Preliminary results and documentation of sediment cores CRP2 and CRP-2A from the Ross Sea off Cape Roberts, Antarctica

The site for CRP-2, 14 km east of Cape Roberts (77.006°S; 163.719°E), was selected to overlap the early Miocene strata cored in nearby CRP-1, and to sample deeper into the east-dipping strata near the western margin ofe he Victoria Land Basin to investigate Palaeogene climatic and tectonic history. CRP-2 was cored from 5 to 57 mbsf (metres below the sea floor) (core recovery 91 %), with a deviation resulting in CRP-2A being cored at the same site. CRP-2A reached down to 624mbsf (recovery 95%), and to strata with an age of c. 33-35 Ma. Drilling took place from 16 October to 25 November 1998, on 2.0-2.2 m of sea ice and through 178 m of water. Core fractures and other physical properties, such as sonic velocity, density and magnetic susceptibility, were measured throughout the core. Down-hole logs for these and other properties were run from 63 to 167 mbsf and subsequently from 200 to 623 mbsf, although density and velocity data could be obtained only to 440 mbsf because of hole collapse. Sonic velocity averages c. 2.0 km S-1 for the upper part of the hole, but there is an sharp increase to c. 3.0 km s-1 and also a slight angular unconformity, at 306 mbsf, corresponding most likely to the early/late Oligocene boundary (c. 28-30 Ma). Velocity then increases irregularly to around 3.6 km s-1 at the bottom of the hole, which is estimated to lie 120 m above the V4/V5 boundary. The higher velocities below 306 mbsf probably reflect more extensive carbonate and common pyrite cementation, in patches, nodules, bedding-parallel masses and as vein infills. Dip of the strata also increases down-hole from 3° in the upper 300 in to over 10° at the bottom. Temperature gradient is 21° k-1. Over 2 000 fractures were logged through the hole. Borehole televiewer imagery was obtained for the interval from 200 to 440 mbsf to orient the fractures for stress field analysis. Lithostratigraphical descriptions on a scale of 1:20 are presented for the full length of the core, along with core box images, as a 200 page supplement to this issue. The hole initially passed through a layer of muddy gravel to 5.5 mbsf (Lithological Sub-Unit or LSU 1.1), and then into a Quaternary diatom-bearing clast-rich diamicton to 21 mbsf (LSU 2. l), with an interval of alternating compact diamicton and loose sand, and containing a rich Pliocene foraminiferal fauna, to 27 mbsf (LSU 2.2). The unit beneath this (LSU 3.1) has similar physical properties (sonic velocity, porosity, magnetic susceptibility) and includes diamictites of similar character to those of LSU 2.1 and 2.2, but an early Miocene (c. 19 Ma) diatom assemblage at 28 mbsf (top of LSU 3.1) shows that this sub-unit is part of the older section. The strata beneath 27 mbsf, primary target for the project, extend from early Miocene to perhaps latest Eocene age, and are largely cyclic glacimarine nearshore to offshore sediments. They are described as 41 lithological sub-units and interpreted in terms of 12 recurrent lithofacies. These are 1) mudstone, 2) inter-stratified mudstone and sandstone, 3) muddy very fine to coarse sandstone, 4) well-sorted stratified fine sandstone, 5) moderately to well-sorted, medium-grained sandstone, 6) stratified diamictite, 7) massive diamictite, 8) rhythmically inter-stratified sandstone and mudstone, 9) clast-supported conglomerate, 10) matrix-supported conglomerate, 11) mudstone breccia and 12) volcaniclastic sediment. Sequence stratigraphical analysis has identified 22 unconformity-bounded depositional sequences in pre- Pliocene strata. They typically comprise a four-part architecture involving, in ascending order, 1) a sharp-based coarse-grained unit (Facies 6,7,9 or 10), 2) a fining-upward succession of sandstones (Facies 3 and 4), 3) a mudstone interval (Facies l), in some cases coarsening upward to muddy sandstones (Facies 3), and 4) a sharp-based sandstone dominated succession (mainly Facies 4). The cyclicity recorded by the strata is interpreted in terms of a glacier ice margin retreating and advancing from land to the west, and of rises and falls in sea level. Analysis of sequence periodicity awaits afirmer chronology. However, apreliminary spectral analysis of magnetic susceptibility for a deepwater mudstone within one of the sequences (from 339 to 347 mbsf) reveals ratios between hierarchical levels that are similar to those of the three Milankovitch orbital forcing periodicities. The strata contain a wide range of fossils, the most abundant being marine diatoms. These commonly form up to 5% of the sediment, though in places the core is barren (notably between 300 and 412 mbsf). Fifty samples out of 250 reviewed were studied in detail. The assemblages define ten biostratigraphical zones, some of them based on local or as yet undescribed forms. The assemblages are neritic, and largely planktonic, suggesting that the sea floor was mostly below the photic zone throughout deposition of the corcd sequence. Calcareous nannofossils, representing incursions of ocean surface waters, are much less common (72 out of 183 samples examined) and restricted to mudstone intervals a few tens of metres thick, but are important for dating. Foraminifera are also sparse (73 out of 135 samples) and represented only by calcareous benthic species. Changing assemblages indicate a shift from inshore environments in the early Oligocenc to outer shelf in the late Oligocenc, returning to inshore in the early Miocene. Marine palynomorplis yielded large numbers of well-preserved forms from most of the 116 samples examined. The new in situ assemblagc found last year in CRP-1 is extended down into the late Oligocene and a further new assemblage is found in the early Oligoccnc. Many taxa are new, and cannot us yet contribute to an improved understanding of chronology or ecology. Marine invertebrate macrofossils, mostly molluscs and serpulid tubes, are scattered throughout the core. Preservation is good in mudstones but poor in other lithologies. Climate on land is reflected in the content of terrestrial palynomorphs, which are extremely scarce down to c. 300 mbsf. Some forms are reworked, and others represent a low growing sparse tundra with at least one species of Nothofagus. Beneath this level, a significantly greater diversity and abundance suggests a milder climate and a low diversity woody vegetation in the early Oligocene, but still far short of the richness found in known Eocene strata of the region. Sedimentary facies in the oldest strata also suggest a milder climate in the oldest strata cored, with indications of substantial glacial melt-water discharges, but are typical of a coldcr climate in late Oligocene and early Miocene times. Clast analyses from diamictites reveal weak to random fabrics, suggesting either lack of ice-contact deposition or post-depositional modification, but periods when ice grounded at the drill site are inferred from thin zones of in-situ brecciated rock and soft-sediment folding. These are more common above c. 300 mbsf, perhaps reflecting more extensive glacial advances during deposition of those strata. Erosion of the adjacent Transantarctic Mountains through Jurassic basalt and dolerite-intruded Beacon strata into basement rocks beneath is recorded by petrographical studies of clast and sand grain assemblages. Core below 310 mbsf contains a dominance of fine-grained Jurassic dolerite and basalt fragments along with Beacon-derived coal debris and rounded quartz grains, whereas the strata above this level have a much higher proportion of basement derived granitoids, implying that the large areas of the adjacent mountains had been eroded to basement by the end of the early Oligocene. There is little indication of rift-related volcanism below 310 mbsf. Above this, however, basaltic and trachytic tephras are common, especially from 280 to 200 mbsf, from 150 to 46 mbsf, and in Pliocene LSU 2.2 from 21 to 27 mbsf. The largest volcanic eruptions generated layers of coarse (up to 1 cm) trachytic pumice lapilli between 97 and 114 mbsf. The thickest of these (1.2 m at 112 mbsf) may have produced an eruptive column extending tens of km into the stratosphere. A source within a few tens of km of the drill site is considered most likely. Present age estimates for the pre-Pliocene sequence are based mainly on biostratigraphy (using mainly marine diatoms and to a lesser extent calcareous nannofossils), with the age of the tephra from 112 to 114 mbsf (21.44k0.05 Ma from 84 crystals by Ar-Ar) as a key reference point. Although there are varied and well-preserved microfossil assemblages through most of the sequence (notably of diatoms and marine palynomorphs), they comprise largely taxa either known only locally or as yet undescribed. In addition, sequence stratigraphical analysis and features in the core itself indicate numerous disconformities. The present estimate from diatom assemblages is that the interval from 27 to 130 mbsf is early Miocene in age (c. 19 to 23.5 Ma), consistent with the Ar-Ar age from 112 to 114 mbsf. Diatom assemblages also indicate that the late Oligocene epoch extends from c. 130 to 307 mbsf, which is supported by late Oligocene nannofossils from 130 to 185 mbsf. Strata from 307 to 412 mbsf have no age-diagnostic assemblages, but below this early Oligocene diatoms and nannofossils have been recovered. A nannoflora at the bottom of the hole is consistent with an earliest Oligocene or latest Eocene age. Magnetostratigraphical studies based on about 1000 samples, 700 of which have so far undergone demagnetisation treatment, have provided a polarity stratigraphy of 12 pre-Pliocene magnetozones. Samples above 270 mbsf are of consistently high quality. Below this, magnetic behaviour is more variable. A preliminary age-depth plot using the Magnetic Polarity Time Scale (MPTS) and constrained by biostratigraphical data suggests that episodes of relatively rapid sedimentation took place at CRP-2 during Oligocene times (c. 100 m/My), but that more than half of the record was lost in a few major and many minor disconformities. Age estimates from Sr isotopes in shell debris and further tephra dating are expected to lead to a better comparison with the MPTS. CRP-2/2A has recorded a history of subsidence of the Victoria Land Basin margin that is similar to that found in CIROS-170 km to the south, reflecting stability in both basin and the adjacent mountains in late Cenozoic times, but with slow net accumulation in the middle Cenozoic. The climatic indicators from both drill holes show a similar correspondence, indicating polar conditions for the Quaternary but with sub-polar conditions in the early Miocene-late Oligocene and indications of warmer conditions still in the early Oligocene. Correlation between the CRP-2A core and seismic records shows that seismic units V3 and V4, both widespread in the Victoria Land Basin, represent a period of fluctuating ice margins and glacimarine sedimentation. The next drill hole, CRP-3, is expected to core deep into V5 and extend this record of climate and tectonics still further back in time.

Upper Cenozoic dinoflagellate cysts from the continental slope and rise off New Jersey

This report contains the occurrence data for dinoflagellate cysts recorded from 163 samples taken from Sites 902 through 906, during Ocean Drilling Program (ODP) Leg 150. The dinoflagellate cyst (dinocyst) stratigraphy has been presented in Mountain, Miller, Blum, et al. (1994, doi:10.2973/odp.proc.ir.150.1994), and was based on these data. This report provides the full dinocyst data set supporting the dinocyst stratigraphic interpretations made in Mountain, Miller, Blum, et al. (1994). For Miocene shipboard dinocyst stratigraphy, I delineated 10 informal zones: pre-A, and A through I, in ascending stratigraphic order. These zones are defined in Shipboard Scientific Party (1994a, doi:10.2973/odp.proc.ir.150.103.1994), and are based on my studies of Miocene dinocyst stratigraphy in the Maryland and Virginia coastal plain (de Verteuil and Norris, 1991, 1992; de Verteuil, 1995). This zonation has been slightly revised (de Verteuil and Norris, 1996), and the new formal zone definitions are repeated below. Each new zone has an alpha-numeric abbreviation starting with "DN" (for Dinoflagellate Neogene). The equivalence between the informal zones reported in Mountain, Miller, Blum, et al. (1994), and the new DN zones is illustrated in Figure 1. For clarity, I delineated both zonations in the range charts that accompany this report (Tables 1-6). De Verteuil and Norris (1996a), using these and other data, correlated the DN zonation with the geological time scale of Berggren et al. (1995). Figure 2 summarizes these correlations and can be used to check the chronostratigraphic position of samples in this report, as determined by dinocyst stratigraphy. A thorough discussion of the basis for, and levels of uncertainty associated with, these correlations to the Cenozoic time scale can be found in de Verteuil and Norris (1996a). The Appendix lists all the dinocyst taxa recorded during shipboard analyses of Leg 150 samples. Open nomenclature is used for undescribed taxa. The range charts and Appendix also include reference to several new taxa that de Verteuil and Norris (1996b) described from Miocene coastal plain strata in Maryland and Virginia. Names of these taxa in Tables 1 through 6 and in the Appendix of this report are not intended for effective publication as defined in the International Code of Botanical Nomenclature (ICBN, Greuter et al., 1994). Therefore, taxonomic nomenclature contained in this report is not to be treated as meeting the conditions of effective and valid publication (ICBN; Article 29).

14C ages of core top samples from Ontong-Java Plateau

Radiocarbon measurements on core tops from the Ontong-Java plateau confirm a previous finding by Berger and Killingley [1982] that at any given water depth, cores taken on the equator have higher accumulation rates and younger core top ages than their off-equator counterparts. Further, these new results fortify the conclusion by Broecker et al. [1991] that the increase in core top radiocarbon age with water depth rules out homogeneous dissolution within the pore waters as the dominant mechanism. Either most of the dissolution must occur prior to burial or it must occur during the first pass through the respiration-CO2-rich upper pore waters after which the calcite grains become armored against further dissolution. A puzzling aspect of this new data set is that despite the sizable difference in accumulation rate, the extent of dissolution as measured by either the CaCO3 content or the ratio of CaCO3 in the >150-µm size fraction to that in the < 63-µm fraction is no different off than on the equator. In order to reconcile the results of this study with those obtained by Hales and Emerson [1996] using in situ electrodes, it is necessary to call upon calcite armoring.

Neogene and Paleocene-Maestrichtian calcareous nannofossils in sediments of the upper continental rise off New Jersey

Maestrichtian to Holocene calcareous nannofossils from two closely spaced sites on the upper continental rise some 100 miles (161 km) southeast of Atlantic City, New Jersey, were zoned in order to help date a major canyon-cutting event in the late Miocene and to delineate and correlate other hiatuses with seismic stratigraphy. Mid-middle Eocene through middle Miocene sediments (Zones CP14 to CN6) were not recovered in these holes, but nearly all other zones are accounted for. The Eocene section is described in a companion chapter (Applegate and Wise, 1987, doi:10.2973/dsdp.proc.93.118.1987). Nannofossils are generally sparse and moderately preserved in the clastic sediments of Site 604. Sedimentation rates are extremely high for the upper Pleistocene (201 m/m.y. minimum) above a hiatus calculated to span 0.44 to 1.1 Ma. The associated disconformity is correlated with local seismic reflection Horizon Pr . Sedimentation rates continue to be high (93 m/m.y.) down to a second hiatus in the upper Pliocene dated from about 2.4 to 2.9 (or possibly 3.3) Ma. The disconformity associated with this hiatus is correlated with local seismic reflection Horizon P2 and regional Reflector Blue, which can be interpreted to mark either the onset of Northern Hemisphere continental glaciation or circulation changes associated with the closure of the Central American Seaway. Sedimentation rates in the pre-glacial lower Pliocene are only about a third those in the glacial upper Pliocene. A prominent disconformity in the upper Miocene marks a major lithologic boundary that separates Messinian(?) glauconitic claystones above from lower Tortonian conglomeratic debris flows and turbidites below. The debris flows recovered are assigned to nannofossil Zones CN8a and CN7, but drilling difficulties prevented penetration of the bottom of this sequence some 100 m below the terminal depth of the hole. Correlation of the lower bounding seismic reflector (M2/Merlin?) to a drift sequence drilled on the lower rise at DSDP Site 603, however, predicts that the debris flows began close to the beginning of the late Miocene (upper Zone CN6 time) at about 10.5 Ma. The debris flows represent a major canyon-cutting event that we correlate with the beginning of the particularly severe late Miocene glaciations believed to be associated with the formation of the West Antarctic Ice Sheet. The existence of these spectacular debris flows strongly suggest that the late Miocene glacio-eustatic low stand occurred during Vail Cycle TM3.1 (lower Tortonian) rather than during Vail Cycle TM3.2 (Messinian) as originally published. Beneath a set of coalesced regional disconformities centered upon seismic reflection Horizon Au, coccoliths are abundant and in general are moderately preserved at Site 605 in a 619-m carbonate section extending from the middle Eocene Zone CP13b to the upper Maestrichtian Lithraphidites quadratus Zone. Sedimentation rates are 37 m/m.y. in the Eocene down to a condensed interval near the base (Zone CP9). A disconformity is suspected near the Eocene/Paleocene boundary. Sedimentation rates for the upper Paleocene Zone CP8 are similar to those of the Eocene, but Zones CP7 and CP6 lie within another condensed interval. The highest Paleocene rates are 67 m/m.y. down through Zones CP5 and CP4 to a major disconformity that separates the upper Paleocene from the Danian. This hiatus spans about 2.6 m.y. (upper Zone CP3 to lower Zone CP2) and corresponds to the major sea-level drop at the base of Vail Cycle TE2.1. As the most prominent break in this Paleogene section, it may correspond to seismic reflection Horizon A* of the North American Basin. Sedimentation rates from this point to the Cretaceous/Tertiary boundary drop to 11 m/m.y., still high for a Paleocene DSDP section. No major break in deposition could be detected at the Cretaceous/Tertiary boundary.

Composition of Cape Leeuwin Manganese nodule deposit off Western Australia

In 1970 a large deposit of ferromanganese nodules was discovered on the floor of the Indian Ocean southwest of Cape Leeuwin by the research vessel USNS Eltanin. This discovery, which was based largely on bottom photographs from about 20 stations, was discussed by Frakes (1975) and Kennett and Watkins (1975, 1976). The photographs suggest that the deposit spreads, nearly continuously, over 900 000km^2, and cores showed that the nodules are essentially confined to the sediment surface. Kennett and Watkins (op. cit.) pointed to the abundance of ripple and scour marks and current-formed lineations on the present surface, and of extensive disconformities in the cores, as evidence of strong present and past bottom currents in the region. They suggested that the current action had resulted in very low sedimentation rates, which had allowed the nodule field, named by them (1976) the 'Southeast Indian Ocean Manganese Pavement', to develop. In early 1976 the authors used the research vessel HMAS Diamantina for a 10-day cruise in the region to sample the nodules in order to study their chemistry and mineralogy. During the cruise 9 stations were occupied, 8 of them successfully (Figure 1), and about 2000 nodules were recovered from the sea bed. The apparatus used was a light box dredge on the ships hydrowire, which had a breaking strain of about one tonne. Although an attempt was made to reoccupy Eltanin photographic stations, it should be noted that positioning was by celestial navigation, so errors of up to 10 km are possible.

Particle camera and settling chamber measurements at site PC (GeoB11834-3) and SC (GeoB12914-4) off Cape Blanc/Mauritania

Particles sinking out of the euphotic zone are important vehicles of carbon export from the surface ocean. Most of the particles produce heavier aggregates by coagulating with each other before they sink. We implemented an aggregation model into the biogeochemical model of Regional Oceanic Modelling System (ROMS) to simulate the distribution of particles in the water column and their downward transport in the Northwest African upwelling region. Accompanying settling chamber, sediment trap and particle camera measurements provide data for model validation. In situ aggregate settling velocities measured by the settling chamber were around 55 m d**-1. Aggregate sizes recorded by the particle camera hardly exceeded 1 mm. The model is based on a continuous size spectrum of aggregates, characterised by the prognostic aggregate mass and aggregate number concentration. Phytoplankton and detritus make up the aggregation pool, which has an averaged, prognostic and size dependent sinking. Model experiments were performed with dense and porous approximations of aggregates with varying maximum aggregate size and stickiness as well as with the inclusion of a disaggregation term. Similar surface productivity in all experiments has been generated in order to find the best combination of parameters that produce measured deep water fluxes. Although the experiments failed to represent surface particle number spectra, in the deep water some of them gave very similar slope and spectrum range as the particle camera observations. Particle fluxes at the mesotrophic sediment trap site off Cape Blanc (CB) have been successfully reproduced by the porous experiment with disaggregation term when particle remineralisation rate was 0.2 d**-1. The aggregation-disaggregation model improves the prediction capability of the original biogeochemical model significantly by giving much better estimates of fluxes for both upper and lower trap. The results also point to the need for more studies to enhance our knowledge on particle decay and its variation and to the role that stickiness play in the distribution of vertical fluxes.

Pollen analysis of surface sediment off northwest Africa

Over 100 samples of recent surface sediments from the bottomn of the Atlantic Ocean offshore NW Africa between 34° and 6° N have been analysed palynologically. The objective of this study was to reveal the relation between source areas, transport systems, and resulting distribution patterns of pollen and spores in marine sediments off NW Africa, in order to lay a sound foundation for the interpretation of pollen records of marine cores from this area. The clear zonation of the NW-African vegetation (due to the distinct climatic gradient) is helpful in determining main source areas, and the presence of some major wind belts facilitates the registration of the average course of wind trajectories. The present circulation pattern is driven by the intertropical front (ITCZ) which shifts over the continent between c. 22° N (summer position) and c. 4° N (winter position) in the course of the year. Determination of the period of main pollen release and the average atmospheric circulation pattern effective at that time of the years is of prime importance. The distribution patterns in recent marine sediments of pollen of a series of genera and families appear to record climatological/ecological variables, such as the trajectory of the NE trade, January trades, African Easterly Jet (Saharan Air Layer), the northernmost and southernmost position of the intertropical convergence zone, and the extent and latitudinal situation of the NW-African vegetation belt. Pollen analysis of a series of dated deep-sea cores taken between c. 35° and the equator off NW African enable the construction of paleo-distribution maps for time slices of the past, forming a register of paleoclimatological/paleoecological information.

Low-temperature experiments of sediment core GeoB6229-6 at the South American continental margin off the Rio de la Plata estuary

With various low-temperature experiments performed on magnetic mineral extracts of marine sedimentary deposits from the Argentine continental slope near the Rio de la Plata estuary, a so far unreported style of partial magnetic self-reversal has been detected. In these sediments the sulphate-methane transition (SMT) zone is situated at depths between 4 and 8 m, where reductive diagenesis severely alters the magnetic mineral assemblage. Throughout the sediment column magnetite and ilmenite are present together with titanomagnetite and titanohematite of varying compositions. In the SMT zone (titano-)magnetite only occurs as inclusions in a siliceous matrix and as intergrowths with lamellar ilmenite and titanium-rich titanohematite, originating from high temperature deuteric oxidation within the volcanic host rocks. These abundant structures were visualized by scanning electron microscopy and analysed by energy dispersive spectroscopy. Warming of field-cooled and zero-field-cooled low-temperature saturation remanence displays magnetic phase transitions of titanium-rich titanohematite below 50 K and the Verwey transition of magnetite. A prominent irreversible decline characterizes zero-field cooling of room temperature saturation remanence. It typically sets out at ~210 K and is most clearly developed in the lower part of the SMT zone, where low-temperature hysteresis measurements identified ~210 K as the blocking temperature range of a titanohematite phase with a Curie temperature of around 240 K. The mechanism responsible for the marked loss of remanence is, therefore, sought in partial magnetic self-reversal by magnetostatic interaction of (titano-)magnetite and titanohematite. When titanohematite becomes ferrimagnetic upon cooling, its spontaneous magnetic moments order antiparallel to the (titano-)magnetite remanence causing an drastic initial decrease of global magnetization. The loss of remanence during subsequent further cooling appears to result from two combined effects (1) magnetic interaction between the two phases by which the (titano-)magnetite domain structure is substantially modified and (2) low-temperature demagnetization of (titano-)magnetite due to decreasing magnetocrystalline anisotropy. The depletion of titanomagnetite and superior preservation of titanohematite is characteristic for strongly reducing sedimentary environments. Typical residuals of magnetic mineral assemblages derived from basaltic volcanics will be intergrowths of titanohematite lamellae with titanomagnetite relics. Low-temperature remanence cycling is, therefore, proposed as a diagnostic method to magnetically characterize such alteration (palaeo-)environments.

Seawater samples along several inshore-offshore transects off NW Africa and SE Africa

For the optimal use in palaeoceanographic studies of the stable oxygen isotopic signal and elemental composition of the calcareous photosynthetic dinoflagellate Thoracosphaera heimii, it is essential to gain detailed information about its calcification depth and spatial distribution. We therefore studied the vertical and horizontal distribution patterns of T. heimii in the upper water column (0-200 m) along three transects: an inshore-offshore gradient off Cape Blanc (CB), a south-north transect from CB to the Portuguese coast and a north-south transect off Tanzania. We compared concentrations of living cysts (cells with cell content) with chlorophyll-a, salinity and temperature measurements at the sampling depth. In order to explore the seasonal variability in cyst production, three transect off CB were sampled at three different times of the year. Living T. heimii cysts were found in the upper 160 m of the water column with highest concentrations in the photic zone indicating that the calcification of T. heimii occurs in the upper part of the water column. Maximal abundances of living cysts were found relatively often in or just above the deep chlorophyll maximum (DCM), the depth of which varies regionally from about 20-40 m off CB to about 80 m off Tanzania and along the transect from CB to the Portuguese Coast. However, there was no significant correlation at the 95% confidence level between the cyst concentrations and temperature, salinity and chlorophyll-a concentrations at the sampling depths observed. In both the Atlantic and Indian Oceans, the highest abundances of T. heimii were observed in regions where the upper water masses contained relatively low nutrient concentrations that are influenced only sporadically, or not at all, by enhanced photic zone mixing related to the presence of upwelling cells or river outflow plumes at or close to the sampling sites. The seasonal production of cysts by T. heimii appears to be negatively related to the presence of upwelling filaments across the sampling sites. Our study suggests that turbulence of the upper water masses is a major environmental factor influencing T. heimii production.

Alkenone data of sediment core GeoB9508-5 from the continental slope off Senegal

In order to investigate a possible link between tropical Northeast (NE) Atlantic sea-surface temperature (SST), Atlantic meridional overturning circulation (AMOC), and drought in the Sahel during the past 44 thousand years (kyr) we used alkenone paleothermometry and d13C of C. wuellerstorfi of a marine sediment core from the continental slope off Senegal. Our data show periods of low SST and reduced AMOC that coincided with drought in the Sahel during North Atlantic Heinrich stadials (HS). The coldest period was HS1 (ca. 15-18 kyr before present, BP) when SST decreased by more than 2°C. Moreover, the SST off Senegal lagged variations in Sahel aridity, which is in agreement with results from a freshwater hosing experiment. We conclude that variations in tropical NE Atlantic SST were not the initial trigger of millennial-scale Sahel droughts of the past 44 kyr. Instead, it is thought that these droughts were induced by substantial coolings of the extratropical North Atlantic.