Problems of the Joint Inversion of Temporal Gravity Variations with the Data on Land and Seafloor Displacements: a Case Study of the Tohoku-Oki Earthquake of March 11, 2011

The methodology for the joint inversion of various types of data on the Tohoku-Oki earthquake is presented. It is shown that satellite geodetic (GPS) data, results of SAR interferometry, and coseismic GRACE gravity anomalies are overall fairly consistent with each other. However, the resolution of these data is not high and the rupture surface model based on them is smoothed. Meanwhile, the measured and calculated data agree well, which creates the illusion that the resulting solution is close to the real process. The solution of the inverse problem using also the ocean bottom station displacements shows that the rupture surface models based on the measurement data on land and/or GRACE gravity anomalies alone may significantly differ from the real coseismic displacement field. If the rupture surface is specified inaccurately, then the high quality of data fitting and the checkerboard test do not guarantee highly accurate solution. A similar situation is also observed in estimating the degree of locking along the plate contact within locked segments of the subduction zone. In the paper it is underlined that the accuracy of the inverse problem solution mostly depends on how closely the geodynamic model fits the real process. For increasing the accuracy and degree of detail of the solution, the physics of the coseismic and post-seismic process need to be more adequately described in parallel with developing a more detailed description of the geological structure of a rupture area with allowance for the discontinuities of the medium and a more detailed description of the distribution of the physical parameters.