LONG-TERM OBSERVATIONS OF PRESSURE AND TEMPERATURE IN HOLE 892 B , CASCADIA ACCRETIONARY PRISM

Two holes drilled into the Cascadia accretionary prism during Ocean Drilling Program (ODP) Leg 146 were instrumented with borehole seals ("CORKs") for long-term monitoring of temperatures and pressures at in situ conditions. We report the results obtained during submersible data recovery operations at the sites in September, 1993, 9.5 months after the CORK instruments were emplaced. The installation at Hole 889C off Vancouver Island was severely damaged during a deployment made very difficult by poor weather and unstable hole conditions; no useful data were recovered there. In contrast, the installation at Hole 892B in the accretionary prism off Oregon produced excellent thermal and pressure data that provide constraints on the hydrogeology at that site. Site 892 is located over the hanging wall of a hydrologically active thrust fault that is penetrated at a depth of about 100 m in the 146-m-deep Hole 892B. In addition, there is a well-defined regional bottom-simulating reflector (BSR) whose depth shoals about 8 m to 72 mbsf at the site, presumably because of the thermal effects of fluid flow in the fault zone. Results of numerical modeling demonstrate that the local shoaling of the BSR is consistent with the effects of recent up-dip fluid flow that initiated roughly 400 yr ago at an average flux per meter along strike of 1 × 10~ ms"'; steady-state flow is precluded. Hole 892B was sealed with a pressure gauge and a 10-thermistor chain extending to a depth of 122 mbsf. Temperatures in the CORKed hole define a generally uniform gradient of about 68 mK m". At the depth of the regional BSR, this gradient gives a temperature identical to that on seawater-methane-hydrate phase boundary at the equivalent pressure. The gradient is significantly greater than that defined by shipboard temperature measurements made in exploratory holes about 200 m to the southwest. The disagreement can be explained if the exploratory holes intersected fault-controlled zones of fluid upflow at shallower depths than the CORKed hole. The gradient defined by the shipboard measurements may reflect locally diminished heat flow in the footwall of the fault. CORK temperatures also define a distinct thermal anomaly at the depth of the fault zone, which is consistent with results of numerical simulations of a transient fluid flow event. The up-dip fluid flux is constrained to be approximately 6 × I0" ms~, nearly two orders of magnitude greater than the average rate inferred from the shoaling of the BSR. Pressures in the sealed hole decayed from an initial ("shut-in") superhydrostatic value of 70 kPa to a low, relatively stable value of 13 kPa within a few months after drilling (lithostatic pressure at 100 mbsf in this hole is about 630 kPa). The initial superhydrostatic value may have been caused by charging of the formation during drilling, although it is more likely that high pressures were present in the fault zone initially and drained after the fault was penetrated, probably to the surrounding formation spanned by the open section of hole where lower fluid pressure may be present. This conclusion is reached by considering the hydraulic transmissivity required to support the high rates of flow inferred from the CORK and BSR data in light of the transmissivity determinations made at both elevated and reduced pressures by Screaton et al. (this volume). Attenuation and phase of the seafloor tidal loading signal recorded in the sealed hole remained constant throughout the 9month recording period at 0.5 and 0.3 hrs (-9° phase lead at 12 hr period), respectively. These characteristics are consistent with the presence of approximately 2% free gas in the pore volume of the sediments below the BSR and above the perforated interval, and high fault-zone transmissivity connecting the perforated interval to the zone above containing free gas.