Xudong Xu and colleagues use clay mineralogical assemblages (mainly smectite) to demonstrate millennial-scale shifts in the ITCZ as documented in a core site from the Ninetyeast Ridge in the northeastern Indian Ocean. According to their interpretation, smectite maxima during the Last Glacial Maximum (LGM) point to regional shifts in the ITCZ. They further argue that glacial-to-interglacial times scales, smectite did not reach the core site during glacials because islands were exposed (due to low sea level) and because chemical weathering (producing smectite in warm times) was diminished. 

The topic of the paper appears suitable for CP. I consider both the overall scientific significance and quality to be good to fair and hence recommend minor to major revisions. A better understanding of latitudinal ITCZ migrations, both on orbital and millennial time scales is indeed desirable, and existing knowledge is either sparse or conflicting. The authors have worked extensively in the general area and topic before and provide sufficient evidence of a sound understanding on regional paleoceanographic change associated with shifts in the ITCZ. To base their whole story around a single, short core (171106) and mainly one property (smectite) with the additional, occasional provenance data (Sr-Nd) is a little thin in my eyes. The chronology is sound and includes age modeling and uncertainty evaluation. The same appears to be the case for the clay mineralogy and provenance tracers.

The whole discussion on smectite suffers, in my opinion, from the fact that the record is too short (MIS 3-1) and does not cover a full glacial-to-interglacial cycle. One cannot see how the response was, for instance, during MIS 6 and 5, which would be critical to know here. Any discussion of orbital variability is hence hampered. This is specifically true for any statement implicating changes in sea-level and and their effect on changing provenance and clay mineralogy (availability or lack of accommodation space on the shelves). The core is apparently only 162 cm long - why? Wasn’t there a longer alternative to conduct such a study?

The same principal problem surrounds the discussion of the smectite peak around the LGM (21-18 ka). If there were a record for termination 2 (MIS 6/5 transition) one could see if there are common rules established during glacial maxima that are either regional or not. MIS 6 and termination 2 have comparable orbital configurations relative to the LGM and termination 1, hence possible shifts of the ITCZ – which are invoked in the discussion – should have been quite similar, at least from a global view, from which regional deviations could then be derived or discussed if present. 

There is the complete lack of mentioning turbidity currents, which provide the vast majority of sediments from the river mouth of the Swatch of no Ground (SONG) to the BoB and neighboring regions. Even though the short core presented here seems to show a rather steady deposition, plumes from turbidity currents will reach the site and affect the clay mineralogy. In this context, it is also odd that none of the IDOP-Expedition 354 studies, which are in close proximity to the core cite, have been cited ore discussed. 

I also miss the discussion on the variability of the Oxygen Minima Zona (OMZ) when the authors invoke the connection to the Northern Hemisphere. What would be the consequences for the area of the core site? For the Arabian Sea, Schulz et al., (1998, Nature) clearly linked the millennial-scale coupling to Greenland to shifts in extent of the OMZ, pointing to changing water mass composition and oxygenation. 

Another lack of discussion surrounds the length of the core, which is only 162 cm. The sample resolution of 1 cm implies a 300-yr resolution, however, bioturbation should mix the sediment over several cms and smear according ages. What are the author’s assessment of this effect and how would it affect their conclusions? 

More specific comments:

Fig. 3

Sr/Nd resolution is too low to determine temporal variations. The only real change happened at 14-15 ka, probably as a result to changes surrounding Meltwater Pulse 1A, which are indicative of a major re-organization of the global thermohaline circulation. However, the data shows that the glacial-to-interglacial and millennial-scale sources likely did not change. Why is that?

Also, the authors mark H1-H4 and discuss the relation to smectite variability. The data does not show that in my opinion. Even if slight shifts are employed to account for potential mismatches in the age model, there is no consistent relationship, i.e. the various clay minerals occur either at the highs, lows or transitions of H events. 

Line 147-149

This is important and needs clarification and elaboration. How far above the surrounding is the core site? As mentioned above the authors completely ignore the possibility of turbidite deposition. Even if only the fine grained upper sediment clouds of distal turbidites (those naturally contain a high percentage of suspended fine material, i.e., clay minerals)  reach the sediment site, it would have a large effect on the clay mineralogy of site 171106. After all, the sites in not far from the channels on the eastern side of the BoB. 

In this context, Fig. 1 is missing a depth legend for both the marine and terrestrial elevations. Figures 1 and 5 could potentially be combined into a single figure.

General comments:

Xu et al. present a very interesting and useful sediment record from Ninetyeast Ridge with measurements of clay mineralogy and Sr-Nd isotopes. Although the time frame of the record is somewhat limited, it does provide some constraints on orbital variation of terrigenous sediment fluxes. The data are of good quality and the interpretations are reasonable. The paper is well within the scope of Climate of the Past and provides another helpful record of the South Asian monsoon, which drivers are important geological questions. Broadly, I think the paper would be improved by clarifying the uncertainties in the interpretation of the proxies, more information about methods, and interpretation of the driving factors (ITCZ versus other influences on precipitation intensity, weathering and sediment transport).

Specific Comments:

Please add more perspective on disentangling the transport limitations (trapping of sediments on the shelf) and weathering (ITCZ migration and rainfall intensity).

How might the transport conditions mask the ITCZ record (e.g. lower smectite transport is interpreted as lower rainfall). It’s great that both factors are mentioned, but I think the text needs to be clarified as to how these factors may result in different interpretations of the same record. This would give more confidence in the rainfall/ITCZ related interpretations.

I think it is important to discuss the potential factors on rainfall and weathering other than north-south ITCZ migration. For instance, Gebregiorgis et al. 2016 QSR suggest that drivers off the South Asian monsoon may be more complex than migration of the ITCZ.  There has been recent work regarding exposure of the Sunda/Sahul Shelves on Indian Ocean hydroclimate (DiNezio and Tierney, 2013 Nat. Geosci; DiNezio et al., 2020 Sci. Adv.; Pico et al., 2020 Paleoc. Paleoclim.), cross-equatorial moisture transport and influence on the monsoon (Clemens et al., 2021, Sci Adv.), the effect of stratification on the monsoon (Tierney et al. 2015, Nat. Geosci), and even a potential El Nino-like mode in the Indian Ocean during the LGM (Thirumalai et al., 2019 Paleoclim. Paleoc.). These are all factors that can influence precipitation that are not necessarily related to just the position of the ITCZ. There is mention of some potential other factors on lines 235-242, but this is not clear and not discussed directly related to your record. It would be good to consider your smectite record in regard to the timing of these other processes; a clearer discussion of these factors would improve the relevance of what sediment flux records like this are actually recording the Indian Ocean.

Technical corrections:

Line 16: “millennial” instead of “millennium”

Line 18: Maybe “supply” or “transport” is better than “supplementation”

Line 20: exposure of Sunda Shelf?

Line 21 “millennial” instead of “millennium”

Line 23 “millennial” instead of “millennium”

Line 30: It would read better to replace “the paleoclimate and paleo-ocean” with “paleoclimate and paleoceanographic conditions”

Line 38: “important” might be a better word than “nonnegligible”

Line 53: I don’t think statements like “discussed as a hot topic” add much to the context. Better to state which studies have discussed provenance in the BoB and the collective contribution of these studies.

Line 65-66: Please explain more about why Ninetyeast Ridge is an ideal location- appropriate sed rates for a gravity core? And only receiving fine-grained hemipelagic sediments because it is bathymetrically above fan sedimentation?

Line 83: Centrifuged? What rpm and duration?

Line 84: hydrochloric acid

Line 86: Were the slides treated with ethylene glycol? Were clay mineral standards used? Or just the Biscaye method?

Line 92: Are these bulk sediments or a specific grain size fraction?

Line 96: How were the Sr and Nd isolated? Info on columns, etc.

Line 137: “cannot” is a strong word, but yes, kaolinite transport can be limited, but some regions of open ocean have substantial aeolian kaolinite

Line 148: I suggest “abyssal plain” instead of “the normal seafloor”

Line 149: turbidity currents

Line 169-170: Not just the narrowing of the straight but exposure of the continental shelves including the Sunda Shelf and all the way up to Myanmar. The relative exposure of 200 km from the current Irrawaddy delta can affect how sediments are trapped on the shelf or delivered to the deep ocean.

Line 182: “transport” is a better word choice than “importation”

Line 229: Please highlight any information about how far northward the ITCZ may have shifted before and after the LGM. Would this bring noticeable change to the Indo-Burma area?

Line 245: Himalayas, mention the G-B transport.

Figure 5. Label the name of your core on the map.