10:15 AM - 10:30 AM
[SGD03-06] Revealing the Earthquake History of Kikai-jima Based on Numerical Simulation of Marine Terrace Formation
Keywords:Kikai-jima, Marine Terrace, History of Earthquakes, Viscoelastic Response, Raised Coral Reef
In the Ryukyu Arc, the subduction of the Philippine Sea Plate raises concerns about the occurrence of large earthquakes. However, due to the lack of clear seismic traces, evaluating seismic potential remains challenging. The Holocene marine terrace of Kikai-jima have garnered attention as traces of past earthquakes, and geological investigations using coral fossils from these terraces are ongoing. Kikai-jima has experienced continuous uplift at an average uplift rate of 2.1 ~ 2.3 mm/yr over the past 100,000 years due to the subduction of the Amami Plateau (Inagaki & Omura, 2006). In addition to these unique crustal deformations, marine processes, such as wave erosion and biological activity, play significant roles in marine terrace formation, making the island’s crustal deformation history still uncertain. To address this issue, this study aims to reconstruct the tectonic deformation history of Kikai-jima by numerical simulating the formation of its four-step marine terrace system, integrating both seismological and geological perspectives.
In this study, we compare marine terrace surfaces generated by numerical simulation with actual topographic data and optimize the parameters using a nonlinear search approach. First, we developed a landform evolution model that considers the effects of coastal processes on landform formation in addition to mantle viscoelastic response and plate subduction-driven crustal deformation. For coastal processes, we adopted the methods of Shikakura (2014), and for viscoelastic response, we employed the approach of Ito et al. (2015). Using the finite element method (FEM), we simulated the spatiotemporal evolution of the landscape over the past 10,000 years on an annual timescale. Furthermore, we conducted parameter studies and optimization using seven key parameters: steady-state uplift rate associated with plate subduction, maximum coral growth rate, maximum erosion rate due to wave action, mantle viscosity, and the timing of three major earthquakes, to determine the optimal values.
Our optimization analysis suggests that it is possible to estimate both the timing and recurrence interval of earthquakes around Kikai-jima. Additionally, we found that the magnitude of past earthquakes necessary for marine terrace formation can also be inferred. Traditionally, time-predictable recurrence models (Shimazaki & Nakata, 1980) have been derived based on the height of marine terrace cliffs and paleoshoreline elevations. However, our results indicate that when considering the subsurface structure of Kikai-jima, existing models may not always be directly applicable. This finding highlights the need for reassessment and potential revision of conventional time-predictable recurrence models.
In this study, we compare marine terrace surfaces generated by numerical simulation with actual topographic data and optimize the parameters using a nonlinear search approach. First, we developed a landform evolution model that considers the effects of coastal processes on landform formation in addition to mantle viscoelastic response and plate subduction-driven crustal deformation. For coastal processes, we adopted the methods of Shikakura (2014), and for viscoelastic response, we employed the approach of Ito et al. (2015). Using the finite element method (FEM), we simulated the spatiotemporal evolution of the landscape over the past 10,000 years on an annual timescale. Furthermore, we conducted parameter studies and optimization using seven key parameters: steady-state uplift rate associated with plate subduction, maximum coral growth rate, maximum erosion rate due to wave action, mantle viscosity, and the timing of three major earthquakes, to determine the optimal values.
Our optimization analysis suggests that it is possible to estimate both the timing and recurrence interval of earthquakes around Kikai-jima. Additionally, we found that the magnitude of past earthquakes necessary for marine terrace formation can also be inferred. Traditionally, time-predictable recurrence models (Shimazaki & Nakata, 1980) have been derived based on the height of marine terrace cliffs and paleoshoreline elevations. However, our results indicate that when considering the subsurface structure of Kikai-jima, existing models may not always be directly applicable. This finding highlights the need for reassessment and potential revision of conventional time-predictable recurrence models.