Sites 1026 and 1027 are separated by 2.2 km along a line perpendicular to the strike of basement structure (Fig. 3), about 6 km from the closest and smallest of the basement outcrops. Site 1026 is located over the crest of a buried linear basement ridge where the sediment thins to 230-250 m. Site 1027 was drilled into an adjacent buried valley, where the sediment thickness reaches more than 600 m.
The simplicity of the local structure and the small size of the volcanic outcrops on this part of the east flank of the Juan de Fuca Ridge make the area an ideal target for seafloor studies. Results of a series of systematic predrilling surveys established several aspects of the fluid flow and discharge system:
1. Fluid circulation within the upper oceanic crust is sufficiently vigorous to maintain relatively uniform temperatures at the sediment/basement interface, despite variations of sediment thickness of more than a factor of 5.
2. In the few locations where samples could be obtained, basement-fluid compositions also appear locally homogeneous.
3. Fluids leak through the sediment "seal" above buried basement ridges at geochemically detectable rates (<1 mm/yr to tens of millimeters per year) that are inversely proportional to the local sediment thickness.
4. Fluids flow through the basement outcrops at rates sufficient to produce warm springs at the seafloor and to generate detectable thermal, chemical, and light transmissivity anomalies in the water column. The total advective heat loss is inferred to be great enough to cause the heat flow in the vicinity of the outcrops to be depressed substantially below the level expected from this 3.5-Ma lithosphere.
5. Discharge through the outcrop has been sufficiently long-lived to allow significant hydrothermal precipitates to accumulate. Cores recovered from the outcrop have been composed of green clay, semilithified black ferromanganese crusts and layers, and reddish iron oxides in indurated sediment.
6. The thermal structures of all three basement outcrops are fundamentally the same, allowing us to conclude that upflow and discharge not only are stable and long-lived, but also are a general consequence of the permeability and temperature structure of what we have referred to as "Permeable Penetrators" in the Scientific Prospectus.
Although the complete burial of the ridges and nearly complete burial of the outcrops in this area make them exaggerated examples, they are representative of a general class of circulation that is present wherever the crustal topographic relief of mid-ocean-ridge flanks is partially or completely filled in by sedimentation. An example of an "early" phase of this ridge type is the region of North Pond on the Mid-Atlantic Ridge flank. A relic, fully buried example was probably intersected at Site 417 on the Mid-Atlantic Ridge flank. A highly focused and very active example was investigated during Leg 139 in the Middle Valley rift.
Examples of topographically influenced circulation within the upper igneous crust occur on the Costa Rica Rift flank (holes in the vicinity of and including Hole 504B) and possibly on the northern flank of the Galapagos spreading center. Again, the high degree of burial makes the Juan de Fuca Ridge flank an exaggerated example of topographically influenced fluid flow, but one that is ideally suited to the study conducted during Leg 168.
Goals of the Rough Basement Transect
Primary objectives at this pair of sites included documenting precisely the lateral gradients in basement-fluid temperature, pressure, and composition to constrain the rates of fluid flow in the uppermost igneous crust; estimating the history of fluid flow in basement and through the sediment section; and determining the degree of hydrothermal alteration of the oceanic crust in this environment. Specific questions include the following:
1.To what degree are basement fluids thermally and chemically homogenized by circulation in this environment?
2.What implications can be drawn about the bulk hydrologic transport properties of the upper crust from constraints gained from observations of lateral temperature, pressure, and compositional gradients?
3.How is permeability distributed within upper basement?
4.To what degree has hydrothermal alteration of the crust proceeded at this 3.5-Ma site, and how has the alteration proceeded with time and burial history?
5.How have physical properties, namely seismic velocity and permeability, changed with alteration?
6.What are the nature and magnitude of the forces that drive fluid flow through the sediment section above basement ridges and through Permeable Penetrators where basement is exposed at the seafloor?
7.What is the source of the fluids that vent through the seafloor at the outcrops? Do they simply come from the inferred homogeneous basement "reservoir" regionally sealed beneath the sediments or is there a component from a deeper source?
8.How are sediments chemically and physically affected by fluid seepage? Can a declining rate of seepage through the sediment section be tracked through the history of burial?
9.What is the nature of recharge? Is seawater supplied to the crust solely by regional diffuse flow through the sediments away from basement highs or do some fluids enter the crust via locally focused pathways through basement?
Scientific Results
Shallow sediment cores recovered from both sites contain interbedded sand turbidites, silt turbidites, clayey silt, and debris-flow deposits with clasts of mud in a matrix of muddy sand. This mixture of sediments was designated lithologic Unit I at Site 1026 (Fig. 8). Interlayered sediments of the same types were designated subunit IA at Site 1027, where they extend to 184 mbsf. Subunit IB at Site 1027 comprises interbeds of silt turbidites within clayey silt and silty clay, extending from 184 to 467 mbsf. The sandy intervals in Unit I were found to be thicker at Site 1027 than at Site 1026, both in recovered cores and as inferred for a large depth interval that yielded little or no recovery in Hole 1027B. This contrast in thickness is consistent with the local seismic structure, which shows Site 1027 located approximately in the center of a broad distributary channel and Site 1026 near the edge where the channel fill onlaps a levee of a neighboring channel system. Despite the poor recovery in sandy intervals, hole conditions in Hole 1027B remained surprisingly good down to basement. Recovery improved substantially below 260 mbsf and remained generally high until basement was reached. Lithologic Unit II was hemipelagic mudstone with variable calcium carbonate content. This unit extends from 467 to 569 mbsf in Hole 1027B. The base of Unit II coincides with the first appearance of basalt within thin interbeds of pelagic and hemipelagic mudstone. Sediment alteration was not discernable with shipboard determinations of bulk mineralogy except in the sections a few meters above basement. The most obvious signs of alteration occur in carbonate-rich sediment in immediate contact with basement. In Hole 1027C, the basal sediments displayed vivid color variations imparted by Fe-Mn oxides.
Species diversity of nannofossils at Sites 1026 and 1027 is low, although a few key species provided some age control. E. huxleyi (0.28 Ma) was identified throughout the 101-m interval drilled in Hole 1026A; the base of this zone was found at about 98 mbsf in Hole 1027B. Other age controls in Hole 1027B include the top of P. lacunosa (~0.46 Ma) at 197 mbsf, the top of Calcidiscus macintyrei (~1.58 Ma) at 282 mbsf, and the base of Gephyrocapsa lumina (~1.68 Ma) at 417 mbsf. The dominance of Reticulofeuesta minuta and R. minutula in the assemblage below 417 mbsf, as well as the absence of R. pseudoumbilicus, suggests an age of late Pliocene for these deeper sediments including those interbedded with the basalt fragments at the bottom of the hole. These ages are consistent with sea-surface magnetic anomalies that constrain the age of basement at 3.47 and 3.55 Ma at Sites 1026 and 1027, respectively. All age constraints are consistent with an increasing sedimentation rate through the history of basement burial.
Hard rock was first drilled in Hole 1026B at 247.1 mbsf (Fig. 9). The first cored basalt recovered from Hole 1026B was from 256.0 mbsf, with basement coring continuing to a depth of 295.2 mbsf for a total of 39.2 m. Total recovery was 5.0%. Coring of basaltic rocks in Hole 1026C commenced where drilling first encountered basement at 228.9 mbsf. Two cores were cut to a total depth of 248.2 mbsf for a total of 19.3 m; total recovery was only 3.5%.
At Site 1027, basalt was recovered during XCB coring where drilling first encountered basement at a depth of 568.3 mbsf in Hole 1027B. After 9.6 m of basement was cored at Hole 1027B, Hole 1027C was begun with the objective of casing into basement and then coring farther below the cemented casing shoe for a CORK installation. In Hole 1027C, diabase was first recovered from 584.8 mbsf, or 9.3 m below the first hard rock felt by the drillers at 575.5 mbsf, and coring continued an additional 47.6 m. Core 168-1027C-1R consisted of diabase, whereas, Core 168-1027C-2R and most of Core 168-1027C-3R contained terrigenous clay and pelagic sediments. The last 28 cm of Core 168-1027C-3R and all of Cores 168-1027C-4R and 168-1027C-5R contained fractured pillow basalts. Total recovery was 10% in the XCB core from Hole 1027B, 73% in a massive diabase unit from Core 168-1027C-1R, 36% and 73% in fractured pillow basalts from Cores 168 1027C-4R and 168-1027C-5R, respectively.
The basaltic rock types recovered at the two Rough Basement sites vary significantly. The rocks from Hole 1026B are divided into three units. Unit 1 comprises slightly to moderately altered aphyric plagioclase-pyroxene basalt. Unit 2 is a highly altered basalt-hyaloclastite breccia, and Unit 3 consists of a slightly to moderately altered aphyric basalt and moderately phyric pyroxene plagioclase-olivine basalt. All three units exhibit sparse, thin glass margins, with all basalts containing sparse (<1%), small (<1 mm diameter) vesicles. The rocks may represent a complex of pillow basalts and interpillow basalt-hyaloclastite breccia. Hole 1026C consists of pillow basalts, designated Unit 1; these are also sparsely phyric, slightly to moderately altered basalts with several occurrences of glassy margins.
At the deeper basement site (Site 1027), four distinct units were identified. In Hole 1027B, Unit 1 is a massive aphyric olivine-plagioclase-pyroxene basalt with slight to moderate alteration. Unit 2, separated from Unit 1 by at least 15 cm of hemipelagic or pelagic carbonate bearing mud, is a basaltic breccia containing fragments of basalt similar to that of Unit 1. The first rock cored in Hole 1027C is Unit 3, which is a fine-grained plagioclase-olivine-pyroxene diabase with an aphanitic lower chilled margin; the upper margin was not sampled. Alteration varies from slight to moderate between the diabasic core and the aphanitic margin. Below Unit 3 is at least 10.6 m (23.0 m according to the driller's record) of terrigenous clay and bedded, altered and deformed carbonate-rich pelagic sediments. Consequently, Unit 3 is interpreted as a sill that was emplaced off-axis. The final 19 m of coring in Hole 1027C recovered portions of a pillow basalt sequence, recognized by abundant glassy chilled margins (every 0.6-1.2 m, on average). These pillow basalts are slightly to moderately altered, aphyric to moderately phyric plagioclase-olivine and plagioclase-olivine-pyroxene basalts. All four igneous units from Site 1027 contain sparse (<1%), small (<1 mm diameter) vesicles.
Major-element X-ray fluorescence analyses show all the rocks to be low-potassium tholeiitic basalts with variable Mg#. The diabase of Hole 1027C and the massive basalts of Hole 1027B, however, differ significantly from the pillow basalts from Holes 1027C, 1026B, and 1026C. The pillow basalt geochemistry defines well-constrained linear variation diagrams, suggesting that these basalts may be related through simple low-pressure fractionation of olivine, plagioclase, and/or clinopyroxene. In contrast, the diabase and Hole 1027B massive basalts have chemical compositions very similar to each other, which lie off the fractionation trends defined by the pillow basalts. This suggests that the two distinct chemical groups represent different magmatic events. The pillow basalts were probably erupted in an axial or near-axial environment, whereas the diabase and massive basalts must have erupted onto or intruded into the sediment section at a later stage at an off-axis location.
Alteration effects in basalt from Sites 1026 and 1027 vary significantly between igneous units, although certain features are ubiquitous. Nearly all samples contain mineral linings or complete mineral fillings of vesicles. Typically, the vesicle filling involves two to four sequential layers. The observed distribution of the various vesicle fills is systematic and is linked to the proximity of the vesicles to veins and alteration halos. In addition, we observed veins and mineral coatings on fractured rock surfaces. A third mode of alteration is the replacement by clays of magmatic phases including phenocrysts, groundmass crystals, cryptocrystalline mesostasis, and glass.
In summary, petrological studies at Sites 1026 and 1027 suggest the presence of two distinct magmatic series with distinct alteration histories. Pillow basalts from Site 1026 and from the deepest cores at Site 1027 represent petrogenetically related mid-ocean-ridge basalts with abundant early oxidative alteration, which is probably the result of extended exposure of the rocks at the seafloor. In contrast, the diabase and the massive basalts from Hole 1027B and the first core of Hole 1027C are petrogenetically distinct from the pillow basalts and represent the products of off-axis magmatism; these rocks did not experience significant oxidative alteration by seawater. Finally, an overprint of calcium carbonate alteration affects all of the rocks at Site 1027, whereas calcium carbonate is only a minor alteration product at Site 1026. The reason for this difference may relate to the location of Site 1026, drilled into a buried ridge, and Site 1027, in a buried valley, and consequent differences in the hydrothermal regime.
Profiles of inorganic pore-water composition at the two sites follow similar trends and reflect reaction with sediments, flow of incompletely altered waters in basaltic basement, water/rock reactions in basement, and temporal variations in the composition of bottom water from glacial and interglacial periods. Although the chemical data do not indicate significant flow of pore fluids vertically through the sediments, when Hole 1026B was reentered immediately before deploying the CORK in that hole, temperatures within the borehole indicated that basement water was flowing from the formation into the hole, and then up through the casing and into the overlying ocean. Basement fluid pressure at this location must therefore be locally superhydrostatic, but the low permeability of the sediment section must limit natural fluid seepage to a rate below that geochemically detectable. The water sampler temperature probe was run into the hole to collect a sample of the discharging basement fluid and to log temperatures in the borehole so the fluid velocity could be estimated. This was the first successful sampling of true basement fluid in the history of the Deep Sea Drilling Project (DSDP) and ODP. The sample was sufficiently large to allow post-cruise carbon-14 analyses to determine the age or a limiting age of the hydrothermal fluid. To a first order, this basement fluid has the same composition as the fluid exiting from the nearest basement outcrop that was sampled by submersible. It differs in several notable respects from the composition suggested by interstitial-water samples collected from sediments just above basement (Fig. 10). Methane concentrations increased from near zero near the seafloor and basement to values on the order of 104 ppm by volume in the middle of the sediment section at Sites 1026 and 1027. Ethane also was present at low concentrations in headspace samples. Pentane and minor light alkene hydrocarbon gases (<1 ppm) were present at 360 mbsf at Site 1027. The methane/ethane ratio decreased from 26,000 at 200 mbsf to 900 at the base of the sediment section.
Physical properties measurements within the sediment section revealed the expected effects of consolidation and lithification with depth, with local variations associated with changes in sand, silt, and clay contents. P-wave velocities measured in the laboratory at atmospheric pressure were near 1.5 km/s at the seafloor and increased to 1.7 to 1.9 km/s near basement. Velocities of the mudstone in lithologic Unit II were found to be anisotropic, with horizontal velocities higher than vertical. MST magnetic susceptibility and gamma-ray attenuation porosity evaluator (GRAPE) data allowed rapid delineation of fining-upward turbidite sequences, as well as relatively homogeneous sandy intervals (where these were recovered) before the core was split, facilitating the selection of whole-round physical properties samples collected for shore-based studies.
As at the previous sites, temperatures were measured in the sediment section with the goal of obtaining accurate determinations of temperature of the sediment/basement interface, which is inferred to be a primary hydrologic contact between the low-permeability sediment cover and high-permeability extrusive igneous rocks beneath. At Site 1026, the temperature at the contact estimated by extrapolation of advance hydraulic piston corer (APC) and Davis/Villinger Temperature Probe (DVTP) measurements is approximately 61°-62°C, only slightly lower than at the extrusive contact at Site 1027, about 63°C (Fig. 11). These temperatures represent only lower limits for those of "hydrologic basement," because the first zone of high permeability at any given location will probably lie deeper than the sediment/basement contact. This is illustrated well by the lack of agreement between the sediment/basement contact temperature estimated at Site 1026 (~61°-62°C) and the measured temperature of discharging basement water (~64°C). Uncertainties of this magnitude preclude resolving whether hydrothermal temperatures are systematically warmer at the ridge or the valley. Resolution of this questionshould be possible following the examination of data from the long-term CORK installations at these sites. At this time it can be concluded only that upper basement temperatures are uniform to within a few degrees.
Successful packer experiments were conducted in both Holes 1026B and 1027C. The packer was first set in casing at the base of Hole 1026B. Basement pressures recorded while the packer element was inflated (isolating the open hole) indicated a slight underpressure after correction for the cold hydrostatic water column in the borehole, although later flow from the formation into the borehole indicated the formation is naturally overpressured. The retention of a basement underpressure at the end of the Hole 1026B packer tests suggests that the CORK experiment should allow accurate determination of ambient basement fluid pressures. The two slug tests are consistent with relatively high basement permeability, although interpretation of packer tests in Hole 1026B will be complicated by the nonideal hole geometry, because much of the hole below the casing shoe was filled with basalt rubble.
In Hole 1027C, the packer was set first in the casing above open hole. Three slug tests and two injection tests indicate relatively high permeability. The packer was also set outside the casing in Hole 1027C, within the massive basalt flow unit recovered in Core 168-1027C-1R. Two additional slug tests and two injection tests also indicate high permeability. There also appeared to be an underpressure in Hole 1027C, although the magnitude of this underpressure was somewhat lower than that in Hole 1026B. An accurate determination of the natural formation pressure will be made during the long-term CORK experiment.
CORK observatories were set in Holes 1026B and 1027C (Fig. 7). Each observatory included a data logger, pressure gauges above and below the borehole seal, a thermistor string with 10 sensors, and an osmotic sampler above a sinker bar at the bottom of the thermistor cable. Poor drilling conditions in Hole 1026B led to the emplacement of a "liner," a piece of old drill pipe attached to a modified mechanical bit release, that was drilled into the formation and left to hold open the hole. Drilling conditions were considerably better in Hole 1027C, and no liner was required to maintain the hole. Once open holes were established, both CORK systems were emplaced with little difficulty. The first data will be downloaded in about 1 yr during an expedition with the remotely operated vehicle Jason.