Estimating the Significant Wave Height in the Caspian Sea Using Single and Combined-Wavelet Artificial Intelligence Methods

The significant wave height is one of the basic parameters for engineering operations of coasts and marine structures. In this study, significant wave height was estimated based on three scenarios including variables 1- waves, 2- meteorological and 3- combination of the first and second scenario, in time steps without time delay, 12 and 24 hours using adaptive neural fuzzy inference system (ANFIS) and gene expression programming (GEP) and combining them with wavelet theory (WANFIS, WGEP). Also, the results indicate that due to de-noising and removal of uncertainty in the data, combined-wavelet models have provided better results than singular models. The performance improvement percentage of WGEP models compared to GEP considering the RMSE criterion was 11%, 35%, and 7% for the first to third scenarios, respectively. Most of the hydrological conditions in the seas depend on temperature changes and the amount of this parameter is an important determining factor in the environmental conditions of each region. Also, changes in temperature and surface wind change the density of sea water. Therefore, the climatic variables of the region can affect different scenarios. The results of this study and the presentation of the governing mathematical relationship for estimating the value of significant wave height by the method of gene expression programming can be very useful in the field of coastal and water resources management and engineering.