106 t (Zierenberg, et al., 1998.). This estimate does not include a significant portion of the mineralization that was deposited below the seafloor in the feeder zone, which is discussed below.
Eight holes were drilled in the vicinity of the BHMS deposit to assess the thickness and lateral extent of the massive sulfide deposit and to determine the nature of the hydrothermal feeder zone in the sediments and basalt underlying the deposit. The BHMS deposit is the result of a complex interaction between hemipelagic and turbiditic sedimentation, igneous activity, and hydrothermal circulation. The deposit includes iron- and zinc-rich massive and semimassive sulfides, a well-developed feeder zone characterized by crosscutting copper-rich veins and sulfide impregnation of sediment, and a deep stratiform zone of copper-rich mineralization that was deposited in a sandy unit that was an important conduit for lateral fluid flow.
Hole 856H was drilled on the topographic high point of the BHMS deposit and is thought to cut through the thickest portion of the sulfide deposit. This hole can be considered as a reference section for comparison of the vertical and lateral variations in the sulfide deposit and associated host rocks (Fig. F7). From top to bottom, the hole penetrated successively: massive sulfide (0-103.6 mbsf), a sulfide feeder zone (103.6-210.6 mbsf), interbedded turbidite and pelagic sediments (210.6-431.7 mbsf), a 39.4 m interval of basaltic sills and sediment, and 28.9 m of basaltic flows.
Most of the massive sulfide unit was drilled during Leg 139 and the material recovered is described in Davis, Mottl, Fisher, et al., (1992) and Mottl, Davis, Fisher, and Slack (1994). The upper 75 m of the massive sulfide shows a variable extent of recrystallization of pyrrhotite to pyrite including significant intervals of massive pyrite with only trace amounts of other sulfide minerals. The lowermost 25 m of massive sulfide is predominantly pyrrhotite-rich massive sulfide with 1%-5% interstitial sphalerite/wurtzite and intermediate solid solution (ISS), most of which has unmixed into isocubanite host grains with fine chalcopyrite exsolution lamellae. Only minor veining and recrystallization of pyrrhotite to pyrite and magnetite is present in this zone, and the lowermost samples of massive sulfide that were recovered contain significantly more isocubanite and chalcopyrite.
The underlying feeder zone is divided into three subunits. The uppermost unit is sulfide-veined siltstone and mudstone (103.6-152.9 mbsf) where the major sulfides are isocubanite, chalcopyrite, and pyrrhotite (Lawrie and Miller, Chap. 5, this volume; Marquez and Nehlig, Chap. 9, this volume) found in subvertical, anastomosing veins. Below this interval, sediments have only minor disseminated and vein-controlled sulfide (152.9-201 mbsf). From 210.6 to 249.0 mbsf is a sulfide-banded sandstone, highly enriched in copper relative to the overlying mineralization, that has been designated the Deep Copper Zone (DCZ). The contact between the DCZ and the underlying sediment is extremely sharp.
The sediment underlying the feeder zone is nonmineralized to slightly mineralized turbidites. Based on hydrothermal alteration and color, three intervals were defined on board ship. The uppermost subunit (210.6-249.0 mbsf) is relatively unaltered sediment composed of gray, fine-sand turbidites. Below this is an interval of greenish gray siltstone and mudstone (249.0-345.5 mbsf) with abundant chlorite. The lowermost sedimentary subunit is again predominantly gray fine sandstone turbidites (3245.5-431.7 mbsf).
The base of the sedimentary section (431.7-471.3 mbsf) has been intruded by five basaltic sills from 1 to >5 m thick. The igneous rocks are variably altered and crosscut by veins containing quartz, chalcopyrite, pyrrhotite, sphalerite, calcite, and epidote. These sills are intercalated with indurated and hydrothermally metamorphosed sediment, and this unit is interpreted to represent transitional oceanic crust similar to, but much thinner than, the sill-sediment complex penetrated by Hole 857A during Leg 139 (Davis, Mottl, Fisher, et al., 1992). Thirty meters of basaltic flows (471.1-500 mbsf) underlying the sill sediment complex was drilled. These flows are similar in mineralogy and alteration to the sills. If these flows represent the top of a standard section of oceanic crusts, then it appears that the Bent Hill deposit formed near the transition from normal oceanic crust to sedimented-rift type crust (Currie and Davis, 1994). However, the hydrothermal activity was clearly later than the formation of the igneous basement, as sufficient time passed for the accumulation of 350 m of turbidites that formed a thermal blanket and hydrologic seal over the more permeable oceanic crust.
The surface of the sulfide deposit exposed at the seafloor is characterized by gossanous iron-oxide fragments, clasts of pyrrhotite-rich massive sulfide, and vuggy pyritic massive sulfide. A carapace of brecciated and clastic sulfide, including oxidized material, overlies the massive sulfide in most of the drill holes. Some of this material seems to represent material formed by oxidative weathering and disaggregation of massive sulfide exposed at the seafloor. Goodfellow et al. (1993) interpreted some sulfide fragments recovered in piston cores from the top BHMS deposit to represent detritus from collapsed sulfide chimneys that have not undergone the extensive hydrothermal recrystallization that characterizes the bulk of the deposit.
Surprisingly, although the massive sulfide mound must have extended to 100 m above the turbidite-covered seafloor, there is essentially no clastic sulfide deposited with the bulk of the turbidites that bury the flanks of the mound. A unit of interbedded clastic sulfide and sediment was recovered in the uppermost 10 m of sediment in holes to the east, west, and south of the sulfide mound (Fouquet, Zierenberg, Miller, et al., 1998). Similar material was recovered from the uplifted flank of Bent Hill in Hole 856B (18-24 mbsf; Rigsby et al., 1994), but the occurrence of this interval in a zone of slumping makes determination of the original stratigraphic depth uncertain (Mottl et al., 1994). The clastic sulfides are predominately sulfide-rich turbidites composed of sand- to clay-sized sulfide grains with blue-green clay similar to that observed in the weathered residue formed by the dissolution of collapsed anhydrite-rich hydrothermal chimneys from the active vent fields (Turner et al., 1993).
The hydrothermal recrystallization of the massive sulfide, the presence of some intercalated sulfide and sediment at the top of the deposit, and the distribution and mineralogy of clastic sulfide around the edges of the mound suggest that a second episode of hydrothermal venting occurred more recently at the Bent Hill deposit. Larger massive sulfide clasts in the turbidites are composed predominantly of interlocking bladed pyrrhotite with pyrite, sphalerite, and chalcopyrite (Goodfellow and Franklin, 1993; Fouquet, Zierenberg, Miller, et al., 1998). Fragments of vuggy or massive pyrite similar to that which is present throughout much of the recrystallized portions of the massive sulfide mound have not been recognized in the clastic sulfide turbidites. This suggests that this interval represents collapsed chimney debris shed from the mound rather than redeposited fragments of clastic sulfide derived by erosion of the mound. Clastic sulfides have higher contents of Pb, As, Sb, and Se than massive sulfide from the Bent Hill deposit (Goodfellow and Franklin, 1993; Fouquet, Zierenberg, Miller, et al., 1998) indicating an increased contribution of metals leached from sediment during this hydrothermal stage. Lead in the clastic sulfides has a higher ratio of radiogenic isotopes compared with the massive sulfide zones (Bjerkgård, et al., in press). Although analyses are limited at present, the base metal ratios of this material are similar to material from the presently active vents in Middle Valley (Ames et al., 1993). This second pulse of hydrothermal flow may have been responsible for the extensive recrystallization of massive sulfide in the hydrothermal mound as well as the abundant anhydrite that fills pore space in many of the massive sulfide samples from the flanks of the mound. The presence of slumped sediment immediately below the intervals of clastic sulfide is consistent with an episode of faulting or intrusion of basalt beneath Bent Hill, either of which could have reinitiated hydrothermal discharge by breaching the silicified caprock that presently forms a hydrologic seal on this hydrothermal system, as discussed below.
The massive sulfide zone in the BHMS deposit is a large mound with relatively steep flanks and a near-horizontal base (Fig. F7). The base of the massive sulfide lens is interpreted to represent a time horizon marking the onset of vigorous hydrothermal venting at this site (Zierenberg et al., 1998). Most of the massive sulfide was deposited above the sediment/water interface and the rate of deposition was faster than the rate of sedimentation such that little to no sediment was incorporated into the core of the deposit. The mound eventually built ~100 m above the surrounding seafloor. The thickness of massive sulfide changes abruptly between Hole 1035G, located 68 m west of Hole 856H, and 9 m further west at Hole 1035A, suggesting that these holes are near the western edge of the sulfide deposit, although thinning caused by normal faulting and slumping is possible. Hole 1035D, drilled 75 m east of Hole 856H, penetrated in excess of 40 m of massive and semimassive sulfide, so the eastern edge of the deposit remains undefined. In contrast to the relatively low-grade pyritic massive sulfide recovered from Holes 1035A and 1035G, mineralization in Hole 1035D resembles the upper portions of the deposit recovered in Hole 856H and has variable contents of pyrrhotite, pyrite, magnetite, sphalerite, and isocubanite-chalcopyrite. Compared to the central part of the mound, this lateral extension is depleted in high-temperature minerals such as chalcopyrite, isocubanite, and pyrrhotite. Authigenic anhydrite and pyrite are common as veins, nodules, and disseminated crystals. Hole 1035F was drilled at the base of the BHMS deposit 60 m south of Hole 856H and penetrated at least 80 m of massive to semimassive sulfide, including massive to semimassive pyrrhotite and pyrite with altered sediment (14.5-22.5 mbsf), vuggy massive pyrite with minor chalcopyrite, anhydrite, and sphalerite (22.5 77 mbsf), and massive to semimassive, fine-grained pyrrhotite and pyrite with white clayey altered mudstone (77-89.9 mbsf). Altered sedimentary intervals intercalated with the sulfide are located toward the edges of the sulfide mound, especially near the lower or upper contacts of massive sulfide with sediment. Local veining and replacement of sediment are indications of subsurface sulfide mineralization.
The lack of interbedded sediment throughout much of the deposit indicates that the bulk of the massive sulfide at Bent Hill formed in a single episode of sulfide mound building that was rapid relative to the rate of turbidite sedimentation. Hydrothermal recrystallization of early deposited sulfide minerals is very common in the massive sulfide. The primary hydrothermal precipitates were dominantly pyrrhotite with less abundant high temperature Cu-Fe sulfide that have bulk-grain compositions within the field for ISS, and sphalerite and/or wurtzite (Davis, Mottl, Fisher, et al., 1992; Krasnov et al., 1994; Duckworth et al., 1994; Fouquet, Zierenberg, Miller, et al., 1998; Lawrie and Miller, Chap. 5, this volume). Much of the deposit has been recrystallized to pyrite ± magnetite, and base metal sulfides show textural evidence for dissolution in some parts of the deposit and late-stage veining and replacement in other horizons. This process is generally referred to as zone refining and has resulted in intervals of higher grade base-metal mineralization, especially toward the top and flanks of the deposit (Davis, Mottl, Fisher, et al., 1992; Fouquet, Zierenberg, Miller, et al., 1998).
We estimate that the Bent Hill Massive Sulfide lens is ~200 m across at its base along the east to west drill hole transect. The north to south dimensions of the deposit are constrained by the lack of massive sulfide in Hole 856B, drilled 153 m north of Hole 856H, and by Hole 1035F, drilled 60 m south of Hole 856H, which penetrated >80 m of massive sulfide. A conservative estimate of the size of the massive sulfide mound can be obtained by assuming the lens has the shape of a half sphere with a diameter of 200 m. Using the measured density of 4.2 g/cm3 (Fouquet, Zierenberg, Miller, et al., 1998), the massive sulfide lens has a minimum size of 8.8
106 t (Zierenberg, et al., 1998.). This estimate does not include a significant portion of the mineralization that was deposited below the seafloor in the feeder zone, which is discussed below.
One of the major accomplishments of Leg 169 was the first successful penetration through the feeder zone mineralization underlying a seafloor massive sulfide deposit. This zone represents the pathway for the hydrothermal fluids that deposited the massive sulfide. Feeder zone mineralization associated with similar ancient massive sulfide deposits on land often represent a significant portion of the economic reserves of deposits. In contrast to some ancient massive sulfide deposits, the feeder zone mineralization under the BHMS deposit has not been metamorphosed or deformed.
Feeder zone mineralization is best represented in core from Hole 856H (100-210 mbsf), which recovered a spectacularly developed sulfide feeder zone (Fig. F8). The feeder zone has been subdivided into three subunits based on the style and intensity of mineralization and alteration (Fouquet, Zierenberg, Miller, et al., 1998; Marquez and Nehlig, Chap. 9, this volume). The upper 45 m of the feeder zone is intensely veined at the top with vein density decreasing downcore. Veins range from at least 8 cm to <1 mm; the thickest veins near the top of this interval typically show crack seal textures indicative of multiple episodes of vein opening because of fluid overpressure, followed by vein-filling mineral precipitation (Fig. F9). The veins are predominantly subvertical in the upper section of the feeder zone. Intergrown isocubanite-chalcopyrite and pyrrhotite are the predominant vein-filling minerals with minor to trace amounts of sphalerite and late-stage marcasite or pyrite ± magnetite replacing pyrrhotite. Nonsulfide minerals are not abundant and include chlorite and quartz, which generally are present along the vein margins. The altered turbiditic sediment that hosts the veins is altered to chlorite, quartz, and fine-grained rutile and titanite (Lackschewitz et al., 2000).
The interval from ~145 to 200 mbsf is less intensely veined. Vein thickness averages ~1 mm. The relative proportion of subhorizontal veins and bedding parallel disseminated sulfide increases progressively down the core. Many subvertical veins branch off into subhorizontal sulfide impregnations in more permeable horizons.
Core recovered below ~200 mbsf is again intensely mineralized, containing up to 50 vol% sulfide minerals (Deep Copper Zone). Sulfide mineralization occurs as impregnations and replacement of the host sediments and is strongly controlled by variation in the original sedimentary textures. Much of the mineralization is developed in medium- to coarse-grained, locally crossbedded, turbiditic sand. The sulfide mineralization locally preserves the original sedimentary structures (Fig. F10). Electron microprobe analyses show that the dominant sulfide mineral is ISS, which typically has unmixed to host grains with a composition near stoichiometric isocubanite (Fe:Cu 2.0-2.1) containing coarse exsolution lamellae of very iron-rich chalcopyrite (Fe:Cu 1.2) (R. Zierenberg, unpubl. data). Pyrrhotite is present but is much less abundant than in the overlying veins; other sulfide minerals are present in only trace amounts. A representative sample of high-grade mineralization from this zone contains 16.1 wt% Cu (Fouquet, Zierenberg, Miller, et al., 1998). The host rock is completely altered to a silvery gray chlorite. Hydrothermal quartz and rutile are present as clear, euhedral crystals disseminated in the chlorite matrix.
Geophysical logging reveals the same type of vertical variation in the feeder zone that was observed in core. From 100 to 145 mbsf, the apparent resistivity reaches extremely high values in response to intense sulfide veining. Resistivity tends to decrease with depth, as do density and acoustic velocity, and porosity reaches the minimum values measured in this hole. The Formation MicroScanner (FMS) images give a very good picture of the vein network and of its change in density with depth. The DCZ is well imaged, and logging indicates that this zone is ~13 m thick. In contrast to the overlying intervals, sulfide veining is essentially absent in the DCZ. Below 210 mbsf, the sediments are only slightly altered, and the logs record changes in structure and lithology of the underlying succession of interbedded mudstones, siltstones, and sandstones. FMS images show a high level of fracturing related to faulting in two intervals (221-239 and 250-270 mbsf) that also have low resistivity and high porosity. However, core recovery from these intervals was poor and rocks with structures indicative of significant faulting were not present in the recovered core. The high gamma-ray counts and low density indicate dominantly clayey sediment in these zones. The tops of these two intervals are marked by a contact surface dipping 50° to the west. Between these two intervals, the FMS maps a succession of fractures dipping the same direction at about 50°-70°. A sharp decrease in X-rays derived from K (measured by the shipboard multisensor track) in core recovered above 292 mbsf may reflect fault-controlled hydrothermal circulation and alteration of the overlying rocks.
Feeder zone mineralization is also well developed in Hole 1035F on the south flank of the massive sulfide but is only weakly developed in Hole 1035D to the east and is absent below the massive sulfide horizons in Holes 1035A and 1035G to the west. The greater extent of the sulfide feeder zone mineralization north-south is consistent with structural control of fluid flow by rift parallel faulting. The DCZ is developed in approximately the same stratigraphic interval in Hole 1035F as observed in Hole 856H (Fig. F7), but pyrrhotite and sphalerite are more abundant than in the center of the system.
Drilling on the east and west flanks of the deposit in Holes 1035D and 1035A, respectively, was terminated near the top of the DCZ horizon (~175 mbsf) because of destruction of the tungsten carbide drill bits used in these holes. Although the core recovery in this interval was low, the material retrieved from this depth is intensely silicified turbiditic mudstone weakly veined by pyrrhotite. Penetration of this silicified zone in Hole 1035G using the more robust tricone bits used to drill Holes 856H and 1035F showed that mineralization at this horizon on the flanks of the deposit is more weakly developed and consists mostly of pyrite and pyrrhotite veins and replacements in silicified mudstone. ISS grains from a sample from this zone have not unmixed and have a composition that is more Fe rich than isocubanite (Fe:Cu 2.4), similar to ISS included in sphalerite from massive sulfide near the top of the deposit that formed as rapidly quenched chimney fragments (R. Zierenberg, unpubl. data; Goodfellow et al., 1993).
The silicified horizon above the DCZ appears to represent an important hydrologic control on the high-temperature hydrothermal system that formed the massive sulfide. The transition from predominantly vertical crack-seal veins in the upper part of the feeder zone to subhorizontal mineralization controlled by sedimentary texture at the base of the feeder zone indicates that cyclic overpressure capable of fracturing the rock only occurred near the seafloor. During periods when the high-permeability pathways represented by the veins were sealed, fluid was forced to flow laterally into the more permeable sandy turbidite units. Conductive cooling of this ponded hydrothermal fluid facilitated silica deposition (Janecky and Seyfried, 1984), thus sealing the top of this interval. Conductive cooling also resulted in precipitation of ISS, which is the least soluble of the sulfide minerals that are present in this deposit. Development of high-grade, copper-rich replacement-style mineralization below the sulfide feeder zone was not anticipated prior to drilling. Permeable horizons below ancient massive sulfide deposits on land represent potential exploration targets that may not have been tested, as exploration drill holes are often terminated once the feeder zone mineralization has been penetrated.
Two lines of evidence suggest that the silicified horizon at the top of the DCZ also controls present-day fluid circulation in this part of Middle Valley. Pore fluid collected from the deepest section of Hole 1035A has distinctly lower Cl and Na and higher Li and K than pore fluids from overlying sediments (Fouquet, Zierenberg, Miller, et al., 1998). The composition of the low-salinity fluid closely matches the composition of hydrothermal fluid collected from the 264°C hydrothermal vent located 300 m south on the flank of the ODP Mound (Butterfield et al., 1994).
Hole 1035F, which penetrated through the silicified zone, was observed by a television camera lowered down the drill string to be vigorously venting hydrothermal fluid. It thus appears that the silicified zone represents an impermeable caprock that prevents fluids from the hydrothermal reservoir zone developed in the upper igneous crust (Davis and Fisher, 1994) from reaching the seafloor, except in areas where the seal has been penetrated by fracturing or, more recently, by drilling (Zierenberg et al., 1998).
A second massive sulfide mound located ~350 m south of the BHMS deposit was investigated in a single drill hole. Hole 1035H was drilled on a relatively flat bench near the southern peak of ODP Mound (Fig. F3). Only one hydrothermal chimney was known to be actively venting in the Bent Hill area before Leg 169. The Lone Star vent is located ~54 m north and 36 m east of Hole 1035H at the north end of ODP Mound. The vent site consists of a single anhydrite chimney issuing fluid at 264°C. A second area of hydrothermal activity was briefly observed during the camera survey before drilling Hole 1035H and was later visited and sampled by submersible. This vent site, Shiner Bock vent, is ~30 m north of Hole 1035H and was also venting 264°C fluid from an anhydrite chimney.
Hole 1035H recovered 238 m of mineralized rock, which included three massive sulfide zones interbedded with feeder zones and weakly mineralized sediments (Fig. F7). Sulfide-veined and impregnated sediment underlies the massive to semimassive sulfide zones and grades into nonmineralized sediment. The massive sulfide consists of coarse-grained pyrite variably infilled and replaced by sphalerite (5%-40%), magnetite (up to 30%), clay minerals, minor chalcopyrite, and traces of galena. Massive sulfide from this mound contains significantly more sphalerite than material recovered from the BHMS deposit.
The uppermost core (0-8.8 mbsf) recovered clastic vuggy pyrite and sphalerite and hydrothermally altered claystone. Between 8.8 and 30 mbsf is a unit of massive sulfide consisting of angular, variably sized clasts composed of pyrite, marcasite, and sphalerite with minor chalcopyrite and isocubanite in a matrix of finer grained clastic sulfides. Magnetite and hematite are present in variable amounts. The major nonsulfide phases are dolomite and ankerite. Underlying this massive sulfide is a sulfide feeder zone where fine sandstone, siltstone, and silty claystone are impregnated and cut by thin veins of pyrrhotite, pyrite, sphalerite, chalcopyrite, and anhydrite (30-55.2 mbsf). The feeder zone grades into weakly mineralized fine sandstone and claystone (55.2-74.6 mbsf).
Below this is a second unit of massive sulfide composed of pyrite, pyrrhotite, and sphalerite (74.6-84.2 mbsf) followed by a second feeder zone where siltstone and fine sandstone are veined and impregnated with pyrrhotite and chalcopyrite (80-123 mbsf). Sedimentary rocks are locally brecciated and hydrothermally altered to chlorite.
A third interval of base metal-rich massive sulfide consisting of compact to vuggy black sphalerite (40%-70% of the sulfides) with pyrite, magnetite, and talc or clay infilling voids was intersected (123-142.3 mbsf). Copper sulfides are a minor component in this interval. At the base of this massive sulfide zone, a short interval of highly altered and mineralized rocks rich in amphibole and epidote was recovered. Deeper in the section is a relatively complex feeder zone (142.3-190.3 mbsf) where three thin zones of massive to semimassive sulfide (between 150 and 154 mbsf, 162 and 163 mbsf, and 180.7 and 182.3 mbsf) are within sulfide-veined sediment. Silty claystone, siltstone, and sandstone are impregnated with Cu-Fe sulfides and hydrothermally altered to chlorite. The semimassive to massive sulfide with altered sediment contains as much as 80% Cu-Fe sulfide. Replacement by sulfide and hydrothermal silica preserve planar laminae and cross-bedding that occurred in the original sediment. This rock appears to be very similar to the banded Cu-Fe sulfide found in the DCZ in Hole 856H. Between 190.3 and 219.1 mbsf the sediments are less intensely mineralized. Interbedded claystone, siltstone, and sandstone are partly silicified and altered to chlorite and contain veinlets and impregnation of Cu-Fe sulfides. The last section was cored in hemipelagic and turbiditic sediment (219.1-247.9 mbsf) that is partly silicified, weakly chloritized, and contains anhydrite molds. Minor disseminated pyrite is present locally in this interval.
Multiple episodes of hydrothermal discharge are clearly evident in the section drilled below the ODP Mound. The presence of three stacked sequences of massive sulfide underlain by feeder zone mineralization encountered over a depth interval of 210 m is an indication of episodic hydrothermal discharge. Lower sulfide horizons are locally very coarse grained and have been extensively recrystallized in response to the passage of hydrothermal fluids that formed overlying massive sulfide zones. Estimation of mineral abundance, supplemented by limited onboard geochemical analyses, shows that the material recovered from this hole is a much higher grade (2.9%-51% Zn, 0.12%-0.62% Cu) than massive sulfide from the Bent Hill deposit. High-grade (8.0%-16.6% Cu), Cu replacement mineralization is present in approximately the same stratigraphic horizon as encountered below the BHMS and raises the question of the possible continuity of the DCZ mineralization between the two deposits (Fouquet, Zierenberg, Miller, et al., 1998).
Drilling at this site also reinitiated hydrothermal venting, similar to that at Hole 1035F. The intensity of the hydrothermal flow out of the 25-cm-diameter borehole was sufficient to carry drill cuttings of sedimentary rock 3 cm in diameter, and massive sulfide fragments up to 0.5 cm in diameter, more than 10 m above the seafloor. This violent venting suggests that the drilling penetrated into an overpressured zone (Zierenberg et al., 1998). The Ridge Inter-Disciplinary Global Experiments (RIDGE) program of the U.S. National Science Foundation initiated an event-response cruise using the Thomas G. Thompson in an effort to sample the vent fluids to establish a baseline for recognizing compositional changes in the vents and to evaluate the microbiological response to initiation of hydrothermal venting. By the time the site was visited with the Canadian remotely operated vehicle ROPOS 3 weeks later, the velocity of venting had slowed considerably and 272°C hydrothermal fluid was venting from a 1-m-high anhydrite-rich chimney that had grown in the bore hole opening (D.S. Kelley and M.D. Lilley, unpubl. data). This chimney had grown to a height of ~9 m when visited by the Alvin submersible in 1998 (R. Zierenberg, unpubl. data).
The Dead Dog vent field is the primary locus of hydrothermal venting in Middle Valley with at least 20 active high-temperature (250° to 276°C; Ames et al., 1993) hydrothermal vents. Investigations during Leg 169 included drilling near an active hydrothermal mound (Site 1036), deployment of pop-up pore pressure instruments (PUPPIs), and reinstrumentation of seal bore holes at Sites 858 and 857.
Site 1036 targeted the active Dead Dog Mound (Fig. F4) in order to (1) constrain the fluid flow rates and pathways for hydrothermal fluid and seawater entrained into the hydrothermal upflow zone, (2) study the mode of formation of the hydrothermal mounds, (3) determine the effects of hydrothermal activity on sediment diagenesis and alteration, and (4) determine the presence, continuity, and nature of a suspected caprock horizon at ~30 mbsf. In order to meet these objectives, we drilled Holes 1035A, 1035B, and 1035C along a northwest-southeast transect from the top to the margin of the 7-m-high Dead Dog active hydrothermal mound. Hole 1036A was located ~9 m west of a 268°C hydrothermal vent and was cored to 38.5 mbsf. Hole 1036B was offset ~37 m to the northwest and cored to a depth of 52.3 mbsf through the edge of the mound. Hole 1036C was offset another 34 m to the northwest and cored to a depth of 54.2 mbsf in an area within the vent field, but between hydrothermal mounds.
The results from the drilling at Site 1036 showed that Holocene- to late Pleistocene-age hemipelagic sequence of flat-lying silty clay has a relatively consistent thickness between holes (25.09 m at Hole 1036A, 25.60 m at Hole 1036B, and 26.70 m at Hole 1036C). The ~7-m-high relief of the Dead Dog Mound is close to the stratigraphic thickness of a unit of chimney-derived rubble that was recovered in Hole 1036A drilled on top of the mound. This unit is a heterogeneous mixture of clasts derived from the collapse of anhydrite chimneys, now partly altered to gypsum, greenish gray clay (probably a Mg-bearing smectite), and fine-grained pyrrhotite with subordinate pyrite and sphalerite. The lowermost 1.6 m of the sequence has a dark color, probably a result of the slow, ongoing process of gypsum dissolution, leaving behind a residue of clay and sulfide. Silty clay of apparent hemipelagic origin was recovered from beneath the chimney debris at Hole 1036A but is present at the surface of the seafloor at Holes 1036B and 1036C. The mound appears to be entirely a buildup of rubble and did not form by inflationary growth through precipitation of authigenic hydrothermal minerals in the subsurface, nor are the mounds formed by tectonic uplift. The Dead Dog Mound is presumably a recently formed feature built upon the Holocene fill of Middle Valley.
Hydrothermal alteration boundaries are zoned away from the upflow conduit that feeds the Dead Dog vent (Goodfellow and Peter, 1994.). The boundaries of an upper authigenic carbonate alteration zone and a lower anhydrite alteration zone deepen away from the mound, reflecting the decreasing geothermal gradients. The core from Hole 1036A contains just a few dolomite nodules in the sediment underlying the chimney-derived rubble. In the core from Hole 1036B, numerous dolomite and calcite nodules are found in the interval from 10.20 to 32.20 mbsf. Hydrothermal alteration also affected the preservation of microfossils in the cores. Core-catcher samples prepared for paleontological analysis were devoid of foraminifers below 9.5 mbsf in Hole 1036A, below 18.6 mbsf in Hole 1036B, and below 35.0 mbsf in Hole 1036C. Siliceous microfossils disappear at even shallower burial depths in all the holes. The depth at which the magnetic susceptibility of sediments decreases to near zero because of hydrothermal alteration (magnetic wipeout zone) also increases with distance from the vent.
Authigenic anhydrite is present as disseminated crystals, nodules, and cement and is a distinctive alteration facies of the hydrothermal system. The depth of first appearance of anhydrite was 9.5 mbsf in Hole 1036A, 18.55 mbsf in Hole 1036B, and 42.20 mbsf in Hole 1036C. Anhydrite forms by the mixing of Ca-rich, sulfate-depleted hydrothermal fluids with sulfate-bearing seawater that is drawn into the upper portions of the hydrothermal vent field in response to rapid fluid upflow (Glenn, 1998; Stein and Fisher, in press). Normal seawater concentrations are present in pore fluids sampled from shallow depths, but indicate seawater recharge and high-temperature sediment alteration at depth (Fouquet, Zierenberg, Miller, et al., 1998). PUPPIs deployed in the area of the drilling transect also recorded differential pore pressure indicative of localized seawater recharge (Schultheiss, 1997). Anhydrite has inverse solubility and becomes less soluble at higher temperatures. Experimental data (Bischoff and Seyfried, 1978) and geochemical modeling (Janecky and Seyfried, 1984) indicate the depth to the top of the anhydrite precipitation zone represents the approximate position of the 160°C isotherm. Fluids with compositions similar to those sampled at Dead Dog vent are present below 40 mbsf in the cores from Hole 1035B. Inflow of cold seawater may also have influenced the abundance and distribution of subseafloor microbes by preventing the flux of H2S though near-surface sediments where in situ temperatures below the upper thermal limits for hyperthermophiles occur (Cragg et al., Chap 2, this volume; Summit et al., Chap. 3, this volume).
Although Hole 1036A is only 9 m away from an active vent, no major hydrothermal vein network was intercepted at depth and the uppermost sediments are not highly lithified by hydrothermal alteration. Only a few hydraulic breccias and some anhydrite veins were recovered. The absence of a stockwork zone underlying the mound and the lack of extensive induration suggest that the Dead Dog hydrothermal mound is an immature feature, consistent with the occurrence of collapsed chimney material overlying very young sediments. Hydrothermal lithification (silicification?) increases with depth to the point that in each of the holes, the XCB cutting shoe was destroyed in cutting the lowermost core at ~50 mbsf.
An important scientific objective for Leg 169 was to replace the existing CORK and thermistor strings installed in Holes 857D and 858G as part of an active hydrologic flow experiment designed to enable temporal monitoring of changes in subseafloor temperature and pore pressure. The operation plan at Hole 857D was to replace the 300-m-long thermistor string installed during Leg 139, which was entirely within the cased section of the hole, with an 898-m-long thermistor string that would provide temperature information extending across the area of the hole with high hydraulic conductivity extending to near the bottom of the hole (Fig. F11). An attempt to replace this thermistor string during Leg 146 had damaged the existing CORK. The borehole seal of the CORK in Hole 858G had also failed and was observed on Alvin dives in 1993 to be venting hydrothermal fluid. An array of ocean bottom seismometers (OBS) had been deployed by scientists at Scripps Institution of Oceanography in advance of the drilling in an attempt to monitor fracturing that could be induced by allowing cold seawater to flow into the hot formations. If the holes were sealed and had recovered to thermal equilibrium with the surrounding sediment, then opening the holes to overlying cold seawater could result in a significant overpressure relative to the hot hydrostatic head. A hole-to-hole hydrologic experiment was planned to monitor the pressure in Hole 858G generated in response to a transient pressure pulse to be induced by removing the CORK from Hole 857D, which is located 1.6 km to the south.
Initial observation of the CORK in Hole 858G from the vibration-isolated television camera revealed no evidence that the fluid flow observed earlier was continuing. Upon recovery of the lower half of the CORK housing, we discovered that the inside housing of the CORK body was coated with hydrothermal precipitates, predominantly anhydrite, but with abundant pyrrhotite and pyrite. Naturally occurring hydrothermal chimneys from the Dead Dog vent field are predominantly composed of anhydrite, with only minor amounts of pyrite and trace amounts of pyrrhotite. A water sampling temperature probe (WSTP) run 20 m into the casing indicated fluid temperatures >220°C prior to failure of an O-ring in the tool. The fluid sample collected contained a mixture of seawater and hydrothermal fluid that extrapolates to a hydrothermal end-member similar to fluids sampled by Alvin in the vent field. A temperature run with the Geothermal Resources Council ultrahigh-temperature multisensor memory dewared tool indicated that the borehole was essentially isothermal at 272°C (near the maximum measured in the vent field) at depths below 85 mbsf. The temperature tool could only be run to 205 mbsf because of an obstruction in the cased portion of the drill hole. A wash core was recovered from an unlined core barrel and is believed to be representative of the material drilled from inside the casing at depths below the blockage at 205 mbsf. The wash core was predominantly loosely aggregated, fine-grained pyrrhotite and pyrite with minor amounts of anhydrite. Fluid sampled from the wash core had a very low magnesium content consistent with a high temperature origin and minimal dilution with seawater. The composition of this fluid sample is very similar to vent fluids collected in the Dead Dog field.
It was anticipated that removal of the CORK from Hole 857D would result in a rapid inflow of water into the hole as observed during Leg 139 by packer and flowmeter experiments. Temperature profiles and fluid samples confirmed that cold seawater was entering the hole. The new thermistor string was installed, and the hole was resealed with a new CORK. Subsequent measurements downloaded from the CORK data loggers provide the first detailed characterization of the downhole temperature profile in a seafloor hydrothermal field (Davis and Becker, 1998) and determined the pressure gradient that drives fluid from the vicinity of Hole 857D toward Hole 858G in the Dead Dog vent field.
Escanaba Trough provides an interesting contrast to Middle Valley. The two sites share many similarities; both are sediment-buried spreading centers covered by ~500 m of Pleistocene turbiditic sediments, and each site hosts older large massive sulfide deposits as well as active hydrothermal vent sites that are not presently depositing massive sulfide. Differences between the sites include a lower spreading rate at Escanaba (2.4 vs. 5.8 cm/yr) and lower background heat flow such that the projected temperatures at the sediment/basalt interface are ~80°C for Escanaba vs. ~270°C for Middle Valley (Davis and Becker, 1994a; Davis and Villinger, 1992). There are pronounced differences in the composition of the hydrothermal deposits as well. Trace element and lead isotope compositions of massive sulfide from Escanaba Trough require a significantly higher contribution of sedimentary-derived components (Zierenberg et al., 1993; Koski et al., 1994). The active hydrothermal vents at Escanaba Trough are lower temperature (220° vs. 285°C) and also have a higher concentration of components derived by hydrothermal leaching from sedimentary sources (Campbell et al., 1994). A primary objective of the drilling was to attempt to understand the reasons for the compositional differences between the Middle Valley and Escanaba Trough hydrothermal systems.
Site 1037 was drilled near the spreading axis ~5 km south of the known area of hydrothermal venting to provide a reference section through relatively unaltered sedimentary fill of Escanaba Trough. This site provided an important insight into the rapid infilling of Escanaba Trough by large scale turbidites. Recovery from Hole 1037B included a relatively complete sedimentary sequence of the Escanaba Trough fill sediment between 0 and 495.6 mbsf (Fouquet, Zierenberg, Miller, et al., 1998). Metamorphosed claystone and siltstone was recovered beneath this sequence between 495.6 and 507.8 mbsf. Basaltic rocks were recovered from 507.8 mbsf to the base of the hole at 546 mbsf.
Zuffa et al. (2000) defined six major lithostratigraphic units on the basis of sand/silt-dominant vs. mud-dominant turbidites and changes in composition and sedimentary structure and magnetic susceptibility data that assisted in distinguishing graded turbidites in fine-grained units.
Unit A (0-1.7 mbsf) consists of Holocene hemipelagic fossiliferous greenish gray clay. A 14C date on planktonic foraminifers indicates a sedimentation rate of 0.17 cm/yr (Brunner et al., 1999), which is consistent with previous determinations of the Holocene sedimentation rate from Escanaba Trough (Karlin and Lyle, 1986).
Unit B (1.7-120.6 mbsf) is the uppermost of a series of fining- and thinning-upward sequences that record extremely rapid burial of Escanaba Trough by megaturbidite beds. In the lowermost 50 m of this unit, only massive fine- to medium-grained sand was recovered. Coring disturbance in this unit precludes definitive determination of the sedimentary history, but Zuffa et al. (2000) interpret this sand interval as the base of a single megaturbidite flow (120.6-63.5 mbsf) that grades upward to pelite. This thick sand-rich interval is at the depth from which petrographically similar sands were recovered from Deep Sea Drilling Project Hole 35, drilled ~15 km to the south. Davis and Becker (1994b) correlated this massive sand interval with an acoustically transparent horizon mapped by single-channel seismic reflection profiling and speculated that it was deposited as the result of the floods draining glacial Lake Missoula. The overlying interval (63.5-1.7 mbsf) contains 10 thinning-upward megaturbidite beds.
Unit C (120.6-177.6 mbsf) is a fining- and thinning-upward sequence similar to Unit B including a 35-m-thick, poorly sorted, fine- to medium-grained sand at the base that Zuffa et al. (2000) interpret as three megaturbidite beds.
Unit D (177.6-366.6 mbsf) is a finer grained unit dominated by mud-rich turbidites. The individual turbidites can exceed 10 m in thickness but generally have only thin basal units of silt or very fine-grained sand.
Unit E (366.6-495.6 mbsf) consists predominantly of turbiditic beds with siltstone bases. Individual depositional units in this interval are thinner than in overlying units. Sedimentary structures typical of turbidites (parallel, wavy, and cross laminations) are more common than in the overlying megaturbidite sequence.
Core recovery was poor between 495.6 and 507.8 mbsf, but most of the cored material is silicified calcareous claystone rubble. These rocks are just above the transition from Escanaba Trough sediment fill to igneous basement and have been thermally altered. Metamorphic minerals include euhedral quartz, zoned plagioclase feldspar, brown biotite, and in the deepest samples, epidote and actinolite (Fouquet, Zierenberg, Miller, et al., 1998).
Brunner et al. (1999) determined sedimentation rates for Hole 1037B based on foraminiferal assemblages and accelerator mass spectrometer 14C dating of planktonic foraminifers. Average depositional rates between 11 ka and 32 ka were >10 m/k.y., an amazingly fast rate for sediments deposited at 3200 mbsl. The ages of the upper 120 m of turbidites coincide with the formation of the channeled scabland deposits formed by catastrophic draining of glacial Lake Missoula. Brunner et al. (1999) interpret the uppermost turbiditic fill of Escanaba to be derived from hyperpycnal turbidite flows generated as jökulhlaups from glacial Lake Missoula entered the ocean near the present mouth of the Columbia River. Previous work by Vallier et al. (1973) had identified the Columbia River Basin as the source of the pyroxene-bearing sands cored in Escanaba Trough. Sedimentary petrography on samples collected during Leg 169 confirms the Columbia River basin as the dominant sedimentary source for most of the Escanaba Trough sediment fill. Zuffa et al. (2000) suggest that the compositional changes in heavy mineral assembles downcore that were previously interpreted to reflect a change with depth to a source area dominated by material derived from the Klamath mountains is instead related to selective dissolution of pyroxene downcore. Based on detailed near-surface seismic reflection profiles, bathymetric data, and sidescan sonar data, Zuffa et al. (2000) determined that turbidites have reached Escanaba Trough by flow down the Cascadia Channel, along the Blanco Fracture Zone, south across the Tufts Abyssal Plain to the Mendocino Escarpment, and eventually northward into Escanaba Trough (Fig. F5), a journey of >1100 km. The thickest sedimentary beds are interpreted to represent frozen turbidites that entered the box canyon defined by Escanaba Trough where they stopped, dropping their entire sedimentary load.
The rapid sedimentation rate in the Escanaba Trough helps to preserve the chemical signatures in pore waters, and pore-water concentration-depth profiles are likely to correlate with in situ diagenetic changes little modified by diffusion. Concentration profiles of several species (Li, Na, and Cl) show local maximum at 360 mbsf. Li and Na also increase as basement is approached, as do Ca, Sr, B, and Cl (James and Palmer, 1999; Gieskes et al., Chap. 1, this volume). Most of the dissolved elements have lower concentrations as basement is approached. In contrast, K and Mg generally have values decreasing below seawater concentrations with depth, but concentrations increase again just above the igneous rocks. Sulfate is depleted within the upper 100 m of the sediment column by bacterial sulfate reduction. Diagenetic reaction in these sediments apparently provides a net source of B, Li, Na, Ca, Sr, and NH4 to pore waters. A decrease in Mg and K concentration with depth is interpreted as being caused by clay-mineral formation.
The organic matter in the sediments of this reference site is generally immature from the top to ~450 mbsf. There are a few horizons with more mature bitumen, which is interpreted to be derived from recycled, older sedimentary detritus carried in by turbidites; however, the low extract yields throughout the hole indicate that there are no petroleum zones. Below 450 mbsf, the organic matter is thermally altered, but low concentrations of bitumen reflect in situ maturation without migration/expulsion, suggesting that the thermal event that resulted in metamorphism of the lowermost sediments was not accompanied by extensive fluid circulation higher in the sediment column. Temperatures recorded during coring of the uppermost sediments show a generally linear increase to 15°C at 85 mbsf, the deepest measurement. Extrapolation of the thermal gradient indicates a temperature of 84°C at 500 mbsf, the approximate depth of the sediment basalt interface (Fouquet, Zierenberg, Miller, et al., 1998). These data are consistent with surface heat-flow measurements (Davis and Becker, 1994b) and indicate that the thermal pulse responsible for organic maturation and sediment alteration has decayed.
The nature of the transition from sediments into the uppermost igneous crust differs from the thick-sheeted sill complex encountered in Middle Valley. Chilled margins and an absence of any recovered sedimentary interbeds suggests that the uppermost basalts are flows. The basaltic rocks recovered from ~530 mbsf to the base of the hole (546 mbsf) range from fine- to coarse-grained and appear to be a single cooling unit, most likely a thick, ponded flow. Although the basalts are generally fresh appearing, there are indications of postcrystallization thermal alteration, consistent with the evidence for metamorphism of the overlying sediments. Alteration minerals are generally vein controlled and include smectite, chlorite, albite, and actinolite/Mg hornblende. Veins completely filled with interpenetrating, fine acicular Mg hornblende crystals up to 2 mm long are locally abundant, indicating the presence of relatively high temperature (~350°C?) hydrothermal fluids. However, the lack of pervasive hydrothermal alteration of the basalt is consistent with a local, short-lived thermal event, perhaps driven by a shallow subvolcanic sill (Fouquet, Zierenberg, Miller, et al., 1998).
A transect of holes (Holes 1038A to 1038I) was drilled across Central Hill in Escanaba Trough with the highest priority being to drill through the massive sulfide deposits and into the alteration zone near the center of the hydrothermal upflow zone (Fig. F6). A primary objective was to establish the causes of the major compositional differences between the deposits at Middle Valley and Escanaba Trough. A series of shallow RCB exploratory holes was targeted primarily at the exposed mounds of massive sulfide in order to establish the extent, composition, and drilling potential and conditions of the massive sulfide in this area prior to starting a deeper drill hole. Drilling conditions, especially in the massive sulfides, proved difficult and our understanding of this site is hampered by generally low core recovery. Regardless, the drilling at Site 1038 provided important information that allows us to interpret the differences between the Middle Valley hydrothermal system and that at Escanaba Trough.
Massive sulfide recovered from Central Hill suggests that the mineralization forms only a thin (5-15 m) veneer over the sediment. No major intersection of massive sulfides was recovered. In each hole started on massive sulfide outcrops, the inferred thickness of the massive sulfides was similar to the extent of sulfide mound build-up above the seafloor as observed on submersible dives. Seafloor observations had indicated that this system was much younger than the massive sulfides at Bent Hill (Zierenberg et al., 1993), and drilling did not encounter any sediment-buried sulfide deposits. The sulfide samples recovered were dominantly massive pyrrhotite with abundant pyrite in some samples. Much of the pyrite formed by replacement of earlier pyrrhotite. Several samples contain abundant sphalerite, and minor amounts of Cu-Fe sulfide are present in some samples. Gangue minerals are minor with barite as the predominant phase (Fouquet, Zierenberg, Miller, et al., 1998).
A characteristic difference between this site and the Bent Hill deposit in Middle Valley is the absence of a well-developed vein-dominated feeder zone in the sediment under the sulfide mounds. Chlorite is the dominant alteration mineral in samples recovered beneath the massive sulfide deposits (Lackschewitz et al., Chap. 6, this volume). Changes in the properties of magnetic minerals from samples collected below massive sulfide mounds also indicate hydrothermal alteration (Urbat et al., 2000). The well-developed hydrothermal lithification, silicification, and sulfide veining that characterizes the feeder zone mineralization beneath the Bent Hill deposit was not encountered at Escanaba Trough. The extensive surface exposure of relatively small massive sulfide mounds coupled with the lack of development of a distinct feeder zone suggests that hydrothermal venting was probably diffuse and short lived. This is consistent with the lack of chimneys, an indication of focused fluid discharge, and the occurrence of thin, high-temperature pyrrhotite crusts on the sediment observed during submersible dives (Zierenberg et al., 1993).
Incomplete core recovery precludes detailed correlation of sedimentary units between Sites 1037 and 1038. Using grain-size and magnetic susceptibility data, a correlation can be made between the upper 100 m of cores from Holes 1038A, 1038B, 1038G, 1038I, and the reference hole (Hole 1037B). Biostratigraphic correlation is precluded because most samples examined from Site 1038 contain very few foraminifers or are barren. Alteration of foraminifers is clearly associated with thermal and hydrothermal effects. However, Hole 1038F provides confirmation that Holocene sedimentation is exceptionally fast, probably >290 cm/k.y.
Stratigraphic correlation is more tenuous below 100 mbsf. Hole 1038I was drilled near the flat-topped apex of Central Hill and is the only hole that penetrated through the sediment fill to igneous rocks presumed to represent the uppermost oceanic crust. The most reasonable interpretation of the sedimentary sequences suggests that each of the lithostratigraphic units defined on board the JOIDES Resolution for Hole 1037B can be correlated with Hole 1038I within the depth of a single coring interval (~10 m). This interpretation implies that the lowermost 100 m of sediment section deposited above the basaltic rocks in Hole 1037B are not present above the basaltic rocks drilled at the base of Hole 1038I. The depth to the seafloor at Hole 1036I is 85 m shallower than at Hole 1037B and the apex of Central Hill has been uplifted ~110 m above the surrounding turbidite fill. Data from detailed seismic profiling (Davis and Becker, 1994b; Zuffa et al., 2000), sediment coring (Normark et al., 1994; Karlin and Zierenberg, 1994), submersible observations (Zierenberg et al., 1994), and sidescan sonar (Ross et al., 1996) indicate that the uplift of Central Hill occurred during the Holocene and postdates the deposition of the valley-filling turbidites.
Near-surface (upper 3 m) sands lack lateral continuity across Central Hill, suggesting that they may have been locally derived, perhaps by slumping off the steep fault scarps that surround the basin. Sediment debris flows recovered by gravity coring indicate that slumping of sediment accompanied uplift of Central Hill (Normark et al., 1994). Submersible observations near the crest of Central Hill revealed steep scarps and erosional channels interpreted to have formed by local mass flow (Zierenberg et al., 1994). The thick sand horizon that characterizes the base of Unit B in Hole 1037B can be correlated with sands recovered at Site 1038. The top of this interval is identified in all holes that penetrated to a depth >80 mbsf. This unit is interpreted to be correlative to the top of the regionally developed transparent seismic layer (Davis and Becker, 1994b). Despite the fact that there is 40 m of relief between Holes 1038B and 1038I, the top of this sand is present within 6.1 m of the same depth in all the holes, which implies that the topographic expression of Central Hill is predominantly caused by slumping along normal faults and that significant sections of the sedimentary cover have not been removed.
A basaltic layer from 1 to at least 5 m thick was intersected in three holes (Holes 1038G, 1038H, and 1038I) at depths between 142 and 162 mbsf. Adjusted for elevation of the top of the hole, this unit is at a near-constant elevation of 3365-3385 m and appears to slope from west to east. A flat-lying seismic reflector interpreted as either a basaltic flow or sill was mapped at the same depth to the east of Central Hill across the innermost rift bounding fault (Zierenberg et al., 1994). This unit could be either a thin flow erupted on sediment or a thin sill intruded near the top of the sand-rich horizon that defines the top of Unit C in Hole 1037B. Recovered fragments of basalt contained quenched margins, but contact relationships with the sediment were disturbed by drilling.
Basalt was also intersected at the base of Hole 1038I (403 mbsf). A contact was recovered that had 2 mm of baked and bleached sediment underlain by 5 mm of fresh glass. Time constraints at the end of the leg permitted only 1.5 m of penetration into the basalt, and all of the recovered core appeared to be from one cooling unit. The depth to this basalt coincides with a seismic reflector that is directly under Central Hill and sits ~100 m above the poorly defined sediment basalt contact in the Central Hill area (Zierenberg et al., 1994). The evidence at hand is consistent with this unit being a basaltic sill intruded near the base of the sediment section. Intrusion of this sill may have been responsible for the localized uplift of Central Hill, in accordance with the models presented by Davis and Becker (1994b), Denlinger and Holmes (1994), and Zierenberg et al. (1994) to account for the topography and hydrothermal history of Escanaba Trough.
Pore fluids collected from Site 1038 show a wide range in chemical compositions and provide important insight into the history of the hydrothermal system that formed the massive sulfide deposits. Comparison of pore fluids with hydrothermal fluids sampled from the active vents in 1988 (Campbell et al., 1994) shows that a hydrothermal component is obvious in all the holes and dominates pore-fluid chemistry at shallow depths below hydrothermally active and inactive sulfide mounds. One of the most surprising results of drilling at Site 1038 was the discovery that pore fluid Cl concentration ranges from 300 to 800 mM, indicating the presence of hydrothermal fluid affected by phase separation. The relationship between the Na and Ca content with Cl fits on a mixing line, indicating a single source for the high- and low-Cl fluids. Because most of the holes are relatively shallow, the high-salinity component dominates pore fluids. Low-salinity pore fluid is particularly evident in sand-rich layers between 80 and 120 mbsf in Holes 1038A, 1038H, and 1038I and is dominant below 190 mbsf in Hole 1038I.
The thermal gradient in the upper part of Hole 1038I is ~2°C/m, but the deepest measurement was at 56.8 mbsf, where a temperature of 115.9°C was measured. The highest vent temperature measured at this site was 220°C (Campbell et al., 1994). Rapid lateral advection of 220°C hydrothermal fluid through the massive sand layer at a depth of ~110 mbsf could conceivably support a 2°C/m thermal gradient and could supply the active hydrothermal vents. However, the hydrothermal vents have salinities above that of seawater, as do pore fluids sampled beneath massive sulfide, whereas the sandy horizon seems to contain low-salinity pore fluids. High-salinity pore fluid is also present immediately below the basaltic sill(?) in Holes 1038H and 1038I (Gieskes et al., Chap. 1, this volume). If the 2°C/m linear temperature gradient measure in the shallow portion of Hole 1038I continued to depth, temperatures at depths greater than ~200 mbsf would be in the range needed for phase separation. Although we were not able to measure temperatures below 57 mbsf in Hole 1038I, it would appear very unlikely that the high thermal gradient continues much deeper. The deepest sediments recovered from Hole 1038I showed less evidence for enhanced lithification and thermal alteration than rocks collected at in situ temperatures of 270°C in Middle Valley. Evidence for high downhole temperatures for holes drilled in Middle Valley, such as melting of core liners or seals, were not encountered at Escanaba Trough.
Even if the present thermal regime is not hot enough to promote phase separation of seawater-derived hydrothermal fluids, the presence of both high- and low-salinity components in the pore waters is an indication that the temperature pulse that caused a phase separation must have been recent relative to the time scale of diffusion of fluids through sediment (a few tens of years?). High temperatures are also indicated by the low concentration of Mg below 100 mbsf and by concentrations of Li and B below 300 mbsf that are far higher than those recorded in the reference hole (James and Palmer, 1999).
The organic geochemical data collected on board ship also support the notion of an intense, but brief, heating of the sediments. High concentrations of thermogenic methane were found in Holes 1038E, 1038H, 1038F, 1038G, and at depths greater than 40 mbsf in Hole 1038I. The presence of benzene and toluene confirm high-temperature cracking of organic matter. Bitumen fluorescence indicates a high maturation temperature of 150° to 250°C. The estimated temperature of organic maturation is higher than in Middle Valley. The regional maturation of organic matter has been rapid for all holes. In Hole 1038I, organic matter maturation occurred in situ without migration and is complete at a depth of ~160 mbsf, indicating high heat flow, but only limited flow of hydrothermal fluid through the sediment (Fouquet, Zierenberg, Miller, et al., 1998).
