing the potential relationships between model biases in different regions and/or in different variables…”

eww.

Perhaps, “Another activity involves analysing relationships between model biases in different regions and/or in different variables…”

I was going to add the “promising” back in, but for that, I think you need a reference for such work in order to demonstrate the promise.

27. “develop more process-oriented metrics making use of the fact that paleoclimatic data document different aspects of..”

Add comma. “metrics, making”.

28. “e.g. cloud feedbacks”

How are you going to measure cloud feedbacks for paleoclimates?

3.2. Analysing the response of the climate system to multiple forcing

29. “..analyse the PMIP4-CMIP6 ensemble so as to separate the impacts of model resolution, content, or complexity on the simulated climate.”

What is “content”. Not a word I know in relation to model development. Complexity is generally used to imply which processes are included. If you mean something more subtle then you need to write precisely what you mean. Do you mean the algorithms? In climate modelling, these are usually called parameterisations.

30. “Many CMIP6 models will include new processes, such as dust, or improved representations of major radiative feedback processes, such as clouds.”

Clouds are fluffy objects in the sky. They are not processes.

31. “These analyses should provide new clues to constraint climate sensitivity.”

You can work this one out yourself!

32. “The issue of climate sensitivity (sensu stricto)”

There are almost as many definitions of climate sensitivity as there are researchers working on climate sensitivity. I do not know which one the sensu stricto is. I think maybe you can just remove the “(sensu stricto)” and everyone will be happy.

3.3 Interactions with other CMIP6 MIPs and the WCRP Grand Challenges

33. “CFMIP includes an idealized experiments”

grammar!!!

34. “It is mandatory that the model version used for the PMIP4-
CMIP6 experiments is exactly the same as for the other CMIP6 experiments, in particular the DECK and historical simulations.”

Rephrase. My understanding is that the only thing that is mandatory for all model versions is the DECK and hist. So you could easily get modellers running just the DECK, hist and paleo, and no other experiments with that model version.

35. “during each of the PMIP4-CMIP6 period”

grammar!!!

One substantive comment.

p12 “No additional interactive component should be included in the model unless it is already included in the DECK version. Such changes would affect the global energetics (Braconnot and Kageyama, 2015) and therefore prevent rigorous analyses integrating across multiple time periods or between MIPs (section 3).

Because of this, even though environmental records show that natural vegetation patterns during each of the PMIP4-CMIP6 periods were different from today, the PMIP4-CMIP6 palaeoclimate simulations should use the same model configuration to the DECK and historical simulations. If the DECK and historical simulations use dynamic vegetation, then the PMIP4-CMIP6 palaeoclimate simulations should do so too. If the DECK and
historical simulations use prescribed vegetation, then the same vegetation should be prescribed in the PMIP4- CMIP6 palaeoclimate simulations.”

Please rephrase this to be consistent with what is written in the other PMIP4 papers. To start with I suggest removing, “Such changes would affect the global energetics (Braconnot and Kageyama, 2015) and therefore prevent rigorous analyses integrating across multiple time periods or between MIPs (section 3). Because of this.”

A main focus of this manuscript is the promotion of PMIP as a series of experiments, the output from which may be compared with real data. If you knowingly implement your protocols to have very incorrect vegetation then these arguments are rendered much less convincing. It is not at all clear here why dynamic vegetation should be permitted but different fixed-vegetation not permitted, especially when many other boundary conditions are changed for the experiments. I went back to the LGM paper, and there it is simply stated that present day vegetation should be used without a reason being given. However, in that paper, modellers are also encouraged to run sensitivity experiments with altered vegetation, as well as the default protocol. This makes sense to me in the following way: the response to vegetation is so variable between models that we want all those who alter their paleo fixed-vegetation to first do the run with present day fixed-vegetation - otherwise we have no hope of understanding the response of the models. Not sure if that is your intended meaning, but rather than attempt to add further explanation, which is not anyway included in the LGM paper, I suggest that here you again encourage modellers to perform the vegetation sensitivity tests. (That these sensitivity tests may be not strictly CMIP6 runs is unimportant - this manuscript has to make sense scientifically!)


Apart from that, it will helpful to fix the following.

p3 l22 “feedbackfeedbacks”

p4 l23 “between climate-ice sheet system”
Missing “the”

l29-31
Can’t understand this sentence:

“ Finally, PMIP is also relevant to the third CMIP6 question “How can we assess future climate changes given climate variability, predictability and uncertainties in scenarios?” through examination of these questions for documented past climate states and via the use of the last millennium simulations as a reference state for natural variability.”

p6 l22 “Mediaeval Climate Anomaly”
Google Trends for the United Kingdom show no hits at all for this phrase, and far, far fewer hits for “Mediaeval” vs. “Medieval”. I think, therefore, it makes no sense to use your alternative spelling.
(The split for paleoclimate and palaeoclimate is much more even!)
(Don’t even think of trying “musaeum”.)

p7 l6-8 “The PMIP4-CMIP6 past1000 protocol will use a new, more comprehensive reconstruction of volcanic forcing (Sigl et al., 2015) and ensures a more continuous transition from the pre-industrial past to the future.”

“ensures” -> “ensure”
or
“will use” -> “uses”

…depending on which tense you want to use. First option sounds a bit better than the second, to me.

l8-9 “The final choices resulted from strong interactions with the groups producing the different forcing fields for the historical simulations (Jungclaus et al, 2017).”

“choices” here is ambiguous. Are you talking about choices you made (perhaps “decisions”?) to produce the forcings for a single protocol, or are you alluding a the range of possible forcings that people running the experiments may choose from?

p8 12-14 “However, as is the case for the Last Interglacial, the PlioMIP simulations were not always derived from the same models as for the CMIP5 future projections.”

perhaps,
“..the PlioMIP simulations were not always performed using the same models that were used for the CMIP5 experiments.”

Apart from awkward syntax, surely PlioMIP models were also different from those used in CMIP5-PMIP3 (e.g. the PMIP3 LGM).

l16-18
“The PMIP4-CMIP6 midPliocene-eoi400 experiment (Haywood et al., 2016) is designed to elucidate the long- term response of the climate system to a near modern concentration of atmospheric CO2 (long term climate sensitivity or Earth system sensitivity).”

I have seen “modern” used in the paleoclimate literature to have a rather vague meaning, sometimes closer to pre-industrial. Here I think you specifically mean close to 400ppm, so it might be clearer to say this…? After all, people may read this paper in 20 years time when CO2 is 500ppm... !

p9 l22 “also run historical experiment, in addition to the piControl one. “

Missing “the”