Application of a nested modelling approach for the analysis of landslide-tunnel systems

The interaction between landslides and transportation infrastructure represents a critical challenge in mountainous regions worldwide, with significant implications for public safety, regional connectivity, and socio-economic sustainability. Professional geoscientists and engineers often take advantage of numerical simulations to investigate the correlation between landslide movement and structural response. However, one of the main challenges of the analysis of landslide-infrastructure systems (LIS) is related to the difficulty of reliably simulating localized infrastructure damage induced by large-scale landslide processes within computationally feasible frameworks. In this work, we develop a nested numerical modelling approach and explore its capability to investigate a real case study, namely the Passo della Morte LIS. The Passo della Morte landslide is a 30-million-cubic-meter active rock-block slide (moving 2-3 cm/year) crossed by a 2200-meter-long road tunnel that, since its construction, began showing evidence of damage and deformation at the intersection with the inferred landslide basal surface. In the developed approach, a landslide-scale model, built using the finite-difference code FLAC3D, is used to capture bulk slope behavior and stress distributions. Custom Python and FISH scripts are then used to extract velocity fields and stress tensors from the landslide-scale model in order to transfer them to a smaller, high-resolution infrastructure-scale model, built using the distinct element code 3DEC. The multi-scale numerical analysis shows that the tunnel model develops roto-translational deformation patterns consistent with field observations, landslide movement patterns, and monitoring data. This work demonstrates that numerical simulations can achieve computational efficiency and physical accuracy achieved when nested approaches are used. It also shows that multi-scale simulations represent a solid pathway for advancing LIS interaction analysis toward mature frameworks, supporting evidence-based infrastructure management in landslide-prone regions.