Sustainable groundwater management using stacked LSTM with deep neural network

Groundwater is a vital water resource and plays a major role in human life, production, irrigation, and development of the country based on economically. Due to irregular rainfall and drought in the summer season, storage of groundwater is essential for the usage of multipurpose. The prediction of groundwater using the spatiotemporal attention mechanism is the main goal of this research. With the rapid growth of urbanization, population, and industrialization, the resource of groundwater has become vulnerable to depletion. Therefore, it is necessary for groundwater resource management in the aspects of quality and quantity. The groundwater and its demand are indirectly proportional to each other. This imbalance criterion brings more problems in groundwater availability. Effective and efficient planning is required to face this dilemma. In facing the groundwater challenge many research works have been implemented. The issues are inefficient and fail to predict the demand for water requirements. To overcome these issues this paper proposed managing groundwater by using stacked LSTM with a deep neural network (SLSTM-DNN) for performing demand prediction in the collected dataset. An accuracy rate of 81.32% was obtained for CNN, 82.34% obtained for DNN, 88.12% obtained for LSTM, and the proposed accuracy rate achieves 91.45%.