Abstract. Prediction of groundwater level is of immense importance and challenges coastal aquifer management with rapidly increasing climatic change. With the development of artificial intelligence, data-driven models have been widely adopted in hydrological process management. However, due to the limitation of network framework and construction, they are mostly adopted to produce only 1 time step in advance. Here, the temporal convolutional network (TCN) and models based on long short-term memory (LSTM) were developed to predict groundwater levels with different leading periods in a coastal aquifer. The initial data of 10 months, monitored hourly in two monitoring wells, were used for model training and testing, and the data of the following 3 months were used as prediction with 24, 72, 180, and 360 time steps (1, 3, 7, and 15 d) in advance. The historical precipitation and tidal-level data were incorporated as input data. For the one-step prediction of the two wells, the calculated R2 of the TCN-based models' values were higher and the root mean square error (RMSE) values were lower than that of the LSTM-based model in the prediction stage with shorter running times. For the advanced prediction, the model accuracy decreased with the increase in the advancing period from 1 to 3, 7, and 15 d. By comparing the simulation accuracy and efficiency, the TCN-based model slightly outperformed the LSTM-based model but was less efficient in training time. Both models showed great ability to learn complex patterns in advance using historical data with different leading periods and had been proven to be valid localized groundwater-level prediction tools in the subsurface environment.

The influence of lunar semidiurnal tides on coastal groundwater aquifers has been conceptualized for decades. However, a thorough understanding of the impact of tides on groundwater dynamics due to the widely distributed waterways and heterogeneous sediments in estuarine aquifers, is still needed. This study shows the tidal impact on groundwater dynamics in the Pearl River estuary in southeast China through wavelet and time series analysis. The groundwater level and electrical conductivity (EC), as well as tidal levels, were monitored in several observation wells and tidal stations to determine how the estuarine groundwater levels respond to tidal forcing. The results show that the groundwater fluctuations have short periodicities of 0.51 and 1 day corresponding to major tidal constituents of M2 (semidiurnal) and K1 and O1 (diurnal) signals, respectively. The significant impacts decrease with increasing distance inland of the locations of the wells. Additionally, the coherence analysis displays a higher correlation between tides and groundwater levels for the spring tide than for the neap tide. The tidal infl uences on groundwater EC are weak compared to those on groundwater levels. In addition, when the tidal level increases, the EC decreases in wells located in the estuarine entrance. This is related to the high salinity of retained paleo-seawater in the strata lenses. A conceptual model is proposed to illustrate the complex groundwater flow dynamics. The model may provide useful insights into the understanding of similar systems located in geographically different coastal regions.

<p>The influences of lunar semidiurnal tides on coastal groundwater aquifers have been conceptualized for decades. However, in estuarine aquifers, comprehensive work is needed to quantify the impact of the tides on groundwater dynamics due to the widely distributed waterways and heterogeneous sediments. Taking the Pearl River estuary in southeast China as a study site, the tidal impacts on the groundwater dynamics have been investigated through wavelet and time series analysis. The groundwater level and electrical conductivity (EC) in four monitoring wells, along with waterway water level (tidal level) at three tidal stations, were monitored every 30 minutes over a 2-month period to determine how nearshore groundwater responds to tidal forcing. The results show that the estuarine groundwater fluctuations have two significant short periodicities (0.51 and 1 day), which correspond to the major tidal constituents in the tides: M<sub>2</sub><sub> </sub>(semidiurnal), K<sub>1</sub><sub> </sub>and O<sub>1</sub> (diurnal) signals. The significant impacts decrease with increasing distance inland of the locations of the wells. Additionally, the coherence analysis displays a higher correlation between tides and groundwater levels for the spring tide than for the neap tide. The tidal influences on groundwater EC are weaker. In addition, when the tide level increases, the EC decreases in the wells located in the estuarine entrance. This phenomenon is related to the high salinity of retained paleo-seawater in the strata lens. A conceptual model is proposed to illustrate the complex groundwater flow dynamics, which provides useful insights into understanding groundwater systems in other geographically similar coastal estuarine regions.</p>