| 12 Jun 2026
Middle to Late Miocene paleoceanographic evolution of the south-eastern Indian Ocean (ODP Site 752) inferred from nannofossil assemblages
Abstract. Understanding how the Earth’s system behaves under climate forcing conditions is critical for predicting how future climate change scenarios may affect the planet. In this regard, the Miocene, with an atmospheric temperature ~4 °C higher than modern and a near-modern ocean configuration, stands out as a potential analogue for future climate projections. Our current understanding of the Miocene derives principally from the Atlantic and Pacific Oceans, whereas the Indian Ocean remains understudied. Nevertheless, different studies performed in the equatorial Indian Ocean have revealed that the multiple climate shifts that occurred during the Miocene modified both the atmospheric and oceanic processes in this basin (e.g., changes in westerlies position, nutrient flux and productivity). In the present study, we investigate how changes in the climate conditions affected the surface ocean system at Ocean Drilling Program (ODP) Site 752, located in the southeastern Indian Ocean, across the Middle to Late Miocene. For this purpose, we present a new quantitative dataset of calcareous nannofossil assemblages, in conjunction with already existing multi-proxy data.
Our results indicate that the warm atmospheric conditions reached during the Miocene Climatic Optimum (MCO), starting at ~16.9 Ma, caused an intensification of the seasonal signal at ODP Site 752 that lasted until 15.32 Ma. After this period, we detected a shift in the calcareous nannofossil assemblage towards species characteristic of low-nutrient and warm-water conditions (e.g., Reticulofenestra pseudoumbilicus, Reticulofenestra haqii). Nevertheless, with the progression towards the Late Miocene Cooling, which started at ~7 Ma, and the increased influence of the cooler and more productive Southern Ocean waters, the calcareous nannofossil assemblage evolved into a community characterised by relatively high-nutrient and cold-water taxa (e.g., Coccolithus pelagicus, Calcidiscus leptoporus). By 10.67 Ma, we recorded a reversion in the nannoplankton assemblage to species typical of low-nutrient and warm-water conditions that lasted until 9.88 Ma. The comparison with available eNd(t) records for Site 752 and 707 (located in the equatorial Indian Ocean) revealed that this change in the assemblage composition responded to a strengthening of the Pacific Ocean water influx occurring at this time. High-nutrient and cold-water conditions were reestablished after 9.88 Ma, lasting until the end of the studied record at ~7.34 Ma. Our evaluation of the primary producer community from a multi-proxy-based perspective revealed a progressive evolution of the surface ocean conditions in the southeastern Indian Ocean region across the Middle to Late Miocene for the first time. Furthermore, the strong correlation with existing proxy data for global climate records suggests that large-scale climate events are the main factors driving the changes in the primary producer community and surface ocean conditions at the Broken Ridge during the Middle to Late Miocene. However, certain regional processes (e.g., strengthening of the Pacific Ocean influx) may also occasionally influence surface ocean conditions. Moreover, this multi-proxy approach allowed us to better understand the Indian Ocean–Southern Ocean interaction, and more specifically, the nutrient transport efficiency between these two ocean basins during a climate change scenario. Specifically, our multi-proxy-based dataset revealed that the Antarctic divergence region experienced a northward migration after the MCO, leading to a deep reorganisation of nutrient transport across the surface southeastern Indian Ocean.
This manuscript presents a new quantitative calcareous nannofossil record from ODP Site 752 spanning the middle to late Miocene and integrates these data with previously published geochemical proxies to reconstruct the evolution of the southeastern Indian Ocean. The dataset itself is valuable because this region remains considerably less studied than the Atlantic and Pacific Oceans, and the new assemblage record represents a useful addition to the Miocene palaeoceanographic literature. The manuscript is generally well written, the figures are of high quality, and the stratigraphic dataset will likely become an important reference for future work.
However, I have substantial concerns regarding the interpretation of the ecological data. Throughout the manuscript, the authors repeatedly infer specific environmental conditions (nitrogen limitation, phosphorus excess, thermocline depth, nutrient availability, seasonality and frontal migration) from changes in calcareous nannofossil assemblages with a degree of certainty that is not supported by either the statistical analyses or the current understanding of Miocene nannoplankton ecology. In many cases, the discussion extends considerably beyond what can be demonstrated from the available evidence. Overall, I believe the dataset is worthy of publication, but the manuscript requires substantial revision before the conclusions can be considered robust. I therefore recommend major revision.
Major comments
1. The ecological and nutrient-regime interpretations are more specific than the available evidence supports (Lines 221–237; 326–528)
The discussion repeatedly interprets the identified assemblages as evidence for relatively specific environmental conditions, including changes in nutrient availability, nitrogen limitation, phosphorus excess, thermocline depth, mixed-layer dynamics, and seasonality. While many of these interpretations are plausible, they are not necessarily uniquely constrained by the assemblage data.
Calcareous nannoplankton ecology is multivariate, and assemblage composition may reflect the combined influence of temperature, nutrient availability, water-column stratification, light availability, carbonate chemistry, evolutionary change, and preservation. I therefore encourage the authors to distinguish more clearly between the observed assemblage changes and the environmental mechanisms proposed to explain them.
For example, R. minuta is interpreted as reflecting nitrogen-limited conditions and, through the proposed relationship between nitrogen limitation and phosphorus excess, TG1 is ultimately characterised as representing “high-nutrient (N-limited) cold waters” (Lines 328–346). This interpretation appears more specific than the available evidence permits. The cited literature provides a plausible mechanistic framework, but the Site 752 data presented here do not directly reconstruct nitrogen or phosphorus availability or their relative balance. Moreover, the description of these waters as simultaneously “high-nutrient” and “N-limited” requires clearer explanation of which nutrient pool is inferred to be elevated and which is limiting.
Similarly, the ecological significance of Reticulofenestra size variation remains complex. The manuscript itself acknowledges contrasting interpretations of large reticulofenestrids, which have been associated with both oligotrophic/deep-nutricline conditions and enhanced nutrient availability (Lines 355–363). The authors should therefore explain more clearly why particular ecological interpretations are preferred for Site 752 and acknowledge where alternative explanations remain possible.
The comparison with independent geochemical proxy records is a strength of the study and can provide support for broader changes in surface-ocean conditions. However, the authors should distinguish between nutrient availability, primary productivity, and export productivity, which are related but not equivalent. Agreement with a productivity or export-productivity proxy may support a broader environmental change but does not necessarily independently validate a specific mechanism such as nitrogen limitation, phosphorus excess, or a particular N:P balance. I therefore suggest moderating the most mechanistically specific interpretations unless they can be supported by additional independent evidence.
2. Definition and interpretation of the “taphogroups” (Lines 198–220; 279–325)
The manuscript uses hierarchical clustering to identify five statistically defined assemblage groups, which are subsequently termed “taphogroups” and assigned specific environmental interpretations. I suggest that the authors clarify their use of the term taphogroup. The groups are defined primarily from variations in taxonomic relative abundance rather than from taphonomic characteristics, yet preservation at Site 752 is described as moderate to moderately poor (Lines 240–242). It is therefore important to clarify whether “taphogroup” is intended to refer to (i) assemblages shaped primarily by taphonomic processes, (ii) fossil assemblages used as palaeoecological groupings while acknowledging taphonomic filtering, or (iii) simply statistically defined assemblage clusters. If the latter, “assemblage group” or “assemblage cluster” may be clearer. If the authors retain “taphogroup”, the term should be clearly defined and its use justified.
I also encourage the authors to be more cautious in treating these groups as discrete ecological states. The reported cophenetic correlation (0.76) indicates a reasonable but not exceptionally strong representation of the original dissimilarities by the dendrogram, while the authors acknowledge weak bootstrap support for individual clusters and relatively high nMDS stress values (0.2565 for two dimensions and 0.2018 for three dimensions). Importantly, these different statistics address different aspects of the analysis: the nMDS stress does not directly test the validity of the hierarchical clusters, and ANOSIM significance indicates differences among the predefined groups but does not, by itself, demonstrate that they represent naturally discrete ecological states. I therefore suggest that the authors better distinguish between statistically defined assemblage groupings and the subsequent environmental interpretation of those groups.
3. Potential influence of preservation on assemblage composition and morphotype identification (Lines 240–242)
The authors state that calcareous nannofossil preservation ranges from moderate to moderately poor, based on observations using light and scanning electron microscopy. Given that the principal interpretations of this study rely on changes in relative abundance, diversity, and Reticulofenestra morphotype distributions, I think the potential influence of preservation warrants more consideration.
Preservation does not necessarily affect all taxa or morphotypes equally, and dissolution, fragmentation, and overgrowth can alter relative assemblage composition. This issue may be particularly relevant to the Reticulofenestra classification used here, because several morphotypes are distinguished not only by size but also by whether the central area is interpreted as open or closed (Lines 183–197). The authors should therefore discuss whether preservation could affect the identification or relative abundance of these morphotypes and whether any systematic changes in preservation coincide with the major assemblage transitions identified in the study.
I am not necessarily suggesting that preservation drives the observed patterns, but the manuscript should provide sufficient evidence to demonstrate that the major assemblage and morphotype changes are primarily ecological rather than taphonomic. If preservation was assessed systematically, the authors should describe the criteria used and, if possible, show whether preservation varies through the studied interval.
4. The definition and robustness of the assemblage clusters require further clarification (Lines 198–220; 279–325)
The overall multivariate approach is broadly appropriate for ecological assemblage data, but several analytical decisions require clearer justification. In particular, the authors should explain the rationale for excluding taxa with an average abundance below 0.5% and discuss whether this filtering materially affects the resulting assemblage structure. More importantly, the basis for selecting the similarity cut-off used to define the five clusters (and the subsequent subdivision of cluster 5) should be explained. At present, it is unclear whether these thresholds were determined using an objective criterion or selected primarily to obtain ecologically interpretable groups.
This is particularly relevant because the authors report weak bootstrap support for individual clusters. I therefore encourage the authors to provide a clearer assessment of cluster robustness. This does not necessarily require repeating the entire analysis using multiple clustering methods, but the authors should demonstrate that the principal assemblage groupings are not highly sensitive to the chosen cut-off or explain the limitations associated with the weaker-supported groups. The methodological rationale for the final number of assemblage groups should be made transparent.
5. Chronological uncertainty should be considered when discussing the timing and apparent synchronicity of events (Lines 159–168 and throughout Section 4.2)
The study uses the astronomically tuned age model of Lyu et al. (2023), with additional biostratigraphic support from Puentes-Jorge et al. (2025). Given that several interpretations in Section 4.2 rely on the apparent temporal correspondence between changes in the Site 752 nannofossil assemblage and climatic or oceanographic events identified in other records, I suggest that the authors briefly discuss the chronological uncertainty associated with these comparisons.
This is particularly relevant where relatively precise assemblage transitions (e.g. 10.67, 9.88 and 9.22 Ma) are interpreted as coincident with changes in independently dated records. The authors do not necessarily need to provide a new quantitative age-model uncertainty analysis, but they should clarify which proxy records share the Site 752 chronology and which are based on independent age models. Where records are independently dated, apparent synchronicity should be interpreted with appropriate caution unless the respective chronological uncertainties are sufficiently small to resolve the proposed temporal relationship.
6. Spatial scale and certainty of some palaeoceanographic interpretations (Lines 47–52; 404–528)
The manuscript uses the Site 752 nannofossil assemblage together with previously published regional proxy records to develop a broader reconstruction of Southern Ocean–Indian Ocean interactions. This is a potentially valuable synthesis. However, some conclusions appear more spatially and mechanistically specific than can be directly constrained by the available evidence. In particular, the statement in the Abstract that the dataset “revealed that the Antarctic divergence region experienced a northward migration after the MCO” (Lines 50–52) appears stronger than the evidence presented. The conceptual model itself appropriately describes the migration of the polar front system as hypothesised (Lines 522–527), and I suggest maintaining a similar level of caution throughout the manuscript.
A change in assemblage composition at Site 752 that is consistent with increased influence of cooler and/or nutrient-rich southern-sourced waters does not, by itself, uniquely demonstrate a northward migration of the Antarctic Divergence. Similar changes at the site could potentially arise from changes in the strength of water-mass advection, frontal intensity, mixing, nutrient supply, or circulation pathways without requiring a simple latitudinal displacement of the frontal system. The authors should therefore distinguish more clearly between what is directly recorded at Site 752 and the broader circulation mechanism proposed to explain that signal. Where the latter is not independently constrained by spatially distributed records, it should be presented as a hypothesis or conceptual interpretation.
Minor comments
Lines 101-114
The introduction would benefit from explaining more clearly the specific scientific significance of Site 752. The authors establish that the Indian Ocean is comparatively understudied, but the rationale for studying this particular site could be developed further. What does the location of Site 752 at Broken Ridge allow the authors to address that cannot be resolved from existing Indian Ocean records, or from better-studied records in the Atlantic and Pacific? Clarifying the site's particular sensitivity to Southern Ocean–Indian Ocean exchange and/or regional circulation changes would provide a stronger motivation for the study and better define its broader palaeoclimatic significance.
Lines 170–177
Please clarify the counting protocol. The authors state that additional transects were counted when the initial count was below 400 specimens, whereas the average count was approximately 600 specimens per sample. Was a minimum count of 400 specimens therefore used as the stopping criterion, with higher counts resulting from completion of the final transect? Clarifying this would make the quantitative methodology more reproducible.
Section 3.3 and corresponding figures
I miss a clear explanation of the point at which the statistically defined clusters are subsequently termed “taphogroups” (TG1–TG5). Please define what is meant by a taphogroup in this study and explain the basis for this transition in terminology. The use of “cluster” and “taphogroup” should also be consistent across the text and corresponding figures.