Event Trees and Epistemic Uncertainty in Long‐Term Volcanic Hazard Assessment of Rift Volcanoes: The Example of Aluto (Central Ethiopia)

Aluto is a peralkaline rhyolitic caldera located in a highly populated area in central Ethiopia. Its postcaldera eruptive activity has mainly consisted of self‐similar, pumice‐cone‐building eruptions of varying size and vent location. These eruptions are explosive, generating hazardous phenomena that could impact proximal to distal areas from the vent. Volcanic hazard assessments in Ethiopia and the East African Rift are still limited in number. In this study, we develop an event tree model for Aluto volcano. The event tree is doubly useful: It facilitates the design of a conceptual model for the volcano and provides a framework to quantify volcanic hazard. We combine volcanological data from past and recent research at Aluto, and from a tool to objectively derive analog volcanoes (VOLCANS), to parameterize the event tree, including estimates of the substantial epistemic uncertainty. Results indicate that the probability of a silicic eruption in the next 50 years is highly uncertain, ranging from 2% to 35%. This epistemic uncertainty has a critical influence on event‐tree estimates for other volcanic events, like the probability of occurrence of pyroclastic density currents (PDCs) in the next 50 years. The 90% credible interval for the latter is 5–16%, considering only the epistemic uncertainty in conditional eruption size and PDC occurrence, but 2–23% when adding the epistemic uncertainty in the probability of eruption in 50 years. Despite some anticipated challenges, we envisage that our event tree could be translated to other rift volcanoes, making it an important tool to quantify volcanic hazard in Ethiopia and elsewhere. Key Points The first published event tree model to quantify volcanic hazard at an African volcano is presented for Aluto (Ethiopia) The event tree facilitates the design of a conceptual model for Aluto and provides a framework to quantify volcanic hazard The impact of epistemic uncertainty is explored by merging diverse data sets, from field data to expert elicitation and global databases