Counterfactual Analysis of Runaway Volcanic Explosions

"Few, beside the student of such phenomena, will be able to realize what would have resulted from a more concentrated liberation of the accumulating energy if, for instance, instead of being separated into moderate earth movements and gas emissions spread over more than a 100 weeks of time, all this energy had been manifested in a single week? It is safe to say that, even if the island itself might still grace the map of the world, there would not have remained upon it so much as one human habitation." Report by Dr. Frank Perret to the Commissioner of Montserrat, 1935 (held in archives of the United Kingdom Foreign & Commonwealth Office). These words were written by the pre-eminent pioneering volcanologist Dr. Frank Perret, commenting on the earthquakes and volcanic unrest on the island of Montserrat in the period 1933-1935. Perret's way of thinking presaged the form of reasoning we identify now as "counterfactual analysis." One of the major challenges in volcano crisis decision-making arises when some significant eruptive activity has begun, and the threat potential exists of a massive sudden eruption. Even a small chance of this happening may warrant an urgent call for evacuation. For most active volcanoes, there is only a very limited geological and historical record upon which to base an estimate of the chance of a massive eruption. However, this database may be expanded by considering analog volcanoes, e.g., within a hierarchical model construct, or by stochastic modeling of past crises at the particular volcano of concern that had the potential for a dangerous event but did not ultimately result in a massive eruption. We adopt the latter approach and present a conceptual framework for the counterfactual analysis of runaway volcanic explosions, with reference to other extreme geohazards and georisks. This innovative type of probabilistic analysis has widespread application and is illustrated with the example of the well-documented 1997 Montserrat Vulcanian explosions sequence. An alternative possible mode of eruptive behavior might have substituted this sequence with fewer but larger explosive eruptions or even by a single runaway extreme event. This latter contingency was considered at the time, and motivated a brief, temporary evacuation of the building then housing the Montserrat Volcano Observatory; in response to the escalating violence of explosions, the base for observatory operations was subsequently re-located much further away from the volcano. Absent counterfactual thinking, it would be easy for volcanologists to fail to articulate fully, for decision makers, the possible scales and implications of plausible, potentially dangerous, future eruptive scenarios, thereby laying themselves open to charges of dereliction of duty or even professional negligence.