AUTHOR=Liang Zehua , Liu Yaping , Hu Hongchang , Li Haoqian , Ma Yuqing , Khan Mohd Yawar Ali 

TITLE=Combined Wavelet Transform With Long Short-Term Memory Neural Network for Water Table Depth Prediction in Baoding City, North China Plain

JOURNAL=Frontiers in Environmental Science

VOLUME=Volume 9 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/environmental-science/articles/10.3389/fenvs.2021.780434

DOI=10.3389/fenvs.2021.780434

ISSN=2296-665X

ABSTRACT=Accurate estimation of water table depth dynamics is essential for water resource management, especially in areas where groundwater is over-exploited. In recent years, as a data-driven model, artificial neural networks (ANNs) have been widely used in hydrological modeling. However, due to the non-stationarity of water table depth data, the performance of ANNs in areas of over-exploitation is challenging. Therefore, reducing data noise is an essential step before simulating the water table depth. This research focuses using of wavelet analysis combined with Long-Short-Term Memory (LSTM) neural network (NN) to establish a data-driven model based on non-stationary time series data of water table depth, which provide a tool for groundwater resource management.  A typical groundwater over-exploitation area, Baoding, North China Plain (NCP), was selected as a study area. To reflect the impact of anthropogenic activities, the variables harnessed to develop the model includes temperature, precipitation, evaporation, and some socio-economic data. The results show that decomposing the time series of the water table depth into three sub-temporal components by Meyer wavelets can significantly improve the simulation effect of LSTM on the water table depth. Additionally, a feedforward neural network (FNN) is used to compare forecasts over 12-months. As expected, Wavelet-LSTM outperforms wavelet-FNN. As the prediction time increases, the advantages of wavelet-LSTM become more evident. The wavelet-LSTM is satisfactory for forecasting the water table depth at most in 6 months. Furthermore, the importance of input variables of wavelet-LSTM is analyzed by the weights of the model. The results indicate that anthropogenic activities such as agricultural production and the steel industry greatly influence the water table depth, especially in the sites close to the Baiyangdian reserve, the largest lake in the North China Plain. This study demonstrates that the wavelet-LSTM model provides an option for water table depth simulation and predicting areas of overexploitation of groundwater.