Useful predictions ahead of large earthquakes and lessons learned for future progress

Abstract

A significant basis for this article is the outcome of 4 multinational research projects which were carried out in South-West Iceland from 1988–2006, focusing on crustal processes that preceded two magnitude-6.6 earthquakes in June 2000. The seismic activity that preceded a magnitude 6.3 double-earthquake in the same seismic zone in 2008 is also significant, as well as other research work which helps to understand how observable crustal processes lead to earthquakes. A significant outcome is that it cannot be assumed that any two earthquakes have the same precursory processes. Therefore, statistical analysis of precursors of past earthquakes is of limited value for predicting future earthquakes. On the other hand, with highly sensitive seismic monitoring it is possible to observe the nucleation process for each specific large earthquake for long enough time for earth-realistic modeling of it and extrapolating towards the earthquake in time and space. In the Iceland crust, with its fluid-rock interactions, pre-earthquake activity on a scale of years is expected. This allows a long-term approach to prediction. We apply historical information and sensor-based data to find probable sources of earthquake nucleation. We monitor the nucleation process for a possibly impending earthquake at these sources, and then model the process to find its governing factors and extrapolate those in time and space, aiming towards finding hypocenter, fault-size, impact, and time of the impending earthquake. We refine our models by predicting frequent medium-sized earthquakes and compare the predictions with measurements. We predict how the possibly impending large earthquake would trigger earthquakes at other locations. Given the complexity of the crust, we must take all observed changes into account when developing models of pre-earthquake processes. The development of a continuously operating geo-watching system is discussed to link scientific evaluations to warnings that can be used by emergency authorities.