Oxygen minimum zones (OMZs) are regions in the ocean where oxygen saturation is at its lowest, typically found at depths below the mixing layer. These zones are crucial for understanding marine ecosystems as they represent the limit of survival for aerobic marine organisms (Slater and Kroopnick, 1984). OMZs are formed in areas of high organic matter remineralization and play a significant role in the cycling of nutrients like phosphate, manganese, and iron (Lewis and Luther Iii, 2000; Kraal et al., 2012). Currently, the Indian Ocean hosts the thickest OMZ, which originates from the Arabian Sea upwelling zone and extends laterally up to 5,000 km into the central Indian Ocean, with a depth range between 200 and 2000 m (Wyrtki et al., 1971; Jensen et al., 2011). The distribution of dissolved oxygen in the oceans is controlled by the balance between water mass ventilation and the remineralization of organic matter. During the Late Miocene to Early Pliocene, a significant oceanographic event known as the “Late Miocene Early Pliocene Biogenic Bloom” (LMBB) occurred, characterized by widespread oxygen depletion in the Pacific and Indian Oceans. This event is believed to have lasted for approximately 5 million years, from 9.0 Ma to 3.5 Ma (Dickens and Owen, 1994; Dickens and Owen, 1999; Hermoyian and Owen, 2001), and was associated with increased productivity in the northern basins of the Indian Ocean, leading to the expansion of the OMZ (Bialik et al., 2020; Auer et al., 2023).

A long-standing hypothesis suggests that the OMZ extended as far south as Broken Ridge (31°S) during the LMBB. This hypothesis is supported by two main lines of evidence: a depletion in manganese and the Mn/Sc ratio in Broken Ridge sediments between 6.5 and 3 Ma (Dickens and Owen, 1994), and a coeval drop in benthic foraminifera (BF) diversity along with an increase in the relative abundance of Bolivina pusilla and Ehrenbergina carinata, which are considered indicators of low-oxygen and high-productivity conditions (Singh et al., 2012; Gupta et al., 2013). However, these species are not exclusive to low-oxygen environments and can also be found under oxic conditions (Geslin et al., 2004). Furthermore, some benthic foraminifera are capable of surviving low-oxygen conditions via denitrification or kleptoplasty, introducing uncertainty into strict species-level oxygen classifications (Glock, 2023).

Late Miocene ocean-atmosphere simulations further support the southward expansion of the OMZ, showing a latitudinal range of oxygen-depleted intermediate waters comparable to the pre-industrial control simulation (Sarr et al., 2022) (Figure 1). However, these simulations assumed a pre-industrial global nutrient supply, which may underestimate the extent of Late Miocene oxygen depletion, as it is generally accepted that surface-ocean nutrient levels were globally higher during this period (Diester-Haass et al., 2002; Auer et al., 2023).

Nevertheless, the hypothesis of OMZ expansion remains a topic of debate. High-resolution productivity reconstructions based on XRF core scanning at ODP Site 752, developed and calibrated against organic carbon content in Lyu et al. (2023), do not indicate a significant productivity increase during the LMBB. This suggests that Broken Ridge has been in a nutrient-poor region of the South Indian gyre since the Middle Miocene. Additionally, sedimentological evidence indicates substantial winnowing at Site 752 during the LMBB (House et al., 1991; Lyu et al., 2023), which could explain the observed decrease in Mn and Mn/Sc ratios through diagenetic changes rather than a reduction in oxygen levels.

To further investigate the potential southward expansion of the Indian Ocean OMZ during the LMBB, we employed a multi-proxy approach focusing on benthic foraminiferal taxonomic and quantitative analyses. Given that winnowing can affect geochemical proxies, our study uses benthic foraminifera to reconstruct changes in bottom-water oxygen concentrations over the past 9 Myr. We use the Enhanced Benthic Foraminiferal Oxygen Index (EBFOI) (Kranner et al., 2022), which allows semi-quantitative reconstruction of dissolved oxygen conditions based on full assemblage data. The EBFOI of the core-top assemblage at Site 752 aligns with today’s well-oxygenated waters (∼4.5–5.0 mL/L), providing a robust benchmark for comparison.

Our reconstructions suggest persistent oxic conditions at Broken Ridge throughout the study interval, challenging the notion of a southward OMZ expansion to Broken Ridge during the LMBB. Instead, we propose that the observed shifts in benthic foraminiferal assemblages are more likely related to variations in the intensity of Tasman Leakage, which influences the characteristics of intermediate waters in the central southern Indian Ocean.

The benthic foraminiferal (BF) assemblages from ODP Site 752 on Broken Ridge reveal dynamic but consistently moderate ecological conditions across the last 9 Myr (Figure 3). BF diversity, evaluated using the Fisher Alpha index, remains moderate (between 5 and 10) throughout the interval, with a notable dip near five at approximately 5.4 Ma (Figure 3e). Infaunal species composition fluctuates, peaking at 72.9% between 4.5 and 4.1 Ma, while stress-associated species remain relatively stable, never exceeding 48.8% (Figures 3c,d). These findings collectively suggest steady and moderate oxygenation levels throughout this time.

The absence of consistent trends or remarkable features in benthic foraminiferal species assemblage during the Late Miocene Biogenic Bloom (LMBB) aligns with the outcome of our cluster analysis (Figure 3a), which identifies two distinct assemblage groups not aligned with LMBB timings. Cluster I (9.1–5.8 Ma) is characterized by the dominance of Hanzawaia boueana, Siphonodosaria spp., Bolivina spathulata, and Bolivina reticulata, whereas Cibicidoides wuellerstorfi, Ehrenbergina carinata, and Textularia spp. Remain scarce or absent. In contrast, Cluster II (5.5–2.2 Ma) exhibits increased occurences of Globocassidulina subglobosa, Cibicidoides mundulus, Oridorsalis umbonatus, and Uvigerina auberiana, alongside reduced abundance of Hanzawaia boueana compared to Cluster I. These cluster-defined groups reflect an assemblage shift that does not coincide with the timing of the LMBB, suggesting that factors other than productivity variations drove the observed ecological transitions. Ecologically, the shift between the two clusters represents a transition from a community dominated by species associated with relatively stable, low-nutrient environments toward one favoring taxa indicative of well-ventilated, oxic intermediate waters.

Oxygen levels reconstructed through Benthic Foraminifera Oxygen Index (BFOI) and Enhanced BFOI (EBFOI) indicate well-oxygenated bottom waters across the studied interval (Figures 3f, 4D). The average EBFOI-based oxygen concentration is 4.2 mL/L, comparable to modern levels on Broken Ridge and persistently above 3.2 mL/L, supporting a high-oxygen environment. The prominence of oxic species, such as Cibicidoides wuellerstorfi and Globocassidulina subglobosa, further substantiates this conclusion. However, we emphasize that the EBFOI should be considered here as a semi-quantitative proxy, useful for tracking relative oxygenation changes rather than absolute concentrations. This is because alternative metabolic strategies (e.g., denitrification, fermentation, kleptoplasty) among benthic foraminifera (Geslin et al., 2004) can decouple past species distributions from oxygen alone. Furthermore, agglutinated taxa are underrepresented in our assemblages due to preservation issues, potentially influencing EBFOI results in the lower oxygen range. Yet, we note that the EBFOI of the youngest sample (∼0.1 Ma) closely aligns with present-day bottom water oxygenation (∼4.5–5.0 mL/L), lending credibility to the entire record’s internal consistency.

BF analyses thus underscore that Broken Ridge’s intermediate waters retained oxic conditions throughout the Late Miocene to Early Pleistocene. We interpret this pattern as the result of sustained influence from southern-sourced, well-oxygenated water masses, likely Tasman Leakage or Antarctic Intermediate Water, rather than from local productivity changes or OMZ expansion. This finding challenges previous assumptions of OMZ expansion and its impact on Broken Ridge, though we emphasize that our interpretation represents a hypothesis grounded in foraminiferal ecology. In the following section, we assess whether sedimentological and geochemical records provide additional support for this interpretation.

Our findings indicate that the influence of the Late Miocene Biogenic Bloom (LMBB) on the southeastern Indian Ocean, especially at Broken Ridge, warrants re-evaluation. Dickens and Owen (1999) emphasized that understanding productivity changes in the central Indian Ocean during the LMBB is critical for assessing broader paleoceanographic impacts. Indeed, the central Indian Ocean is a dynamically complex region where multiple water masses converge and interact. During the Miocene–Pliocene transition, the intermediate water mass configuration of the Indian Ocean was shaped by the competing contributions of Antarctic Intermediate Water (AAIW), Indonesian Intermediate Waters (IIW), and Tasman Leakage (Christensen et al., 2021). Quaternary reconstructions of intermediate water dynamics (You, 2000; Piotrowski et al., 2009; Struve et al., 2022; Raddatz et al., 2023) highlight the role of remote sourcing and ventilation in controlling bottom water oxygenation across the Indian Ocean—mechanisms that likely also extend to earlier intervals.

In previous work (Lyu et al., 2023), we presented an XRF-derived paleoproductivity proxy from Site 752 (also shown in Figure 4A), which demonstrated consistently low productivity levels during the LMBB. The XRF-derived productivity index was developed by utilizing a group of metallic micronutrients (Cd, Cu, Ni, Zn, Cr and Br) divided by Ca counts. This paleoproductivity index was calibrated against directly measured organic carbon contents from discrete samples at ODP Site 752 (Lyu et al., 2023). This proxy suggests that Broken Ridge likely remained an oligotrophic zone within the central South Indian gyre since the Late Miocene, positioned away from nutrient-rich waters at the subtropical front to the south and from continental nutrient sources to the north.

Previous benthic foraminifera assemblage studies at Broken Ridge, however, came to different conclusions. Indeed, studies by Singh et al. (2012) and Gupta et al. (2013) suggested that the OMZ extended southward to Broken Ridge, based on reduced BF diversity and increased low-oxygen taxa during the LMBB. These low-oxygen indicators, including Uvigerina peregrina and Trifarina angulosa, correlate with oxygen depletion and heightened productivity. Although our assemblages also show the presence of these taxa in Late Miocene and Early Pliocene samples, their relatively low abundance—alongside a strong presence of oxic species—supports our alternative interpretation. The dominance of oxic species throughout this period suggests persistently well-oxygenated conditions, casting doubt on the OMZ expansion hypothesis at Broken Ridge (Figure 3f).

While our assemblage data diverge from previous interpretations in the above-described aspect, they are not entirely incompatible with Singh et al. (2012) and Gupta et al. (2013). Moreover, we highlight that our observations are consistent with the “little-faunal-turnover” conclusion by Ridha et al. (2019) for Neogene benthic foraminiferal assemblages at Broken Ridge. All Broken Ridge benthic assemblage studies, including ours, report increased low-oxygen, high-productivity taxa alongside higher stress and infaunal taxa abundances from 5.1 to 4.1 Ma (Figures 3c,d), indicating a minor decline in oxygen, but within overall oxic conditions. BF diversity, shown by the Fisher Alpha index (Figure 3e), decreases slightly between 7.0 and 4.1 Ma, consistent with earlier studies. However, our cluster analysis (Figure 3a) does not align this trend with the LMBB period. Additionally, XRF-derived productivity proxies (Figure 4A) do not suggest a substantial LMBB effect at Site 752, likely due to the unique paleoceanographic configuration at Broken Ridge (Lyu et al., 2023; Pillot et al., 2023). Together, these findings imply that neither productivity increases nor oxygen depletion drove BF distributions on Broken Ridge across this interval.

Instead, we propose that the primary factor influencing BF variability at Site 752 was Pacific-sourced Tasman Leakage (TL) waters. XRF geochemical and grain-size analyses (Lyu et al., 2023; Figure 4B) show an increase in winnowing from approximately 7 Ma, aligning with the onset of the modern TL circulation (Christensen et al., 2021). TL currents intensified between seven and 4 Ma, bringing low-nutrient, higher-oxygen waters to Broken Ridge and favoring oxic BF species. The peak winnowing effect, as indicated by grain-size and BF distribution data, occurs around 4 Ma, coinciding with increased Cibicidoides wuellerstorfi, shifts in stress species, and infaunal species composition (Figures 3b,c). C. wuellerstorfi’s epiphytic lifestyle suggests adaptation to stronger bottom currents introduced by TL (Linke and Lutze, 1993). The decrease in Mn and Mn/Sc, initially interpreted as indicative of decreasing oxygen levels according to (Dickens and Owen, 1999), could instead be explained by a change in sediment composition due to winnowing, where finer, Mn-enriched particles are removed, while Sc remains relatively evenly distributed across the sediment fraction. This reinterpretation is supported by the sedimentological changes at Site 752 and does not require invoking redox-driven early diagenesis.

Further evidence to support TL’s role in shaping Broken Ridge’s oxygen levels comes from carbonate MARs. This proxy is commonly used to quantify productivity throughout the LMBB (Lyle et al., 2019; Karatsolis et al., 2022; Sutherland et al., 2022; Gastaldello et al., 2023). Carbonate MARs at Broken Ridge peak around 5.5 Ma (Figure 4C), yet show no sustained increase during the LMBB, reinforcing a generally low-productivity environment at Broken Ridge. The steady CaCO₃ content over this period suggests that MAR variability mainly reflects sedimentation rate changes rather than shifts in nutrient availability or organic matter deposition.

Our model, which highlights a strong TL influence dominating intermediate water conditions over Broken Ridge since the Late Miocene (Christensen et al., 2021; Lyu et al., 2023), provides a more compelling explanation for the high-oxygen conditions observed at Broken Ridge during the Late Miocene and Early Pliocene than a southward expansion of the OMZ. TL likely acted as a barrier to OMZ encroachment, maintaining a nutrient-poor, well-oxygenated environment over Broken Ridge throughout the LMBB and beyond. Arguably, this interpretation is solely based on indirect faunal and sedimentological evidence. Future work should include direct geochemical tracers (e.g., δ¹³C_benthic, εNd(t), Mg/Ca, clumped isotopes, total organic carbon, and total nitrogen) to robustly test the role of TL and water mass sourcing in maintaining oxic conditions at Broken Ridge. When confirmed, this mechanism underscores the role of TL as a key factor in controlling southeastern Indian Ocean oxygenation patterns at intermediate depths (Figure 5).

Detailed taxonomic and quantitative analyses of benthic foraminifera from ODP Site 752 (Broken Ridge) revealed no evidence of oxygen depletion in the upper intermediate waters of the southeastern Indian Ocean during the LMBB. Instead, the new evidence presented here suggests that oxic conditions persisted at intermediate depths in the last 9 Myr at Broken Ridge. We interpret these findings as indicating that the Indian Ocean OMZ did not reach Broken Ridge during the LMBB. Rather, we propose that modern-style TL significantly influenced benthic ecosystems at Broken Ridge since its onset in the Late Miocene (Figure 5). The benthic foraminifera distribution observed here is likely linked to the vigor of bottom water currents and the supply of intermediate-depth waters. Thus, we propose a new mechanistic model in which high-intensity TL limits the expansion of the OMZ in the southeastern intermediate-depth Indian Ocean.

Tasman Leakage is characterized by a bottleneck configuration south of Australia, with the Australian continent to the north and the Antarctic frontal system to the south. The intensity of TL and inter-basin exchange through this pathway from the Pacific to the Indian Ocean is shaped by the interplay between tectonic configuration and the latitudinal position of climate belts. During the LMBB (Figures 5C,D), global climate was warm enough to keep the Antarctic frontal systems far south of southern Australia, enabling intense TL flow (Christensen et al., 2021; Lyu et al., 2023). This flow of TL waters prevented the expansion of oxygen-poor waters from the northern Indian Ocean into the southeastern Indian Ocean at the same depth.

While we interpret this TL-driven mechanism as the most plausible explanation for sustained oxygenation, we recognize that this conclusion is based on indirect evidence from benthic assemblages, sedimentological trends, and geochemical proxies. Future studies using isotopic water mass tracers will be essential to confirm the extent and timing of TL’s influence on intermediate water masses in the Indian Ocean. That being said, our interpretation is supported by the alignment of benthic faunal shifts, geochemical evidence, and sedimentary patterns with known TL intensification. We therefore conclude that the Early Pliocene represented an optimal Earth System configuration for maximizing TL intensity—shaping benthic ecological conditions at ODP Site 752 and buffering this site from the OMZ expansion observed farther north in the Indian Ocean.