Predicting Earthquakes With the MRI Algorithm

INTRODUCTION Earthquake prediction is still in its pre-feasibility stage. Empirical techniques based on the concept of seismic gaps (Fedotov, 1965; Sykes, 1968) are difficult to put in quantitative terms. The disagreement, before the destructive 1985 Mexico earthquake, on whether the Michoacan, Mexico gap had rupture potential is an example. Some seismologists actually argued that the gap was aseismic and incapable of generating major earthquakes. Such ambiguous interpretations of seismic gaps are not easily avoided in the present state of the art. Yet most large earthquakes continue to occur in seismic gaps. This may or may not have something to do with the fact that many segments of plate boundaries, perhaps the majority, have been designated as seismic gaps by someone. On the other hand, the gaps assigned maximum seismic potential ( e.g. , by Nishenko and McCann, 1981) correlate rather poorly with the locations of major earthquakes. The MRI algorithm (Lomnitz, 1993)...