(Table 1) Age determination and isotopic signature of last glacial Patagonian caly/silt

Ice cores provide a record of changes in dust flux to Antarctica, which is thought to reflect changes in atmospheric circulation and environmental conditions in dust source areas (Forster et al., 2007; Diekmann et al. 2000, doi:10.1016/S0031-0182(00)00138-3; Winckler et al., 2008, doi:10.1126/science.1150595; Reader et al., 1999, doi:10.1029/1999JD900033; Mahowald et al., 1999, doi:10.1029/1999JD900084; Petit et al., 1999, doi:10.1038/20859; 1990, doi:10.1038/343056a0 Delmonte et al., 2009, doi:10.1029/2008GL033382; Lambert et al., 2008, doi:10.1038/nature06763). Isotopic tracers suggest that South America is the dominant source of the dust (Grousset et al., 1992, doi:10.1016/0012-821X(92)90177-W; Basile et al., 1997, doi:10.1016/S0012-821X(96)00255-5; Gaiero et al., 2007, doi:10.1016/j.chemgeo.2006.11.003), but it is unclear what led to the variable deposition of dust at concentrations 20-50 times higher than present in glacial-aged ice (Petit et al., 1990, doi:10.1038/343056a0; Lambert et al., 2008, doi:10.1038/nature06763). Here we characterize the age and composition of Patagonian glacial outwash sediments, to assess the relationship between the Antarctic dust record from Dome C (refs Lambert et al., 2008, doi:10.1038/nature06763; Wolff et al., 2006, doi:10.1038/nature04614) and Patagonian glacial fluctuations (Sugden et al., 2005; McCulloch et al., 2005, doi:10.1111/j.0435-3676.2005.00260.x; Kaplan et al., 2008, doi:10.1016/j.quascirev.2007.09.013) for the past 80,000 years. We show that dust peaks in Antarctica coincide with periods in Patagonia when rivers of glacial meltwater deposited sediment directly onto easily mobilized outwash plains. No dust peaks were noted when the glaciers instead terminated directly into pro-glacial lakes. We thus propose that the variable sediment supply resulting from Patagonian glacial fluctuations may have acted as an on/off switch for Antarctic dust deposition. At the last glacial termination, Patagonian glaciers quickly retreated into lakes, which may help explain why the deglacial decline in Antarctic dust concentrations preceded the main phase of warming, sea-level rise and reduction in Southern Hemisphere sea-ice extent (Wolff et al., 2006, doi:10.1038/nature04614).

Plant-wax, alkenone and major element analyses of ODP Site 659 for the Last Glacial cycle and associated coretop data

We present a hydrologic reconstruction of the Sahara-Sahel transition, covering the complete last glacial cycle (130 ka), based on a combination of plant-wax-specific hydrogen (dD) and carbon isotopes (d13C). The dD and d13C signatures of long-chain n-alkanes from ODP Site 659 off NW Africa reveal a significant anti-correlation. Complementary to published pollen data, we infer that this plant-wax signal reflects sensitive responses of the vegetation cover to precipitation changes in the Sahel region, as well as varying contributions from biomes north of the Sahara (C3 domain) by North-East Trade Winds (NETW). During arid phases, especially the northern parts of the Sahel likely experienced crucial water stress, which resulted in a pronounced contraction of the vegetation cover, thus reducing the amount of C4 plant waxes from the region. The increase in NETW strength during dry periods further promoted a more pronounced C3-plant-wax signal derived from the North African C3 plant domain. During humid periods, the C4-dominated Sahelian environments spread northward into the Saharan realm, in association with lower NETW inputs of C3 plant waxes. Arid-humid cycles deduced from plant-wax dD are in accordance with concomitant changes in weathering intensity reflected in varying major element distributions. Environmental shifts are generally linked to periods with large fluctuations in Northern Hemisphere summer insolation. During Marine Isotope Stages 2 and 3, when insolation variability was low, coupling of the hydrologic regime to alkenone-based estimates of NE Atlantic sea-surface temperatures becomes apparent.

Changes in bottom water corrosiveness and carbonate ion concentrations in the sub-Antarctic Atlantic during the last glacial period

The glacial climate system transitioned rapidly between cold (stadial) and warm (interstadial) conditions in the Northern Hemisphere. This variability, referred to as Dansgaard-Oeschger variability, is widely believed to arise from perturbations of the Atlantic Meridional Overturning Circulation. Evidence for such changes during the longer Heinrich stadials has been identified, but direct evidence for overturning circulation changes during Dansgaard-Oeschger events has proven elusive. Here we reconstruct bottom water [CO3]2- variability from B/Ca ratios of benthic foraminifera and indicators of sedimentary dissolution, and use these reconstructions to infer the flow of northern-sourced deep water to the deep central sub-Antarctic Atlantic Ocean. We find that nearly every Dansgaard-Oeschger interstadial is accompanied by a rapid incursion of North Atlantic Deep Water into the deep South Atlantic. Based on these results and transient climate model simulations, we conclude that North Atlantic stadial-interstadial climate variability was associated with significant Atlantic overturning circulation changes that were rapidly transmitted across the Atlantic. However, by demonstrating the persistent role of Atlantic overturning circulation changes in past abrupt climate variability, our reconstructions of carbonate chemistry further indicate that the carbon cycle response to abrupt climate change was not a simple function of North Atlantic overturning.

Mean sortable silt values from Holocene and glacial Scotia Sea sediments

The Antarctic Circumpolar Current is key to the mixing and ventilation of the world's oceans. This current flows from west to east between about 45° and 70° S connecting the Atlantic, Pacific and Indian oceans, and is driven by westerly winds and buoyancy forcing. High levels of productivity in the current regulate atmospheric CO2 concentrations. Reconstructions of the current during the last glacial period suggest that flow speeds were faster or similar to present, and it is uncertain whether the strength and position of the westerly winds changed. Here we reconstruct Antarctic Circumpolar Current bottom speeds through the constricting Drake Passage and Scotia Sea during the Last Glacial Maximum and Holocene based on the mean grain size of sortable silt from a suite of sediment cores. We find essentially no change in bottom flow speeds through the region, and, given that the momentum imparted by winds, and modulated by sea-ice cover, is balanced by the interaction of these flows with the seabed, this argues against substantial changes in wind stress. However, glacial flow speeds in the sea-ice zone south of 56° S were significantly slower than present, whereas flow in the north was faster, but not significantly so. We suggest that slower flow over the rough topography south of 56° S may have reduced diapycnal mixing in this region during the last glacial period, possibly reducing the diapycnal contribution to the Southern Ocean overturning circulation.

Mineral dust deposition in interglacial and glacial climate conditions: model results

The global aerosol/climate model ECHAM5-HAM is used in order to investigate the dust cycle for four interglacial and one glacial climate conditions. The 20-year time-slices are the pre-industrial control (CTRL), mid-Holocene (6000 years BP), last glacial inception (115000 years BP), Eemian (126000 years BP) and Last Glacial Maximum (LGM) (21000 years BP) time intervals. The study is focused on the Antarctic region. The model is able to reproduce the magnitude order of dust deposition globally for the pre-industial and LGM climates. Correlation coefficient of the natural logarithm of the observed and modeled values is 0.78 for the CTRL and 0.81 for the LGM. For the pre-industrial simulation the model overestimates observed values in Antarctica by a factor of about 2-3 due to overestimation of the Australian dust source and too high wet deposition in the Antarctica interior. In the LGM, the model underestimates dust deposition in eastern Antarctica by a factor of about 4-5 due to underestimation of the South American dust source. More records are needed to validate dust deposition for the past interglacial time-slices. The modeled results show that dust deposition in Antarctica in the past interglacial time-slices is higher than in the CTRL simulation. The largest increase of dust deposition in Antarctica is simulated for the LGM, showing about 10-fold increase compared to CTRL.

Glacial-interglacial record of ODP Site 181-1119 in the Canterbury Bright, Southwest Pacific Ocean

Ocean Drilling Program (ODP) Site 1119 is located at water depth 395 m near the subtropical front (STF; here represented by the Southland Front), just downslope from the shelf edge of eastern South Island, New Zealand. The upper 86.19 metres composite depth (mcd) of Site 1119 sediment was deposited at an average sedimentation rate of 34 cm/kyr during Marine Isotope Stages (MIS) 1-8 (0-252 ka), and is underlain across a ~25 kyr intra-MIS 8 unconformity by MIS 8.5-11 (277-367 ka) and older sediment deposited at ~14 cm/kyr. A time scale is assigned to Site 1119 using radiocarbon dates for the period back to ~39 ka, and, prior to then, by matching its climatic record with that of the Vostok ice core, which it closely resembles. Four palaeoceanographic proxy measures for surface water masses vary together with the sandy-muddy, glacial-interglacial (G/I) cyclicity at the site. Interglacial intervals are characterised by heavy delta13C, high colour reflectance (a proxy for carbonate content), low Q-ray (a proxy for clay content) and light delta18O; conversely, glacial intervals exhibit light delta13C, low reflectance, high Q-ray and heavy delta18O signatures. Early interglacial intervals are represented by silty clays with 10-105-cm-thick beds of sharp-based (Chondrites-burrowed), shelly, graded, fine sand. The sands are rich in foraminifera, and were deposited distant from the shoreline under the influence of longitudinal flow in relatively deep water. Glacial intervals comprise mostly micaceous silty clay, though with some thin (2-10 cm thick) sands present also at peak cold periods, and contain the cold-water scallop Zygochlamys delicatula. Interglacial sandy intervals are characterised by relatively low sedimentation rates of 5-32 cm/kyr; cold climate intervals MIS 10, 6 and 2 have successively higher sedimentation rates of 45, 69 and 140 cm/kyr. Counter-intuitively,and forced by the bathymetric control of a laterally-moving shoreline during G/I and I/G transitions, the 1119 core records a southeasterly (seaward) movement of the STF during early glacial periods, accompanied by the incursion of subtropical water (STW) above the site, and northwesterly (landward) movement during late glacial and interglacial times, resulting in a dominant influence then of subantarctic surface water (SAW). The history of passage of these different water masses at the site is clearly delineated by their characteristic delta13C values. The intervals of thin, graded sands-muds which occur within MIS 2-3, 6, 7.4 and 10 indicate the onset at times of peak cold of intermittent bottom currents caused by strengthened and expanded frontal flows along the STF, which at such times lay near Site 1119 in close proximity to seaward-encroaching subantarctic waters within the Bounty gyre. In common with other nearby Southern Hemisphere records, the cold period which represents the last glacial maximum lasted between ~23-18 ka at Site 1119, during which time the STF and Subantarctic Front (SAF) probably merged into a single intense frontal zone around the head of the adjacent Bounty Trough.

Measurements of stable carbon isotope ratios of CO2 over the last 24000 years and CO2 concentration measurements on Antarctic ice cores using three different methods

The stable carbon isotope ratio of atmospheric CO2 (d13Catm) is a key parameter in deciphering past carbon cycle changes. Here we present d13Catm data for the past 24,000 years derived from three independent records from two Antarctic ice cores. We conclude that a pronounced 0.3 per mil decrease in d13Catm during the early deglaciation can be best explained by upwelling of old, carbon-enriched waters in the Southern Ocean. Later in the deglaciation, regrowth of the terrestrial biosphere, changes in sea surface temperature, and ocean circulation governed the d13Catm evolution. During the Last Glacial Maximum, d13Catm and atmospheric CO2 concentration were essentially constant, which suggests that the carbon cycle was in dynamic equilibrium and that the net transfer of carbon to the deep ocean had occurred before then.

Stable isotope ratios of foraminifera from last glacial-interglacial sediments of the southern high latitudes

A set of numerical equations is developed to estimate past sea surface temperatures (SST) from fossil Antarctic diatoms. These equations take into account both the biogeographic distribution and experimentally derived silica dissolution. The data represent a revision and expansion of a floral data base used previously and includes samples resulting from progressive opal dissolution experiments. Factor analysis of 166 samples (124 Holocene core top and 42 artificial samples) resolved four factors. Three of these factors depend on the water mass distribution (one Subantarctic and two Antarctic assemblages); factor 4 corresponds to a 'dissolution assemblage'. Inclusion of this factor in the data analysis minimizes the effect of opal dissolution on the assemblages and gives accurate estimates of SST over a wide range of biosiliceous dissolution. A transfer function (DTF 166/34/4) is derived from the distribution of these factors versus summer SST. Its standard error is +/- 1°C in the -1 to +10 °C summer temperature range. This transfer function is used to estimate SST changes in two southern ocean cores (43°S and 55°S) which cover the last climatic cycle. The time scale is derived from the changes in foraminiferal oxygen and carbon isotopic ratios. The reconstructed SST records present strong analogies with the air temperature record over Antarctica at the Vostok site, derived from changes in the isotopic ratio of the ice. This similarity may be used to compare the oceanic isotope stratigraphy and the Vostok time scale derived from ice flow model. The oceanic time scale, if taken at face value, would indicate that large changes in ice accumulation rates occurred between warm and cold periods.

Planktic and benthic 14C reservoir ages, calibrated by a suite of 14C plateaus in the glacial-to-deglacial Suigetsu atmospheric 14C record

We here present a compilation of planktic and benthic 14C reservoir ages for the Last Glacial Maximum (LGM) and early deglacial from 11 key sites of global ocean circulation in the Atlantic and Indo-Pacific Ocean. The ages were obtained by 14C plateau tuning, a robust technique to derive both an absolute chronology for marine sediment records and a high-resolution record of changing reservoir/ventilation ages (Delta14C values) for surface and deep waters by comparing the suite of planktic 14C plateaus of a sediment record with that of the atmospheric 14C record (Sarnthein et al., 2007, doi:10.1029/173GM13). Results published thus far used as atmospheric 14C reference U/Th-dated corals, the Cariaco planktic record, and speleothems (Fairbanks et al., 2005, doi:10.1016/j.quascirev.2005.04.007; Hughen et al., 2006, doi:10.1016/j.quascirev.2006.03.014; Beck et al., 2001, doi:10.1023/A:1008175728826). We have now used the varve-counted atmospheric 14C record of Lake Suigetsu terrestrial macrofossils (Ramsey et al., 2012, doi:10.1126/science.1226660) to recalibrate the boundary ages and reservoir ages of the seven published records directly to an atmospheric 14C record. In addition, the results for four new cores and further planktic results for four published records are given. Main conclusions from the new compilation are: (1) The Suigetsu atmospheric 14C record on its varve counted time scale reflects all 14C plateaus, their internal structures and relative length previously identified, but implies a rise in the average 14C plateau age by 200-700 14C yr during LGM and early deglacial times. (2) Based on different 14C ages of coeval atmospheric and planktic 14C plateaus, marine surface water Delta14C may have temporarily dropped to an equivalent of ~0 yr in low-latitude lagoon waters, but reached >2500 14C yr both in stratified subpolar waters and in upwelled waters such as in the South China Sea. These values differ significantly from a widely assumed constant global planktic Delta14C value of 400 yr. (3) Suites of deglacial planktic Delta14C values are closely reproducible in 14C records measured at neighboring core sites. (4) Apparent deep-water 14C ventilation ages (equivalents of benthic Delta14C), deduced from the sum of planktic Delta14C and coeval benthic-planktic 14C differences, vary from 500 up to >5000 yr in LGM and deglacial ocean basins.

Glacial-interglacial sediment record from the North Atlantic

Studies from the subtropical western and eastern Atlantic Ocean, using the 231Pa/230Th ratio as a kinematic proxy for deep water circulation, provided compelling evidence for a strong link between climate and the rate of meridional overturning circulation (MOC) over the last deglaciation. In this study, we present a compilation of existing and new sedimentary 231Pa/230Th records from North Atlantic cores between 1710 and 4550 m water depth. Comparing sedimentary 231Pa/230Th from different depths provides new insights into the evolution of the geometry and rate of deep water formation in the North Atlantic during the last 20,000 years. The 231Pa/230Th ratio measured in upper Holocene sediments indicates slow water renewal above ?2500 m and rapid flushing below, consistent with our understanding of modern circulation. In contrast, during the Last Glacial Maximum (LGM), Glacial North Atlantic Intermediate Water (GNAIW) drove a rapid overturning circulation to a depth of at least ?3000 m depth. Below ~4000 m, water renewal was much slower than today. At the onset of Heinrich event 1, transport by the overturning circulation declined at all depths. GNAIW shoaled above 3000 m and significantly weakened but did not totally shut down. During the Bølling-Allerød (BA) that followed, water renewal rates further decreased above 2000 m but increased below. Our results suggest for the first time that ocean circulation during that period was quite distinct from the modern circulation mode, with a comparatively higher renewal rate above 3000 m and a lower renewal rate below in a pattern similar to the LGM but less accentuated. MOC during the Younger Dryas appears very similar to BA down to 2000 m and slightly slower below.