Near-Surface Full-Waveform Inversion of DAS Data—A Case Study from an Alpine Landslide

Full-waveform inversion (FWI) is an advanced, albeit computationally demanding, technique that utilizes the complete seismic waveform to generate detailed reconstructions of subsurface velocity models. Traditionally, its application has relied on data acquired from seismographs, which are often limited by sparse acquisition geometries. Distributed acoustic sensing (DAS) presents a promising alternative, enabling seismic data acquisition with metre-scale spatial resolution along fibre-optic cables. At the same time, DAS introduces specific challenges: it records strain rather than particle displacement, averages deformation over a finite gauge length, and exhibits direction- and curvature-dependent sensitivity. In this study, we present the results of a 3-D elastic FWI tailored to DAS data acquired on the Cuolm da Vi landslide—one of the largest actively deforming alpine landslides. Three different initial velocity models, constructed using a combination of 2-D multichannel analysis of surface waves (MASW) and 3-D travel-time tomography, are tested to identify a suitable starting point for multiscale inversion. The resulting P- and S-wave velocity models exhibit detailed structural features and are consistent with independent geological observations, while achieving low data misfits. Our findings underscore the viability of applying elastic FWI to DAS datasets and offer practical guidance for future applications in similarly complex and data-limited settings.