Palaeogeography, Palaeoclimatology, Palaeoecology 397 (2014) 20–30

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology

j ourna l homepage: www.e lsev ie r .com/ locate /pa laeo

Major and trace element geochemistry of Bay of Bengal sediments:Implications to provenances and their controlling factors

Gyana Ranjan Tripathy a,b,⁎, Sunil Kumar Singh a, V. Ramaswamy b

a Geosciences Division, Physical Research Laboratory, Ahmedabad 380009, Indiab National Institute of Oceanography, Dona Paula, Goa 403004, India

⁎ Corresponding author at: Present address. AIRIE ProgrColorado State University, CO 80523, USA.

E-mail address: grtripathy@gmail.com (G.R. Tripathy

0031-0182/$ – see front matter © 2013 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.palaeo.2013.04.012

a b s t r a c t

a r t i c l e i n f o

Article history:Received 21 May 2012Received in revised form 28 February 2013Accepted 14 April 2013Available online 22 April 2013

Keywords:Bay of Bengal sedimentsGeochemistryProvenancesHimalayaDeccan TrapsClimate

Major and trace elements of sediments from a ~13-m long piston core (SK187/PC33) from the western Bay ofBengal have been investigated in this study to infer about the changes in provenances and related controllingfactors during the last glacial–interglacial period. Factor analysis of these geochemical dataset ascertainsdominant role of riverine supply of sediments in regulating the geochemistry of SK187/PC33 sediments.The Al-normalized major (K and Ti) and trace elemental (Cu and Cr) ratios of these marine sediments fallwithin the ranges observed for their major provenances, viz. sediments from the Ganga, Brahmaputra andthe Godavari–Krishna (GK) rivers and depth profiles of these ratios showed significant variations withsynchronous excursion at around the last glacial maxima (LGM), implying a change in their provenanceswith relatively reduced Himalayan contribution during this climatic event. Inverse model calculation ofAl-normalized elemental ratios of the sediments estimated an average sediment contribution of 66 ± 13%and 34 ± 13% from the Himalayan and the peninsular Indian rivers to the core site respectively. Consistentwith the depth profiles of elemental ratios, the estimated sediment contributions from the Himalayan riversare observed to decrease by ~30% from the Himalayan rivers, particularly that of the Ganga, to the studiedlocation during the LGM. Lowering of sediment supply from the Himalaya (Ganga) during the LGM is dueto weakening of south-west monsoon and reduction in available exposure area for weathering due to extentof glacier cover. Outcomes of this study underscore the strong linkage between erosion and climate.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Sediments of the western Bay of Bengal (BoB) are supplied mainlyby the major rivers draining the Himalaya (e.g. the Brahmaputra andthe Ganga) and the peninsular India (e.g. the Godavari, Krishna andthe Mahanadi). Presently, these rivers together supply about1350 million tons of sediments per year to the BoB, which accounts for~8% of the total riverine supply to the oceans (Milliman and Syvitski,1992; Milliman, 2001). This sediment load is disproportionally (~4times) higher compared to the fraction of areal coverage of these riversto the global drainage area underscoring the important role of physicalerosion in these river basins on the global sedimentary budget. Further,the erosion patterns of these river basins and their spatio-temporal var-iations have received significant attention due to their key role in theglobal carbon budget (Galy and France-Lanord, 1999; Dalai et al., 2002;Das et al., 2005; Singh et al., 2005; Rahaman et al., 2009; Huh, 2010;Krishnaswami and Tripathy, 2012). Researches based on sediment fluxand their chemical and isotopic properties have demonstrated that theongoing physical erosion in these river basins is by and large regulated

am, Department of Geosciences,

).

rights reserved.

by climate (rainfall) over millennium timescale (Islam et al., 1999;Singh and France-Lanord, 2002; Goodbred, 2003; Singh et al., 2008;Chakrapani and Saini, 2009; Panda et al., 2011). It has also been docu-mented that the climate over this region has varied appreciably over katimescale (Duplessy, 1982; Prell and Kutzbach, 1987; Sarkar et al.,1990; Tiwari et al., 2005). It is therefore, expected that physical erosionpattern over these river basins could also have varied in the past in re-sponse to these climate changes. Sediment archived in the western BoBcan provide useful information on the paleo-erosion pattern of theseriver basins and therefore on climate–erosion link, which continues tobe a topic of debate (Tripathy et al., 2011a).

Existing studies on the reconstruction of erosion pattern of river ba-sins from the Himalaya and the peninsular India over ka timescale haveled to diverging views. Temporal variations in the sedimentological andSr–Nd isotopic properties of the BoB sediments indicated relative changesin the supply of sediments fromdifferent sourceswhich have been attrib-uted to variability in monsoon intensity in the past (Ahmad et al., 2005;Kessarkar et al., 2005). Prakashbabu et al. (2010), based on down-corevariations in the granulometric, geochemical and mineral magnetism ina core from the eastern BoB observed a shift in sediment provenancefrom the Brahmaputra to the Ganga since early Holocene which hasbeen attributed to climatic changes. More recently, Sr–Nd isotopic studyof the Bay of Bengal sediments by Tripathy et al. (2011b) observed

21G.R. Tripathy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 397 (2014) 20–30

reduced sediment supply from the Himalaya to the western BoB duringlast glacial maximum (LGM) due to weakening of south-west monsoon(Prell and Kutzbach, 1987) and larger glacier cover over the HigherHimalaya (HH) (Owen et al., 2002); all these studies providing additionalsupport for the climate–erosion link. The hypothesis of climate–erosionlink was further strengthened by the studies of Clift et al. (2008) andRahaman et al. (2009), both of which showed varying erosion rate overthe Higher and the Lesser Himalaya in response to climatic changes dur-ing last tens of ka. In contrast to these results, there are also lines of evi-dence that suggest lack of major changes in sediment provenanceduring climatic events, challenging the climate–erosion relation. For ex-ample, Galy et al. (2008) based on Sr–Nd isotopic composition of BoB sed-iments showed no systematic change in sediment source despite changesin climate. Similar studies on tracking provenances of sediments from theEastern Arabian Sea also have not indicated anymajor changes in supplyfrom the peninsular Indian river basins over a ka timescale (Kessarkar etal., 2003; Goswami et al., 2012). In order to evaluate the impact of climatechange on the erosion patterns of the Himalaya and peninsular Indianriver basins, it would be interesting to compare the temporal variationsin relative sediment contribution from these rivers to the western BoB.

Earlier studies on the provenances of BoB sediments reliedmainly ontheir clay mineral and Sr–Nd isotopic composition. Prior to this, therehave been a few studies on the chemical composition of BoB sediments(Sarin et al., 1979; Colin et al., 2006; Prakashbabu et al., 2010), whichprovide useful information about both chemical and physical erosionpatterns of the river basins. The present study attempts to track theprovenances of sediments from the western Bay of Bengal usingdown-core variations of their major and trace elemental compositionand to reconstruct the paleo-erosion pattern of the river basins drainingthe Himalaya and peninsular India. Further, this study employsan integrated approach of factor analysis and inverse modeling to

Fig. 1. Location of the sediment core SK187/PC33 in the western BoB. The major sources ofDeccan Traps and alluvium (Godavari–Krishna). These basins are also shown.

quantitatively estimate the sediment contribution from various sourcesto the core site and their temporal variations. The results of this study ontemporal variations in provenances are used to assess the role of climatein regulating the erosional pattern of the river basins of the Himalayaand peninsular India.

2. Material and methods

The sediments analyzed in this study are from a ~13-m long pistoncore (SK187/PC33; Tripathy et al., 2011b) raised from the western BoB(Fig. 1; 16° 16′N, 84° 30′ E; water depth: 3003 m) during the 187th ex-pedition of ORV Sagar Kanya. This core was examined onboard for someof its physical properties (P-wave velocity, bulk density and gamma rayattenuation) using a Geotek®multi-sensor core logger. Thesemeasure-ments indicate that the core is turbidite-free and preserves a continuousrecord of sedimentation. The core was sub-sampled at 1–2 cm intervaland stored at the National Institute of Oceanography core repository.Chemical analysis and Sr–Nd isotopic composition of the sedimentswere carried out at an interval of ~50 cm, at the Physical Research Lab-oratory, India. Tripathy et al. (2011b) in an early study reported thedepth–age model, magnetic susceptibility and Sr–Nd isotopic composi-tion of these sediments. Present study is limited to major and trace ele-ment composition of the sediment core and its potential as a tool toreconstruct the paleo-erosion pattern of the river basins supplying sed-iments to the core-site. The depth–age model for the SK187/PC33 wasdeveloped based on the 14C ages of inorganic carbon (CaCO3) and mag-netic susceptibility stratigraphy. The 14C depth–age model was devel-oped after critically evaluating the impact of contribution from detritalcarbonates (Tripathy et al., 2011b). This model suggests that sedimentsdeposited in the depth interval 600–700 cmbsf corresponds to the LGM.The sedimentation rate of the core shows significant variation with

sediments to this location are the rivers draining the Himalaya (Ganga–Brahmaputra),

22 G.R. Tripathy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 397 (2014) 20–30

depth, the rate is lower during the LGM (0.2 mm/yr) compared to thatafter this climatic event (2.7 mm/yr).

The sub-samples were powdered using an agate mortar, washedwith de-ionized water to remove the sea-salts and ashed at 600 °C tooxidize organic matter. The sea-salt and organic-matter free sedimentswere dissolved completely with repeated acid digestion using HF–HNO3–HCl. These solutions of sediments were used to measure theirmajor and trace elemental concentrations. Na and K were measuredusing Flame AAS, whereas the concentrations of other elements (Ca,Mg, Al, Fe, Ti, V, Ba, Cr, Sr, Mn, Cu, Co, Ni and Zn) were analyzed usingICP-AES (Jobin-Yvon, Ultima). Among these, the Al, Fe and V concentra-tions have already been reported by Tripathy et al. (2011b). Totalcarbon and nitrogen in the sediments (at ~10 cm interval) wereanalyzed using CN analyzer (FISONS NA1500) using the method ofBhushan et al. (2001), whereas carbonate contents were determinedusing a Coulometer (Model 5012, UIC Inc., Illinois, USA). Organic carbon(Corg) was taken as difference betweenmeasured total carbon and inor-ganic carbon (Cinorg) concentrations. A few samples were analyzed inreplicate to determine the precision of measurements; these resultsyield coefficient of variation of about 5% or better for various elements.Major and trace element compositions of USGS G-2 reference standardwere also measured during the course of the study to check on the ac-curacy of analyses; this was found to be better than ±10% for most ofthe elements (Tripathy, 2011).

3. Results

3.1. Major and trace elements

The concentrations of major and trace elements in sediments ofthe SK187/PC33 core are given in Table 1. The major elements, Na(0.54–2.39 wt.%), K (1.03–3.08 wt.%), Ca (0.74–29.37 wt.%), Mg(0.80–3.39 wt.%), Al (2.50–9.19 wt.%), Fe (1.84–6.54 wt.%), Ti (0.16–

Table 1Major and trace elemental composition of sediments from the SK187/PC33 core.

Sample ID Na K Ca Mg Ala Fea Ti

(wt.%)

10–12 cm 1.14 1.95 0.97 1.93 8.96 6.54 0.5718–20 cm 1.05 2.19 1.37 1.29 7.44 4.05 0.5628–30 cm 1.12 2.20 1.16 1.98 8.83 6.42 0.57109–111 cm 1.03 2.14 0.74 1.64 7.95 6.22 0.55149–151 cm 1.13 2.43 0.87 1.80 8.85 5.63 0.55201–203 cm 1.08 2.49 1.96 1.69 8.47 4.88 0.45221–223 cm 0.54 1.07 22.7 1.15 2.50 2.17 0.37249–251 cm 1.16 2.33 1.64 1.80 8.82 5.78 0.50300–302 cm 0.97 1.99 2.31 1.62 7.89 5.57 0.45320–322 cm 1.10 1.88 13.4 2.59 6.28 3.97 0.34350–352 cm 1.29 2.70 7.48 3.37 8.63 4.69 0.38400–402 cm 1.36 2.07 10.8 2.68 6.41 3.87 0.36440–442 cm 0.74 1.03 29.4 1.62 2.84 1.84 0.16450–452 cm 0.99 2.43 4.25 1.66 8.27 4.97 0.44500–502 cm 1.02 3.08 2.75 1.94 8.22 4.47 0.50550–552 cm 1.68 2.03 6.11 2.10 6.15 2.52 0.27600–602 cm 1.42 2.04 11.5 3.36 6.80 4.42 0.50650–652 cm 0.97 2.01 6.24 2.11 7.30 4.88 0.41700–702 cm 1.10 2.15 4.62 2.11 7.77 4.84 0.48750–752 cm 1.46 2.40 2.87 1.70 7.75 3.84 0.38800–802 cm 1.35 2.42 8.12 3.39 8.13 4.86 0.31849–851 cm 1.11 2.05 2.92 1.54 6.81 3.57 0.40899–901 cm 1.02 1.69 16.2 2.60 5.57 3.69 0.29949–951 cm 1.06 2.52 2.59 1.78 7.87 4.82 0.39999–1001 cm 0.84 1.57 4.79 1.70 4.55 2.63 0.211049–1051 cm 1.16 2.67 2.60 1.70 8.07 4.46 0.381099–1101 cm 1.18 3.03 2.40 1.85 8.72 4.98 0.431149–1151 cm 1.65 2.61 2.25 1.45 7.55 3.68 0.421199–1201 cm 2.39 1.97 1.89 0.80 6.00 1.86 0.221219–1221 cm 1.06 2.72 2.25 1.94 9.19 4.87 0.43

a Al, Fe and V data are from Tripathy et al. (2011b).

0.57 wt.%) andMn (0.03–1.09 wt.%), show significant down-core var-iations in the core. The average carbonate-free concentration of majorelements (excluding Ca and Mn) in the core, as expected, is withinthe range of values (Table 2) reported for its possible sources, viz. sed-iments from the Ganga, Brahmaputra and the GK Rivers. The Ca con-centration (0.74–29.4 wt.%) in some of the bulk sediment sections ishigher compared to that reported for the Himalayan and the peninsu-lar river sediments (Table 2) and is attributable to supply of CaCO3

from marine biogenic sources. The carbonate-free concentrations ofMn, Sr and Ba are not calculated as these elements are known to beassociated with carbonates in sea sediments (Morse and Mackenzie,1990). Carbonate-free concentrations of trace elements also fall with-in range of values reported for the Ganga, Brahmaputra and the GKsediments (Table 2), excluding Zn which seems to have marginallyhigher concentration in the SK187/PC33 sediments. The cause for en-richment of Zn in the SK187/PC33 core is unclear, however, the ob-served good correlation (r ~0.6; excluding two outliers) of Zn withCinorg indicates possible association of Zn with biogenic carbonates(Boyle, 1981).

3.2. CaCO3, organic carbon (Corg) and total N contents and Corg/N ratio

The concentrations of CaCO3, Corg, and total nitrogen, and Corg/Nratios with depth in the SK187/PC33 core are listed in Table 3. The cal-cium carbonate content of the sediments shows a wide range from0.2 to 54 wt.% with majority of samples (97 out of 123) havingb10 wt.%. This temporal trend in CaCO3 (Fig. 2) is attributable to var-iations in the relatively supply of CaCO3 from detrital and biogeniccarbonates. The mean CaCO3 content for the SK187/PC33 sediments(~7 wt.%) is higher compared to that of sediments of the peninsularIndian rivers (~3 wt.%; Das and Krishnaswami, 2007) and the LowerMeghna (~3 wt.%; Galy and France-Lanord, 2001), combined flow ofthe Ganga and Brahmaputra river system in Bangladesh. Two samples

Ba Cr Mn Sr Co Cu Ni Zn Va

(μg/g)

358 117 10946 97 32 66 75 152 130407 96 1190 111 21 31 48 140 90401 117 1400 105 30 62 69 156 125385 124 1727 93 32 70 89 130 119461 120 1351 105 27 57 77 120 115449 109 856 133 22 45 61 138 106239 35 2361 2749 16 8 32 56 30559 110 977 133 26 66 79 132 112533 103 1014 150 26 59 78 124 108363 85 1685 531 23 36 68 397 77365 86 1689 500 22 44 63 546 78520 95 1572 186 24 40 79 461 78231 37 3111 4159 10 16 28 250 34383 102 1648 262 25 44 68 104 105463 89 941 123 25 36 64 99 99365 60 662 213 16 16 39 333 49409 107 1589 434 26 41 70 545 104551 90 2543 318 33 56 87 188 115489 88 1782 233 33 53 79 158 112429 67 1440 173 21 27 45 110 83406 103 1926 254 29 40 77 467 89401 71 1765 187 21 25 46 144 79365 66 1396 886 24 37 62 418 56497 84 1217 150 33 44 68 114 103338 50 1096 170 15 27 38 278 53437 77 883 144 22 37 43 97 94487 83 1026 153 24 41 53 109 103409 61 615 147 19 24 37 83 73340 26 284 156 9 7 15 93 35434 83 753 123 22 42 47 160 96

Table 2Comparison of average chemical compositions (±1σ) of sediments of the SK187/PC33 core with their major provenances, viz. sediments from the Himalayan (Ganga andBrahamaputra) and peninsular India rivers. The numbers in the parentheses indicate the Al-normalized ratio for the element.

Gangaa Brahmaputraa Godavari–Krishnab SK187/PC33d

Total Carbonate-corrected

Al (wt.%) 7.9 ± 2.2 (1) 8.0 ± 1.1 (1) 8.0 ± 0.4 (1) 7.9 ± 2.2 (1) 8.0 ± 1.2Fe 4.6 ± 2.0 (0.57 ± 0.11) 4.1 ± 0.7 (0.51 ± 0.02) 6.4 ± 0.3 (0.80 ± 0.05) 4.6 ± 2.0 (0.60 ± 0.11) 4.8 ± 1.1Ti 0.5 ± 0.1 (0.06 ± 0.02) 0.4 ± 0.5 (0.054 ± 0.004) 0.8 ± 0.1 (0.10 ± 0.01) 0.5 ± 0.1 (0.06 ± 0.02) 0.5 ± 0.1Mg 1.5 ± 0.5 (0.18 ± 0.02) 1.6 ± 0.4 (0.19 ± 0.03) 1.5 ± 0.1 (0.19 ± 0.02) 1.5 ± 0.5 (0.29 ± 0.12) 2.2 ± 0.8K 2.3 ± 0.5 (0.30 ± 0.06) 2.6 ± 0.3 (0.32 ± 0.02) 1.7 ± 0.01 (0.21 ± 0.01) 2.3 ± 0.5 (0.31 ± 0.04) 2.4 ± 0.4Ca 2.1 ± 1.0 (0.30 ± 0.18) 1.4 ± 0.2 (0.18 ± 0.04) 2.0 ± 0.3 (0.25 ± 0.04) 2.1 ± 1.0 (1.27 ± 2.40) 1.9 ± 2.2e

Na 0.7 ± 0.3 (5.34 ± 1.52) 1.3 ± 0.1 (5.13 ± 0.22) 0.6 ± 0.3 (0.08 ± 0.04)c 0.7 ± 0.3 (0.17 ± 0.06) 1.3 ± 0.3Ba μg/g 448 ± 107 (59 ± 14) 510 ± 51 (64 ± 4) 205 ± 3 (26 ± 1) 448 ± 107 (59 ± 12) –

V 100 ± 46 (12 ± 4) 93 ± 19 (12 ± 1) 192 ± 23 (24 ± 3) 100 ± 46 (12 ± 2) 97 ± 25Cr 95 ± 35 (11 ± 3) 102 ± 33 (13 ± 3) 129 ± 13 (16 ± 2) 95 ± 35 (12 ± 2) 94 ± 24Cu 40 ± 29 (4 ± 2) 32 ± 17 (4 ± 2) 91 ± 20 (11 ± 3) 40 ± 29 (5 ± 2) 44 ± 16Ni 47 ± 25 (6 ± 2) 49 ± 19 (6 ± 2) 84 ± 9 (11 ± 1) 47 ± 25 (8 ± 2) 67 ± 20Zn 90 ± 34 (11 ± 2) 86 ± 20 (11 ± 2) 151 ± 25 (19 ± 3) 90 ± 34 (32 ± 25) 245 ± 184Sr 88 ± 23 (12 ± 5) 136 ± 19 (17 ± 5) 140 ± 43 (18 ± 5) 88 ± 23 (115 ± 322) –

Co 15 ± 6 (1.8 ± 0.6) 16 ± 4 (2.0 ± 0.2) 30 ± 3 (3.8 ± 0.4) 15 ± 6 (3.3 ± 0.9) 26 ± 6Mn 908 ± 248 (121 ± 48) 1000 (127 ± 18) 800 ± 100 (100 ± 13) 908 ± 248 (267 ± 289) –

Errors here represent variability in samples from the corresponding river basin.a Garzanti et al. (2011).b Pattan et al. (2008).c Das and Krishnaswami (2007).d Present study.e Total Ca–Ca from carbonates.

23G.R. Tripathy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 397 (2014) 20–30

of SK187/PC33 (221–223 and 440–442 cm) have anomalously highCaCO3 (~50 wt.%). It is pertinent to mention here that the Sr and Ndisotopic composition for these two samples also show anomalous sig-natures (Tripathy et al., 2011b) that hint a unique flood event.

The Corg and total nitrogen for these sediments range from 0.1 to7 wt.% and 0.01 to 0.4 wt.% respectively. Except for three outliers (330–332 cm, 440–442 cm and 480–482 cm), the samples have b2 wt.% ofCorg with an average of 0.8 ± 0.4 wt.%, consistent with previouslyreported organic carbon content in BoB sediments (Galy et al., 2008).The average Corg concentration of the SK187/PC33 sediments is interme-diate to those of the river sediments of Lower Meghna (0.6 wt.%; Aucouret al., 2006) and Godavari (2.5 wt.%; Gupta et al., 1997), possibly hintingat dominancy of terrigenous organic carbon at this core site. The averagetotal nitrogen content of the sediments analyzed in this study is 0.1 wt.%.The Corg/N (atomic ratio) values of the sediments show awide depth var-iation (1.3–31, excluding the two anomalous samples)with amean valueof 14 ± 5. This average Corg/N ratio for the SK187/PC33 sediments iscomparable with that of the river sediments of the Ganga (~12.3), andBrahmaputra (~13.5) at their mouths (Subramanian et al., 1985), buthigher than that at the Godavari outflow (6.7; Balakrishna and Probst,2005). These measured parameters (carbonate, Corg, Corg/N), thoughexhibit depth variations; show no systematic evolution during last gla-cial–interglacial cycle (Fig. 2). The application of the depth profile ofCorg abundance in the SK187/PC33 core to infer the sediment provenanceis limited because ofmultiple sources of carbon to the core site and issuespertaining to their post depositional preservation (Burdige, 2007). Theδ13C of the Corg may provide a better resolution of the provenance. It ishowever, recognized that the use of Corg to infer erosion patterns requiresdistinguishable link betweenCorg anddetrialmaterials, the co-variation ofCorg and Si/Al ratio in the Ganga–Brahmaputra river sediments (Galy etal., 2007) may provide such a link. Information on such trends for thepeninsular Indian rivers is required to further pursue the application ofCorg as a proxy for provenances of BoB sediments.

4. Discussions

4.1. Detrital component

The chemical composition of marine sediments is governed bymixing proportions of supply from the detrital, authigenic (hydrogenous

and biogenic) and hydrothermal sources. In the Bay of Bengal at the coresite, the chemistry of sediments is expected to be dominated by detritaland biogenic components (Kessarkar et al., 2005; Pattan et al., 2008). Theuse of the SK187/PC33 sediment core to retrieve information aboutpaleo-erosion pattern of river basins (Himalaya-vs-peninsular) requiresdata on the composition of detrital material in sediments both from theBay and the river basins (source materials). Further, it is important thatthe different source materials have measurably different compositionand that source signatures are well preserved in the Bay sediments, re-quirements that would provide better estimates of their contributions.Frequently, aluminum (Al), a weathering resistant major element, hasbeen used for reference and normalization to investigate detrital contri-bution of other elements to sediments. The weathering resistant natureof Al is also evident from the low mobilization of Al by rivers such asthe G–B; the dissolved flux of Al transported by the Himalayan rivers isonly b0.1% of its particulate flux (Galy and France-Lanord, 2001).

The chemical composition of bed rocks are likely to alter during theweathering processes in river basins, mainly due to loss of mobile ele-ments (Das and Krishnaswami, 2007) and size sorting (Horowitz andElrick, 1987; Bouchez et al., 2011). On the contrary, the chemical compo-sition of river sediments is likely to remain mostly unaltered in the seaand hence, comparison of chemistry of river sediments with that ofthe marine sediments can provide useful information on temporal vari-ations in their provenances. An assessment of the extent of weatheringloss undergone by detrital components of sediments in the SK187/PC33 can be made by comparing the CIA (Chemical Index of Alteration)values of sediment sections with those of source materials. CIA providesinformation on the intensity of chemical weathering the sediments haveundergone (Nesbitt and Young, 1982) and is calculated as:

CIA ¼ 100� Al2O3

Na2Oþ K2Oþ CaO � þAl2O3

� �ð1Þ

where, the oxide abundances are in molar units and CaO* represents thesilicate derived CaO concentrations in the sediments. CaO* is the differ-ence between the CaO content measured in the total sediments andthat of biogenic origin, the later determined from the concentration of in-organic carbonates. The estimation of biogenic CaO requires knowledgeof the composition of the inorganic carbonates, particularly the presenceof dolomites. In the absence of such information it is customary to use a

Table 3CaCO3, Corg and N concentrations and Corg/N ratio of BoB sediments from the SK187/PC33 core.

Depth(cm)

CaCO3 Corg TN Corg/N Depth(cm)

CaCO3 Corg TN Corg/N

(wt.%) (wt.%)

2–4 3.30 0.70 0.07 10.0 620–622 9.70 0.60 0.06 10.010–12 1.70 1.10 0.12 9.2 630–632 10.4 0.90 0.08 11.318–20 1.50 0.80 0.07 11.4 640–642 6.90 0.30 0.03 10.028–30 1.10 1.00 0.08 12.5 650–652 14.1 1.10 0.09 12.238–40 2.10 0.80 0.06 13.3 660–662 10.1 1.20 0.12 10.048–50 3.00 0.30 b.d. 670–672 7.00 0.60 0.05 12.058–60 2.40 0.70 0.05 14.0 680–682 11.2 0.80 0.07 11.468–70 2.80 0.70 0.07 10.0 690–692 6.10 0.40 0.03 13.378–80 3.70 0.50 0.04 12.5 700–702 10.2 0.90 0.09 10.089–91 2.20 1.20 0.13 9.2 710–712 6.00 – – –

109–111 1.20 1.30 0.12 10.8 720–722 4.90 0.10 b.d. –

111–113 1.50 1.50 0.13 11.5 730–732 5.70 – – –

121–123 1.10 1.10 0.07 15.7 740–742 8.50 – – –

129–131 2.50 0.60 0.05 12.0 750–752 5.90 0.50 0.03 16.7141–143 1.20 1.10 0.07 15.7 760–762 7.50 – – –

149–151 1.50 0.10 0.09 1.1 770–772 16.2 – – –

161–163 1.60 1.60 0.42 3.8 780–782 13.7 0.60 0.03 20.0169–171 1.10 1.40 0.13 10.8 790–792 6.20 – – –

181–183 4.90 0.90 0.07 12.9 800–802 8.30 1.10 0.12 9.2189–191 0.90 1.20 0.12 10.0 810–812 8.00 – – –

201–203 3.80 1.20 0.08 15.0 819–821 6.50 1.30 0.11 11.8209–211 19.6 0.50 0.06 8.3 829–831 5.20 – – –

221–223 50.5 0.30 0.02 15.0 839–841 5.20 – – –

229–231 2.20 1.20 0.12 10.0 849–851 5.20 0.60 0.05 12.0241–243 1.10 0.60 0.11 5.5 859–861 5.30 – – –

249–251 3.10 1.30 0.12 10.8 869–871 5.40 – – –

260–262 2.90 1.10 0.09 12.2 879–881 6.40 1.50 0.12 12.5270–272 14.5 0.50 0.04 12.5 889–891 6.90 – – –

280–282 2.40 1.20 0.10 12.0 899–901 27.9 – – –

290–292 3.90 – – – 909–911 7.60 0.30 0.03 10.0300–302 4.90 1.00 0.10 10.0 919–921 10.7 0.30 0.03 10.0310–312 7.70 0.30 0.08 3.8 929–931 4.90 – – –

320–322 23.0 1.10 0.07 15.7 939–941 5.10 – – –

330–332 13.9 2.90 0.23 12.6 949–951 5.70 1.00 0.09 11.1340–342 6.00 0.80 0.04 20.0 959–961 11.4 – – –

350–352 6.00 0.50 0.04 12.5 969–971 8.70 – – –

360–362 6.00 1.20 0.07 17.1 979–981 15.6 – – –

370–372 7.10 0.50 0.05 10.0 989–991 13.1 – – –

380–382 7.40 – – – 999–1001 7.00 1.20 0.08 15.0390–392 6.20 1.50 0.14 10.7 1009–1011 6.70 – – –

400–402 13.8 0.80 0.06 13.3 1019–1021 7.30 0.10 0.01 10.0410–412 5.70 0.60 0.05 12.0 1029–1031 7.30 – – –

420–422 6.00 0.40 0.03 13.3 1039–1041 5.10 – – –

430–432 15.9 0.70 0.06 11.7 1049–1051 6.30 0.40 0.03 13.3440–442 53.9 6.70 0.04 168 1059–1061 5.40 – – –

450–452 10.0 1.30 0.11 11.8 1069–1071 5.30 – – –

460–462 6.60 0.20 0.02 10.0 1079–1081 6.20 0.60 0.04 15.0470–472 8.30 1.10 0.09 12.2 1089–1091 5.50 – – –

480–482 3.00 2.10 0.08 26.3 1099–1101 5.60 0.70 0.04 17.5490–492 2.90 1.70 0.08 21.3 1109–1111 4.80 – – –

500–502 7.10 0.60 b.d. – 1119–1121 5.00 1.00 0.08 12.5510–512 4.50 0.80 0.07 11.4 1129–1131 4.50 – – –

520–522 7.30 1.20 0.08 15.0 1139–1141 4.20 – – –

530–532 0.20 – – – 1149–1151 3.30 0.20 0.03 6.7540–542 12.9 0.80 0.15 5.3 1159–1161 3.30 – – –

550–552 5.20 0.20 b.d. – 1169–1171 2.60 – – –

560–562 7.70 0.90 0.07 12.9 1179–1181 1.60 0.20 – –

570–572 7.30 – – – 1189–1191 1.60 – – –

580–582 12.1 1.10 0.06 18.3 1199–1201 1.70 0.20 – –

590–592 21.6 – – – 1209–1211 4.00 – – –

600–602 14.3 0.80 0.21 3.8 1219–1221 3.90 0.60 0.05 12.0610–612 11.2 0.80 0.06 13.3

b.d.: below detection.

24 G.R. Tripathy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 397 (2014) 20–30

modified version of CIA, designated as CIA* calculated using the followingrelation:

CIA� ¼ 100� Al2O3

Na2Oþ K2Oþ Al2O3

� �: ð2Þ

The CIA* values of G–B river suspendedmatter (Rai, 2008) and thoseof the sediment section from SK187/PC33 are nearly identical, withvalues of 73 ± 2 and 72 ± 4 respectively, but are marginally higherthan that of the source rocks, the crystallines of theHigher (65) and Less-er (58 ± 7) Himalaya (France-Lanord and Derry, 1997; Singh et al.,2008). This is an indication that there is measurable loss of some of thecations during the transformation of Himalayan crystallines tosuspended sediments of the G–B river (Rai, 2008). This is unlike the G–B bank sediments which do not show any discernible weathering loss(Singh et al., 2008). The CIA* for river sediments from the peninsular re-gion showwider range from 80 to 96 for the Krishna basin that predom-inantly drain through the Deccan Traps (Das and Krishnaswami, 2007).If, however, only the sediments from the mainstream of the Krishnaare considered, the average CIA* value comes out to be 87 ± 4 (DasandKrishnaswami, 2007),which is higher than that of theDeccan basalts(~75; Das and Krishnaswami, 2007). It is seen from the comparison ofthe CIA* that these values for the SK187/PC33 sediments (72 ± 4) ex-pectedly lie between the values of the Himalayan (73 ± 2) and theDeccan river (87 ± 4) suspended matter/sediments. This observationsupports that the composition of the source material is preserved inthe BoB sediments and therefore, can be used to track their provenancesand their relative contributions to the core location in the past.

Fig. 3 shows the interrelation of Al with a few selected elements ofSK187/PC33 sediments. The correlation coefficient (r) between Al andthe various elements plotted in Fig. 3 is (~0.7 or higher), statistically sig-nificant to better than 99% confidence level. These r values thoughmostly decrease (>0.6) if the two outliers (221–223 cm and 440–442 cm) are excluded from statistical analysis, they are still significantto 99% confidence level. This observation of strong co-relation betweenAl and various elements in Fig. 3 suggests that the abundances of theseelements in sediments of the SK187/PC33 core site is governed mainlyby their supply from detrital sources and therefore, has the potentialto provide information about changes in the sediment provenances inthe past. Earlier studies by Sarin et al. (1979) also have demonstratedthat elements such as Fe, Mg and Cr are strongly correlated with Alunderscoring the importance of their supply from continental alumino-silicates to the BoB. Further, it is interesting to observe strong correla-tion of Fe with Ti (r = 0.8) and Cr (r = 0.9; Table 4) in the SK187/PC33 sediments. This observation is consistent with the presence ofweathering resistant Fe–Ti minerals, such as rutile and ilmenite, in theriver sediments from the Himalaya (Garzanti et al., 2011) and the Dec-can Traps (Das and Krishnaswami, 2007). Correlation between Fe andCr in the SK187/PC33 is attributable to substitution of trivalent Crwith Fe(III) in the mineral structure (Schwertmann and Cornell,2000). However, the observed significant correlation of Fe in theSK187/PC33 sediments with some other trace elements, e.g. Ni, Cu,and Co (Table 4) is intriguing. One likely explanation for this correlationis substitution of Fe(III) from octahedral position by these cations viaproton capture (Sidhu et al., 1980; Carvalho-e-Silva et al., 2003).Taken together, these observations on Fe, Al and Ti correlations indicatethat these elements are primarily from continental sources.

In order to better constrain the contributing factors towards geo-chemistry of the SK187/PC33 sediments, factor analysis of the geo-chemical dataset (Table 1) have been carried out in this study. Thisanalysis extracted four major factors (Table 5) with Eigen valuesgreater than 1 and these factors together can explain ~85% of the var-iance of the dataset. Among these four factors, the dominating factor(Factor 1) explains ~50% of the total variance and is primarily loadedwith lithogenic elements, e.g. Al, Ti and Fe (Fig. 4). Dominant loadingof elements, e.g. Al and Ti, those frequently reflect the detrital compo-nent (Fig. 4) clearly assigns the Factor 1 as the detrital source. Thisfactor also characterized with higher loading for various trace ele-ments (Ba, Cr, Co, Cu, Ni and V; Table 5) of the SK187/PC33 core, im-plying their dominant supply also through terrigenous sources(Fig. 4). This inference is consistent with strong correlation of thesetrace elements with Al (Fig. 3; Table 4). Higher loading of Fe in the

Fig. 2. Concentration–depth profiles of carbonate, Corg, N and Corg/N of the sediments from the core SK187/PC33. The Corg/N ratio centers around 12 ± 4, with two samples are out-side this range; these samples are not considered for data analysis.

25G.R. Tripathy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 397 (2014) 20–30

Factor 1 with weak loading for Mn ensures that Fe in these marinesediments is primarily of detrital origin. It is interesting to notice thatinsignificant amount of Fe and trace elements in the Bay of Bengal sed-iments are supplied through formation of Fe–Mn oxides, which is animportant geochemical process in the oceans. This observation of min-imal association of trace elements with the Fe–Mn oxides in the west-ern Bay of Bengal is consistent with earlier reported insignificantcorrelation (r = 0.17) between Fe and Mn in the surface sedimentsfrom the Godavari–Krishna river shelf (Pattan et al., 2008). Along withdetrital supply, the geochemistry of the BoB sediments are minutelyalso regulated by carbonates (Factor 2), and organic matters (Factor4). It is unclear to assign a source for ‘Factor 3’which is primarily loadedwith Zn,Mg and K (Table 5). One possible explanation for Factor 3 couldbe their association with micro-nutrients that substitute some of theseelements in the marine sediments (Huang and Conte, 2009). Resultsfrom the factor analysis of the geochemical dataset for the SK187/PC33 core ascertain that most of these elements are primarily suppliedvia their detrital sources (Table 5; Fig. 4).

4.2. Temporal variations of elemental ratios

It is evident from the above discussion that the geochemistry of theSK187/PC33 sediments preserves the signature of their sources andtheir variations in their relative fluxes in the past. Temporal variationsin the sediment sources to the core location can be examined throughchanges in Al-normalized elemental ratios of the sediments. Fig. 5shows the depth profiles of Al-normalized ratios of K, Ti, Cu and Cr,along with the Ti/Ca ratio and CIA* of the SK187/PC33 sediments. Thestrategy for using these Al-normalized elemental ratios to track theprovenances lies in the fact that these ratios are appreciably differentfor the Himalayan (Ganga–Brahmaputra) and the peninsular Indian(Godavari–Krishna) river sediments (Table 2). All these elemental ratiosshow clearly discernible variations with the depth in the SK187/PC33core, attributable to changes in the mixing proportion of sedimentsfrom these two sources in the past. For example, the K/Al varies from0.22 to 0.37 in the SK187/PC33 sediment core; the lower value being

close to theGK sediments and the higher valuewithin errors of G–B sed-iments (Table 2). Besides minor excursions, the K/Al ratios of the sedi-ments corresponding to the LGM show systematically lower values(Fig. 5). A likely cause for the lower K/Al ratios during LGM (Fig. 5) is rel-ative decrease in sediment supply from the Himalayan river sedimentswhich have characteristically higher K/Al ratios compared to GK riversediments. K/Al ratios of sediments are prone to alteration during chem-ical weathering and hence, these observations are further evaluatedthrough various other elemental proxies. The decline in the sedimentsupply from the Himalayan rivers during the LGM also seems to be evi-dent from the relatively higher Ti/Al ratios in the BoB sediments duringthe LGM albeit larger scatter (Fig. 5). The Ti/Al ratios of the Ganga(0.06 ± 0.02) and Brahmaputra (0.054 ± 0.004) river sediments arelower compared to that of the GK river sediments (0.10 ± 0.01). Rela-tively higher Cr/Al and Cu/Al ratios of the SK187/PC33 sediments duringthe LGM (Fig. 5) also lend support to the hypothesis for reduced sedi-ment contribution from the G–B rivers during this climatic event. TheCu/Al and Cr/Al ratios for the Ganga (4 ± 2 and 11 ± 3) and Brahmapu-tra (4 ± 2 and 13 ± 3) river sediments are lower compared to that ofthe GK river sediments (11 ± 3 and 16 ± 2). Taken together, thedepth profile of K/Al, Ti/Al, Cu/Al and Cr/Al for the SK187/PC33 coreseems to indicate that the relative supply of sediments from the G–B re-duced during the last glacialmaxima (Fig. 5). This conclusion is also con-sistentwith the results of Tripathy et al. (2011b), which based on Sr–Ndisotopes of sediments from the western BoB indicated lowering of sedi-ment supply from the Himalaya during LGM.

The depth profile of Ti/Ca ratio for the SK187/PC33 core showed aninteresting shift towards higher values during the Holocene (Fig. 5). Tiinmarine sediments is derived fromdetrital sources,whereas Ca in sed-iments depends on its contributions from detrital and biogenic supply.A critical inspection of depth profile of CaCO3 indicates relativelylower concentration in the SK187/PC33 sediments during the Holocene(Fig. 2), which in turn likely to increase the Ti/Ca ratios in the upper sec-tions of the core. The exact cause for decrease in CaCO3 in the Holoceneis intriguing; however these changes might be linked with change inthe marine productivity pattern during this period. On the contrary to

Fig. 3. Scatter diagram of Al concentration with those of other major and trace elements in the SK187/PC33 core. The r (correlation coefficient) values for these correlations aresignificant to better than 99% confidence level, indicating their dominant supply from detrital sources.

26 G.R. Tripathy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 397 (2014) 20–30

the trend of Ti/Ca ratio, the depth profile of CIA* showed near constanttrend, except an elevated hump immediate after the LGM (Fig. 5). Theaverage CIA* values before the LGM (70 ± 4) overlaps with that afterthe climatic event (73 ± 3).

4.3. Quantification of source contributions using an inverse model

Efforts are made in this study to quantify the contribution fromthe major provenances for the SK187/PC33 sediments using an in-verse model. Inverse modeling of geochemical dataset (Tripathy andSingh, 2010) iteratively solves a set of mass balance equation startingfrom pre-assigned elemental ratios for the sources and it provides therelative contribution from each of the sources to the core site. In thiswork this model is used to apportion the contributions from three riv-erine sources, Ganga, Brahmaputra and GK rivers. For this, mass bal-ance equations for only those elements (viz. Al, Fe, Ti, Cr, Co, Cu andV) that have high (>0.80) factor loading in Factor 1 (Table 5) areused.

The mass balance equation for an element (X) in the BoB sedimentcan be written as:

X ¼X3i¼1

Xi � f ið Þ ð3Þ

and

X3i¼1

f ið Þ ¼ 1 ð4Þ

where, Xi represents for the elemental concentration of source i andi = 1, 2, and 3 stands for the Ganga, Brahmaputra and the GK riversrespectively. fi is the fractional contribution of sediments to the coresite from the source i.

For Al-normalized ratios of X, Eq. (3) can be re-written as,

X.

Al

� �� Al ¼

X3i¼1

X.

Al

� �i� Ali � f i

� �: ð5Þ

As concentrations of Al for all the three sources (Table 2) are sim-ilar, Eq. (5) can be reduced to,

X.

Al

� �¼

X3i¼1

X.

Al

� �i� f i

� �: ð6Þ

Eqs. (4) and (6) for elements X = Fe, Ti, Cr, Co, Cu and V are usedin the present model to estimate the source contributions. Thea-priori compositions for the Al-normalized ratios of the three sources

Table 4Correlation coefficient between chemical constituents of sediments from the SK187/PC33 core. Two anomalous samples (221–223 cm and 440–442 cm) are not considered informing this matrix.

Na K Ca Mg Al Fe Ti Corg Cinorganic Ba Cr Mn Sr V Co Cu Ni Zn

Na 1.00K −0.02 1.00Ca 0.00 −0.40 1.00Mg −0.11 −0.02 0.73 1.00Al −0.22 0.70 −0.56 −0.02 1.00Fe −0.54 0.26 −0.33 0.12 0.79 1.00Ti −0.39 0.30 −0.46 −0.10 0.71 0.81 1.00Corg −0.60 −0.30 0.11 0.14 0.07 0.48 0.22 1.00Cinorganic −0.18 −0.37 0.92 0.54 −0.51 −0.25 −0.38 0.24 1.00Ba −0.28 0.30 −0.18 −0.04 0.37 0.38 0.32 0.17 −0.04 1.00Cr −0.53 0.15 −0.16 0.27 0.64 0.90 0.81 0.51 −0.15 0.34 1.00Mn −0.16 −0.23 −0.07 0.11 0.22 0.42 0.33 0.27 −0.06 −0.17 0.34 1.00Sr −0.05 −0.34 0.88 0.60 −0.43 −0.24 −0.38 0.07 0.87 −0.27 −0.19 −0.05 1.00V −0.55 0.29 −0.40 0.03 0.76 0.94 0.86 0.48 −0.27 0.50 0.87 0.39 −0.35 1.00Co −0.57 0.09 −0.03 0.30 0.50 0.82 0.62 0.52 0.09 0.46 0.75 0.41 −0.02 0.84 1.00Cu −0.59 0.04 −0.20 0.16 0.60 0.94 0.71 0.58 −0.11 0.44 0.86 0.40 −0.13 0.88 0.83 1.00Ni −0.56 −0.05 0.16 0.44 0.37 0.76 0.53 0.60 0.22 0.49 0.83 0.31 0.10 0.73 0.88 0.86 1.00Zn 0.08 −0.31 0.84 0.89 −0.38 −0.21 −0.35 0.06 0.58 −0.26 0.01 0.02 0.67 −0.32 −0.03 −0.11 0.19 1.00

27G.R. Tripathy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 397 (2014) 20–30

are from Table 2. The optimizationmethod used for the inverse modelis the same that used successfully by Tripathy and Singh (2010) forapportioning the sources of the dissolved solutes of the Ganga head-waters. The model based on this iterative approach yields the contri-bution of sediments from the riverine sources to the SK187/PC33sediments.

The inverse-model results on the sediment contributions from theGanga, Brahmaputra and the GK rivers to the core site are listed inTable 6. On average, the Himalayan rivers supply 66 ± 13% of sedi-ments to the core site with 35 ± 16% derived from the Ganga. TheGK rivers supply 34 ± 13% sediment to the SK187/PC33 core site.These estimates are consistent with that reported earlier using theSr–Nd isotopic composition of these sediments (Tripathy et al.,2011b). Depth profile of sediment contributions from the Himalayanrivers to the SK187/PC33 core site showed large variability (Fig. 6).The sediment supply from the Himalayan (Ganga and Brahmaputra)rivers is observed to be lower during the LGM compared to their cor-responding range observed during Pre-LGM period. The contributionfrom the Himalayan rivers reduced from ~70% during Pre-LGM to~50% during the LGM. This declining trend is also noticed in thecase of the Ganga; however, it is interesting to note that the sediment

Table 5Factor analysis of geochemical dataset of the SK187/PC33 core extracts four factorswith Eigen values >1.

Factor 1 Factor 2 Factor 3 Factor 4

Eigen value 9.0 3.3 2.3 1.5

% variance 47 17 12 8

Na 0.01 −0.57 0.55 0.19K 0.68 −0.41 0.30 0.02Ca −0.77 0.56 −0.08 0.16Mg 0.20 0.59 0.56 0.51Al 0.92 −0.19 0.14 −0.01Fe 0.93 0.24 −0.14 −0.06Ti 0.81 0.06 −0.31 −0.12CaCO3 −0.76 0.52 −0.26 0.02Corg 0.21 −0.49 −0.51 0.66TN 0.21 −0.49 −0.51 0.66Ba 0.74 −0.11 0.18 −0.07Cr 0.89 0.30 −0.07 0.10Mn 0.12 0.40 −0.42 0.09Sr −0.76 0.38 −0.40 0.07Co 0.83 0.42 −0.06 −0.04Cu 0.86 0.38 −0.14 −0.06Ni 0.77 0.55 0.01 0.10Zn −0.15 0.53 0.62 0.52V 0.96 0.13 −0.19 −0.01

contribution from the Brahmaputra remained nearly the same duringthe LGM and the Pre-LGM periods (Fig. 6).

Results from the inverse model showed about 30% (from ~70% to50%) decline in the sediment contribution from the Himalayan rivers,mainly the Ganga, to the BoB during the LGM, consistent with thatreported earlier based on Sr and Nd isotopic composition (Tripathyet al., 2011b). The possible cause for this change in the Himalayanerosion during this climate event could be due to weakening of south-west monsoon (Duplessy, 1982) and the larger glacial extent (Owenet al., 2002) over the HH. Presently, the sediment load of the Himala-yan rivers is by and large regulated by the south-west monsoon. It hasbeen well documented that the intensity of the SWmonsoon was rel-atively lower during the LGM (Duplessy, 1982; Prell and Kutzbach,1987; Sarkar et al., 1990). Therefore, it is likely that the relative sup-ply of sediments from the G–B rivers to the BoB could have decreasedin response to weakening of south-west monsoon. Along with the SWmonsoon, the river basins in the Brahmaputra and the peninsularIndia also receive rainfall during north-east monsoon. Hence, the sed-iment supply from these (Brahmaputra and GK) river basins showedminimal change during the LGM. Furthermore, the extent of glacial

Fig. 4. Loading for various elements in the dominant factor (that explains about half ofthe observed total variance) extracted from the factor analysis of geochemistry of theSK187/PC33 sediments. This factor is primarily loaded with elements (Al and Ti) thatfrequently serve as detrital indicators, assigning this factor to be the detrital sourceof these sediments. Most of other trace elements are also found to have higher factorloadings for this factor.

Fig. 5. Depth profiles of Al-normalized ratios of K, Ti, Cr and Cu of the SK187/PC33 sediments. Discernible changes in these elemental ratios during LGM indicate relative variationsin sediment provenances to the core site, with reduced contribution from the Himalayan rivers during this climatic event. Down-core variation in Ti/Ca and CIA* is also shown here;the recent increase in Ti/Ca ratio is due to lower CaCO3 during this period, possibly hinting at change in marine productivity pattern. Two anomalous samples are not considered inthis depth profile. The bold curve drawn through the data in the figure is the 3-point moving average.

28 G.R. Tripathy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 397 (2014) 20–30

cover over the HH (Owen et al., 2002) was higher during LGM due toits high elevation which reduces the exposed area for weathering. TheHH is the dominant sediment supplier from the Himalaya to the BoBand hence, larger glacial cover over this geological unit is likely tolower the supply of sediment from the Himalaya to the Bay of Bengal.Observations from this study confirm that the Himalayan erosion isregulated by the changes in the regional climate, hinting towards aclimate–erosion link. This conclusion of climate–erosion link isconsistent with results from some of the earlier studies on the

Table 6Results from the inverse model on sediment contributions (in %; Errors are in ±1σ)from the Ganga, Brahmaputra and the Godavari–Krishna river systems to the SK187/PC33 core site.

Sample ID Ganga Brahmaputra G–B riversa GK rivers

10–12 cm 21 ± 6 30 ± 8 51 ± 10 49 ± 418–20 cm 17 ± 6 60 ± 8 77 ± 10 23 ± 328–30 cm 20 ± 6 34 ± 8 54 ± 10 46 ± 4109–111 cm 14 ± 6 26 ± 8 40 ± 10 60 ± 5149–151 cm 21 ± 6 38 ± 8 61 ± 10 39 ± 4201–203 cm 30 ± 6 39 ± 8 69 ± 10 31 ± 3249–251 cm 31 ± 6 25 ± 8 56 ± 10 44 ± 4300–302 cm 26 ± 6 26 ± 8 52 ± 10 48 ± 4320–322 cm 27 ± 6 34 ± 8 61 ± 10 39 ± 4350–352 cm 54 ± 5 17 ± 6 71 ± 8 29 ± 3400–402 cm 25 ± 6 33 ± 8 59 ± 10 41 ± 4450–452 cm 30 ± 6 36 ± 7 66 ± 10 34 ± 3500–502 cm 30 ± 6 44 ± 7 74 ± 9 26 ± 3550–552 cm 53 ± 5 34 ± 6 87 ± 8 13 ± 2600–602 cm 11 ± 6 48 ± 7 59 ± 9 42 ± 4650–652 cm 26 ± 6 20 ± 8 46 ± 10 54 ± 5700–702 cm 28 ± 6 25 ± 8 53 ± 10 47 ± 4750–752 cm 48 ± 5 32 ± 6 80 ± 8 20 ± 2800–802 cm 48 ± 5 17 ± 6 66 ± 8 34 ± 3849–851 cm 31 ± 6 48 ± 7 79 ± 9 21 ± 3899–901 cm 40 ± 5 16 ± 7 56 ± 9 44 ± 4949–951 cm 38 ± 6 23 ± 7 61 ± 9 39 ± 4999–1001 cm 43 ± 5 18 ± 7 62 ± 9 38 ± 41049–1051 cm 46 ± 5 26 ± 7 72 ± 8 28 ± 31099–1101 cm 44 ± 5 27 ± 7 71 ± 9 29 ± 31149–1151 cm 49 ± 5 34 ± 6 83 ± 8 17 ± 21199–1201 cm 86 ± 4 11 ± 5 97 ± 6 3 ± 21219–1221 cm 51 ± 5 23 ± 6 74 ± 8 26 ± 3

a Sum of contributions from the Ganga and the Brahmaputra rivers.

paleo-erosion pattern of the Himalaya (Goodbred, 2003; Ahmad etal., 2005; Kessarkar et al., 2005; Clift et al., 2008; Rahaman et al.,2009; Prakashbabu et al., 2010; Tripathy et al., 2011b). However,these outcomes on linkage of erosion with climate are not concurrentwith some of existing studies (Colin et al., 1999; Pierson-Wickmannet al., 2001; Kessarkar et al., 2003; Goswami et al., 2012) that showinsignificant changes in the provenances despite climatic changesduring last few ka. The possible reason for this difference is unclear;one possible explanation could be that the studies showing no ero-sion–climate link are primarily from locations that receive sedimentsfrom limited sources and hence, change in erosion pattern in re-sponse to climatic changes may not be appreciated in these locationsas the geochemical proxies only preserve signatures of relative (butnot absolute) changes in sediment supply.

Alongwith themajor excursion during the LGM, the results from theinverse model also indicated fall in sediment contribution from theGanga and the G–B to the core site during theHolocene (Fig. 6). Howev-er, the estimated sediment contribution from the Ganga and theG–B forthese (Holocene) samples overlap within their associated uncertainty,implying insignificant change in the sediment supply from the G–B dur-ing the Holocene. This inference also draws support from minimalchange observed for the Sr and Nd isotopic composition of these sedi-ments (Tripathy et al., 2011b) during this period. Combining thesemulti-proxy observations, we argue the need of minimizing uncer-tainties associatedwith the geochemical composition of sediment prov-enances (river sediments) to precisely estimate their contributions tothe core. Further, this study warrants the requirement of combinedstudy of elemental and Sr–Nd isotopic compositions of marine sedi-ments to better constrain the relative supply of sediments from sourcesto the core site.

5. Conclusions

This study investigated the major and trace element geochemistryof sediments from a ~13-m long piston core (SK187/PC33) from thewestern BoB to track temporal changes in their sediment sources.Factor analysis of the chemical composition of these sediments sug-gests that their geochemistry is controlled mainly by their supplyfrom detrital sources, viz. sediments derived by the Himalayan andthe peninsular Indian rivers. Inverse model calculation and depth

Fig. 6. Estimated sediment contributions from the Himalayan rivers to the core site of SK187/PC33 using an inverse model. The bold curve line is the 3-point average of the dataset.Besides two anomalous samples (221–223 cm, 440–442 cm), one outlier sample (1199–1201 cm) with extremely high contribution from the Ganga is not included in this trend.

29G.R. Tripathy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 397 (2014) 20–30

profiles of Al-normalized elemental ratios indicate 30% reduction inthe sediment supply from the Himalaya during the LGM. This reduc-tion in sediment supply from the Himalaya is attributable to weaken-ing of south-west monsoon and the large glacial extent over the HHduring this climatic event. These inferences are consistent with ourearlier studies using Sr–Nd isotope systematics on the same samplesand support the hypothesis of link between climate and the Himala-yan erosion on a ka timescale.

Acknowledgment

Discussions with S. Krishnaswami helped improving this manu-script. We thankfully acknowledge M. M. Sarin for his constructiveviews during this work. Thanks to Ravi Bhushan for his help duringthe carbon and nitrogen analyses. We thank the editor and two anon-ymous reviewers for their constructive suggestions. Financial assis-tance under the BENFAN project by the Ministry of Earth Sciences,India is highly appreciated. This is NIO contribution no. 5378.

References

Ahmad, S.M., Anil Babu, G., Padmakumari, V.M., Dayal, A.M., Sukhija, B.S., Nagabhushanam,P., 2005. Sr, Nd isotopic evidence of terrigenous flux variations in the Bay of Bengal:implications of monsoons during the last 34,000 years. Geophysical Research Letters32, 1–4.

Aucour, A., France-Lanord, C., Pedoja, K., Pierson-Wickmann, A.-C., Sheppard, S.M.F.,2006. Fluxes and sources of particulate organic carbon in the Ganga–Brahmaputrariver system. Global Biogeochemical Cycles 20 (2). http://dx.doi.org/10.1029/2004GB002324.

Balakrishna, K., Probst, J., 2005. Organic carbon transport and C/N ratio variations in alarge tropical river: Godavari as a case study, India. Biogeochemistry 73, 457–473.

Bhushan, R., Dutta, K., Somayajulu, B.L.K., 2001. Concentrations and burial fluxes of or-ganic and inorganic carbon on the eastern margins of the Arabian Sea. Marine Ge-ology 178, 95–113.

Bouchez, J., Gaillardet, J., France-Lanord, C., Maurice, L., Dutra-Maia, P., 2011. Grain sizecontrol of river suspended sediment geochemistry: clues from Amazon river depthprofiles. Geochemistry, Geophysics, Geosystems 12 (Q03008), 1–24.

Boyle, E., 1981. Cadmium, zinc, copper, and barium in foraminifera tests. Earth andPlanetary Science Letters 53, 11–35.

Burdige, D., 2007. Preservation of organic matter in marine sediments: controls, mech-anisms, and an imbalance in sediment organic carbon budgets? Chemical Reviews107, 467–485.

Carvalho-e-Silva, M., Ramos, A., Tolentino, H., Enzweiler, J., Netto, S., Alves, M., 2003. In-corporation of Ni into natural goethite: an investigation by X-ray absorption spec-troscopy. American Mineralogist 88, 876–882.

Chakrapani, G.J., Saini, R., 2009. Temporal and spatial variations in water discharge andsediment load in the Alaknanda and Bhagirathi Rivers in Himalaya, India. Journal ofAsian Earth Sciences 35, 545–553.

Clift, P.D., Giosan, L., Blusztajn, J., Campbell, I.H., Allen, C., Pringle, M., Tabrez, A.R.,Danish, M., Rabbani, M.M., Alizai, C.A., Lückge, A., 2008. Holocene erosion of theLesser Himalaya triggered by intensified summer monsoon. Geology 36, 79–82.

Colin, C., Turpin, L., Bertaux, J., Desprairies, A., Kissel, C., 1999. Erosional history of theHimalayan and Burman ranges during the last two glacial–interglacial cycles.Earth and Planetary Science Letters 171, 647–660.

Colin, C., Turpin, L., Blamart, D., Frank, N., Kisel, C., Duchamp, S., 2006. Evolution ofweathering patterns in the Indo-Burman Ranges over 280 kyr: effects of sedimentprovenance on 87Sr/86Sr ratios tracer. Geochemistry, Geophysics, Geosystems 7.http://dx.doi.org/10.1029/2005GC000962.

Dalai, T.K., Krishnaswami, S., Sarin, M.M., 2002. Major ion chemistry in the headwatersof the Yamuna river system: chemical weathering, its temperature dependenceand CO2 consumption in the Himalaya. Geochimica et Cosmochimica Acta 66,3397–3416.

Das, A., Krishnaswami, S., 2007. Elemental geochemistry of river sediments from theDeccan Traps, India: implications to source of elements and their mobility duringbasalt–water interaction. Chemical Geology 242, 232–254.

Das, A., Krishnaswami, S., Sarin, M.M., Pande, K., 2005. Chemical weathering in theKrishna basin and western ghats of the Deccan traps, India: rates of basaltweathering and their controls. Geochimica et Cosmochimica Acta 69, 2067–2084.

Duplessy, J.C., 1982. Glacial to interglacial contrasts in the northern Indian Ocean. Na-ture 295, 494–498.

France-Lanord, C., Derry, L., 1997. Organic carbon burial forcing of the carbon cyclefrom Himalayan erosion. Nature 390, 65–67.

Galy, A., France-Lanord, C., 1999. Weathering processes in the Ganges–Brahmaputrabasin and the riverine alkalinity budget. Chemical Geology 159, 31–60.

Galy, A., France-Lanord, C., 2001. Higher erosion rates in the Himalaya: geochemicalconstraints on riverine fluxes. Geology 29, 23–26.

Galy, V., France-Lanord, C., Beyssac, O., Faure, P., Kudrass, H., Palhol, F., 2007. Efficientorganic carbon burial in the Bengal Fan sustained by the Himalayan erosional sys-tem. Nature 450, 407–410.

Galy, V., Francois, L., France-Lanord, C., Faure, P., Kudrass, H., Palhol, F., Singh, S.K., 2008.C4 plants decline in the Himalayan basin since the Last Glacial Maximum. Quater-nary Science Reviews 27, 1396–1409.

Garzanti, E., Ando, S., France-Lanord, C., Censi, P., Vignola, P., Galy, V., Lupker, M., 2011.Minerological and chemical variability of fluvial sediments 2. Suspended-load silt(Ganga–Brahmaputra, Bangladesh). Earth and Planetary Science Letters 302,107–120.

Goodbred, S.L., 2003. Response of the Ganges dispersal system to climate change: asource-to-sink view since the last interstade. Sedimentary Geology 162, 83–104.

Goswami, V., Singh, S.K., Bhushan, R., Rai, V.K., 2012. Temporal variations in 87Sr/86Srand εNd in sediments of the southeastern Arabian Sea: impact of monsoon and sur-face water circulation. Geochemistry, Geophysics, Geosystems 13, Q01001. http://dx.doi.org/10.1029/2011GC03802.

Gupta, L., Subramanian, V., Ittekkot, V., 1997. Biogeochemistry of particulate organicmatter transported by the Godavari river, India. Biogeochemistry 38 (2), 103–128.

Horowitz, A., Elrick, K., 1987. The relation of stream sediment surface area, grain sizeand composition to trace element geochemistry. Applied Geochemistry 2,437–451.

Huang, S., Conte, M., 2009. Source/process apportionment of major and trace elements insinking particles in the Sargasso sea. Geochimica et Cosmochimica Acta 73, 65–90.

Huh, Y., 2010. Estimation of atmospheric CO2 uptake by silicate weathering in theHimalayas and the Tibetan plateau: a review of existing fluvial geochemical data.In: Clift, P., Tada, R., Zheng, H. (Eds.), Monsoon evolution and tectonics—climatelinkage in Asia: Geological Society, London, Special Publications, 342, pp. 129–151.

30 G.R. Tripathy et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 397 (2014) 20–30

Islam, M.R., Begum, S.F., Yamaguchi, Y., Ogawa, K., 1999. The Ganges and Brahmaputrarivers in Bangladesh: basin denudation and sedimentation. Hydrological Processes13, 2907–2923.

Kessarkar, P.M., Rao, V.P., Ahmad, S.M., Anil, Babu G., 2003. Clay minerals and Sr–Ndisotopes of the sediments along the western margin of India and their implicationfor sediment provenance. Marine Geology 202, 55–69.

Kessarkar, P.M., Rao, V.P., Ahmad, S.M., Patil, S.K., Anil Kumar, A., Anil Babu, G.,Chakraborty, S., Soundar Rajan, R., 2005. Changing sedimentary environment dur-ing the Late Quaternary: sedimentological and isotopic evidence from the distalBengal Fan. Deep Sea Research 52, 1591–1615.

Krishnaswami, S., Tripathy, G.R., 2012. Chemical and physical erosion of river basins ofIndia: links to climate and tectonics. Proceedings of the Indian National ScienceAcademy 78, 299–311.

Milliman, J.D., 2001. River inputs. Encyclopedia of Ocean Sciences, pp. 2419–2427.Milliman, J.D., Syvitski, J., 1992. Geomorphic/tectonic control of sediment discharge of

sediment discharge to ocean: the importance of small mountainous rivers. Journalof Geology 100, 525–544.

Morse, J., Mackenzie, F., 1990. Geochemistry of Sedimentary Carbonates. Elsevier, NewYork (707 pp.).

Nesbitt, H., Young, G., 1982. Early Proterozoic climates and plate motions inferred frommajor element chemistry of lutites. Nature 299, 715–717.

Owen, L., Finkel, R., Caffee, M., 2002. A note on the extent of glaciation throughout theHimalaya during the global Last Glacial Maximum. Quaternary Science Reviews 21,147–157.

Panda, D., Kumar, A., Mohanty, S., 2011. Recent trends in sediment load of the tropical(Peninsular) river basins of India. Global and Planetary Change 75, 108–118.

Pattan, J.N., Parthiban, G., Prakashbabu, C., Khadge, N., 2008. A note on geochemistry ofsurface sediments from Krishna–Godavari basin, East coast of India. Journal of theGeological Society of India 71, 107–114.

Pierson-Wickmann, A., Resiberg, L., France-Lanord, C., 2001. Os–Sr–Nd results fromsediments in the Bay of Bengal: implications for sediment transport and the ma-rine Os record. Paleoceanography 16, 435–444.

Prakashbabu, C., Pattan, J., Dutta, K., Basavaiah, N., Ravi Prasad, G., Ray, D.K., Govil, P.,2010. Shift in detrital sedimentation in the eastern Bay of Bengal during the latequaternary. Journal of Earth System Science 119, 285–295.

Prell, W.L., Kutzbach, J.E., 1987. Monsoon variability over the past 150,000 years. Jour-nal of Geophysical Research 92, 8411–8425.

Rahaman, W., Singh, S.K., Sinha, R., Tandon, S.K., 2009. Climate control on erosion dis-tribution over the Himalaya during the past 100 ka. Geology 37, 559–562.

Rai, S.K., 2008. Isotopic and Geochemical Studies of Ancient and Modern Sediments.Ph.D. Thesis MS University, Vadodara (195 pp.).

Sarin, M.M., Borole, D.V., Krishnaswami, S., 1979. Geochemistry and geochronology ofsediments from the Bay of Bengal and the equatorial Indian Ocean. Proceedingsof the Indian Academy of Sciences 88 A, 131–154.

Sarkar, A., Ramesh, R., Bhattacharya, S.K., Rajagopalan, G., 1990. Oxygen isotope evi-dence for a stronger winter monsoon current during the last glaciation. Nature343, 549–551.

Schwertmann, U., Cornell, R., 2000. Iron Oxides in the Laboratory: Preparation andCharacterization, 2nd ed. Wiley-Vch, Weinheim (188 pp.).

Sidhu, P., Gilkes, R., Posner, A., 1980. Behavior of Co, Ni, Zn, Cu, Mn, and Cr in magnetiteduring alteration to maghemite and hematite. Soil Science Society of America Jour-nal 44, 135–138.

Singh, S.K., France-Lanord, C., 2002. Tracing the distribution of erosion in the Brahma-putra watershed from isotopic compositions of stream sediments. Earth and Plan-etary Science Letters 202, 645–662.

Singh, S.K., Sarin, M.M., France-Lanord, C., 2005. Chemical erosion in the easternHimalaya: major ion composition of the Brahmaputra and δ13C of the dissolved in-organic carbon. Geochimica et Cosmochimica Acta 69, 3573–3588.

Singh, S.K., Rai, S.K., Krishnaswami, S., 2008. Sr andNd isotopes in river sediments from theGanga basin: sediment provenance and spatial variability in physical erosion. Journalof Geophysical Research 113, F03006. http://dx.doi.org/10.1029/2007JF000909.

Subramanian, V., Richey, J., Abbas, N., 1985. Geochemistry of river basins in the Indiansubcontinent: part II. Preliminary studies on the particulate C and N in the Ganga–Brahmaputra river system. Mitteilungen aus dem Geologisch-PaläontologischenInstitut der Universität Hamburg 58, 513–518.

Tiwari, M., Ramesh, R., Somayajulu, B.L.K., Jull, A.J.T., Burr, G.S., 2005. Early deglacial(~19–17 ka) strengthening of the northeast monsoon. Geophysical Research Let-ters 32, 1–4.

Tripathy, G.R., 2011. Isotope Geochemistry of Black Shales and Recent Marine Sedi-ments. Ph. D. Thesis MohanLal Sukhadia University (207 pp.).

Tripathy, G.R., Singh, S.K., 2010. Chemical erosion rates of river basins of the Ganga sys-tem in the Himalaya: reanalysis based on inversion of dissolved major ions, Sr, and87Sr/86Sr. Geochemistry, Geophysics, Geosystems 11, Q03013. http://dx.doi.org/10.1029/2009GC002862.

Tripathy, G.R., Singh, S.K., Bhushan, R., Ramaswamy, V., 2011a. Sr–Nd isotope composi-tion of the Bay of Bengal sediments: impact of climate on erosion in the Himalaya.Geochemical Journal 45, 175–186.

Tripathy, G.R., Singh, S.K., Krishnaswami, S., 2011b. Sr and Nd isotopes as tracers ofchemical and physical erosion. In: Baskaran, M. (Ed.), Handbook of EnvironmentalIsotope Geochemistry. Springer-Verlag, pp. 521–551. http://dx.doi.org/10.1007/978-3-642-10637-8_26.