The stress field for the new models was extrapolated to spring 1999, with the additional stresses due to plate motion since 1912. In the "improved model", shear stress inside the SISZ are only high at small places and at some places at the northern and southern margins of the seismically active zone. This is similar in the "short rupture model". In general, in the west and in the east, where the SISZ meets the ridge segments, stresses are concentrated. Is this a weakness of the models or is it the real situation?
An argument against high stress there is, that there was no large event (M6.5) since 1706 near the ridge segments. The origin of the stress concentrations at the end of the SISZ, i.e. at the tips of the adjoining ridges, is the fact that the ridges do not extend to infinite depth, but are assumed to reach only 7 km depth and enter than an inelastic, hot region not capable of supporting stresses for time periods of years. Deeper penetration of the brittle layer there would homogenize the stress field between rift tips at some average value. Entailed is the following: In the models introduced here, high initial stresses at the tips are needed to obtain high enough stresses in the center of the SISZ (see section about the initial stress field above). Further, high tensional stresses at the rifts mean high initial stress in 1706 compared to the annual increase by plate motion, as long as the rifting speed is kept constant. Indeed, if one compares the situation in 1912 with that calculated for 1999 the stress build-up by plate motion is very low. It would mean that the stress release by earthquakes in a series of events as it was observed and is modelled here, would be followed by a period of quiescence until the stress level of 1706 is reached again.
On the other hand, there are some indications that the present stress release indeed mainly takes place at the ends of the SISZ: Interestingly, in 1987, there was a strong earthquake (MS=5.8) at 63.91N, 19.78W (198, -9) near Vatnafjöll (see Figure 27) at the east end of the SISZ. This event was not included in the modelling as its magnitude was below M=6. Nevertheless, it occurred in a region of high stress in the models. In 1998, there were 2 stronger events at the Hengill triple junction at the western end of the SISZ: June 4 (M=5.1) and November 13 (M=5) both accompanied by a lot of smaller events.
Considering this conflicting information, another (extreme) model will be computed during the next months using a much deeper width of the brittle layer at the rifts with the consequences of lower stresses at the ends of the SISZ and a stronger influence of the plate motion over that of the initial stress amplitudes.