9:30 AM - 9:45 AM
[SCG55-03] Understanding forearc deformation of northeastern Japan in megathrust earthquake cycle with nonlinear rheology
Keywords:Earthquake cycle, Power-law rheology, Postseismic deformation, 2011 Tohoku-oki earthquake, Northeastern Japan, GNSS observations
The geodetic observations during the last century indicated a significant rate of crustal subsidence (~0.5 cm/year) at the forearc of northeastern (NE) Japan before the 2011 Tohoku-oki earthquake (Nishimura, 2014 JDR and references therein). On the contrary, the forearc of NE Japan experienced a substantial uplift rate of ~10 cm/year in several years following the earthquake. Such interseismic subsidence and postseismic uplift of NE Japan forearc are the results of interactions between earthquake-induced processes (viscoelastic relaxation and afterslip) and interplate coupling, as hypothesized by several past studies (Nishimura, 2014 JDR; Hashima & Sato, 2017 EPS; Sagiya, 2015 GENAH; Sasajima et al., 2019 Sci. Rep.). However, the time-dependent changes of forearc crustal deformation over a megathrust earthquake cycle are still poorly understood. Recent studies highlighted that the nonlinear rheology of the Japanese mantle could largely contribute to the postseismic deformation of the 2011 Tohoku-oki earthquake (Agata et al., 2019 Nat. Commun.; Muto et al., 2019 Sci. Adv.; Fukuda & Johnson, 2021 JGR; Dhar et al., 2022 GJI). By incorporating such nonlinear mantle rheology and interplate coupling of Japan Trench, we developed a numerical model to simulate 3000 years of forearc crustal deformation over five megathrust earthquake cycles.
We used the observed GNSS displacement rate (provided by GEONET) at ~10 years before and after the Tohoku earthquake to constrain our model. The viscoelastic mantle and plate interface were discretized into volume elements and surface elements, respectively, to simulate the distributed stress responses based on analytical solutions (same used in Muto et al., 2019 Sci. Adv.). The viscoelastic relaxation of the Japanese mantle was modeled using power-law Burgers rheology, which was previously used for explaining ~5 years of postseismic deformation after the Tohoku earthquake (Agata et al., 2019 Nat. Commun.; Muto et al., 2019 Sci. Adv.; Dhar et al., 2019 GJI). A coseismic source model of the Tohoku earthquake (Iinuma et al., 2012 JGR) was introduced on the plate interface every 600 years, similar to the strategy of Sasajima et al. (2019 Sci. Rep.). The slip deficit (= plate convergence – afterslip, Sasajima et al., 2019 Sci. Rep.) on the plate interface during the earthquake cycle was modeled using elastic coupling (Nishimura, 2014 JDR) and stress-driven afterslip (rate-strengthening friction law, Muto et al., 2019 Sci. Adv.). The parameters related to the power-law Burgers rheology and friction law are optimized against the late interseismic (1997–2001) and early postseismic (2011–2014) periods along a latitudinal profile near the main rupture zone.
Our model successfully reproduces the observed GNSS velocities (both horizontal and vertical) at late interseismic and early postseismic periods. The effective viscosity of the mantle wedge is ~5×1018 Pa·s shortly after the earthquake and increases by ~1–2 orders of magnitude in the late interseismic period. The slip deficit rate is -60–100 cm/year in the early postseismic period (roughly consistent with Iinuma et al., 2016 Nat. Commun.) and +2.4–8.3 cm/year in the late interseismic period. Moreover, our model reproduces the long-term forearc uplift, consistent with the geological observations. Hence, our proposed model can help reconcile the mismatch between the long-term (~10000 years) forearc uplift and short-term (~100 years) subsidence rate at the forearc of NE Japan.
We used the observed GNSS displacement rate (provided by GEONET) at ~10 years before and after the Tohoku earthquake to constrain our model. The viscoelastic mantle and plate interface were discretized into volume elements and surface elements, respectively, to simulate the distributed stress responses based on analytical solutions (same used in Muto et al., 2019 Sci. Adv.). The viscoelastic relaxation of the Japanese mantle was modeled using power-law Burgers rheology, which was previously used for explaining ~5 years of postseismic deformation after the Tohoku earthquake (Agata et al., 2019 Nat. Commun.; Muto et al., 2019 Sci. Adv.; Dhar et al., 2019 GJI). A coseismic source model of the Tohoku earthquake (Iinuma et al., 2012 JGR) was introduced on the plate interface every 600 years, similar to the strategy of Sasajima et al. (2019 Sci. Rep.). The slip deficit (= plate convergence – afterslip, Sasajima et al., 2019 Sci. Rep.) on the plate interface during the earthquake cycle was modeled using elastic coupling (Nishimura, 2014 JDR) and stress-driven afterslip (rate-strengthening friction law, Muto et al., 2019 Sci. Adv.). The parameters related to the power-law Burgers rheology and friction law are optimized against the late interseismic (1997–2001) and early postseismic (2011–2014) periods along a latitudinal profile near the main rupture zone.
Our model successfully reproduces the observed GNSS velocities (both horizontal and vertical) at late interseismic and early postseismic periods. The effective viscosity of the mantle wedge is ~5×1018 Pa·s shortly after the earthquake and increases by ~1–2 orders of magnitude in the late interseismic period. The slip deficit rate is -60–100 cm/year in the early postseismic period (roughly consistent with Iinuma et al., 2016 Nat. Commun.) and +2.4–8.3 cm/year in the late interseismic period. Moreover, our model reproduces the long-term forearc uplift, consistent with the geological observations. Hence, our proposed model can help reconcile the mismatch between the long-term (~10000 years) forearc uplift and short-term (~100 years) subsidence rate at the forearc of NE Japan.