Application of ANN and Wavelet Conjunction Model in Forcasting Short-Term Sea Level Variations (Case Study: Chabahar Port)

Exact determination of sea level in coastal area is so important in coastal engineering applications. Due to sea level forecasting, observation of sea flow and its surface variation, wave height, wind velocity, and the tide phenomena have been possible. These observations have a significant role in coastal management and planning. ANN and Wavelet Conjunction Model (WCM) ability in short-term forecasting sea level in the Chabahar port is evaluated in the present paper. Error coefficient statistical parameters (RMSE, E) are used as an index to compare aforementioned models with neutral network and linear regression methods. Available data of sea level which are taken hourly are implemented as model inputs and the next twelve hours are predicted using mentioned model. Comparison of ANN and WCM results with other models indicates better performance of mentioned model, in forecasting short-term sea level in this station. Error coefficient value, E, is 0.989, 0.878, and 0.848 in ANN and WCM, neural network, and linear regression, respectively. Using of wavelet transform and decomposition of sea level time-series into subseries with useful information and different frequences, process of forecasting is improved.