Using machine learning method to forecast river water levels in the Bac Hung Hai irrigation system in Vietnam

Abstract

In recent years, the application of the Machine Learning (ML) method in analyzing and studying hydrological problems is increasingly becoming common. The numerical models based on ML algorithms have been widely used for predicting river water levels or flowrate. This paper proposes a new approach using one of the applications of deep learning models to predict river water levels in irrigation systems. A predictive model has been developed based on the Long Short-Term Memory (LSTM) neural networks to forecast the water levels upstream of Tranh Culvert in the Bac Hung Hai irrigation system in Vietnam. The findings of this study indicate that although only a modest amount of data is required, the proposed model produced superior results. These results can be used to construct an operating regime for irrigation sluice gates in the Bac Hung Hai system.