Surface Productivity–Bottom-Water Oxygenation Decoupling During the Last Deglaciation in the Southeastern Arabian Sea

1. Introduction

The Arabian Sea hosts the world’s thickest oxygen minimum zone (OMZ), which accounts for up to about 40% of global pelagic denitrification while occupying less than 2% of the ocean area [1,2]. The OMZ emerges from high oxygen demand, fueled by monsoon-driven productivity, mixed with slow ventilation of intermediate waters, hence its extent and intensity respond to both the Indian monsoon and large-scale ocean circulation [2–4].

Sediment records reveal that the Arabian Sea monsoon, productivity, and OMZ changed on millennial timeframes in tandem with the North Atlantic climate. Laminated, organic-carbon-rich strata in the northeastern Arabian Sea correspond with Greenland interstadials [5]. Nitrogen-isotope data from the Oman margin show that denitrification and productivity were linked to Dansgaard-Oeschger events via the strength of summer-monsoon upwelling [6–8], and Holocene times with a weak summer monsoon coincided with cold periods in the North Atlantic [9].

Productivity alone does not determine oxygen levels at intermediate depths. Changes in the ventilation of thermocline and intermediate waters have been documented for the Indian Ocean on centennial to millennial timescales [10], and [11] attributed a mid- to late Holocene strengthening of the OMZ in the northwestern Arabian Sea, despite weaker monsoon winds, to changes in intermediate-water circulation rather than monsoon intensity. Model tests also suggest that stronger summer monsoon winds increase both oxygen consumption and ventilation, with the net effect varying by depth [2]. Separating productivity from ventilation requires the use of a proxy that measures bottom-water oxygen independent of surface circumstances.

Benthic foraminiferal assemblages are often employed for this purpose because species differ in their susceptibility to low oxygen as well as their response to organic matter flux [12,13].  In the northern Arabian Sea, they were utilized to reconstruct organic matter flux, bottom-water oxygenation, and OMZ intensity [14,15]. Along the eastern coast, benthic assemblages from the Goa shelf link oxic-dysoxic changes during the Holocene to monsoon variability and poorly ventilated intermediate water [16]. Planktic foraminifera from the same samples add surface-ocean context: G. bulloides is an established indicator of upwelling driven by the southwest monsoon [9] whereas G. ruber indicates warm, stratified surface water.

The majority of these reconstructions came from the northern and western Arabian Seas, or from the shelf. There are far fewer deglacial benthic records from the top slope of the southeastern edge, where the OMZ is narrower and moderately oxygen-depleted, and a poleward undercurrent seasonally raises oxygen levels at intermediate depth [17]. Because oxygen levels are closer to the critical thresholds for benthic foraminifera, this environment may be particularly susceptible to changes in ventilation and oxygen demand.

Here we present a high-resolution benthic and planktic foraminiferal record from a gravity core SK 313/GC01 taken at 513 m off Kerala, spanning the last 14.4 kyr. Our objectives are to (1) document changes in the benthic assemblage across the Bølling–Allerød (B/A), Younger Dryas (YD), Early Holocene (EH), Mid-Holocene (MH) and Late Holocene (LH); (2) evaluate how bottom-water oxygenation relates to surface-ocean conditions inferred from planktic foraminifera; and (3) test whether the OMZ in the southeastern Arabian Sea responded to deglacial climate events in the same way as in the northern basin.

2. Study area

The 4.8 m gravity core SK-313/GC-01 was retrieved from the southeastern Arabian Sea (10°30.11′ N, 75°22.50′ E) off Kerala, at around 513 m water depth, during the 313th expedition of ORV Sagar Kanya conducted by the National Center for Polar and Ocean Research (NCPOR), Goa, India (Figure 1). The site is located on the continental slope, inside the depth range of the Arabian Sea OMZ.

The regional climate is influenced by the seasonally reversed Indian monsoon. The southwest (summer) monsoon provides strong winds and copious rainfall over the Western Ghats, which provide freshwater and terrigenous material to the Kerala shelf [18] via numerous west flowing rivers, whereas the northeast (winter) monsoon is drier and has weaker winds [19]. Surface circulation also undergoes a pronounced seasonal reversal. The Summer Monsoon Current (SMC) flows eastward across the southern Indian Ocean during the summer monsoon, whereas the Winter Monsoon Current (WMC) flows westward during winter [19]. Along the western margin of India, the West India Coastal Current (WICC) flows equatorward during the summer monsoon and poleward during the winter monsoon [19–21].

The site lies within the depth range of the Arabian Sea OMZ, but in its southern, weaker part. Year-round ship observations of the eastern Arabian Sea show that the OMZ (< 20 µM O₂) occupies about 60-1350 m over an annual cycle and thins from north to south, that its perennial core (< 5 µM O₂) ends near 15° N, and that a poleward undercurrent raises oxygen in the 150-500 m layer by about 5 µM between 6.5° and 13° N [17]. At 10.5° N and 513 m, our core is just below the layer most impacted by this undercurrent and south of the perennial core OMZ, where very minor changes in ventilation or oxygen demand can shift bottom-water oxygen levels above thresholds that are important to benthic foraminifera.

Previous Holocene benthic foraminiferal research on the eastern margin has concentrated on the Goa shelf and upper slope, where oxic-to-dysoxic shifts have been connected to weakening of the summer monsoon and poorly ventilated intermediate water [16]. The deglacial history of the southeastern slope, particularly the response to the B/A and YD, is less well recorded, which motivates the current study.

3. Materials and methods

3.1 Sampling and age model

The core was sampled at [2–5] cm intervals, giving 168 samples with benthic and planktic counts. The chronology rests on five AMS ¹⁴C dates (NZA 61165–61169) at 3.5, 143, 247, 357 and 459 cm. Ages were calibrated and combined into an age–depth model with the Bayesian software Bacon (Blaauw and Christen, 2011). The model covers 14.4 kyr to the present. The youngest sample dated by the model is 1.1 ka, so the LH is represented by samples between 4.2 and 1.1 ka. We follow the climatic intervals B/A (14.4–12.9 ka), YD (12.9–11.7 ka), EH (11.7–8.2 ka), MH (8.2–4.2 ka) and LH (4.2–0 ka).

3.2 Foraminiferal analysis

The sediment samples were dried, weighed, disaggregated, and wet-sieved through a 63 µm sieve. The >63 µm fraction was dried and inspected under a stereomicroscope using conventional benthic foraminiferal processing and census procedures [22]. Benthic foraminifera were recognized at the species or genus level, with an average count of 280 specimens per sample (range 186-336). Planktic foraminifera were enumerated in the same samples and expressed as percentages of the six species (G. bulloides, G. ruber, Neogloboquadrina dutertrei, Globorotalia menardii, Trilobatus sacculifer, and Orbulina universa).

3.3 Data analysis

Benthic species were converted to relative abundances (percentages). We classified Bolivina spathulata and B. subspinescens as low-oxygen, Uvigerina peregrina and U. proboscidea as organic-flux, and Bulimina marginata and Cibicides spp. individually. Diversity was measured by the number of species and the Shannon index (H). Ward’s hierarchical clustering was used to investigate assembly zones, followed by principal component analysis of square-root-transformed percentages. Pearson correlation coefficients were determined for taxa, diversity, and planktonic percentages [23]. Interval means are the arithmetic averages of all samples within each climatic interval.

4. Results

The benthic composition dramatically changes at the B/A-YD-EH sequence before stabilizing throughout the Holocene (Figure 2, Table 1). Three taxa groups dominate the record: Bolivina spp., Bulimina marginata and Uvigerina spp. (U. peregrina, U. proboscidea), with Cibicides and Ammonia as minor components.

4.1 Bølling–Allerød (14.4–12.9 ka)

The B/A is composed of two different components. At the onset, B. marginata dominates (up to 60% at around 14.2 ka), followed by Uvigerina (up to 28% at 13.9 ka) and, in the oldest sample, Ammonia (approximately 20%). Bolivina spp. began to replace this assemblage at 13.7 ka, accounting for more than 60% of 16 samples between 13.7 and 13.0 ka. As a result, Bolivina ranges from 2 to 85%, with a mean of 52%. G. ruber is already high (mean 57%), but G. bulloides is modest (17%, peak 22% at 13.5 ka).

4.2 Younger Dryas (12.9–11.7 ka)

The YD represents the most severe interval in the dataset. Bolivina spp. dominate all samples (59-92%, mean 81%), with 29 of 30 samples exceeding 60%. B. marginata (0-3%) and Uvigerina (1-5%) are almost extinct, cibicide use is at its lowest (2.2%), and diversity is at its lowest (H = 0.8). G. ruber achieves its highest mean (61%), whilst G. bulloides remains low (17%). With a mean spacing of 40 years, this is also the most well-resolved interval.

4.3 Early Holocene (11.7–8.2 ka)

Bolivina dominance ends suddenly around 11.8 ka, coinciding with the YD-Holocene boundary. Bolivina declines to an average of 18% (5-57%), whereas Uvigerina peaks at 63% at 11.1 ka. B. marginata recovers to a mean of 24%, with a peak of 53% at 9.5 ka. Diversity returns to H = 2.0, and cibicides rise to 7.2%. G. ruber decreases to 32%, while G. bulloides increases to a mean of 21% (max 34% at 10.4 ka).

4.4 Mid-Holocene (8.2–4.2 ka)

The MH assemblage is stable and dominated by Uvigerina (mean 37%, peak 56% at 5.2 ka) and B. marginata (34%, 16–49%), which account for around 70% of the fauna. Bolivina is modest (3-25%; average: 10%). G. bulloides continues to increase (mean 27%; max 45% at 5.1 ka), while G. ruber is at its lowest (24%).

4.5 Late Holocene (4.2–0 ka)

The LH assemblage is similar to that of the MH (B. marginata 35%, Uvigerina 30%, Bolivina 14%), with Uvigerina reaching a peak of 65% at 3.6 Ka. The greatest alteration occurs in the planktic fauna, with G. bulloides reaching its highest mean (40%; maximum 57% at 2.75 ka). The record’s youngest sample is 1.1 ka.

4.6 Relationships between taxa

Bolivina abundance is substantially negatively connected with Uvigerina (r = -0.83), B. marginata (r = -0.82), Shannon H (r = -0.83), and marginally adversely with Cibicides (r = -0.50). Bolivina correlates positively with G. ruber (r = +0.79) and negatively with G. bulloides (r = -0.45). Uvigerina and B. marginata are positively correlated (r = +0.55). Because nearby samples are not independent, these coefficients reflect the intensity of the associations rather than their formal significance.

5. Discussion

5.1 Oxygen-depleted bottom waters during the late B/A and YD

The dominance of Bolivina spp. between about 13.7 and 11.8 ka indicates the most severe oxygen deficit in the previous 14.4 kyr at this site. Bolivina species are opportunists that adapt to extremely dysoxic, organic-rich substrates [12–14], and in this interval they replace the Uvigerina and B. marginata assemblage that characterises the rest of the record. The simultaneous decline in diversity (H = 0.8), the near-absence of epifaunal Cibicides, and the fauna’s near-monospecificity all point to a stressed benthic habitat with low oxygen at the sea floor. The event is abrupt on both ends, with Bolivina peaking at 60% within a few samples at approximately 13.7 ka and then decreasing again at about 11.8 ka, and it is best resolved in the YD, where the mean sample spacing is around 40 years.

The onset does not correlate with the beginning of the B/A (14.4 ka), but rather begins around 700 years later, in the late B/A. The early B/A is instead dominated by B. marginata (up to 60%), indicating a substantial organic-matter flux under oxygen circumstances that were still sufficient for species less resistant of oxygen depletion than Bolivina. Oxygen levels thus dropped in two stages: an organic-flux-rich era before 13.7 ka, followed by severe depletion until 11.8 ka.

5.2 Surface–bottom decoupling

A basic explanation for low bottom-water oxygen is increased export output driven by monsoon upwelling. Our planktic data do not support this for the YD and late B/A: G. bulloides remains low (mean 17%), while G. ruber, an indicator of warm, stratified, nutrient-poor surface water, is at its highest (57-61%). Bolivina is correlated positively with G. ruber (r = +0.79) and negatively with G. bulloides (r = -0.45). The low-oxygen state at 513 m appears to have been separated from the intense upwelling at the surface.

Two non-exclusive explanations are worth examining. First, reduced ventilation of intermediate waters may have reduced oxygen delivery to the slope even in the absence of increased production [2,10] resulting in greater organic-matter depletion with the same flux. Second, a stratified surface layer absorbing freshwater and terrigenous organic matter from the Western Ghats rivers may have supplied organic material to the sea floor without the G. bulloides characteristic of upwelling. These must be distinguished using independent proxies such as benthic and planktic stable isotopes, total organic carbon with source indicators, redox-sensitive trace elements, and sediment accumulation rates.

Records from the northern Arabian Sea link a strong OMZ and high production to warm interstadials and a strong summer monsoon [5,6], and model experiments indicate that the OMZ expands at depth when summer-monsoon winds intensify [2]. Because the YD is a cold interval, generally associated with a poor summer monsoon [9], this framework predicts a weaker OMZ, yet the southeastern site shows its strongest depletion then. This indicates that the southeastern OMZ reacts to additional controls, including intermediate-water ventilation [10,11], Alternatively, regional variances in oxygen supply could outweigh the monsoon signal. This is a hypothesis to be explored with additional cores from the margin, rather than a conclusion based on a single record.

5.3 Early Holocene re-oxygenation

The rapid end of Bolivina dominance at around 11.8 ka, within the age range of the YD-Holocene boundary, indicates that oxygenation was restored within a few millennia. The Uvigerina peak of 63% at 11.1 ka, followed by a B. marginata peak at 9.5 ka, suggests that organic matter flux remained high when oxygen became sufficient for these taxa. At the same time, G. ruber decreases from 61% to 32%, while G. bulloides increases (peak 34% at 10.4 ka), indicating a shift from a stratified surface ocean to more productive, upwelling-influenced conditions, as G. bulloides is an established indicator of upwelling driven by the southwest monsoon [9].

5.4 Stable Mid- to Late Holocene conditions

The MH and LH are very uniform: Uvigerina and B. marginata account for approximately 65-70% of the assemblage, with Bolivina accounting for less than 30%. This shows that bottom-water oxygen at 513 m remained over the threshold, favoring Bolivina. The gradual increase in G. bulloides to about 40% in the LH has no counterpart in the benthic assemblage, which could imply that benthic fauna responded to organic-matter flux and oxygen conditions that varied little over this time period, or that the lower sample resolution (about 115 years) and relative abundances obscured changes. to such changes or that its sensitivity is not captured by relative abundances. Absolute abundances and accumulation rates are required to assess this. The stability also contrasts with the northwestern Arabian Sea, where OMZ strengthening from the mid- to late Holocene was linked to changes in intermediate-water circulation [11] implying that the southeastern margin was less vulnerable.

5.5 Limitations

Several caveats apply. Because relative abundances are closed data, the YD rise in Bolivina must inevitably reduce the percentages of other species. Because no data on benthic foraminiferal accumulation rates, preservation indices, or independent oxygen proxies are available, selective dissolution cannot be ruled out as a source to the YD signal. Planktic data are limited to the six most abundant species. The base of the record, with Ammonia contributing up to 20%, most likely represents shelf specimens moved downslope when sea levels were lower, and its age is near to the lowest dated level. Finally, the age model is based on five dates, therefore event boundaries are subject to age errors of a few hundred years.

6. Conclusions

The benthic foraminiferal record of SK-313/GC-01 shows that oxygen at 513 m off Kerala was lowest between about 13.7 and 11.8 ka, across the late Bølling–Allerød and the Younger Dryas.

1. B/A (14.4–12.9 ka): an early B. marginata-dominated, high-organic-flux assemblage is replaced from about 13.7 ka by Bolivina spp.
2. YD (12.9–11.7 ka): Bolivina spp. average 81% (59–92%), Uvigerina and B. marginata nearly vanish and diversity is lowest (H = 0.8), indicating severe oxygen depletion.
3. Surface–bottom decoupling: the low-oxygen interval coincides with high G. ruber (≈61%) and low G. bulloides (≈17%), so it was not driven by upwelling-related productivity; reduced intermediate-water ventilation and terrestrial organic input are testable alternatives.
4. EH (11.7–8.2 ka): Bolivina dominance ends at about 11.8 ka and is followed by an Uvigerina peak (63% at 11.1 ka), with increasing G. bulloides.
5. MH and LH (8.2–0 ka): a stable B. marginata–Uvigerina assemblage with low Bolivina (10–14%), while G. bulloides rises to about 40%.

Together, these results indicate that the southeastern Arabian Sea OMZ responded strongly to deglacial climate change, in a way that cannot be explained by surface productivity alone. Accumulation rates, stable isotopes and organic geochemistry from the same core are the next steps to separate productivity from ventilation.

Author statement

We declare that this manuscript is original it has not been published before.

Authors’ contributions

Conceptualization: Dhiraj Shinde, V. Yoganandan. Data curation, Formal analysis and investigation, Methodology: Dhiraj Shinde; Supervision: V. Yoganandan, Validation: Visualization: Dhiraj Shinde ; Writing—original draft preparation: Dhiraj Shinde. All authors read and approved the final manuscript.

Acknowledgement

We sincerely thank the anonymous reviewers for their careful evaluation of our manuscript and for providing constructive, insightful, and valuable comments.

Declaration of Competing Interest- The authors declare there is no competing interest.

Funding Statement- This research received no specific grant from any funding agency

 

References

1. Bange HW, Naqvi SWA, Codispoti LA. The nitrogen cycle in the Arabian Sea. Progress in Oceanography 2005;65:145–58. https://doi.org/10.1016/j.pocean.2005.03.002.
2. Lachkar Z, Lévy M, Smith KS. Strong Intensification of the Arabian Sea Oxygen Minimum Zone in Response to Arabian Gulf Warming. Geophysical Research Letters 2019;46:5420–9. https://doi.org/https://doi.org/10.1029/2018GL081631.
3. Shinde D, Veeran Y, Tripathi S, et al. Millennial-Scale Denitrification Dynamics in the Arabian Sea: A Paleo-Bibliometric Review. The Bioscan 2025;20:761–70. https://doi.org/10.63001/tbs.2025.v20.i03.pp761-770.
4. Shinde D, Veeran Y. Ecological-Isotopic Evidences for Deglacial to Holocene Changes in Arabian Sea Denitrification and Oxygenation. Open Journal of Marine Science 2026;16:254–76. https://doi.org/10.4236/ojms.2026.164015.
5. Schulz, von Rad, Erlenkeuser, et al. Correlation between Arabian Sea and Greenland climate oscillations of the past 110,000 years. Nature 1998;393:54–7. https://doi.org/10.1038/31750.
6.   Altabet MA, Higginson MJ, Murray DW. The effect of millennial-scale changes in Arabian Sea denitrification on atmospheric CO2. Nature 2002;415:159–62. https://doi.org/10.1038/415159a.
7. Altabet MA, Murray DW, Prell WL. Climatically linked oscillations in Arabian Sea denitrification over the past 1 m.y.: Implications for the marine N cycle. Paleoceanography 1999;14:732–43. https://doi.org/10.1029/1999PA900035.
8.   Altabet MA, Francois R, Murray DW, et al. Climate-related variations in denitrification in the Arabian Sea from sediment 15N/14N ratios. Nature 1995;373:506–9. https://doi.org/10.1038/373506a0.
9. Gupta AK, Anderson DM, Overpeck JT. Abrupt changes in the Asian southwest monsoon during the Holocene and their links to the North Atlantic Ocean. Nature 2003;421:354–7. https://doi.org/10.1038/nature01340.
10.   Böning P, Bard E. Millennial/centennial-scale thermocline ventilation changes in the Indian Ocean as reflected by aragonite preservation and geochemical variations in Arabian Sea sediments. Geochimica et Cosmochimica Acta 2009;73:6771–88. https://doi.org/10.1016/j.gca.2009.08.028.
11. Das M, Singh RK, Gupta AK, et al. Holocene strengthening of the Oxygen Minimum Zone in the northwestern Arabian Sea linked to changes in intermediate water circulation or Indian monsoon intensity? Palaeogeography, Palaeoclimatology, Palaeoecology 2017;483:125–35. https://doi.org/10.1016/j.palaeo.2016.10.035.
12. Murray JW. Ecology and Applications of Benthic Foraminifera. Cambridge: Cambridge University Press; 2006. https://doi.org/10.1017/CBO9780511535529.
13.   Kaiho K. Benthic foraminiferal dissolved-oxygen index and dissolved-oxygen levels in the modern ocean. Geology 1994;22:719–22. https://doi.org/10.1130/0091-7613(1994)022%3C0719:BFDOIA%3E2.3.CO;2.
14. Dulk M den, Reichart GJ, Heyst S van, et al. Benthic foraminifera as proxies of organic matter flux and bottom water oxygenation? A case history from the northern Arabian Sea. Palaeogeography, Palaeoclimatology, Palaeoecology 2000;161:337–59. https://doi.org/https://doi.org/10.1016/S0031-0182(00)00074-2.
15. Dulk M den, Reichart GJ, Memon GM, et al. Benthic foraminiferal response to variations in surface water productivity and oxygenation in the northern Arabian Sea. Marine Micropaleontology 1998;35:43–66. https://doi.org/https://doi.org/10.1016/S0377-8398(98)00015-2.
16.    Saravanan P, Gupta AK, Li X, et al. Reconstructing hypoxia on the eastern Arabian sea continental shelf during the past ∼7400 years: A multi-proxy approach. Quaternary International 2024;707:1–12. https://doi.org/https://doi.org/10.1016/j.quaint.2024.05.007.
17. Sudheesh V, Gupta GVM, Reddy Y, et al. Oxygen minimum zone along the eastern Arabian Sea: Intra-annual variation and dynamics based on ship-borne studies. Progress in Oceanography 2022;201:102742. https://doi.org/https://doi.org/10.1016/j.pocean.2022.102742.
18. Santhikrishnan S, Jyothibabu R, Sajeev R, et al. Impact of river discharge and suspended sediments on the nearshore benthic environment along the Indian southwest coast. Science of The Total Environment 2024;954:176455. https://doi.org/https://doi.org/10.1016/j.scitotenv.2024.176455.
19. Shankar D, Vinayachandran PN, Unnikrishnan AS. The monsoon currents in the north Indian Ocean. Progress in Oceanography 2002;52:63–120. https://doi.org/https://doi.org/10.1016/S0079-6611(02)00024-1.
20. Shetye SR, Shenoi SSC, Gouveia AD, et al. Wind-driven coastal upwelling along the western boundary of the Bay of Bengal during the southwest monsoon. Continental Shelf Research 1991;11:1397–408. https://doi.org/https://doi.org/10.1016/0278-4343(91)90042-5.
21. Narvekar J, D’Mello JR, Prasanna Kumar S, et al. Winter-time variability of the eastern Arabian Sea: A comparison between 2003 and 2013. Geophysical Research Letters 2017;44:6269–77. https://doi.org/https://doi.org/10.1002/2017GL072965.
22. Schönfeld J, Alve E, Geslin E, et al. The FOBIMO (FOraminiferal BIo-MOnitoring) initiative—Towards a standardised protocol for soft-bottom benthic foraminiferal monitoring studies. Marine Micropaleontology 2012;94–95:1–13. https://doi.org/https://doi.org/10.1016/j.marmicro.2012.06.001.
23. Hammer O, Harper DAT, Ryan PD. PAST: Paleontological Statistics Software Package for Education and Data Analysis n.d.