Probabilistic Inversion of tTEM for Improved Mapping of Landslide-Associated Glaciotectonic Structures

Recent studies have identified approximately 3,200 – primarily slow-moving - landslides across Denmark. Many of the larger landslides are related to areas with glaciotectonic thrusting of plastic clays. GEUS is currently mapping selected landslide areas to improve the understanding of the geological factors driving slope instability. Here, we focus on data from the Røsnæs peninsula, where landslides in a residential area show movements in the order of 0.5–12 cm per year. The Røsnæs area have been mapped with tTEM data. In this presentation, we will compare different inversion strategies including the industrial standard SCI inversions with different constraints and a probabilistic inversion workflow, developed in the research project ‘INTEGRATE’’. The probabilistic inversion is performed using a set of geological rules as prior information and a regional resistivity-lithology relationship. The probabilistic framework generates an ensemble of geologically and geophysically consistent models, providing robust uncertainty quantification that supports subsequent interpretation and decision-making. By applying the probabilistic inversion workflow, we managed to map and model glaciotectonic thrust structures in agreement with the directions of the glacial lineations in the surface terrain.