The Analysis And Mechanism Research Of Postseismic Deformation Based On GPS Coordinate Time Series

Since the 1990 s,GPS has been widely used in many fields of geodesy and geodynamics due to its high-precision,large-scale,and all-weather characteristics,revealing many geophysical phenomena that are difficult to recognize by other means.The position changes observed by GPS fully capture the crustal deformation over the entire earthquake cycle,including the interseismic deformation caused by the crustal stress accumulation,the coseismic deformation caused by the sudden rupture on the fault,and the postseismic deformation of the crust and upper mantle gradually returning to a steady state.Therefore,GPS has become one of the main monitoring methods of crustal deformation.In recent years,due to the wide establishment of GPS continuous stations,the time series with high time resolution captures the clear postseismic deformation process,which has attracted the attention of many researchers.On the one hand,from the perspective of GPS time series analysis,detecting the postseismic deformation in the time series and establishing a suitable model are important and inevitable in the future GPS time series analysis work,for the sake of whether the maintenance of the reference frame or the establishment of the mathematical model of the GPS time series.On the other hand,from the perspective of the physical mechanisms behind the postseismic deformation,the spatiotemporal evolution of postseismic deformation observed by GPS time series directly provides the insight into the physical state of the lithosphere.With the constaint by GPS observed postseismic deformation,we can studies the postseismic mechanisms such as the afterslip and viscoelastic relaxation,and look into the mechanical properties of the fault zone and lithospheric rheology.The processing and analysis of postseismic deformation based on GPS time series are the basis,while the postseismic mechanisms are the physical explanation of the former,the two promote the insigt of each other.The research on GPS time series data processing and the underlying physical mechanism for postseismic deformation is systematic and has important research significance.Therefore,the main work of this paper is focused on the processing and analysis of GPS postseismic deformation time series and the investigation of postseismic deformation mechanisms.Based on the GPS data processing and time series analysis,this article focuses on postseismic deformation extraction,and writes a set of GPS time series processing software for the postseismic deformation analysis.On this basis,the time and space characteristics of postseismic deformation are then analyzed.In addinational,taking the 2015 Mw7.8 Gorkha earthquake in Nepal as an example,the postseismic deformation mechanism was studied with the extracted GPS deformation constrained.The specific content and results include the following aspects:(1)Preprocessing is the basis of GPS time series analysis and an important part of the procedure,including gross error detection,spatial filtering,and gap interpolation.The procedure is written to detect and eliminate gross errors using the sliding window method;integrates three spatial filtering methods including the stacking,principal component analysis and Karhunen-Loeve transform,and comparations among these methods are made;Integrates an interrpolation method based on model and spatial correlation in order to avoid the influence of local signals,(2)Due to the wide application of GPS continuous stations,more and more stations are affected by earthquakes.Screening and correctly modeling the affected stations has become an important task of time series analysis.In order to avoid the complicated work of manual screening and possible omissions,this paper proposes a comprehensive testing method to automatically detect the stations which are affected by coseismic and postseismic deformations,and proves the reliability and efficiency of this method through multiple earthquake examples.In addition,comprehensively consider the characteristics of the data and the spatial correlation of postseismic deformation,this paper proposes an iterative PCA parameter estimation method,which is proved to be more stable and reliable than single station modeling or two-step PCA modeling w ith 1000 sets of data synthesized by the Monte Carlo method.(3)According to 37 earthquakes distributed around the world,uses the procedure to automatically detect the stations affected by the earthquake and invoke the iterative PCA method to estimate the model parameters.Statistics of the decay constants of different earthquakes and different models reveal that: the decay constants of different earthquakes are different and have no obvious relationship with factors such as epicenter,magnitude,depth,etc.;analysis of the time characteristics of the decay constant manifests that the longer the time after earthquake,the longer the observation time window involved in the calculation,the larger the e estimated decay constant.(4)At present,postseismic deformation in Nepal is mainly based on elastic earth layered model or elastic half-space model inversion for afterslip,and layered model or adding viscoelastic layers in the middle and lower crust beneath the Tibet Plateau to simulate the postseismic viscoelastic relaxation.Many evidences such as the receive function,geodetic inversion,resistivity,and temperature profile indicate the inhomogeneous properities beneath the Tibet Plateau,and the properties of the media control how fault dislocations are transmitted to the surface deformation.Different from previous studies,this paper establishes 3D finite element model that considers topographic undulation,earth curvature and anisotropic media to probe the afterslip and viscoelastic relaxation after the Nepal earthquake.The results show that the postseismic deformation observed by GPS does not match the direction and magnitude of the viscoelastic relaxation,which can be better explained by the afterslip occurring below the rupture zone.The time evolution of the posseismic deformation shows that the postseismic deformation gets slower as time.Afterslip has been dominant during the 4.8 years after the Gorkha earthquake.The magnitude of viscoelastic relaxation is small but its contribution to the postseismic deformation slightly increases.Due to the complexity of anisotropic rocks,it is difficult to obtain an analytical solution for the pore rebound of the anisotropic model.This paper differences coseismic deformation of different Poisson's ratios under the elastic model to estimate the pore rebound,and finds the pore rebound is relatively small and contributes little to the postseismic deformation.(5)Through the comparisons of different models,evaluate the influence of terrain undulation,curvature,anisotropy,rupture model on viscoelastic relaxation,afterslip and pore rebound.The study found that: terrain and earth curvature have relatively small effects on viscoelastic relaxation and after-slip;the viscosity model has a large impact on viscoelastic relaxation and cannot be ignored;The Burgers body with dual viscoelastic properties and the Maxwell body have the basically same postseismic relaxation pattern but big difference in magnitude.Burgers body is a superposition of exponentially decayed Kelvin body and linearly increasing Maxwell body,which relaxes faster and greater after an earthquake;Rupture models is an important factor affecting viscoelastic relaxation and pore rebound,since it direct determines the magnitude and pattern of the viscoelastic relaxation and pore rebound.