Revisiting the Tidal Activation of Seismicity With a Damage Mechanics and Friction Point of View

It seems plausible that solid Earth tides should trigger earthquakes. Most of the time, tides are responsible for the largest temporal variations in crustal stresses. Diurnal and semi-diurnal tides have amplitudes reaching 30 mbar and peak loading rates on the order of 20 mbar/hr. These tidal oscillations are superimposed on a tectonic loading rate of 1 mbar/hr or less. Except for the case of moonquakes, there appear to be no published correlations between seismicity and Earth tides that can withstand rigorous statistical scrutiny. Recently the critical point model of earthquakes has been invoked to suggest that the ratio of seismicity during times of increased tidal loading to times of decreased tidal loading takes on anomalously large values as the preparatory region of the earthquake approaches a critical state. This idea is based on the non-linearity of stress-strain behavior near macroscopic failure. The loading and unloading response become increasingly asymmetric as failure is approached. The tidal loading referred to here is the Coulomb effective stress appropriate for the fault orientation and slip direction characteristic of the region where the large earthquake occurred. This hypothesis is referred to as LURR (Load Unload Response Ratio). We have re-analyzed the data from several examples for which published LURR results suggested anomalous behavior before large earthquakes. Our analysis shows that the LURR function is largely controlled by the occurrence, randomly with respect to tidal phase, of a few moderate earthquakes. We conclude that the reported anomalous behavior of LURR prior to large earthquakes is of no predictive significance beyond confirming that foreshocks often precede large earthquakes. It may be that in the Earth time dependent failure processes preclude tides from having a significant effect on earthquake occurrence. In the course of reviewing the LURR hypothesis we have identified what we believe is a basic flaw. While it is true that loading and unloading are asymmetric, this asymmetry exists only when new damage is being produced during the loading part of the cycle. This is known as the Kaiser effect in laboratory studies of acoustic emissions. Purely cyclic loading follows a hysteresis loop. If tectonic loading is added to a cyclic load, then the asymmetric response occurs only during that portion of each cycle when the load exceeds the previous maximum. In the case of Earth tides this effect should produce a periodicity of roughly 15 days. If there are departures from a strictly linear tectonic loading rate, then the tidal effect on seismicity will not even be periodic, and thus not easily detected with spectral methods. We propose a novel procedure to test this new hypothesis.