The results of geochemical analyses for the acid-soluble and residual fractions are listed in Table 1 and Table 2, respectively. The REE concentrations were excluded from this paper, and will be discussed elsewhere.
All samples examined in this study contain considerable amounts of calcareous material, reflecting their principal lithology of nannofossil-rich to nannofossil-bearing clayey sediments (Paull, Matsumoto, Wallace, et al., 1996). We avoided a strong acid for carbonate dissolution because it would have attacked the clay minerals and Mg-silicates. However, acetic acid may extract exchangeable heavy metals adsorbed on clays (e.g., Filipek and Owen, 1979), and recovery for the individual carbonate mineral phase (calcite, dolomite, siderite, and rhodochrosite) is unknown.
A preliminary extraction study was performed for the different reaction times using 1.0 M acetic acid on previously analyzed sediment samples. Table 3 shows the relative peak heights (counts/s) for calcite, dolomite, siderite, and rhodochrosite by semiquantitative X-ray diffraction analysis on the same powdered samples for geochemical analyses, a diatom-bearing nannofossil-rich clay of Unit I (Sample 164-997A-11H-1, 81-83 cm; 81.21 mbsf), and a diatom-rich clay of Unit III (Sample 164-997A-22X-6, 12-14 cm; 179.92 mbsf). The extraction was conducted for 14, 24, and 48 hr and 9 days at room temperature. The result permits us to conclude that calcite and dolomite were completely dissolved by the treatment adopted in this study (extraction for 2 nights), whereas siderite and rhodochrosite were not dissolved or even may have been reprecipitated to some extent. Although these latter two carbonate phases are rare above 110 mbsf depths, the downhole decrease in dolomite is paralleled by an increase in siderite (Paull, Matsumoto, Wallace, et al., 1996).
The section investigated here comprises well-mixed fine-grained silty-clayey detritus with varying amounts of biogenic and diagenetic carbonates. Calcite rather than dolomite, siderite, or rhodochrosite is the dominant carbonate mineral. This is shown in Figure 3, where the stratigraphic variation of geochemically calculated amounts of calcite and dolomite are plotted, based on the major cations (Ca, Mg, and Fe) in the acid-extracted fraction. The calcite content fluctuates generally between 5 to 30 wt% and rarely drops below 5 wt%, whereas dolomite tends to decrease from 4 wt% downhole to 2 wt%. An almost equivalent results was obtained with shipboard geochemical analysis for CaCO3 (Paull, Matsumoto, Wallace, et al., 1996). Although the values are too small for shipboard XRD quantification, the trends are consistent with the onboard estimates of stratigraphic variations for these two carbonate minerals.
Although the acetic acid-extracted fraction may contain acid-soluble iron sulfides and some of the absorbed species, their contribution seems to be negligible because our rinsing process extracted most of the exchangeable ions from the powdered sample, and the Fe-monosulfide is almost completely removed by reduction below 4 mbsf at Site 997 (Paull, Matsumoto, Wallace, et al., 1996). Thus, the acid-extracted fraction is regarded as having been derived from the major carbonate phase comprising calcite and dolomite, whereas siderite and rhodochrosite remained part of the residual detrital fraction.
Figure 4 shows the stratigraphic elemental variation from Holes 997A and 997B, where concentration in the bulk (sum of the acid-extracted and residual fractions) and detrital (residue from the extraction) fractions are plotted respectively. Contribution in the acid-extracted fraction is represented as the difference of the two plots.
Vanadium, Cr, Zr, Nb, Rb, and Hf are almost exclusively represented by the residual "detrital" fraction. Scandium, Ba, Mo, and As show small but significant carbonate contributions that occur at limited intervals as an addition or an anomaly where up to 20% of Sc and Ba, 95% of Mo, and 60% of As are associated in the soluble phase. The enrichment of Sc and As typically occurs below 550 mbsf depths (Fig. 4B, J).
Manganese exhibits extensive fluctuation downhole, while the insoluble residue varies very gradually (Fig. 4E). Extremely high total contents, particularly more than 1500 ppm, are always caused by acid-soluble fraction. Although the extracted fraction could not recover siderite and rhodochrosite, as stated before, siderite occurs throughout Unit III up to the top of the unit (107.83 mbsf). Because our extraction did not affect ferromanganese oxide phases, extreme enrichments of Mn in the extracted fraction can be attributed to siderite and possibly rhodochrosite.
The remaining elements (i.e., P, Mn, Co, Ni, Cu, Zn, Sr, Y, Cd, Pb, Th, and U) occur to some degree in the acid-extracted fraction and appear to have little correlation with carbonate content. Several elements are enriched in the acid-extracted fraction at some depths where the carbonate contents are relatively low. In particular, phosphorus is relatively enriched in middle Unit III at 450 to 540 mbsf, and the calcite content is generally lower than 15%.
Shipboard observations and analyses have shown the Blake Ridge sections to be fairly homogeneous and to contain a relatively constant amount of terrigenous detrital material among the sites. Although the residual "detrital" fraction given by the acid extraction may reflect the composition of terrigenous detritus plus noncarbonate biogenic fractions, the trace elements discussed here represent the geochemical signatures of the siliciclastic detritus. The mineralogical composition by X-ray diffraction analyses is discussed in a companion paper (Lu et al., Chap. 14, this volume). Stratigraphic variations of each element are plotted in Figure 4.
Most of the elements do not exhibit notable downhole enrichment nor depletion except for numerous spikes contained in Units I and II. Chromium and Mn are the exceptions and tend to be enriched downhole (Fig. 4D, E). Also notable for the Mn profile is the frequent occurrence of spiky peaks, sometimes close to 1000 ppm, in the deeper section. Associated acid-extracted fraction was also enriched in Mn at the greater magnitude. As shown earlier, rhodochrosite is not fully subject to our acid extraction. Therefore, these discontinuous enrichments of Mn in the deeper section may reflect diagenetic formation of manganese-rich carbonate (Matsumoto, 1992).
The chromium profile is much different from that of Ni, which exhibits a constant value of concentration through Unit III. Figure 5 shows the Cr/Ni profile, which further demonstrates an elevated ratio in middle Unit III to Unit II. The maximum ratios extend to 5.0, which are quite larger than average hemipelagic shales but comparable with the average value of the Canadian Shield (1.8 to 4.0; Fahrig and Eade, 1973; Shaw et al., 1976).
Thorium and Sc downhole concentrations are comparable, but Sc is enriched toward the bottom of the section below 620 mbsf (Fig. 4U, B). The Th/Sc ratio is relatively constant between 0.6 to 0.8, which is significantly lower than the average fine-grained clastic rock of continental crust composition (1.0; McLennan, 1982) because of slightly low Th concentrations (Fig. 6). This result suggests that the terrigenous clastics of Site 997 comprise an admixture of continental crust and more mafic volcanic materials that have lower Th/Sc values (i.e., Th/Sc = 0.16 for andesitic crust and Th/Sc < 0.01 for oceanic crust; Taylor and McLennan, 1985).
Uranium, another incompatible element, remains at relatively constant concentration throughout Unit III (up to 2 ppm), whereas it decreases to 1 ppm within Units I and II (Fig. 4V). Figure 7 shows downhole variation of the Th/U value, which is relatively constant within Unit III, ranging between 4.0 to 5.0. Higher ratios (>6.0) frequently occur in Unit II, which are probably caused by relatively lower sedimentation rates during the Pleistocene because long interaction with bottom seawater enables removal of oxidized U4+ (McLennan and Taylor, 1980). Concentrations of Th and U remain relatively constant in the section except within the uppermost part of Unit III (120 to 140 mbsf), where Th and U concentrations are elevated by a factor of two to three.
Based on the downhole variations of trace element concentrations in the insoluble residual fraction, the following four stratigraphic intervals have been recognized. Each interval is described by association of key elements and/or notable behavior of elements.
Interval I extends from the top of the section to 183 mbsf (Cores 164-997A-1H to 24X). Interval I comprises lithologic Units I, II, and the uppermost part of Unit III, in which most of the element contents in insoluble fraction exhibit notable fluctuation. The carbonate content of Interval I is much higher than the rest of the section below, introducing a further complication, caused by fluctuations in the bulk sediment chemical composition. This interval is characterized by a slightly but significantly enriched element suite of P, Cu, V, Hf, Y, Zr, Rb, Hf, Pb, and Th, while the other elements are depleted relative to the lower part of the section.
The mass accumulation rate for the noncarbonate fraction in this interval are generally quite low compared with the lower section, (<7 g/cm2/k.y.; Paull, Matsumoto, Wallace, et al., 1996). A distinct increase in mass accumulation rate occurs near the basal portion of Interval I, which suggests a changes in terrigenous sediment transport and origin. Nannofossil biostratigraphy indicates that the bottom of Interval I is located in the late Pliocene, at ~3 Ma if the linear sedimentation rate is applicable (Paull, Matsumoto, Wallace, et al., 1996). It also coincides with the interval of elevated abundance of diatoms (147 to 183 mbsf). The shoaling of the Isthmus of Panama is believed to have taken place approximately at the same time (Keigwin, 1982) and could have initiated or rearranged the near-shore circulation system of the Gulf Stream. It is not clear whether the enhanced sedimentation rate and diatom abundance are related.
Among the fluctuations of the elements in Interval I, noted enrichments in Sc (~25 ppm), V (~220 ppm), Hf (~6 ppm), Pb (~65 ppm), Th (~30 ppm), and U (~5 ppm) occur between 120 to 142 mbsf, associated with depletion in Rb, Zr, and Nb. The combination of enriched and depleted elements suggests an introduction of different terrigenous sediments.
The interval between 183 to 440 mbsf (Cores 164-997A-25X to 54X and 164-997B-1X to 5X) is recognized as Interval II. Geochemically, this interval is characterized by rather constant downhole concentrations. The top of Interval II is marked by decreases of V, Cu, Rb, Zn, Rb, and Ba, with increase of Hf. The acid-soluble Ba occurs only in this interval. Mass accumulation rates for the noncarbonate fraction in this interval is close to 20 g/cm2/k.y., but increases toward 30 g/cm2/k.y. in Interval III below (Paull, Matsumoto, Wallace, et al., 1996). Interval II coincides with the zone of gas hydrate occurrence. The trace element data do not have geochemical features that are diagnostic of gas hydrate occurrence.
Interval III extends from 440 to 618 mbsf (Cores 164-997B-8X to 29X) in lithologic Unit III. Scandium, V, Y, Zr, and Rb abundances decrease considerably at the top of the interval. Chromium increases by 20 ppm at the upper boundary, which accounts for the elevated Cr/Ni value (> 4.0) throughout the interval.
Interval IV comprises the bottom of the section below 618 mbsf (Cores 164-997B-30X through 47X). This interval is characterized by variable and elevated abundances of Mn, Rb, Sc, and to a lesser extent Y, Co, and Ni. Chromium abundance stays relatively constant, making the Cr/Ni value notably reduced in Interval IV.
There are no diagnostic features that enable us to identify sediment sources for any of the intervals, but the secular shift of abundances in the lithophilic elements clearly indicates compositional changes have occurred in terrigenous component of the Blake Ridge sediments.