Japan quake model tackles earthquake occurrence probability
A new earthquake and tsunami model for Japan being developed by Impact Forecasting, the catastrophe modelling development centre of Aon Benfield, will differentiate itself from other models in the region in the way it tackles earthquake occurrence probability in the Nankai Trough, a submarine trough located south of the Nankaidō region of Japan’s island of Honshū.
Jeongmin Han, earthquake model developer for Aon’s Impact Forecasting team, explained to SIRC Today that one of the potential shortcomings in the current modelling of earthquakes is the ability to model recurrent Nankai Trough earthquakes.
This has been a challenge for a number of years but it is something Impact Forecasting has attempted to address in the new model that it is working on, which should be released in Q2 2018.
“There is not only a considerable degree of uncertainty associated with the occurrence probability of future Nankai events, but also the overall model results can vary significantly depending on assumptions made on the likelihood and severity of the events.
“This is especially the case when there is a limited number of earthquake observations available and the recurrence interval between the events is variable,” Han said.
“In Impact Forecasting’s new Japan earthquake and tsunami model, the earthquake occurrence probability in the Nankai Trough is evaluated by accounting for uncertainties associated with the timing of past observations and the variability of recurrence intervals.”
He explained that this approach is unique and will set the new model, once complete, apart from others that are available.
“This modelling approach with a unique view on Nankai Trough makes the Impact Forecasting model differentiated from the others including the National Seismic Hazard Maps developed by the Headquarters for Earthquake Research Promotion, a governmental organisation in Japan, for disaster preparation and planning purposes,” Han said.
“Another example of limitations in modelling is that no one model fits all insurance companies in Japan. Individual insurance companies have their unique risk profile, spatial distribution of exposure and insurance policy conditions in case of mutual companies.
“In order to reflect insured loss expectancy properly, Impact Forecasting’s model was calibrated based on clients’ claims histories and its model performance was tested through back-testing.”
Get the latest re/insurance news sent to your inbox every day - Sign up to our free email newsletters
Other stories from the SIRC Day Two newsletter