Studies on typhoon-wave prediction by Artificial Neural Networks at Longdong, Taiwan

Based on typhoon-wave data, collected from June 16 to August 17, 2004, during Typhoons Mindulle, Kompasu, and Rananim at the Longdong buoy station in northeast Taiwan, this study simulates typhoon-wave predictions using an Artificial Neural Network (ANN), and compares the results with the SWAN model. The results show that simulated data are consistent with measured data, and can effectively predict significant wave heights and periods with lead time six-hour. The root-mean-square errors are about 24.8cm and 0.54s.