Detail Abstract

Prediction of water level in irrigation system using the hydraulic and data-driven models: an application in the Bac Hung Hai irrigation System

Abstract

In the irrigation system, the water level is an important indicator because of its role in various purposes such as the operation of different structures like culverts, sluice gates, and pumps as well as water management, and water supply for different water–consuming sectors. Accurate estimations and predictions of water levels in agricultural irrigation systems are pre-requisite for the successful implementation of water resources planning and management tools. In this context, one-dimensional hydraulic (i.e., MIKE 11) and data-driven models (consisting of a traditional multiple linear regression and a recurrent long – short term memory) are implemented together for predicting the water level in irrigation systems. Different statistical metrics including the root mean square error (RMSE), mean absolute error (MAE), mean error (ME), Nash–Sutcliffe efficiency (NSE), Pearson’s correlation coefficient (r), and Willmott’s index are implemented for quantitatively assessing the agreement between estimated and observed water levels at various locations in the period from 2000 to 2021 (with an interval time of 6 hours). The Bac Hung Hai irrigation system (BHHIS) is used to demonstrate the capability of models. The results showed that both hydraulic and data-driven models can estimate water levels with high accuracy. The latter is strongly believed to provide a better understanding of water resources in the system as well as to support decision-making in water resources planning and management under the effects of climate change, urbanization, and human activities.