Earthquake triggering.

Abstract

In this thesis, I present an observational study of multiplet earthquakes as a step towards understanding earthquake triggering. Several subsets of different seismicity catalogs are analyzed for earthquake clustering: (i) a global seismicity catalog of large $(\rm M\sb{s} \ge 7.1)$ earthquakes; (ii) catalogs of large, shallow (depth $<$ 70km) and deep (depth $\ge$ 70km) earthquakes; (iii) regional catalogs in the Philippines and Kurils. Different space, time and magnitude criteria are explored to determine the degree of earthquake clustering for each sub-catalog. The observed earthquake clustering is compared against a null hypothesis: the Poisson process. Multiplets are earthquake clusters which cannot be explained by the Poisson process. We interpret multiplet occurrence as evidence for earthquake triggering. We find multiplets in all the sub-catalogs. The spatial distance within multiplets ranges from across the world, to several kilometers apart and the time within multiplets ranges from days to a few years. Moreover, multiplet occurrence is more common in some areas of the world than in others. Overall, across the global catalogs, approximately 10% of all large earthquakes are triggered by previous earthquakes. Detailed investigations of the source time functions of two multiplet sequences, the Kurils and the Philippines subduction zones revealed that earthquakes within multiplet sequences cannot be differentiated from non-multiplet earthquakes based on the properties of their source time function. This suggests that the rupture process of multiplet and non-multiplet earthquakes are the same. Three earthquake triggering mechanisms are explored: (i) earthquake to earthquake triggering; (ii) triggering due to external forces; (iii) triggering due to human activities. Earthquake to earthquake triggering via static and dynamic stress transfer imparts the most significant stress change and indeed, this thesis shows evidence for this mode of earthquake triggering. Each mechanism, however, changes the existing stress state on faults so that it is either brought closer or further from failure. In fact, if the fault is close enough to failure, an increase in stress by a tiny amount may be sufficient to trigger an earthquake.