Marine Micropaleontology 101 (2013) 115–126

Home , Indonesian Throughflow, Timor Sea

Contents lists available at SciVerse ScienceDirect

Marine Micropaleontology

journal homepage: www.elsevier.com/locate/marmicro

Indonesian Throughﬂow and monsoon activity records in the Timor Sea since the last glacial maximum

X. Ding a,⁎, F. Bassinot b, F. Guichard b, N.Q. Fang a a China University of Geosciences (Beijing), Beijing 100083, China b Laboratoire des Sciences du Climat et de l'Environnement Domaine du CNRS, Gif-sur-Yvette 91198, France article info abstract

Article history: Indonesian Throughflow (ITF) is known to play an important role in the heat exchange between the Pacific Received 26 February 2012 and the Indian Oceans. However, our understanding of the long-term evolution of the ITF and, in particular, Received in revised form 7 February 2013 the mechanism of heat transport is limited. Here, we present a high-resolution foraminifera-based Accepted 10 February 2013 multi-proxy study in the main ITF outflow area of the Timor Sea, to reconstruct the ITF variability and to un- derstand the relationship between the ITF changes and monsoon activity from the last glacial maximum Keywords: (LGM) to the Holocene. Our results show that when the strong surface water ITF occurs, high productivity Timor Sea LGM is related to the mixing of the upper water column owing to the wind-driven upwelling rather than the Holocene shoaling of the depth of thermocline (DOT). By contrast, the DOT is affected more strongly by the ITF than Planktonic foraminifera by the monsoonal wind-driven upwelling in the Indonesian Seas. During the LGM (23–19 ka) and middle Ho- Indonesian Throughflow locene (8–6 ka), warm surface water ITF was dominated owing to the lowered sea level and (or) the higher Monsoon steric height difference between the western Pacific and eastern Indian Oceans as a result of the strong south- Sea level changes east monsoon. During the early Holocene (11–8 ka) and late Holocene (last ~6 ka), because of the postglacial high sea level, the strong northwest monsoon and heavy rains, large amounts of freshwater flowed into the Java Sea from the South China Sea (SCS). The freshwater plug at the southern tip of the Makassar Strait blocked the warm surface flow, thus initiating the enhanced thermocline ITF. In the Timor Sea, the changes in the vertical profile of the ITF were influenced by the glacio-eustatic sea-level changes that have modified the geometry of the pathways within the Indonesian Seas, as well as by the monsoon activity which was modulated by the changes in the insolation with a precessional cyclicity. © 2013 Elsevier B.V. All rights reserved.

1. Introduction mainly within the thermocline rather than at the sea surface (Gordon et al., 2003; Potemra et al., 2003; Song and Gordon, 2004; Gordon, Today, the wind stress between the Pacific Ocean and the Indian 2005), and thus the net effect in terms of heat transport to the Indian Ocean maintains a sea-level height difference between these two Ocean could be negative instead of being positive. oceans (Bray et al., 1996), as seen in the Indonesian Throughflow The major component of the ITF is the Mindanao Current that (ITF), a network of currents, surface and thermocline waters that originates from the upper thermocline of the North Pacific and is are transported from the western equatorial Pacific Ocean into the In- transported into the Indonesian Seas through the Makassar Strait dian Ocean (Hirst and Godfrey, 1993; Linsley et al., 2010). The ITF is (Gordon, 1986; Murray and Arief, 1988; Gordon and Fine, 1996) the only low-latitude connection along the return branch of the (Fig. 1). Within the Makassar Strait, the 680-m-deep Dewakang sill Great Conveyor Belt, which ultimately brings upper thermocline and permits only the upper thermocline waters to enter the Flores Sea surface waters from the Pacific to the North Atlantic (Gordon, 1986; and flow eastward to the Banda Sea, or to directly exit into the Indian Hirst and Godfrey, 1993; Bray et al., 1996; Gordon and Fine, 1996; Ocean via the shallow Lombok Strait (Sprintall et al., 2009). Today, Müller and Opdyke, 2000). The annual mean heat transport through the ITF transport of warm, low-salinity water into the Indian Ocean the Indonesian Throughflow region (about 1.4×1015 W) represents averages ~16 Sv (1 Sv=106 m3 s−1) per year (Gordon and Fine, a heat sink for the upper Pacific Ocean and is an important heat 1996; Schiller et al., 1998; Gordon et al., 2003). Mooring measure- source for the Indian Ocean (Schiller et al., 1998; Ganachaud and ments show that only a small portion (~1.7 Sv) (Murray and Arief, Wunsch, 2000). However, modeling experiments and recent oceano- 1988) of the waters flowing through the Makassar Strait across the graphic measurements indicate that the modern ITF transport occurs Indonesian Seas directly enters the Indian Ocean via the Lombok Strait (with a sill depth of 350 m) between the islands of Bali and ⁎ Corresponding author. Tel. +861082334643. Lombok (Fig. 1). The largest part of these waters turns eastward E-mail address: [email protected] (X. Ding). into the Banda Sea and Flores Sea before spreading into the Indian

Fig. 1. Locations of core sites MD98-2172, SHI9034 and SHI9022 and those referred to in the text, with main oceanographic surface current (black line arrows), thermocline current (black dashed arrows) and main geographic locations mentioned in the text. Core SHI9016 was analyzed by Spooner et al. (2005) and core MD01-2378 by Xu et al. (2006).

Ocean through two main pathways: the Ombai Strait (with a sill depth foraminifera, Linsley et al. (2010) suggested that the freshening of the of 3250 m and a yearly average flow of ~5.0±1 Sv (Molcard et al., surface ocean in the southern Makassar Strait 9.5 ka ago increased 2001)) and the Timor Sea (with a sill depth of 1890 m and a yearly av- the northward pressure gradient and inhibited the flow of warmer erage flow of 7.0 Sv (Creswell et al., 1993)) (Fig. 1). These ITF waters surface-layer water into the Indian Ocean. Thus, 9.5 ka may have flow into the Indian Ocean as parts of the west-flowing South Java Cur- marked the initiation of the thermocline-enhanced cool ITF transport rent, the South Equatorial Current, and the south-flowing Leeuwin Cur- that is observed today. Xu et al. (2006, 2008) used a multi-proxy (plank- rent that runs along the western Australian margin (Gordon and Fine, tonic foraminiferal census data and oxygen isotopic and Mg/Ca records) 1996; Siedler et al., 2001)(Fig. 1). approach to reconstruct changes in the vertical profile of the ITF (depth The modern climate of the Indonesian Seas is dominated by bi-annual of the thermocline and changes in sea surface and upper thermocline monsoonal shifts. Heavy rain accompanies northwesterly winds between temperature), as well as monsoonal wind and precipitation variations in November and March (Austral summer), during the Northwest (NW) the Timor Sea during Terminations I and II. These studies indicated that Monsoon. The dry season corresponds to the Southeast (SE) Monsoon pe- the vertical structure of the ITF probably varied considerably over preces- riod from May to September (Austral winter) (Spooner et al., 2005). sional and glacial–interglacial time scales, with the thermocline flow The intertropical convergence zone (ITCZ), a pressure trough dominating during warm periods, i.e. the thermocline shoaled from the where the southeast and northeast trade winds meet, usually lies last glacial maximum (LGM) to the Holocene. However, through recon- about 10°–15° north of the equator in the Austral winter and migrates structions of the vertical structure of the water column in the Banda Sea southward, close to or over northern Australia in the Austral summer over the last ~80 ka using the abundance ratio of the planktonic forami- (Spooner et al., 2005). During the Austral summer, the NW Monsoon nifera thermocline and mixed-layer dwellers, Spooner et al. (2005) gathers large amounts of moisture while crossing the sea from the showed that the mixed layer was thinner during the LGM, but thickened Asian high-pressure belt on its way to the ITCZ, which has shifted south- at the beginning of the Holocene. The latter may have been related to the ward. At the ITCZ, the moisture-laden air rises, resulting in heavy rains strengthened surface water ITF in the early Holocene. Owing to the differ- (van der Kaars et al., 2000). During the Austral winter, the SE Monsoon ent climate proxies used in these studies, the existing data appear to be originates from the Southern Hemisphere high-pressure belt and is rel- inconsistent. atively dry and cool (van der Kaars et al., 2000; Spooner et al., 2005). A multi-proxy study in the main ITF outflow area and Java upwell- The modern ITF is also closely related to the Asian monsoon dy- ing area is presented here, with the goal of comparing the results namics. The main flow of the ITF in the key passages of the Makassar obtained using different climate proxies and reconstructing the ITF and Timor Straits shows a strong seasonal variability (Gordon et al., variations, hence improving our understanding on the relationship 1999; Potemra et al., 2003). During the NW Monsoon (Austral sum- between the ITF changes and monsoon activity from the LGM to the mer), a thermocline flow of relatively cool water dominates, as the Holocene. Our main objectives are to track changes of the ITF outflow warm surface flow becomes blocked by the development of a fresh- during a period of extreme climate change and sea-level variations water plug at the southern tip of the Makassar Strait, driven by mon- and to assess links between high- and low-latitude climate changes. soonal winds from the Java Sea (Gordon et al., 2003; Gordon, 2005). As a result, the tropical Indian Ocean is cooled rather than warmed 2. Materials and methods by the ITF (Song and Gordon, 2004). On the Milankovitch time scale, two main mechanisms may affect Core MD98-2172 (8°31′S, 128°09′E) was obtained from the Timor the ITF: (1) orbitally driven, low-latitude changes in insolation that Sea at a water depth of 1768 m during the International Marine Glob- affect monsoon dynamics and (2) glacio-eustatic sea-level changes al Change Study (IMAGES) cruise IV of the R/V Marion Dufresne that modify the geometry of the pathways within the Indonesian (Fig. 1). This Calypso giant piston core is 54 m long, but only work Seas. We have very limited understanding of the ITF evolution at carried out on the upper 7.5 m of the core is discussed in this paper. this time scale. Using the records of the δ18O and Mg/Ca of planktonic Core MD98-2172 was sampled at 2 to 10-cm intervals for stable X. Ding et al. / Marine Micropaleontology 101 (2013) 115–126 117 isotope analysis of foraminifera and at 10-cm intervals for foraminifer Table 1 census, carbonate content measurements, and grain size analysis. Radiocarbon dates from cores MD-982172, SHI9022 and SHI9034. After adjusted by subtracting 400 years for the marine reservoir effect, 14C dates were calibrated using After being dried and weighed, samples were washed through a the radiocarbon calibration program Fairbanks0701. 150-μm sieve. The coarse fraction was dried at 40 °C overnight, and 14 14 then split with a micro-splitter to provide a subsample with at least Lab code Depth C dates C cal 300 whole tests of foraminifera which were identified and counted. Fora- (cm) (year BP) (year BP) miniferal fragments were also counted to provide a dissolution index MD98-2172 – (Thunell, 1976) and assess any possible bias resulting from the preferen- GifA 101513 0 2 1330±80 845±87 GifA 101512 350–351 9120±100 9702±164 tial dissolution of the more fragile planktonic species. Planktonic forami- GifA 101511 418–419 11,310±130 12,803±109 niferal census data were used to estimate the depth of the thermocline GifA 101510 465–466 13,270±290 14,973±382 (DOT), based on the transfer function developed by Andreasen and GifA 101509 511–512 15,210±180 17,726±375 Ravelo (1997) for the tropical Pacific Ocean, where the 18 °C isothermal GifA 101508 581–582 18,250±190 21,159±380 is arbitrarily defined as the DOT. Today, the position of the 18 °C isother- SHI9022 mal in the Timor Sea is located at 170 m, corresponding to the depth GifA 95334 84–85 3800±70 3644±89 range at which water temperature changes most rapidly (Xu et al., GifA 95335 154–155 5580±70 5930±74 2006). The transfer function equation has a standard error of ±22 m, GifA 95336 204–205.5 6880±80 7397±69 – and an additional ±5 m error is introduced by low species counts in GifA 95339 385 386 9800±90 10,630±131 GifA 95340 474.5–475.5 14,850±130 17,053±255 the core-top database (Andreasen and Ravelo, 1997). GifA 95341 504.5–505.5 15,110±130 17,529±288 A total of 10–15 specimens of planktonic foraminifer Globigerinoides GifA 95342 594.2–595.8 18,620±160 21,713±267 ruber in the size range of 250–315 μmand4–6 specimens (>250 μm) GifA 95343 694.2–695.8 26,290±300 31,112±352 of the epifaunal benthic foraminifer Cibicides wuellerstorfi were handpicked and then ultrasonically cleaned in methanol for less than SHI9034 GifA 9504 34–36 3110±60 2802±53 10 s. Carbon and oxygen stable isotopes were measured using a GifA 9505 44–46 3930±60 3813±82 Finnigan MAT-251 mass spectrometer at the Laboratoire des Sciences GifA 9506 105–106 5960±70 6340±62 du Climat et de l'Environnement, Gif-sur-Yvette, France. The external GifA 9507 125–126 6610±60 7122±87 – reproducibility is ~0.06‰ for δ18O and 0.03‰ for δ13C. Isotopic data GifA 9508 360 362 12,360±90 13,780±83 GifA 9510 390–391 12,760±90 14,209±78 are reported relative to the PDB standard through calibration to the GifA 9511 430–431 14,350±100 16,267±177 NBS19 standard. AMS 14C ages were obtained on monospecific samples of G. ruber (>250 μm) picked from six intervals. The specimens were ultrasonically cleaned in distilled water and then analyzed with the Tandetron foraminiferal δ18O records. Depths were converted to calendar ages Accelerator in Gif-sur-Yvette. Foraminifera 14C ages were adjusted for by linear interpolation between stratigraphic control points. Ages the apparent reservoir effect on the ages of surface seawater by for the older samples in these cores were extrapolated by assuming subtracting 400 years; then, dates were calibrated using the radiocar- a constant sedimentation rate throughout the LGM. Clark et al. bon calibration program Fairbanks0701 (Fairbanks et al., 2005). The (2009) recently proposed to extend the duration of the LGM based ages are given in Table 1. on the careful study of maximum extension of ice sheets, and Carbonate weight content (in percent) was measured on dry bulk suggested that it lasted from 26.5 to 19 ka (cal yr BP). However, in samples using a standard titration method at the School of Marine the present study, we decided to stick to the time interval 23–19 ka Sciences, China University of Geosciences (Beijing), and the analytical defined by the EPILOG working group (Mix et al., 2001). This would error was ±0.2%. allow us to remain in phase with numerous studies over the last de- For grain size analysis, 0.5–1 g samples were placed in beakers with cades in which the definition had been used. ~25 ml water. 2–3 drops of pure hydrogen peroxide and 1–2ml10% The temporal resolution between samples is ~200 years for δ18O hydrochloric acid solutions were added in the beakers. After the carbon- records, ~400 years for the planktonic foraminifera census in core ate and organic carbon fractions were completely removed, the grain-size MD98-2172 and ~400 years in core SHI9034, and the average sedi- (>65 μm, 65–3 μmandb3 μm) distribution of the non-carbonate, terrig- mentation rate is ~26 cm/kyr in core MD98-2172 and ~22 cm/kyr enous material was analyzed with a Mastersizer 2000 laser grain size an- in core SHI9034. The resolution in the upper 500 cm (after 17.5 ka) alyzer at the School of Marine Sciences, China University of Geosciences of core SHI9022 is also ~200 years, but lower downcore, and the av- (Beijing). erage sedimentation rate is ~22 cm/kyr. Two other piston cores, SHI9022 and SHI9034, were also studied. The first one was obtained from the Timor Sea (11°35′S, 122°03′E, 3.2. Stable isotope record water depth 2313 m, length 7.11 m), and the second from the Java upwelling area (9°09′S, 111°01′E, water depth 3330 m, length 8.84 m) Over the time interval covered, the planktonic δ18O values varied be- during a joint French–Indonesian marine geological research program tween −0.89‰ and −2.93‰ in core MD98-2172, between −0.83‰ in February 1990 (Fig. 1). The paleoceanographic proxy records of and −2.97‰ in core SHI9022, and between −0.74‰ and −3.35‰ in cores SHI9022 and SHI9034 (the G. ruber δ18O and the planktonic fora- core SHI9034 (Fig. 2). The planktonic and benthic δ18O values were minifera census data) had previously been used to analyze the changes heavier during the LGM (Fig. 2). Between 19 and 11 ka, the planktonic in the Indonesian Seas heat transport pathways (Ding et al., 2002, and benthic δ18O curves displayed sudden distinct declines. During 2006). The AMS 14C ages of the two cores were recalibrated using Fair- the interval of global warming from the end of the LGM approximately banks0701 (Fairbanks et al., 2005). 19 ka to the early Holocene 11 ka, virtually every component of the climate system underwent large-scale changes (Clark et al., 2012). This 3. Results dramatic time of global change was triggered by changes in insolation (Clark et al., 2012). The maximum of summer solstice insolation at 3.1. Age models 65°N is at 11 ka (Laskar et al., 2004). Here, we discuss this 19–11 ka interval as the last deglaciation, and the interval after 11 ka as the Holocene. The age models for cores MD98-2172, SHI9022, and SHI9034 were For core MD98-2172, the planktonic and benthic δ13C values var- based on the calibrated radiocarbon dates (Table 1) and planktonic ied between 0.64‰ and 1.84‰ and between −0.29‰ and 0.48‰, 118 X. Ding et al. / Marine Micropaleontology 101 (2013) 115–126

18 13 13 Fig. 2. Plots of planktonic and benthic foraminifera δ O and δ C and the carbon isotope difference (Δδ CPF−BF) between planktonic and benthic foraminifera of cores MD-982172, and the δ18O contrast of three cores MD98-2172, SHI9022, and SHI9034. Gray shadings mark the Last Glacial Maximum (LGM, 23–19 ka), the late deglaciation after the meltwater pulse (MWP)-1A (14.5–11 ka) and strong Southeast (SE) Monsoon periods (~8–6 ka) respectively.

respectively (Fig. 2). The difference in δ13C values between planktonic decrease during deglaciation and the early Holocene, and an increase 13 and benthic foraminifera (Δδ CPF−BF) was high during the LGM and again about 8–6 ka; Figs. 3 and 4). during 7–5 ka in the Holocene (Fig. 2). In cores MD98-2172 and SHI9022, the abundance of Neogloboquadrina dutertrei was high during the LGM, reaching a striking maximum peak in the early deglaciation (Figs. 3 and 4), 3.3. Planktonic foraminifera abundance and decreasing abruptly after ~14.5 ka, then continuing to decrease steadily and slowly into the Holocene. In core SHI9034, on the other Variations of planktonic foraminiferal abundances in MD98-2172, hand, the abundance of this species was high during the LGM and degla- SHI9022 and SHI9034 are plotted in Figs. 3–5. The abundance of ciation, and decreased in the Holocene (Fig. 5). Globigerina bulloides in core MD98-2172 is high during the LGM and The evidence for polar ice-margin retreat occurring between 20 at about 8–6 ka in the Holocene, but low during the late deglaciation and 19 ka indicates that the 19 ka meltwater pulse (MWP), which (Fig. 3). The G. bulloides peak abundance at about 8–6 ka is even more represents a rapid 10 m rise in sea level from the LGM lowstand obvious in core SHI9022 (Fig. 4). For core SHI9034, collected in the sometime between 20 and 19 ka, originated from the Northern Hemi- Java upwelling area, the abundance of G. bulloides is higher, and the sphere ice sheets (Clark et al., 2009). In the Southern Hemisphere, an peak abundance of this species in the Holocene appeared earlier, with abrupt rise in sea level at 14.5 ka is referred to as MWP-1A (Clark et higher values and longer duration (about 10.5–6 ka) than in cores al., 2009). MD98-2172 and SHI9022 (Fig. 5). However, the variation in G. bulloides Similar to N. dutertrei, the dextrally coiled Neogloboquadrina abundance during the LGM in core SHI9022 does not resemble that ob- pachyderma was most abundant during the LGM and in the early de- served in core MD98-2172 (Figs. 3 and 4). glaciation but decreased rapidly after ~14.5 ka and was only a minor Variations in the abundance of Globigerinita glutinata from core species during the Holocene interval in the three cores (Figs. 3–5). SHI9022 is similar to what was observed for the G. bulloides record In cores MD98-2172 and SHI9022, the abundance of Globigerinoides from core MD98-2172 (high abundance in the LGM, followed by a sacculifer was low during the LGM, but it increased gradually since the

Fig. 3. The abundance of the important planktonic foraminifera species of core MD98-2172. Gray shadings mark the Last Glacial Maximum (LGM, 23–19 ka), the late deglaciation after the meltwater pulse (MWP)-1A (14.5–11 ka) and strong Southeast (SE) Monsoon periods (~8–6 ka) respectively. X. Ding et al. / Marine Micropaleontology 101 (2013) 115–126 119

Fig. 4. The abundance of the important planktonic foraminifera species of core SHI9022. Gray shadings mark the Last Glacial Maximum (LGM, 23–19 ka), the late deglaciation after the meltwater pulse (MWP)-1A (14.5–11 ka) and strong Southeast (SE) Monsoon periods (~8–6 ka) respectively. deglaciation MWP-1A ~14.5 ka, and reached a maximum peak during the content decreased rapidly across the deglaciation and then increased early Holocene, and then declined abruptly around 8–6 ka. After ~5 ka, gradually again in the Holocene, with a noticeable, striking minimum at the abundance of G. sacculifer decreased towards the top of both cores about 9–6.5 ka. The 3–65 μmmedium-grainandb3 μm fine-grain varia- (Figs. 3 and 4). This phenomenon is different from that of core SHI9034, tion displayed a marked anticorrelation, with the former showing high in which G. sacculifer was most abundant during ~5–3kaperiod(Fig. 5). content in the LGM and low contents from the deglaciation, and the latter showing the opposite (Fig. 6). 3.4. Carbonate content 4. Discussion Carbonate content of core MD98-2172 was low during the LGM, then increased noticeably after ~17 ka (Fig. 6) and reached the 4.1. High productivity periods highest values of 43.5% during the late Holocene. 13 The foraminifer fragment content indicates that carbonate disso- Δδ CPF−BF values have been used for qualitative estimates of ex- lution was lower in the LGM and the deglaciation and higher during port productivity in tropical and subtropical oceans, with greater the Holocene, especially its latter part after ~3 ka (Fig. 6). These re- values during the period representing enhanced productivity (Jian 13 sults suggest, therefore, that dissolution played a minor role in con- et al., 2001). The Δδ CPF−BF values of core MD98-2172 were high trolling the CaCO3 percentage with the exception of the latter part during the LGM and ~7–5 ka in the Holocene (Fig. 2). of the Holocene, when the carbonate content was inversely correlated G. bulloides mainly occurs in subpolar regions and is also commonly with the foraminifer fragment content. encountered in upwelling areas and boundary currents in low-latitude regions where surface productivity is high (Bé, 1977; Duplessy et al., 3.5. Sediment grain size changes 1981; Prell and Curry, 1981; Brock et al., 1992; Martinez et al., 1998; Pflaumann and Jian, 1999). G. glutinata is known to have a wide latitu- Granulometric analysis performed on the noncarbonate fraction of dinal distribution. It can tolerate a rather extensive range of tempera- MD98-2172 samples indicates that the silt fraction (3–65 μm) is the larg- tures and salinities and is moderately susceptible to dissolution. Its est contributor to the sediments (≥60% in mass of the carbonate-free distribution resembles that of G. bulloides by having a high abundance fraction), the clay fraction (b3 μm) represents about 30%, and the in mid to high latitudes and also in upwelling regions of low latitudes >65 μm coarse fraction represents only between 0.5% and 6% (Fig. 6). that are characterized by fertile waters (Fairbanks et al., 1982; Thunell The coarse fraction content was higher in the LGM. The coarse-grain and Reynolds, 1984; Martinez et al., 1998; Pflaumann and Jian, 1999;

Fig. 5. The abundance of the important planktonic foraminifera species of core SHI9034. Gray shadings mark the Last Glacial Maximum (LGM, 23–19 ka), the late deglaciation after the meltwater pulse (MWP)-1A (14.5–11 ka) and strong Southeast (SE) Monsoon periods (~8–6 ka) respectively. 120 X. Ding et al. / Marine Micropaleontology 101 (2013) 115–126

18 Fig. 6. Plots for the planktonic foraminifera δ O, CaCO3 content, foraminifera fragment content and the >65 μm, 3−65 μm, and b3 μm grain-size for core MD98-2172. Gray shadings mark the Last Glacial Maximum (LGM, 23–19 ka), the late deglaciation after the meltwater pulse (MWP)-1A (14.5–11 ka) and strong Southeast (SE) Monsoon periods (~8–6ka) respectively.

Kawahata et al., 2002). The abundances of the two species G. bulloides the LGM and the early deglaciation in cores MD98-2172 and SHI9022 and G. glutinata were high during the LGM and at about 8–6kainthe (Figs. 3 and 4; Table 2). Holocene, low during the late deglaciation and early Holocene in cores The highest abundance of the warm-water species G. sacculifer is re- MD98-2172 and SHI9022 (Figs. 3 and 4; Table 2). stricted to oligotrophic and mesotrophic conditions. The abundance of N. dutertrei is known to be a tropical to temperate species that shows this species can reach high values in areas with very low export produc- a tolerance to large sea-surface temperature (SST) and sea-surface salin- tion rates (Žarić et al., 2005). In cores MD98-2172 and SHI9022, the ity (SSS) changes and is abundant in active current systems, along conti- abundance of G. sacculifer was low during the LGM and around 8–6ka nental margins, and in upwelling regions (Fairbanks et al., 1982; Curry et (Figs. 3 and 4; Table 2). 13 al., 1983; Cannariato and Ravelo, 1997; Kawahata et al., 2002). The abun- In summary, our Δδ CPF−BF data from core MD98-2172 and dance of N. dutertrei is the highest in high productivity areas with a changes of the abundances in the productivity indicator species of warm temperature and a low salinity (Bé, 1977; Prell and Curry, 1981; planktonic foraminifera for cores MD98-2172 and SHI9022 all indi- Thunell and Reynolds, 1984; Thiede and Jünge, 1992; Pflaumann et al., cate that productivity was high during the LGM and early deglacia- 1996; Hilbrecht, 1997; Martinez et al., 1998; Pflaumann and Jian, 1999). tion, then decreased after the deglaciation MWP-1A event, and The dextrally coiled N. pachyderma, an important constituent of subpolar became low during the Holocene, except a brief increase around and transitional assemblages, also occurs in low abundances in the equa- 8–6 ka in the Timor Sea. torial Indian Ocean (Bé, 1977). This is attributed to cool-water upwelling in coastal regions of the tropical and subtropical Atlantic (Pflaumann and 4.2. Changes in the depth of thermocline (DOT) Jian, 1999). In the northern South China Sea, N. pachyderma (dex.) has been found in small numbers along with G. bulloides, representing the The DOT controls the vertical distribution of planktonic foraminif- coolest faunal assemblage there (Pflaumann and Jian, 1999). The two spe- era in the upper water column of the oceans (Bé, 1977; Ravelo et al., cies N. dutertrei and N. pachyderma (dex.) were the most abundant during 1990). G. ruber and G. sacculifer live in the upper layer of the ocean and

Table 2 Average abundance of the productivity indicator species and the depth of the thermocline in the various intervals for cores MD98-2172, SHI9022 and SHI9034 from the LGM to the Holocene.

Time Core Average abundance of the productivity indicator species (%) DOT (m)

G. bulloides N. dutertrei N. Pachyderma (dex.) G. glutinata G. sacculifer Ranging Average

Late Holocene ~6 ka MD98-2172 8.3 8.1 1.4 16.3 7.1 127–179 155 SHI9022 8.0 10.8 2.5 18.8 5.1 147–204 174 SHI9034 10.0 15.7 0.7 6.7 8.2 92–163 126 Middle Holocene 8–6 ka MD98-2172 10.2 6.2 2.3 17.1 7.1 164–197 183 SHI9022 8.3 11.3 2.7 16.6 7.9 175–211 190 SHI9034 21.7 18.6 0.8 8.0 6.0 166–210 187 Early Holocene 11–8 ka MD98-2172 8.5 5.2 2.2 15.7 9.6 159–190 175 SHI9022 6.4 9.4 2.8 13.3 11.8 155–198 175 SHI9034 30.0 18.7 0.5 7.6 4.7 113–191 158 Late deglaciation 14.5–11 ka MD98-2172 6.6 8.1 3.7 15.5 6.3 162–193 185 SHI9022 4.7 11.8 2.8 15.2 9.0 158–199 179 SHI9034 19.4 23.0 0.5 7.9 4.0 100–156 122 Early deglaciation 19–14.5 ka MD98-2172 8.6 12.7 6.1 14.0 3.6 158–196 173 SHI9022 5.5 17.0 4.7 18.2 3.8 132–201 173 SHI9034 15.1 23.0 12.2 8.6 3.1 91–135 112 LGM 23–19 ka MD98-2172 9.9 11.3 5.4 19.6 1.9 156–202 179 SHI9022 4.9 12.5 5.7 23.3 1.7 184–197 190 SHI9034 12.3 22.5 11.9 9.4 1.6 83–175 109 X. Ding et al. / Marine Micropaleontology 101 (2013) 115–126 121 are, therefore, mixed-layer dwelling species. N. dutertrei, N. pachyderma, the early Holocene ~10–5 ka, the DOT was shallow throughout core G. bulloides and Globorotalia crassaformis are thermocline-dwelling SHI9034 (Fig. 7, Table 2). species, living below the mixed layer (Ravelo et al., 1990; Huang et al., 2002; Spooner et al., 2005). In general, mixed-layer-dwelling species increase in abundance when the DOT deepens, whereas 4.3. Indonesian Throughflow and monsoon thermocline-dwelling species increase when it shoals (Ravelo et al., 1990; Huang et al., 2002). Consequently, the abundance of mixed-layer- 4.3.1. Last glacial maximum (23–19 ka) or thermocline-dwelling planktonic foraminiferal species can be used as The non-carbonate sediment grain size plots for core MD98-2172 a qualitative index of DOT variations (Ravelo et al., 1990; Xu et al., 2006). show that the >65 μm and 3–65 μm grain size content displayed The thermocline-dwelling foraminifera in core MD98-2172 were peak values during the LGM; in contrast, the b3 μm fine-grain content most abundant during the LGM and early deglaciation, and they de- was relatively low (Fig. 6). These results indicate that sediments were clined rapidly after the deglaciation MWP-1A interval. Instead, the coarsened with the mass terrestrial discharge owing to the declining mixed-layer dwellers were most abundant during the late deglacia- sea level during the LGM. tion and Holocene intervals (Fig. 7). This is similar to the changes of The high-productivity species abundance of the planktonic forami- 13 the adjacent core SHI9016 (Spooner et al., 2005)(Fig. 7). However, nifera in both core MD98-2172 and core SHI9022 and the Δδ CPF−BF for core MD98-2172, the abundance plot of mixed-layer-dwelling of core MD98-2172, all indicate that productivity was high during the foraminifera does not parallel the DOT values estimated using the trans- LGM. This is in accord with previous studies (Martinez et al., 1998, fer function from Andreasen and Ravelo (1997), especially during the 1999; Takahashi and Okada, 2000; Gingele et al., 2001; Holbourn et LGM (Fig. 7). al., 2005; Spooner et al., 2005), which indicates a strong SE Monsoon ac- The Mg/Ca temperature difference from surface- and thermocline- tivity or enhanced mixing during the LGM and an increasing supply of dwelling planktonic foraminifera (thermal gradient ΔT(G. ruber− terrigenous nutrition owing to the drop in the sea level. However, dur- Pulleniatina obliquiloculata)) is interpreted as changes in the DOT, with ing this time interval, the Austral summer insolation (20°S, Dec.) was large ΔT values reflecting a shallow DOT (Anand et al., 2003; Xu et al., high compared to the early Holocene (Berger, 1978)(Fig. 8), which like- 2006). Xu et al. (2006) compared the DOT estimated by using the trans- ly resulted in strong NW Monsoon activity and heavy rains over land fer function with ΔT(G. ruber−P. obliquiloculata) in the Timor Sea during Termi- masses. nation II. The results showed the variations in ΔT were consistent with The vegetation over land masses indicates drier climates culminat- fauna-based DOT estimates. For the last ~23 ka, DOT values estimated ing at the LGM (Barmawidjaja et al., 1993; van der Kaars and Dam, by using the transfer function for core MD98-2172 and core SHI9022 1995; Wang et al., 1999). Studies in Australia by Magee et al. (1995), exhibited similar changes to those of ΔT(G. ruber−P. obliquiloculata) for core Magee and Miller (1998), Veeh et al. (2000), van der Kaars and De MD01-2378 (Xu et al., 2006)(Fig. 7). Especially in core SHI9022, the Deckker (2002), Hesse and McTainsh (2003),andHesse et al. (2004) latest Holocene DOT at ~176 m obtained by using the transfer function also indicated drier conditions during MIS 3 and 2, suggesting greater is close to the modern 18 °C isotherm depth (~170 m) in the Timor influence of the SE Monsoon. Wyrwoll and Miller (2001) and van der Sea (Xu et al., 2006). Thus, this may suggest that the use of the transfer Kaars and De Deckker (2002) estimated the NW Monsoon “switching function is a better approach for reconstructing DOT variability than on” in Australia at 14 ka. They presumed that the southward shift of using the abundance of the mixed-layer or thermocline-dwelling fora- the ITCZ in the Austral summer may have been considerably restricted, minifera in the Indonesian area. and the ITCZ lay north of the Banda Sea during glacial times, causing the The changes in DOT of core MD98-2172 and core SHI9022 exhibit trade winds to blow across the Banda Sea (Spooner et al., 2005; Xu et al., high-frequency variations but the longer term oscillations are similar 2006)(Fig. 9). This essentially has the effect of setting a “perpetual” SE to those of ΔT(G. ruber−P. obliquiloculata) of core MD01-2378. Three main Monsoon into operation (Barrows and Juggins, 2005). plateaus are recognized at 23–19 ka, 14.5–11 ka, and 8–6 ka, separat- We note that, unlike today when the SE Monsoon promotes a thinner ed by periods of shoaling of the thermocline between 19 and 14.5 ka, mixed layer and a shallower thermocline depth, the DOT records for both 11–8 ka, and since 6 ka (Fig. 7, Table 2). However, the data for core cores MD98-2172 and SHI9022 were deeper during the LGM (Fig. 8). The

SHI9034 do not parallel the DOT changes in the other two cores DOT indicated by ΔT(G. ruber−P. obliquiloculata) of mixed-layer-dwelling spe- MD98-2172 and SHI9022. Apart from a remarkable deepening in cies G. ruber and thermocline-dwelling species P. obliquiloculata in core

Fig. 7. The abundance of the planktonic foraminifera thermocline and mixed-layer dwellers of cores MD98-2172 and SHI9016 (Spooner et al., 2005). The depths of the thermocline

(DOT) of the cores MD98-2172, SHI9022, and SHI9034. The ΔT(G. ruber−P. obliquiloculata) of the core MD01-2378 (Xu et al., 2008). Gray shadings mark the Last Glacial Maximum (LGM, 23–19 ka), the late deglaciation after the meltwater pulse (MWP)-1A (14.5–11 ka) and strong Southeast (SE) Monsoon periods (~8–6 ka) respectively. The arrows indicate that the DOTs shoal. 122 X. Ding et al. / Marine Micropaleontology 101 (2013) 115–126

Fig. 8. The ΔT(G. ruber−P. obliquiloculata) shown is for core MD01-2378 (Xu et al., 2008). The depths of the thermocline (DOT) and the abundances of G. bulloides from cores MD98-2172, SHI9022, and SHI9034. The insolation curves for 20°S Dec. (red line) and 30°N Jun. (blue line) (Berger, 1978) are also shown for comparison. Gray shadings mark the Last Glacial Maximum (LGM, 23–19 ka), the late deglaciation after the meltwater pulse (MWP)-1A (14.5–11 ka) and strong Southeast (SE) Monsoon periods (~8–6 ka) respectively. The arrows indicate that the DOTs shoal.

MD01-2378 situated at the main ITF outflow (Xu et al., 2008) also showed N. pachyderma (dex.) decreased abruptly after 14.5 ka. At the same a similar shift. time, the abundance of the high-productivity species G. bulloides Today, the South China Sea (SCS) is a main heat and freshwater con- dropped dramatically whereas that of the oligotrophic species G. sacculifer veyor, as it provides a major pathway for surface waters flowing from rose in both cores MD98-2172 and SHI9022 (Figs. 3 and 4). Furthermore, 13 the west Pacific into the Indonesian Seas (Qu et al., 2009). The export for core MD98-2172, the Δδ CPF−BF value also dropped gradually of fresh surface water from the SCS via the Java Sea to the Bali Sea was (Fig. 2), the coarse-grain content decreased rapidly, and b3 μm fine- shut off during the LGM, when the Sunda land and large parts of the grain content was high across the deglaciation (Fig. 6). Java Sea were exposed (De Deckker et al., 2002). Previous studies in- To explain these observations, we propose that the sea level rose cluding circulation modeling experiments suggested that the ITF was rapidly after 19 ka, resulting in the widening of the main ITF outflow dominated by surface flow during the LGM, when trade winds intensi- pathways, which made it possible for warm, low-salinity surface fied, precipitation over Borneo decreased (De Deckker et al., 2002; water to pour into the eastern Indian Ocean through the Lombok Partin et al., 2007), and the passage to the SCS was blocked (Tozuka et and Ombai Straits, and the Timor Sea, which diverted the surface al., 2007). Furthermore, the higher steric height difference between water ITF away from the west Pacific Ocean. Furthermore, the declin- the western Pacific and the eastern Indian Ocean owing to a “perpetual” ing productivity indicates a weaker “perpetual” SE Monsoon, which SE Monsoon, and the narrowed main ITF outflow pathways resulting probably decreased the steric height difference between the west Pa- from the lowered sea level, would lead to a strong surface water ITF cific and the eastern Indian Ocean. Thus, the decreasing water pres- and deepening of DOTs during the LGM (Figs. 8 and 9). sure gradient between the two ocean basins caused surface water The modern hydrography and environments of the Timor Sea ITF weakening and DOT shoaling during the early deglaciation in show that a high nutrient content (annual mean ~30 μmol/l) and a the Timor Sea (Fig. 9). low oxygen level (annual mean ~1.6 ml/l) occur below water depth ~400 m, but the nutrient content in the water column below ~50 m increases gradually from 2 to 30 μmol/l along with the water depth 4.3.3. Early Holocene (11–8 ka) (Levitus, 1998; Ding et al., 2006). During the LGM, the average DOTs The DOT estimated for cores MD98-2172 and SHI9022 shoaled of the MD98-2172 and SHI9022 core sites were ~180 m or deeper during the early Holocene. This phenomenon is even better observed (Table 2); hence intensified mixing of the upper water column at the core MD01-2378 site for which the DOT was indicated by the owing to a “perpetual” SE Monsoon wind-driven upwelling may ΔT(G. ruber−P. obliquiloculata) (Fig. 7). The abundances of foraminifer spe- have elevated the nutrient content, leading to high productivity. cies indicative of high productivity conditions in cores MD98-2172 Therefore, it is inferred that, because of the stronger surface water and SHI9022 all show a slightly declined productivity during this pe- ITF, the high productivity was related to the mixing of the upper riod (Figs. 3 and 4). At the same time, the content of the oligotrophic water column due to the wind-driven upwelling, but it is not neces- species G. sacculifer in the two cores reached a peak value (Figs. 3 and 4). sarily related to shallow DOTs. Instead, the DOTs were affected by The >65 μm coarse-fraction content rose again in the early Holocene the ITF more than by the monsoonal wind-driven upwelling in the In- (Fig. 6). donesian Seas. It was estimated that the Australian NW Monsoon switched on at 14 ka (Wyrwoll and Miller, 2001; van der Kaars and De Deckker, 4.3.2. Deglaciation (19–11 ka) 2002). Spooner et al. (2005) also suggested that the NW Monsoon The DOTs reconstructed from cores MD98-2172 and core seemed to strengthen at 10.3 ka, indicating that the bi-annual mon- SHI9022 shoaled during the early deglaciation and became stabi- soonal system was most intense at this time. The strong NW Monsoon lized after 14.5 ka. This is consistent with the DOT indicated by and heavy rains may have resulted in increasing freshwater and de- the ΔT(G. ruber− P. obliquiloculata) found in core MD01-2378 (Xu et al., creasing surface productivity during the NW Monsoon season at 2008)(Fig. 8). After reaching their maximum values in the early degla- that period and carried even coarser sediments from the western ciation, the abundances of the high-productivity species N. dutertrei and land to the studied area. X. Ding et al. / Marine Micropaleontology 101 (2013) 115–126 123

Fig. 9. The various scenarios postulated for the ITF during the LGM (23-19 ka), deglaciation (19-11 ka), early Holocene (11-8 ka), middle Holocene (8-6 ka) and late Holocene (last ~6 ka). The position of Intertropical Convergence Zone in the Austral summer during the LGM (e) is referred from De Deckker et al. (2002) and Xu et al. (2006).

At about 9.5 ka, when the postglacial sea level rise reached −30 m the Makassar Strait blocked the warm surface flow, thus initiating the below the present sea level (Lambeck and Chappell, 2001; Peltier, enhanced flow at lower, thermocline depths seen in the modern ITF 2002; Peltier and Fairbanks, 2006; Xu et al., 2008; Linsley et al., 2010), (Tozuka et al., 2007; Xu et al., 2008; Linsley et al., 2010), and conse- the South China and Indonesian Seas became connected, freshwater ex- quently the DOT shoaled (Figs. 8 and 9). port from the SCS to the Java Sea was initiated through the open connection of the Karimata Strait, and, together with the strong NW 4.3.4. Middle Holocene (8–6 ka) Monsoon and heavy rains, large amounts of freshwater flowed into During the 8–6 ka period, the DOT at the MD98-2172 and SHI9022 the Java Sea from the SCS. The freshwater plug at the southern tip of core sites deepened again (Fig. 7), with the abundance of foraminifer 124 X. Ding et al. / Marine Micropaleontology 101 (2013) 115–126 species indicative of high productivity conditions showing a prominent since the LGM. We found that the importance of deep versus shallow peak and the content of the oligotrophic species declining obviously at transport as well as the intensity of ITF activity did vary through time 13 the two core sites (Figs. 3 and 4). Meanwhile, the Δδ CPF−BF showed and this is related to sea-level variations and monsoon activity over the a high value in core MD98-2172 (Fig. 2), which indicated that produc- last ~23 ka. tivity rose once more. However, for the SHI9034 core situated in the Our findings suggest that surface ocean productivity was high dur- south Java Sea region and affected by the upwelling from the SE Mon- ing the LGM, and was related to the behaviors of a “perpetual” SE soon wind, the DOT deepened gradually in the early Holocene, reaching Monsoon resulting from the ITCZ being deflected northward. Warm, high values (maximum DOT) around ~8 ka (Fig. 8), with the abundance surface water ITF transport was dominated by the higher steric height of the high-productivity species G. bulloides rising in the period and difference between the western Pacific and the eastern Indian Oceans reaching a maximum around ~8 ka (Fig. 8). owing to a “perpetual” SE Monsoon and the narrowed main ITF out- The solar insolation curve shows that the Boreal summer values in- flow pathways caused by the lowered sea level. creased during the early Holocene (Fig. 8). In the eastern Indian Ocean, The warm, surface water ITF transport decreased after ~19 ka the enhanced marine productivity was directly related to strengthening MWP in the Timor Sea. This resulted from a sea level rise which wid- of coastal upwelling during periods of increased Boreal summer insola- ened the main ITF outflow pathways. The warm, low-salinity surface tion and associated SE Monsoon strength with a precessional cyclicity water ITF was diverted away from the western Pacific Ocean via the (Andruleit et al., 2008; Lückge et al., 2009). The Boreal hemisphere sum- widened pathways. In addition, the weakened “perpetual” SE Mon- mer insolation controls the SE Monsoon strength. However, in the soon likely diminished the steric height difference between the west- Timor Sea, the intensity of the SE Monsoon indicated by the high pro- ern Pacific and the eastern Indian Oceans. The decrease in water ductivity reached its highest values during 8−6 ka, lagging behind the pressure gradient between the two ocean basins led the surface variations of Boreal summer insolation. Consequently, in the early Holo- water ITF to weaken. At ~9.5 ka, when the postglacial sea level reached cene ~11–8 ka, the SE Monsoon was insufficiently strong to influence −30 m below its present level, large amounts of freshwater flowed the main ITF outflow in the Timor Sea, but it had obviously affected into the Java Sea from the SCS. The freshwater plug located at the the Java upwelling area. By analogy to today's seasonal variations, southern tip of the Makassar Strait blocked the warm surface flow, with an increasing SE Monsoon, strong westward currents along the thus initiating the thermocline ITF seen in the modern Indonesian Seas. southern coast of Java occurred, so that the sea-level steric height differ- During the period of ~8–6 ka, the intense SE Monsoon controlled ence between the southern and northern coasts of Java was increased, by the high Boreal hemisphere summer insolation increased the steric implying that the surface water ITF was boosted in the Java upwelling height difference between the western Pacific and the eastern Indian area, and the DOT at the SHI9034 core site deepened (Figs. 8 and 9). Oceans, and more saline Banda Sea water transferred into the south- During the 8–6 ka period, the strongest SE Monsoon appeared in ern Makassar Strait due to an intense SE Monsoon. At this time the the area, with the influence of the SE Monsoon not only restricted freshwater plug at the southern tip of the Makassar Strait that had to the Java upwelling area but also expanding into all the Indonesian blocked the warm surface flow disappeared. Thus, the surface water Seas. As a result the steric height difference between the western Pa- ITF intensified momentarily around ~8–6 ka. cific and eastern Indian Oceans must have increased. It is also inferred Over the last ~6 ka, strong NW Monsoon activity resulted in not only that the intense SE Monsoon transferred more saline Banda Sea water heavy rains but also large amounts of freshwater being transported into the southern Makassar Strait, and the freshwater plug at the from the SCS to the Java Sea. The freshwater plug at the southern tip southern tip of the Makassar Strait blocking the warm surface flow of the Makassar Strait blocked the warm surface flow, and consequently from the west disappeared. Thus, the surface water ITF intensified the thermocline ITF enhanced during the last ~6 ka. momentarily, resulting in a deepened DOT and high productivity In the Timor Sea, the changes in the vertical profile of the ITF were over all core sites in the study area (Figs. 8 and 9). influenced by the glacio-eustatic sea-level changes that have modified the geometry of the pathways within the Indonesian Seas, as well as 4.3.5. Late Holocene (last ~6 ka) by the monsoon activity, which was modulated by the changes in the After ~6 ka, the DOT shoaled gradually throughout the late Holocene insolation with a precessional cyclicity. in the study area (Fig. 7). The abundance of deep-water-dwelling species G. menardii and P. obliquiloculata in cores MD98-2172, SHI-9022 Acknowledgments and SHI-9034 rose clearly during this period. The rise may have been – partially caused by dissolution (Figs. 3 5). The abundances of the We are grateful to CNRS and CEA for AMS 14C and LSCE for the geo- high-productivity species N. dutertrei and N. pachyderma (dex.) were chemical analysis; Dr. Marie-France Loutre of Université Catholique de – low (Figs. 3 5). Louvain of Belgium for providing the insolation data and Dr. Liping The NW Monsoon gradually strengthened owing to the increase of Zhou of Peking University for the discussions. We thank Dr. Richard the Austral summer insolation during the late Holocene (Berger, Jordan and two anonymous reviewers for their critical remarks and con- 1978)(Fig. 8), which caused heavy rains, with large amounts of fresh- structive suggestions. This work was supported by the K. C. Wang Foun- fl water owing into the Java Sea from the SCS. The freshwater plug at dation of Centre National de la Recherche ScientifiquedeFrance,the the southern tip of the Makassar Strait blocked the warm surface National Natural Science Foundation of China (grant no. 40676034), fl ow, and thus the thermocline ITF developed again after ~6 ka and the fundamental research project of the State Oceanic Administra- (Fig. 9). tion of China “the paleoclimate research of the eastern Indian Ocean”. Of interest is that a detailed investigation of Termination II Cores were taken by IMAGES cruise IV of the R/V Marion Dufresne and – (135 128 ka) from core MD01-2378 within the same study area SHIVA cruise, a joint French-Indonesian marine geological research pro- (Xu et al., 2006) further indicated that the vertical structure of the gram. We express our gratitude to the crews of the survey ships for their fl Timor out ow during the MIS 5e sea-level highstand was dominated invaluable help. by thermocline flow, as occurs today (Gordon et al., 2003).

References 5. Conclusions Anand, P., Elderﬁeld, H., Conte, M.H., 2003. Calibration of Mg/Ca thermometry in plank- Information obtained from cores MD98-2172, SHI9022, and SHI9034 tonic foraminifera from a sediment trap time series. Paleoceanography 18 (2), 1050. http://dx.doi.org/10.1029/2002PA000846. collected in the Timor Sea and Java upwelling area reveal the evolution of Andreasen, D.J., Ravelo, A.C., 1997. Tropical Paciﬁc Ocean thermocline depth recon- surface hydrographic, and ITF and monsoon activity in the Timor Sea structions for the last glacial maximum. Paleoceanography 12 (3), 395–413. X. Ding et al. / Marine Micropaleontology 101 (2013) 115–126 125

Andruleit, H., Lückge, A., Wiedicke, M., 2008. Late Quaternary development of the Java Kawahata, H., Nishimura, A., Gagan, M.K., 2002. Seasonal change in foraminiferal pro- upwelling system (eastern Indian Ocean) as revealed by coccolithophores. Marine duction in the western equatorial Pacific warm pool: evidence from sediment Micropaleontology 69, 3–15. trap experiments. Deep-Sea Research II 49, 2783–2800. Barmawidjaja, B.M., Rohling, E.J., van der Kaars, W.A., Grazzini, C.V., Zachariasse, W.J., Lambeck, K., Chappell, J., 2001. Sea level change through the last glacial cycle. Science 1993. Glacial conditions of the northern Molucca Sea region (Indonesia). 292, 679–686. Palaeogeography, Palaeoclimatology, Palaeoecology 101, 147–167. Laskar, J., Robutel, P., Joutel, F., Gastineau, M., Correia, A.C.M., Levrard, B., 2004. A long Barrows, T.T., Juggins, S., 2005. Sea-surface temperatures around the Australian margin term numerical solution for the insolation quantities of the Earth. Astronomy and and Indian Ocean during the Last Glacial Maximum. Quaternary Science Reviews Astrophysics 428, 261–285. http://dx.doi.org/10.1051/0004-6361:20041335. 24 (7–9), 1017–1047. Levitus, 1998. World Ocean Atlas. http://www.ifremer.fr/sismer/references/climatologies/. Bé, A.W.H., 1977. An ecological, zoogeographical and taxonomic review of recent Linsley, B.K., Rosenthal, Y., Oppo, D.W., 2010. Holocene evolution of the Indonesian planktonic foraminifera. In: Ramsay, A.T.S. (Ed.), Oceanic Micropaleontology, 1. Ac- Throughflow and the western Pacific warm pool. Nature Geoscience 3, 578–583. ademic Press, London, pp. 1–100. Lückge, A., Mohtadi, M., Rühlemann, C., Scheeder, G., Vink, A., Reinhardt, L., Wiedicke, Berger, A., 1978. Long-term variation of daily insolation and Quaternary climatic M., 2009. Monsoon versus ocean circulation controls on paleoenvironmental con- changes. Journal of the Atmospheric Sciences 35, 2362–2367. ditions off southern Sumatra during the past 300,000 years. Paleoceanography Bray, N.A., Hautala, S., Chong, J., Pariwono, J., 1996. Large-scale sea level, thermocline, 24, PA1208. http://dx.doi.org/10.1029/2008PA001627. and wind variations in the Indonesian Throughflow region. Journal of Geophysical Magee, J.W., Miller, G.H., 1998. Lake Eyre palaeohydrology from 60 ka to the present: Research 101 (C5), 12239–12254. beach ridges and glacial maximum aridity. Palaeogeography, Palaeoclimatology, Brock, J.C., McClain, C.R., Anderson, D.M., Prell, W.L., Hay, W.W., 1992. Southwest mon- Palaeoecology 144, 307–329. soon circulation and environments of recent planktonic foraminifera in the north- Magee, J.W., Bowler, J.M., Miller, G.H., Williams, D.L.G., 1995. Stratigraphy, sedimentology, western Arabian Sea. Paleoceanography 7 (6), 799–813. chronology and palaeohydrology of Quaternary lacustrine deposits at Madigan Gulf, Cannariato, K.G., Ravelo, A.C., 1997. Pliocene–Pleistocene evolution of eastern tropical Lake Eyre, South Australia. Palaeogeography, Palaeoclimatology, Palaeoecology 113, Pacific surface water circulation and thermocline depth. Paleoceanography 12 3–42. (6), 805–820. Martinez, J.I., Taylor, L., De Deckker, P., Barrows, T.T., 1998. Planktonic foraminifera Clark, P.U., Dyke, A.S., Shakun, J.D., Carlson, A.E., Clark, J., Wohlfarth, B., Mitrovica, J.X., from the eastern Indian Ocean: distribution and ecology in relation to the Western Hostetler, S.W., McCabe, A.M., 2009. The last glacial maximum. Science 325, 710–714. Pacific Warm Pool (WPWP). Marine Micropaleontology 34, 121–151. Clark, P.U., Shakun, J.D., Baker, P.A., 2012. Global climate evolution during the last de- Martinez, J.I., De Deckker, P., Barrows, T.T., 1999. Palaeoceanography of the last glacial glaciation. Proceedings of the National Academy of Sciences of the United States maximum in the eastern Indian Ocean: planktonic foraminiferal evidence. of America 1–9. http://dx.doi.org/10.1073/pnas.1116619109. Palaeogeography, Palaeoclimatology, Palaeoecology 147, 73–99. Creswell, G., Frische, A., Peterson, J., Quadfasel, D., 1993. Circulation in the Timor Sea. Mix, A.C., Bard, E., Schneider, R., 2001. Environmental processes of the ice age: land, Journal of Geophysical Research 98 (C8), 14379–14389. oceans, glaciers (EPILOG). Quaternary Science Reviews 20, 627–657. Curry, W.B., Thunell, R.C., Honjo, S., 1983. Seasonal changes in the isotopic composition Molcard, R., Fieux, M., Syamsudin, F., 2001. The Throughflow within Ombai Strait. of planktonic foraminifera collected in Panama Basin sediment traps. Earth and Deep-Sea Research I 48, 1237–1253. Planetary Science Letters 64 (1), 33–43. Müller, A., Opdyke, B.N., 2000. Glacial–interglacial changes in nutrient utilization and De Deckker, P., Tapper, N.J., van der Kaars, S., 2002. The status of the Indo-Pacific Warm paleoproductivity in the Indonesian Throughflow sensitive Timor Trough, eastern- Pool and adjacent land at the Last Glacial Maximum. Global and Planetary Change most Indian Ocean. Paleoceanography 15, 85–94. 35, 25–35. Murray, S.P., Arief, D., 1988. Throughflow into the Indian Ocean through the Lombok Ding, X., Guichard, F., Bassinot, F., Labeyrie, L., Fang, N.Q., 2002. Evolution of heat trans- Strait, January 1985–January 1986. Nature 333, 444–447. port pathways in the Indonesian Archipelago during last deglaciation. Chinese Partin, J.W., Cobb, K.M., Adkins, J.F., Clark, B., Fernandez, D.P., 2007. Millennial-scale Science Bulletin 22, 1912–1917. trends in west Pacific warm pool hydrology since the Last Glacial Maximum. Ding, X., Bassinot, F., Guichard, F., Li, Q.Y., Fang, N.Q., Labeyrie, L., Xin, R.C., Adisaputra, Nature 449. http://dx.doi.org/10.1038/nature06164. M.K., Hardjawidjaksana, K., 2006. Distribution and ecology of planktonic foraminif- Peltier, W.R., 2002. On eustatic sea level history: last Glacial Maximum to Holocene. era from the seas around the Indonesian Archipelago. Marine Micropaleontology Quaternary Science Reviews 21, 377–396. 58, 114–134. Peltier, W.R., Fairbanks, R.G., 2006. Global glacial ice volume and Last Glacial Maximum Duplessy, J.C., Delibrias, G., Turon, J.L., Pujol, C., Duprat, J., 1981. Deglacial warming of the duration from an extended Barbados sea level record. Quaternary Science Reviews northeastern Atlantic Ocean: correlation with the paleoclimate evolution of European 25, 3322–3337. continent. Palaeogeography, Palaeoclimatology, Palaeoecology 35, 121–144. Pflaumann, U., Jian, Z., 1999. Modern distribution patterns of planktonic foraminifera in Fairbanks, R.G., Sverdlove, M., Free, R., Wiebe, P.H., Bé, A.W.H., 1982. Vertical distribu- the South China Sea and western Pacific: a new transfer technique to estimate re- tion and isotopic fractionation on living planktonic foraminifera from the Panama gional sea-surface temperatures. Marine Geology 156, 41–83. Basin. Nature 198, 841–844. Pflaumann, U., Duprat, J., Pujol, C., Labeyrie, L.D., 1996. SIMMAX: a modern analog tech- Fairbanks, R.G., Mortlock, R.A., Chiu, T.C., Cao, L., Kaplan, A., Guilderson, T.P., Fairbanks, nique to deduce Atlantic sea surface temperatures from planktonic foraminifera in T.W., Bloom, A.L., 2005. Marine radiocarbon calibration curve spanning 0 to deep-sea sediments. Paleoceanography 11 (1), 15–35. 50,000 years B.P. based on paired 230Th/234U/238U and 14C dates on pristine corals. Potemra, J.T., Hautala, S.L., Sprintall, J., 2003. Vertical structure of Indonesian Throughflow Quaternary Science Reviews 24, 1781–1796. in a large-scale model. Deep-Sea Research II 50, 2143–2161. Ganachaud, A., Wunsch, C., 2000. Improved estimates of global ocean circulation, heat Prell, W.L., Curry, W.B., 1981. Faunal and isotope indices of monsoonal upwelling: transport and mixing from hydrographic data. Nature 408, 453–457. Western Arabian Sea. Oceanologica Acta 4, 91–98. Gingele, F.X., De Deckker, P., Hillenbrand, C.D., 2001. Late Quaternary fluctuations of Qu, T., Song, Y.T., Yamagata, T., 2009. An introduction to the South China Sea throughflow: the Leeuwin Current and palaeoclimates on the adjacent landmasses: clay min- its dynamics, variability, and application for climate. Dynamics of Atmospheres and erals evidence. Australian Journal of Earth Sciences 48, 867–874. Oceans 47, 3–14. http://dx.doi.org/10.1016/j.dynatmoce.2008.05.001. Gordon, A.L., 1986. Interocean exchange of thermocline water. Journal of Geophysical Ravelo, A.C., Fairbanks, R.G., Philander, G., 1990. Reconstructing tropical Atlantic hy- Research 91, 5037–5046. drography using planktonic foraminifera and an ocean model. Paleoceanography Gordon, A.L., 2005. Oceanography of the Indonesian Seas and their throughflow. 5, 409–431. Oceanography 18 (4), 14–27. Schiller, A., Godfrey, J.S., McIntosh, P.C., Meyers, G., Wijffels, S.E., 1998. Seasonal near- Gordon, A.L., Fine, R.A., 1996. Pathways of water between the Pacific and Indian Oceans surface dynamics and thermodynamics of the Indian Ocean and Indonesian in the Indonesian Seas. Nature 379, 146–149. Throughflow in a global ocean General Circulation Model. Journal of Physical Gordon, A.L., Susanto, R.D., Ffield, A.L., 1999. Throughflow within Makassar Strait. Geo- Oceanography 28 (11), 2288–2312. physical Research Letters 26, 3325–3328. Siedler, G., Church, J., Gould, J., 2001. Ocean circulation & climate: observing and Gordon, A.L., Susanto, R.D., Vranes, K., 2003. Cool Indonesian Throughflow as a conse- modeling the Global Ocean, International Geophsics Series. Academic Press. A quence of restricted surface layer flow. Nature 425, 824–828. Harcourt Science and Technology Company 77, 303–314. Hesse, P.P., McTainsh, G.H., 2003. Australian dust deposits: modern processes and the Song, Q., Gordon, A.L., 2004. Significance of the vertical profile of the Indonesian Quaternary record. Quaternary Science Reviews 22, 2007–2035. Throughflow transport to the Indian Ocean. Geophysical Research Letters 31, Hesse, P.P., Magee, J.W., van der Kaars, S., 2004. Late Quaternary climates of the Australian L16307. http://dx.doi.org/10.1029/ 2004GL020360. arid zone. Quaternary International 118–119, 87–102. Spooner, M.I., Barrows, T.T., De Deckker, P., Paterne, M., 2005. Palaeoceanography of Hilbrecht, H., 1997. Morphologic gradation and ecology in Neogloboquadrina pachyderma the Banda Sea, and Late Pleistocene initiation of the Northwest Monsoon. Global and N. dutertrei (planktic foraminifera) from core top sediments. Marine Micropale- and Planetary Change 49 (1–2), 28–46. ontology 31, 31–44. Sprintall, J., Wijffels, S.E., Molcard, R., Jaya, I., 2009. Direct estimates of the Indonesian Hirst, A.C., Godfrey, J.S., 1993. The role of Indonesian Throughflow in a global ocean Throughflow entering the Indian Ocean: 2004–2006. Journal of Geophysical Re- GCM. Marine Micropaleontology 23, 1057–1086. search 114, C07001. http://dx.doi.org/10.1029/2008JC005257. Holbourn, A., Kuhnt, W., Kawamura, H., Jian, Z., Grootes, P., Erlenkeuser, H., Xu, J., 2005. Takahashi, K., Okada, H., 2000. The paleoceanography for the last 30,000 years in the Orbitally paced paleoproductivity variations in the Timor Sea and Indonesian Southeastern Indian Ocean by means of calcareous nannofossils. Marine Micropa- Throughflow variability during the last 460 kyr. Paleoceanography 20, PA3002. leontology 40, 83–103. http://dx.doi.org/10.1029/2004PA001094. Thiede, J., Jünge, B., 1992. Faunal and floral indicators of ocean coastal upwelling (NW Huang, B.Q., Jian, Z.M., Cheng, X.R., Wang, P.X., 2002. Foraminiferal responses to up- African and Peruvian continental margins). In: Summerhayes, C.P., Prell, W.L., welling variations in the South China Sea over the last 220000 years. Marine Mi- Emeis, K.C. (Eds.), Upwelling systems: evolution since the Miocene: Geological cropaleontology 47, 1–15. Society London Special Publication, 64, pp. 47–76. Jian, Z., Huang, B., Kuhnt, W., Lin, H., 2001. Late Quaternary upwelling intensity and east Thunell, R.C., 1976. Optimum indices of calcium carbonate dissolution in deep-sea sed- Asian monsoon forcing in the South China Sea. Quaternary Research 55, 363–370. iments. Geology 4, 525–527. 126 X. Ding et al. / Marine Micropaleontology 101 (2013) 115–126

Thunell, R.C., Reynolds, L.A., 1984. Sedimentation of Planktonic foraminifera: seasonal Wang, X., van der Kaars, W.A., Kershaw, P., Bird, M., Jansen, F., 1999. A record of fire, veg- changes in species flux in the Panama Basin. Micropaleontology 30, 243–262. etation and climate through the last three glacial cycles from the Lombok Ridge core Tozuka, T., Qu, T., Yamagata, T., 2007. Dramatic impact of the South China Sea on the Indo- G6-4, eastern Indian Ocean, Indonesia. Palaeogeography, Palaeoclimatology, Palaeo- nesian throughflow. Geophysical Research Letters 34, L12612. http://dx.doi.org/10. ecology 147, 241–256. 1029/ 2007GL030420. Wyrwoll, K.H., Miller, G.H., 2001. Initiation of the Australian summer monsoon van der Kaars, W.A., Dam, M.A.C., 1995. A 135,000-year old record of vegetational and 14,000 years ago. Quaternary International 83–85, 119–128. climatic change from the Bandung area, West-Java, Indonesia. Palaeogeography, Xu, J., Kuhnt, W., Holbourn, A., Andersen, N., Bartoli, G., 2006. Changes in the vertical pro- Palaeoclimatology, Palaeoecology 117 (1–2), 55–71. file of the Indonesian Throughflow during Termination II: evidence from the Timor van der Kaars, S., De Deckker, P., 2002. A Late Quaternary pollen record from deep-sea core Sea. Paleoceanography 21, PA4202. http://dx.doi.org/10.1029/2006PA001278. FR10/95, GC17 offshore Cape Range Peninsula, northwestern Western Australia. Re- Xu, J., Holbourn, A., Kuhnt, W., Jian, Z.M., Kawamura, H., 2008. Changes in the thermo- view of Palaeobotany and Palynology 120 (1–2), 17–39. cline structure of the Indonesian outflow during Terminations I and II. Earth and van der Kaars, S., Wang, X., Kershaw, P., Guichard, F., Setiabudi, D.A., 2000. A late Quater- Planetary Science Letters 273, 152–162. nary palaeoecological record from the Banda Sea, Indonesia: patterns of vegetation, Žarić, S., Donnar, B., Fischer, G., Mulitza, S., Wefer, G., 2005. Sensitivity of planktic fora- climate and biomass burning in Indonesia and northern Australia. Palaeogeography, minifera to sea surface temperature and export production as derived from sedi- Palaeoclimatology, Palaeoecology 155, 135–153. ment trap data. Marine Micropaleontology 55, 75–105. Veeh, H.H., McCorkle, D.C., Heggie, D.T., 2000. Glacial/interglacial variations of sedimentation on the West Australian continental margin: constraints from excess 230Th. Marine Geology 166, 11–30.