Following a review of recent developments by social theorists, concerned with the policy and practice of natural hazard reduction it is argued that neither the so-called dominant nor the radical perspectives, are sufficient on their own to answer the questions now being addressed by applied volcanologists during the International Decade for Natural Disaster Reduction. In order to make useful recommendations to policy-makers, which are both reliable assessments of eruption risk and at the same time sensitive to probable interactions of future eruptions with the physical environment of the study region and its peoples, a conflation of traditional hazard analysis and vulnerability analysis is required.

This view is tested by means of a case study of evacuation planning at Furnas Volcano in the Açores. As well as supporting the argument from theory, this example leads to the proposition that research on evacuation planning and civil defence should proceed in an ordered sequence. First hazard analysis (i.e. hazard mapping) should be carried out in the traditional manner. Next possible interactions with the physical environment should be specified and a preliminary plan of evacuation routes drawn up. Before detailed recommendations about the logistics of evacuation are made, however, it is vital that such factors as the demographic, socio/economic and cultural/behavioural characteristics of the population at risk are fully considered.​