Water Temperature Prediction Using Improved Deep Learning Methods through Reptile Search Algorithm and Weighted Mean of Vectors Optimizer

• Published in: Journal of marine science and engineering Vol. 11; no. 2; p. 259
• Main Authors: Ikram, Rana Muhammad Adnan, Mostafa, Reham R., Chen, Zhihuan, Parmar, Kulwinder Singh, Kisi, Ozgur, Zounemat-Kermani, Mohammad
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.01.2023
• Subjects: Accuracy; Air temperature; Algorithms; Analysis; Aquatic ecosystems; Aquatic resources; Artificial neural networks; Case studies; China; Datasets; Decision trees; Deep learning; deep learning methods; Ecosystem management; Environmental impact; Environmental impact assessment; Errors; Hydrology; Long short-term memory; Machine learning; Management; Mathematical models; Mathematical optimization; Neural networks; Optimization; Optimization algorithms; Precipitation; Predictions; reptile search algorithm; Reptiles; Reptiles & amphibians; Rivers; Search algorithms; Sediments; Statistical analysis; Strategic management; Stream discharge; Stream flow; Streamflow; Vectors; Water; Water resources; Water temperature; Weather forecasting; weighted mean of vectors optimizer; China; Bailong River; Yangtze River
• ISSN: 2077-1312; 2077-1312
• DOI: 10.3390/jmse11020259

Precise estimation of water temperature plays a key role in environmental impact assessment, aquatic ecosystems’ management and water resources planning and management. In the current study, convolutional neural networks (CNN) and long short-term memory (LSTM) network-based deep learning models were examined to estimate daily water temperatures of the Bailong River in China. Two novel optimization algorithms, namely the reptile search algorithm (RSA) and weighted mean of vectors optimizer (INFO), were integrated with both deep learning models to enhance their prediction performance. To evaluate the prediction accuracy of the implemented models, four statistical indicators, i.e., the root mean square errors (RMSE), mean absolute errors, determination coefficient and Nash–Sutcliffe efficiency were utilized on the basis of different input combinations involving air temperature, streamflow, precipitation, sediment flows and day of the year (DOY) parameters. It was found that the LSTM-INFO model with DOY input outperformed the other competing models by considerably reducing the errors of RMSE and MAE in predicting daily water temperature.