Interpretation of postseismic deformation with a viscoelastic relaxation model

Abstract

Aleutian tide gages registered coseismic subsidence of 10–15 cm at three locations during the great earthquakes of 1957 and 1965. For approximately 5–6 years after this initial subsidence the data indicate postseismic uplift of the islands to approximately their initial level. Similar postseismic uplift in other island arcs has previously been interpreted as due to fault creep along the deep end of the rupture plane and along extensions of it. Since the lithosphere below island arcs and above deep earthquakes has anomalously low Q values, it must be modeled as a volume of anomalously low viscosity. This anomalous volume is located adjacent to the deep end of great subduction zone earthquake ruptures. By modeling the lithosphere and asthenosphere as elastic solids, with an anomalous viscoelastic inclusion below the island arc volcanoes, we find that the observed postseismic surface deformation is a corollary of the known island arc structure. A satisfactory fit to the uplift following the 1957 earthquake is found for a viscoelastic volume with 80‐km width extending from 50‐ to 200‐km depth. Similarly fitting the post‐seismic uplift following the 1964 Alaskan earthquake, we find an anomalous volume northwest of Anchorage with a width of 40 km and a depth extending from 20 to 80 km. These smaller dimensions suggest that the viscous volume pinches out near the northern end of the 1964 rupture, where the chain of volcanoes ends. The viscosity of the anomalous material is estimated to be from 1.2×1019 to 2.2×1019 P. We propose that the assumption of fault creep on deep and sometimes distant extensions of the rupture plane is a less‐plausible mechanism for postseismic surface deformation than the viscoelastic relaxation model.