Triggered earthquake probability forecasting: The development of a new model for earthquake recurrence

Abstract

This project set out to determine the probability, on a global scale, of large earthquakes being triggered by large earthquakes; “triggered earthquakes” being those occurring close in time and space to the parent earthquake but not close enough to be considered aftershocks.

The occurrence of such triggered events was investigated by removing aftershocks – “declustering” – from a global catalogue of M ≥ 5 events recorded between 1964-2000. We term the set of spatially proximal events from which aftershocks and triggered events have been removed a supercuster without aftershocks. Subsets of the global catalogue with the same spatial limits as superclusters without aftershocks but from which aftershocks and triggered events have not been removed are termed superclusters with aftershocks.

It was found that the distinction between “aftershocks” and “triggered events” was arbitrary. Consequently we make no distinction between these two types of post-mainshock earthquake.

The greatest weakness in our modelling is the assumption of physical connectedness between all events in a supercluster, whose size is defined by its largest event. The model will not be able to be used in practice until a combined spatial/temporal model is developed.