It is well known that seismogenic faulting is normally associated with zones of fluid overpressure (Guđmundsson 1999; Guđmundsson 2000a; Guđmundsson 2000b); in absence of fluids there would be hardly any tectonic earthquakes. The origin and magnitude of the fluid overpressure are, however, uncertain. For example, for the San Andreas fault, there are two models; one assuming the fluids to be of local and shallow origin, the other assuming the fluids to be of deep origin.
In order to throw light on the potential fluid overpressure in a major fault zone, as a part of a general study on the effects of fluid pressure on the probability of fault-slip, field measurements were made of 1717 mineral-filled veins in the damage zone in deeply eroded and well-exposed parts of the Tjörnes fracture zone, particularly on the Flateyjarskagi peninsula (Guđmundsson 1999; Guđmundsson 2000c; Guđmundsson et al. 2000). Most veins are composed of quartz, chalcedony and zeolites, strike roughly parallel or perpendicular to the fault zone, and are members of dense palaeo-fluid transporting networks. A common vein frequency in these networks is 10 veins per meter. Cross-cutting relationships indicate that 79% of the veins are extension (mode I) cracks; 21% are shear cracks. The apertures of most veins, measured as mineral-fill thicknesses, are from 0.1 mm to 85 mm, and the aperture frequency distribution is a power law. The outcrop trace lengths of 384 veins (of the 1717) could be measured accurately. These 384 veins are mostly small and range in length from 2.5 cm to 400 cm, in aperture from 0.01 cm to 0.9 cm, and have an average length/aperture ratio of about 400. Using simple fracture mechanics models, and the appropriate elastic properties of the host rock, this length/aperture ratio indicates and average fluid overpressure during vein formation of 20 MPa. If this fluid pressure acted on a potential fault plane, the effective normal stress across that plane would be zero or negative and, therefore, reduce the driving shear stress needed to trigger slip on that plane to only 4-6 MPa. This compares well with the most common stress drops, estimated at 3-6 MPa, for earthquakes worldwide.
Another potentially strong effect of fluid pressure on the probability of fault-slip in seismic zones is dyke emplacement in the nearby volcanic zones (Guđmundsson 2000d). In this model, dyke injection (and normal faulting) in the volcanic systems can lock or unlock the Húsavík-Flatey fault and the central parts of the South Iceland seismic zone. Dyke injection in the parts of the North and East volcanic zones between the Húsavík-Flatey fault and the South Iceland seismic zone tends to open (unlock) these zones and trigger seismogenic faulting. By contrast, dyke injection south of the South Iceland seismic zone and north of the Húsavík-Flatey fault tends to lock these faults and suppress their seismogenic faulting. Similarly, dyke injection in the north part of the West volcanic zone tends to lock, but dyke injection in its south part (including the Reykjanes peninsula) to unlock, the South Iceland seismic zone. Locking by dyke injection, however, is always temporary because plate pull gradually relaxes the compressive stresses generated by the dykes.
In terms of this model, the largest historical eruption in Iceland, Laki 1783, may have triggered the largest known earthquake sequence in S-Iceland, that of 1784. Conversely, the Húsavík-Flatey fault has recently experienced locking by dyke injection. There was considerable seismicity associated with the Húsavík-Flatey fault until early 1976. Then dyke injection and normal faulting in the northernmost part of the Krafla volcanic systems generated horizontal compressive stresses encouraging sinistral movement on the otherwise dextral Húsavík-Flatey fault, thereby locking the fault. Renewed seismicity on the Húsavík-Flatey fault, since February 1994, indicates that the Húsavík-Flatey fault is currently being unlocked by normal plate-pull movements which gradually relax the horizontal compressive stresses generated by the 1976 dyke. The unlocking began at the westernmost part of the Húsavík-Flatey fault, at the greatest distance from the 1976 dyke.