Item

Abstract

Time series of interferometric SAR (InSAR) images offer the potential to detect and monitor surface deformation with high spatial resolution, even for slow deformation processes. Current InSAR time series processing techniques are however often biased towards an assumption of linear deformation. When estimating the different contributions to the interferometric phase, which is necessary to seperate deformation from other influences, but also to determine the noise level and subsequently the stability and reliability of a pixel, either linear or spatially correlated deformation is assumed to model the signal. Since deformation fields can display complex behavior in time and space, a simple linear assumption can lead to a disregard of pixels with nonlinear deformation or removal of the nonlinear parts of the signal during processing. We aim to improve on previous work by analyzing the signal throughout all Interferograms with a principal component analysis to differentiate various types of deformation before modeling the signal and determining the stability of a pixel. We test and compare our approach with established methods in terms of time coherent pixel density and stability on Sentinel-1 InSAR data from 2015 to 2022 above the storage cavern field Epe in North Rhine Westphalia, Germany. Epe displays a spatially and temporally complex surface deformation field consisting of linear, as well as seasonal and cavern pressure dependent contributions as shown in previous studies of the area. The reliability of the time series results is verified with data from GNSS stations and levelling campaigns in the study area.