The Use of Neural Networks for Modeling the Movement of Surface Water Masses in Enclosed Sea Areas
Purpose: The article presents the use of neural networks to predict the parameters of the movement of surface water masses in enclosed sea areas. Design/Methodology/Approach: The input data were meteorological parameters recorded at the stations Trzebież and Świnoujscie. The output data were the parameters of moving drifters, obtained because of an experiment in 2018 in the waters of the Szczecin Lagoon. The model uses Multi-Layer Perceptron networks with different activation functions. As a criterion for selecting the best networks, the highest correlation statistics for the test and validation sample were used. Findings: As a result of the research, predictions of the speed and direction of surface water masses were obtained based on the meteorological conditions recorded on the outskirts of the studied reservoir. Originality/value: The presented research is a new application of artificial neural networks in security in restricted waters. The results obtained in the study may be beneficial for the maritime administration, which is responsible for the safety of navigation in the studied water area. The model can be used to design a survivor's route or a contamination route.