1 Introduction

The South Asian monsoon, characterized by the seasonal reversal of surface winds and associated precipitation (IPCC 2014), affects the lives of more than a billion people. Yet, the projections using state-of-the-art climate models exhibit a high degree of uncertainty in determining local precipitation. Accurately projecting future monsoonal variability and the changes in regional precipitation is thus difficult to achieve (Turner and Annamalai 2012). For this reason, a number of studies have used geological records to examine the responses of the South Asian monsoon to climate change. In particular, long-term changes in monsoonal runoff are estimated from the detrital flux in the Bengal Fan sediments (Hovan 1995; Galy and France-Lanord 2001; Ali et al. 2021).

The Bay of Bengal (BOB), the northeastern part of the Indian Ocean, receives a large amount of detritus from major rivers draining the Himalayas (e.g., Ganges, Brahmaputra, and Meghna). However, the source of the sediments is not exclusive to the Himalayas as evidenced by differences in their radiogenic isotope (Sr–Nd) compositions (Colin et al. 1999; Joussain et al. 2016; Ali et al. 2021). Instead, a significant input of detritus from the Myanmar region (e.g., through the Irrawaddy and Salween rivers), where monsoon rainfall is particularly intense, is suggested based on the Sr–Nd compositions of the BOB sediments. The changes in the runoff of the major rivers in the BOB, therefore, can be estimated quantitatively by evaluating the detritus contribution to sedimentary records from various rivers with different drainage areas in the BOB. Nevertheless, the fan-dominated depositional processes of the central BOB often complicate the recovery of continuous, high-resolution sedimentary records. For this reason, the Ocean Drilling Program (ODP) and the International Ocean Discovery Program (IODP) drilled the Ninetyeast Ridge (NER) where the ridge-top location of the NER has prevented the turbiditic sediment input and the resultant truncation of the underlying sequence: ODP Site 758 and IODP Site U1443 (5° 23.05′ N, 90° 21.67′ E) (Pierce et al. 1989; Clemens et al. 2016; see locations in Fig. 1). According to the paleoceanographic variability and the detrital input from the Himalayan erosion for the late Neogene (Gourlan et al. 2010; Bolton et al. 2013; Ali et al. 2021; Bretschneider et al. 2021), these sites are governed by the Indian monsoon regime’s precipitation and open-ocean dynamics regulated by its current system.

Fig. 1
figure 1

Core location of HI1710-MC1 and schematic map of surrounding area. The study sites of cited literature are also shown (see legend) with provinces suggested by Colin et al. (1999). The map was generated with Ocean Data View software (Schlitzer 2020) using 6′ × 6′ gridded bathymetry data from GEBCO 2014 database (https://www.gebco.net/data_and_products/gridded_bathymetry_data/) and modified based on Ali et al. (2021). NER Ninetyeast Ridge, BOB Bay of Bengal

Full size image

We collected two multiple cores (34 cm- and 35-cm long) at the northern NER using the R/V ISABU of the Korea Institute of Ocean Science and Technology (KIOST, launched in 2017) during the 2017 research cruise in the Indian Ocean (project title: Marine environmental change and the evolution of Indian monsoon). To investigate the surface ocean environment and the provenance of the detrital components, we used the multi-proxy dataset from the 35-cm long undisturbed multiple core. We evaluated the origin of the study site’s biogenic sediment and compared the analyzed radiogenic isotope (143Nd/144Nd and 87Sr/86Sr) and clay mineral compositions of detrital components to those of the adjacent basins (i.e., the BOB and Andaman Sea). We addressed the provenance and the relative contribution of the Himalayan and Myanmar detrital components during the Holocene. This study provides background data for the reconstruction of changes of runoff in the major river systems under the influence of the Indian monsoon system. The sedimentary records spanning the last 13 kyrs will serve as a baseline dataset for the long-term paleoceanographic reconstruction.

2 Study Site

A multiple core HI1710-MC1 (5° 23.010′ N, 90° 21.738′ E, 2925 m water depth) was collected from the NER in the BOB during the HI1710 expedition of the R/V ISABU in 2017 (Fig. 1). The NER formed before the middle Eocene by age-progressive hotspot volcanism that is currently beneath the Kerguelen Plateau (Royer et al. 1991; Sager et al. 2010; Clemens et al. 2016). The ridge-top location of the site allows preservation of a continuous sedimentary record suitable for paleoceanographic study. Further, the surface salinity at the study site remains relatively constant throughout the year despite the strong monsoonal precipitation, yielding exceptionally strong north–south salinity gradient and the seasonal variation in the local salinity of the BOB and Andaman Sea (Fig. 2). Thus, we considered this site to be at the southern limit influenced by the Indian summer monsoon.

Fig. 2
figure 2

Statistical mean salinity data on 1° grid for 1981–2010 from World Ocean Atlas 2018 (Zweng et al. 2018). The study site (HI1710-MC1) is shown as an open diamond (DJF December–January–February, MAM March–April–May, JJA June–July–August, SON September–October–November)

Full size image

3 Methods

The 35 cm-long multiple core was halved, photographed, and sectioned at 1-cm intervals at the Library of Marine Samples (LIMS), South Sea Branch, KIOST. The sectioned subsamples were stored in 50 mL plastic jars and then freeze-dried. To separate sand-sized foraminifera from the clay, approximately 1 g of bulk sediments was soaked in 5% sodium hexametaphosphate solution and then sieved using tap water into > 125 μm, 125–63 μm and < 63 μm fractions. More than 100 specimens of planktic foraminifera (mixed species) were isolated from the > 125 μm fraction of four selected intervals for an accelerator mass spectrometer (AMS) radiocarbon dating (Table 1). Radiocarbon age determinations were processed using an AMS at the Beta Analytics, USA and the Rafter Radiocarbon Laboratory of National Isotope Centre, New Zealand. Calibration of the radiocarbon ages was performed with the Calib 7.0 (http://calib.org) software and the Marine20 calibration curve (Stuiver and Reimer 1993; Heaton et al. 2020; Stuiver et al. 2021). The local reservoir effect (ΔR) of 17 ± 70 year was applied which was reported for Nicobar Island (9° N, 94° E) (Southon et al. 2002).

Table 1 Analytical results of HI1710-MC1
Full size table

The surface ocean productivity was inferred from the biogenic components. The bulk sediment was dried at 50 °C and ground using mortar and pestle for the analyses of total carbon (TC), total inorganic carbon (TIC), and total nitrogen (TN) contents. The TC and TN contents of every subsample were measured with a Flash 2000 Series Elemental Analyzer at Pusan National University (PNU) and was reported as a percentage of sediment dry mass. The analytical precisions were less than ± 0.1% and ± 0.01%, respectively. The TIC content was measured with a UIC CM5012 CO2 coulometer at PNU (Engleman et al. 1985). The analytical precisions were less than ± 0.1%. The TIC was multiplied by 8.333 (the molecular weight ratio between CaCO3 and C) to calculate the calcium carbonate content. The TIC was subtracted from the TC to estimate the total organic carbon (TOC) content.

To investigate detrital provenance, the inorganic silicate fraction (hereafter the detrital fraction) was extracted from the bulk sediments using a method outlined by Hovan (1995) at KIOST. The < 63 µm sample was treated with a 10% acetic acid to remove the carbonate fraction. Subsequently, to remove Fe–Mn oxides and hydroxides, the sample was treated with a hot sodium citrate–sodium dithionite solution buffered with sodium bicarbonate. The biogenic silica component was then removed with an 80 °C sodium hydroxide solution. An analysis of the 143Nd/144Nd and 87Sr/86Sr ratios of bulk inorganic silicate fraction, including isotopic separation and multi-collector thermal ionization mass spectrometric (TIMS; VG54-30, Isoprobe-T) analysis, was performed at the Korea Basic Science Institute. The detailed methodology was outlined by Cheong et al. (2016). The replicate analysis of NBS987 and JNdi-1 provided the mean values of 143Nd/144Nd = 0.512110 ± 20 (n = 5, 2σ), 87Sr/86Sr = 0.710251 ± 70 (n = 5, 2σ). For convenience, we expressed the 143Nd/144Nd ratios as εNd; i.e., the deviation from a chondritic uniform reservoir (εNd = (143Nd/144Nd/0.512638 − 1) × 104) (Jacobsen and Wasserburg 1980).

The clay mineralogy was determined from the < 2 μm clay fraction, which was separated according to the Stoke’s Law, spread onto glass slides, and allowed to air dry. Analyses were performed on air-dried and ethylene glycol-treated mounts using the SIMENS/Bruker D5005 X-ray diffractometer with CuKα radiation in 40 kV, 40 mA, and 3–30° (2θ) conditions at the Department of Geology, Gyeongsang National University. The relative abundances of four major clay minerals (smectite, illite, kaolinite and chlorite) were semi-quantitatively estimated using the Eva 3.0 software with the empirical factors of Biscaye (1965). The relative proportions of kaolinite and chlorite were determined based on the ratio from the 3.58 Å (kaolinite 002) and 3.54 Å (chlorite 004) peak areas.

4 Results

The HI1710-MC1 sediments consisted of a homogeneous pale brown nannofossil ooze with foraminifera and clay without any noticeable lithological changes. Table 1 summarizes the analytical results of HI1710-MC1 sediment as shown in Fig. 3. Table 2 presents the four AMS 14C dates of planktic foraminifera. The core extended to 14,213 cal yr before present (BP), and the average linear sedimentation rate (LSR) of the composite section was estimated to be 2.4 cm/kyr. The increase in LSR in the uppermost layer (Fig. 3a) is attributable to the high water content and minimal compaction. The age model was established using linear interpolation between measured intervals (Fig. 3a). The CaCO3 contents ranged between 56.1 and 64.7 wt% (avg. 61.0 wt%) (Fig. 3b). The TOC and TN were remarkably low (< 0.5% and < 0.07%, respectively) and TN contents were below the limit of detection in several samples (Fig. 3c). The measured εNd and 87Sr/86Sr varied between − 11.8 and − 12.4, and 0.7260 and 0.7289, respectively (Fig. 3d, e). The most common clay mineral in the < 2 μm fraction was illite (av. 50.1%), followed by smectite (av. 22.7%), kaolinite (av. 14.6%), and chlorite (av. 13.6%) (Fig. 3f). The relative abundance of each clay mineral group did not change significantly through the analyzed section.

Fig. 3
figure 3

Analytical results of HI1710-MC1 plotted against depth. a Depositional ages, b CaCO3 content, c total organic carbon (TOC) and total nitrogen (TN) contents, d 87Sr/86Sr of lithogenic component, e εNd of lithogenic component and f clay mineral assemblages. C chlorite, K kaolinite, Il illite, Sm smectite

Full size image
Table 2 AMS radiocarbon ages of selected intervals for HI1710-MC1
Full size table

5 Discussion

5.1 Surface Ocean Environment at the Ninetyeast Ridge

The CaCO3 content in marine sediment shows spatial and temporal variation which reflects its production, preservation, and dilution by non-carbonate materials (Arrhenius 1952; Farrell and Prell 1989; Naidu et al. 1993). In the BOB, the sediment discharged from the major rivers (e.g., the Ganges–Brahmaputra, Irrawaddy, etc.) often exert a major control on the CaCO3 concentration in marine sediment, especially at the sites proximal to the continent or under the influence of the submarine fan deposition (Kolla et al. 1976; Suresh and Bagati 1998). For instance, the CaCO3 content at the western BOB sites close to the Indian continent is low (generally < 10%) during the Holocene, but it was higher during the last glacial maximum (LGM). This is due to the weakened summer monsoon and the resultant decrease in sediment discharge (Suresh and Bagati 1998; Phillips et al. 2013). Such effect of terrigenous dilution decreases greatly toward the offshore area as inferred from the reduced contribution of the lithogenic component in suspended sediments (e.g., Rixen et al. 2019), and it is often considered insignificant in the central ocean basin. Carbonate dissolution has been considered as a significant control on the CaCO3 content as well, especially in the carbonate-undersaturated deep ocean (e.g., Peterson and Prell 1985).

The CaCO3 content of HI1710-MC1, obtained from the central ocean ridge, does not vary greatly for the last 14.4 kyrs (Fig. 3b). Minor variability in centimeter scale could be attributed to the millennial- or centennial-scale climate variability (e.g., 8.2 ka event; Banerji et al. 2020); however, the low sedimentation rate (ca. 2.4 cm kyr−1) at the study site does not resolve such events precisely. Still, CaCO3 was lower in the 17–30 cm interval, which corresponded to the depositional ages of 6–11 ka. As previously determined carbonate preservation state at the NER sites suggested that the dissolution maxima have often occurred at the glacial initiations (Peterson and Prell 1985), this early- to mid-Holocene minimum is not attributable to the carbonate dissolution. Instead, this time period is marked by a strengthened ISM precipitation (Fleitmann et al. 2003; Hong et al. 2003). Thus, the observed CaCO3 decrease probably indicates an increased terrigenous dilution during the early- to mid-Holocene. Overall, the CaCO3 content was rather high (avg. 61.0%) compared to those (0–50%) from the surrounding basin of the BOB (Kolla et al. 1976). This indicates that the ISM-related river discharge is not particularly distinctive in the bulk composition of HI1710-MC1 throughout the study period. The other biogenic components, the TOC and TN contents, were very low (< 0.5%) and decreased downcore (Fig. 3c). This tendency is not attributed to the paleoproductivity change, but to the decay of organic material (i.e., remineralization) after burial. The strong correlation of the TOC with the TN (r = 0.88, p < 0.001) suggests their common origin. The TOC/TN (C/N) ratios were lower than ten throughout the core, indicating the aquatic (marine) origin of the organic matter (Meyers 1994).

The small influence of terrigenous dilution on the CaCO3 variation, the low C/N ratio of organic material, and the low linear sedimentation rate (ca. 2.4 cm/kyr) imply that there is only a minor control of river discharge from the regional surface ocean environment, which is also supported by the small seasonal variability of sea surface salinity between 33.5 and 34.0 psu (Fig. 2). Bolton et al. (2013) found the diminished upper water column stratification at the NER site (ODP site 758) during the ISM strengthening events with an analysis of a 500,000-year foraminifera oxygen isotope record. Such phenomenon in the central BOB does not correspond to the enhanced surface water stratification in the northern BOB and Andaman Sea sites, indicated by the approximated decrease in sea surface salinity, during the strengthening of ISM in the mid-Holocene (Gebregiorgis et al. 2016; Sijinkumar et al. 2016). Our results support Bolton et al. (2013)’s suggestion that the large-scale surface wind system and ocean dynamics, rather than regional land–ocean interaction represented by freshwater discharge, largely regulates the surface ocean environment at the NER.

5.2 Detrital Provenance During the Holocene: 1. Sr–Nd Isotope

The Sr–Nd radiogenic isotope compositions of detrital sediments are useful indicators of sediment provenance (e.g., Nakai et al. 1993; Chen et al. 2007; Chen and Li 2013). Although the Himalayas produces an enormous amount of detritus, which is supplied to the entire BOB, the geology of catchment areas of individual river systems is different from one another. This results in the regional variation of BOB detrital sediment in their composition. For instance, the Ganges–Brahmaputra-lower Meghna (G-B-M) rivers mainly drain the Himalayas, whereas the Irrawaddy and Salween (I-S) rivers incorporate detritus from the Indo-Burma Ranges (IBR), Myanmar Central Basin, and Shan Plateau. According to the earlier work by Colin et al. (1999), the Eastern Indian Ocean was divided into four provinces based on the detrital Sr–Nd and clay mineralogy: the northern BOB, the western BOB, the eastern BOB, and the Andaman Sea (Figs. 1 and 4) (Colin et al. 1999; Gourlan et al. 2010; Tripathy et al. 2011; Ali et al. 2021). The western BOB sediments, including the samples from the levee of an active channel of the Bengal Fan (Tripathy et al. 2011; Lupker et al. 2013), are characterized by Nd–Sr isotope composition similar to that of the Lower Meghna river developed downstream of the Ganges and Brahmaputra river confluence (Fig. 4). This suggests the dominance of detritus supplied by these rivers in the western BOB. The western BOB is also fed by the peninsular rivers of India draining into the Archean Shield (e.g., Godavari and Krishna) represented by εNd and 87Sr/86Sr signatures similar to G-B-M rivers (Fig. 4). However, sediment discharge of these rivers is much smaller (ca. 10%) compared to G-B-M rivers (Mouyen et al. 2018). The influence of sediments from the mainland India is considered insignificant in the entire western BOB, as also supported by clay mineral composition discussed in the next section.

Fig. 4
figure 4

Crossplot for 87Sr/86Sr and εNd of lithogenic component of HI1710-MC1. Compositions of surface sediments, including Holocene sediment in Bay of Bengal and Andaman Sea are also shown (Colin et al. 1999; Ahmad et al. 2009; Tripathy et al. 2011; Lupker et al. 2013; Ali et al. 2015; Joussain et al. 2016; Miriyala et al. 2017) (Supplementary Table A1). Composition of the bulk detrital (Gourlan et al., 2010) and clay (< 2 µm) fraction from the uppermost interval of the ODP Site 758 are shown as open and closed squares, respectively (Ali et al. 2021; Song et al. 2021). The average compositions of major river sediments and possible source regions (shaded area) are shown together and provided as supplementary data (Galy and France-Lanord 2001; Singh and France-Lanord 2002; Allen et al. 2008; Najman et al. 2008; Singh et al. 2008; Gourlan et al. 2010; Licht et al. 2013; Giosan et al. 2018; Ali et al. 2021)

Full size image

The eastern BOB and the Andaman Sea sediments are characterized by more radiogenic Nd and less radiogenic Sr isotope signatures than those of the western BOB sediments (Fig. 4). This was attributed to the input of IBR sediments with radiogenic Nd and unradiogenic Sr signatures that were discharged into the Andaman Sea shelf and the easternmost part of the BOB via the I-S rivers (Colin et al. 1999; Joussain et al. 2016; Ali et al. 2021; Bretschneider et al. 2021). While the I-S rivers are known to discharge mostly in the Andaman Sea (Colin et al. 1999; Ali et al. 2015; Lee et al. 2020), the appearance of intermediate Nd–Sr isotope compositions, between the Andaman Sea and the western BOB sediments, suggests significant input of sediments from the I-S rivers to the eastern BOB (Fig. 4). Even the northern BOB sediments, sampled near the G-B-M river mouth, reveal isotopic compositions between the eastern BOB and Meghna river sediments (Fig. 4). This suggests that the I-S and IBR components have influence in the wide region of the eastern BOB.

The Sr and Nd isotope compositions of HI1710-MC1 are similar to those of northern BOB sediment (Fig. 4), an indication of their common source area, which is primarily the High and Tethys Himalaya and secondarily the IBR regions. The heavy monsoonal precipitation on the eastern Himalaya and IBR regions allows intense sediment discharge through I-S rivers, which is comparable (~ 70%) to those of G-B-M river system (Chapman et al. 2015). Although most of them would be primarily deposited into the Andaman Sea basin, a significant amount of I-S river sediments is transported to the eastern BOB (Ali et al. 2021; Bretschneider et al. 2021). Ours and the northern BOB samples had less radiogenic εNd and more radiogenic 87Sr/86Sr than those of the eastern BOB sediments, which indicate the smaller contribution of the I-S river sediments in the remote area from the Irrawaddy river mouth. Our Nd–Sr isotope data suggest that the G-B-M river system with secondary input from I-S river system strongly influences the northern and eastern BOB, including the NER.

Eolian mineral dust and volcanic materials might have been supplied to the study site as minor component. The modern dust deposition in the southern BOB is estimated as 0.1–0.2 g cm−2 ky−1 by dust models (Jickells et al. 2005), which seems significant compared to the total lithogenic flux of 0.2–0.8 g cm−2 ky−1 during the Late Quaternary measured in the deep-sea sediments at the adjacent ODP Site 758 (5° 23.05′ N, 90° 21.67′ E; Hovan and Rea 1992; Song et al. 2021). Dust supplied to the surface ocean is, however, subjected to the various internal processes (e.g., biological scavenging, chemical degradation, physical advection, etc.), so as not to be fully exported to the bottom. For instance, the measured inorganic silicate deposition flux at the northeastern Pacific site (KODOS 02-01-02, 16° 12′ N, 125° 59′ W), where the terrigenous mineral transport is nearly dominated by the eolian process, was 0.012 g cm−2 ky−1 (Hyeong et al. 2006). This is significantly below the model estimation of 0.02–0.05 g cm−2 ky−1 (Jickells et al. 2005). In this context, the actual amount of dust deposition at the study site is likely to be much lower than the model estimation. The amount of volcanic material from the adjacent Sunda arc would be very minor in the NER as well, as volcanic detritus is rarely found, even at the arc-proximal Nicobar Fan sites (IODP Site U1480 and U1481; see locations in Fig. 1) (Pickering et al. 2020). Given the tremendous amount of sediment input from the Himalaya, the BOB is primarily filled by riverine fluxes from the adjacent landmasses and the following hemipelagic processes, including deep-sea current, suspension-settling, etc.

5.3 Detrital Provenance During the Holocene: 2. Clay Mineralogy

Clay mineral assemblage can be used to estimate the relative contribution of sediment supplied by major rivers. The northern and eastern BOB surface sediments were plotted into two groups (EBOB-1 and EBOB-2) with distinct clay mineral compositions (Fig. 5). The clay mineral composition of the EBOB-2 group (Li et al. 2017, 2018; Sun et al. 2020) was estimated by applying the weighting factors of Biscaye (1965) for different minerals. All the samples in the EBOB-2 group plotted close to our data (Fig. 5), which is indicative of their common provenance. The clay mineral composition of another group (EBOB-1) that included the northern BOB sediments was estimated without the weighting factors of Biscaye (1965), and could not be compared with our data directly.

Fig. 5
figure 5

Ternary diagram for clay mineral assemblages of HI1710-MC1 (solid diamonds) plotted together with the average compositions of river mouth and Bay of Bengal sediments. Irrawaddy: Irr, blue cross (Rodolfo 1969); Ganges–Brahmaputra-lower Meghna: G-B-M (Khan et al. 2019), red cross; Godavari-Krishna: G-K, yellow cross (Bejugam and Nayak 2017); eastern Bay of Bengal (BOB) surface sediment: green shade and green circles labelled as EBOB-1 (Colin et al. 1999; Joussain et al. 2016) and orange shade and triangles labelled as EBOB-2 and triangles (Li et al. 2017, 2018; Sun et al. 2020). Note the composition of ODP Site 758 sediment (open circle) (Ali et al. 2021) included in the EBOB-1 group (See Fig. 1 for sampling locations)

Full size image

In the BOB, dominant clay minerals are illite and smectite with minor contribution of chlorite and kaolinite. The major sources of illite and smectite are the G-B-M and I-S river systems, respectively (Colin et al. 1999; Joussain et al. 2016; Khan et al. 2019; Sun et al. 2020). The G-B-M river system is the sole known source dominated by illite (Table 3). The dominance of illite over smectite in HI1710-MC1 sediment thus indicates a significant influence of the G-B-M rivers. If we assume G-B-M rivers and I-S rivers as the two major detrital sources, the clay mineral composition of HI1710-MC1 site can be explained with the mixing of ~ 70% of G-B-M river and ~ 30% I-S river sediments (Fig. 5), which agrees with the estimation based on Sr–Nd isotope compositions. However, the kaolinite/chlorite ratio (K/C) of 0.9 ~ 1.3 in our samples is not explained with mixing of two sources of G-B-M and I-S rivers with an average K/C ratio of 0.4 and 0.8, respectively, which indicates a minor influence of G-K river sediments with an average K/C ratio of 5.6 (Table 3). Thus we cannot preclude influence of G-K river sediments that cannot be deciphered with Nd–Sr isotope composition because of similar composition between G-B-M and G-K river sediments.

Table 3 Average clay mineral compositions of major rivers surrounding the Bay of Bengal (data
Full size table

The Nd–Sr isotopic and clay mineral compositions provide mutually supportive source regime of the detrital components, mixing of G-B-M rivers (~ 70%) and I-S river sediments (~ 30%), but minor input of G-K sediments indicated in the K/C ratio. The clay mineral composition reported at the neighboring ODP Site 758 on the NER is notable (Ali et al. 2021). At this site, smectite has been the predominant clay mineral (39 ~ 69%) for the last 27 Myrs, which is significantly different from those (23% of smectite and 50% of illite on average) of the study site. Ali et al. (2021) did not, however, apply the weighting factors of Biscaye (1965) for semi-quantitative estimation. It means that the smectite content was overestimated in Ali et al. (2021) like the other data in the EBOB-1 group. Nevertheless, the Nd–Sr isotope compositions of bulk detrital component at ODP 758 site showed excellent agreement with ours (Fig. 4) and support our conclusions on the detrital provenance of the HI1710-MC1 site. Our consistent results suggest that the study site at the NER can serve as an archive for the paleoceanographic studies on the Indian monsoon system, especially, in terms of relative changes in rainfall intensity in the G-B-M rivers and I-S rivers drainage areas, as the site receives its detrital components from these two river systems with different Nd–Sr isotopic and clay mineral compositions.

6 Conclusions

The surface ocean production and detrital provenance of the last 13 kyrs were investigated using a 35 cm-long multiple core retrieved at the northern Ninetyeast Ridge in the southern Bay of Bengal (BOB). At the study site, surface ocean environment and detrital provenance have not responded sensitively to the Indian Summer Monsoon and largely remained stable for the last 13 kyrs. The small decrease in CaCO3 content was attributed to the strengthened ISM during the mid-Holocene, though not prominent as previously reported in the continental margin sites of the BOB. Radiogenic Sr–Nd isotope and clay mineral compositions indicate two major sources of detrital sediments for the last 13 kyrs: primarily from the Himalaya transported via Ganges–Brahmaputra-Meghna rivers (~ 70%) and secondly from the Indo-Burman Ridges region via Irrawaddy-Salween rivers (~ 30%). This Holocene record provides a basis for the study of Late Quaternary variability in Indian monsoon precipitation pattern, and resultant runoff to the BOB.