Identification of Tectonic Signals in GPS Positional Time Series

During the last decades, the development of space geodesy resulted in an increased amount and quality of crustal deformation observations. In particular, instruments of the Global Positioning System (GPS) provide positional time series constituted by crustal displacement signals carrying information about different geophysical processes. For instance, signals associated to tectonic processes that occur during the different stages of the seismic cycle. The aim of this work is to efficiently separate out and characterize the different signals present in GPS time series with the ultimate goal of obtaining precise crustal surface observations that can be used to better understand the geophysical processes occurring during the seismic cycle. We use a Basis Pursue scheme in which we identify and separate the different signals in the positional time series from a vast library of ad-hoc basis functions. We use an optimization algorithm (SPGL1, Van der Berg & Friedlander, 2008) that allows for a sparse representation of the positional time series by selecting the minimum amount of basis functions that are needed to represent such signal. We apply the methodology to GPS observations from the Chilean National Seismological Center to separate and analyze the tectonic and non tectonic signals present in such observations.