Geologic processes at convergent plate margins control geochemical cycling, seismicity, and deep biosphere activity within subduction zones. The study of input into a convergent plate margin by sampling the downgoing plate provides the geochemical reference necessary to learn what factors influence the production of suprasubduction zone crust and mantle in these environments. The study of the output in terms of magma and volatiles in volcanic arcs and backarc basin settings constrains processes at work deep in the subduction zone, but these studies are incomplete without an understanding of the throughput, the nature of geochemical cycling that takes place between the time the subducting plate enters the trench and the time it reaches the zone of magma genesis beneath the arc. Tectonically induced circulation of fluids at convergent margins is a critical element in the understanding of chemical transport and cycling within convergent plate margins and, ultimately, in understanding global mass balance (e.g., COSOD II, 1987; Langseth et al., 1988; Kulm and Suess, 1990; Langseth and Moore, 1990; Martin et al., 1991). In the shallow to intermediate suprasubduction zone region, dehydration reactions release pore fluids from bound volatiles in oceanic sediments and basalts of the downgoing plate (Fryer and Fryer, 1987; Peacock, 1987, 1990; Mottl, 1992; Liu et al., 1996). Fluid production and transport affect the thermal regime of the convergent margin, metamorphism in the suprasubduction zone region, diagenesis in forearc sediments, biological activity in the region, and, ultimately, the composition of arc and backarc magmas. Furthermore, these fluids, their metamorphic effects, and the temperature and pressure conditions in the contact region between the plates (the décollement) affect the physical properties of the subduction zone, where most major earthquakes occur.
The discovery of Earth's deep biosphere is recognized as one of the most outstanding breakthroughs in the biological sciences. The extent of this biosphere is currently unknown, but we are becoming increasingly aware that life has persisted in environments ranging from active hydrothermal systems on mid-ocean ridges to deep ocean sediments, but so far no detailed investigations have been made of the potential for interaction of the deep biosphere with processes active in convergent plate margins.
Determining unequivocally the composition of slab-derived fluids and their influences over the physical properties of the subduction zone, biological activity, or geochemical cycling in convergent margins requires direct sampling of the décollement region. To date, studies of décollement materials, mass fluxes, and geochemical interchanges have been based almost exclusively on data from drill cores taken in accretionary convergent margins (e.g., Kastner et al., 1993; Carson and Westbrook, 1995; Maltman et al., 1997). Large wedges of accreted sediment bury the underlying crystalline basement, making it inaccessible to drilling, and the wedges interact with slab-derived fluids, altering the original slab signal. The dehydration reactions and metamorphic interchanges in intermediate and deeper parts of the décollements have not been studied in these margins. By contrast, nonaccretionary convergent margins permit direct access to the crystalline basement and produce a more pristine slab-fluid signature for two reasons: the fluids do not suffer interaction with a thick accretionary sediment wedge, and they pass through fault zones that have already experienced water-rock interactions, thus minimizing interaction with subsequently escaping fluids. Regardless of the type of margin studied, the deeper décollement region is directly inaccessible with current or even foreseeable ocean drilling technologies. A locality is needed where some natural process brings materials from great depths directly to the surface. The Mariana convergent margin provides precisely the sort of environment needed.
The Mariana convergent plate margin system is nonaccretionary, and the forearc between the trench and the arc is pervasively faulted. It contains numerous large (30 km diameter and 2 km high) mud volcanoes (Fryer and Fryer, 1987; Fryer, 1992, 1996) (Fig. F2). The mud volcanoes are composed principally of unconsolidated flows of serpentine mud with clasts consisting predominantly of serpentinized mantle peridotite (Fryer, Pearce, Stokking, et al., 1990). Some have also brought up blueschist materials (Maekawa et al., 1995; Fryer and Todd, 1999). Faulting of the forearc to great depth produces fault gouge that when mixed with slab-derived fluids generates a thick gravitationally unstable slurry of mud and rock that rises in conduits along the fault plane to the seafloor (Fig. F3) (Fryer, 1992, 1996). These mud volcanoes are our most direct route to the décollement and, episodically, through protrusion events, open a window that provides a view of processes and conditions at depths as deep as 35 km beneath the forearc.
Prior to Leg 195, only one other active serpentine mud volcano (Conical Seamount) (see Fig. F2) had ever been sampled by drilling (Fryer, Pearce, Stokking, et al., 1990). Little was then known of either the processes that formed such seamounts, their distribution, and their relation to the tectonics of the forearc region or of the potential for understanding the deeper forearc processes they reflect. Recent advances in the understanding of the structure and tectonic evolution of nonaccretionary forearcs, the nature of geochemical cycling within them, and the various active thermal, hydrologic, metamorphic, and biological processes involved in the formation of mud volcanoes permitted the planning of comprehensive studies of the intermediate-depth processes occurring within the "subduction factory." By revisiting descriptions of serpentine melanges and "sedimentary" serpentinite terranes from past literature (Lockwood, 1972), we now realize that serpentine mud volcanism in convergent margin settings is not merely a local curiosity of the Mariana system but occurs worldwide.
Site 1200 is located on a 200-m-high summit knoll on South Chamorro Seamount at 13°47´N, 146°00´E in a water depth of 2910 m, ~125 km east of Guam in the western Pacific Ocean. It lies 85 km from the trench, where the depth to the downgoing slab is ~26.5 km, based on studies by Isacks and Barazangi (1977). Pore fluids collected in gravity cores from this seamount exhibit a strong slab signal. It is the only known site of active blueschist mud volcanism in the world and supports the only documented megafaunal assemblages associated with serpentine/blueschist mud volcanism (Fryer and Mottl, 1997).
Side-scan surveys of this seamount (Fig. F4) show that the southeastern sector of the edifice has collapsed, and debris flows of serpentine material (dredged in 1981 and observed during Shinkai 6500 dives in 1995) blanket the inner slope of the trench from the summit of the seamount to the trench axis. The summit knoll sits at the apex of the sector collapse, and its formation was most likely initiated in response to the collapse. Submersible observations show that the knoll's surface is broken into uplifted slabs of cohesive serpentine mud (Fryer, 1996) separated by meter-deep fissures with crosscutting orientations. Medium blue-green to dark blue serpentine mud and clasts of metamorphosed rocks are exposed. Low-temperature springs in the fissures support a vigorous biological community of mussels, gastropods, worm tubes, and galatheid crabs (Fryer and Mottl, 1997). The mussels are likely of the genus Bathymodiolus, one which contains methylotrophic symbionts in its gills and requires high ambient concentrations of methane in its feeding source (Fryer and Mottl, 1997). The pore fluid composition profiles and the presence of a vigorous biological community at the surface suggest that the summit knoll is a currently active seep region. The interior of the seamount shows little structure on six-channel seismic reflection profiles (Fryer and Mottl, 1997) (Fig. F5). This seamount is likely an active serpentine mud volcano similar to Conical Seamount, drilled during Leg 125 (Sites 778-780), and thus provides an excellent drill target for studies of the active processes of these mud volcanoes. It has the strongest slab signature in pore fluids from among the seamounts sampled in 1997 and is comparable to Conical Seamount in the strength of its slab signal.
The overall objectives of drilling at South Chamorro Seamount were to (1) study geochemical cycling and mass transport in the subduction zones and forearcs of nonaccretionary convergent margins; (2) determine the spatial variability of slab-related fluids within the forearc environment as a means of tracing dehydration, decarbonation, and water-rock reactions both in the subduction zone and the overlying suprasubduction zone environments; (3) study the metamorphic and tectonic history of nonaccretionary forearc regions; (4) investigate the physical properties of the subduction zone and their influence on dehydration reactions and seismicity; and (5) investigate biological activity associated with subduction zone material from great depth.
To achieve these scientific objectives, operations during Leg 195 were designed to recover sufficient materials to permit petrologic and mineralogic characterization of the serpentine mud flow units, to analyze their pore fluid compositions, to collect any biological material contained in the muds, and to deploy a long-term geochemical observatory at South Chamorro Seamount.
The primary objective at South Chamorro Seamount was to deploy a long-term geochemical observatory in a cased reentry hole in the central conduit of the serpentine mud volcano. The reentry hole for the installation of a downhole thermistor string, pressure sensor, and osmotic fluid samplers was designed to be sealed with a circulation obviation retrofit kit (CORK). The techniques that were used to install the CORK were similar to those used for successful installations during Legs 139, 164, 168, and 174B (Davis et al., 1992). The hole at Site 1200 was CORKed with a thermistor cable to obtain a long-term record of the temperature variations in the sealed hole as the natural hydrologic system reestablishes itself after drilling. This installation will provide a long-term record of (1) the rebound of temperatures toward formation conditions after the emplacement of the seal; (2) possible temporal variations in temperature and pressure due to lateral flow in discrete zones, regional and/or local seismicity, and short-term pressure effects; and (3) the composition of deep circulating fluids obtained with the osmotic samplers. Data from the downhole instruments will be collected during an NSF-funded Jason/DSL 120 cruise that is tentatively scheduled to be conducted 18 months after Leg 195.
The drill site on South Chamorro Seamount was designed to help assess the variability of fluid transport and composition within the forearc. Previous field studies indicate that most of the fluid flow in the Mariana forearc is channeled along forearc faults and fault-controlled conduits in mud volcanoes. The pore fluid compositions are expected to vary depending on the nature of the channeling structures (diffuse network of small faults, major faults, and mud volcano conduits). In particular, fluids ascending through mud volcano conduits along well-established paths in contact with previously metamorphosed wall rock should carry the most pristine slab signature. This was certainly the case at Conical Seamount, drilled during Leg 125. The summit Site 780 produced by far the purest deep slab-derived fluids, based on their much lower chlorinity and higher K, Rb, B, H2S, and sulfate, whereas the flank Sites 778 and 779 produced combinations of slab-derived fluid and seawater that had reacted with peridotite and basalt at shallower crustal levels (Mottl, 1992).
Although total fluid budgets are difficult to ascertain in any convergent margin, they are likely to be more readily determined at nonaccretionary active margins because the hydrologic flow systems operate on longer timescales than do those at accretionary margins. Attempts to determine the total fluid budgets at accretionary active margins have been hindered by the presence of lateral heterogeneity and transient flow processes. Lateral heterogeneity results in different flow rates and compositions along the strike of the margin. Transient flow apparently results largely from the valvelike influence of the accretionary complexes themselves.
Sediment properties vary with fluid pressure, and fluid pressure varies as a function of fluid production rate and transient hydrologic properties. Thus, the accretionary system acts as both a seal and a relief valve on the fluid flow system. The absence of such a short timescale, fluid pressure, and formation-properties modulator at nonaccretionary systems should allow fluids to escape more steadily. To test this hypothesis, the physical nature of fluid flow at nonaccretionary settings must be determined. Then fluid budgets can be constructed to determine whether the expected long-term flow is consistent with observations or if the flow must occur in transient pulses. The CORK experiment planned for the South Chamorro Seamount site will address this problem.
The composition of slab-derived fluids and deep-derived rock materials may differ along the strike of the forearc, reflecting regional variations in composition within the slab and suprasubduction zone lithosphere. The pore fluids from several of the forearc mud volcanoes already sampled are chemically distinct, and it was anticipated that the pore waters from South Chamorro Seamount would also be chemically distinct. These differences are probably associated not only with the depth to the slab but also with the physical conditions under which water-rock reactions occur and the variations in the regional composition of the plate and overriding forearc wedge.
The geochemistry of the fluids from Conical Seamount is described in detail in several publications (Fryer et al., 1990; Haggerty, 1991; Haggerty and Chaudhuri, 1992; Haggerty and Fisher, 1992; Mottl, 1992; Mottl and Alt, 1992). These papers show the origin of the Conical Seamount fluids to be from dehydration of oceanic crustal basalt and sediment at the top of the subducting lithospheric slab. The compositions of the fluids from the PACMANUS hydrothermal field and seamounts farther south are reported in Fryer et al. (1999). Pore fluids from these indicate a slab source, as shown by their lower chlorinity and higher K and Rb, similar to that observed at Conical Seamount by Mottl (1992).
The composition of slab-derived and deep-derived metamorphosed rock is useful in defining geochemical processes and estimates of the thermal and pressure regime at depth, and thus, for determining the physical properties of the décollement region. It was hoped that it would also be possible to constrain some of the pressure and temperature conditions under which certain dehydration reactions take place in the subducted slab. Pore fluids from Ocean Drilling Program (ODP) Site 780 at the summit of Conical Seamount are unusual because of geochemical and physical processes at depth. The observed enrichments in alkali elements and B in fluids from Site 780 are unambiguous indicators of a source temperature in excess of 150°C, yet the fact that these elements are depleted at Sites 778 and 779 on the flanks of Conical Seamount, relative to their concentrations in seawater, indicates that the deep slab signal can be readily overprinted by local peridotite-seawater reactions at lower temperatures. Not all chemical species are affected by this overprinting, however (i.e., sulfur isotopic composition of dissolved sulfate) (Mottl and Alt, 1992). Thus, to avoid potential reactions between sediment and slab-derived fluids, we planned to collect fluids from mud volcano conduits, where continued focused flow provides a pathway for slab-derived "basement" fluids to reach the seafloor.
Studies of deep-derived minerals and metamorphic rock fragments brought to the surface in mud flows in serpentine seamounts can be used to constrain the pressure and temperature regimes under which the metamorphism that formed them took place. It is known, for instance, that the minimum pressures of formation for incipient blueschist materials from Conical Seamount are 6-7 kbar (Maekawa et al., 1995). Similarly, from the paragenesis of crossite schist recovered in cores from South Chamorro Seamount, it can be shown that pressures >7 kbar are consistent with their metamorphism. Examination of a more extensive collection of the muds and clasts from South Chamorro Seamount should make it possible to quantify the assemblages of muds and xenoliths present in the flows and constrain the ranges of pressure and temperature that exist in the source regions for these materials.
Interest in the deep subsurface biosphere has grown dramatically as a result of recent studies linking extreme environments to the first living organisms that inhabited the Earth. The search for the last common ancestor in the geologic record is moving toward high-temperature environments, such as those at spreading centers and hotspots both on the ocean floor and on land. Microbes and microbial products are abundant in oceanic hydrothermal environments and are presumed to be representative of a community of thermophilic and hyperthermophilic organisms that originated beneath the seafloor (Fisk et al., 1998). Microbes are also involved in the transformation of minerals in the oceanic crust and in the cycling of elements in the crust; however, the origin of these microbes is much more controversial.
Drilling at Chamorro Seamount provides a unique opportunity to determine the nature of microbiological activity in a very different kind of extreme environment, the high-pH, low-temperature environment associated with serpentine/blueschist mud volcanism (Fryer and Mottl, 1997; Fryer et al., 1999), and to reexamine the hypothesis that microbes are capable of using alternative energy sources that would support a heterotrophic subsurface ecosystem. In addition, because the pore fluids are more pristine in nonaccretionary convergent margins, it should be easier to assess from the chemistry of both the muds and the fluids whether organic syntheses capable of supporting life are active in these settings.
Understanding the origin of the deep biosphere is a fundamental ODP objective and will further address the compelling question of whether life arose in extreme environments rather than on the surface of the early Earth. Although several experimental studies indicate that a thermophilic origin of life is possible, definitive proof must await an assessment of the full range of conditions in which life exists and the nature of life in these environments.
The mechanics and rheology of serpentine muds in the Mariana forearc seamounts control the processes that formed the seamounts and their morphology. It was thus planned to conduct a rheological study of the serpentine muds to place realistic constraints on the mechanisms governing the ascent of the muds to the surface, the maintenance of the conduits, and the construction of the seamounts.
Shipboard torsion-vane testing during Leg 125 at Conical Seamount in the Mariana forearc and at Torishima Forearc Seamount in the Bonin forearc showed that the serpentine muds are plastic solids with a rheology that bears many similarities to the idealized Cam clay model and is well described by critical-state soil mechanics (Phipps and Ballotti, 1992). These muds are thus orders of magnitude weaker than salt and are, in fact, comparable in strength to common deep-sea pelagic clays. The rate at which the muds rise relative to the fluids will likely influence the water-rock reactions and the character of the slab signal in fluids from these mud volcanoes. Better constraints on the nature of the fluids will permit a more accurate determination of the physical conditions of the décollement, where the fluids originate.
After steaming to the site and lowering the pipe to the summit of the seamount, we planned to conduct a brief seafloor television survey of the conduit to locate the springs and mussel beds identified in Shinkai 6500 dives and to identify sites near the springs for rotary core barrel (RCB) coring and logging and relatively clast-free sites for advanced piston corer/extended core barrel (APC/XCB) coring, jet-in tests, and the establishment of a reentry hole. After conducting a jet-in test to establish the depth of the first casing string for the reentry hole, we planned to core and log an RCB pilot hole to 450 meters below seafloor (mbsf) to determine the nature of the formation and the ease of drilling on the seamount, a major concern because drilling during Leg 125 at Conical Seamount had been plagued with drilling problems.
Depending on the results of the RCB hole, we then planned to offset and jet in a reentry cone and 20-in casing to ~25 mbsf and then drill a hole in stages to 420 mbsf for the CORK installation. In anticipation of hole instability problems, an elaborate casing program was envisioned, with cemented 16-in casing to 200 mbsf followed by 10.75-in casing to 400 mbsf, including 23 m of screened casing and a casing shoe at the bottom to prevent the serpentine mud from slowly invading the installation from below. After the hole was drilled and cased, the instrumentation installed, and the remotely operated vehicle (ROV) platform emplaced, we planned to drill an APC/XCB hole to 420 mbsf to collect a continuous, undisturbed section for petrologic and pore water studies. By drilling the APC/XCB hole last, the time allocated for APC/XCB coring could be held in reserve as contingency time if it took longer than anticipated to drill the reentry hole and install the observatory. Not surprisingly, the actual operations at Site 1200 unfolded rather differently.
The JOIDES Resolution arrived at Site 1200 (proposed Site MAF-4B) at 2100 hr on 11 March 2001. Following a 4-hr camera survey, Hole 1200A was spudded adjacent to a vent mussel community at the top of South Chamorro Seamount at 2200 hr on 11 March with the RCB and cored to a depth of 147.2 mbsf (Fig. F6; Table T1). High torque and lost rotation at this depth resulted in a stuck drill string that ultimately forced us to abandon the hole. Recovery in Hole 1200A was poor (147.2 m cored; 9.4% recovered), with little recovery of the mud matrix material surrounding the hard ultramafic clasts. The drill string was pulled out of the hole, and the seafloor was cleared at 1535 hr, ending Hole 1200A.
The ship was then offset 25 m to the east, and Hole 1200B was spudded at 1640 hr on 13 March. Our intention was to wash to 147.2 mbsf and then start coring to the target depth of 450 m. By 1615 hr on 14 March, the hole had been advanced to a depth of 98.0 mbsf. Once again, high torque and overpull began to plague the hole. Despite consistent mud sweeps and multiple reaming attempts, the hole could not be stabilized. At 0645 hr on 14 March, we decided to abandon further drilling/coring efforts in the pilot hole and to start the reentry hole because it was obvious that deep penetration could not be achieved without the use of casing.
Once the drill string was pulled out of the hole with the top drive, the ship was offset to the north and Hole 1200C was spudded at 0450 hr with the 20-in casing attached to the reentry cone. After a total of 18.25 hr of drilling with an 18.5-in bit, 22-in underreamer, and drilling motor, the reentry cone base reached the seabed, placing the 20-in casing shoe at 23.7 mbsf. Drilling of the 22-in hole for the 16-in intermediate casing string, without the motor this time, advanced smoothly and without incident. At 0130 hr on 18 March, the hole reached a depth of 140 mbsf. While we washed/lowered the casing string into the hole, however, the casing shoe encountered an obstruction 6 m off bottom that prevented the casing hanger from landing. After three futile hours, the drill string was recovered and one joint of casing was removed from the string. The shortened string was run back into the hole and washed down without incident. With the 16-in casing shoe placed at a depth of 107.4 mbsf, the casing string was cemented in place. The drilling process was then reinitiated using a 14.75-in drill bit and 20-in underreamer, dressed with 20-in cutter arms, and advanced to a depth of 266.0 mbsf. The penetration rate deteriorated to zero at that point, and a subsequent wiper trip found 9 m of soft fill at the bottom of the hole that was easily removed by circulation. Upon reaching the rig floor, the underreamer was missing two out of three cutters. Because the hole had penetrated below the base of the summit knoll but could not be deepened further, the 10.75-in casing was deployed, reaching a depth of 224.0 mbsf by 0300 hr on 24 March without incident. However, progress beyond this point was not possible, and the string was pulled out of the hole to remove three joints of casing. The shortened casing string finally placed the shoe at 202.8 mbsf, with the screened interval extending from 202.3 to 148.8 mbsf (Fig. F7).
The CORK assembly, with two stands of 5.5-in drill pipe used as the stinger, was run in the hole, and the drill string was lowered to 53.0 mbsf. This left the CORK shy of landing out by ~8 m. At 1300 hr on 25 March, the thermistor/osmotic sampler assembly was slowly run in the hole, and at 1430 hr the data logger was landed in the CORK body. After 1 hr of pressure transducer calibration, the data logger was latched and the CORK body was then lowered the final few meters and latched in place as in Figure F8. The CORK installations at Hole 1200C ended with a successful free-fall deployment of the ROV platform. The pipe was then returned to the surface; not counting the pilot holes, nine round pipe trips totaling 52 pipe-km of tripping pipe were required to deploy the observatory.
After the successful installation of the geochemical observatory, the vessel was offset 40 m to the south and Hole 1200D was spudded with the APC to sample the serpentine muds for petrology, pore water, and microbiology studies at 1415 hr on 26 March. Coring continued to a depth of 44.4 mbsf using the APC advance-by-recovery method. Hard clasts were drilled with an XCB center bit assembly. After coring was halted by a hard clast at 44.4 mbsf, the XCB was used to deepen the hole. After advancing 9.0 m, however, the penetration rate fell to zero and the decision was made to abandon Hole 1200D.
The vessel was then offset back to the location of the earlier identified mussel beds, and Hole 1200E was spudded with the APC for further pore water studies. Coring with the APC/XCB continued, again using the advance-by-recovery method for the APC cores, to a depth of 50.4 mbsf, where the scientific objectives of the hole were met.
The ship was then offset 20 m to the north, and Hole 1200F was spudded at 1315 hr. Coring proceeded until the time allocated for operations at Site 1200 ran out. The hole depth reached 16.3 mbsf, with the recovery of APC Cores 195-1200F-1H through 3H. At 0200 hr on 29 March 2001, the ship was under way to the Guam pilot station.
The principal objective at Site 1200 was to install a borehole geochemical observatory at the summit of South Chamorro Seamount to sample fluids from the décollement below the Mariana forearc. Although this objective was achieved, no data will be recovered from the observatory until it is revisited by an ROV in 2003. As was to be expected in such an exotic environment, however, the drilling and coring undertaken to install the observatory and document its setting produced unexpected and often spectacular results.
Perhaps the most fundamental achievement at Site 1200 was the documentation of the muds, xenoliths, and fluids rising to the surface from the mantle and the décollement zone through the central conduit of the mud volcano. The recovery in all cored holes consisted of poorly sorted, dark blue-gray to black serpentine mud breccia (Fig. F9) in which the muds are composed predominantly of silty clay-sized serpentine, and the clasts, which range up to a meter across, consist largely of serpentinized ultramafics.
Well-preserved and diversified subtropical assemblages of planktonic foraminifers and calcareous nannofossils were found in the top 0.1-0.3 m of the holes that were APC cored (Holes 1200D, 1200E, and 1200F), indicating that the surface of the summit is blanketed with a veneer of calcareous microfossil-bearing deposits. A few species of benthic foraminifers are also present in small quantities in all holes. Samples farthest away from the vent communities contain more abundant, diversified, and better-preserved microfossil fauna. The downcore sections in these three holes are virtually barren of microfossils, except for a peculiar interval with folded color bands between 11 and 13 mbsf in Hole 1200D, which is interpreted as a paleosurface that has been covered or folded into the mud by the flow of serpentine. This interval contains abundant and diversified calcareous microfossils comparable to the core tops. The major difference is that these fossils tend to be robust species overgrown by calcite on the original structures, whereas the more delicate species have been dissolved. All the fossils are late Quaternary in age.
With the exception of the two calcareous intervals noted above, which are light yellow-brown to pink due to their microfossil content, and a 10- to 20-m-thick zone of strongly reduced, black serpentine muds starting about a meter below the seafloor, the mud breccia in the conduit displays no stratigraphy, which is consistent with its mud volcano origin. It has been divided into two facies, however, on the basis of the abundance of clasts (<10% and 10%-30%); based on recovery, the average is ~7%. The smaller ultramafic clasts tend to be angular with planar external surfaces along early serpentine veins, whereas the larger clasts are subrounded to rounded, suggesting comminution by collisions during ascent.
The mineralogy and composition of the muds and clasts almost all attest to a deep origin along the décollement zone or the overlying mantle, regardless of the particle or clast size, from silty clay to boulder. With the exception of aragonite produced at the mudline, where rising pore fluids interact with seawater to produce carbonates and rare zeolites (analcime) found farther down in the section, X-ray diffraction (XRD) analysis shows that the muds are dominated by serpentine minerals formed by the hydration of ultramafics + accessory glaucophane, spinel, garnet, chlorite, and talc derived from the metamorphism of mafic rocks along the décollement. Optical examination on board ship suggests lizardite/antigorite > chrysotile > brucite.
The dual origin of the materials making up the mud breccia is even more clearly revealed in the grit fraction (0.1-1.5 cm). About 90% by volume of the grit fraction consists of partially to completely serpentinized ultramafic rocks, but 10% consists of metabasites, including glaucophane schist (Fig. F10), crossite/white-mica/chlorite schist, chlorite schist (Fig. F11), white-mica schist, and amphibolite schist containing blue-green to black amphibole and interstitial mica. These lithologies, especially the blueschists, are indicative of a high-pressure, low-temperature origin (Fryer et al., 1999; Fryer and Todd, 1999; Todd and Fryer, 1999), and we interpret them as metamorphosed basic rocks from the descending slab. About 1% of the serpentine mud consists of blue sodic amphibole. Analysis of similar grains in gravity cores from South Chamorro Seamount showed a crossitic composition (Fryer et al., 1999). The mineral grains are zoned with blue rims and lighter blue-green cores, implying relatively rapid ascent with the rising serpentine muds. The rationale for this interpretation is that if the grains had been in contact with rising fluids having the extreme compositions observed in the pore water analyses (see below) for geologically significant periods of time, they would have likely back-reacted and would show retrograde metamorphic effects.
Although retrograde reactions are generally sluggish, the primary reason for this is the lack of reactive fluids in a system that has previously experienced prograde regional metamorphism. Such metamorphism drives volatiles out of the rocks, resulting in a dry system, which is far less likely to undergo retrograde metamorphism despite changes in pressure and temperature. The presence of highly reactive fluids in intimate contact with the serpentine muds at Site 1200, however, would make the possibility of retrograde reactions far more likely. None of the materials studied previously by Fryer and colleagues (Fryer et al., 1999; Todd and Fryer, 1999) have ever shown any indication of retrograde effects. The mineral grains from Site 1200 also lack retrograde effects.
Although the grit fraction contains a rich and varied population of high P-T samples from the décollement, such samples appear to be absent from the large clasts, suggesting that the metabasites reaching the surface are preferentially smaller pieces. The abundance of phyllosilicate minerals in the schists may contribute to the comminution of these samples as they rise from the source region. Pressure release as the rock fragments rise may cause the phyllosilicates to expand and disintegrate, and the continual collisions and mechanical grinding within the rising muds may cause the more friable rocks to break up into small fragments.
As noted earlier, about 7% of the material recovered at Site 1200 consists of large clasts of partially to completely serpentinized ultramafic rocks from the mantle wedge under the Mariana forearc. Whereas serpentinization has been extensive in all samples (40%-100%; average = ~75%), there are sufficient relict minerals present in many of the samples to assess the original grain size (0.01-5.0 mm) and to determine that harzburgite is the dominant protolith, dunite is much less common, and lherzolite is rare. This is consistent with the whole-rock chemistry of the samples, which suggests that the ultramafics underwent 20%-25% melt extraction at some point during the formation of the arc (Fig. F12). The actual percentage of relict minerals is extremely variable, with olivine ranging from 0% to 40%, orthopyroxene (enstatite) from 0% to 35%, clinopyroxene from 0% to 5%, and chrome spinel from 0% to 3%, depending on the original mineralogy and the degree of serpentinization. In general, olivine and enstatite were the least stable minerals, with olivine altering readily to serpentine (lizardite) or brucite + magnetite and enstatite altering to serpentine ± tremolite/actinolite, whereas clinopyroxene and spinel were usually the most resistant to alteration. Olivine often developed mesh and hourglass textures during serpentinization, whereas enstatite is commonly replaced by bastitic textures. The serpentinization appears to have occurred in stages, because lizardite veins are often cut orthogonally by chrysotile veins, producing spectacular "Frankenstein veins" consistent with uniform dilation during late-stage serpentinization along grain boundaries. Interestingly, most of the ultramafics and all of the dunites show evidence of deformation prior to serpentinization: the olivines commonly show kink banding and granulation and the enstatites (or their bastite replacements) often display undulatory extinction.
The pore waters from Site 1200 revealed two distinct phenomena, a deep-sourced fluid that is believed to be upwelling from the top of the subducting slab 25-30 km below the seafloor and a new and exotic extremophile microbial community at 0-20 mbsf that is chemically manipulating its environment. As can be seen in Figure F13, most pore water vs. composition profiles for the site represent nearly ideal advection-diffusion curves and the gradients in the top few meters are so steep that they can only be maintained by upwelling from below. The deep fluid is similar in many ways to that sampled at Conical Seamount to the north during Leg 125 (Fryer, Pearce, Stokking, et al., 1990). It has a pH of 12.5 because it is in equilibrium with brucite, making it, along with the Conical Seamount fluids, the most alkaline pore water ever sampled in the deep sea. The pore water is also enriched in (mainly carbonate) alkalinity (60 mmol/kg), Na (610 mmol/kg), Na/Cl (1.2), K (19 mmol/kg), B (3.2 mmol/kg), ammonia (0.22 mmol/kg), methane (2 mmol/kg), and C2 through C6 hydrocarbons, all components that are virtually absent in depleted harzburgites and therefore require a different source. The pore water is highly depleted in Mg, Ca, Sr, and Li and has low concentrations of Si, Mn, Fe, Ba, and phosphate. It is slightly depleted in chloride (510 mmol/kg in seawater) and enriched in sulfate (by 7% relative to chloride). This chloride depletion is much smaller than in the deep fluid from Conical Seamount, suggesting that the Conical conduit is more heavily serpentinized and less reactive, allowing more of the H2O from the deep source to arrive at the seafloor without being lost to serpentinization along the way, or that the fluids are rising more rapidly at Conical Seamount and have had less time to react.
Pore water composition vs. depth profiles also reveal that these deep fluids feed an active microbial community that is oxidizing light hydrocarbons from the fluid while reducing sulfate within the black serpentine mud in the upper 20 mbsf. This is a true extremophile community, operating at and probably driving the pH to 12.5, thus perpetuating its own ecosystem. Sulfate reduction is most active at two levels. Microbes within the upper level at 3 mbsf reduce seawater sulfate that diffuses downward against the ascending flow. Those within the lower level at 13 mbsf reduce sulfate that is supplied from the subducting slab by the upwelling fluid. As organic carbon is virtually absent within the depleted, serpentinized harzburgite, the microbes rely on methane and the C2 through C6 thermogenic hydrocarbons for their source of organic carbon and ammonia for their source of nitrogen. Both are supplied by the upwelling fluid. The microbial community intercepts these nutrients and effectively traps them within the ecosystem, where they can be recycled and continually enriched. This process explains the enrichment in organic carbon in the uppermost sediment. Iron sulfides and CaCO3 in the form of aragonite needles and chimneys are also enriched there by reaction between the ascending fluid, the microbial community, and the overlying seawater.
As would be expected, the physical properties of the serpentine muds and clasts at Site 1200 are quite different and both are strongly influenced by the properties of serpentine. The velocities of the clasts, for example, range from 3.8 to 5.5 km/s and average 4.9 km/s, consistent with extensive serpentinization. Whereas the mud and the clasts have the same average grain density (2.64 g/cm3), the average bulk densities of the clasts and the muds are lower and quite different, 2.49 and 1.87 g/cm3, respectively. This is due primarily to differences in porosity between the clasts, which have low porosities, and the muds, which range from 40% to 60% porosity. Assuming the mud constitutes 93% of the breccia, the material in the conduit has an average density of 1.91 g/cm3. If the average density of the crust is 2.75 g/cm3, then the buoyancy of the serpentine mud breccia in the upper crust would be ~0.8 g/cm3 before consolidation, or four times the density contrast between the salt in diapirs and most sedimentary rocks. Similarly, the buoyancy of completely consolidated serpentine mud breccia with a bulk density of 2.64 g/cm3 (the grain density) in fresh ultramafics (~3.2 g/cm3) would be ~0.5 g/cm3, or more than twice the buoyancy of salt in sedimentary rocks. Whereas the average shear strength of the serpentine mud (52.5 kPa) is high for sediments, it is orders of magnitude lower than that for rocks, which is consistent with their extrusion on the seafloor as mud volcanoes.
Interestingly, although vent communities are observed on the summit, the borehole temperatures are very low in the upper 50 m in all of the holes measured at Site 1200, ranging from 2° to 3°C, or ~1.3°C above seafloor values. The heat flow values are quite variable, however, with those measured in Holes 1200A and 1200E near the springs averaging 15 mW/m2, considerably below the global average of 50 mW/m2, whereas the value measured in Hole 1200F, which was farther away, was ~100 mW/m2. The thermal conductivities of the muds are not unusual, ranging from 1.04 to 1.54 W/(m·K) with an average of ~1.32 W/(m·K), but the hydraulic conductivities are extremely low, ~0.6 cm/yr.
Unlike the physical properties discussed above, the magnetic properties of the clasts and muds are similar: the average natural remanent magnetization (NRM) intensities of both are high (0.49 and 0.44 A/m for the clasts and muds, respectively), as are their average susceptibilities (5.58 x 10-3 and 6.81 x 10-3, respectively). In both cases, the magnetization disappears at a Curie temperature of 585°C, indicating that the dominant magnetic mineral is magnetite produced by serpentinization. The only significant difference is that the NRM in the serpentine muds is unstable, whereas that in the clasts tends to be stable, with a high Koenigsberger ratio (average = 2.4). The NRM inclination and declination vary randomly with depth in single long pieces, however, indicating that the magnetization was acquired (i.e., serpentinization occurred) over a relatively long period or when the rock was being deformed or tumbled in the décollement or the conduit of the seamount.