On probability in risk analysis of natural disasters

Abstract

PurposeThe purpose of this paper is to show how the common practice of applying the frequency interpretation of probability in risk analysis of so‐called low‐probability and high‐consequence disasters can prove to be flawed, and to present a possible remedy.Design/methodology/approachThe common practice is reviewed by using the Åknes case from Norway where an up to 100 million m3 rock slide is threatening one of Norway's most visited tourist sites, Geiranger. The same case is also reworked using the alternative approach and then a comparison is made. The study is therefore a comparative study.FindingsThe paper clearly shows the fallacy of using the frequency interpretation of probability in cases where the data are limited because the natural disasters under study appear very rarely. By exploiting the fact that responsible decision‐makers in public offices cannot claim that human losses today are worse than human losses tomorrow (human lives cannot be discounted, as it were), the alternative approach provides much more realistic decision‐support.Practical implicationsThe paper presents a new approach to analyzing the risk of low probability, high impact natural disasters that can be readily applied in other low probability, high consequence cases.Originality/valueAs far as is known, the paper presents an original contribution to the analysis of risk of low probability, high consequence natural disasters in that it is shown that the commonly used frequency interpretation of probability can prove to be flawed in such cases. An alternative approach is provided.