A Seismic-Wave-Induced Fluid Pressure Increase: An Alternative Hypothesis for the Strain Transient at Long Valley Caldera Following the 1992 M7.3 Landers, California Earthquake

Microseismicity at Long Valley caldera triggered by the 1992 M7.3 Landers, California, earthquake was accompanied by a transient contractional strain change recorded by the borehole dilatometer (POPA) just outside the caldera rim. The exact mechanism of the strain transient has not been established. Within Long Valley caldera, large distant earthquakes, including the Landers earthquake, have induced fluid pressure drops in non-thermal wells and fluid pressure increases in thermal wells. At each responding well, the fluid-pressure changes have a characteristic time history that is well matched by the diffusion of an abrupt pore pressure change near, but not at, the observation well. The transient contractional strain change induced by the Landers earthquake has a similar time history, suggesting that the contractional strain signal could be due to earthquake-induced fluid pressure changes near the Devil's Postpile. This hypothesis can be evaluated using fluid pressure data collected since November, 1999, in a borehole 30 m from the strainmeter. The POPA strainmeter exhibits a seasonal signal with an amplitude of 2 - 5 x 10^-6, reaching peak annual extension between November and March and peak annual contraction between May and July. Fluid pressure also varies seasonally about 2 m every year, and peak contractional strain occurs within a few weeks of peak fluid pressure. The seasonal increase in fluid pressure due to spring snowmelt also coincides with increased rates of contraction at POPA each year. These data indicate that the strainmeter records contractional strain in response to increasing subsurface fluid pressure, with a ratio of approximately 1.5 x 10^-6/m H₂O. Comparison of the strain data with time series of snow water content at Mammoth Pass shows that snowpack loading contributes 10% or less of the seasonal variation at POPA. The strain transient induced by the Landers earthquake was approximately 0.25 x 10^-6, and could have been produced by a fluid pressure increase of 0.17 m, about half the size of water level changes inside Long Valley caldera induced by the same earthquake. Increasing fluid pressure compresses the relatively impermeable strainmeter and surrounding grout, producing a contractional output, which under these circumstances does not represent the undrained extensional strain in the formation. Co-located fluid pressure and strain measurements are necessary in order to understand the physical effects of seismic waves at locations of remotely triggered seismicity.