Crustal deformation in extensional regimes: Iceland, Nevada and SW Turkey

Abstract

In 1991 a 23-point, 30 x 25 km, GPS geodetic network was established in the Hengill ridge-ridge-transform triple junction, SW Iceland. The GPS data were processed using the Bernese V3.2 software. The ambiguity-fixed solution yielded scaled formal errors of less than 1 cm in the horizontal and about 1 cm in the vertical. The effect of ocean loading on the vertical component of the GPS measurements, a hither-to ignored effect in GPS data, was examined and found to be negligible for surveys of this type. Extensional deformation processes in the western Basin and Range province were studied by analysing the co- and postseismic deformation associated with the 1954 Rainbow Mountain - Fairview Peak - Dixie Valley sequence of M>6.0 earthquakes. Levelling and triangulation data which constrain the coseismic deformation are fitted well by modelling uniform slip on rectangular planar dislocations embedded in an elastic half-space. The best-fitting fault geometries dip at 50 to 80 and extend to depths of 8 to 14 km. A simultaneous inversion of the triangulation and levelling data which spanned the coseismic period reveals that slip determined geodetically tends to be equal to, or greater than, the surface offsets. Using the single value decomposition method to invert the data allowed determination of where the coseismic slip was resolvable given the data distribution. The static stresses changes induced by each earthquake in the 1954 sequence were calculated using the source models derived from the geodetic modelling. After the first earthquake each event in the 1954 sequence was preceded by a static stress increase of 104 to 105 Pa. Thus, it appears that static stresses may have played an important role in triggering the later earthquakes. The vertical postseismic deformation which followed the earthquakes was modelled assuming stress redistribution in the Earth's crust through viscous flow in the lower crust. Calculation of surface deformation caused by the analastic response to a normal faulting earthquake was performed using the finite element method. The viscosity which best fit the postseismic levelling data was 1x10(^20) Pa s at depths of 10 to 30 km. GPS measurements made between 1988 and 1992 show that north-south aseismic extension within SW Turkey is occurring at rates of 11.7+5 mm/yr. The deformation extension was modelled along 2-dimensional north-south profiles using the finite element method. These calculations show that the postseismic deformation created by eight earthquakes of Ms6.5 and greater within SW Turkey this century accounts for 10% of the north-south extension rate measured by GPS. The rest can be explained as postseismic deformation from earthquakes in earlier centuries. Modelling indicates that north-south extension rates decrease from west to east across the province. Reoccupation of the GPS network in the year 2000 will yield a clear regional strain signal, given the errors of the 1989 survey.