確率台風モデルとニューラルネットワークを用いた台風の将来変化を考慮した高潮予測

抄録

In order to project long-term changes in storm surge characteristics under future climate condition, the present study has developed two types of statistical storm surge models using an artificial neural network (NN). The proposed NN model, which uses only tropical cyclone information, enables to estimate large number event of storm surges using the stochastic tropical cyclone model. The long-term impact assessment of storm surge around Japan can be estimated by the NN and the stochastic tropical cyclone model considering future change of tropical cyclone characteristics.