Reviewer Comments

General Comments

The manuscript entitled “ Indian Ocean variability changes in the Palaeoclimate Model Intercomparison Project by Chris Brierley et al. explored the changes in the tropical Indian Ocean simulated by the Palaeoclimate Model Intercomparison Project (PMIP4). The manuscript is well written, and the findings from this study have implications for model development and point to the importance of understanding the internal variability of the climate system. Hence, I recommend this study for publication in EGUsphere. But before publishing, I would like to mention a few points observed in this study. I hope that adding those points can improve the manuscript.

Specific Comments

1. The introduction section of the manuscript gives an interesting description of the Indian Ocean Dipole and Indian Ocean Basin Mode. However, the authors didn’t explain the reason for choosing these two modes when there are others like Subtropical Indian Ocean Dipole (SIOD) (Behera et al.,2001 ), IOD Modoki (Endo et al. 2015), and Indian Ocean Tripole (IOT, Zhang et al., 2020) exists. Hence, it is wise to add a few more sentences to the first paragraph of the introduction. Explain why you have considered IOD and IOBM so that readers will find continuity in reading.

2. The study discusses the precipitation and temperature variability over the Indian and African regions. But in many figures, the variability of these parameters is noticed over the Arabian Peninsula regions, particularly during the Last Interglacial period. Hence, it would be worth adding the descriptions related to Arabian Peninsula to the manuscript.

Technical corrections

1. In the figure caption.5, Check line 4 and correct the spelling of simulation.

This manuscript presents simulated Indian Ocean variability in the latest PMIP4 multi-model ensembles. The periods investigated are two characterised by orbital forcing—midHolocene and lig127k, and two by substantial global cooling and warming—LGM and abrupt4xCO2. The responses of the mean climate and three interannual modes (the IOD, IOBM and Nino-like mode) compared to the those during the pre-Industrial were studied. The results show quite diverse responses across experiments and models. The robust changes are only present in limited scenarios for not all modes, for instance, weakening of the IOD in the midHolocene and strengthening of Nino-like mode in the lig127k. The relationship between the mean state changes and climate variability changes and between ENSO amplitude and that of the Indian climate modes were examined. The authors reported that the relationship in model simulations is not always in agreement with proxy data, such as IOD-zonal SST gradient relationship. 

The paper is nicely written and very informative. It clearly represents a large and substantial amount of effort and careful analysis. The results will be of value to the paleoclimate community by raising new questions. I have a few comments, mostly surrounding clarity of figures and presentation, that would improve the paper. 

1. My only major comment is for the definition of climate mode indices. All these modes are indicated by area-average SST anomalies of the Indian Ocean, often with overlapped regions (Sec 2.3). The Nino-like mode index is also derived in a similar way due to data availability. I am not sure if theses indices are significantly different from one another. Maybe this needs to be tested within uncertainty. A big concern is that this can partly contribute to their insignificant changes in different scenarios. Another potentially useful test is the mean change in SST and rainfall patterns of IOBM and Nino-like mode (similar to Fig. 5), and see the patterns are distinguished between modes.

2. L215. This statement may not be accurate. In LGM simulations, the presence of continental ice-sheets at higher latitudes can also impact the tropical climate through the atmospheric circulations, at least for the tropical Pacific (e.g. Lee et al. 2015; Lu et al. 2016).

3. The direct comparison of a transient state of warming of abrupt4xCO2 (year 101-150) and an equilibrium state of cooling of LGM can be misleading, e.g. due to delayed response in deeper ocean and thus the ocean stratification. This difference should be pointed out when discussing the results.

Fig. 3 Please give the meaning of black contour in the figure caption.

Fig. 4 and others Please describe how you calculate the vertical profiles, and the meaning of shorter dashed lines. For each dot, does its location on x-axis mean anything?

Fig. 7 There is no black solid line here, and it should be in Fig. 6.

refs

Lee, S. Y., Chiang, J. C., & Chang, P. (2015). Tropical Pacific response to continental ice sheet topography. Climate Dynamics, 44(9), 2429-2446.

Lu, Z., Liu, Z., & Zhu, J. (2016). Abrupt intensification of ENSO forced by deglacial ice-sheet retreat in CCSM3. Climate dynamics, 46(5), 1877-1891.

Review: Indian Ocean variability changes in the Palaeoclimate Model Intercomparison Project, by Brierley et al.

The manuscript discusses different modes of Indian Ocean variability in relation to different climatic forcing simulated within PIM4. I have few critical comments and some minor comments which are listed below. It is an interesting work and I think it should be published in Climate of the Past after the revision.

Main comments:

1. The first main comment I have is about the writing structure. The paper contains a lot of information and the reader can easily lose the track of what is changing what with the current structure. For instance if one does not know much about the IOBM and its changes, they should go back and forth in the paper to keep the track of the story. Moreover, the paper slowly deviates from the main focus (given in title/abstract) and conclusions scatter a bit throughout the paper. I suggest if the authors could re-structure the paper and text to ease the read a bit, for instance (just a suggestion):

i. start with the definition of the variability modes early in the paper → ii. then continue how these modes look in observations/reanalysis and discuss their impact in positive/negative phases if relevant → iii. comparison to pre-industrial simulations + discussion → iv. comparison to and difference/changes between the different forcing (both in mean state and variability) + discussion → v. potential impact of those changes in terms of temperature and precipitation (regression analysis) + discussion. Move section 4.3 for Nino-like mode to section 3.3; the connection to ENSO keeps coming in different parts of the paper, try to move all the ENSO-related story to section 4.2; move the definition of the modes in page 7 to page 2 or 3 where you mentioned them first.

2. The periods/frequencies of the modes are not mentioned. Next to the changes in modes’ amplitude, their periods (or number of occurrence per 100 years) and how they change in different climate, should be discussed.

3. This is a question more than a comment. For the variability in the Pacific and Atlantic Oceans, it is common to use the EOF modes to represent the modes of variability. I wonder how this will be for the Indian Ocean? I am not asking for new analysis, but it will be helpful if you could cite relevant studies which discuss the Indian Ocean variability using EOF (if any). The advantage of using EOF is that you can see if the patterns and variance of dominant modes and how they vary for different climatic forcing.

4. I think a plot (maybe for appendix) showing the standard deviation of temperature among the ensemble members for each experiment will be helpful but it is not necessary.

Other comments:

Title: needs to be revised (e.g. Indian Ocean variability in PMIP4 simulations)

Line 72: “...Indian Ocean variability.” of what? SST variability?

Line 80: “...capture the key processes...” like which processes?

Line 122: I am not sure if point c is correct. “that” should be than.

Line 200-213 and 230-247: hard to follow, I suggest to re-write (recalling main comment)

Line 268: “Having established…” I think such sentences would help in re-structuring (main comment).

Line 287-288: composite analysis versus regression: I am not sure if I agree with that sentence, and both techniques should have similar main features. You could also use difference-correlation in this part of the analysis.

Line 360: “positive relationship” you mean positive correlation?

Figure 2: color bar scale needs to be changed to show the anomalies better for 4xCO2