Wave height forecasting by the transfer function model

Three completing methods were introduced in this paper. The data were completed from 1983 to 1988 at the KL and PTC stations. The completed data were called original data. The completing methods were the transfer function with fixed-parameters then built the Transfer Function (TF) model between the two stations. The TF model was used to estimate the wave data of the future. The difference of the estimated data and the original data were conferred in this paper. Two types of forecasting differences were discussed here: (1) the input data year increases from 1 to 6 years, and the difference of the estimated data and the original data were then compared. (2) Use one-year data as the input series to forecast the data of the next five years, and then compare the differences of the estimated data and the original data. After comparing the differences we can know the results. The mean value and the standard deviation of the differences decrease and go more stable as the quantity of the input data year increases. The mean value and the standard deviation of the differences increase and go more unstable as the quantity of the forecasted data year increases. This model can be used to forecast the wave data.