Fluid flow and pressure distribution within active faults are essential but poorly constrained parameters that affect fault zone processes. Observations on active margins have shown that manifestations of fluid seepage at the seafloor are commonly associated with active tectonic features and that episodic flow occurs in fault zones [Carson and Screaton, 1998]. Notably, geochemical and geophysical evidence for rapid flow from seismogenic depths channeled along thrusts has been obtained in ODP drill holes on the Cascadia margin [Davis et al., 1995; Sample, 1996]. Sibson [Sibson, 1981] broadly classified fluid-fault coupling processes as seismic pumping and fault valve mechanisms. Since then, a number of physical models have been proposed to explain pressure transients or fluid discharge associated with seismic (and aseismic) slip: poroelasticity and pressure diffusion [Davis et al., 2001; Ge and Stover, 2000; Muir-Wood and King, 1993], damage and fluidization due to ground shaking [Gavrilenko et al., 2000; Wang et al., 2001], fracturing/sealing cycles [Barton et al., 1995; Husen and Kissling, 2001; Renard et al., 2000; Sleep and Blanpied, 1994] and solitary waves [Henry, 2000; Rice, 1992]. However, in general, the relationship between episodes of fluid flow and occurrences of fault sliding remains to be defined. While any or all may occur, the differences can potentially be resolved through long term flow monitoring. For example, permeability changes effect the tidal response of a seep while poroelastic effects do not (e.g., Elkhoury et al., 2006; Tryon et al., 2002).

Coupling between deformation and fluid flow may lead to post seismic fluid release, to precursor events, as well as to systematic variations of flow rates, fluid chemistry and pore pressure during inter-seismic phases. Evidence for changes in subsurface water chemistry associated with tectonic activity has been noted in a wide variety of geological environments [Biagi et al., 2004; Italiano et al., 2001; Sano et al., 1998]. For example, significant progressive elevations in Cl-, Mg, SO4, and Sr2+ were observed prior to the Mw 6.9 1995 Kobe earthquake based on the analysis of bottled water taken from a 100 m deep water well drilled near the epicenter [Tsunogai and Wakita, 1995]. The bottled water gave a several year record with an exponential change in chemistry leading up to the point of the event itself.

We hypothesize that fluid seeps fed by high permeability conduits located within fault zones are sensitive to the state of stress in the fault zone and, thus, may respond to processes occurring there much as on-land water wells respond to both short and long term strain. However, non-seep environments also respond hydrologically to strain events. Dilation and compression associated with tectonic activity occurs throughout the sediment column and while pressure changes are attenuated at shallow depths due to differences in matrix elastic properties, strain is still detectable at the surface. For example, an array of fluid flow meters detected a similar pattern of inflow and outflow events on the outer rise off Costa Rica correlated with subduction zone earthquake activity [Tryon and Brown, 2005]. Such short-term (days to weeks) flow events can only be associated with shallow (<5-10 m) strain due to the rapid increase in hydraulic impedance with greater depth. These flow meters were merely recording the flow response due to shallow pore pressure changes induced by aseismic outer rise extension events. Identifying and understanding these sorts of responses relies primarily on a comparative study of flow, pore pressure, and seismic data from the project.