17:15 〜 19:15
[SCG61-P17] Function modeling of postseismic GNSS time series following the 2010 Mw 8.8 Chile Maule Earthquake
キーワード：余効変動、粘弾性緩和、余効すべり、GNSS時系列関数モデル
The time-series analysis based on Global Navigation Satellite System (GNSS) dataset is one of the key methods to understand the postseismic surface deformation after the megathrust earthquake.
Previous studies employed combinations of exponential and logarithmic decay functions to predict the GNSS time series following the megathrust earthquake (Tobita, 2016 EPS; Fujiwara et al., 2022 EPS; Freed et al., 2017 EPSL). However, this empirical combination of exponential and logarithmic functions may struggle to distinguish the dominating mechanisms for crustal deformations such as viscoelastic relaxation and afterslip. To address this issue, Dhar and Muto (2023 GRL) proposed a physic-based function model which constitute of two decay functions: one represents viscoelastic relaxation as power-law Burgers flow and another represents afterslip as velocity-strengthening frictional slip. These decay functions incorporate the laboratory-derived constitute laws which were used to predict the postseismic deformation model of the 2011 Tohoku-oki earthquake by mechanical modeling (Muto et al.., 2019 Sci. Adv.; Dhar et al., 2022 GJI; Agata et al., 2019 Nat. Commun.)
In this study, we employ the physics-based function model to analyze the postseismic GNSS time series following the 2010 Mw 8.8 Maule earthquake in Chile and evaluate the robustness of the model. By comparing our results with those from the Tohoku-oki earthquake (Dhar and Muto, 2023 GRL), we aim to identify key similarities and differences in the postseismic deformation between cold (Tohoku) and warm (Chile) subduction zones. Additionally, we compare our time-series model with a mechanical model that incorporates realistic subduction zone geometry with nonlinear rheology of rocks (Weiss et al., 2019 Sci. Adv.) to highlight their similarities and differences. Through this analysis, we seek to improve our understanding of mantle flow properties and the slip behavior along plate interfaces in different subduction zone.
The time series analysis based on the model proposed by Dhar and Muto (2023 GRL) predicts the 13-year time series of the postsemisic deformation of the 2010 Maule earthquake accurately. The mean residuals in the prediction were 0.20 cm/year, 0.42 cm/year, and 0.25 cm/year for the north–south, east–west, and up–down components, respectively. Generally, the accurate prediction of the vertical displacements has been difficult until now due to reasons such as rheological heterogeneity, while this study demonstrates a robust prediction even for the vertical displacements. Moreover, the close comparison between the time-series analysis and the postseismic deformation model of the 2010 Mw 8.8 Maule earthquake by Weiss et al. (2019 Sci. Adv.) suggests that the horizontal displacement by afterslip show similar patterns, but uplift and subsidence were reversed in the vertical displacement. On the other hand, horizontal displacement by viscoelastic relaxation by mechanical model (Weiss et al., 2019 Sci. Adv.) shows an overall westward motion, whereas our time series analysis shows a northeastward motion along the coast. These differences may be due to the absence of steady-state velocities associated with plate subduction in our model. This study further elaborates that the functional model can help forecast future ground motions and understand the deformation mechanisms related to megathrust earthquakes.
Previous studies employed combinations of exponential and logarithmic decay functions to predict the GNSS time series following the megathrust earthquake (Tobita, 2016 EPS; Fujiwara et al., 2022 EPS; Freed et al., 2017 EPSL). However, this empirical combination of exponential and logarithmic functions may struggle to distinguish the dominating mechanisms for crustal deformations such as viscoelastic relaxation and afterslip. To address this issue, Dhar and Muto (2023 GRL) proposed a physic-based function model which constitute of two decay functions: one represents viscoelastic relaxation as power-law Burgers flow and another represents afterslip as velocity-strengthening frictional slip. These decay functions incorporate the laboratory-derived constitute laws which were used to predict the postseismic deformation model of the 2011 Tohoku-oki earthquake by mechanical modeling (Muto et al.., 2019 Sci. Adv.; Dhar et al., 2022 GJI; Agata et al., 2019 Nat. Commun.)
In this study, we employ the physics-based function model to analyze the postseismic GNSS time series following the 2010 Mw 8.8 Maule earthquake in Chile and evaluate the robustness of the model. By comparing our results with those from the Tohoku-oki earthquake (Dhar and Muto, 2023 GRL), we aim to identify key similarities and differences in the postseismic deformation between cold (Tohoku) and warm (Chile) subduction zones. Additionally, we compare our time-series model with a mechanical model that incorporates realistic subduction zone geometry with nonlinear rheology of rocks (Weiss et al., 2019 Sci. Adv.) to highlight their similarities and differences. Through this analysis, we seek to improve our understanding of mantle flow properties and the slip behavior along plate interfaces in different subduction zone.
The time series analysis based on the model proposed by Dhar and Muto (2023 GRL) predicts the 13-year time series of the postsemisic deformation of the 2010 Maule earthquake accurately. The mean residuals in the prediction were 0.20 cm/year, 0.42 cm/year, and 0.25 cm/year for the north–south, east–west, and up–down components, respectively. Generally, the accurate prediction of the vertical displacements has been difficult until now due to reasons such as rheological heterogeneity, while this study demonstrates a robust prediction even for the vertical displacements. Moreover, the close comparison between the time-series analysis and the postseismic deformation model of the 2010 Mw 8.8 Maule earthquake by Weiss et al. (2019 Sci. Adv.) suggests that the horizontal displacement by afterslip show similar patterns, but uplift and subsidence were reversed in the vertical displacement. On the other hand, horizontal displacement by viscoelastic relaxation by mechanical model (Weiss et al., 2019 Sci. Adv.) shows an overall westward motion, whereas our time series analysis shows a northeastward motion along the coast. These differences may be due to the absence of steady-state velocities associated with plate subduction in our model. This study further elaborates that the functional model can help forecast future ground motions and understand the deformation mechanisms related to megathrust earthquakes.