Ragnar Stefánsson
Iceland Meteorological Office;
A recent successful stress forecast of a magnitude 5 earthquake, November 13, 1998 in SW-Iceland, puts on the agenda how we can in an effective way follow up such forecasts to approach a better estimate on location, size and time of an eventually forthcoming large earthquake. The forecast was revealed two weeks ahead of the earthquake by Professor Stuart Crampin of University of Edinburgh, based on stress induced variations in S-wave splitting time, observed in the permanent seismic network in Iceland, the SIL network.
Indications of fast build up of stress, approaching fracture criticality, should be followed up by "real time research", i.e. fast evaluations of continuous observations and modelling by well established methods. This implies that during the earthquake cycle, information is utilized continuously, and to a large extent automatically, to gradually create a more realistic model of an approaching earthquake.
We will describe results from research we did before, during and after the November 13, 1998 earthquake and the earthquake sequence that followed it. These include distribution and fault plane solutions of small earthquakes and GPS measurements in a large area surrounding the fault that ruptured in the main shock. Changes in fault plane solutions in the last two days before the main shock as well as clustering of microearthquake activity in the area, reveal patterns that may be linked to instability in the nucleation zone of the earthquake.
Using automatic methods to monitor and recognize such patterns during the earthquake cycle, will be important in our quest to know where, how and when a destructive earthquake will take place.