Deformation of the Japanese Islands and seismic coupling: an interpretation based on GSI permanent GPS observations

The entire area of the Japanese Islands has been covered by the permanent GPS observation network of the Geographical Survey Institute since 1994, In this paper we use a solution for the vectors of motion during 1995 for a selection of 116 stations to discuss the origin of the observed deformation field. We refer the displacement field to Eurasia using the VLBI-determined motion of Kashima and demonstrate that other choices such as the Okhotsk or North American plates for north Japan are not compatible with the data. 1 yr GPS velocities are much higher than geological constraints would allow because these short-term measurements include transient elastic deformation. However, the good qualitative agreement between the observed geodetic deformation tensors and those inferred from active faults and earthquakes suggests that the Quaternary permanent deformation is essentially the result of the transfer of part of the subduction-induced elastic deformation into permanent plastic deformation. We then compute the elastic deformation of the Japanese Islands caused by interseismic loading of the Pacific and Philippine subduction planes. The geometry of the coupled zone and its downward extension are determined from the distribution of earthquakes for the Pacific slab. For the Philippine slab we use the geometry proposed by Hyndman et al. (1995). These elastic models account for most of the observed velocity field if the subduction movement of the Philippine Sea Plate is 100 per cent locked and if that of the Pacific Plate is 75-85 per cent locked. We note that the boundaries of the areas where significant elastic deformation is predicted (more than 10 mm yr(-1) of motion with respect to Eurasia) coincide with the main zones of permanent deformation: the Eastern Japan Sea deformation zone for the Pacific subduction elastic deformation field and the Setouchi/MTL deformation zone for the Nankai field. Each zone probably accommodates 10-15 mm yr-l of motion in the long term (convergence in the Eastern Japan Sea; strike-slip in the Setouchi/MTL zone), To account for this deformation, the effect of elastic loading from the trench must be combined with 5-10 mm yr(-1) of motion of the Amur Plate with respect to Eurasia. Because loading during the subduction earthquake cycle causes an increase in stress in the Eastern Japan Sea and Setouchi/MTL deformation zones, the probability of earthquake occurrence in these zones may be higher near the end of the cycle.