Inferring 2D Local Surface-Deformation Velocities Based on PSI Analysis of Sentinel-1 Data: A Case Study of Oraefajokull, Iceland

Two-dimensional deformation estimates derived from Persistent Scatterer Interferometric (PSI) analysis of Synthetic Aperture Radar (SAR) data can improve the characterisation of spatially and temporally varying deformation processes of Earth's surface. In this study, we examine the applicability of Persistent Scatterer (PS) Line-Of-Sight (LOS) estimates in providing two-dimensional deformation information, focusing on the retrieval of the local surface-movement processes. Two Sentinel-1 image stacks, ascending and descending, acquired from 2015 to 2018, were analysed based on a single master interferometric approach. First, Interferometric SAR (InSAR) deformation signals were corrected for divergent plate spreading and the Glacial Isostatic Adjustment (GIA) signals. To constrain errors due to rasterisation and interpolation of the pointwise deformation estimates, we applied a vector-based decomposition approach to solve the system of linear equations, resulting in 2D vertical and horizontal surface-deformation velocities at the PSs. We propose, herein, a two-step decomposition procedure that incorporates the Projected Local Incidence Angle (PLIA) to solve for the potential slope-deformation velocity. Our derived 2D velocities reveal spatially detailed movement patterns of the active Svinafellsjokull slope, which agree well with the independent GPS time-series measurements available for this area.

Automated summary

In this study, two-dimensional deformation estimates derived from Persistent Scatterer Interferometric (PSI) analysis of Synthetic Aperture Radar (SAR) data were examined for their improved characterization of spatially and temporally varying deformation processes of Earth's surface. By analyzing Sentinel-1 SAR image data from 2015 to 2018, the study obtained 2D vertical and horizontal surface deformation velocities, which showed good agreement with independent GPS measurements.