Identification and Analysis of Recurringly Occluded Persistent Scatterers, with Application to Displacement Monitoring in the Oetztal Alps

The Persistent Scatterer Interferometry (PSI) is a powerful remote sensing technique for Earth surface displacement monitoring with potential sub-millimeter accuracy. It is based on identifying and analyzing point scatterers which provide coherent backscatter over time. Standard PSI approaches require these Persistent Scatterers (PSs) to be stable throughout the whole monitoring period. However, PSs can appear or disappear over time, due to construction work, natural decorrelation, or temporary occlusion by floods or snow. The integration of such Temporary PSs (TPSs) is needed to establish an optimum measurement point network. While a few TPS-integrating PSI algorithms have been proposed before, these were exclusively evaluated at construction sites. Here, we evaluate the capabilities of an existing algorithm to identify and analyze recurringly occluded PSs (ROPSs), which might be seasonally snow-covered or flooded. We show, based on simulated data, that a newly proposed CuSum algorithm outperforms the approach implemented in the tested algorithm at identifying the coherent time series segments of ROPSs. The displacement rate estimation becomes increasingly instable with increasing occlusion periods. Inadequately sampled seasonal movements can also lead to false displacement estimations, which must be identified and discarded. Finally, we present results from a Sentinel-1 experiment over the seasonally snow-covered Oetztal Alps. Integrating ROPSs significantly increases the measurement point density at many locations across the study area, compared to the European Ground Motion Service. While the algorithm did not identify coherent segments in each summer for most of the ROPS, their integration still leads to a significant information gain.