MATERIALS AND METHODS

In addition to this data report, sample specifications and data lists can be extracted from the on-line PANGAEA data information system (www.pangaea.de).

Site 1089

Site 1089 was drilled in a water depth of 4624 m north of the Agulhas Ridge (40°56.18´S, 09°53.64´E). Four drill holes penetrated a Pliocene to Holocene sequence of calcareous muds (Shipboard Scientific Party, 1999a). The study of Kuhn and Diekmann (2002) was concentrated on the upper part of the spliced composite section between 0 and 90 meters composite depth (mcd), corresponding to the last 580 k.y. (Hodell et al., 2001; Cortese and Abelmann, 2002). For the upper 14 mcd, we refer to sediment core PS2821-1, recovered with a piston corer at the same position during a presite survey with research vessel Polarstern (Kuhn et al., 1998). The depth scale of core PS2821-1 was converted to the Site 1089 mcd scale (Table T1) by the correlation of downcore reflectance and magnetic susceptibility data from both records.

Samples for bulk analyses were taken at 10-cm intervals, representing time steps between 0.25 and 1.00 ka. Mineralogical and granulometric analyses were conducted every 10 cm in the upper 28.4 mcd (marine isotope Stage [MIS] 1 to MIS 6) and at lower temporal resolution in the older strata: every 20 cm down to 68.8 mcd (MIS 7 to MIS 11) and every 30 cm below.

Site 1090

Three holes (1090B, 1090D, and 1090E) were drilled at Site 1090 (42°54.8´S, 8°54.0´E; 3702 m water depth) that yielded a spliced Quaternary section down to 44 mcd (Shipboard Scientific Party, 1999b) used for the MPT study (Diekmann and Kuhn, 2002). Because the uppermost part of the Site 1090 section seems to be disturbed, we connected it with nearby sediment core PS2489-2 (42°52.4´S, 8°58.4´E, 3794 m water depth), taken during a presite survey with Polarstern (Gersonde, 1995). We spliced the Site 1090 record at 12.40 mcd with the core PS2489-3 record, corresponding to an age of 408 ka. Both records show good overlap in their benthic foraminiferal 18O records for the interval between 6.3 and 16.8 mcd (340-560 ka) (Becquey and Gersonde, 2002).

Samples were taken at 10-cm intervals, representing—with a few exceptions—time steps between 5 and 15 k.y. for the interval below 1200 ka (35.5 mcd) and time steps between 1.5 and 6.0 k.y. for the younger interval.

Bulk Sediment Parameters

Bulk sediment composition was analyzed on freeze-dried and ground subsamples. Carbonate, organic carbon, and nitrogen were measured on LECO carbon element analyzers (CS-125, CS-400, and CNS-2000). The percentage of carbonate was calculated from the difference between percentage bulk carbon and percentage organic carbon, multiplied by 8.33. Biogenic silica (opal) was measured by an automated leaching method with a relative analytical precision of 4%-10% (Müller and Schneider, 1993). The proportion of nonopaline and noncalacareous constituents is regarded as the lithogenic siliciclastic or terrigenous sediment fraction.

For calculating mass accumulation rates ([MAR] in grams per square centimeter per thousand years) of sediment components, linear sedimentation rates must be multiplied with values of dry bulk density (DBD) and the proportion of the sediment component. DBD was inferred from dry sediment density, measured with a Micromeritics AccuPyc 1330 pycnometer, and shipboard measured wet bulk densities by gamma ray attenuation densitometer with a multisensor track at 2 cm intervals (Shipboard Scientific Party, 1999a). At Site 1090, based on the good correlation between carbonate concentrations and grain densities in shipboard samples, DBD was calculated from wet bulk density (gamma ray attenuation densitometer) and grain density was estimated from carbonate concentration. The inferred high-resolution dry bulk densities for the sample sets used for this study agree well with widely spaced shipboard measurements of dry bulk density in the course of physical properties investigations (Shipboard Scientific Party, 1999b).

Clay and Silt Mineralogy

Mineralogical analyses were carried out by X-ray diffraction measurements on random silt mounts and on glycolated preferentially oriented clay mounts. Sample preparation and semiquantitative evaluation of X-ray diffractograms followed techniques explained in detail elsewhere (Ehrmann et al., 1992; Petschick et al., 1996). Mineral proportions were calculated from weighted peak areas recorded in the X-ray diffractograms. Quartz/feldspar ratios (Qz/Fsp) refer to the quotient of the 4.26-Å peak area times five of quartz divided by the 3.19- to 3.24-Å double peak area of plagioclase and K-feldspar. The relative abundance of the clay mineral groups in the clay fraction is summed to 100% from weighted peak areas recorded in the X-ray diffractograms (Biscaye, 1965); the 17-Å peak area for smectite, the 10-Å peak area times four for illite, and the 7-Å peak area for kaolinite and chlorite subdivided in proportion to the relative areas of their 3.57- and 3.54-Å peaks, respectively. The 5-Å/10-Å peak intensity ratio is an indicator for illite chemistry. Values <0.15 represent Fe-Mg-rich illites (biotitic illite) and those values >0.40 are indicative for the presence of Al-rich illites (muscovite and sericite) (Esquevin, 1969). Relative analytical precision for minor clay components is 8%-14% and 6%-9% for major clay components (Ehrmann et al., 1992).

Grain-Size Analyses

For the granulometric characterization of terrigenous silt and clay, subsamples were washed through a 63-µm mesh for grain-size separation of sand and gravel from the mud fraction, silt, and clay. The fine fraction was treated with 3% hydrogen peroxide solution and 10% acetic acid for the removal of organic carbon and carbonate and for disaggregation. Silt and clay was separated at 2 µm in settling tubes. The silt fraction was dispersed in sodium polyphosphate solution and measured with a Micromeritics SediGraph 5000E to determine the grain-size distribution in 1/10 steps. We used silt samples instead of mud samples to get a better resolution of coarse silt and to avoid flocculation effects caused by the high abundance of smectite in the samples (Stein, 1985).

Since we only analyzed carbonate-free samples, the measured grain-size distributions might be biased by the individual grain-size distributions of opaline particles. However, opal concentrations are generally below 15 wt% in the carbonate-free samples, ruling out a significant impact of opal on terrigenous (Terr) grain-size pattern (Fig. F1). Moreover, terrigenous/opal ratios vary independently from silt grain-size parameters (Kuhn and Diekmann, 2002). The grain-size distributions of unleached carbonate-free samples were discussed in terms of silt/clay ratios and proportions of noncohesive sortable silt (particles >10 µm) in the <63-µm fraction (Kuhn and Diekmann, 2002). The latter parameter has been established as a qualitative proxy of relative changes in bottom-current strengths (McCave et al., 1995).

NEXT