Physical mechanism-driven extraction model for post-seismic deformation recorded by continuous GNSS observations

Abstract

Post-seismic deformation caused by large earthquakes is one of the most effective parameters for investigating the frictional behavior of faults and the rheological characteristics of the lithosphere. This study introduces a modified functional model that accounts for both afterslip and viscoelastic relaxation effects, incorporates a physical mechanism, and effectively disentangles distinct mechanisms of post-seismic deformation captured by continuous GNSS observations.A parameter optimization solution method based on the Bayesian framework was adopted to obtain optimal estimates and corresponding confidence intervals. The effectiveness of proposed model was validated through simulation experiments. The proposed method was applied to study post-seismic deformations from the 2015 Nepal MW 7.8 earthquake, and the different characteristics of post-seismic deformations caused by different mechanisms were discovered at different stations. The afterslip effect has the duration of 2.67 years, whereas the viscoelastic relaxation effect will continue for 150.8 years. Moreover, the calculated equivalent viscosity coefficient in the southern Tibet region was 1.1 ± 0.3 × 1018 Pa·s. Our proposed model can more accurately and objectively separate different mechanisms from the post-seismic displacements observed by GNSS than previous one, and thus improve the accuracy of fault kinematics and regional lithosphere rheology investigations.