Climate Assessment of rainfall-induced Landslide Hazard and Risk: Assessing Past Simulations and Future Projections

Beyond their immediate environmental impact, landslides pose significant social and economic challenges for vulnerable communities. These highly dynamic natural hazards are mostly triggered by rainfall in many regions worldwide, making it crucial to understand the relationship between precipitation and landslide occurrence. This understanding is key to enhancing the accuracy and reliability of systems that assess and mitigate landslide risks.

To explore this relationship, this study employs a framework similar to that of Berenguer et al. (2015) and Palau et al. (2020) for real-time application. The system integrates landslide susceptibility information with precipitation inputs to generate maps with a qualitative classification of the warning level in four classes. For this analysis, this system combines 3-hourly accumulated precipitation simulations from the EURO-CORDEX dataset (with a resolution of 12.5 km x 12.5 km) and the European Landslide Susceptibility Map (200 m x 200 m), developed by Wilde et al. (2018) and Günther et al. (2014). By merging the fine spatial resolution of the susceptibility dataset with the temporal resolution of precipitation data, the system provides a dynamic representation of landslide hazards that accounts for local susceptibility and precipitation variability.

The framework is designed for application at various scales, from the European level to specific regions. For this study, Catalonia (NE Spain) is the focus area due to the availability of a landslide inventory, which allows for initial validation of the system's preliminary results. Although the inventory has some limitations—such as incompleteness and biases towards events near transportation networks and urban areas—it offers valuable data for validating the framework and identifying its strengths and weaknesses. A historical precipitation dataset (1976–2005) serves as input for simulating past landslide hazards, laying the groundwork for analyzing long-term trends. By comparing simulated precipitation-induced landslides with reported events, insights into the relationship between precipitation patterns and landslide occurrences.

To assess future risks, climate projections from 2011 to 2100 across various timeframes and scenarios are analyzed. Temporal variations in precipitation are examined on monthly and seasonal scales to understand how shifting precipitation patterns may affect landslide-prone conditions in specific regions. This methodology can be improved by incorporating socio-economic risk indicators, such as population density, infrastructure vulnerability, and the economic value of exposed assets. This integration helps transition the focus from hazard assessment to risk analysis, reflecting the potential severity of consequences. Such analysis could help in developing proactive risk management strategies, providing valuable insights into the future dynamics of landslide hazards under changing climatic scenarios.