Spectral wave analysis of using MIKE 21 SW and artificial neural network (ANN)

Degradation of the marine surface with offshore wind causes irregular waves. The longer blowing wind speed, the greater the wave height. However, surface of water against the wind, inducing the pressure difference between the back made him wrong. Thus, analysis of irregular waves formed by transferring the energy provided by wind is difficult, because random variables play a role in the generation of irregular waves and it is possible to calculative for uncertainity methods. Spectral wave analysis is one of the most important factors for the use of applications of coastal and ocean engineering such as, design of coastal and marine structure, water supply, navigation,wildlife habitat and commercial fishery. MIKE 21 SW software and Artificial Neural Network (ANN) were used for spectral wave analysis of multi-stations in the lake Okeechobee. Similarly, C. P. Tsai, C. Lin&J.-N. Shen (2002) applied the ANN technique to forecast among multi-stations. A. Altunkaynak and K.-H. Wang ( 2012) have used Artificial Neural Network (ANN) and integrated Geno-Kalman Filtering techniques to estimate the significiant wave height at stations LZ40, L006, L005 and L001 in Lake Okeechobee. ANN model and GKF model compared. The results showed that expert system techniques can perform succesfully to estimate the significiant wave height in shallow lakes which has multi-stations. A. Altunkaynak (2013) studied on geno-multilayer perceptron to predict the significiant wave height in multi-stations located in the Lake Okeechobee. The results were divided into two training algorithms which is named genetic algorithm and back propagation algorithms. To predict the suspended solid concentration in one of multi-stations located in the Lake Okeechobee, Florida, triple diagrams model was developed by A. Altunkaynak and K.-H. Wang (2010). In this research, MIKE 21 SW and Artificial Neural Network (ANN) have been used for prediction of significiant wave height. Spectral wave analysis and generating model is possible to MIKE 21 SW. With program that allows you to make numerical modelling as well as uncertainity methods, depending on many parameters. MIKE 21 SW supplies modelling of physical phenomena such as foaming effects, bottom friction effects, refraction caused by depth, reflection and waves caused by wind effects. Generating model, wind speed and wind direction were considered, especially. Also, Artificial Neural Network (ANN) is used for prediction of significant wave height from wind speed and direction. ANN could be used for the solution of most of engineering problems as well as it offers an effective way to solution without physical information requirements. In this study, ANN and MIKE 21 SW was applied to four stations (LZ40, L006, L005 and L001) located in the Lake Okeechobee, Florida. A comparison between the result of two different spectral wave model shows that MIKE 21 SW can perform very well in predicting the significant wave height and produce stronger correlation coefficent and coefficient of efficiency than Artificial Neural Network. With programs such as MIKE that allows to make numerical modeling besides uncertainity methods, depending on many parameters and spectral wave analysis and generating model becomes possible. Modeling of physical phenomena such as foaming effects, bottom friction effects, refraction caused by depth, reflection and waves caused by wind effects in areas such as wetlands and coastal areas, seas, lakes interact with each other, is important. Alternatively, the Artificial Neural Networks (ANN) is used. Artificial Neural Networks could be used for the solution of most of engineering problems as well as it offers an effective way to solution without physical information requirements. In this study, spectral wave analysis is developed using MIKE21 SW programming and Artificial Neural Networks (ANN) for Lake Okeechobee. Comparison of model results, the accuracy of wave height data were verified and examined, depending on which parameters should be studied. Global Mapper program was used to identify the bathymetry. Bathymetry information is the important process of model set up to obtain reliable calculating the results by minimising the errors. Creating the mesh network, triangular mesh was chosen to find the best result for lake Okeechobee by using iteration. Bathymetry map was obtained by interpolating based on triangular mesh network The wind speed and wind direction data, the significant wave height data and current data were used for LZ40, L006, L005 and L001 stations in the lake Okeechobee and spectral wave analysis was developed using MIKE21 software and artificial neural networks (ANN) method. Lake Okeechobee is located in the US state of Florida which is located south central of Florida between 26° 41' and 27° 12' North, 80° 36' and 81° 05' West (Jin et al., 2003). Lake Okeechobee is one of the most important water resource. Lake surface area is 1890.7 km2 (730 mi2) and an average depth is 2.74 m (9 feet) ((SFWMD), 2014). It is the largest fresh water lake in the southeastern United States ((SFWMD), 2014) and Kissimme river is the most important water resource of Okeechobee lake. Wind is one the most important parameter in terms of mixtures and movement of water for Okeechobee lake where is a quite shallow. In this study, bathymetry data were taken from website of (sofia.usgs.gov/publications/maps/lakeokeebathy/ index.html) South Florida Information Access (SOFIA) to prepare the bathymetry map. Raw data file (ptokeeft.xyz) containing longitude, latitude depth data was downloaded. Wind, significant wave height (Altunkaynak, 2013) and current data (Altunkaynak and Wang, 2010) were taken from respectively, LZ40 where it is located at 26°52' 56.00"N ve 80°47'15.01"W longitude and latitude, L006 where it is located at 26° 49' 24'' N ve 80° 47' 16'' W longitude and latitude, L005 where it is located at 26°57'32.97"N ve 80°58'40.02"W longitude and latitude and L001 where it is located at 27° 8'24.00"N ve 80°47'24.00"W longitude and latitude in the lake Okeechobee. Section 2 describes the data and methods used in this study. The programs used and the evalution of model performance methods are explained. In Section 3, the spectral wave analysis application is performed and measured values were compared with predicted wave heights using by MIKE 21 SW and ANN. The prediction of significiant wave height were verified in terms of accuracy based on statistical criteria. In Section 4, the results are evalulated.