Task 2: Station/EQ relationship

Start: March 1997 (month 13)
End: February 1998 (month 24)
Responsible partner: UEDIN.DGG
Cooperative partner: IMOR.DG

Time-delays and polarizations have been measured for all suitable data in the period January 1996 - May 1997, inclusive. Suitable events are those recorded with station-to-epicenter distance less than hypocentral depth. These are within the shear-wave window which ensures that the shear-waves are not distorted by surface conversions. This constraint places a severe restriction on the number of events that can be used for shear-wave splitting analysis. Also, to identify temporal trends, there needs to be sufficient activity, spread in time, near the station. These criteria were only fulfilled at stations SAU and BJA.

Time-delay measurements are analyzed for temporal trends in Figures 18 and 19. At each station, the mean circular polarization direction is

  

Figure: Shear-wave splitting at SAU from January 1, 1996 to June 1, 1997. Variation of normalized time-delays with time, and polar equal-area maps out to 45$^{\circ }$ of shear-wave polarizations with dotted line indicating average direction, for (a) ray paths in bands with incidence 0$^{\circ }$ to 15$^{\circ }$ to the crack face (sensitive to crack density), and for (b) ray paths in bands with incidence 15$^{\circ }$ to 45$^{\circ }$ to the crack face (sensitive to aspect ratio). Lines are linear least-square fits to data before and after the Vatnajökull eruption (October 1996). Error bars are approximate.

  

Figure 19: Shear-wave splitting at BJA from January 1, 1996 to June 1, 1997. Diagrams as for Figure 18. The dashed lines are linear least-square fits for a shorter time period. The absence of data for March, April and May 1996 is due to instrumental problems.

calculated, and time-delay measurements with polarizations within the standard deviation of this direction are selected (at SAU and BJA this is polarizations N44$^{\circ }$E ${\pm}$ 20$^{\circ }$). The time-delay measurements (normalized over straight-line path distance) are then separated into two bands. The bands are defined by incidence to the vertical plane of symmetry parallel to the mean polarization direction (interpreted as the strike of aligned near-vertical cracks), previous studies have shown that the outer band of time-delays is expected to be the most sensitive to changes in crack aspect-ratio, the result of increasing stresses. At SAU, there is a clear increase in time-delays from about May 1996 to the beginning of October 1996 in the outer band of measurements. After this time, the delays gradually decrease. During the same period, the delays in the inner band remain roughly constant. This suggests that the change in delays could be a result of increasing aspect-ratio of the microcracks with the stress build-up (and subsequent release). The tectonic event likely to responsible for this is the Vatnajökull eruption that started on September 30, 1996, the fourth largest eruption in Iceland this century. The volcano is over 160 km away, so the effect is also expected to be visible on other stations during this period.

The only other station with sufficient data during this time period is BJA. BJA is near the centre of the SISZ (Figure 20) and is in a much

  

Figure 20: Map of SW-Iceland showing all seismicity recorded during the period January 1996 - May 1997 (inclusive). Red triangles are SIL stations that were deployed for the whole period, and blue triangles are stations added to the network during this period. The large black circles show the locations of the Bárdarbunga and Grímsvötn volcanoes beneath the Vatnajökull ice sheet.

more complicated tectonic area than SAU. The data does not show the same trend as SAU over the period May to October 1996, but there is an average increase in delays when a shorter time period is taken (mid-July to October 1996). For this time period, there is an increase in the delay measurements in both bands of data, suggesting there was also some increase in crack density. This trend may be related to the changing stress state prior to the eruption. However, there was also a magnitude 4.2 event, only 7 km from the BJA on October 14, 1996. It is possible that this increase of delay times over the shorter period is related to this more local event, rather than the eruption (200 km from BJA).

Data from station KRI is also shown in Figure 21. The measurements at this station have

  

Figure: Normalized time-delays at KRI from February 15, 1996 to March 15, 1997, as for Figure 18. Lines are linear least-square fits before and after the magnitude 3.9 event, 7 km from KRI (no data recorded in the 0$^{\circ }$-15$^{\circ }$ band).

polarizations N51$^{\circ }$E${\pm}$ 28$^{\circ }$. In February 1997 there was an increase in activity just north of the station which culminated in a magnitude 3.9 event on February 23, 1997, 7 km from KRI. This is the largest event in this area for at least two years. There is some evidence from the time-delay data in Figure 21 that there is an increase in delays prior to this larger event. As can be seen from the projection, the events all have similar locations, hence similar ray paths that make the changes in delays all the more significant, as they are less subject to variations caused by different ray paths.