Site 1254 is located ~1.5 km arcward from the deformation front at a water depth of 4183 m, close to the holes drilled at Site 1040 during Leg 170 (Kimura, Silver, Blum, et al., 1997). Hole 1254A is positioned ~15 m west of Hole 1040C, and Hole 1054B is ~50 m northeast of Hole 1040C (Figs. F3, F23, F24). Therefore, all comparisons to Leg 170 results are to Hole 1040C at Site 1040, as it was the only one which penetrated the décollement and underthrust.

The primary objective of Site 1254 was to investigate a fault zone in the prism, investigate the décollement, and install a long-term observatory for monitoring of fluid flow, pressure, and temperature in the décollement. Results from Site 1040 (Kimura, Silver, Blum, et al., 1997) and seismic data (Fig. F24) provided the framework for drilling the sedimentary sequence and the interpretation of pore fluid geochemistry and structure. Site 1040 geochemical anomalies suggest that deeply sourced fluids, perhaps from seismogenic depths, are migrating along the décollement and prism fault. Site 1254 was intended to investigate in detail the structure and geochemistry of these zones and install an observatory in the décollement. Although perturbed by drilling disturbance, high recovery at Site 1254 enabled detailed structure observations where they were considered reliable and higher-resolution chemical sampling than was possible during Leg 170. It is also possible to better correlate intervals of maximum fluid flow to specific structural horizons.

The seismic record (Fig. F24) in the vicinity of Site 1254 (CMP number 3130) shows no coherent reflections above the décollement. This reflects the general chaotic sedimentary pattern observed in cores from Hole 1254A. The first prominent reflector relevant for drilling objectives is at 6 s TWT which marks the boundary between margin sediments and the underthrust sequence, which was cored at 361 mbsf. The prism fault zone is not imaged in the seismic data.

After setting the reentry cone at Hole 1254A, we cored the prism fault zone (150 to 230 mbsf) and the décollement (300 to 367.5 mbsf) with the RCB. Recovery averaged at ~88% throughout the cored interval. With generally good hole conditions we planned to case the hole with 10³/₄-in casing. However, after running the casing to 232 mbsf, the casing could no longer advance and had to be pulled up. Soon it became clear that the reentry cone had hung up on the casing; when the sections that were jammed into the cone were pulled up into the moonpool, it became obvious that the casing had collapsed in the throat of the reentry cone for unknown reasons. Hole 1254B, the second attempt for a CORK-II installation, was offset 50 m to the northeast. However, drilling conditions there prevented us from deepening the hole to >278 mbsf, when the drill string got stuck during several attempts to deepen the hole. Therefore, we decided to install the osmotic fluid sampler in the upper fault zone with the screen located at 225 mbsf; this interval cored and analyzed at Hole 1254A was not recored because of time constraints. The depth for the screen was determined by inference from the geochemical results of Hole 1254A, which indicate that deeply sourced fluids containing thermogenic hydrocarbons are present in the target zone. After a successful installation of the 10³/₄-in casing, the installation of the CORK-II failed as it got stuck ~20 m above the final depth. Attempts to penetrate further probably caused the 4¹/₂-in casing to break right below the CORK-II head. Thus, we had to abandon Hole 1254B with ~20 m of casing sticking out of the reentry cone.

In total, we drilled 367.5 m at Site 1254, with 140.5 m cored and 227 m drilled and washed. Because of the nature of the tectonic structures encountered, part of the core was heavily disturbed by RCB drilling, which makes structural and paleomagnetic studies especially difficult. However, the generally good recovery with an average of 89% allowed extensive whole-round sampling of the cored sections for pore water and organic geochemistry in addition to shipboard sampling for physical property and paleomagnetic studies as well as personal samples for postcruise studies.

The sedimentary sequence recovered at Site 1254, Subunit P1B after Leg 170, is dominated by structureless and typically unsorted dark greenish gray claystones with variable, subsidiary quantities of silt and rare interbedded volcanic ashes, sandstone, and clasts, spanning a badly dated sequence of presumed Pliocene–Pleistocene age (Fig. F25). Recovered cores often show moderate to extreme degrees of drilling disturbance, nonetheless, coherent fragments of more lithified sedimentary rocks do indicate that much of the section is either massive or slightly mottled, which is suggestive of moderate bioturbation.

The dominance of clay minerals within the sequence is readily apparent from smear slides, as is the downcore decrease in volcanic ash. Fresh volcanic glass is present at low (<10%) and moderate levels (<30%) above 230 mbsf, becoming heavily altered deeper (>300 mbsf) in the section. The continental provenance of the sediments cored in Hole 1254A is clear from the abundance of quartz and feldspar grains and also from the bright, brownish red biotite mica flakes that are found at all stratigraphic levels. The terrigenous nature of the sediments is confirmed by the very low biogenic component (<5%) of the sediment, restricted to occasional nannofossils above 200 mbsf and below 360 mbsf. Below Section 205-1254A-15R-2 (360.62 mbsf), the proportion of diatoms increases sharply (>10%). The appearance of diatoms is considered important for understanding the structure of the forearc prism because the uppermost sedimentary subunit in the subducting Pacific stratigraphy (Subunit U1A) recorded high percentages of diatom abundance (Kimura, Silver, Blum, et al., 1997).

Redeposited blocks of shallow water peloid limestones, lithified prior to incorporation within mudstones, are found throughout the section, which is consistent with fluidized gravity and debris flows being the dominant mode of sedimentation. The cobbles show evidence for a shallow-water depositional environment, identified by shallow-water bivalve shell fragments and small gastropods.

Compared to the sequence of well-preserved tephra found at ODP Sites 1039 and 1253 on the subducting Cocos plate, there is little well-preserved tephra stratigraphy found at Site 1254. Although occasional, thin altered ash layers are recognized, they are rare, typically <2 cm thick and often completely altered to claystone. Volumetrically the tephra represent <1% of the total section. Two thicker coherent ash layers are recorded at Site 1254 (205-1254A-5R-8, 14–20 cm, at 193.49 mbsf and 8R-8, 22–65 cm, at 222.37–222.80 mbsf). Both the thicker ashes preserve relatively fresh glass shards and are interpreted to be the product of primary airfall deposition followed by settling through the water column. The base of the tephra recovered in Section 205-1254A-8R-8 was not recovered, resulting in a minimum thickness estimate of 43 cm. Because Site 1254 is ~150 km from the nearest arc volcano in Central America, this thickness at this range indicates that this must have been a very large eruption, comparable to the Minoan Ash from Santorini as the closest analogue (Watkins et al., 1978). Major and trace element analyses of this tephra (interval 205-1254A-8R-8, 22–65 cm) characterize its source as being the volcanic arc of Central America.

Coring at Site 1254 targeted two different structural domains based on Site 1040 results: (1) a fault zone from 150 to 223 mbsf containing fractured sediment and locally steep bedding dips called the prism fault zone and (2) the décollement zone from 300 to 368 mbsf (Fig. F25). A variety of deformation structures is present at Site 1254, and description of deformation was based on breccia size, foliation, hardness of breccia clasts, and polished surfaces. Because structural observations in poorly lithified material require good quality cores and the recovered cores are sometimes severely disturbed by drilling, it is difficult to distinguish natural from drilling-induced features.

Cores from 150 to 223 mbsf show various degrees of deformation, with the intensity of deformation, particularly breccciation and brittle shearing increasing downward, reaching a peak at ~219 mbsf. Deformation is discontinuous, being focused along sheared horizons, 20 cm to 2 m thick. These horizons are characterized by stratal disruption, foliated breccia with fragments as small as a few millimeters in length, brittle shear zones, deformation bands, and distinctly inclined bedding. Concentration of deformation structures at ~210 and 219 mbsf documents that this is indeed a fault zone that has a distinct geochemical anomaly as discussed below. A well-preserved foliated breccia (interval 205-1254A-R8-1, 0–24 cm; 213 mbsf) indicates reverse movement, based on Riedel shears, which may imply that the fault system is reverse. Paleomagnetic reorientation of this shear zone suggests that the fault is a northeast- or southwest-dipping feature, implying that it is a thrust fault (Fig. F25) that strikes parallel to the deformation front.

The second interval cored started at 300 mbsf, and well-preserved structures are observed starting at Core 205-1254-11R (319.30 mbsf) (Fig. F25). Cores typically show pervasive drilling disturbance, previously described during Leg 170 as "spiraliferous" (Kimura, Silver, Blum, et al., 1997), consisting of a spiral rotation of clay-rich sections. Despite the drilling disturbance, some bedding plane orientations were observed. Bedding and fissility show various dips, indicating heterogeneity of deformation, but the paleomagnetic reorientation shows that they consistently dip northeast or southwest, parallel to the deformation front. The recovered section from 319.30 to 367.50 mbsf is characterized by intense deformation. The deformation is heterogeneous, and brecciation, usually associated with a strong foliation, is the basis for dividing the deformed interval in two zones.

The upper zone from 319.30 to 328.90 mbsf is characterized by generally increasing brecciation with depth, producing fragments of <0.3 cm. Foliation is common throughout Core 205-1254-11R resulting in a clear alignment of clasts, which are equidimensional, but internally strongly foliated. Below 324.15 mbsf (Core 205-1254A-12R) deformation sharply decreases, and consolidated and coarsely brecciated sand layers become common. These sandstone layers have steeply dipping laminations and a few web structures. We interpret this well-defined change in deformation intensity to mark the top of a relatively less-deformed rock volume that may be the footwall of the fault identified between 319. and 328.9 mbsf and may be related to the décollement zone. This indicates a more articulate structural geometry than that observed at Site 1040 (Kimura, Silver, Blum, et al., 1997; Tobin et al., 2001).

The upper boundary of the décollement zone at 338.5 mbsf is defined by the increasing amount of deformation in Core 205-1254A-R13. The definition of the décollement upper boundary is always difficult to place because the deformation gradually increases in its intensity toward the zone of concentrated shear; a sharp increase in deformation is not observed between Cores 205-1254A-12R and 13R. The décollement zone itself is heterogeneous, with a general downward increase of brecciation intensity, fragment aspect ratio, and hardening of the sediments. Despite the good recovery, "spiraliferous" drilling disturbance affects the cores, even though less extensive than at Site 1040. Unlike Site 1040, "spiraliferous" disturbance is not concentrated in the lowermost part of the décollement zone. Brecciation can be pervasive and severe with fragments characterized by polished surfaces; the development of scaly fabric is precluded by the abundant silt and sand in the sediments. From 354.8 to 355.9 mbsf sandstone layers are brecciated and foliated. At 360.60 mbsf the appearance of diatoms in the sediments marks the lithologic boundary with the hemipelagic Subunit U1A of the underthrust (Figs. F25, F26). The lithologic boundary is present below 50 cm of finely brecciated sand and 10 cm of highly sheared clay indicating a surface of ductility contrast which appears as a major structural discontinuity. The hemipelagic sediments below the lithologic boundary are still intensely deformed and brecciated with aligned clasts showing a strong internal foliation (Fig. F27). The base of the décollement is placed at 364.2 mbsf and is below the lithologic boundary. Deformation starts to decrease and becomes discrete below 364.2 mbsf, where intact sediments are separated by 3- to 8-cm-thick brittle shear zones producing gouge or Riedel shears (Fig. F28). These brittle shear zones show exceptionally consistent normal movement and landward dips when reoriented to the geographical coordinates. The hemipelagic sediments above 364.2 mbsf are also deformed by normal faults, a few of them are present as conjugate features. At Site 1254 the décollement zone has a thickness of 25.7 m. In this analysis the décollement has cut down into the uppermost underthrust section incorporating a small amount (4.2 m) of Subunit U1A into its base. The complex geometry of the décollement system at Site 1254 contrasts with that described at Site 1040 and 1043, where the top of the décollement was placed at an increase in brecciation, even though somewhat arbitrarily, and the lithologic boundary between the prism and the hemipelagic subunit coincides with the base of the décollement.

Paleomagnetic measurement on archive-half sections and discrete samples are severely degraded by pervasive drilling disturbance and drill string overprints. Natural remanent magnetization inclinations are highly variable after alternating-field demagnetization and make the identification of magnetic polarity and the construction of a magnetostratigraphy impossible. However, the data were useful in carefully selected intervals to reorient core segments for structural interpretation. Demagnetization curves of discrete samples from the prism sediments (Subunit P1B) are often poorly behaved, indicating that they have a very unstable magnetization. Two significant high magnetic intensity and susceptibility zones were observed in the intervals from 184 to 202 mbsf and from 310 to ~350 mbsf. The interval of the first anomaly is close to the prism fault zone at ~210 to 220 mbsf, and the second anomaly is within the décollement zone. These variations suggest changes in concentration, grain size, and chemical components of magnetic minerals related to lithology and/or chemical alteration perhaps related to fluid flow.

A total of twenty 35- to 45-cm whole rounds were sampled at Site 1254 for pore fluid geochemistry. Pore waters were analyzed for Ca, Mg, K, Na, B, Ba, Fe, Mn, Sr, H₄SiO₄, NH₄⁺, and SO₄^2– concentrations (Fig. F25). Li, Ca, K, Mg, and Na were analyzed in "real time" on the shipboard ICP-AES between 305 and 366 mbsf to identify the horizons of maximum fluid flow within the décollement zone based on correlation to nearby Site 1040. The "real-time" chemical analyses were available 2 hours after core recovery and, together with careful observations of hydrocarbon gas concentrations and penetration rate, helped to identify the top of the underthrust section.

The pore fluid salinity in the prism sediments (Subunit P1B) is lower than that of seawater by 20% and focused excursions of higher dilutions up to a maximum of 29% are present at 218 and ~351 mbsf (13 m from base of the décollement zone). The two main focused salinity minima also show propane, lithium, and calcium concentration maxima, as well as Mg/Ca, potassium, and magnesium minima. The geochemical excursions between 210 and 218 mbsf are present within a highly fractured interval interpreted as a fault zone, whereas the excursions at ~351 mbsf coincide with a brecciated sandy interval that is moderately indurated. A small peak in calcium, lithium, and propane concentrations is present at 330 mbsf, and it is associated with another sandy, brecciated interval in the décollement. These data suggest, together with results from the entire interval cored during Leg 170, that fluid has migrated along conduits and permeated the lower half of the deformed wedge. Assuming that the geothermal gradient is ~20°–30°C/km, the source region must be present at >4 km depth because the minimum temperature required for thermogenic gas formation is 90°–100°C. The minimum in potassium concentrations at 218 and 351 mbsf further suggests that the deformed sediments have been permeated by a fluid from an elevated temperature source of 80°C to 120°C where the illitization reaction is effective, which consumes potassium. Also, the K depletion signature of this fluid provides an approximate upper limit to the temperature at the source of ~<150°C. Above this temperature, fluid-rock reactions leach potassium from the rocks. Lithium, like potassium, is partitioned into solids at low to moderate temperatures. At higher temperatures, >100°C but <250°C, lithium is released into the fluid phase (Chan and Kastner, 2000). The precise threshold temperatures for the partitioning of Li and K into the solid or fluid phases are as yet unknown. Clays and other silicate dissolution or alteration releases boron into the fluid phase; however, clay, especially illite, formation consumes boron and may be responsible for the low boron concentrations within the deformed sediments. The deeply sourced fluid, however, is not enriched in dissolved silica.

Geochemical excursions in calcium, lithium, propane, potassium, and magnesium are present at ~218 mbsf within the prism fault zone at Site 1254. Similar increases in calcium, lithium, and propane concentrations, as well as marked decreases in magnesium and potassium concentrations, were observed at an observed prism fault zone at Site 1040; however, it was present between ~180 and 200 mbsf. Therefore, the upper geochemical boundary at Hole 1254A is ~20 m below the same boundary observed during Leg 170.

The geochemical boundary at ~218 mbsf separates intervals with pore fluid chemistry typical of clay-rich sediments from those permeated by a fluid from an elevated temperature source, and it seems to be independent of any lithology. Bulk sediment chemistry is also relatively homogenous throughout the entire prism. Changes in pore water chemistry in a lithologically and chemically homogenous sediment section likely result from fluid advection into the lower half of the deformed sediment section. The chemical changes observed at the base of the fault zone (conduit) at ~218 mbsf are similar to those observed near the bottom of the décollement zone associated with the fluid anomaly in a sandy interval. Except for the biogeochemical components, the pore fluid concentration depth profiles of the underthrust section are similar to those at the reference Site 1039. The concentrations themselves are slightly different in magnitude than at Site 1039, presumably reflecting the changes in solubilities and dissolution rates of the major sediment components under the new pressure regime as they are underthrust. In contrast to Site 1039, the higher ammonium concentrations and the absence of sulfate at the interface between décollement and the underthrust sediments reflect the fact that all the sulfate is reduced at Site 1254 by microbiological activity. Sulfate reduction thus reaches completion in the uppermost few meters of the underthrust hemipelagic section, resulting in somewhat elevated methane concentrations within the zero-sulfate depth interval. These geochemical patterns are similar to those observed at Hole 1040C.

Volatile hydrocarbon gases were sampled by headspace and vacutainer techniques at a higher frequency than pore water samples to assist in determining the exact depths of the inferred fluid conduits associated with fault zones discovered at Site 1040. Analyses of the vacutainer samples (Fig. F25) show that the gas mainly consists of methane but also contains considerable amounts of higher alkanes up to pentane. Methane concentrations were very high (7–9 x 10⁵ ppmv) throughout the cored interval but dropped to ~4 x 10⁴ ppmv directly below the décollement zone at 364 mbsf. Propane, which is a strong indicator of deeply sourced fluids because of its thermogenic origin (>90°–100°C required), shows one peak at 216 mbsf and another in the basal part of the décollement zone at 355 mbsf, with maximum levels of 326 and 370 ppmv, respectively. These high propane concentrations correlate with structurally identified fault zones. Similar patterns, at much lower concentrations, were also observed in the headspace gas samples.

Samples for microbiological investigations were taken and either frozen or fixed for postcruise ATP quantification, DNA assessment, or cell counts. Samples of drilling water were frozen to evaluate contamination of cores. The chemical tracer for quantifying microbiological contamination was not deployed during coring at Site 1254 because of concern that the trace element chemistry of the PFT may affect postcruise pore fluid geochemical analyses. Particulate tracer tests yield fluorescent microsphere counts suggesting very low to no particulate contamination in the interior of the microbiology whole rounds.

Porosities and bulk densities at Site 1254 (Fig. F25) exhibit trends similar to those seen at Hole 1040C. Variations in porosity and density within the structurally defined décollement zone correlate with core descriptions: in general, zones of lower porosity (40%–45%) correspond to zones characterized by "spiral" deformation interpreted as drilling disturbance; zones of higher porosity (50%–55%) correspond to zones characterized by brecciation. Porosity is also low (42%–44%) between 358 and 361 mbsf, within and adjacent to a zone of localized shear. Porosity increases and bulk density decreases sharply below 361 mbsf across the lithologic boundary between prism sediments and Pleistocene diatomaceous claystone.

We attempted three downhole measurements of formation temperature and pressure, two with the DVTPP at 50 and 200 mbsf and one at 150 mbsf with the Davis-Villinger Temperature Probe (DVTP). The temperature measurement at 200 mbsf was the only deployment with an interpretable decay curve and indicated a temperature of 3.59°C. This is in good agreement with measurements from Site 1040C. All pressure measurements were unsuccessful as a result of tool movement when in formation. However, pressures measured at the mudline and bottom of the hole are in very good agreement with expected hydrostatic pressures expected at that depth which clearly demonstrates that the pressure measurements are reliable.

In summary, the analyses of structural fabric and geochemical anomalies allowed us to identify a geochemical boundary at ~218 mbsf that separates pore fluids typical of clay-rich sediments above from those permeated by a fluid from a source at elevated temperature below within a lithologically homogenous section. At ~338.5 mbsf a fault marks the upper boundary of the décollement zone, which extends into the upper meters of the underthrust sequence at 364.2 mbsf. Maximum pore fluid chemical anomalies, indicative of active fluid flow, appear to preferentially follow zones characterized by brittle fabric. Analysis of cores from the two intervals allowed us to select the optimal depth interval for the long-term borehole fluid sampler experiment. However, because of unstable hole conditions, two attempts to install a CORK-II failed and Site 1254 had to be abandoned because of time constraints.