Spatial and temporal analysis of ground deformation data for the characterization of natural and anthropogenic sources

Ground displacement measurements are fundamental for investigating the surface effects of numerous natural and anthropogenic processes acting within the same region. Spatial geodesy measures the displacement of the ground due to the sum of multi-scale processes, i.e. processes that occur at different spatial and temporal scales. The joint action of these phenomena can generate surface deformations characterized by constant trends or transients over time or even by cyclical variations that generate seasonal signals in the displacement time series, with an annual or multi-annual period. Separating the contribution of each phenomenon in the displacement measurements is a complicated objective to achieve because it is necessary to identify within the GNSS and InSAR time series the signals associated with the various processes and to have a large amount of information relating to the geological, geophysical and hydrological characteristics.

The target area of this work (coastal area of the Po Plain, Italy) is affected by various processes of natural and anthropogenic origin, such as the subsoil water pumping, the compaction of sediments throughout the plain area, the hydrocarbon cultivation at the numerous onshore and offshore active concessions, and also the active tectonic process linked to the convergence between the Northern Apennines and the Adriatic plate.

Aim of this work is to develop a systematic method of analysis both at regional and local scales of the GNSS and InSAR displacement time series using signal decomposition techniques to identify the main ongoing deformation processes. Extracted signals are compared with the time series of all available physical, hydrological, geophysical and geological parameters to identify the main deformation sources causing the observed displacements. 

In particular, considering the differences in lengths and temporal samplings among the datasets, all the measurements have been standardized in the same formats through an open-source code, allowing for the comparison among the different types of data to investigate any associations and correlations, and executing also a data quality analysis. Furthermore, a Matlab-based code has been developed to quickly and automatically analyze the InSAR displacement time series. The code provides information on linear, non-linear, cyclic and/or seasonal components, by using frequency analysis (spectral analysis via Lomb-Scargle periodogram to evaluate most significant components and their periodicity), and by means of the estimate the Non-Linearity Index (INL), defined as the ratio between the long-term signal variability and the high-frequency noise variability. Such a code is general and could be applied to several areas of interest.