An approach for Canadian river hydrokinetic resource assessment with synthetic aperture radar satellite images

• Hydrokinetic Resource Assessment with SAR Satellites (HyRASS) is introduced. • One classifier identifies whitewater river sections in summer imagery. • One classifier identifies river ice and open water in winter imagery. • Whitewater, ice, and water combined to identify persistent, stable, high-flow water. • Persistent, stable, high-flow water is required for hydrokinetic turbine deployment. This paper describes a novel approach that uses Synthetic Aperture Radar (SAR) satellite imagery to identify river sections with persistent, stable, high-flow water. Such sections represent candidate sites for the generation of green electricity through the deployment of hydrokinetic turbines. Canada-wide mapping of potential hydrokinetic turbine sites requires Earth Observation satellite technology. The application of SAR satellites is advantageous because of their capacity to acquire high-quality images independent of weather and light conditions. Our SAR image approach includes two classifiers. The first identifies whitewater river sections in images acquired during summer. The second uses images acquired during winter to identify river ice and open water. Both classifiers are developed using a machine-learning algorithm that identifies classification thresholds in a multidimensional space made up of advanced SAR image parameters. The whitewater classifier makes use of backscatter and image texture parameters to achieve 92.6% accuracy. The ice–water classifier makes use of backscatter in two polarizations and a polarimetric decomposition to achieve 89.0% accuracy. In combination, the classes identified enable the mapping of river sections with persistent, stable, high-flow water, i.e., candidate hydrokinetic turbine sites. Our SAR image approach is named HyRASS—Hydrokinetic Resource Assessment with SAR Satellites.