The June 2000 earthquake sequence in South Iceland

The South Iceland seismic zone (SISZ) is an E-W trending transform zone that connects the Western volcanic zone and Reykjanes peninsula in the west to the Eastern volcanic zone in the east (Figure 1). The SISZ is approximately 70 km long and 10 to 20 km wide. The SISZ accommodates the relative plate motion along an array of N-S trending right-lateral strike slip faults instead of having only one long E-W trending left-lateral fault [Einarsson and Eiríksson (1982)]. This behaviour is termed bookshelf faulting [Sigmundsson et al. (1995)]. Destructive earthquake sequences are historically known to occur in the SISZ at intervals of 45 to 112 years [Einarsson et al. (1981)]. The earthquake sequences usually consist of several earthquakes with magnitudes over 6.5 that occur on N-S trending faults.

Such a sequence started at 15:40:41 GMT on June 17, 2000, with an earthquake of moment magnitude M_W=6.5. The hypocenter was located at 64.97$^\circ$N, 20.37$^\circ$W and 6.3 km depth. A second event of M_W=6.4 followed at 00:51:47 GMT on June 21, 2000. The hypocenter was located at 63.98$^\circ$N, 20.71$^\circ$W and depth 5.1 km [Stefánsson et al. (2000)]. The epicenters are marked with large stars in Figures 36 and 37. The two large earthquakes occurred on N-S trending faults as indicated by location of the aftershocks [Stefánsson et al. (2000)], geodetic measurements [Árnadóttir et al. (2001)], [Pedersen et al. (2001)] and mapping of surface fractures [Einarsson et al. (2000)].

  

Figure 35: Time series of VOGS covering the period of the June 2000 earthquakes. The times of the June 17 and 21 earthquakes are noted with vertical lines. Vertical axes have been adjusted for visual clarity.

A significant coseismic deformation signal was observed at all operational ISGPS stations for both the main earthquakes. Figure 35 shows an 80 day long time window from VOGS covering the times of the earthquakes. The coseismic displacements due to each of the large earthquakes can easily be separated. The weighted average of the coordinates is calculated for three time intervals, using data from 10 days before the earthquakes, three days between the earthquakes and 10 days after the earthquakes. The coseismic displacements for each earthquake are then obtained by calculating the diffrences of the coordinates. REYK is assumed to remain fixed during the earthquakes. Results are listed in Table 9. The errors are estimated as four times the formal error resulting from the weighted average (see Section 3.1). Vertical displacements were insignificant at all stations. The cosesismic displacements observed at the operational stations are shown as vectors in Figures 36 and 37. Station HLID was not in operation at the time of the earthquakes. HVOL was not in operation until June 15, 2000, when a new wind generator was installed. Thus there are only two days of data behind the weighted average of coordinates prior to the earthquakes at HVOL.

HOFN was excluded from routine processing of the data from June 7 until June 16. That is why there are missing values for HOFN in Table 9. When HOFN was included in the daily processing, the formal errors of the daily coordinate results were larger by a factor of two and offsets as large as 20 mm west and 10 mm north (at SOHO) were observed. The offsets occurred 10 days prior to the June 17 earthquake and were at first interpreted as a precursor for the large earthquake sequence, but when HOFN was excluded from the processing the formal errors came back to normal values and the spurious offsets disappeared from the time series. It is not known why HOFN had such an impact on the coordinate results but perhaps it is due to some kind of instrumental error.

 

Table 9: Coseismic displacements at the continuous GPS stations for the events on June 17 and 21, assuming REYK is fixed. Uncertainties are at the 1$\sigma$ level.

	June 17 [mm]		June 21 [mm]		Combined displacement [mm]
Station	East	North	East	North	East	North
VOGS	18.8 $\pm$ 0.6	-10.4 $\pm$ 1.1	8.4 $\pm$ 0.7	-0.1 $\pm$ 1.2	27.3 $\pm$ 0.5	-10.5 $\pm$ 0.9
HVER	6.5 $\pm$ 0.7	-4.5 $\pm$ 1.1	-1.9 $\pm$ 0.8	6.7 $\pm$ 1.2	4.6 $\pm$ 0.6	2.2 $\pm$ 0.9
OLKE	-10.0 $\pm$ 0.7	2.9 $\pm$ 1.2	-13.3 $\pm$ 0.8	11.1 $\pm$ 1.3	-23.2 $\pm$ 0.6	14.0 $\pm$ 0.9
THEY	12.4 $\pm$ 0.8	-11.9 $\pm$ 1.1	6.7 $\pm$ 0.9	-6.8 $\pm$ 1.3	19.1 $\pm$ 0.7	-18.6 $\pm$ 1.0
SOHO	9.3 $\pm$ 0.8	-7.5 $\pm$ 1.1	5.3 $\pm$ 0.8	-7.5 $\pm$ 1.2	14.7 $\pm$ 0.7	-15.1 $\pm$ 0.9
HVOL	8.9 $\pm$ 1.0	-6.9 $\pm$ 1.4	4.1 $\pm$ 0.9	-4.7 $\pm$ 1.2	13.0 $\pm$ 0.9	-11.5 $\pm$ 1.3
HOFN	-	-	3.6 $\pm$ 1.1	-4.2 $\pm$ 1.2	-	-

 

The displacements, assuming REYK is fixed, at VOGS are larger on June 17 than on June 21, although the June 21 earthquake is closer to VOGS and the earthquakes have similar magnitudes. This implies that other earthquakes on the same day contributed to the observed signal on June 17. In Figure 36 three stars noting earthquakes with moment magnitudes greater than 4.5 are on the Reykjanes peninsula. These earthquakes are believed to be triggered by the large earthquake (K. Vogfjörð, personal communication 2002) and occur at times (from east to west): 15:41:06, 15:41:11 and 15:45:27 GMT. InSAR interferometric results of the Reykjanes peninsula and GPS network measurements carried out in July 2000 and April to June 2001 show clearly significant deformation associated with the Reykjanes peninsula earthquakes on June 17 [Pagli et al. (2002)], [Árnadóttir et al. (2002)]. The largest deformation is observed near Kleifarvatn.

Figures 36 and 37 (black arrows) show that the coseismic offsets at HVER are not consistent for the two earthquakes. One would expect the direction of the displacements to be similar. Árnadóttir et al. (2001) modelled the geometry and displacements of the faults using GPS network campaign data from 1995, 1999 and 2000 (June 19-30). The displacements at the ISGPS stations are compared with modelled displacements supplied by Þóra Árnadóttir for each large earthquake in Figures 36 and 37, assuming REYK is fixed. The observed offsets at stations to the east (THEY, SOHO and HVOL) due to the June 17 earthquake (Figure 36, Table 9) agree

  

Figure 36: Observed (black arrows) and calculated (red arrows) horizontal coseismic displacements for the June 17 main shock [Árnadóttir et al. (2001)], assuming REYK is fixed. Large star notes the location of the June 17 main event and small black stars note significant events triggered by the June 17 earthquake. Red line by the large earthquake shows the causative fault as modelled by [Árnadóttir et al. (2001)]. Note the discrepancy for stations in the Hengill area (OLKE, HVER and VOGS).

  

Figure 37: Same as Figure 36, for the earthquake on June 21. The only station that does not fit well to the model is HVER.

fairly well with the model except at HVOL. The fit is very poor for stations OLKE, HVER and VOGS. This is to be expected for station VOGS since its offset is affected by the three earthquakes on Reykjanes peninsula. Indeed, preliminary modelling of coseismic displacements resulting from the Reykjanes peninsula earthquakes using GPS network measurements are in good agreement with the offset observed at VOGS (Þ. Árnadóttir, personal communication 2002). The predicted displacement field for the three Reykjanes peninsula quakes is not large enough to explain the coseismic deformation at OLKE and HVER. A number of earthquakes occurred in the Hengill area on June 17, as mentioned in Section 4.3, along an E-W trend. The earthquake data from the June swarm has not been fully processed and earthquakes with magnitudes similar or less than the Reykjanes earthquakes might be hidden in the data, specially if the earthquake was triggered by the S-wave of the main June 17 event (K. Vogfjörð, personal communication 2002). Presently there are two earthquakes with local magnitude M_L=3.5 and one with local magnitude M_L=4.3 in the SIL database on June 17 in the Hengill area. It is tempting to conclude that the anomalous coseismic deformation signal observed at OLKE and HVER results from cumulative local deformation sources.

The reference station REYK is within the deformation field for the June 17 and 21 earthquakes. The coseismic displacements of REYK can be estimated with three different methodes: 1) using the model of [Árnadóttir et al. (2001)]; 2) HOFN is well outside the deformation field of the earthquakes so changes in the baseline REYK-HOFN for the June 21 earthquake can be attributed to REYK; 3) using data from international analysis centers that include REYK in their processing. The predicted coseismic displacements for REYK are 1 mm west, 2 mm north the June 17 event and 3 mm west, 3 mm north for the June 21 event (Þ. Árnadóttir, personal communication 2002). The predicted displacements for HOFN are less than 1 mm (Þ. Árnadóttir, personal communication 2002). We can thus attribute the coseismic displacements observed at HOFN (Table 9) to REYK resulting in 3.6 $\pm$ 1.1 mm west and 4.2 $\pm$ 1.2 mm north coseismic displacement for REYK due to the June 21 earthquake, in good agreement with the predicted displacements. Visual inspection of time series of REYK from international data processing centers like [SOPAC (2002)] and [MIT (2002)] reveals offsets ranging from 5 to 8 mm towards west and from 4 to 7 mm towards north for the combined displacements due to the earthquakes.