UC Berkeley

Coseismic static stress changes and dynamic stresses generated by seismic waves cause changes in water level in wells and in streamflow and alter hydrogeologic properties such as permeability and porosity. These hydrologic responses occur over a wide range of timescales, from as short as the duration of shaking to permanent changes, and over a wide range of length scales, from within the fault zone to more than 10. 000. km away. Consequently, they provide an opportunity to study hydrogeologic processes and properties at spatial and temporal scales that are often difficult to study. Water-level changes in wells can reflect the pattern of volumetric strain caused by seismic slip. The sign and amplitude of observed water-level changes in many wells, however, are not consistent with this pattern of strain. Instead, such anomalous changes in water level can be caused by shaking-induced changes in unconsolidated sediments through consolidation, liquefaction, or even dilation close to the rupture. Anomalous changes in water level may also reflect changes in permeability or storage properties. Changes in streamflow may similarly have multiple explanations. Many recorded examples of streamflow changes can be explained by changes in vertical permeability caused by shaking. Other hydrologic phenomena that can be attributed to dynamic, rather than static, strains include liquefaction, the eruption of mud volcanoes from depths greater than hundreds of meters, and changes in the eruption interval of geysers. Because the stability of faults is sensitive to pore pressure, changes in hydrogeologic properties or pore pressure induced by earthquakes can, in turn, influence seismicity. It is thus not unreasonable to expect that some earthquakes may have a hydrologic origin. Hydrologic processes can greatly amplify small strains, and some hydrologic responses appear to have very small strain thresholds. This provides some hope that hydrologic and hydrogeochemical monitoring may detect any hypothetical, diagnostic precursory strains. Although it is unclear whether future documented hydrologic precursors could actually be used for earthquake prediction, they may at least provide new insight into the physics of earthquakes.