Understanding Controls on Disaster Risk

Climate change, rapid urbanization, and socioeconomic inequalities exacerbate uncertainties in risk assessments by altering hazard intensities, exposure distributions, and vulnerability dynamics. Understanding the drivers of risk requires moving beyond static and siloed risk assessments to frameworks that capture the dynamic interactions between risk components. Sensitivity analysis helps identify which variables are most influential, providing the foundation for landslides and floods risk models that can adapt to the uncertainties. We aim to develop a conceptual framework for integrating sensitivity analysis into risk assessments to facilitate nuanced risk evaluations considering transient risk controls.

We will create a risk index by combining key factors representing hazard, exposure, and vulnerability. We will evaluate the non-linear relationships and complex interactions among these factors using machine learning models. We will quantify the contribution of each variable to the risk outcomes. We will test the proposed framework using high-resolution global datasets on flood and landslide hazards, population grids, building heights, and socioeconomic vulnerability. This procedure will enhance model interpretability and help determine the most influential drivers. The planned methodology should be scalable to other hazard types and urban contexts, providing a flexible approach for future risk assessments.

This work highlights the importance of rethinking disaster risk frameworks to inform more responsive and adaptive risk reduction strategies. By emphasizing risk sensitivity, our goal is to support evidence-based policymaking and resource allocation, strengthening preparedness and resilience in urbanizing landscapes.