Seismicity Induced by Groundwater Recharge at Mt. Hood, Oregon, and its Implications for Hydrogeologic Properties.

Earthquakes induced by human-caused changes in fluid pressure have been documented for many years. Examples include seismicity induced by filling reservoirs and by fluid injection or extraction. Less well-documented are seismic events that potentially are triggered by natural variations in groundwater recharge rates (e.g., Wolf et al., BSSA, 1997; Jimenez and Garcia-Fernandez, JVGR, 2000; Audin et al., GRL, 2002). Large groundwater recharge rates can occur in Volcanic Arcs such as the Oregon Cascades where annual precipitation is > 2 m of which > 50 % infiltrates the ground mostly during snowmelt in spring. As a result, infiltration rates of > 1 m per year concentrated during a few months can occur. Near-surface porosities are about 5-10 %. Thus, groundwater levels may fluctuate annually by about 10-20 m resulting in seasonal pore fluid pressure variations of about 1-2 x 10⁵ Pa. Such large-amplitude, narrow-duration fluid pressure signals may allow investigation of seismicity induced by pore fluid pressure diffusion without the influence of engineered systems such as reservoirs. This kind of in-situ study of natural systems over large representative elementary volumes may allow determination of hydrologic parameters at spatial and temporal scales that are relevant for regional hydrogeology. Furthermore, natural hydrologic triggering of earthquakes that persist for decades provides insight into the state of stress in the crust and suggest long-term near-critical failure conditions. Here, we approximate the temporal variations in groundwater recharge with discharge in runoff-dominated streams at high elevations that show a peak in discharge during snow melt. Seismicity is evaluated as time series of daily number of earthquakes and seismic moments. Both stream discharge and seismicity are compared at equivalent frequency bands by applying segmented least-squares polynomial fits to the data. We find statistically significant correlation between groundwater recharge and seismicity at Mt. Hood, Oregon. We can use the time lag of about 120 days between the two records to estimate the regional hydraulic diffusivity (1 m²/s) and other hydrogeologic parameters (permeability ≈ 10^-13 m², vertical matrix compressibility ≈ 10^-10 m²/N). These values are comparable with our results from coupled heat and groundwater flow studies that are based on bore hole temperature data at Mt. Hood.