Near real-time monitoring of volcanic deformation and lava flows from multi-angular Capella Space SAR imagery
摘要
Monitoring deformation and surface changes is crucial for rapid hazard assessment during volcanic eruptions. This can be achieved by detecting changes in synthetic aperture radar (SAR) images. However, most traditional methods require SAR images acquired with an identical observation geometry, limiting the temporal resolution to the repeat cycle of the satellite/constellation (typically a few days). Here, we present a novel method based on a single Capella Space high-resolution SAR amplitude image acquired with a viewing geometry that is not necessarily known in advance. We compare this image to a high-resolution digital elevation model (DEM) acquired months or years before the eruption. The SAR image is correlated against a synthetic image generated from knowledge of the SAR sensor geometry and the DEM. We show that the correlation score allows to map surface changes, such as new lava flows. Moreover, the estimated shifts between the real and synthetic images provide two independent components of the ground displacement. We combine multiple images of arbitrarily different geometries to retrieve the three components of displacement. We apply the method to the Piton de la Fournaise volcano (La Réunion island) and validate our results with “ground truth” data including daily-resolved lava flow maps and Global Navigation Satellite System (GNSS) displacements. The method achieves a precision of