The use of belief-based probabilistic methods in volcanology: Scientists' views and implications for risk assessments

This paper constitutes a philosophical and social scientific study of expert elicitation in the assessment and management of volcanic risk on Montserrat during the 1995–present volcanic activity. It outlines the broader context of subjective probabilistic methods and then uses a mixed-method approach to analyse the use of these methods in volcanic crises. Data from a global survey of volcanologists regarding the use of statistical methods in hazard assessment are presented. Detailed qualitative data from Montserrat are then discussed, particularly concerning the expert elicitation procedure that was pioneered during the eruptions. These data are analysed and conclusions about the use of these methods in volcanology are drawn. The paper finds that while many volcanologists are open to the use of these methods, there are still some concerns, which are similar to the concerns encountered in the literature on probabilistic and determinist approaches to seismic hazard analysis.

► We show that volcanologists are divided in their acceptance of probabilistic methods.
► We analyse the use of expert elicitation on Montserrat.
► We present data from a survey, interviews and archival analysis on this topic.
► We conclude that these methods are very useful for policy purposes.