The late Quaternary succession at Site 1017 consists primarily of siliciclastic silty clay to clayey silt, with minor amounts of biogenic calcium carbonate (<10%; foraminifers and nannofossils), and rare biogenic opal (~2%; chiefly fragments of sponge spicules, with fewer diatoms; Lyle, Koizumi, Richter, et al., 1997; Tada et al., Chap. 25, this volume). Rare, centimeter-scale, bioclastic or siliciclastic sand layers are concentrated between two intervals, 320-500 and 1390-1470 cmbsf, representing MISs 2 and 4, respectively—both sea-level lowstands. Sand layers are typically sharp based and some are normally graded. Sands are primarily quartzofeldspathic and contain variable, generally minor, quantities of mica and abraded bioclasts. Trace amounts of amphibole, augite, and a green garnet (possibly uvarovite) are present in most sands and coarse silt fractions.
With the exception of sand layers and sparse, millimeter-scale, silty laminations discernible by X-rays, the upper Quaternary sequence is bioturbated and massive. Near-continuous burrowing is displayed by computed tomographic (CT) vertical X-rays of sediment slabs and CT-scans of bedding parallel slices (Behl et al., 1997; Morris and Behl, 1998; Fig. 2). Chondrites and Planolites trace fossils are present, but most common are unidentified millimeter- to centimeter-diameter vertical to nearly horizontal burrows in addition to pervasive mottling. Ongoing bioturbation (up to tens of centimeters deep) likely obliterated all but the most rapidly deposited and buried sedimentary structures.
The mean grain size for the entire measured spectrum ranges from 7.4 to 30.3 µm though the succession, with an average value of 19.5 µm (Table 1; Fig. 3). The mean of the sortable silt fraction (10-63 µm) ranges from 17.6 to 33.9 µm (average 24.8 µm). The degree of sample-to-sample variation is fairly high, regularly fluctuating 2-4 µm between adjacent samples, yet certain trends and cyclicities can be discerned in the succession, especially within a smoothed record (21-pt. running average; i.e., 63 cm or ~3 k.y.). The uppermost 2200 cm (0 to ~113 ka) is distinctly coarser (20.0 µm average whole spectrum; 25.2 µm average sortable silt) than the lowermost 400 cm (~113-130 ka, 16.1 µm average whole spectrum, and 21.8 µm average sortable silt). The upper section consists of ~12 irregular coarse to fine cycles (smoothed record) with the lower interval spanning another two. Cycles range from ~120 to ~250 cm in thickness with an average period of ~8.5 k.y. in the upper section and 11.2 k.y. in the lower part (Fig. 4). Stratification is poorly developed, and the cycles expressed by variation in mean grain size and sorting are not visually distinct or sharply bounded.
The degree of sediment sorting (as shown by the standard deviation of the grain-size distribution) is inversely related to grain size (Fig. 3, Fig. 4). Each coarser interval is associated with a corresponding increase in the standard deviation of the grain size. The sediment is best sorted near the top and bottom of the succession at ~0-220, 1900-1990, and 2160-2487 cmbsf (0-10 ka/Holocene, 97-102 ka/MIS 5c, and 111-130 ka/MIS 5e, respectively; Fig. 3, Fig. 4).
Grain-size distributions of samples were analyzed to resolve the relative contributions and modes of different grain-size components. In particular, we focused on two intervals in the Holocene and MIS 3, including every other 3-cm increment (6-cm spacing). Most samples are unimodal or bimodal and have fine tails, allowing us to resolve the spectrum into two or three components with individual log normal distributions (Fig. 5; see Tada et al., Chap. 25, this volume, for details). The principal mode is located between 8 and 15 µm (i.e., chiefly fine silt); lesser components have modes between 2 and 4 µm (i.e., clay) and 39 and 73 µm (coarse silt to very fine sand; Table 2). The fine silt component makes up 59%-86% of the samples by volume; clay, 5%-27%; and coarse silt, 0%-28%. The modes and contribution of each of these components varies within the stated ranges. The greatest variation is in the abundance of the coarse silt to fine sand fraction, which covaries with the mean grain size (Set LB; Fig. 6, Fig. 7). This relationship suggests that the principal control of textural variation at Site 1017 is the presence or absence of the coarse silt fraction.
Superimposed on the larger scale trends and meter-scale cycles are decimeter-scale oscillations in mean grain size, standard deviation, and the contribution of the coarse silt fraction. These cycles and the negative correlation between grain size and "sorting" are well illustrated in two expanded profiles through the Holocene and part of MIS 3 (Fig. 6, Fig. 7). Recognition of the bimodal or trimodal grain-size distribution demonstrates that a larger standard deviation does not reflect the degree of sediment sorting but, instead, the relative abundance of the three grain-size components.