Interpretation of Crustal Deformation following the 2011 Tohoku-oki Megathrust Earthquake by the Combined Effect of Afterslip and Viscoelastic Stress Relaxation

Abstract:

Crustal deformation following the 2011 Tohoku-oki megathrust earthquake, occurred at the North American-Pacific plate interface, has been revealed by GPS measurement on land (Geospatial Information Authority of Japan) and GPS/Acoustic measurement on seafloor (Japan Coast Guard). The essential causes of the postseismic crustal deformation are considered to be slow afterslip at the downdip extension of the main rupture zone and viscoelastic relaxation of stress changes induced in the asthenosphere. Crustal responses to the afterslip and the viscoelastic relaxation are different in both space and time. So, given proper plate interface geometry and proper crust-mantle rheological structure, we can estimate unbiased spatiotemporal distribution of afterslip through the inversion analysis of observed geodetic data. In the present analysis, we used a 3-D realistic model developed by Hashimoto et al. (2004) for plate interface geometry and a standard elastic-viscoelastic layered model, consisting of a 60 km-thick elastic surface layer and a Maxwell-type viscoelastic substratum with the viscosity of 10¹⁹ Pa s, for crust-mantle rheological structure. First, following Noda et al. (2013), we transformed the GPS displacement data on land into the average strains of triangular elements composed of adjacent three GPS stations. Then, by applying a sequential method of stepwise (every two months) inversion to the strain data, we estimated the spatiotemporal distribution of afterslip together with coseismic slip distribution. The estimated results show that significant afterslip has proceeded for the first one and a half years at the downdip extension of the main rupture zone off Iwate and Miyagi with decaying its rate. Finally, based on the estimated results, we computed postseismic offshore crustal movements by using the same elastic-viscoelastic structure model and compared them with seafloor geodetic observations (Watanabe et al., 2014). The good agreement between the computed crustal movements and the observations in both spatial pattern and transient behavior demonstrates the validity of our inversion results for coseismic slip and afterslip distributions.