Abstract
The Global Positioning System (GPS) is widely used for measuring crustal movements from varieties of geophysical origins. Nonlinear movement in the observations draws increasing attention owing to the improved measurement precision. We developed an innovative method capable of detecting nonlinear motion from the observation time series. By implementing this new technique, we identified several nonlinear episodic events in the noisy time series that were ignored previously. Two types of nonlinear motion with different mechanisms were presented in this dissertation. In Greenland, the nonlinear motion is presented by accelerating uplift in the vertical GPS time series. The accelerating uplift of the continental crust is caused by accelerating melting of the ice cap near the stations. Another form of nonlinear motion expressed by the slow slip events usually last for several weeks. The displacements in the time series are caused by slow reverse movement to the plate convergence direction on the subduction fault plane. We identified a series of such events in Costa Rica and modeled the surface displacement data. We conclude that the slow slip events in this area will release seismic energy accumulated on the fault plane in a nondestructive way, thus reduce the seismic hazards in Costa Rica.