A low-latitude Upper Pleistocene oxygen isotope stack

Below oxygen isotope stage 16, the orbitally derived time-scale developed by Shackleton et al. (1990) from ODP site 677 in the equatorial Pacific differs significantly from previous ones (e.g. Kominz and Pisias, 1979 doi:10.1126/science.204.4389.171; Morley and Hays, 1981 doi:10.1016/0012-821X(81)90034-0, Imbrie et al. 1984), yielding estimated ages for the last Earth magnetic reversals that are 5-7% older than the K/Ar values (Mankinen and Dalrymple, 1979 doi:10.1029/JB084iB02p00615; Berggren et al., 1985; Harland and Armstrong, 1989) but are in good agreement with recent Ar/Ar dating (Baksi et al., 1991; 1992 doi:10.1126/science.256.5055.356; Spell and McDougall, 1992 doi:10.1029/92GL01125). These results suggest that in the lower Brunhes and upper Matuyama chronozones most deep-sea climatic records retrieved so far apparently missed or misinterpreted several oscillations predicted by the astronomical theory of climate. To test this hypothesis, we studied a high-resolution oxygen isotope record from giant piston core MD900963 (Maldives area, tropical Indian Ocean) in which precession-related oscillations in delta18O are particularly well expressed, owing to the superimposition of a local salinity signal on the global ice volume signal (Rostek et al., 1993 doi:10.1038/364319a0). Three additional precession-related cycles are observed in oxygen isotope stages 17 and 18 of core MD900963, compared to the SPECMAP composite curves (Imbrie et al., 1984; Prell et al., 1986 doi:10.1029/PA001i002p00137), and stage 21 clearly presents three precession oscillations, as predicted by Shackleton et al. (1990). The precession peaks found in the delta18O record from core MD900963 are in excellent agreement with climatic oscillations predicted by the astronomical theory of climate. Our delta18O record therefore permits the development of an accurate astronomical time-scale. Based on our age model, the Brunhes-Matuyama reversal is dated at 775 +/- 10 ka, in good agreement with the age estimate of 780 ka obtained by Shackleton et al. (1990) and recent radiochronological Ar/Ar datings on lavas (Baksi et al., 1991; 1992; Spell and McDougall, 1992). We developed a new low-latitude, Upper Pleistocene delta18O reference record by stacking and tuning the delta18O records from core MD900963 and site 677 to orbital forcing functions.

Persistent monsoonal forcing of Mediterranean Outflow dynamics during the late Pleistocene

Reconstruction of Mediterranean Outflow variability.

Osmium isotope record of Late Pleistocene sediments

We report high temporal resolution osmium isotopes records of bulk sediment and sediment leachates from DSDP Site 480 (Gulf of California) over the last 30 ka; from ODP Site 849 (Eastern equatorial Pacific) from the last 200 ka and from ODP Site 1002C (Cariaco Basin) across the 9-17 ka time interval in order to critically evaluate claims of a global 10% shift in the 187Os/188Os of seawater from glacial to interglacial intervals. We use organic-rich continental margin sites and carbonate-rich pelagic sites to isolate the temporal variations of the osmium seawater isotopic composition. Our results reveal that variations in 187Os/188Os fail to correlate with global changes in temperature across glacials/interglacials cycles as previously claimed. Instead, these results indicate differences of a few percent in the measured 187Os/188Os between each oceanic basin. We argue that these differences strongly suggest that seawater is not well homogenized with respect to its Os isotope composition.

Plio-Pleistocene age-depth points and occurrence of hiatuses in ODP sites of the Southern Ocean

Changes in Atlantic deep water circulation were reconstructed by comparing the benthic foraminiferal delta13C record at ODP Site 1090 in the South Atlantic with similar records from the North Atlantic (Sites 982, 607, 925, 929) and deep Pacific (Site 849) oceans. Important deep water circulation changes occurred in the early Pleistocene at 1.55 Myr and during the Mid-Pleistocene Transition at 0.9 Myr. At 1.55 Myr, glacial delta13C values in the Southern Ocean became significantly lower than those in the deep Pacific, establishing a pattern that persisted throughout the late Pleistocene. We propose that the lowering of delta13C values of Southern Component Water (SCW) at this time resulted from expansion of sea ice and reduced ventilation of deep water during glacial periods after marine isotope stage 52. Accompanying this change in Southern Ocean deep water circulation was enhanced interhemispheric coupling between the North and South Atlantic after 1.55 Myr. At ~0.9 Myr, the magnitude of glacial-to-interglacial variabilityin delta13C increased and shifted to a longer frequency (100 kyr) along with oceanic delta18O (ice volume). Calculation of percent Northern Component Water (NCW) using Site 1090 as the SCW end member yielded 20-30% less reduction of NCW during glacial periods of the late Pleistocene. Also, a trend toward reduced glacial suppression of NCW during the past 400 kyr is not evident. The apparent decoupling of ice volume and deep water circulation reported previously maybe an artifact of using a Pacific, rather than a Southern Ocean, carbon isotopic record to calculate past mixing ratios of NCW and SCW.

Feldspar Pb-isotope ratios for earliest Pleistocene sediments of the sub-polar North Atlantic Ocean

Relatively little is known in detail about the locations of the early Pleistocene ice-sheets responsible for ice-rafted debris (IRD) inputs to the sub-polar North Atlantic Ocean during intensification of northern hemisphere glaciation (iNHG). To shed new light on this problem, we present the first combined in-depth analysis of IRD flux and geochemical provenance of individual sand-sized IRD deposited in the sub-polar North Atlantic Ocean during the earliest large amplitude Pleistocene glacial, marine isotope stage (MIS) 100 (~2.52 Ma), arguably the key glacial during iNHG. IRD provenance is assessed using laser ablation lead (Pb) isotope analyses of single feldspar grains. We find that the Pb-isotope composition (206Pb/204Pb, 207Pb/204Pb and 208Pb/204Pb) of individual ice-rafted (>150 µm) feldspars deposited at DSDP Site 611A, ODP Site 981 and IODP Site U1308 during MIS 100 records a shift from predominantly Archaean-aged circum-North Atlantic Ocean continental sources during early glacial ice-rafting events to dominantly Palaeozoic and Proterozoic-aged sources during full glacial conditions. The distribution of feldspars in Pb-Pb space for full glacial MIS 100 more closely resembles that documented for feldspars deposited at the centre of the last glacial IRD belt (at IODP/DSDP Site U1308/609) during ambient (non-Heinrich-event) ice-rafting episodes of MIS 2 (~23.8 ka) than that documented for MIS 5d (~106 ka). Comparison of our early Pleistocene and last glacial cycle datasets suggests that MIS 100 was characterised by abundant iceberg calving from large ice-sheets on multiple continents in the high northern latitudes (not just on Greenland).

CaCO3, foraminifer fragmentation, benthic carbon isotopes, coarse lithic counts and radiogenic isotopes (Sr, Nd, Pb) of the Pliocene and earliest Pleistocene terrigenous component (3.3-2.4 Ma) of IODP Site 306-U1313

We present Plio-Pleistocene records of sediment color, %CaCO3, foraminifer fragmentation, benthic carbon isotopes (d13C) and radiogenic isotopes (Sr, Nd, Pb) of the terrigenous component from IODP Site U1313, a reoccupation of benchmark subtropical North Atlantic Ocean DSDP Site 607. We show that (inter)glacial cycles in sediment color and %CaCO3 pre-date major northern hemisphere glaciation and are unambiguously and consistently correlated to benthic oxygen isotopes back to 3.3 million years ago (Ma) and intermittently so probably back to the Miocene/Pliocene boundary. We show these lithological cycles to be driven by enhanced glacial fluxes of terrigenous material (eolian dust), not carbonate dissolution (the classic interpretation). Our radiogenic isotope data indicate a North American source for this dust (~3.3-2.4 Ma) in keeping with the interpreted source of terrestrial plant wax-derived biomarkers deposited at Site U1313. Yet our data indicate a mid latitude provenance regardless of (inter)glacial state, a finding that is inconsistent with the biomarker-inferred importance of glaciogenic mechanisms of dust production and transport. Moreover, we find that the relation between the biomarker and lithogenic components of dust accumulation is distinctly non-linear. Both records show a jump in glacial rates of accumulation from Marine Isotope Stage, MIS, G6 (2.72 Ma) onwards but the amplitude of this signal is about 3-8 times greater for biomarkers than for dust and particularly extreme during MIS 100 (2.52 Ma). We conclude that North America shifted abruptly to a distinctly more arid glacial regime from MIS G6, but major shifts in glacial North American vegetation biomes and regional wind fields (exacerbated by the growth of a large Laurentide Ice Sheet during MIS 100) likely explain amplification of this signal in the biomarker records. Our findings are consistent with wetter-than-modern reconstructions of North American continental climate under the warm high CO2 conditions of the Early Pliocene but contrast with most model predictions for the response of the hydrological cycle to anthropogenic warming over the coming 50 years (poleward expansion of the subtropical dry zones).

Pleistocene sedimentology of ODP Hole 133-820A from the northeastern Australian margin

A principal objective at Site 820, situated on the outer shelf, upper slope of the northeastern Australian continental margin, was to test the relationships between changes in Pleistocene sea level and sedimentary packages produced on a mixed carbonate-siliciclastic continental margin. To this end, we have examined the downcore distribution of grain size, magnetic susceptibility, and calcium-carbonate content throughout Hole 820A and, in particular, the top 35 meters below the seafloor (mbsf). These data are compared with variations in the oxygen-isotope signal defined for the same hole and are interpreted as indicating sea-level oscillations. The distribution of sand, mud, calcium carbonate of the mud fraction and total sample, and magnetic susceptibility during the last 20,000 yr defines the position of a sea-level regression (41,000-18,000 yr B.P.), a lowstand, early (18,000-9,400 yr B.P.) and late transgressions (9400-900 yr B.P.), and a highstand (4900 yr to the present). The regression is seen first in a high-carbonate content peak. Calcium carbonate constituents mainly comprise skeletal carbonate grains, with abundant planktonic and benthic foraminifers, and lime muds. The lowstand is characterized by a maximum abundance of the sand fraction, which contains dominantly skeletal carbonate grains and a minor abundance of lithoclasts. Sand-sized terrigenous sediments are proposed to have bypassed the continental shelf during a lowstand of sea level. Sedimentation rates throughout the regression and lowstand are low (3.0 cm/k.y.). The early transgression, marked by highest values in magnetic susceptibility, displays a rapid increase in sedimentation rate that coincided with an increase in terrigenous mud. Highest sedimentation rates of 82.3 cm/k.y. occurred during the late transgression, with increasing percentages of lime-mud. A decrease in noncarbonate constituents in the mud fraction during the late transgression and highstand of sea level is thought to be the result of restricted inner-shelf sedimentation of terrigenous sediments. The same relationship is also seen in the major sea-level oscillation, which is interpreted as isotope stage 6.

Plio-Pleistocene Mg/Ca and TEX86 sea surface temperature records of ODP Sites 184-1143 and 165-999,

The western warm pools of the Atlantic and Pacific oceans are a critical source of heat and moisture for the tropical climate system. Over the past five million years, global mean temperatures have cooled by 3-4 °C. Yet, current reconstructions of sea surface temperatures indicate that temperature in the warm pools has remained stable during this time. This stability has been used to suggest that tropical sea-surface temperatures are controlled by some sort of thermostat-like regulation. Here we reconstruct sea surface temperatures in the South China Sea, Caribbean Sea and western equatorial Pacific Ocean for the past five million years, using a combination of the Mg/Ca, TEXH86-and Uk'37 surface temperature proxies. Our data indicate that during the period of Pliocene warmth from about 5 to 2.6 million years ago, the western Pacific and western Atlantic warm pools were about 2 °C warmer than today. We suggest that the apparent lack of warming seen in the previous reconstructions was an artefact of low seawater Mg/Ca ratios in the Pliocene oceans. Taking this bias into account, our data indicate that tropical sea surface temperatures did change in conjunction with global mean temperatures. We therefore conclude that the temperature of the warm pools of the equatorial oceans during the Pliocene was not limited by a thermostat-like mechanism.