Determination of a Permeability-depth Curve for the Oregon Cascades Employing Numerical, Analytical, and Statistical Methods

We employ analytical and numerical models as well as statistical methods and signal processing techniques to study both pore-fluid pressure diffusion and coupled groundwater and heat transfer in the subsurface. Both processes are used to infer the large-scale vertical permeability distribution and other hydrogeologic parameters in the Oregon Cascades volcanic arc. Multiple temperature-depth profiles and their deviations from a linear(conductive) relationship provide insight into advective heat transfer and related groundwater flow patterns and velocities. In addition, the coupled modeling approach provides background heat-flow rates of 0.080 < H < 0.130 W/m² for the study region where heat flow is otherwise typically masked by cold groundwater recharge. We also infer regional scale groundwater recharge rates of about 1 m/year. Statistical, Monte Carlo, and signal processing techniques are used to investigate if seasonal groundwater recharge due to spring snow melt enhances earthquake occurrence after a pressure-diffusion related time lag, during summer. We find statistically-significant cross correlations between groundwater recharge and earthquake occurence at Mt. Hood, Oregon, with a time lag of about 151 days. This phase lag and a mean earthquake depth of about 4.5 km imply an average permeability of about 10^-15 to 10^-14 m² for a depth range of about 2 < z < 5 km. Combining the heat-advection study, the hysroseismicity study, and other investigations allows us to infer permeability, k, as a function of depth, z. We suggest approximately log(k) = -13 to log(k)= -18.3 m² for z=0 to z=12 km, respectively, for the Oregon Cascades. Our results agree with values compiled by Manning and Ingebritsen (Rev. Geophys., 1999) for continental crust in general. However, we suggest an exponental k(z)-curve for z < 0.8 km and a power law relationship for z > 0.8 km.