Horizontal surface velocity and strain patterns near thrust and normal faults during the earthquake cycle: The importance of viscoelastic relaxation in the lower crust and implications for interpreting geodetic data

Abstract

In recent years, more and more space-geodetic data on the surface deformation associated with earthquakes on intracontinental normal and thrust faults have become available. However, numerical models investigating the coseismic and postseismic deformation near such faults in a general way, i.e., not focused on a particular earthquake, are still sparse. Here we use three-dimensional finite element models that account for gravity, far-field ("regional") extension/shortening and postseismic relaxation in a viscoelastic lower crust to quantify the surface deformation caused by an M w~7 earthquake on a dip-slip fault. The coseismic deformation is characterized by horizontal shortening in the footwall of the normal fault and extension in the hanging wall of the thrust fault-consistent with elastic dislocation models, geological field observations, and GPS data from earthquakes in Italy and Taiwan. During the postseismic phase, domains of extensional and contractional strain exist next to each other near both fault types. The spatiotemporal evolution of these domains as well as the postseismic velocities and strain rates strongly depend on the viscosity of the lower crust. For viscosities of 10 18 -10 20 Pa s, the signal from postseismic relaxation is detectible for 20-50 years after the earthquake. If GPS data containing a postseismic relaxation signal are used to derive regional rates, the stations may show rates that are too high or too low or even an apparently wrong tectonic regime. By quantifying the postseismic deformation through space and time, our models help to interpret GPS data and to identify the most suitable locations for GPS stations.

Improving deformation models by discounting transient signals in geodetic data: 1. Concept and synthetic examples

“…Sometimes this correction is performed by modeling the postseismic signal as a logarithmic function of the time, in other cases by just ignoring the first weeks, or months, of GPS time series after an earthquake greater than a threshold magnitude (e.g., m ≥ 5.0) [ Riguzzi et al, ]. However, the effects of earthquakes that occurred before the survey began are often not considered [ Tong et al , ], even though signals from postseismic relaxation may be detectible for 20–50 years after an earthquake of moderate magnitude [ Hampel and Hetzel , ]. In the postseismic relaxation of a viscoelastic lower crust near normal and thrust faults, domains of extensional and contractional strain rate exist next to each other.…”

“…In the postseismic relaxation of a viscoelastic lower crust near normal and thrust faults, domains of extensional and contractional strain rate exist next to each other. If GPS velocities used to derive strain rates and long‐term tectonic moment rates contain postseismic relaxation signals, the resulting deformation pattern may show rates that are too high or too low or even imply the wrong tectonic regime [ Hampel and Hetzel , ]. In areas affected by moderate to large earthquakes, it is inappropriate to consider GPS velocities and related errors as representative of the long‐term behavior [ Freymueller et al , ].…”

Quantitative Prediction of Lower Order Faults Based on the Finite Element Method: A Case Study of the M35 Fault Block in the Western Hanliu Fault Zone in the Gaoyou Sag, East China

“…Finite element method is widely used in numerical simulations of the stress field, due to its accurate quantitative and discrete approximation, and it is playing an increasingly important role in oil and gas exploration and development and in identifying mechanisms of fault formation (Bertoluzza & Perotti, ; Candela et al, ; Hampel & Hetzel, ; Khodaverdian et al, ; Liu, Ding, Yang, et al, ). Numerical simulations of tectonic stress fields combine rock mechanics experimental data and finite element numerical calculations (Ghisetti, ; Takada et al, ).…”

Improving deformation models by discounting transient signals in geodetic data: 2. Geodetic data, stress directions, and long‐term strain rates in Italy

“…For example, postseismic relaxation following any m = 6.0–6.5 earthquake is expected to perturb the geodetic strain rates by a given percentage for a longer time in a slow‐deforming region like Italy than in fast‐deforming plate‐boundary regions. The consequence is a higher risk of inferring incorrect long‐term strain rates and kinematics, as explained by Hampel and Hetzel []. Coping with transients is thus essential, especially if long‐term earthquake rates are to be inferred from the resulting strain rate map.…”

“…Whereas the identification of stations showing anomalous contraction due to postseismic relaxation in the Basin and Range required a series of surveys covering a distance of 1000 km [ Hetland and Hager , ], available profile lengths in Italy are much shorter. Hampel and Hetzel , [] report this as a further barrier to detecting anomalous velocities and correcting long‐term strain rate calculations, especially across the Apennines.…”