Geodetic inference on decadal afterslip following the 2011 Tohoku-oki earthquake

The postseismic deformation in the aftermath of the 2011 Tohoku-oki earthquake showed a stronger surface movement in northeastern Japan (NE) with subsequent decays over time. In response to the coseismic stress perturbation, afterslip on the megathrust interface is held responsible for the short-term deformation while viscoelastic relaxation in the surrounding lithosphere largely contributes to the long-term crustal deformation (e.g., Ozawa et al. 2012, JGR). On the contrary, decade-long studies on the postseismic model implied the prevalence of viscoelastic flow during the early phase of postseismic deformation (e.g., Sun et al. 2014, Nature, Watanabe et al. 2014, GRL, Freed et al. 2017, EPSL; Muto et al. 2016, GRL). Although geodetic displacement at any GNSS station may not indicate the single domination of either viscoelastic relaxation or afterslip over the longer period after the earthquake, the densely deployed nationwide GNSS observations (GEONET) till ~2021 provides a definite opportunity to resolve the contributions of various source mechanisms and their evolution over time.

 

Time series of geodetic observations are mainly explained using a numerical simulation of the source mechanisms (e.g., Agata et al. 2019 Nat. commun.; Luo & Wang 2021, Nat. Geosci.; Muto et al. 2019, Sci. Adv.; Fukuda & Johnson 2021, JGR) or non-linear regression of a fitting function (Tobita 2016, EPS). Utilizing the lesson learnt from the postseismic model built on laboratory-derived constitutive laws, we proposed an analytical fitting function for the GNSS time-series over the NE Japan. We deploy statistical approaches to ensure its stability and robustness. Our analytical function can be used to fit and predict the postseismic displacements at GNSS stations and understand the relative contributions of source mechanisms in lesser efforts.  We conclude that the afterslip at the downdip of the main rupture zone may continue for several decades following the megathrust earthquake. The decade-long records of repeating earthquakes on the plate boundary reiterate a similar conclusion concerning the longer persistence of afterslip in the Japan subduction zone (Igarashi & Kato 2021, Commun. Earth Env.; Uchida 2019, PEPS).

 

Our results also show that viscoelastic relaxation dominates immediately following the mega-earthquake at most inland GNSS stations. This conclusion can be supported by comparing the geodetic displacements with aftershock decay patterns (Morikami & Mitsui 2020, EPS), including recently developed stress-dependent postseismic deformation models (Agata et al. 2019, Nat. Commun; Fukuda & Johnson 2021, JGR; Muto et al. 2019, Sci. Adv).  Nevertheless, the previous studies indicate a change in the dominant mechanism of the postseismic deformation after the year ~2013-2015, particularly evident in the vertical motion (Morikami & Mitsui 2020, EPS; Yamaga & Mitsui 2019, GRL). We suggest that the transient deformation of the viscoelastic mantle decayed significantly during the ~3-4 years of the postseismic period, allowing the afterslip rate to supersede. The higher uplift rate along the Pacific coast of NE Japan, even after a decade, may reflect the shift in the dominant mechanism to the afterslip, persisting at the downdip of the main rupture zone.