Gravity core SK 218/1 (8.2 m length) was collected from 3,307 m water depth in the western BoB (14°02′06″ N, 82°00′12″ E; ~204 km offshore) by ORV Sagar Kanya in 2005 (Figure 1).

BoB is a semi-enclosed basin in the eastern part of the northern Indian Ocean. Surface hydrography of BoB is driven by seasonally reversing monsoon wind circulations. During the summer season (June–September), a huge amount of freshwater and sediments are discharged into the bay through major rivers Ganga–Brahmaputra, Irrawaddy–Salween from the Himalayas and Mahanadi, Krishna–Godavari, Cauvery rivers from the Indian peninsula (Bird et al., 2008). The freshwater discharge and overhead precipitation freshen the surface layers by 3–7 psu, inducing strong stratification in the upper water column (Prasanna Kumar et al., 2002). The surface stratification restricts the wind driven mixing to a shallow depth (<20 m), thereby, inhibits the injection of nutrients from the subsurface into the euphotic zone leading to an oligotrophic condition and low primary productivity in BoB. In the winter season (December–February), surface stratification weakens and winter monsoon winds induced mixing invokes vertical transport of nutrients to the surface, enhancing primary productivity (Kay et al., 2018).

The water masses in BoB are a combination of locally and externally sourced (Schott and McCreary, 2001). The upper 100 m of the water column consists of low-saline Bay of Bengal waters (BBW) derived mainly from river discharge into the bay. The water masses at thermocline depth up to about 1,000 m are Red Sea Intermediate Water (RSIW) and Indonesian Intermediate Water (IIW). IIW is derived from North Pacific Central Water, which enters BoB via the Indonesian Throughflow (You, 1998). The water mass between 1,500 and 3,800 m is the Indian Deep Water (IDW) comprising mainly of North Atlantic Deep Water (NADW) and Circumpolar Deep Water (You and Tomczak, 1993; Tomczak and Godfrey, 2003). The Antarctic Intermediate Water (AAIW) lies just above the NADW at 1000–1500 m depth in the Indian Ocean up to 10° S (Lynch-Stieglitz et al., 1994). The Antarctic Bottom Water (AABW) is reported from water depths below 3,800 m (Tomczak and Godfrey, 2003; Figure 1).

A total of 158 sediment samples at regular intervals of 2–4 cm were taken for benthic foraminiferal studies to obtain an average temporal resolution of ~288 years. For separation of foraminiferal tests, sediment samples were processed following the standard micropaleontological techniques (e.g., Singh et al., 2018). Census counts of benthic foraminifera were made on >125 μm size fraction. Based on census data, the relative abundances of each benthic foraminiferal species were calculated. The taxonomic classification of benthic foraminiferal species is based on Schwager (1877) revised by Srinivasan and Sharma (1980), Brady (1884) revised by Barker (1960), Loeblich and Tappan (1988, 1990), and Sen Gupta (2002).

The age model of the core adopted in this study was established by Govil and Naidu (2011) based on 8 AMS ¹⁴C dates and δ¹⁸O record of Globigerinoides ruber correlated with δ¹⁸O_c global isostack curve of Martinson et al. (1987) (Table 1).

A total of 136 species representing 78 genera of benthic foraminifera is recorded (Appendix 1). Temporal variation in relative abundances of major benthic foraminiferal groups and their important constituent species are shown in Figures 2–4. Selected ecologically sensitive benthic species recorded in this study are illustrated in Plate 1.

The benthic foraminiferal assemblage comprises of various taxa belonging mainly to uvigerinids, buliminids, cibicidids, miliolids, cassidulinids, and bolivinids (in decreasing order of relative abundances). Other quantitatively significant taxa of the assemblage are Gyroidinoides spp., Oridorsalis spp., Chillostomella oolina, Eggrella brady, Pullenia spp., and Fursenkoina spp.

The uvigerinid group, a major component of benthic assemblages is represented mainly by Uvigerina peregrina, Uvigerina hispida, and Neouvigerina proboscidea. The relative abundance of total uvigerinids in the core varies between 0 and 76 % with maximum abundances occurring during 21–23 and 26–27 ka (Figure 4). There has been a significant reduction in the abundance of uvigerinids during 14–15, 23–24 ka, and between 38 and 40 ka. The Holocene in general is characterized by a very low abundance of the uvigerinids. The constituent species U. peregrina and U. hispida both follow a pattern similar to that of the total uvigerinids (Figures 2, 4). The uvigerinid taxa are known to be related to both the organic carbon flux and sea bottom oxygen level (Lutze et al., 1986; Hermelin and Shimmield, 1990; Jannink et al., 1998; Singh et al., 2015). The uvigerinids have preference for high organic carbon flux to the sea floor and low-to-moderate oxygen levels.

The buliminid group comprising genera Protoglobobulimina and Bulimina is the next important component of the benthic assemblages. The major constituent species of this group are Bulimina aculeata, Protoglobobulimina pacifica, Protoglobobulimina pupoides, and Buliminella tenuata. The buliminid group shows its high abundances during 12–13, 24–25, around 28 ka, between 30 and 37 ka and between 40 and 43 ka (Figure 4). Its abundance declines significantly during the Holocene (except for the period since 3 ka) and the last glacial maximum (LGM; 19–22 ka). However, B. aculeata, P. pacifica, and B. tenuata show an increase in their abundances during brief intervals within the Holocene (Figure 2). The tapered/cylindrical tests of the buliminids are suggestive of their preferences for infaunal habitat. Hence, their distribution pattern in the core is possibly controlled by the organic matter content in sediment and dissolved oxygen concentrations (e.g., Phleger and Soutar, 1973; Kaiho, 1994). B. aculeata, the major constituent of the buliminid population is considered to be a tolerant species to the low-oxygen environment (den Dulk et al., 2000).

The cibicidid group, which is represented mainly by Cibicides wuellerstorfi and Cibicides robertsonianus shows large variation in its abundance down core. The total abundance of cibicidids was very low between 16 and 45 ka, as compared with the Holocene (Figure 4). There has been a rapid increase in its abundance from 11 ka, reaching to its maximum during 6–8 ka. The variation pattern of C. robertsonianus abundance appears to be almost similar to that of the total cibicidids. The pattern of C. wuellerstorfi abundance, however, is slightly different with its abundance maxima occurring between 40 and 41, 19 and 22, 13.5 and 16 ka, and during 8–9 and 1–3 ka (Figure 3). Cibicides generally prefer epifaunal habitat (Kaiho, 1994; 1999) and are known to be associated with the high-oxygen sea bottom condition (Corliss and Fois, 1990; Singh et al., 2015; Verma et al., 2018; Verma and Singh, 2019).

The miliolids are relatively abundant in the Holocene when compared with the late glacial period. Its high abundance is also recorded during 42–44 ka of MIS 3 (Figure 3). The miliolids are known to have epifaunal habitat and are considered to be associated with oxic environment and low organic carbon flux (Kaiho, 1994, 1999; Singh et al., 2015; Verma et al., 2018).

Temporal variations in abundance patterns of other important groups (cassidulinids and bolivinids) and Gyroidinoides spp., Pullenia spp., and Oridorsalis spp. all show significant changes in their abundances down-core (Figures 3, 4) and the interval of major changes (highs and lows) are associated possibly to variation in sea bottom condition mainly organic matter flux and oxygen level.

The cassidulinids are known to prefer infaunal habitat, moderate to high organic matter flux, and low-oxygen environment (Kaiho, 1994; Almogi-Labin et al., 2000; den Dulk et al., 2000; Singh et al., 2015). The bolivinids being infaunal species are associated with low-oxygen condition and high organic matter content in the sediment. Pullenia prefers infaunal microhabitat in high productive regions (Gooday, 1994), suboxic condition and is less tolerant to oxic environment. Although the environmental preferences of Oridorsalis are not fully known, some studies have shown its association with the low productive condition (Kuhnt et al., 1999). Nevertheless, its abundant occurrence in high productivity areas has also been reported (Woodruff, 1985). Gyroidinoides species have a preference for shallow infaunal habitat and suboxic condition (Kaiho, 1994; Guichard, 1997). Fursenkoina and Chilostomella oolina, both are strikingly absent or rarely present in the Holocene sediment of the examined core. The abundance patterns of these taxa indicate their peak abundances during certain time intervals of the late glacial period (Figure 3). Earlier studies have shown these taxa to be associated with very low sea bottom oxygen condition (Kaiho, 1994, 1999).

Benthic foraminifera depending upon their adaptation capabilities to variation in sea floor environment and microhabitat preferences develop various functional morphologies (e.g., Singh et al., 2015; Verma et al., 2018). Therefore, various morphotypes shown by benthic foraminifera and their association with specific microhabitats can be potentially used in paleoenvironmental reconstructions (e.g., Nigam et al., 2007; Singh et al., 2015). Epifaunal, shallow infaunal, and deep infaunal are the broad microhabitat categories of benthic foraminifera (e.g., Corliss and Fois, 1990; Jorissen et al., 1995). The epifauna are those living on the sediment surface, shallow infauna within sediment top 2 cm, and deep infauna below 2 cm (Buzas et al., 1993). As it is difficult to distinguish clearly the shallow infauna from epifauna, generally both together are classified as epifauna (Jorissen et al., 1995). We grouped quantitatively significant benthic foraminiferal taxa into two broad categories: epifauna and infauna based on the existing knowledge of their microhabitat preferences (Figure 5). Down core variation in abundance patterns of these two groups reflects prominent changes on glacial/interglacial scale and also on the millennial scale. The Holocene period is characterized by a high abundance of epifaunal taxa and corresponding low abundances of infaunal taxa except for the late Holocene showing an opposite pattern since 3 ka. The abundance of infaunal taxa was generally high during the late glacial period between 21 and 45 ka. At the millennial timescale, the faunal record shows a conspicuous decline in infaunal taxa during 38–38.5, 28–29, 23.5–25, and 19–21 ka. During these time intervals, the abundances of epifaunal taxa increased. Additionally, there have been periods of a major increase in epifauna and corresponding decline in infauna between 13 and 15, 40 and 41, and 43 and 44 ka. The period between 10.5 and 13 ka is characterized by a conspicuous reduction of epifauna taxa and corresponding increased abundance of infauna (Figure 5).

Epifaunal taxa generally show varying morphologies such as plano-convex, biconvex, milioline, and rounded trochospiral. The infaunal species commonly prefer to have tapered/cylindrical, flattened ovoid, spherical, or rounded plani-spiral tests (Singh et al., 2015). Singh et al. (2015) in their recent study from the AS OMZ suggested benthic foraminifera with plano-convex tests (epifaunal) and tapered/ cylindrical (infaunal) to be more sensitive to changes in bottom oxygen level coupled with organic matter flux. The relative abundance profiles of these two ecologically sensitive benthic morphotype categories in the examined core are shown in Figure 5. Abundance records of these morphogroups reveal that the pattern of variation in tapered/cylindrical category is broadly opposite to the plano-convex record. The variation pattern of tapered/cylindrical category mirrors to the pattern of infaunal abundance record both on glacial/interglacial and millennial scales. The abundance of tapered/ cylindrical forms was high during the late glacial, whereas, plano-convex tests were more abundant during the Holocene (Figure 5).

The benthic foraminiferal distribution in deep sea is primarily controlled by primary productivity-related organic matter flux to the sea floor and dissolved oxygen level (Otto et al., 1995; den Dulk et al., 1998, 2000; von Rad et al., 1999; Schmiedl and Leuschner, 2005; Bharti and Singh, 2013; Singh et al., 2015). In this study, we evaluate the benthic foraminiferal response to the past changes in surface productivity in western BoB. We compared the abundance records of major benthic foraminiferal groups (uvigerinids, buliminids, bolivinids, and cassidulinids) with the record of mixed layer eutrophic planktic foraminiferal assemblage (relative abundances of Globigerina bulloides, Globigerinita glutinata, and Globigerina falconensis), which has been suggested to be linked to the surface primary productivity (Singh et al., 2011, 2018; Verma et al. under revision); Figure 4). From the faunal abundance records, it is evident that the intervals of increased abundances of these benthic groups broadly correspond to the peak abundances of eutrophic planktic species. It is to be noted that the abundance of eutrophic planktic species is high during the late glacial period as compared with the Holocene. The millennial-scale variations in both the planktic and benthic assemblage records broadly correspond to the Northern Hemisphere cold stadials and warm interstadials (Figure 5). However, the temporal variation in abundance of uvigerinid and buliminid groups appears to be governed by the surface primary productivity-driven organic carbon flux to the sea floor. The primary productivity in BoB is mainly modulated by the summer monsoon-induced precipitation and continental runoff. Earlier studies have shown low primary productivity and thus low export flux of organic carbon to the sea floor, during the periods of intensified summer monsoon when sea surface was more stratified due to increased fluvial discharge into the bay (e.g., Narvekar and Prasanna Kumar, 2006; Da Silva et al., 2017). In contrast, the surface productivity-induced organic flux enhanced during the periods of weakened summer monsoon resulting in less stratified surface waters coupled with winter monsoon intensification invoking vertical advection of nutrient-rich waters to the surface mixed layer.

Abundances of uvigerinids and buliminids are generally low during the Holocene and warm B/A periods when summer monsoon was intensified leading to the low primary production and reduced export flux of organic matter. Conversely, the group cibicidids show its maximum abundances during the periods of low primary productivity.

It is also noticed that the patterns of temporal variation in benthic foraminiferal groups individually vary and do not always match each other. These differences in the individual pattern may be related to their differential responses to the surface productivity-related organic fluxes. Additionally, it is plausible that changes in deep-water oxygenation/ventilation through externally sourced water masses might have influenced their distribution patterns.

Benthic foraminifera are sensitive to changes in dissolved oxygen concentration over the sea floor. They show different microhabitat preferences in response to changes in ocean-bottom oxygen level and adapt distinct morphological features (Corliss, 1985; Corliss and Chen, 1988; Corliss and Emerson, 1990). Previous studies have indicated a good correlation between microhabitat preferences and morphological characteristics of benthic foraminifera and sea bottom oxygen condition (e.g., Corliss and Fois, 1990; Kaiho, 1994). Therefore, the abundance records of different microhabitat categories and morphogroups are the potential proxies to reconstruct the deep-sea oxygenation history (Figure 5). This approach has been applied to the late Quaternary benthic foraminiferal assemblages of the AS to decipher temporal changes of the OMZ intensity in the eastern AS (e.g., Nigam et al., 2007; Mazumder and Nigam, 2014; Singh et al., 2015; Verma et al., 2018). In this study, the relative abundance records of two broad groups of microhabitat categories (epifauna and infauna) in combination with the two predominant morphogroups (tapered/cylindrical and plano-convex) (Tables 2, 3) show major changes on glacial/interglacial and millennial scales. Epifaunal taxa have a preference for high-oxygen condition over the sea floor, whereas infaunal taxa prefer a low-oxygen habitat associated with a decrease in dissolved oxygen content within the sediment (e.g., Corliss and Chen, 1988; Kaiho, 1994). High abundances of tapered/cylindrical benthic foraminiferal tests in sediment are suggestive of low bottom oxygen condition, whereas increased abundances of plano-convex tests indicate the presence of an oxygenated environment at the sea floor (Corliss and Fois, 1990; Gooday, 1994; Verma et al., 2018). The predominance of epifaunal taxa and benthic foraminifera with plano-convex tests during the Holocene suggests an oxygenated sea bottom environment compared with the late glacial period between 21 and 45 ka characterized by a general increase in abundances of infauna, tapered/cylindrical tests, and very low abundances of plano-convex tests. A rapid increase in epifaunal taxa corresponding to a significant decline of infaunal and tapered/cylindrical tests between 13 and 15 ka (equivalent to the warm B/A event) reflects a major shift in the benthic environment from a low-oxygen to the oxygenated condition. From the faunal record, it is clearly evident that the oxygenated benthic environment rapidly switched to a low-oxygen condition again during 10.5–13 ka, a period corresponding to the cold Younger Dryas (YD) event. Further, the relative abundance records of benthic microhabitat and morphotype categories for late glacial period depicts millennial-scale oscillations in the late glacial period, apparently in concert with the Northern Hemisphere Heinrich and Dansgaard–Oeschger (D-O) events (Figure 5). These short-term fluctuations in benthic records are obviously related to the changes in bottom water oxygen condition, which seems to be more pronounced between 40 and 45 ka.

In order to have a better reconstruction of past changes in bottom oxygen condition, we used oxygen-sensitive benthic species of assemblages and classified them into low-oxygen and high-oxygen groups based on their association with changing dissolved oxygen levels (e.g., Douglas, 1981; Hermelin and Shimmield, 1990; Kaiho, 1994; Jannink et al., 1998; Singh et al., 2015) (Table 4; Figure 6). Temporal changes in relative abundances of these oxygen-sensitive benthic foraminiferal groups exhibit major variations during 3–10.5 ka when low-oxygen taxa were significantly replaced by the benthic species associated with a high-oxygen environment. The faunal proxy record, however, further suggests a relatively low-oxygen condition between 3 and 5 ka (Figure 6).

Our study reveals that the sea bottom oxygen condition in the western BoB oscillated in concert with the Northern Hemisphere climate events. The bottom water was relatively oxygen-poor during the time intervals corresponding broadly to the North Atlantic Heinrich events H1, H2, H3, and H4 and D-O cold stadials as compared with the warm interstadials. Apparently, the change in bottom water oxygen condition during the H4 event was more pronounced than H1–H3. Furthermore, the benthic environment at the core site during the cold YD event is also characterized by low-oxygen condition. Prior to YD, the bottom water was oxygenated during the period corresponding to the warm B/A. It is also evident from the faunal record that the bottom water was relatively oxygen-poor during the LGM (19 - 22 ka). However, it appears that the bottom water during 19–21 ka was more oxygenated than during 21–22 ka suggesting two-step changes in sea floor environment within the LGM. The other most striking feature of major fluctuation in oxygen condition during MIS 3 is noticed between 35 and 45 ka. The faunal data reveal a high oxygenated environment at around 38, 40, and 42.5 ka which broadly corresponds to the D-O interstadials 8, 9, and 11, respectively (Figure 6).

The bottom water oxygen condition is driven mainly by two factors: (1) amount of organic matter to the sea floor and magnitude of its degradation and (2) deep-water circulation (Joos et al., 2003; Rathburn et al., 2018). By comparison of benthic foraminiferal record with the surface productivity record (mixed layer eutrophic planktic foraminifera abundance), it can be inferred that the bottom oxygen condition at the core site is governed primarily by the surface productivity-related export flux of organic carbon to the sea floor. The respiration of sinking organic matter and the release of CO₂ lead to a low-oxygen condition on the sea floor. The benthic environment was more oxygenated (as evident by high abundances of high-oxygen taxa) when surface productivity and export flux of organic carbon was low, a condition associated with the intensified SW summer monsoon precipitation resulting in high fluvial discharge and stratified surface waters in BoB. On the other hand, the export flux of organic carbon was enhanced during the intensified NE winter monsoon wind-induced vertical mixing, coupled with a weakened summer monsoon. Additionally, the deep-water circulation might have also influenced the oxygen level of bottom water mass at the core site. Previous studies have shown that the relative proportions of NADW and southern sourced AAIW and AABW modulated the bottom water oxygen condition in the northern Indian Ocean both on the glacial/interglacial and millennial scales (e.g., Ahmad and Labeyrie, 1994; Piotrowski et al., 2009; Ma et al., 2019 and references therein). A significant reduction in NADW export to the Indian Ocean during the cold North Atlantic events (Heinrich and YD events) has been suggested (Naqvi et al., 1994; Piotrowski et al., 2009). These studies provided ample evidences of increased propagation of NADW to BoB during the Holocene (Piotrowski et al., 2009; Ma et al., 2019) compared with the late glacial period. It is plausible, therefore, that the deep water at the core site was more ventilated by NADW during the Holocene. We also suggest that the high-oxygen condition during the warm D-O interstadials of MIS 3 is most likely related to the effective propagation of NADW to the BoB. This is in line with the previous results indicating an enhanced influence of NADW during the warm B/A event (Piotrowski et al., 2009). However, more research on high-resolution timescale is required to better understand the relative contributions of NADW and AAIW/AABW flow in modulating the deep ocean oxygenation history in BoB in concert with the D-O oscillations.