Assessing Railway Exposure to Rapid Flow-Like Landslides: A National-Scale Methodology

Geo-hydrological hazards, particularly rapid flow-like landslides, present a critical challenge for transportation infrastructures globally. These phenomena pose severe risks due to their ability to propagate rapidly and cause extensive damage to railway tracks, vehicles, and human life. Climate change exacerbates these risks by intensifying precipitation patterns, further increasing landslide frequency and impact.

This study introduces an innovative methodology for assessing the exposure of railway infrastructure to rapid flow-like landslides on a national scale [1]. Applying this methodology to Italy's extensive railway network, we integrate statistical and conceptual models, utilizing digital elevation models (DEMs) and landslide inventories to identify landslide source areas, simulate runout paths, and evaluate exposure. The results yield susceptibility and exposure maps that highlight vulnerable railway segments and provide a foundation for risk mitigation and resource allocation.

The methodology involves distinguishing between hillslope and channelized landslides, each with unique source area characteristics and propagation behaviors. Channelized landslides, often occurring within confined channels, exhibit longer runout distances and lower reach angles compared to hillslope phenomena, which are more dispersed and occur on open slopes. This distinction allows for tailored modeling approaches to improve the accuracy of predictions. Validation using an independent landslide dataset confirmed the model's robustness, achieving Area Under the Receiver Operating Characteristic (AUROC) curve values between 0.7 and 0.95 in most regions, demonstrating its effectiveness for large-scale assessments. However, in areas where model performance was lower, biases in the validation dataset, such as inconsistent landslide classifications or incomplete coverage, were often identified as contributing factors.

Key findings indicate that approximately 20.1% of the Italian railway network exhibits exposure values exceeding 0.5, with 13.4% classified as highly exposed (exposure >0.75) to rapid flow-like landslides. Regions such as Trentino-Alto Adige, Campania, and Sicily are particularly affected due to their geomorphological and climatic conditions. This highlights the urgent need for targeted interventions to safeguard critical infrastructure and minimize disruptions to transportation services.

The study emphasizes the utility of high-quality landslide inventories and DEMs in developing predictive models applicable at national scales. The outputs enable stakeholders to prioritize interventions, such as reinforcing vulnerable railway segments, implementing early warning systems, and optimizing maintenance schedules. These measures not only mitigate immediate risks but also contribute to long-term infrastructure resilience. Furthermore, the methodology’s adaptability makes it applicable to other linear infrastructures and regions facing similar hazards, showcasing its potential for broader implementation.

Marchesini et al., Eng. Geol. 332 (2024) https://doi.org/10.1016/j.enggeo.2024.107474