U007-04
Evacuation during pandemic: Competing priorities during compounded crises

Wednesday, 9 December 2020: 11:00
Bill Fry, GNS Science, Lower Hutt, New Zealand and Laura S L Kong, UNESCO IOC - NOAA, Honolulu, HI, United States
Abstract:
The recent establishment of social distancing as a safety measure for epidemic crisis calls into question the utility of conservatism in the establishment and execution of tsunami evacuation guidelines. Typically, evacuation decisions are underpinned by the pragmatic approach of mitigating uncertainty in tsunami forecasts by liberal application of evacuation orders in borderline cases. In effect, it is usually better to over-evacuate than under-evacuate.

Of course, coastlines need to be evacuated if there is an obvious land threat (commonly defined as >1m tsunami height at the coast). The problem arises when >1m is not obvious, but possible. This situation typically forces a judgement call based on a previously considered set of uncertainties and conservatism that go into threat maps. There is a knife-edge balance for many borderline evacuation decisions and judicious conservatism yields a large likelihood of unnecessary evacuations. In the Covid19 reality, this knife-edge balance has shifted and encourages a rethink of tsunami decision criteria when competing life-safety demands are present, such as during a pandemic, when evacuation could produce significant fatalities.

We have the opportunity prior to the onset of future tsunami to assess embedded conservatism in response rules. In some situations, this could be manifest as a delay of warning issuance for borderline cases, allowing uncertainties in forecasts to diminish through analysis of additional data. In extremely dangerous pandemics, an adjustment of threat bands might be warranted. This could be accomplished by applying a decision tree in which estimated fatalities from evacuation due to the compounding hazard (increasing due to evacuation) and tsunami impacts (diminishing due to evacuation) are compared to estimated tsunami fatalities without an evacuation. While this seems obvious, adequate treatment of scientific uncertainties prior-to and during a response is not trivial. This requires establishment of a framework in which to apply this logic prior to the response.