In this study, we aimed to develop and assess a hydrological model using a deep learning algorithm for improved water management. Single-output long short-term memory (LSTM SO) and encoder-decoder long short-term memory (LSTM ED) models were developed, and their performances were compared using different input variables. We used water-level and rainfall data from 2018 to 2020 in the Takayama Reservoir (Nara Prefecture, Japan) to train, test, and assess both models. The root-mean-squared error and Nash–Sutcliffe efficiency were estimated to compare the model performances. The results showed that the LSTM ED model had better accuracy. Analysis of water levels and water-level changes presented better results than the analysis of water levels. However, the accuracy of the model was significantly lower when predicting water levels outside the range of the training datasets. Within this range, the developed model could be used for water management to reduce the risk of downstream flooding, while ensuring sufficient water storage for irrigation, because of its ability to determine an appropriate amount of water for release from the reservoir before rainfall events.

Various transient test-based techniques have been developed for detecting leaks in pipeline systems, and most of these depend on a signal that is reflected by the leak. In field pipes, this signal is easily disturbed by high frequency noise, which can prevent detection. This paper presents a leak detection method using transient pressure damping by energy dissipation from the leak, which is minimally affected by the noise. The numerical simulation confirms that the method can be applied within the range of practical pipe scales, provided that the flow rate is reduced by restricting the valve. The model is verified, and leak location is estimated within an error of <6% of the total length despite the noise from the loops. This method offers a simple and easy means to detect leaks without numerical simulation, although it is limited to single pipeline systems.