A systematic review of neural network applications for groundwater level prediction

• Published in: Discover applied sciences Vol. 7; no. 9; pp. 942 - 30
• Main Authors: Afful, Samuel K., Boateng, Cyril D., Ahene, Emmanuel, Aryee, Jeffrey N. A., Wemegah, David D., Gidigasu, Solomon S. R., Britwum, Akyana, Osei, Marian A., Gilbert, Jesse, Touré, Haoulata, Mensah, Vera
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.09.2025; Springer Nature B.V; Springer
• Subjects: Accuracy; Applied and Technical Physics; Aquifers; Artificial intelligence; Artificial neural networks; Artificial neural networks (ANNs); Back propagation networks; Chemistry/Food Science; Climate variables; Datasets; Earth Sciences; Engineering; Environment; Environmental factors; Environmental impact; Forecasting; Groundwater; Groundwater data; Groundwater level (GWL) forecasting; Groundwater levels; Heuristic methods; Hydrology; Long short-term memory; Machine learning; Materials Science; Meta-analysis; Neural networks; Neural networks (NNs); Predictions; Review; Root-mean-square errors; Systematic review; Water shortages; Groundwater level (GWL) forecasting; Climate variables; Review; Artificial neural networks (ANNs); Neural networks (NNs)
• ISSN: 3004-9261; 2523-3963; 3004-9261; 2523-3971
• DOI: 10.1007/s42452-025-06817-5

Physical models have long been employed for groundwater level (GWL) prediction. Recently, artificial intelligence (AI), particularly neural networks (NNs), has gained widespread use in forecasting GWL. Forecasting of GWL is essential to enable the analysis, quantifying, and management of groundwater. This systematic review investigates the application of NNs for GWL prediction, focusing on the architectures of the various NN models employed. The study utilizes the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) methodology to screen and synthesize relevant scientific articles. Various NN architectures, such as artificial neural networks (ANNs), feedforward neural networks (FFNNs), backpropagation neural networks (BPNNs), long short-term memory (LSTM), and hybrid models, were analyzed. The results from the systematic review indicate a growing preference for hybrid models, which effectively capture hidden relationships between GWL and environmental factors. The root mean square error (RMSE) emerges as the predominant performance metric, highlighting its significance in evaluating NNs. Results from the review also highlight the significance of comprehensive, long-term datasets covering a decade for robust trend analyses and accurate predictions. The findings contribute to a deeper understanding of new trends in groundwater research such as the application of neural networks for prediction problems in groundwater research. In conclusion, a hybrid metaheuristic algorithm produced more efficient results emphasizing their efficacy. In addition, lagged values were essential input for GWL prediction. The paper addressed both technical nuances and broader environmental implications. Article Highlights Neural networks help predict groundwater levels, with hybrid models improving accuracy. Long-term data (10 + years) is essential for reliable groundwater level forecasting. Data quality, preprocessing techniques, and optimal hyperparameter selection improves prediction accuracy.