Analysis and automatic detection of lava flows using SAR backscatter applied to the 2017 eruption of Erta 'Ale Volcano, Ethiopia

Mapping lava flows from a range of conditions and environments is critical for understanding flow behaviour and hazard assessment, but ground-based observations can be challenging especially in remote regions. Here, we use 39 SAR backscatter images from the COSMO-SkyMed satellite to measure the extent and surface properties of lava flows from the 2017-2019 eruption of Erta ‘Ale volcano, Ethiopia. Using pairs of images, we map the flow outlines and observe both increases and decreases in backscatter which we attribute to differences in surface roughness. We apply a sequential analysis technique, CUSUM, and find that it can automatically identify flows from the background with 0.88 efficiency. However, simple lava flow simulations (Q-LavHA) could not reproduce flow extents with reasonable input parameters, highlighting the lack of appropriate calibrations. This stresses the importance of thorough volcanic observations and measurements to correctly determine lava flow parameters for use in volcanic modelling and hazard assessments.