Can decision theory help end-users take the appropriate action in an emergency?

Harvey, N. J. ORCID: https://orcid.org/0000-0003-0973-5794, Western, L. M., Dacre, H. F. ORCID: https://orcid.org/0000-0003-4328-9126 and Capponi, A. (2022) Can decision theory help end-users take the appropriate action in an emergency? Bulletin of the American Meteorological Society, 103 (10). E2176-E2187. ISSN 1520-0477 doi: 10.1175/bams-d-21-0258.1

Abstract/Summary

Making decisions about the appropriate action to take when presented with uncertain information is difficult, particularly in an emergency response situation. Decision makers can be influenced by factors such as how information is framed, their risk sensitivity and the impact of false alarms. Uncertainty arising from limited knowledge of the current state or future outcome of an event is unavoidable when making decisions. However, there is no universally agreed method on the design and presentation of uncertainty information. The aim of this article is to demonstrate that decision theory can be applied to an ensemble of plausible realisations of a situation to build a transparent framework which can then be used to determine the optimal action by assigning losses to different decision outcomes. The optimal action is then visualized, enabling the uncertainty information to be presented in a condensed manner suitable for decision makers. The losses are adaptable depending on the hazard and the individual operational model of the decision maker. To illustrate this approach, decision theory will be applied to an ensemble of volcanic ash simulations used for the purpose of airline flight planning, focussing on the 2019 eruption of Russian volcano Raikoke. Three idealised scenarios are constructed to show the impact of different loss models on the optimal action. In all cases, applying decision theory can significantly alter the regions, and therefore potential flight tracks, identified as potentially hazardous. Thus we show that different end users would and should make different decisions when presented with the same probabilistic information based on their individual user requirements.

Item Type	Article
URI	https://reading-clone.eprints-hosting.org/id/eprint/106773
Identification Number/DOI	10.1175/bams-d-21-0258.1
Refereed	Yes
Divisions	Science > School of Mathematical, Physical and Computational Sciences > Department of Meteorology
Uncontrolled Keywords	Atmospheric Science
Publisher	American Meteorological Society
Download/View statistics	View download statistics for this item