The main features of the hypothetized model (Figure 2).
The general ongoing processes are on one hand the build-up of shear strain following the next strain release before, and on the other hand build-up of pore pressure, especially fast in the area where the pore pressure was released in the last earthquake sequences. Inside the zone pore pressures will be highest in areas where no significant earthquakes have occurred for a long time. Such areas will be very strong towards E-W transversal motion because of the relatively strong N-S component in the largest horizontal compression. Pores which have directions perpendicular to the compression will have a tendency to close.
Small earthquakes are frequent near the bottom of the seismogenic zone, because of high stress gradient there where material flow prevails below the seismogenic zone boundary Stefansson/etal:1993. Small cracks stick into the hard stuff. Gradually one of these cracks (possibly N-S crack because of history and because of lengthening of pores in notherly direction) may take over. It opens farther into the hard stuff and pore fluids from its sides also from below flow into it. The pore pressures in its neighbourhood lower and ease further opening of this crack. A local zone of weekness has been created in the hard stuff which may take up some of the general strain build-up in the area. Gradually it may increase to an elongated crack of N-S direction at the bottom of the seismogenic crust, below a several hundred years old N-S earthquake fault.
For a long time, however, maybe 140 years Stefansson/Halldorsson:1988, a general shearing will prevail in a large area around the zone to take up the general 2 cm plate motion. This will be the general characteristics untill relatively fast motion starts somewhere within the sheared area to take up, most probably including a large earthquake, a significant part of the total strain energy.
This N-S expansion causes compression across the zone and thus deviates the NE-SW compression to become less than 45 E from N, Figure 2. The optimal direction for making an earthquake on a N-S fault is close to 25-30 E from N. The earthquake releases the deviatoric stresses by right lateral faulting on the N-S fault which also becomes longer as also the pressures that were built up perpendicular to the zone are locally released resulting in N-S expansion. It is most probable that the nucleus of the large N-S earthquake is in the narrow SISZ, but to release all the accumulated strain energy it will extend tens of kilometers outside the zone. The N-S extension of the E-W zone, a minor rifting of the zone near the origin of the earthquake, will migrate along the zone. The resulting reduction of N-S compression in the zone will open the way for stable E-W transverse motion along the zone. It will not create large earthquakes on E-W faults as there are no long E-W earthquake faults available in the zone. Rather the material within the zone will bend elastically to start with and by laminar motion on many E-W boundaries inside the zone. However, this mostly aseismic E-W shearing of the zone will propagate westward (in fact also eastward) and will hit the next candidate for a large N-S earthquake, i.e. it will slow down there while increaseing the deviatoric stresses that have already built up there. The rifting apart of the zone thus increases the probability that earthquakes will occur along the zone, especially at places where local strain has already built up, and this will be on N-S faults.
After the earthquake, fluids start to flow from below into the created earthquake crack (figure 2). It expands it gradually. Most of the expansion and strain change caused by the intrusion would take days. In the Vatnafjöll earthquake of 1987 Agustsson/etal:1996 the duration time for most of the expansion was of the order of one day. However, it is possible that such a process might be a much slower in the main area of SISZ than in Vatnafjöll, which is in a more volcanically active area and thus may have better developed fluid roots. This flow of material after the earthquake will increase the compression around the N-S fault and help to create a stable condition near the earthquake source for a long time, maybe for hundreds of years or until pore pressures below the seismogenic zone have increased again so new faulting is initiated at this particular N-S fault by the combined efforts of shearing and pore pressure.