SEDIMENTARY GEOCHEMISTRY

The sedimentary geochemistry analytical program for Site 1149 consisted of XRF analyses to determine the bulk major- and trace-element composition of 19 samples distributed throughout the entire sequence. We provide in Table T13 a brief description of the lithologies from which the analyzed samples were taken. The shallowest sample was collected at 2.18 mbsf, whereas the deepest was taken at 397.97 mbsf, <10 m from the sediment/basalt interface (Table T14). The downhole coverage provided by these samples allows for a preliminary assessment of the chemical composition of the complete sequence, with the exception of the fracture-filling calcareous marlstones of Unit V.

Chemical data are given in Table T14 (carbonate-poor samples) and Table T15 (carbonate-rich samples). For the carbonate-rich sediments, only the major elemental ratios discussed below are reported because calibration difficulties for the high-carbonate matrix preclude determination of the absolute concentrations. In most cases, data are normalized to Al (or Al₂O₃) concentration in order to remove the effects of compositional dilution by carbonate (in the limestones) or silica (in the cherts and porcelanites).

Four signals dominate the record of the bulk chemistry of the sediment. In stratigraphic order, these are: (1) metalliferous sedimentation through Unit IV, (2) the enrichment in biogenically associated elements such as Ba and Sr through Units IV and III, (3) the authigenic clays of Unit II, and, finally, (4) the influence of volcanic ash in Unit I. Each of these geochemical signatures is discussed below.

The distribution of Fe/Al shows a very well defined decrease upsection away from the basement, which we interpret as recording the waning influence of hydrothermal metalliferous sources as Site 1149 traveled away from the ridge crest (Fig. F57). Although most of this decrease occurs through Unit IV, it appears that Fe/Al values in excess of those in average shale are maintained through Unit III as well. Based on the few shipboard analyses through this particular interval, however, it is difficult to determine where the Fe/Al ratios reach values near that of average shale. Higher frequency postcruise study should resolve this issue.

The distribution of MnO/Al₂O₃ also records maximum values nearest the basement (Fig. F57), although the continual decrease upsection is not as smooth as the Fe/Al trend. Nonetheless, the very high values in the lowermost section are consistent with a metalliferous (ridge) source of the Mn in Unit IV.

Through Unit IV, P₂O₅/Al₂O₃ values also record a sharp decrease with increasing distance above the basement (Fig. F57) and maintain relatively high values through Unit III (the latter is discussed in "Biogenic Sedimentation"). We interpret the decrease in P₂O₅/Al₂O₃ to reflect the scavenging of dissolved PO₄^3- from seawater by metalliferous plume particles, a process that becomes less significant with distance from the ridge. This scavenging is observed in modern ridge systems as well as in other sedimentary sequences, and is an important sink in terms of global budgets of PO₄^3- (e.g., Feely et al., 1996).

This data set indicates that there is a substantial hydrothermal metalliferous contribution to the sediments below 350 mbsf. A measurable, yet less important, component appears to be present from at least 350 to 280 mbsf, whereas the influence through deposition of Unit III requires further study.

The depositional process that resulted in the cherts and porcelanites of Unit III, as well as the carbonates of Unit IV, supplied SiO₂, Ba, Sr, and P to the seafloor to a greater extent than if sedimentation had occurred only by terrigenous or metalliferous inputs. The cherts and porcelanites of Unit III are identified by the high SiO₂, with respect to average shale (Fig. F58). Normalizing SiO₂ to Al₂O₃ indicates that even though carbonate is the dominant lithology through Unit IV, there is still enough biogenic silica in these sediments to elevate the SiO₂/Al₂O₃ ratio significantly above that of average shale. The high variability of SiO₂/Al₂O₃ within Unit IV indicates local variations in silica diagenesis, the degree of which will be better determined with higher frequency sampling during postcruise studies.

Barium and strontium are also supplied to the seafloor by biogenic sedimentation. Barium is relatively refractory compared to SiO₂; at this site, where dissolved SO₄^2- is not significantly depleted from the interstitial waters (see "Diagenesis of Biogenic Silica and Carbonate"), barium may be a general indicator of biogenic input. Indeed, Ba (Table T14) and Ba/Al (Fig. F59) are significantly elevated through Units III and IV and increase steadily toward basement. Strontium (Table T14) and Sr/Al (Fig. F59) also record this deep increase, reflecting an increase in carbonate with depth in the sediment column. There are several processes that affect the sediment composition through Unit IV. With respect to Ba, the deepest samples probably also record a hydrothermal source, and postcruise statistical analyses will target resolving these different sources of Ba. With respect to Sr, the decrease of Sr/Al with distance above the basalt is consistent with thermal subsidence of the oceanic crust away from the ridge and below the CCD. Thus, the sediments at Site 1149 provide an excellent chemical record from which we hope to resolve metalliferous and biogenic sedimentation, as well as the depth of the CCD during the time period within which Unit IV was deposited.

As described in "Unit II" and "Ash Alteration and Formation of Authigenic Clays," the authigenic clays of Unit II are a unique sedimentary unit at Site 1149. Solid-phase chemical changes are generally much less sensitive to diagenetic change than the interstitial waters because of the high distribution coefficients between the solid and aqueous phases. Nonetheless, the preliminary bulk XRF data record authigenic clay formation. Because clay mineral authigenesis often involves cation exchange (see "Ash Alteration and Formation of Authigenic Clays"), tracking variations in sedimentary K, Na, and Mg (Fig. F60) provide an important data set complementary to the interstitial water chemistry.

The authigenic clays of Subunit IIB are enriched in K relative to the clays of Unit I, but, they appear depleted in Na with respect to Unit I, which may include a contribution from sea salt. The clearest signal is seen in the MgO/Al₂O₃ ratio (Fig. F60), which shows an increase completely bounded by Unit II. These changes in the alkalies and Mg are also apparent in the interstitial water profiles (see "Ash Alteration and Formation of Authigenic Clays"). Finally, it appears these clays are depleted in SiO₂ (Fig. F58).

The pelagic clays of Unit I include many discrete ash layers (see "Unit I"). Whereas many elements and elemental ratios record values near that of average shale through Unit I (Figs. F57, F58, F59, F60), the inclusion of ash in the sediment affects its bulk chemical composition for other elements. We exploit this effect by using the chemical deviations from average shale to infer the degree to which dispersed ash is included in the sediment. In this context, we refer to "dispersed" ash as the ash that is not contained in visually obvious discrete layers (which are commonly up to several centimeters thick). This dispersed-ash component includes discrete layers that have been pervasively bioturbated as well as ash that has been deposited directly to the seafloor as part of the pelagic rain. In either case, this dispersed component is quantitatively significant; smear-slide studies (see "Abstract") estimate that 25%-35% of the sediment may consist of ash particles.

We provide a first-order estimate of the amount of discrete ash in Unit I through a normative approach. Whereas postcruise study of other elements, along with statistical analyses, will yield a more robust result than that described below, these preliminary results yield reasonable estimates that are consistent with the smear-slide studies. We first estimate the amount of "terrigenous clay and ash" by normalizing the concentration of Al in each sample with that in average shale, according to

Although it may seem inappropriate to use the Al concentration in an average shale for a sediment that is known to include both terrigenous clay and ash, in practice, the influence of the ash on the bulk content of Al is minimal because Al in ash is relatively near to that in average terrigenous materials. Through Unit I, this calculation yields values of ~85-90 wt% for the terrigenous + ash fraction, indicating that diatomaceous material, oxides, structurally bound water, and perhaps other non-Al-bearing phases are present to ~10-15 wt%. Because the SiO₂/Al₂O₃ ratio in Unit I is very similar to that of average shale (Fig. F58), the contribution of diatoms to the sediment is relatively small. The calculation yields unrealistically large (i.e., >100%) values of terrigenous material through Unit II, confirming that these pelagic clays (which have been significantly diagenetically altered) are not composed of average shale type material. Because of this, we are unable at this point to quantify the abundance of dispersed ash in Subunit IIA. The concentration of terrigenous material decreases through the cherts, porcelanites, limestones, and metalliferous sediments of Units III and IV, recording compositional dilution.

How quantitatively significant is the dispersed ash in Unit I? The Al/Ti and Nb/Al ratios in Unit I are significantly different from those in average shale, both suggesting the presence of dispersed ash (Fig. F61). Al/Ti values are higher than average shale, whereas Nb/Al values are lower. In consideration of the near constancy in Al concentrations between terrigenous matter and ash, these contrasts reflect a depletion in both Ti and Nb, which in turn result from the presence of both basaltic and rhyolitic compositions of the dispersed ash.

Because Nb is depleted in arc ashes and enriched in average shale (19 ppm; Taylor and McLennan, 1985) (0.5-6 ppm; Taylor and Nesbit, 1998), we use the Nb concentrations of the bulk sediment along with the results of Equation 1 to calculate the amount of dispersed ash in a given sample according to

which assumes a constant value of 19 ppm Nb in average shale and zero in ash. Equation 2, therefore, subtracts the terrigenous component from the mixed "clay and ash" result of Equation 1. Ideally, we would like to be able to use in Equation 2 a chemical element that had zero concentration in the ash, but our limited data set at this point precludes doing so. Because actual Nb concentration in the ashes may be as much as 4-6 ppm, the third term in Equation 2 overestimates the amount of terrigenous clay; thus, the final calculation yields a minimum estimate of dispersed ash.

The results of this calculation for the samples in Unit I indicate that these sediments contain 35-50 wt% dispersed ash (Fig. F62). These values are somewhat higher than the smear-slide estimates of 25%-35%, which tend to underestimate ash, not only because fine-grained particles are underestimated but also because some of the originally ashy material may be altered to clay and thus be difficult to distinguish from terrigenous material. Thus, given the uncertainties in our calculations, we believe these results are likely to be accurate to within 5-10 wt%.

Overall, the preliminary sedimentary chemistry reveals several critical intervals that should be targeted during postcruise studies. In stratigraphic order, Site 1149 records a well-developed hydrothermal metalliferous sedimentary profile that documents the decreasing influence of plume precipitation with lateral distance from the ridge. Diagenetic enrichments of silica, and perhaps other elements through the cherts and porcelanites of Unit III, modify to a certain degree the primary depositional signal beneath the elevated high productivity of equatorial deposition, although some paleoproductivity indicators (e.g., Ba/Al) may prove useful. Authigenic clay formation in Unit II is identified in both the solid and aqueous phases. Lastly, the high proportion of dispersed ash in Unit I, which is both visually identified in smear slides and chemically quantifiable on the basis of bulk chemical analyses, can perhaps be more accurately partitioned into basaltic and rhyolitic contributions by expanding the chemical data set with further shore-based analyses.