Analysis of the coseismic slip behavior for the MW=9.1 2011 Tohoku-Oki earthquake from satellite GOCE vertical gravity gradient

Over the past decade, the three largest and most destructive earthquakes in recent history with associated tsunamis occurred: the Mw = 9.2 Sumatra-Andamam in 2004, then the Mw = 8.8 Maule in 2010, and finally the Mw = 9.1 Tohoku- Oki in 2011. Due to the technological and scientific developments achieved in recent decades, it has been possible to study and model these phenomena with unprecedented resolution and precision. In addition to the coseismic slip models, for which joint inversions of data from various sources are carried out (e.g., teleseismic data, GNSS, INSAR, and Tsunami, among others), depicting the space-time evolution of the rupture, we have high-resolution models of the degree of interseismic coupling (based on GNSS) and also maps of seismic b-value changes. Among these advances, new Earth gravity field models allow mapping densities distribution homogeneously and with a resolution (in wavelengths) of approximately the large rupture areas of great megathrust earthquakes. In this regard, the maximum resolution of GOCE-derived static models is in the order of lambda/2 asymptotic to 66 km, while GRACE monthly solutions are in the order of lambda/2 asymptotic to 300 km. From the study of the static and dynamic gravitational field, it has been possible to infer mass displacements associated with these events, which have been modeled and compared to the deformation inferred using other methods, yielding very good results. In this work we study the kinematic behavior of the rupture process for one of these largest events, the Mw = 9.1 Tohoku-Oki 2011 earthquake, employing the vertical gradient of gravity derived from the GOCE satellite, finding that the maximum slip occurred close to a lobe of minimum Tzz, as was observed for other case-studies in other subduction-related settings studied in previous works (e.g., the Maule earthquake and the Sumatra-Andaman earthquake, among others). In addition, from the rupture propagation using kinematic models, it can be observed that the rupture is arrested when it approaches high-density structures and, it is enhanced when connecting with lobes of low vertical gravity gradient. We also mapped a block expressed as a low Tzz lobe, developed along the marine forearc, which is controlled by a parallel-to-the-trench normal fault that accommodates subsidence during the interseismic period, as it is coupled with the subducted slab. Then, after rupturing the plate interface, this block is decoupled promoting tectonic inversion and uplift. In this way, the hypothesis that the density structure along the forearc is the ultimate first-order factor that governs the rupture process is reinforced.