METHODS

Textural analysis and bulk and clay mineralogy were run with systematic sampling frequency of two samples per core except when crossing specific layers (e.g., Horizons A, B, and B´ and the BSR) defined on the three-dimensional MCS profiles (Tréhu et al., 2002; Tréhu, Bohrmann, Rack, Torres, et al., 2003). These intervals of interest were sampled at a rate of one sample per section. We collected samples (10-cm³ tubes) for sedimentology and geochemistry analyses and for rock magnetic measurements (7-cm³ cubes) (Gràcia et al., 2003; Larrasoaña et al., this volume) from the same stratigraphic intervals so that the data are directly comparable.

Grain-size analyses were performed using a settling tube for the coarse-grained (>50 µm) fraction (Gibbs, 1974) and SediGraph 5100 for the silt and clay (<50 µm) fractions (Micromeritics, 1978). The division at 50 µm instead of 63 µm was established in order to have enough sediment samples for the settling tube, since a correct measurement requires a minimum 0.2 g of material. Although measured separately for each sample, sediment fractions were integrated in a single textural distribution using specific software. The SediGraph method determines the size distribution of particles as a function of the attenuation of an X-ray beam that crosses a transparent cell containing the samples in a dispersant suspension. The X-ray absorption is then converted into weight percentage of grain size on the basis of the sedimentation principle (Stoke's law of particles settling). This technique provides a rapid and accurate method (instrumental error = <1%) for grain-size analyses (e.g., Stein, 1985; Bianchi et al., 1999). We used the SediGraph because it is capable of sensing the total amount of material present and gives the whole size spectrum with satisfactory resolution >1 µm.

The grain-size distribution can provide information on particle size availability in the depositional system and on the type and competency of the processes operating in the system. Textural statistical parameters, such as mean size, standard deviation (sorting), kurtosis (peakness of the curve), and skewness (symmetry of the curve) are sensitive to environmental processes (e.g., Camerlenghi et al., 1995). Statistical parameters were calculated using the method of moments (McManus, 1988) on sample populations containing one-half -interval classes in all fractions. The Swan et al. (1979) sediment sorting classes were applied as follows: 0.5–0.8 for well-sorted sediments, 0.8–1.4 for moderately sorted sediments, 1.4–2 for poorly sorted sediments, and 2–2.6 for very poorly sorted sediments. Sand fraction and very coarse silt components (>50 µm) were identified using a binocular microscope, and relative abundances of components were estimated by counting a minimum of 300 grains per sample. We identified the following components: biogenic fraction (pelagic and benthic foraminifers, radiolarians, sponge spicules, etc.), light minerals (quartz, mica, and feldspar), heavy minerals, rock fragments, and diagenetic minerals (i.e., pyrite). Calcium carbonate content was determined using a Bernard calcimeter by the acid leaching method (Milliman, 1974).

Magnetic susceptibility was measured at the paleomagnetic laboratory of the Institute of Earth Sciences "Jaume Almera" (CSIC-UB) in Barcelona (Spain) (see Larrasoaña et al., this volume). The low-field magnetic susceptibility was measured with a KLY-2 susceptibility bridge using a field of 0.1 mT at a frequency of 470 Hz and was normalized by the dry weight of the samples.

Bulk and clay mineral compositions were obtained by X-ray diffraction (XRD). For bulk mineralogy, the samples were air dried, ground in an agate mortar, and packed into Al sample holders for XRD analyses. For clay mineral analyses, the carbonate fraction was removed using acetic acid. Clays were deflocculated by successive washing with demineralized water after carbonate removal. The <2-µm fraction was separated by centrifuge at 900 rpm for 1.3 min, and the clay fraction was smeared onto glass slides. Separation of the clay fraction and preparation of the bulk samples for XRD analyses were performed following the international recommendations compiled by Kisch (1991).

X-ray diffractograms were obtained using a Philips PW 1710 diffractometer with CuK radiation and an automatic slit. Scans were run from 2° to 64°2 for bulk sample diffractograms and untreated clay preparations and from 2° to 30°2 for glycolated clay fraction samples. Diffractogram interpretations and semiquantitative analyses were performed considering the integrated peak area using specific software (Xpowder, www.xpowder.com). The clay mineral proportion was estimated from the glycolated diffractogram. The 10-Å peak was used for illite, and the 7-Å peak was used for the total amount of chlorite and kaolinite, using the peak ratios at 3.54 Å and 3.58 Å, respectively, to differentiate these minerals. The 17-Å peak was used for smectites. The area between the illite peak at 10 Å and the smectite peak at 17 Å, which corresponds to the illite-smectite mixed layers, was not estimated because the peaks were usually masked by the smectite and illite peaks. The estimated semiquantitative analysis error for bulk mineralogy absolute values is 5%. In the case of clay mineralogy, the estimated semiquantitative analysis error ranges from 5% to 10%.