Strategies for PS processing of large sentinel-1 datasets

Several advanced DInSAR techniques have been used to map surface deformations due to volcanism, active tectonics, landslides, subsidence, and uplift as well as to monitor the deformation of critical infrastructure such as bridges and dams. Recently, studies have explored the potential of these techniques to be integrated into a permanently operating monitoring system. ESA's Sentinel-1 satellites have been providing SAR images for such a purpose since 2014. Nowadays, it is easy to access more than 230 SAR images of any area of interest, and update this dataset every six days with a new image. Due to the high frequency of the data acquisition, the question arises on how to best handle such a dataset. Is it suitable to always consider the whole available dataset or would a partial processing of the dataset and combining the results at a later point be more appropriate? To answer these questions, three different processing strategies are investigated in this paper. The first is a continuously growing dataset and for the second and third strategy, the dataset was divided into sub-stacks with and without overlap. In this study, the key parameters of each strategy are analyzed. In addition, the size of the sub-stacks is varied and the results are compared.