An improved method for conveying earthquake loss data utilising Monte Carlo simulations

Abstract

A simple and intuitive method for estimating the probability distribution of cumulative direct earthquake losses is presented. The method is a Monte Carlo simulation which treats the occurrence of earthquake events as a random Poisson process. Examples carried out using this method show that the probabilistic skew in a cumulative loss distribution is greatest for short considered time periods and for regions with lower overall seismic hazard. Disaggregation and sample loss functions in time are used to examine which intensity level events contributed most to the overall cumulative loss estimates and how often these events occurred. Errors and limitations of the model are briefly discussed, with the conclusion that the presented method is potentially a valuable means of conveying earthquake loss information, as an addition to an expected loss term and/or loss hazard curve.