Dynamic Modeling Of Postseismic Deformation Following The 2015 Mw 7.8 Nepal Earthquake Using Finite Element Method

Due to the frictional coupling between plates,part of the subduction energy is stored as elastic strain.Eventually the increasing stress reaches some threshold associated with the fault strength and an earthquake occurs.The underground structure will produce viscoelastic deformation response because the stress disturbance of the upper crust and lower mantle around the coseismic rupture caused by the large earthquake,which will lead to the long-term crustal post-seismic deformation following the earthquake.At present,it is recognized that the main dynamic processes controlling the post-seismic deformation include the afterslip on the fault plane,the viscoelastic relaxation of regional lower crust and upper mantle,and the poroelastic rebound.On the contrary,we can further understand the dynamic process by studying post-seismic deformation,and then constrain the regional lithospheric rheological structure and fault plane frictional characteristics.On 25 April 2015,the Mw7.8 Nepal earthquake occurred on the Himalayan main thrust fault zone in the southern Tibet Plateau.It was the largest earthquake in the Himalayan region in the past 80 years and caused a large number of casualties.A cluster of aftershocks followed the mainshock,including the largest Mw 7.3 event which occurred 17 days later beyond the eastern end of the mainshock rupture.After the earthquake,a number of GPS stations in Nepal and southern Tibet recorded significant post-seismic deformation,which provides a good opportunity for us to understand the post-seismic dynamic process and to explore the lithospheric rheological structure in the southern Tibet Plateau.Based on the finite element method,this thesis adopts GPS observation data 5 years after the earthquake to model and analyze its post-seismic dynamics process,and mainly includes the following works:(1)We collected the data of continuous station in the Nepal,and added 7 new campaign station data in the southern Tibet.GAMIT 10.7 was used to process GPS phase data to obtain position time series in the ITRF2014.Then,the post-seismic deformation in the first 5 years after the Nepal earthquake was isolated by removing the irrelevant signals such as long-term linear rate,seasonal variations due to annual and semi-annual hydrological loading and offsets caused by equipment changes or earthquakes.(2)Based on the flat-ramp-flat fault geometry,the subduction zone FEM model was constructed and grid meshing was performed.In order to more accurately model the post-seismic deformation,we re-estimated the coseismic slip of the mainshock and the largest aftershock(Mw 7.3)by inverting the GPS and InSAR data.To validate the coseismic elastic behavior in our finite element model,we compared the predicted surface displacements with coseismic observation.(3)Viscoelastic relaxation model.The Maxwell body was used to represent the upper mantle of the India and the Tibet Plateau,and the viscosity coefficient was fixed to be 1×1020 Pa s.The Burgers body was used to represent the lower crust of the Tibet Plateau.Firstly,the upper crust thickness(We)and the lower crust steady-state viscosity(?m)were constrained using the observations of 5 years after the earthquake,then the We is fixed as 25 km,and we used the observed data of 2-5 years after the earthquake to constrain the ?m.We find that the single viscoelastic relaxation model cannot explain the GPS observed deformation(4)Stress-driven afterslip model.Rate-strengthening friction law was added in the PyLith software code to model the afterslip.To initiate the afterslip evolution,the stress perturbations due to the mainshock and the largest Mw7.3 aftershock were firstly calculated using the PyLith code.Assuming an uniform normal stress of 100 MPa,we constrained two parameters:the initial slip rate v0 and friction parameters a?.The best-fit to the horizontal GPS time series is found when v0=300 mm/yr and a?=0.25 MPa.The results show that afterslip mainly occur on the downdip of the coseismic rupture and possible stable-sliding region between the rupture zones of the mainshock and the largest aftershock.The modeled deformation can explain the near-field observed data well,but underestimate the far-field observations.(5)Combined mechanism model.Considering the combined effects of viscoelastic relaxation and post-seismic afterslip,we developed a combined model including three parameters:the steady state viscosity of Tibet's lower crust,reference slip rate and frictional parameter.We obtain the optimal model by evaluating the minimum misfit between the observed and predicted time series using a simple grid search algorithm.Through a global grid search,we found ?m=3 × 1018 Pa s,v0=150 mm/yr,a?=0.15 MPa.The simulated postseismic displacements of the combined model are well consistent with the observed data on the aspect of temporal evolution and spatial distribution.We concluded that the afterslip has a significant effects in near-to intermediate-field postseismic deformation in the initial 4 years after the Gorkha earthquake,in the following periods,viscoelastic relaxation is the dominant long-term contributor.This work systematically modeled and tested the post-seismic deformation processes of the Nepal earthquake,which is of great significance for further understanding the post-seismic dynamic mechanism,the seismicity of the Himalayan region and the uplift of the Tibet Plateau.