Task 1: Identify optimal stations and search for precursors

Start: June 1996 (month 4)
End: February 1997 (month 12)
Responsible partner: UEDIN.DGG
Cooperative partner: IMOR.DG
Shear-wave splitting is sensitive to changes in stress, for example, changes prior to a large earthquake. The two principal arrivals identifiable on local earthquake records are the direct compressional P-wave, and transverse shear-wave. In anisotropic media, such as triaxially stressed microcracked rock, the shear-waves split into two orthogonally polarized waves that propagate with slightly different velocities. Careful examination of seismograms can identify the polarization direction of the first (or fast) shear-wave, and the time-delay between the two arrivals (Figure 16). The time-delay is

  

Figure 16: Example of shear-wave splitting at station SAU. On the right-hand side are the three recorded components. On the left hand side, the rotated horizontal components clearly show the time-delay between the two shear-wave arrivals. This event occurred at 07:28:30.4, July 10, 1996, at 9.8 km depth, 5 km from SAU.

particularly sensitive to changes in the anisotropy.

We have used earthquake data provided by IMOR.DG to ascertain which of the SIL stations record sufficient numbers of events suitable for shear-wave splitting analysis, and which of those are suitable for looking for precursory changes. Figure 17 shows those stations with

  

Figure: Shear-wave splitting polarizations measured at SIL stations. All data are within the shear-wave window (45$^{\circ }$), January 1, 1996 - June 1, 1997.

polarization measurements during the period January 1996 to May 1997, inclusive. Many of the stations show a good alignment of polarizations, indicative of seismic anisotropy with transverse isotropy and a horizontal axis of symmetry. The polarizations show the average direction of the maximum horizontal stress is NE-SW.

During the first year, time-delay analysis was concentrated on station SAU as it is close to a seismically active area, and at the easternmost end of the South Iceland seismic zone (SISZ) where a larger earthquake might be expected. Data was initially analyzed from May 1996. The results appeared to show a trend in the time-delay results associated with the eruption beneath the Vatnajökull ice cap in October 1996 (as reported previously). This study has now been continued, at SAU and at other stations, as described under Task 2.