|Institution:||The Ohio State University|
|Keywords:||Geophysics; GPS; time series; Geodetic; accelerating trend; trajectory model; GNET; CAP|
|Full text PDF:||http://rave.ohiolink.edu/etdc/view?acc_num=osu1276727414|
Geodesists and geophysicists engaged in crustal motion geodesy monitor the position (or displacement) time series associated with thousands of GPS stations worldwide. These time series are useful for studying a wide range of geodynamic phenomena including plate motion, mountain building, the earthquake deformation cycle, postglacial rebound, and environmental loading. Station coordinate time series are expressed in a spatial reference frame which is typically a global, earth-centered, earth-fixed (ECEF) reference frame. The motion of a station in a given reference frame can be referred to as the trajectory of that station. The great majority of station trajectory models in use within the geodetic community are linear models, which consist of three component or sub-models characterizing: (i) the trend of displacement over time, (ii) jumps or discontinuities in the time series, and (iii) annual oscillations. In this thesis, we use a constant velocity model for most trends, but a polynomial function for trends with time-varying velocity, Heavyside function to implement jumps if and when jumps are required, and a truncated Fourier series, typically composed of just annual and semi-annual terms to compose our standard linear trajectory model. We first illustrate the use of a polynomial trend model with reference to a GPS station of our GNET project which is well known to have a time-varying velocity, particularly in the vertical component. We then consider an original problem: quantifying time-changes in the velocity of a station COYQ near Coyhaique in southern Chile from CAP project which manifests postseismic transient deformation in the aftermath of the great 1960 Chile earthquake. We have shown that station trajectory models in which the secular trend of displacement can be represented as a polynomial function of time can be very useful for modeling GPS time series obtained in areas undergoing accelerating ice loss, and in areas undergoing postseismic transient deformation a decade or more after a great earthquake. Of course the great majority of GPS stations can be characterized perfectly adequately using a constant velocity trend model. But the areas in which this is not true are areas of considerable geodynamic interest. This thesis presents a new tool for studying those areas.