In Japan, the fundamental disaster management plan was modified after a heavy rainfall event in 2015. According to the updated plan, the transfer of flood disaster risk to non-life insurance is promoted by the Japanese government. Thus, the importance of flood risk modeling for the insurance industry has increased.
Winds are expected to become even stronger, resulting in higher storm surges, when the central pressure of the typhoon is intensified. Furthermore, it is possible for an insurance system to experience peak risk when such damage occurs simultaneously. Hence, refining the assessment method of storm surge risk is very important.
An insurance company to which storm surge risk is transferred needs to assess not only the infrequent risks, for managing the risk of the company, but also the expected value of the estimated loss, for evaluating the insurance premium. However, only a few studies have assessed storm surges by stochastic approaches.
In this study, storm surge losses along the coast of Tokyo Bay are predicted using the output of a stochastic typhoon model for 10,000 years. Storm surge losses due to 600 typhoons potentially causing storm surge damage for 10,000 years are calculated. Exceedance probability curves (EP curves) of estimated storm surge loss for each asset are created. Expected loss and the loss of representative return periods are evaluated based on these EP curves. We successfully determined the expected loss with a small calculation load.