Comprehensive multi-hazard risk assessment in data-scarce regions – a study focused on Burundi

Abstract. The increased occurrence of multiple cascading and compounding hazards underlines the importance of integrated- and multi-hazard-based assessment approaches for the development of thorough strategies towards disaster resilience. To this purpose, a national-scale multi-hazard risk assessment was conducted between September 2020 and December 2021 for Burundi, focusing on the natural hazards flooding, torrential rains, landslides, earthquakes, and strong winds. This integrated multi-hazard assessment resulted in comparable nationwide provincial and commune-scale Annual Average Loss (AAL) values, further aggregated to provide a preliminary estimate of the resulting overall risk. Historical climatology (1990–2019) was computed, and a preliminary evaluation of the potential effects of climate change in the future period (2020–2049) was carried out. Data availability and reliability were challenging throughout the whole assessment and were tackled by integrating local authoritative sources with international and global resources. An up-to-date exposure model was implemented and complemented by an indicator-based socioeconomic vulnerability assessment. Furthermore, a data-driven statistical susceptibility model for shallow landslides has been derived at national scale. The consequent multi-hazard risk assessment provides an approximate picture of the expected nationwide risk distribution in economic terms. The results should support the identification of priority areas and actions for disaster risk management. From a research perspective, the paper provides a transparent, hazard-wise framework that harmonises heterogeneous hazard, exposure and vulnerability information into comparable risk metrics in data-scarce environments. From a practice perspective, the resulting risk estimates are intended as a pragmatic baseline for multi-hazard comparison and prioritisation, supporting the identification of areas where DRR efforts and data improvements may be most impactful.