1. Introduction
The South Asian or Indian monsoon system represents one of the basic elements of global atmospheric circulation. However, little is known about its evolution and variability over long periods of geologic time. Previous palaeoceanographic and palaeoclimatic reconstructions from the western and northern Arabian Sea, using planktic foraminiferal assemblages and geochemical proxies, provided deep insights into South Asian / Indian monsoon variability on long- to short-term timescales (orbital to sub-orbital and centennial to sub-centennial scales) (e.g. Clemens et al. Reference Clemens, Prell, Murray, Shimmield and Weedon1991; Anderson & Prell, Reference Anderson and Prell1993; Vénec-Peyré et al. Reference Vénec-Peyré, Caulet and Grazzini1995; Naidu & Malmgren, Reference Naidu and Malmgren1996; Reichart et al. Reference Reichart, Lourens and Zachariasse1998; Schulz et al. Reference Schulz, Von Rad and Erlkeuser1998; von Rad et al. Reference Von Rad, Schulz, Riech, Den Dulk, Berner and Sirocko1999; Schulte & Müller, Reference Schulte and Müller2001; Jung et al. Reference Jung, Ivanova, Reichart, Davies, Ganssen, Kroon, Van Hinte, Clift, Kroon, Gaedicke and Craig2002; Gupta et al. Reference Gupta, Anderson and Overpeck2003; Ivanova et al. Reference Ivanova, Schiebel, Singh, Schmiedl, Niebler and Hemleben2003; Ivanochko et al. Reference Ivanochko, Ganeshram, Brummer, Ganssen, Jung, Moreton and Kroon2005). Earlier studies suggested that fossil planktic foraminiferal assemblages in the eastern Arabian Sea (EAS) record primarily winter monsoon induced changes in surface water oceanographic conditions (Singh et al. Reference Singh, Kroon and Ganeshram2006, Reference Singh, Jung, Darling, Ganeshram, Ivanochko and Kroon2011, Reference Singh, Jung, Anand, Kroon and Ganeshram2018), because winter monsoon winds play a dominant role in surface hydrography and biological productivity. Most of the previous palaeoceanographic reconstructions based on planktic foraminiferal assemblages and geochemical proxies (e.g. stable isotopes and Mg/Ca-based sea surface temperature (SST)) from the EAS mainly focused on the last glacial–interglacial cycle (Cayre & Bard, Reference Cayre and Bard1999; Thamban et al. Reference Thamban, Purnachandra Rao, Schneider and Grootes2001; Ivanova et al. Reference Ivanova, Schiebel, Singh, Schmiedl, Niebler and Hemleben2003; Pattan et al. Reference Pattan, Masuzawa, Naidu, Parthiban and Yamamoto2003; Banakar et al. Reference Banakar, Oba, Chodankar, Kuramoto, Yamamoto and Minagawa2005; Chodankar et al. Reference Chodankar, Banakar and Oba2005; Guptha et al. Reference Guptha, Naidu, Haake and Schiebel2005; Singh et al. Reference Singh, Kroon and Ganeshram2006, Reference Singh, Jung, Darling, Ganeshram, Ivanochko and Kroon2011, Reference Singh, Jung, Anand, Kroon and Ganeshram2018; Anand et al. Reference Anand, Kroon, Singh, Ganeshram, Ganssen and Elderfield2008). However, no record of millennial- to orbital-scale variations from this region is currently available extending beyond the last glacial–interglacial cycle, particularly for the early Pleistocene, which has been suggested as a period when a major reorganization of atmospheric circulation from a weak to a strong zonal Walker circulation took place (Etourneau et al. Reference Etourneau, Schneider, Blanz and Martinez2010). For that reason, sediment records retrieved at IODP Site U1457 in the Laxmi Basin of the EAS provide an opportunity to reconstruct the history of oceanographic and monsoon variations on tectonic, orbital to sub-orbital timescales further back in geological time (Pandey et al. Reference Pandey, Clift and Kulhanek2015).
Here, we present planktic foraminiferal assemblage records together with stable carbon and oxygen isotopes, Mg/Ca-based SSTs and seawater δ18O estimates in order to decipher changes in the regional surface hydrography and productivity in the EAS during the early Pleistocene and, hence, changes in strength of the South Asian / Indian monsoonal circulation. This study also aims to examine possible atmosphere–ocean–climate linkages between the South Asian / Indian monsoon system and low- and high-latitude climate components.
2. Study area
The climate and surface hydrography in the Arabian Sea are primarily controlled by the seasonally reversing South Asian / Indian monsoon wind patterns (Fig. 1). The monsoon winds produce distinct seasonal and spatial variations in surface ocean hydrography, circulation, nutrient distribution and productivity (e.g. Banse, Reference Banse1987; Brock et al. Reference Brock, Sathyendranath and Platt1994; Luis & Kawamura, Reference Luis and Kawamura2004). The advection of upwelled water from the western Arabian Sea during the summer SW monsoon season (June to September) and convective mixing during the winter NE monsoon (November–February) modulate biological productivity in the EAS. During the summer monsoon season, a clockwise circulation develops in the Arabian Sea (Schott, Reference Schott1983), and the eastern branch of this anticyclonic circulation forms the West India Coastal Current (WICC) (Shetye et al. Reference Shetye, Gouveia, Shenoi, Michael, Sundar, Almeida and Santanam1991) in the EAS. The equatorward-flowing WICC in summer carries high-salinity waters from the northern Arabian Sea towards the south (Prasanna Kumar et al. Reference Prasanna Kumar, Narvekar, Kumar and Shaji2004). As the wind system reverses during the winter, the WICC starts flowing northward against the alongshore component of the wind stress (Shetye et al. Reference Shetye, Gouveia, Shenoi, Michael, Sundar, Almeida and Santanam1991) carrying the low salinity waters from the Bay of Bengal (Prasanna Kumar et al. Reference Prasanna Kumar, Narvekar, Kumar and Shaji2004). The dry northeasterly winter monsoon winds cause cooling of surface waters and vertical mixing with an injection of nutrient-rich subsurface waters into the photic zone (Banse, Reference Banse1987; Lévy et al. Reference Lévy, Shankar, André, Shenoi, Durand and De Boyer Montégut2007; Koné et al. Reference Koné, Aumont, Lévy, Resplandy, Wiggert, Hood, Wajih, Naqvi, Brink and Smith2009). This results in an increase in primary productivity in the northeastern and EAS (Banse & McClain, Reference Banse and McClain1986; Madhupratap et al. Reference Madhupratap, Kumar, Bhattathiri, Kumar, Raghukumar, Nair and Ramaiah1996). The seasonally reversing winds also modulate the mixed layer depth (MLD) in the region (Banse, Reference Banse, Haq and Milliman1984). Recent studies have shown spatio-temporal variability of physical and biological parameters of MLD in the EAS (Shankar et al. Reference Shankar, Remya, Vinayachandran, Chatterjee and Behera2015; Vijith et al. Reference Vijith, Vinayachandran, Thushara, Amol, Shankar and Anil2016). These studies suggest inhibition of MLD deepening in the south of the EAS due to advection of low-salinity water carried by the poleward-flowing WICC (Vijith et al. Reference Vijith, Vinayachandran, Thushara, Amol, Shankar and Anil2016). This in turn inhibits the entrainment of nutrients into the mixed layer, resulting in low chlorophyll concentration in the southern part of EAS (Vijith et al. Reference Vijith, Vinayachandran, Thushara, Amol, Shankar and Anil2016). Satellite-derived chlorophyll data reveals a regional difference in chlorophyll concentration within the EAS during winter, with high concentration in the northern EAS (Banse & McClain, Reference Banse and McClain1986).
Map showing the location of IODP Site U1457. Schematic representation of dominant wind directions (dashed yellow lines) and main surface ocean circulation (white lines) in the Arabian Sea during the monsoons (as inferred from Schott & McCreary, Reference Schott and McCreary2001; Singh et al. Reference Singh, Jung, Darling, Ganeshram, Ivanochko and Kroon2011; Cabarcos et al. Reference Cabarcos, Flores, Singh and Sierro2014). Map indicates chlorophyll concentration (a) and sea surface temperature (SST) (b) during the summer and winter monsoons (sources: Chl a: NASA/SeaWiFS; SST: US National Oceanic and Atmospheric Administration). WICC: West Indian Coastal Current; NMC: Northeast Monsoon Current.
IODP Site U1457 in the Laxmi Basin is located at the eastern periphery of the oligotrophic part of the central Arabian Sea (Fig. 1). A subtle increase in productivity at the site during the summer monsoon season might be related to the advection of nutrient-rich upwelling filaments from the western Arabian Sea and/or to a higher supply of nutrients because of increased fluvial runoff into the EAS during the rainy summer monsoon season. On the contrary, an increase in productivity as revealed by the chlorophyll concentration data and cooler SST is recorded during the winter season. Annual primary productivity in the vicinity of the IODP Site U1457 is estimated to be lower (∼166 g C m−1 a−1), as compared to the northern and western Arabian Sea (∼230–250 g C m−1 a−1) (Madhupratap et al. Reference Madhupratap, Kumar, Bhattathiri, Kumar, Raghukumar, Nair and Ramaiah1996; Rixen et al. Reference Rixen, Haake and Ittekkot2000; Ivanova et al. Reference Ivanova, Schiebel, Singh, Schmiedl, Niebler and Hemleben2003). The MLD in winter is reported to be slightly deeper (60–80 m) than during the summer (40–80 m) (Rao et al. Reference Rao, Molinari and Festa1989; Madhupratap et al. Reference Madhupratap, Kumar, Bhattathiri, Kumar, Raghukumar, Nair and Ramaiah1996).
3. Material and methods
3.a. Site description and age model
IODP Site U1457 (17° 9.9486′ N, 67° 55.8121′ E; water depth 3534 m) is located within the Laxmi Basin in the EAS, ∼491 km from the Indian coast and ∼750 km from the present-day Indus River mouth (Fig. 1; Pandey et al. Reference Pandey, Clift and Kulhanek2015). The Indus River is the major source of sediment supply to the Laxmi Basin (Pandey et al. Reference Pandey, Clift and Kulhanek2015). Minor contributors of sediment to the Laxmi Basin are rivers flowing through the Western Ghats of India, mainly the Narmada and Tapti. Three holes were drilled at this site. The sedimentary section from 346 to 426 mbsf (metres below the seafloor) at Hole U1457C was selected for the present study. Sedimentation rates at Hole U1457C range from ∼4 cm ka−1 in the lower part of the examined section to ∼58 cm ka−1 in the upper part of the section, probably because of its location in the distal Indus fan (Pandey et al. Reference Pandey, Clift and Kulhanek2016). Higher sedimentation rates in the upper part of the section might be attributed to a higher supply from the Indus around that time. The sediment consists mainly of foraminifera-rich nannofossil ooze, interbedded with silty clay and sandy silt layers. Sandy silt layers characterized by terrestrial detrital sediment are interpreted as turbidites (Pandey et al. Reference Pandey, Clift and Kulhanek2015). The age model of the examined section is based on four calcareous nannofossil biostratigraphic datums (Pandey et al. Reference Pandey, Clift and Kulhanek2015; Table 1). A short interval (∼0.45 Ma) of non-deposition between 403.83 mbsf and 415.35 mbsf has been suggested based on biostratigraphic and lithological shipboard findings (see Pandey et al. Reference Pandey, Clift and Kulhanek2016 for details). However, the available biochronological datums are not sufficient to confirm this hiatus.
Calcareous nannofossil datums used for age model of the examined section of Site U1457
3.b. Analytical methods
In total, 268 samples at 20 cm regular intervals were analysed for planktic foraminiferal assemblages. Dried samples were washed through wet sieving over a 63 µm sieve, following conventional processing technique. Residues (>63 µm) were dry sieved over a 125 µm sieve. We generated planktic foraminiferal census data from an aliquot of 250–300 specimens of the >125 µm size fraction of each sample. Those samples containing a low number of planktic foraminifera tests were used completely for counting. For species identification, the taxonomic concepts of Kennett & Srinivasan (Reference Kennett and Srinivasan1983) and Hemleben et al. (Reference Hemleben, Spindler and Anderson1989) were followed. Based on the census counts, the absolute abundance of each species was estimated (number of specimens / g dry sediment >125 µm).
Planktic foraminifera respond rapidly to changes in surface hydrography and primary productivity induced by variation in seasonal monsoon wind circulation, as reflected by variations in species abundance and chemical composition of their tests (e.g. Hutson & Prell, Reference Hutson and Prell1980; Steens et al. Reference Steens, Ganssen, Kroon, Summerhayes, Prell and Emeis1992; Peeters & Brummer, Reference Peeters, Brummer, Clift, Kroon, Gaedicke and Craig2002; Anand et al. Reference Anand, Kroon, Singh, Ganeshram, Ganssen and Elderfield2008). Therefore, planktic foraminiferal assemblages in sediment have great potential for reconstructing past surface ocean conditions (e.g. Ravelo et al. Reference Ravelo, Fairbanks and Philander1990; Kroon et al. Reference Kroon, Steens, Troelstra, Prell and Niitsuma1991; Singh et al. Reference Singh, Kroon and Ganeshram2006, Reference Singh, Jung, Anand, Kroon and Ganeshram2018). We examined variations of foraminiferal proxies (absolute abundances of ecologically sensitive species and mixed layer eutrophic and oligotrophic species abundance ratio) to understand the temporal variability of the surface water hydrography, viz. stratification, vertical mixing and nutrient conditions in response to changes in the intensity of the seasonal monsoon circulation. The abundance ratio of Globigerina bulloides to Globigerinoides ruber is used as a proxy for palaeo-productivity variation associated with seasonal upwelling and/or vertical mixing versus stratification of upper water column (Conan & Brummer, Reference Conan and Brummer2000; Singh et al. Reference Singh, Jung, Darling, Ganeshram, Ivanochko and Kroon2011). The advantage of this palaeo-productivity proxy is that it is unaffected by dissolution artefacts in the sediment, as both species are equally sensitive to dissolution (Cullen & Prell, Reference Cullen and Prell1984; Conan et al. Reference Conan, Ivanova and Brummer2002). Hence, G. bulloides / G. ruber ratio is a robust proxy and this has been used previously for palaeoceanographic reconstructions in the EAS (Singh et al. Reference Singh, Kroon and Ganeshram2006, Reference Singh, Jung, Darling, Ganeshram, Ivanochko and Kroon2011, Reference Singh, Jung, Anand, Kroon and Ganeshram2018).
As the site lies above the modern calcium carbonate compensation depth (CCD), dissolution of calcite tests is expected to be minimal. Nevertheless, the degree of preservation of planktic foraminiferal assemblage has been evaluated by the record of absolute abundance ratio of common solution-susceptible species to solution-resistant species (Fig. 2). We also determined the dissolution index (DI) (number of resistance species tests / total number of common species tests × 10, following Berger (Reference Berger1973)) (Fig. 2).
Records of (a) total planktic foraminifera (no. × 1000 g−1), (b) dissolution-susceptible species (no. × 1000 g−1), (c) dissolution index (Berger, Reference Berger1973), (d) dissolution-resistant species (no. × 1000 g−1), (e) absolute abundance ratio of dissolution-susceptible species and dissolution-resistant species from IODP Hole U1457C.
Isotope analyses of Globigerinoides sacculifer-quadrilobatus of the 250–350 µm fraction were performed at MARUM, University of Bremen, on Finnigan MAT 251 mass spectrometers with Kiel I or Kiel III devices. Isotope values were calibrated against the international Vienna Pee Dee Belemnite (VPDB) standard. The internal carbonate standard is a Solnhofen Limestone, which is calibrated to the National Bureau of Standards (NBS) 19 standard. The long-term analytical precision was better than ±0.07 ‰.
SST was reconstructed based on shell Mg/Ca of the planktic foraminiferal species G. sacculifer-quadrilobatus. For each sample, c. 30–40 specimens of G. sacculifer-quadrilobatus were picked out of the 250–350 µm fraction. Foraminiferal tests were cleaned in successive steps following the cleaning protocol developed by Barker et al. (Reference Barker, Greaves and Elderfield2003) and analysed with an inductively coupled plasma optical emission spectrometer (ICP-OES; Agilent Technologies, 700 Series with autosampler ASX-520 Cetac and micro-nebulizer) at MARUM – Center for Marine Environmental Sciences, University of Bremen, Germany. Instrumental precision of the ICP-OES was examined by analysis of an in-house standard solution with a Mg/Ca of 2.93 mmol mol−1 after every five samples with a long-term average of 2.94 mmol mol−1 (long-term standard deviation of 0.062 mmol mol−1). To allow inter-laboratory comparison, we analysed an international limestone standard (ECRM752-1) with a reported Mg/Ca of 3.75 mmol mol−1. The long-term average of the ECRM752-1 standard, which is routinely analysed twice before each batch of 50 samples in every session, is 3.78 mmol mol−1 (1σ = 0.066 mmol mol−1). Mn/Ca, Fe/Ca and Al/Ca were determined together with Mg/Ca because clay contamination and the occurrence of syn-sedimentary and post-depositional Mn-oxide precipitates and Mn-rich carbonate coatings can exert a significant control on Mg/Ca ratios, resulting in elevated Mg/Ca ratios and, by inference, overestimated SSTs. Our results indicate no significant Mg contributions by Mn oxides or Mn-rich carbonates or clays because ratios for Mn/Ca and Fe/Ca were <0.1 mmol mol−1, and not detectable for Al/Ca. The Mg/Ca ratios are not affected by the occurrence of post-depositional and syn-sedimentary precipitated Mn oxide and Mn-rich carbonate coatings, or by post-depositional partial dissolution. The G. sacculifer-quadrilobatus Mg/Ca estimates were converted to SSTs using the multiple-species equation of Anand et al. (Reference Anand, Elderfield and Conte2003) (Mg/Ca (mmol mol−1) = 0.347 exp (0.09 SST)). The errors of the temperature reconstructions are estimated by propagating the errors introduced by the Mg/Ca measurements and the Mg/Ca temperature calibration (see Mohtadi et al. Reference Mohtadi, Prange, Oppo, De Pol-Holz, Merkel, Zhang, Steinke and Lückge2014). The resulting errors for G. sacculifer-quadrilobatus are on average 1.08 °C.
Seawater δ18O (a proxy for salinity) was calculated by removing the temperature-driven component of changes in the δ18O G. sacculifer-quadrilobatus record using the temperature–δ18O seawater relationship given by Bemis et al. (Reference Bemis, Spero, Bijma and Lea1998) (δ18Oseawater =(T − 16.5 + 4.8 * δ18Ocalcite)/4.8 + 0.27). The component of seawater δ18O that is attributed to changes in the local hydrology was then calculated by subtracting the influence of continental ice volume taken from Bintanja & van de Wal (Reference Bintanja and Van de Wal2008).
4. Results
4.a. Planktic foraminiferal assemblages
The planktic foraminiferal assemblages recorded in this study comprise a total of 35 species, of which the most abundant are Globigerina bulloides, Globigerinoides ruber, Globigerinita glutinata, Globigerinoides trilobus, Globigerinoides sacculifer, Neogloboquadrina dutertrei and Pulleniatina obliquiloculata. Temporal variations in absolute abundances of these species are presented in Figure 3. The absolute abundance of total planktic foraminifera (number per g sediment >125 µm) varies between 0 and 99,000 specimens, with significantly lower abundances between 1.85 Ma and 2.7 Ma, followed by a prominent shift to higher values after 1.65 Ma (Fig. 2). Absolute abundances of the seven most dominant species remained low until 1.85 Ma and increased during the period 1.55–1.65 Ma (Fig. 3). G. glutinata is the most abundant species of the assemblage, followed by G. ruber, G. bulloides, N. dutertrei, P. obliquiloculata, G. trilobus and G. sacculifer (in decreasing order of absolute abundances). The interval prior to 1.85 Ma is characterized by an overall low abundance of planktic foraminifera, with very low occurrence of G. bulloides and N. dutertrei. Between 1.55 Ma and 1.65 Ma, high absolute abundances of G. bulloides and G. glutinata are recorded. Relative abundance records of G. glutinata and G. bulloides in general reflect an opposite pattern to G. ruber (Fig. 5 below). The foraminiferal assemblages comprise both mixed layer eutrophic (G. bulloides, G. glutinata) and oligotrophic (G. ruber, G. sacculifer, G. trilobus) species. Temporal variation patterns in abundance ratios of mixed layer eutrophic to oligotrophic species (a proxy for mixed layer trophic conditions) and G. bulloides to G. ruber (a proxy for surface water stratification) show major changes between 1.55 Ma and 1.65 Ma (Fig. 4 below). Before 1.85 Ma, the values of these proxy records are generally low as compared to the younger interval (1.55–1.65 Ma). The temporal variation pattern of the G. bulloides / G. ruber ratio broadly follows that of mixed layer eutrophic/oligotrophic species abundance ratio.
Absolute abundances of planktic foraminifera: (a) G. bulloides, (b) G. glutinata, (c) G. ruber, (d) G. trilobus, (e) G. sacculifer, (f) P. obliquiloculata, (g) N. dutertrei.
Foraminiferal, oxygen and carbon isotope, and SST records. (a) Total planktic foraminiferal abundance, (b) absolute abundance ratio of G. bulloides and G. ruber, (c) absolute abundance ratio of mixed layer eutrophic and oligotrophic species, (d) δ13C (‰), (e) Mg/Ca-based SST record in °C, (f) δ18O record of G. sacculifer-quadrilobatus, (g) seawater δ18O.
4.b. Stable oxygen, carbon isotopes, Mg/Ca-based SST estimates and seawater δ18O estimates
The oxygen isotope record of the surface planktic foraminifera G. sacculifer-quadrilobatus shows significant variations, with a maximum value of −0.51 ‰ and a minimum value of −2.18 ‰ (Fig. 4). However, the isotope record reflects a different pattern of variations at 1.55–1.65 Ma and 1.85–2.7 Ma. During the period 1.55–1.65 Ma, most of the samples show lighter values (<−1.50 ‰) with a mean value of −1.74 ‰. δ18O in this time interval varies between −1.38 ‰ and −2.19 ‰. In contrast, δ18O values during the period 1.85–2.7 Ma, except for a few samples, are relatively heavier (>1.50 ‰) with an average value of −1.21 ‰. The difference of 0.53 ‰ between the average values of the two time slices is intriguing. The carbon isotope record also shows major variations during the two time intervals. The average value δ13C between 1.85 Ma and 2.7 Ma is lighter (1.09 ‰), as compared to that between 1.55 Ma and 1.65 Ma (1.21 ‰). The δ13C varies between −0.16 ‰ and 2.14 ‰, with significantly higher values between 1.55 Ma and 1.65 Ma.
The G. sacculifer-quadrilobatus Mg/Ca-based SST record reveals that sea surface temperature at the examined site varies between 24.8 °C and 29.2 °C during the investigated time period 1.55–2.7 Ma (Fig. 4). The record of seawater δ18O estimates shows lighter values between 1.55 Ma and 1.65 Ma compared with the time period prior to 1.85 Ma, which is suggested to reflect a period of stronger monsoonal rainfall and/or riverine runoff.
5. Discussion
5.a. Preservation vs dissolution of planktic foraminifera
Because the examined site is located in the distal Indus fan, it is assumed that the foraminiferal abundance record is influenced by terrestrial input resulting in dilution of pelagic sediment components deposited at the sea floor (e.g. Singh et al. Reference Singh, Kroon and Ganeshram2006). Hence, the interval prior to 1.85 Ma, which is characterized by an overall low abundance of planktic foraminifera, may represent a period of higher supply of terrigenous-siliciclastic material to IODP Site U1457. We therefore suggest that the assemblage within a few intervals in the lower section (1.85–2.7 Ma) that have very low planktic foraminifera abundances is probably related to the dilution of faunal population caused by high terrestrial input to our site. In addition, deepening of the carbonate lysocline in the past may have influenced the preservation of foraminiferal tests. Dissolution effects on foraminiferal tests during these intervals can also not be ruled out, as is evident from the DI record (Fig. 2). Moreover, the intervals with moderate to high abundances, although not as high as recorded in the younger interval between 1.55 Ma and 1.65 Ma, probably suggest better preservation of foraminiferal tests and minimum sediment dilution. The preservation of foraminifera was good, along with minimal dilution of sediment in the period between 1.55 Ma and 1.65 Ma.
5.b. Seasonal monsoon wind induced changes in surface water hydrography and productivity
The planktic foraminiferal assemblages, oxygen and carbon isotopes and Mg/Ca SST records obtained for the studied section at IODP Site U1457 reveal significant fluctuations in the surface hydrography and productivity conditions in the Laxmi Basin during the early Pleistocene between c. 1.5 Ma and 2.7 Ma. Our proxy records indicate two distinct regimes of monsoon wind induced changes in upper water column structure and surface productivity during the periods 1.85–2.7 Ma and 1.55–1.65 Ma. We suggest that general low absolute abundances of the planktic foraminifera species G. glutinata, G. bulloides and N. dutertrei between 1.85 Ma and 2.7 Ma indicate stratified, nutrient-poor (oligotrophic) and low-productivity surface water conditions. This assertion is based on the fact that today G. glutinata, G. bulloides and N. dutertrei are associated with high-productivity environments in the Arabian Sea. G. glutinata feeds upon diatoms and is considered to be linked with the winter diatom bloom in the NE Arabian Sea because of vertical mixing and entrainment of nutrients into the mixed layer (Schulz et al. Reference Schulz, Von Rad, Ittekkot, Clift, Kroon, Gaedicke and Craig2002). Therefore, high (low) abundances of this species are suggestive of high (low) surface water productivity conditions, related to winter monsoonal wind induced overturning. G. bulloides is considered as a eurythermal species occurring abundantly in cold, nutrient-rich surface mixed layer waters (Sautter & Thunell, Reference Sautter and Thunell1991) associated with upwelling (e.g. Cullen & Prell, Reference Cullen and Prell1984; Anderson & Prell, Reference Anderson, Prell, Prell and Niitsuma1991; Kroon et al. Reference Kroon, Steens, Troelstra, Prell and Niitsuma1991; Naidu & Malmgren, Reference Naidu and Malmgren1996). In addition, G. bulloides has also been suggested to flourish in winter monsoon wind induced high surface productivity conditions (Singh et al. Reference Singh, Jung, Darling, Ganeshram, Ivanochko and Kroon2011, Reference Singh, Jung, Anand, Kroon and Ganeshram2018). N. dutertrei is a eutrophic thermocline dwelling species (e.g. Ravelo et al. Reference Ravelo, Fairbanks and Philander1990; Ravelo & Fairbanks, Reference Ravelo and Fairbanks1992) and it dominates when nutrients are brought into the photic zone, enhancing the primary productivity (Fairbanks et al. Reference Fairbanks, Sverdlove, Free, Wiebe and Bé1982; Schiebel et al. Reference Schiebel, Waniek, Bork and Hemleben2001; Ishikawa & Oda, Reference Ishikawa and Oda2007). It calcifies within the thermocline in association with the chlorophyll maximum (Sautter & Thunell, Reference Sautter and Thunell1991). In tropical areas, the abundance variation of N. dutertrei has been considered to be related to the changes in thermocline depth (Andreasen & Ravelo, Reference Andreasen and Ravelo1997; Jian et al. Reference Jian, Wang, Chen, Li, Zhao, Bühring, Laj, Lin, Pflaumann, Bian, Wang and Cheng2000).
The interval between 1.85 Ma and 2.7 Ma was a period when surface water was more stratified and nutrient-poor (oligotrophic), resulting in overall low productivity. Today, the surface water hydrography and productivity conditions in the EAS are primarily influenced by the winter NE monsoonal wind, which invokes vertical mixing leading to injection of nutrients to the photic zone and thus increased seasonal winter productivity. By analogy to the modern conditions, we therefore infer that the winter monsoon circulation was generally weak during this time interval. Low average values of δ13C also support our interpretation of low-productivity conditions during this period (Fig. 4). Heavier seawater oxygen isotope values compared to the time period 1.55–1.65 Ma might suggest that rainfall and/or continental runoff was lower during this period, indicating a weaker summer monsoon. Taken together, our faunal records, combined with the isotope data, suggest a weaker summer (SW) and winter (NE) monsoon during the time period 1.85–2.7 Ma. Our faunal, isotopic, Mg/Ca-based SST and seawater δ18O records show a prominent shift between ∼1.65 Ma and 1.85 Ma, which is interpreted to reflect a major reorganization of the South Asian / Indian monsoon circulation around that time (Fig. 4).
Overall higher absolute abundances of G. glutinata, G. bulloides and N. dutertrei between 1.55 Ma and 1.65 Ma are suggested to indicate nutrient-rich (eutrophic) and high-productivity surface water conditions due to intensification of the monsoonal winds (Fig. 4). On the other hand, the absolute abundance of G. ruber, which is a mixed layer species and usually associated with oligotrophic surface water conditions, also significantly increased during this period. In addition, average lighter seawater δ18O values between 1.55 Ma and 1.65 Ma suggest increased freshening of surface waters, most likely associated with increased fresh-water runoff from the continent due to enhanced South Asian / Indian summer monsoon precipitation. However, a detailed comparison of our proxy records reveals that periods of heavier G. sacculifer-quadrilobatus δ18O values and high relative abundances (%) of G. glutinata and N. dutertrei are associated with low relative abundances of G. ruber (Fig. 5). We submit that the G. sacculifer-quadrilobatus δ18O record reveals glacial–interglacial variations for the interval 1.55–1.65 Ma because the amplitude of changes is similar to that found in the ‘LR04’ benthic δ18O record stack for that time period (Lisiecki & Raymo, Reference Lisiecki and Raymo2005). Although only fragmentary, our Mg/Ca-based SST record reveals a c. 3°C glacial/interglacial amplitude which is similar to that of the Holocene – Last Glacial Maximum amplitude in the EAS (Anand et al. Reference Anand, Kroon, Singh, Ganeshram, Ganssen and Elderfield2008) (Fig. 5). We suggest that periods of high relative abundances of G. glutinata and N. dutertrei represent glacial periods with less stratified surface water, high nutrient levels and high productivity. We suggest that the enhanced productivity during those periods was triggered by the intensification of the NE winter monsoon winds. This assertion is based on the fact that present-day primary productivity changes in the EAS are strongly coupled to the intensity of the NE winter monsoon wind, affecting both SST and MLD in the Arabian Sea, and thus control the entrainment of nutrients into the mixed layer. A link between NE monsoon wind strength, SST and productivity has been observed in the northern Arabian Sea on glacial/interglacial timescales during the late Pleistocene (Rostek et al. Reference Rostek, Bard, Beaufort, Sonzogni and Ganssen1997; Schulte et al. Reference Schulte, Rostek, Bard, Rullkötter and Marchal1999; Schulte & Müller, Reference Schulte and Müller2001). The high productivity in the N and NE Arabian Sea during the glacial period has been related to the intensified NE monsoon winds inducing convective overturning and injection of nutrient-rich subsurface waters to the euphotic zone (Rostek et al. Reference Rostek, Bard, Beaufort, Sonzogni and Ganssen1997; Reichart et al. Reference Reichart, Lourens and Zachariasse1998). Records of high productivity in the EAS during glacial stages of the late Pleistocene have also been attributed to the intensification of NE winter monsoon winds (e.g. Ivanova et al. Reference Ivanova, Schiebel, Singh, Schmiedl, Niebler and Hemleben2003; Singh et al. Reference Singh, Kroon and Ganeshram2006, Reference Singh, Jung, Darling, Ganeshram, Ivanochko and Kroon2011, Reference Singh, Jung, Anand, Kroon and Ganeshram2018). Other studies have further suggested relatively low surface water productivity during interglacials, which were mainly caused by surface water stratification driven by high fluvial discharge/runoff from rivers flowing through the Western Ghats, mainly the Narmada and Tapti, when the SW monsoon was intensified and the NE monsoon circulation was weak (Singh et al. Reference Singh, Kroon and Ganeshram2006, Reference Singh, Jung, Anand, Kroon and Ganeshram2018; Cabarcos et al. Reference Cabarcos, Flores, Singh and Sierro2014). In contrast, the surface water was more stratified and oligotrophic during periods when our planktic foraminifera stable oxygen isotope record reveals lighter values, which are most likely associated with interglacial periods (Fig. 5). It is intriguing to note that high abundances of G. bulloides occur during periods of lighter δ18O values between 1.55 Ma and 1.65 Ma, which are interpreted to reflect interglacials (see above). We suggest that the SW summer monsoon was intensified during the interglacial periods and stronger upwelling appeared in the western Arabian Sea. Filaments of this upwelling probably reached the area of coring, resulting in a moderate increase in productivity and thus relatively higher abundances of G. bulloides during interglacial periods. In addition, a moderate increase in productivity might have been caused by enhanced fluvial runoff which supplied more nutrients to the EAS due to an interglacial strengthening of the SW summer monsoonal rainfall. However, the latter scenario cannot be fully proven by our datasets because our seawater δ18O record (proxy for salinity and thus changes in monsoonal rainfall) is of low resolution (Fig. 4). In summary, our proxy records suggest an overall weaker NE and SW monsoon during 1.85–2.7 Ma, and an intensification of the NE and SW monsoon between 1.55 Ma and 1.65 Ma.
Faunal, stable oxygen isotope and SST records for 1.55–1.7 Ma. Planktic foraminiferal (N. dutertrei, G. glutinata, G. ruber and G. bulloides) relative abundances, Mg/Ca-based SST record and δ18O record of G. sacculifer-quadrilobatus. Dark bands represent interglacials (IG).
5.c. Monsoon dynamics during the early Pleistocene
Our proxy records of the studied section at IODP Site U1457 reveal changes in upper water column structure occurring between c. 1.65 Ma and 1.85 Ma, most likely due to an intensification of the winter South Asian / Indian monsoon circulation system. Our study suggests that prior to 1.85 Ma, NE monsoon circulation was generally weaker than after 1.65 Ma. After 1.65 Ma, higher magnitudes of variations in productivity, SST and surface water stratification / vertical mixing on glacial/interglacial scale are suggestive of the intensification of the NE monsoon. Indian / South Asian summer monsoon variations have been reconstructed using marine (biogenic and lithogenic indices from the Arabian Sea; Clemens et al. Reference Clemens, Murray and Prell1996) and continental (Heqing palaeolake, pollen, total organic carbon (TOC), geochemical proxies; An et al. Reference An, Clemens, Shen, Qiang, Jin, Sun, Prell, Luo, Wang, Xu, Cai, Zhou, Liu, Liu, Shi, Yan, Xiao, Chang, Wu, Ai and Lu2011) records during the Pliocene–Pleistocene. A long-term decrease in the lithogenic grain-size fraction indicates that the persistent growth of the Northern Hemisphere (NH) ice sheets since 3.5 Ma has weakened the intensity of the Indian summer monsoon (Clemens et al. Reference Clemens, Murray and Prell1996). On glacial–interglacial timescales, the Heqing palaeolake record indicates the importance of interhemispheric forcing in driving Indian / South Asian summer monsoon variability at glacial–interglacial timescales (An et al. Reference An, Clemens, Shen, Qiang, Jin, Sun, Prell, Luo, Wang, Xu, Cai, Zhou, Liu, Liu, Shi, Yan, Xiao, Chang, Wu, Ai and Lu2011). An et al. (Reference An, Clemens, Shen, Qiang, Jin, Sun, Prell, Luo, Wang, Xu, Cai, Zhou, Liu, Liu, Shi, Yan, Xiao, Chang, Wu, Ai and Lu2011) found a dominant NH forcing that results in large-amplitude fluctuations of the ISM variation when ISM minima were coincident with NH ice volume maxima after c. 1.82 Ma. Increasing NH glaciation decreases the strength of the Indian low, resulting in a weakening of the ISM with the growth of NH ice sheets (An et al. Reference An, Clemens, Shen, Qiang, Jin, Sun, Prell, Luo, Wang, Xu, Cai, Zhou, Liu, Liu, Shi, Yan, Xiao, Chang, Wu, Ai and Lu2011). Although not conclusive, we suggest that increasing NH glaciations strengthened the NE (winter) winds, resulting in a higher magnitude of variations in productivity, SST and surface water stratification / vertical mixing after 1.65–1.85 Ma.
In addition, several studies have suggested drastic changes in atmospheric circulations in low- to mid-latitude regions since the mid-Pliocene (e.g. Jia et al. Reference Jia, Chen and Peng2008; Etourneau et al. Reference Etourneau, Martinez, Blanz and Schneider2009; Trauth et al. Reference Trauth, Larrasoaña and Mudelsee2009). Recent studies propose that the zonal and meridional atmospheric circulations (Walker Circulation and Hadley Cell), which are related to the E–W SST gradient along the equatorial Pacific, strengthened in the early Pleistocene after 2.2 Ma (e.g. Etourneau et al. Reference Etourneau, Schneider, Blanz and Martinez2010). Maximum E–W SST differences along the equatorial Pacific have been reported between 2.2 Ma and 1.5 Ma. A modern-like Walker Circulation with strong seasonality of the monsoon system was established during this period. This interpretation seems to be consistent with our findings and we could also assume that the strengthening of the Walker and Hadley circulations and the associated trade winds may have significantly amplified the intensity of the monsoonal winter winds in the EAS between 1.55 Ma and 1.65 Ma, resulting in the strengthening of surface water mixing and productivity.
6. Conclusions
Here we have presented the first evidence for changes in strength of the winter (NE) South Asian / Indian monsoon during the early Pleistocene. Based on the reconstruction of surface water conditions at IODP Site U1457 located in the EAS, we identified an intensification of the winter (NE) monsoon between 1.55 Ma and 1.65 Ma. This would probably have been initiated between ∼1.65 Ma and 1.85 Ma, associated with an increasing NH glaciation after c. 1.82 Ma. In addition, this major reorganization in South Asian / Indian monsoon circulation and intensification of the winter monsoon winds might be attributed to the development of strong meridional and zonal atmospheric circulations (i.e. Walker Circulation and Hadley Circulation). On the basis of our low-resolution records we cannot resolve the influence of interhemispheric forcing in driving the South Asian/winter monsoon variability at glacial and interglacial timescales during the early Pleistocene. Better-resolved sedimentary archives are needed to confirm and understand the evolution and development of the South Asian/winter monsoon during the early Pleistocene.
Author ORCIDs
Arun Deo Singh, 0000-0003-1287-8067.
Acknowledgements
Samples for this project were provided by the International Ocean Discovery Program. We are grateful to M. Mohtadi, J. Groeneveld and H. Kuhnert for supporting the isotope and trace element analyses. R.K.S. acknowledges CSIR for the Junior Research Fellowship. This study is also supported by the DST–PURSE Grant to A.D.S. We are grateful to the shipboard team (co-chief scientists, staff scientist and other scientists, technical staff and crew members on board JOIDES Resolution) of IODP Expedition 355. We thank the Editor and two anonymous reviewers for constructive comments and suggestions, which helped us to improve the manuscript. We also thank Pradyumna Singh for his help in identification of foraminifera, and Jayu Narvekar for preparation of the climatological maps.
