We had hoped to use several additional stations in the shear-wave window at SAU to investigate why observed time-delays between split shear-waves in Iceland are approximately twice those usually observed elsewhere, and why time delay values show large scattering. Unfortunately, financial constraints prevented installation of additional SIL stations.
The range of time-delays between split shear-waves is found to vary between different regions worldwide and sometimes between different stations in the same region. Time-delays vary with crack density, crack aspect ratio, and isotropic P- and S-wave velocities. These effects can be calculated, and are reasonably well understood. Time-delays are also observed to be higher in regions with high heat flow (Crampin 1994) and this is believed to be the main reason for the larger values observed in Iceland. Also following the increase in Band-1 time-delays before the Vatnajökull eruption in 1996, the level of time-delays has been decreasing for both Band-1 and Band-2 by approximately 2 ms/km over 1997-1998. It would appear that the crust has been slowly adjusting to the strain released by the eruption as the mid-Atlantic ridge gradually takes up the change in strain.
However, the large scatter of time-delays and polarizations is difficult to explain. During the last two big earthquakes (November 1998 near BJA and June 2000 near SAU) it was observed that the scatter in time-delays increases considerably with the onset of the earthquake and the large scatter continues for several weeks thereafter (Figures 10a and 10b). It is hoped that the controlled source SMSITES project will help to resolve this difficulty.