Abstract: Since the past decade, geodetic techniques are widely used to gain important information for the monitoring and modeling of the deformation of the Earth at different length and time scales. Although the GNSS derived estimates of the Earth crust velocity are becoming more and more reliable, advanced data analysis techniques are needed to recognize geophysical features in the GNSS time-series, e. g., non linear behaviors, discontinuities in the signal and in its derivative, i.e., in the velocity. Unfortunately these phenomena are often hidden in the time-series noise and external information, as seismic events, are not always known. The main focus of this work is the detection of signal discontinuities in GNSS time-series through the use of advanced analysis techniques: the wavelets, the Bayesian and the variational methods. The Mumford and Shah (Commun Pure Appl Math 42:577-685, 1989) and the Blake and Zisserman (Visual reconstruction, 1987) variational models for signal segmentation can detect signal discontinuities in an explicitly way. The Blake and Zisserman (Visual reconstruction, 1987) model can also detect discontinuities of the signal first derivative, i.e., velocity abrupt changes can be detected. At first, to prove and assess the capability to detect discontinuities correctly, the methods have been applied to some Cascadia (North America) time-series, characterized by well known aseismic deformations. A second test area has been taken into account: the Calabrian Arc subduction zone, in southern Italy. The analyzed Italian GNSS time-series are characterized by very weak and noisy signals and the geodynamic of the area is mostly unknown. When present, discontinuities are expected to be very small and compatible with the signal noise. This motivates the use of advanced data analysis techniques to investigate the presence of discontinuities. At the moment, the analysis of the Italian time-series has revealed several discontinuities which nature cannot be labeled easily as geophysical or geodetic.
Advanced Techniques for Discontinuity Detection in GNSS Coordinate Time-Series. An Italian Case Study
BORGHI, ALESSANDRA;CANNIZZARO, LETIZIA;
2012-01-01
Abstract
Abstract: Since the past decade, geodetic techniques are widely used to gain important information for the monitoring and modeling of the deformation of the Earth at different length and time scales. Although the GNSS derived estimates of the Earth crust velocity are becoming more and more reliable, advanced data analysis techniques are needed to recognize geophysical features in the GNSS time-series, e. g., non linear behaviors, discontinuities in the signal and in its derivative, i.e., in the velocity. Unfortunately these phenomena are often hidden in the time-series noise and external information, as seismic events, are not always known. The main focus of this work is the detection of signal discontinuities in GNSS time-series through the use of advanced analysis techniques: the wavelets, the Bayesian and the variational methods. The Mumford and Shah (Commun Pure Appl Math 42:577-685, 1989) and the Blake and Zisserman (Visual reconstruction, 1987) variational models for signal segmentation can detect signal discontinuities in an explicitly way. The Blake and Zisserman (Visual reconstruction, 1987) model can also detect discontinuities of the signal first derivative, i.e., velocity abrupt changes can be detected. At first, to prove and assess the capability to detect discontinuities correctly, the methods have been applied to some Cascadia (North America) time-series, characterized by well known aseismic deformations. A second test area has been taken into account: the Calabrian Arc subduction zone, in southern Italy. The analyzed Italian GNSS time-series are characterized by very weak and noisy signals and the geodynamic of the area is mostly unknown. When present, discontinuities are expected to be very small and compatible with the signal noise. This motivates the use of advanced data analysis techniques to investigate the presence of discontinuities. At the moment, the analysis of the Italian time-series has revealed several discontinuities which nature cannot be labeled easily as geophysical or geodetic.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
