Hazards SEES: Persistent volcanic crises resilience in the face of prolonged and uncertain risk

Unlike many geological hazards, volcanic eruptions are almost unique in being preceded by weeks or months of clear precursors, which offer the potential for successful early interventions to reduce risk. The impacts of eruptions should therefore be, in principle, amongst the easiest to mitigate of all the great natural hazards. Nevertheless, impacts of volcanic disasters are often out of proportion to the magnitude of the eruptions. Eyjafjallajökull (Iceland), for example, resulted in a cost of $5 billion due to the cancelation of 180,000 aircraft flights during April and May 2010. Volcanic crises carry a host of scientific, engineering, mathematical, and societal problems, along with very high levels of uncertainty. The long-term nature of volcanic unrest, the multi-hazard character of eruptions, major uncertainty in data and models, and the diverse responses of decision makers and other stakeholders, make volcanic hazards uniquely complex amongst natural disasters. This project is developing new understanding of the eruptive record, volcanic plumbing, and the social response and resilience to volcanic threats at Kilauea (Hawaii) and Long Valley (California). Lava from Kilauea currently threatens a town of 950 residents and may soon isolate a large community of 10,500. Heightened seismic activity (including as recent as September 2014) and gas discharges at Long Valley suggest possible magma movement in the crust. Eruptions there can be highly explosive, with tephra plumes greater than 20 km high, and highly destructive pyroclastic density currents. Kilauea and Long Valley represent the spectrum of volcanic crises that pertain to the United States. The prolonged nature of unrest (eruptive or not) in both locations has resulted in mixed social responses, including complacency and denial. This project is developing new and enhanced computational models of eruptive processes to provide predictive geophysical simulations. We are integrating data and models through innovative hierarchical Bayesian statistical modeling in order to fully incorporate and quantify uncertainties. Our social science research is characterizing in greater detail stakeholders' understanding of, and preparedness for, the two types of unrest and volcanic systems, and the mental models that affect how they use scientific information. We are also developing and testing for effectiveness different methods of portraying hazard probabilities and uncertainties to stakeholders, as informed by our social science results. Results from this project will be broadly transferable to other volcanic hazards and will, more generally, inform societal resilience to other natural hazards with long lead times.