Tidal stress changes within the Earth due to the gravitational attraction of the Sun and Moon can exceed 20 mbar/h, whereas tectonic stress rates are of the order of 0.2 mbar/h. Although the absolute magnitude of tidal stresses is less than 1% of the typical stress drops observed in earthquake, a correlation between the tidal stress history on a fault plane and the time of failure should exist if (1) tidal and tectonic stresses are combined in a simple manner, (20 the tectonic stress buildup over time is smooth, and (3) failure occurs immediately upon attainment of some critical stress.

To search for possible tidal triggering of intermediate and deep focus earthquakes, 7359 events were selected from a global catalog and subdivided into 30 regions. Each event was assigned a tidal phase relative to both the semidiurnal and biweekly tidal stress curves at the time of occurrence. All tidal phase distributions were checked for a preferred phase orientation using the Rayleigh test and for nonrandomness using the chi-square test. Results of this analysis of the semidiurnal Earth tide show that three regions have a preferred phase direction and four regions display nonrandomness significant at the 90% confidence level, which would be expected for this number of random data sets. The biweekly tide, when normalized for the unequal distribution of biweekly phase over time after correction for clustering of events in some of the data sets, shows only three non-random data sets at the 95% level as defined by the chi-square test. Thus intermediate and deep focus earthquakes in most subduction zones occur independently of both the semidiurnal and biweekly tidal stress cycles, and we may conclude that failure below 70 km does not occur at the instant a stress curve, the simple sum of tectonic and tidal stresses, reaches the yield strength of subcrustal rocks. This implies that either tectonic stress rates prior to failure are not smooth and are unaffected by predictable sources of stress or that failure does not necessarily occur when a certain critical stress is reached.

These may in turn be due to some time delay characteristic of deep failure, to the stress history of the material, to the diffusion of fluids, or to the mode of failure itself, at present a unique and unexplained phenomenon at such high temperatures and pressures. ¿ American Geophysical Union 1987