Long-range volcanic risk assessments are needed in many areas to protect people and facilities from harm. These risk assessments are guided by past patterns in volcanic activity, which are often used to determine a qualitative range of likely future hazards. The geologic and historic record of past eruptions, however, is often incomplete and rarely accounts for smaller events. Risk assessments using only this imperfect record are biased toward larger eruptions and provide an inaccurate basis to make public health and safety decisions. Models for volcanic processes can be used to overcome this limitation and calculate a likely range of future hazards that includes an appropriate number of smaller events. By sampling a range of parameter values, the models also can determine uncertainty in the hazards analysis and provide a clear basis to evaluate differences in expert judgment.

Due to large uncertainties in future eruption conditions, semi-empirical process models using geologic parameters appear more amenable to hazard assessments than complex thermo-fluid dynamical models using first-order properties. Site-specific ranges for parameters such as eruption volume, hillslope, or wind speed can be readily determined through direct measurement or expert judgment, in contrast to more generalized thermodynamic and fluid flow parameters. Once bounded, parameter ranges are sampled in a large number of model realizations using stochastic methods. The calculated results provide a statistically significant population of conditional hazards. Probabilities of the initiating eruption can be combined with the hazard frequency and associated societal impact to fully quantify a range of risk. This range allows planning, mitigation, or evacuation decisions to be made at levels of confidence determined by the affected society, using risk metrics comparable to other important cost-benefit decisions (e.g., likelihood of fatality per year). Semi-empirical models for lava and debris flows, and tephra falls, currently are available and can be adapted to use a probabilistic risk assessment methodology.

Work supported by U.S. Nuclear Regulatory Commission (Contract NRC-02-97-009) and is an independent product of CNWRA that does not necessarily reflect NRC views or regulatory positions.