Anelastic deformation in Iceland studied using GPS: with special reference to post-tectonic motion following the 1975-1985 krafla rifting episode, and isostatic rebound

Abstract

The Krafla volcanic system is a spreading segment in north Iceland. A decade-long crustal spreading episode began there in 1975. Up to 8 m of rift-normal surface widening occurred along an 80-90 km-long section of the plate boundary. Isostatic uplift in the vicinity of the melting icecap Vatnajōkull has been proposed. A third GPS survey of a regional network surrounding the Krafla system was conducted in 1992. In 1991 a 10-point GPS network was installed and measured for the first time around Vatnajōkull. The 1991 and 1992 GPS data were processed using the Bernese software. Differencing the 1992 results with those from 1987 and 1990 revealed a regional deformation field with a maximum, rift-normal expansion rate of 4.4 cm/yr near the rift, decreasing to 3 cm/yr at large distances. The time-averaged spreading rate in north Iceland, 1.8 cm/yr, cannot account for this deformation. The vertical deformation field reveals regional uplift throughout the network area, at its maximum closest to the rift and decreasing with distance. Three different models were applied to study the postdyking ground deformation, (1) continued opening at depth on the dyke plane in an elastic halfspace, (2) stress redistribution in an elastic-viscous layered medium, and (3) stress redistribution in an elastic layer over a viscoelastic halfspace. The latter model was developed by extending mathematical techniques previously used to model surface displacements resulting from thrust faulting to the case of dyke emplacement. For the model of continuous dyking at depth, a range of dykes will fit the deformation field. Using the elastic-viscous model, the motion 1987-1990 and 1990-1992 is simulated adequately given the survey errors, but the 1987-1992 deformation is poorly fitted, suggesting that a more realistic geophysical model is required. Using the elastic-viscoelastic approach the effects of historical episodes in the region were subtracted from the observed displacement fields and the remaining motion was modelled as relaxation following the recent Krafla rifting episode. The best-fit model involves a halfspace viscosity of 1.1 x 10(^18) Pa s. The vertical field is noisy, but indicates that the Krafla dyke complex rifted the entire elastic layer. Isostatic uplift centred on Vatnajőkull is inconsistent with the vertical deformation field. The model suggests that the Krafla volcano became inactive after 1988/1989. The model further predicts that the width of the "plate boundary zone" is greater than that of Iceland itself.