Abstract
A consistent and comprehensive understanding of risk related to multiple natural hazards threatening population, property and infrastructure is a pillar of the Sendai Framework for Disaster Risk Reduction, and lays at the base of most risk mitigation initiatives. This is particularly challenging since the relevant hazards (e.g., earthquakes, floods or landslides) are often different in terms of recurrence periods, spatial footprint and intensity. Also, the analysis and assessment of risk is based in turn on the knowledge of factors other from hazard, namely exposure and vulnerability. Exposure refers to the communities, assets and systems that are exposed to the hazards and susceptible to damage. This information should be available at the spatial scale and resolution that is most suitable to the considered hazards, and be up-to-date (or reasonably recent to ensure representativeness). Exposure is also tightly connected to the chosen vulnerability modelling approach, hence proving a critical component in the risk assessment pipeline to grant reliable estimates. While gathering a satisfactory exposure dataset from authoritative sources is usually difficult, the task might prove unfeasible in many economically developing countries, where the sheer amount of technical and economical resources needed to collect and maintain such information (e.g., population and housing census, transportation infrastructure, etc.) might exceed the capacity of the local institutions, especially in countries undergoing rapid changes due to urbanization processes. In the last decade, fortunately, several international and global projects made available a significant wealth of data, free of cost and often at regional and global scale. This has proven invaluable to carry out small- and large-scale risk assessment with unprecedented resolution and geographical coverage. However, often the specific characteristics of these datasets and their inherent limitations are not easy to be taken into consideration, therefore paving the way to unwanted biases in the resulting estimates.
In our contribution we will report on the activities carried out to develop exposure information for multi-hazard risk assessment in Burundi, focusing on the integration of multiple information sources for population distribution and road infrastructure, and on their use in probabilistic landslides risk assessment. The limitations and perspectives of the use of open data and tools will be presented and discussed, along with the role of transparent and reproducible research in using, updating and sharing such data.
The described research activities have been carried out within the framework of an international project funded by the International Organization of Migration (IOM) and coordinated by IDOM (Spain).