研究期間：10108~10207;Typhoons frequently cause casualties, property and economic losses in Taiwan. Some flood risk models have been developed to assess the losses for Taiwan’s property insurance industry. However, potentially extreme rainfalls would be ignored and assessments of property losses might be underestimated if only about 240 typhoon events recorded in Taiwan during 1960–2010 are considered in the flood risk models. This work generated a sufficient number of stochastic rainfall events, including potentially extreme rainfall events using a truncated lognormal distribution and a correlation matrix. Generated events will basically retain historical rainfall characteristics such as mean, standard deviation and spatial distribution. A practical case study is carried out that includes different generated rainfall events. Comparisons of event cases and their application for flood risk assessment for an insurance portfolio are made. The generated cases, which have extreme rainfall, produce higher losses than the loss from the historical case. This outcome will be a practical supplement to flood-risk assessments by considering the influence of potentially extreme future rainfall events.
