Ca/Mg
ratios from -0.13 to -1.30 (Table 4).Leg 167 drilling afforded the unique opportunity to investigate interstitial water geochemistry at sediment depths accessible only by scientific ocean drilling (to hundreds of meters sub-bottom depth). Leg 167 sites covered a latitudinal range of ~10°, a water depth range from ~970 to ~4200 m, and distance offshore from ~50 to ~360 km. The sites were arrayed in three onshore-offshore transects and along a north-south transect of coastal sites (Fig. 1), allowing comparisons along gradients of factors controlling sediment supply (e.g., upwelling and primary productivity, terrestrial sediment sources) and those affecting regeneration processes in the water column and near the sediment-water interface (e.g., site water depth).
Previous studies of interstitial water geochemistry in this region, with depth of penetration often limited by coring techniques, generally covered the upper tens to hundreds of centimeters of the sediment. These studies, benefiting from high to very high sampling resolution, yielded great insight into early diagenetic processes, especially with regard to organic matter degradation and related redox reactions, early mineral authigenesis, and transition metal cycling (e.g., Brooks et al., 1968; Sholkovitz, 1973; Shaw et al., 1990; Kastner, 1995; Reimers et al., 1996; McManus et al., 1997). Previous scientific ocean drilling in this region yielding interstitial water profiles is summarized in Table 2. Leg 167 analyses cover significantly greater geographic and depth ranges with a comprehensive suite of geochemical analyses.
We focus on an overview of the interstitial water geochemistry from Leg 167 sites. Interstitial water samples from all sites were analyzed aboard ship for pH, salinity (by refractometer), chlorinity, sodium, alkalinity, sulfate, phosphate, ammonium, silicate, calcium, magnesium, strontium, lithium, and potassium (Lyle, Koizumi, Richter, et al., 1997; see "Explanatory Notes" chapter for details of methods and analytical figures of merit and see respective site chapters for results for individual sites). We discuss the observations in the context of the three onshore-offshore transects (Gorda, Conception, and Baja Transects, from north to south) and the north-south coastal transect (Fig. 1; Table 2). We focus on organic matter diagenesis (sulfate, alkalinity, phosphate, ammonium), on highlights of major element chemistry (chloride at the Eel River Basin site; Ca and Mg in the coastal transect sites), and on opaline silica dissolution and diagenesis (dissolved silicate coupled with thermal gradients).
A major control on the interstitial water geochemistry of the California margin sites is the oxidation of organic matter by the sequence of reactions typically observed in marine sediments (e.g., Froelich et al., 1979; aerobic respiration, denitrification, manganese reduction, iron reduction, sulfate reduction, etc.). Interstitial water sampling at these sites was geared toward capturing large-scale gradients with depth, and therefore did not have the high resolution near the sediment-water interface typical of shallower coring studies. Nevertheless, these profiles allow us to define the depth range of sulfate reduction, its regional variation, and the accompanying geochemical changes in a way not possible with shorter cores.
Sulfate reduction is an important process in many of these sites, marked by the consumption of sulfate and the production of alkalinity, dissolved phosphate, and dissolved ammonium from the oxidation of organic matter (Table 3). In all except the two deepest water sites farthest offshore (Sites 1021 and 1010), sulfate concentrations drop below the analytical detection limit (typically ~1 mM) within the upper 10-100 m sub-bottom depth (Fig. 7; Table 3). One site, Site 1011, displayed a reappearance of sulfate deeper in the sediment column. Presumably, early in the history of sediment accumulation at this site, the entire depth range was in diffusive communication with seawater, limiting the extent of sulfate reduction. After sufficient sediment thickness accumulated, sulfate reduction initiated. Sulfate reduction in the mid-depth range of the sediment is presently consuming sulfate from above, as well as residual sulfate at depth.
The depth to sulfate depletion is shallowest in sites closer to shore and increases with distance offshore in the three transects (Fig. 7A-C). Sulfate depletion is incomplete at Site 1021 (with sulfate decreasing from seawater values around 28 mM to 19 mM at depth) and at Site 1010 (decreasing to 25 mM), with most of the decrease in the upper 100 m at both sites. Along the coast there is no simple north-south pattern in the depth to sulfate depletion. All of the coastal transect sites reach nondetectable sulfate levels shallower than 50 m (Fig. 7D). These decreases in dissolved sulfate are accompanied by increases in alkalinity, dissolved phosphate, and dissolved ammonium.
Dissolved alkalinity profiles are influenced by alkalinity production during organic matter oxidation by sulfate reduction and by alkalinity consumption in authigenic mineralization reactions. Dissolved alkalinity reaches peak concentrations in depth zones coincident with, and extending deeper below, the depth of sulfate disappearance (Fig. 8; Table 3). Maximum alkalinity values are highest in the coastal transect sites and lowest in the sites in deepest water farthest offshore. The highest alkalinity is reached in the Eel River Basin site (Site 1019), where chloride and salinity profiles indicate a significant, although not well understood, influence of gas hydrate formation and dissociation processes on the interstitial water profiles (phosphate and ammonium are also strikingly high for this site; see following section for discussion of the chlorinity profile). Below the zone of maximum concentration, alkalinity typically decreases with increasing depth. Alkalinity profiles can also be influenced by ion exchange reactions and authigenic mineralization processes.
Dissolved phosphate is produced by organic matter degradation and by release during iron reduction and can be consumed by authigenic mineralization reactions. Phosphate profiles show substantial increases with depth, to concentrations as high as 200 µM in coastal transect sites, and up to 10-15 µM in the deepest water sites (Fig. 9; Table 3). Phosphate concentrations for the Eel River Basin site (Site 1019) increase to 340-400 µM. Depth zones for maximum phosphate concentrations are generally shallower than, or coincident with, the shallowest portions of the maximum alkalinity zones (Table 3). After reaching these maxima, phosphate concentrations decrease with increasing depth to values <10 µM and as low as 2-3 µM below 150-200 m in the offshore sites (Fig. 9A-C). Phosphate profiles in all the coastal transect sites behave similarly to the more offshore sites, although the range of concentrations below 150-200 m is much larger for the coastal sites (Fig. 9D). Generally, larger phosphate maxima result in higher phosphate concentrations at depth. Presumably, phosphate uptake with increasing sediment depth below the depth zone of its maximum release by organic matter degradation is related to authigenic mineralization of phosphate-rich phases like carbonate fluorapatite, demonstrated to be significant in even the uppermost sediments of the Santa Barbara Basin (Reimers et al., 1996).
Ammonium is produced from organic matter regeneration by sulfate reduction, and its profiles can be influenced by ion exchange processes. Ammonium concentrations generally increase with increasing depth, to maximum values >5 mM and up to >35 mM in the Coastal Transect sites, and up to 3 mM in the deeper water sites (Fig. 10). The depth zone of maximum ammonium concentrations is generally deeper than the zone of maximum alkalinity values. A few sites show small decreases in ammonium at greater depth, and the Animal Basin site (Site 1011) shows a substantial decrease coincident with increasing sulfate concentrations in that site below 150 m (see Table 3 footnote). There is no simple relationship between ammonium and phosphate in interstitial waters throughout the depth range sampled because of the changing balance of influences of organic matter regeneration, ion exchange, and authigenic mineralization processes and their differing effects on the two nutrient element profiles.
The significance of organic matter supply to the sediments in driving organic matter regeneration is demonstrated by a comparison of peak alkalinity values with average organic carbon mass accumulation rates (Corg MAR; Fig. 11) over the drilled interval. There is a quasi-linear relationship between peak alkalinity and Corg MAR, with the two sites with the lowest values for these (Sites 1010 and 1021) being the two sites where sulfate reduction did not go to completion (Table 2). The Eel River Basin site (Site 1019) has higher alkalinity relative to Corg MAR than the other sites, and the Santa Lucia Slope site (1017) and Delgada Slope site (Site 1022) have relatively low alkalinity values. Site 1022 is characterized by a very thin Pleistocene section (~1 m), underlain by upper Pliocene sediments. Thus, the interstitial chemistry at this site may reflect a continued influence of diffusive exchange with seawater, reducing the maximum alkalinity value observed compared to other sites with comparable Corg MAR.
Calcium and magnesium profiles at these sites reflect the varying influence of basalt alteration in the underlying oceanic crust (resulting in linearly correlated Ca increases and Mg decreases with depth), authigenic mineral precipitation (possibly including calcium carbonate, dolomite, and carbonate fluorapatite), and ion exchange reactions, especially for Mg. Most Ca profiles decrease to a minimum and then increase deeper in the section (Fig. 12A). Minimum Ca concentrations are reached at depths close to or just below the depths at which sulfate is reduced to nondetectable levels. Minimum Ca concentrations are generally shallower than or at the top of the depth zone of maximum alkalinity values (Table 4). In general, the larger the peak alkalinity value (Table 3), the lower the minimum Ca concentration. The minimum in dissolved Ca near the depth of the alkalinity maximum is consistent with calcite precipitation. The deeper portions of the Ca profiles, with Ca increases correlated with Mg decreases, reflect the diffusional influence of alteration reactions in the underlying basaltic crust.
Mg profiles are
considerably more complex than Ca profiles, reflecting the competing influences
of ion exchange reactions, authigenic mineral precipitation, and the diffusional
influence of alteration reactions in the underlying basaltic crust. As a
consequence, depending on the site, Mg profiles display a variety of patterns (Fig.
12B). However, in the majority of sites, below the depth zone of the Ca
minimum, increases in Ca are linearly correlated with decreases in Mg,
consistent with alteration reactions in the underlying basalt and with the
conservative behavior of Ca and Mg in these portions of the profiles. Ca
gradients range from 1.2 to 11 mM/100 m, compared to the mean pelagic site
gradient of ~3-4 mM/100 m (Lawrence and Gieskes, 1981); Mg gradients range from
-0.87 to 12 mM/100 m, with Ca/Mg
ratios from -0.13 to -1.30 (Table 4).
All sites show an increase in dissolved silicate to values >1000 mM with increasing depth, consistent with the solubility of biogenic opal (Fig. 13A). There is a wide spread in dissolved silicate at any given depth across the range of sites. Downhole temperature measurements taken during Leg 167 at each drill site allowed the construction of a thermal gradient for each site. The site with the highest thermal gradient (Site 1020; 189°C/km) has the highest dissolved silicate at any given depth >50 mbsf, and the site with the lowest thermal gradient (Site 1018; 32°C/km) has the lowest dissolved silicate concentration. Using the bottom-water temperatures and these gradients, we can define modern sediment temperature.
The dissolved silicate gradients vs. temperature for the different sites (Fig. 13B) cluster much more tightly than those vs. depth, especially in the deeper sediment sections. This indicates the first-order control of temperature on biogenic opal solubility as an influence on the interstitial water geochemistry. The different behavior of silicate vs. temperature <10°C, with two apparent clusters, may indicate the further influence of sediment lithology on opal dissolution.
The temperature history of the sites gives further insight into diagenetic alteration of biogenic opal to porcellanite and chert. Knowing the modern heat-flow gradient and thermal conductivity at each site lets us use the decay in conductive heat flow through time to hindcast the temperature of a sediment layer in the drill holes (Fig. 14). Conductive heat flow, like ridge-crest topography, decreases proportional to the square root of the crust age (Anderson and Hobart, 1976) for ocean crust. Continental heat-flow patterns are more complicated, but similar t1/2 behavior should occur since the last major rifting event or volcanic episode for drill sites in the California Borderlands or on a continental margin. Knowing the modern geothermal gradient, one can hindcast the gradient in the past, provided that most of the heat loss was conductive. The temperature of any particular sediment layer depends upon its burial depth at the time of interest and the geothermal gradient.
We have hindcast the sediment temperature at the first diagenetic silica layer (opal-CT or chert) based upon the burial history and the change in heat flow (Fig. 14). Cherts or porcellanite are only found where sediment temperatures become higher than 24°-25°C (Sites 1010, 1011, 1016, 1020, and 1022). Site 1021 has no cherts but high dissolved silicate in pore waters and high biogenic opal contents. The base of the drilled sediment section at Site 1021 has never been warmer than 19°C, however. This would suggest that ~25°C is a critical temperature range for chertification of sediments.
At many of the drill sites there is a significant depth offset between the first chert or porcellanite and this hypothesized diagenetic horizon, suggesting that biogenic opal in the sediments shallower than the first chert should also be in the process of converting. There must either be a significant time lag between passage of the critical temperature and significant chertification, or other factors must also play a significant role. Earlier work has established that sediment composition likely plays a role (Kastner et al., 1977).
The California and Cascadia margins are an important observatory to study gas hydrates, because although there is high methane production all along the margin, gas hydrates are only known to occur north of the Mendocino Fracture Zone (Field and Kvenvolden, 1985; Brooks et al., 1991; MacKay et al., 1994; Hovland et al., 1995; Kastner et al., 1995; Spence et al., 1995; Tréhu et al., 1995; Yuan et al., 1996). For this reason, we looked for geophysical signatures of gas hydrates during the site surveys for Leg 167 and prepared a small-scale geochemical sampling program for any drill site where we found evidence for their existence.
One of the surprising features of the California margin is the lack of any bottom-simulating reflector (BSR) south of the Mendocino Fracture Zone despite high burial rates of organic matter and high methane contents of the sediments (Lyle, Koizumi, Richter, et al., 1997). During the Leg 167 site survey cruise (EW9504) we specifically looked for any evidence of BSRs, yet located a BSR only at Site 1019, in the Eel River Basin to the north of the Mendocino Fracture Zone. Nevertheless, high methane occurred at all the coastal sites, ranging up to 160,000 ppm in headspace samples at Site 1017 (Conception Transect; Lyle, Koizumi, Richter, et al., 1997). Although enough methane was present at both Sites 1014 and 1017 to cause a stressed APC core to blow apart on recovery, no BSRs were detected at either site. Nor did we find any evidence for gas hydrates at these drill sites from the logging or the shipboard pore-water geochemical program.
We can speculate that gas hydrates only occur north of the Mendocino Fracture Zone because of the lack of a concentrating mechanism for methane to the south, most likely the lack of strong focused fluid flow through the sedimentary section. North of the Mendocino Fracture Zone the sediments are exposed to compression and dewatering because of convergence between the Gorda/Juan de Fuca Plates and North America, which could cause methane to flow to the hydrate formation zone. The BSR region we drilled at Site 1019 is at the continental slope break, and the sediments beneath the western edge of the basin have been compacted and dewatered. South of the Mendocino Fracture Zone, compression is lower because most of the margin has now been incorporated into the Pacific Plate.
The Eel River Basin has long been known to contain gas hydrates (Field and Kvenvolden, 1985; Brooks et al., 1991). We picked a drill-site location for Site 1019 in a region known to have a BSR. Because of safety considerations Site 1019 was not located on the strongest BSR but on an area with a weak reflector on both seismic crosslines (Fig. 15, after Gallaway, 1997). Based on a 266-ms two-way traveltime to the BSR from the seafloor and an average sedimentary seismic velocity of 1600 m/s, the base of the hydrate layer should be around 212 mbsf. Because the site was moved for safety reasons, the BSR was slightly deeper than we had planned when we originally designed the drilling and the strength of the BSR was weaker.
One of the important features from the seismic survey is the patchiness of the BSR (Fig. 15). We do not observe large areas with consistent amplitude on the BSR near Site 1019 but instead observe small areas (on the order of a few square kilometers) with high-amplitude reflections separated by areas of low reflection strength or no BSR. Transitions from a high-amplitude BSR and no BSR can be abrupt. Because the BSRs drilled on Leg 146 in other parts of the Cascadia Subduction Zone are caused by trapped free gas under the clathrates (MacKay et al., 1994), the patchiness should be caused by variability in the amount of trapped methane. We do not know why there should be patchiness but suspect a highly variable supply of methane to the clathrate-forming regions, implying a variable fluid-flow regime, and/or an episodic mode of formation, where individual gas release events from deeper in the section randomly occur in space and time. Further work to resolve this issue should include a high-resolution, three-dimensional seismic survey in the region around Site 1019.
Because of the presence of the BSR at about 212 mbsf, we sampled the deeper sediment section at Site 1019 at a resolution of one interstitial water sample per core, or about three times the normal resolution. One of the primary geochemical tracers of the influence of methane hydrate dissociation below the BSR is a general freshening of the interstitial water with depth, and the presence of methane hydrates in sediments above the BSR can be marked by anomalously low chlorinity spikes as a consequence of methane hydrate dissociation during recovery (e.g., Blake Ridge Sites 994, 995, 997; Paull, Matsumoto, Wallace, et al., 1996). The Eel River Basin site interstitial water chemistry displays a striking decrease in chlorinity with increasing depth (Fig. 16), with correspondingly large reductions in salinity and sodium. Chlorinity values decrease to ~356 mM, about 65% of normal seawater values below 125 m, significantly lower than the background values of ~500 mM in the methane hydrate zone of the Blake Ridge sites.
There is no significant change in the Cl profile near the estimated depth of the BSR, nor were any unusually low-chlorinity spikes observed as artifacts of hydrate decomposition during recovery. The Cl profiles are remarkably similar to those from other drill sites along the Cascadia Subduction Zone (Leg 146, Yuan et al., 1996). At Sites 889/890, with a strong BSR, chlorinity decreased from seawater concentrations around 550 mM to about 350 mM. Assuming that this represents dilution by the volume of hydrate dissociating during recovery, Yuan et al. (1996) calculated that hydrate as a percentage of pore space varied from near 0 at the top of the section to up to 35% by the BSR. One of the features they noted, which is also apparent at Site 1019, is a lack of change in the Cl profile across the BSR. They attributed this phenomenon to an upward migration of the base of the hydrate stability field associated with a deglacial change in water temperatures.
Defining the influence of methane hydrate dissociation as the source of freshening with depth at Site 1019, as opposed to the contribution of other low-chlorinity sources, will require other isotopic (deuterium and oxygen isotopes) and elemental (iodide and bromide) profiles on these samples.
We attempted to log in situ physical properties across the BSR zone as part of the downhole logging program at Site 1019 but achieved ambiguous results. Part of the problem was because the hole depth was designed for a shallower BSR region than the actual Site 1019 location. We do not know if we completely logged the critical region of the BSR.
Figure 17 shows the Cl profile compared to the resistivity log and reprocessed seismic velocity log (M. Lyle, unpubl. data). The seismic velocity log had inconsistent traveltimes, and only the traveltimes from the two longest paths were used, corrected for borehole size, to determine the traveltime within the formation. A formation velocity was also computed from the difference in traveltimes between these two paths. All are shown in Figure 14. The two separate logging passes over Hole 1019C agreed in detail.
The corrected seismic velocity does not show any obvious major increase over the supposed hydrate interval (the interval where Cl is near its minimum) as would be expected if gas hydrates were cementing the formation. Instead there are small variations on the order of 50-100 m/s throughout the Cl-minimum zone. There is a major drop in velocity on both passes near the start of the logging runs (below 200 mbsf), which might represent the base of the hydrate zone but could also be an artifact from beginning the log.
Because hydrates are nonconductive, a massive hydrate layer appears as a high-resistivity and high-velocity layer in logs (Mathews and von Huene, 1985; Mathews, 1986). Although we find significant resistivity anomalies in the logs, they are not associated with velocity anomalies. The major resistivity anomaly occurs at the beginning of the Cl minimum, which suggests a common factor controlling both.
The lack of high-velocity anomalies within the hydrate layer and the lack of significant changes in resistivity is similar to the logging data from Sites 889 and 892 on the Cascadia Subduction Zone (MacKay et al., 1994). The BSR was difficult to locate with standard logs but was very apparent with a vertical seismic profile (VSP). The Eel River Basin BSR is thus in a common class with others found along the Cascadia Subduction Zone—it has a distinct seismic signature and a strong Cl anomaly, but is poorly defined by the standard shipboard logs. We assume that vertical seismic profiles would detect the base of the BSR at Site 1019.
There are several obvious areas for further exploration with the interstitial water data set. The profiles of alkalinity, phosphate, and ammonium can be compared within and across sites to examine the influence of organic matter degradation at each site, including investigating the rates of organic matter degradation and the relative patterns of behavior of carbon (alkalinity) and the nutrient elements P and N. This will help to address questions about the composition of organic matter being regenerated and the ultimate fate (e.g., retention in sediments, loss to the water column) of P with sedimentary redox state. The major element chemistry, especially that of Ca and Mg, points to the significance of authigenic mineralization reactions in influencing the geochemistry of these sites, including the profiles of alkalinity and phosphate. Development of records of other tracers (e.g., oxygen and deuterium isotopes, bromide, and iodide) would give greater insight into the geochemical environment, especially at the Eel River Basin site. The integration of the thermal properties of the sites with the interstitial water silicate distributions has helped to define the significance of temperature as a control on biogenic opal dissolution and its alteration. This topic deserves further consideration, including comparison to other regions with well-defined temperature gradients and interstitial silicate profiles.
