Post-seismic and Inter-seismic Deformation Associated With Long-rupture (~900 km) Great Subduction Earthquakes

Some great subduction earthquakes rupture very long segments of plate boundaries. The 1960 Chile (Mw=9.5) and 1964 Alaska (Mw=9.2) earthquakes both ruptured fault segments about 900 km long. The 1700 Cascadia (Mw~9) earthquake was inferred to have invovled most of the ~1000 km long margin. These earthquakes are expected to induce prolonged post-sesimic deformation. In both Chile and Alaska, GPS measurements indicate that although coastal sites are moving landward as expected near a locked subduction fault, inland sites are moving in the opposite direction. The seaward motion of the inland sites is interpreted to be a delayed response to the previous great earthquake. Rupture during the earthquake instantaneously stretches the forearc seaward and induces a static elastic shear stress in the deeper part of the fault and the upper mantle. Subsequent stress relaxation allows the inland areas to move seaward to "catch up" with the coseismic motion while the coastal sites are already moving landward due to the locking of the fault. The observed deformation pattern yields information on the viscosity of subdbuction zone upper mantle. Using a mantle viscosity of 3x10¹⁹ Pa s in a 3-D viscoelastic finite element model, we can explain the GPS-observed Chile postseismic deformation. A similar model is developed for Alaska. Using a similar viscosity for the Cascadia margin, crustal deformation observed at present, 300 years after the previous great earthquake, can be well explained. In the present model, the velocity-strengthening behavior of the aseismic part of the fault is approximated using a thin viscoelastic layer, but the long-term post-seismic and inter-seismic deformation is not sensitive to details of the short-term (a few years) behavior of the fault.