Site 897 is situated in the Iberia Abyssal Plain over a north-south basement ridge (Fig. 2), within the presumed ocean/continent transition (OCT) zone off western Iberia (see "Introduction" chapter, this volume). The site is one of a transect of drill sites across the OCT designed to study the petrological changes in the basement rocks within the OCT as a means of identifying the processes that accompanied continental break-up and the onset of steady-state seafloor spreading. Geophysical modeling had predicted that the ridge lay at, or close to, the ocean/continent boundary (Whitmarsh et al., 1990; Whitmarsh et al., 1993) and, by analogy with samples obtained from a 100-km-long ridge within the ocean/continent transition west of Galicia Bank (Boillot, Winterer, Meyer, et al., 1987; Boillot, Comas, et al., 1988), might consist of serpentinized peridotite (Beslier et al., 1993). Four holes were drilled at 40°50.3'N, 12°28.4'W to test this hypothesis, with the primary scientific objective of penetrating basement to a depth sufficient to firmly establish its character. Up to 694 m of Pleistocene to Early Cretaceous age sediments were cored in Holes 897A, 897C, and 897D. Up to 143 m of basement were cored in Holes 897C and 897D.
The evolution of this site revealed in our cores began with exposure on the sea bed of a serpentinized peridotite. About 90% of the peridotite sampled is undifferentiated harzburgite or lherzolite, whose mean original composition was 70%-80% olivine, 15%-20% pyroxene, and l%-2% spinel. These rocks are heterogeneous and range from pyroxene-rich peridotite to dunite. The coexistence of plagioclase and spinel in the rocks suggests that they last equilibrated at low pressure (900-1000 MPa or about 30 km depth). The wide range of peridotite compositions and the locally high proportion of plagioclase (up to 40%) suggest that the peridotite may have experienced some melting and even mobility of magma. About 10% of the peridotite was brecciated during and after serpentinization. The brecciation ranges from pervasive fracturing with serpentine and calcite veining to gravel-sized serpentinized peridotite clasts embedded in a calcite and/or serpentine matrix. In a process yet to be determined, but that probably involved the (hydrothermally driven) circulation of large quantities of seawater, the serpentine in Hole 897D shallower than 760 mbsf has been almost completely replaced by calcite. Locally, the breccias show a well-developed foliation, associated with a late-stage low-temperature shear deformation event. Structural analysis of tension gashes in the cores suggests that the principal extensional direction was horizontal during the late brittle deformation event. While the peridotites have experienced late low-temperature deformation, they exhibit almost no signs of high-temperature ductile deformation. At present, it is not possible to date the emplacement of the peridotite, nor is it possible to be precise about the mechanism of its emplacement. Nevertheless, the tectonometamorphic evolution of these mantle rocks is compatible with an uplift under a rift zone, like the peridotite bounding the western Galicia Bank margin. The extensive serpentinization is the result of the peridotite having reacted with seawater. The altered rocks have a range of physical properties that depends on their alteration state. Densities and velocities lie in the range of 2.3 to 2.5 g/cm3 and 2.8 to 7.1 km/s, respectively. Alteration also was accompanied by a reduction in magnetic susceptibility.
The oldest sediment recovered at this site is a mass flow deposit, found immediately above the basement, that contains peridotite blocks and a variety of sediments of late Hauterivian(?), early Barremian to late Aptian age that youngs upward. This age range includes the estimated Barremian (130 Ma) time of first seafloor spreading at this latitude (Whitmarsh et al., 1990). As the sequence youngs upward systematically, rather than showing a disordered age pattern, it is probable that several debris flows were emplaced during the Early Cretaceous, not in post-late Aptian time, and so are related to the last stages of rifting and the earliest seafloor spreading. The provenance of the sediments in the flows is uncertain. The weak disruption, poor sorting, and absence of graded deposits suggest a relatively short transport path; yet, the deposit contains clasts of continental basement material that must have come from a continental margin or from submarine continental fault blocks. The peridotite must have been exposed at the seafloor during the late Hauterivian(?), early Barremian to late Aptian time, perhaps along a fault scarp, to have contributed clasts to the flows. An east-west seismic reflection profile shows that the crest of the ridge, 1.0 km east of the site, is capped by middle Eocene sediments. Thus, the problem is how to explain a gravity-driven deposit, which is much older than early Eocene, when seismic reflectors, inferred to be roughly contemporary with its age of emplacement, now appear to be several hundred meters deeper than the ridge crest in the basins that flank the ridge. If the ridge was at its present elevation with respect to the adjacent basins when the mass flows were deposited (and assuming that the mass flow could travel only 100 to 200 m uphill), then either all the material is of local origin (i.e., from the same ridge) or, more speculatively, the nonperidotite material was transported from the continental crust to the north, along the axis of the rift itself. Had the peridotite not yet been fully uplifted to form the present-day ridge when the mass flows occurred, then the nonperidotitic material might have arrived from almost any direction. However, the implied upbending of reflectors toward the ridge, which should have accompanied any such uplift, is not seen. Either hypothesis implies that the peridotite ridge was exposed at the sea bed for several tens of millions of years.
Serpentinized peridotite breccias have been found frequently both at other DSDP/ODP sites where serpentinite was drilled (see Juteau et al., 1990, for a review) and elsewhere in ophiolites. During Leg 125, clasts of serpentinized tectonized peridotite and metabasalt were found at the base of Sites 778, 779, and 783, drilled on the flanks of seamounts behind the Izu-Bonin and Mariana arcs (Fryer, Pearce, Stokking, et al., 1990). Serpentinite breccias also were recovered from young Atlantic oceanic crust drilled during Leg 82 (Bougault, Cande, et al., 1985) and from the backarc Tyrrhenian Sea drilled during Leg 107 (Kastens, Mascle, et al., 1987). Surprisingly, no breccia was found at Site 637, which was drilled on the flank of a peridotite ridge off Galicia Bank (Boillot, Winterer, Meyer, et al., 1987). In the western Alps, a widespread ophicalcite sediment breccia exists around the Queyras Ophiolite, a relict of the Ligurian Tethys continental margin (Tricart and Lemoine, 1991); similar deposits also are known in the northern Apennines. Thus, serpentinite breccias are known from many different marine geological environments and attest to the ease with which serpentinized peridotite can become disaggregated when exposed on the seafloor. This may be because such outcrops are subject to extensive brittle deformation during their emplacement. The unusual, but not unique, aspect of the Site 897 mass flow deposit is the variety of sediments of different ages entrained with the serpentinite clasts.
Following the deposition of the mass flow deposit, a sequence of 29 m of unfossiliferous coarse-grained clastic rocks and claystones was deposited. Deposition below the CCD and the constraints posed by the age of immediately overlying beds suggest it was deposited during Eocene-late Paleocene time, when the regional CCD was relatively shallow (Tucholke and Vogt, 1979). Some of the sediment was deposited as debris flows or turbidites. Clasts suggest the existence of contemporary basaltic volcanism and the reworking of continental basement material.
The claystones pass abruptly upward into a 260-m-thick sequence of calcareous graded beds and carbonate-poor pelagic/hemipelagic sediments deposited below the CCD during middle Eocene to middle Miocene time. These sediments were deposited by turbidity currents and have been reworked by contour currents. Evidence of contour currents is sparse in cores of early Miocene age, but the calcareous turbidites persist. A 10.5-Ma hiatus in deposition began in the middle
Miocene and continued until the late Miocene. This hiatus correlates with the angular unconformity between the regional acoustic formations 1A and 1B, defined by Groupe Galice (1979), which is visible in multichannel seismic reflection profiles across the site. The unconformity represents the onlapping of horizontally bedded turbidites on to reflectors that were folded and uplifted by northwest-southeast compression on the western Iberia margin during the Betic compressional phase in southern Spain. This sediment deformation appears to be closely correlated with a monoclinal fold frequently found over the peridotite basement ridge within the Iberia Abyssal Plain (Masson et al., in press).
The lithology remained unchanged after the hiatus until the earliest Pliocene, when about 290 m of Pliocene to Pleistocene, fine-grained terrigenous turbidites, having thin intervening layers of nannofossil clay, began to accumulate at close to 60 m/m.y. The site appears to have been above the CCD. The reason for this sudden increase in sedimentation is not clear, but is possibly related to climatic changes.
The principal results from this site can be summarized as follows: