THE SOUTH ICELAND SEISMIC ZONE

According to the NUVEL 1A plate model (DeMets et al. 1994) the direction of the divergent motion across the plate boundary of the Eurasian relative to the N-American plate in Iceland is N103$^{\circ}$E at a velocity of 1.85 cm/year. If all of this motion is taken up by the SISZ, which is oriented almost due EW, the relative left-lateral motion across SISZ would be approximately 1.8 cm/year and a NS opening component across the zone could be around 0.4 cm/year. Historically most of the earthquakes in this EW left-lateral shear zone have had right-lateral motion on NS striking faults as observed on the surface. Faults due to known historical earthquakes are found side by side at less than 5 km distance from each other (Einarsson 1991; Einarsson et al. 1981; Stefánsson et al. 1993). There are indications that the seismic cycle, i.e. the time period of strain build-up and of break-up (strain release) of the whole SISZ is 140 years. This suggestion was based on historical seismicity and tectonic considerations (Stefánsson and Halldórsson 1988). Volcanic pulsations in the most active volcanic zone (EVZ) of Central Iceland, inferred from tephrastratigraphic studies coincide with this estimate of the seismic cycles (Larsen et al. 1998). On the basis of the estimated magnitudes (Ms) of historical earthquakes, the total seimic moment release during a 140 year period is approximately 10²⁰ Nm (Halldórsson 1987; Stefánsson and Halldórsson 1988). Assuming that the seismogenic crust in the 70 km long zone is only 10 km thick (Stefánsson et al. 1993; Tryggvason et al. 2002) and total displacement across the zone is 1.85 cm/year, the maximum moment build-up over a 140 year period would be $6.5\ast10^{19}$ Nm. There is a discrepancy in the results of the calculations above in such a way that slightly more moment seems to be released in earthquakes than accumulated from plate motion. It would be more natural to expect the opposite, as some of the strain energy could be released aseismically. One explanation could be that the cumulative moment of the historical earthquakes has been overestimated. Another possibility is that the thickness of the elastic/brittle crust is underestimated. The thickness is estimated from the depth limit of microearthquake activity. But it has also been suggested that a part of the strain energy build-up comes from a local sources, due to fluid-induced expansion near the bottom of the seismogenic crust (Stefánsson and Halldórsson 1988; Stefánsson 1999a; Stefánsson 1999b).