Integrating Eco-DRR into landslide susceptibility assessment: The critical role of eco-environmental factors

Understanding the factors driving landslide susceptibility is essential for improving risk assessment and disaster management. Traditional assessments often emphasize structural factors such as topography and geology, while overlooking eco-environmental variables. In this case study from western Rwanda, we propose a multidimensional landslide susceptibility assessment framework grounded in Ecosystem-based Disaster Risk Reduction (Eco-DRR) principles, using a Random Forest model. The framework integrates 60 variables across nine interrelated dimensions, including structural (topography, geology, hydrology) and eco-environmental factors (soil health parameters, vegetation indices, landscape composition, and temporal dynamics). We use six model configurations combining different sets of these dimensions to assess their contribution to model performance. Our results show that models integrating both structural and eco-environmental factors achieve higher accuracy (up to 93 %) than those using only structural factors (83 %). Key predictors included traditional factors like slopes, alongside eco-environmental variables such as soil moisture or land use transition. These findings highlight the value of incorporating Eco-DRR principles in landslide susceptibility assessment by identifying key eco-environmental indicators that improve predictive accuracy. This provides actionable insights for decision-makers to design targeted ecosystem-based interventions. We provide quantitative evidence supporting recent conceptual frameworks that emphasize the importance of eco-environmental factors in landslide processes. By demonstrating the added predictive value of a multi-dimensional approach, this study provides a strong empirical foundation for enhancing disaster prevention, landscape management, and the practical implementation of Eco-DRR strategies in landslide-prone areas in landslide-prone regions like Rwanda.