Assessing volcanic hazard and exposure in a data poor context: Case study for Ethiopia, Kenya, and Cabo Verde

Volcanoes produce a wide variety of hazards across varying spatial and temporal scales. When data are scarce on past eruptions and hazards, it can falsely imply low hazard recurrence and create challenges for robust hazard and risk assessment. Data quality and quantity vary considerably across different regions, volcanoes, and eruptions. Yet, there is a need for regional to global scale information on volcanic hazard and risk, where consistent and reproducible methods are applied. Such information is used by international stakeholders to inform funding priorities, risk reduction policies, and to highlight data and knowledge gaps, contributing towards the Sendai Framework's Sustainable Development Goals. Challenges in gathering this information can be most problematic where large populations are exposed to potential volcanic hazards but there are few comprehensive eruptive histories, as in sub-Saharan Africa. Here, we present a unique study to evaluate hazard and exposure for nine volcanoes in Ethiopia, Kenya and Cabo Verde, as part of an international project to develop multi-hazard Disaster Risk Country Profiles. We applied a two-stage expert elicitation process to volcanoes for the first time, and coupled the results with vent mapping, numerical hazard modelling, and GIS analysis of eight exposure categories to identify where high volcanic hazard and exposure coincide. Testing the sensitivity of our findings to input assumptions, to better understand where uncertainties lay, showed that improving our knowledge of past eruption volumes, frequencies, and dates was key to reducing uncertainty. Expert elicitations proposed that Fogo, Cabo Verde, is the most likely to erupt (eruption on average every 25 years), while Fentale (Ethiopia), Longonot and Suswa (Kenya) were elicited to have the greatest probability for a large explosive (VEI ≥ 4) eruption (on average every 400 years). Menengai and Longonot produce the larger exposure values across most VEI scenarios and categories of exposure, but population and GDP exposure was also large for more distal tephra fall and flows at Corbetti and Suswa, with order of magnitude increases expected between 2010 estimates and 2050 projections. Potentially high impact scenarios include tephra being dispersed across large cities (e.g. Nairobi, 55 km from Suswa) and key infrastructure (e.g. geothermal power station ∼2.5 km from Aluto), as well as important tourist destinations, seats of government and emergency management operations (e.g. islands east of Fogo). This study provided the first hazard and exposure assessment of its kind for these volcanoes and drew attention to volcanic risk at the levels required to inform policy and future in-country funding opportunities.