Reservoir monitoring using satellite SAR and GNSS: a case study in southern Italy

Abstract

Recently, it has been demonstrated that it is possible to relate water levels of a reservoir with its dam displacements. Water levels were determined via remote sensing, while dam displacements were measured via Global Navigation Satellite System (GNSS). Results have shown that displacements and water levels are correlated. Water levels at the Magazzolo reservoir in southern Italy were firstly retrieved using two remote sensing approaches: by visual matching between the reservoir shoreline and contour lines, and by evaluating the surface extent via unsupervised classification to estimate the water levels with an area/depth relation. Dam displacements were measured using GPS receivers on the dam and a permanent station from a GNSS Continuously Operating Reference Stations (CORS) network, about 30 kilometers away. Subsequently, two other remote sensing approaches were tested to detect reservoir levels; the first based on shape similarity indices, while the second on the evaluation of the average distance between a reservoir shoreline and contour levels. First results were extracted from a Landsat 8 optical image acquired during a clear sky day. Within this work, algorithms for water level retrieval have been tested and validated under different conditions over a more consistent satellite dataset including Sentinel-1A Synthetic Aperture Radar (SAR) images acquired from October 2014 to September 2015. The dataset is also used to analyses dam displacements via Interferometric SAR (InSAR), to be compared with the effects of water level fluctuations on the dam. First results suggest that it is possible to correlate dam displacements and water levels derived by the same dataset. However, it is shown that displacements also depend on meteorological forcing