Enhancing Monthly Water Level Forecasting Through CNN-LSTM and SVR Models: Implications for Sustainable Water Resource Management

Abstract

Predicting water level using hybrid models is critically important for effective water resources management, flood control, and disaster prevention. Hybrid models combine the strengths of different modeling techniques—such as machine learning algorithms and physics-based hydrological models—to achieve higher accuracy and reliability in water level forecasts, especially in complex and nonlinear river and lake systems. This study examines the efficacy of a hybrid CNN-LSTM model in comparison to Support Vector Regression (SVR) for forecasting water levels in Beyşehir Lake and Atatürk Dam from 2003 to 2025. Accurate prediction of water levels is crucial for sustainable water resource management and flood risk mitigation, and hybrid models like CNN-LSTM are particularly valuable due to their ability to capture complex temporal patterns in hydrological data. The study employs univariate time series modeling, where the optimal lag is determined using the correlation coefficient to enhance prediction accuracy. Performance was evaluated using RMSE, R², and MAPE, with CNN-LSTM achieving superior results—for Beyşehir Lake, RMSE of 1.390 m, R² of 0.834, and MAPE of 2.252%; for Ataturk Dam, RMSE of 0.357 m, R² of 0.894, and MAPE of 1.028%—demonstrating the model’s strong predictive capability. These findings highlight the importance of leveraging hybrid deep learning approaches for improved water level forecasting, enabling more effective water management strategies over long-term periods.

Keywords

• Water level prediction
• Hybrid Model
• Ensemble models
• Nonlinear dynamics
• Hydrological data

References

1. [1] Ozdemir, S., Yaqub, M., & Yildirim, S. O. (2023). A systematic literature review on lake water level prediction models. Environmental Modelling & Software, 163, 105684.
2. [2] Castillo-Botón, C., Casillas-Pérez, D., Casanova-Mateo, C., Moreno-Saavedra, L. M., Morales-Díaz, B., Sanz-Justo, J., ... & Salcedo-Sanz, S. (2020). Analysis and prediction of dammed water level in a hydropower reservoir using machine learning and persistence-based techniques. Water, 12(6), 1528.
3. [3] Li, G., Liu, Z., Zhang, J., Han, H., & Shu, Z. (2024). Bayesian model averaging by combining deep learning models to improve lake water level prediction. Science of the Total Environment, 906, 167718.
4. [4] Čule, I. S., Ožanić, N., Volf, G., & Karleuša, B. (2025). Artificial neural network (ANN) water-level prediction model as a tool for the sustainable management of the Vrana Lake (Croatia) water supply system. Sustainability, 17(2), 1-19.
5. [5] Karsavran, Y. (2024). Comparison of ANN and SVR based models in sea level prediction for the Black Sea coast of Sinop. Turkish Journal of Maritime and Marine Sciences, 10(1), 49-56.
6. [6] Santos, C. A. G., Ghorbani, M. A., Abdi, E., Patel, U., & Sadeddin, S. (2025). Estimating water levels through smartphone-imaged gauges: a comparative analysis of ANN, DL, and CNN models. Water Resources Management, 39(4), 1639-1654.
7. [7] Xu, Y., He, C., Guo, Z., Chen, Y., Sun, Y., & Dong, Y. (2023). Simulation of water level and flow of catastrophic flood based on the cnn-lstm coupling network. Water, 15(13), 2329.
8. [8] Sharafkhani, F., Corns, S., & Holmes, R. (2024). Multi-step ahead water level forecasting using deep neural networks. Water, 16(21), 3153.

9. [9] Şener, E., Şener, Ş., & Bulut, C. (2023). Assessment of heavy metal pollution and quality in lake water and sediment by various index methods and GIS: A case study in Beyşehir Lake, Turkey. Marine Pollution Bulletin, 192, 115101.
10. [10] Awad, M., Khanna, R., Awad, M., & Khanna, R. (2015). Support vector regression. Efficient learning machines: Theories, concepts, and applications for engineers and system designers, 67-80.
11. [11] Zhang, F., & O'Donnell, L. J. (2020). Support vector regression. In Machine learning (pp. 123-140). Academic Press.
12. [12] Ali, M., Nayahi, J. V., Abdi, E., Ghorbani, M. A., Mohajeri, F., Farooque, A. A., & Alamery, S. (2025). Improving daily reference evapotranspiration forecasts: Designing AI-enabled recurrent neural networks based long short-term memory. Ecological Informatics, 85, 102995.
13. [13] Chicco, D., Warrens, M. J., & Jurman, G. (2021). The coefficient of determination R-squared is more informative than SMAPE, MAE, MAPE, MSE and RMSE in regression analysis evaluation. Peerj computer science, 7, e623. DOI: 10.7717/peerj-cs.623.
14. [14] Saunders, L. J., Russell, R. A., & Crabb, D. P. (2012). The coefficient of determination: what determines a useful R2 statistic?. Investigative ophthalmology & visual science, 53(11), 6830-6832. DOI: 10.1167/iovs.12-10598.
15. [15] Cho, M., Kim, C., Jung, K., & Jung, H. (2022). Water level prediction model applying a long short-term memory (lstm)–gated recurrent unit (gru) method for flood prediction. Water, 14(14), 2221.
16. [16] Ruma, J. F., Adnan, M. S. G., Dewan, A., & Rahman, R. M. (2023). Particle swarm optimization based LSTM networks for water level forecasting: A case study on Bangladesh river network. Results in Engineering, 17, 100951.
17. [17] Li, H., Zhang, L., Yao, Y., & Zhang, Y. (2025). Prediction of water levels in large reservoirs base on optimization of deep learning algorithms. Earth Science Informatics, 18(1), 121.
18. [18] Guo, H., Chen, Z., & Teo, F. Y. (2024). Intelligent water quality prediction system with a hybrid CNN–LSTM model. Water Practice & Technology, 19(11), 4538-4555.
19. [19] Huang, H., Wang, Z., Liao, Y., Gao, W., Lai, C., Wu, X., & Zeng, Z. (2024). Improving the explainability of CNN-LSTM-based flood prediction with integrating SHAP technique. Ecological Informatics, 84, 102904.
20. [20] Zhu, Y. (2024). Application of a QPSO-optimized CNN-LSTM model in water quality prediction. Discover Water, 4(1), 100.
21. [21] Süme, V., Yılmaz, E., Marangoz, H.O., (2025), Daneshfaraz Ebadzadeh, P., Shoaling and Sedimentation Dynamics in Fishery Shelters, A Case Study of Sandıktaş, Journal of Marine Science and Engineering 13 (4), 779
22. [22] Süme, V., Tansel, B., (2016), Capacity building for field inspections: A comprehensive assessment tool for monitoring structural integrity and sediment capture performance of T-head groins, Ocean & Coastal Management 125, 20-28
23. [23] Süme, V., Marangoz, H.O., (2024),Monitoring Sea Level Changes Along the Coast of Rize, Turkey Throughout the Year, Journal of Anatolian Environmental and Animal Sciences 9 (4), 785-794