Decision-making on Population Return At the End of an Eruption

Abstract:

If an eruption sequence is protracted and episodic, continuing for many months, or even a number of years, it can be very difficult for authorities to decide when an evacuated population might be allowed to return. A precautionary approach may be appropriate as a general principle, especially if the societal implications are deemed to be manageable. However, where the livelihoods of the evacuees are critically at stake, such as with farmers, and the economic costs of continuing evacuation are high, a risk-informed approach to supporting decision-making might be warranted. Such an approach has been developed, based on concepts drawn from cost-benefit analysis.

On any given day during an extended volcanic crisis, a decision-maker may wish to consider whether to allow then for the return of evacuees. A delay of a decision by one day will increase the evacuation cost by one day’s worth of accommodation and economic disruption. On the other hand, there is a small chance that there might be an onset to another eruption on this given day, so there would be a safety benefit in the delay.

The task of quantifying this safety benefit requires a coherent synthesis of all the volcano monitoring data time series, together with an assessment of the medium and long-term historical and geological data, so that the probability of each of the alternative eruptive scenarios can be scientifically estimated. Bayesian belief networks provide a convenient framework for this synthesis. For each scenario, the potential casualty implications for returned evacuees need to be assessed.

The application of this end of eruption methodology is illustrated and compared with the equivalent decision-making procedure at the start of an eruption.