Climate Change Research

Climate Change Research

Evaluation of the integration of LSTM Recurrent Neural Network with deep learning metaheuristic algorithms for flood modeling in the Taleghan watershed

Document Type : Original Article

Authors
Ramtin Tavoosi Rad 1
Mohammad Ansari ghojghar 2
1 Master Candidate, Department of Reclamation of Arid and Mountainous regions Engineering, Faculty of Natural Resources, University of Tehran, Karaj, Iran
2 Assistant Professor, Department of Reclamation of Arid and Mountainous regions Engineering, Faculty of Natural Resources, University of Tehran, Karaj, Iran
10.30488/ccr.2026.562118.1317
Abstract
Accurate river flow forecasting is a fundamental challenge in water resources management and flood warning system design. In this study, aiming to evaluate the efficiency of new deep learning models in predicting daily maximum flow, four models including LSTM, GRU-LSTM, ConvLSTM and S-LSTM were compared in five hydrometric stations of the Taleghan watershed. After data pre-processing and extraction of time-lag scenarios, the models were evaluated based on RMSE, MAE, NSE, and correlation coefficient indices in the training and test sets. The results showed that the time-memory-based models LSTM and GRU-LSTM provided significantly more accurate performance than the spatiotemporal models and developed versions. The GRU-LSTM model recorded the best accuracy with NSE higher than 0.95 and very low error, but statistical analysis showed that its difference with the LSTM model is not significant, so that the LSTM algorithm offers almost the same level of accuracy with less computational cost. In contrast, ConvLSTM and S-LSTM performed weaker and the greater dispersion of points in the correlation diagrams indicated their limitations in modeling univariate time series. Visual analysis of the correlation diagrams also showed that LSTM and GRU-LSTM had the most matching with the correlation line and were able to reconstruct temporal patterns of flow, especially peaks and troughs. Considering the balance between accuracy, stability and computational cost, the LSTM model is proposed as the final and optimal option for predicting daily maximum flow in the Taleghan watershed
Keywords
Flow forecasting
Water resources management
Flood warning systems
Novel models

Subjects

1.    انصاری قوجقار، م.، عراقی نژاد، ش.، بذرافشان، ج.، زهرایی، ب. و پارسی، ا. (1403). ارزیابی کارایی مدل هیبریدی GRU-LSTM  در پیش بینی طوفان‌ها‌ی گرد و غبار (مطالعه موردی: استان خوزستان)، مجله تحقیقات منابع آب ایران، 17(1)، 16-32. magiran.com/p2291298
2.    بهرامی، ع. و خزینه‌نژاد، ح.(1399). پیش بینی دبی جریان سیلاب با مدل بیان ژن و برآورد هیدروگراف آن با استفاده از دو مدل گری و گاما (مطالعه موردی : حوضه آبخیز قره سو کرمانشاه). نشریه هواشناسی و علوم جو، 3(3)، 275-293.
3.    توکلی م، امیراحمدی ا، و گلی مختاری، ل.(1403). ارزیابی، پیش‌بینی و تحلیل منطقه‌ای سیلاب با استفاده از مدل‌های داده‌کاوی در حوضه آبریز فریزی. فصل‌نامه تحقیقات جغرافیایی، ۳۹ (۲) :۱۶۱-۱۶۸. ‎DOI: 10.58209/geores.39.2.161
4.    حسین زاده ف، عبادی ح،و کیانی ع.(1401). ادغام داده‌ها‌ی سنجش‌ازدوری و هواشناسی جهت پیش بینی زمانی سیلاب با استفاده از الگوریتم یادگیری عمیق. مهندسی فناوری اطلاعات مکانی، ۱۰ (۲) :۱۹-۳۸. DOI:10.52547/jgit.10.2.19
5.    حنیفی نیا, عبدالعزیز و عبقری, هیراد . (1404). پیش‌‌بینی مناطق مستعد سیل با استفاده از مدل‌‌های یادگیری ماشین خطی تعمیم یافته و بیشینه آنتروپی. مخاطرات محیط طبیعی، 14(43)، 19- 34. doi: 10.22111/jneh.2024.47730.2021
6.    زینعلی، م.، بزرگ حداد، ا.، یاسی، م. و علیزاده، ح. (1403). ارزیابی کارایی مدل LSTM در پیش بینی جریان روزانه ورودی به مخازن سدها، مجله تحقیقات آب و خاک ایران، 55(10)، 1863-1883. magiran.com/p2813028
7.    غفاری رزین، س.ر.، وثوقی، ب. و هوشنگی، ن. (1402). ارزیابی کارایی مدل شبکه عصبی حافظه کوتاه مدت طولانی در پیشبینی سری زمانی یونوسفر و مقایسه آن با مدلهای GRNN، GIM و NEQUICK. اطلاعات جغرافیایی، 32(126). DOI:10.22131/SEPEHR.2023.547749.2839
8.    کاظمی، ح.، منصوری، ن. و جوزی، س.ع. (1400). پهنه‌بندی خطر سیلاب در شهر نوشهر با استفاده از مدل‌های یادگیری ماشین. مسکن و محیط روستا، 40(176)، 71-86. DOI: 10.22034/40.176.71
9.    Akandeh, A., & Salem, F.M. (2019). Slim LSTM NETWORKS: LSTM_6 and LSTM_C6. 2019 IEEE 62nd International Midwest Symposium on Circuits and Systems (MWSCAS), 630-633.
10. Al-Qaderi, M. K., & Rad, A. B. (2018). A Multi-Modal Person Recognition System for Social Robots. Applied Sciences, 8(3), 387. https://doi.org/10.3390/app8030387
11. C.Oddo, P., D.Bolten, J., V.Kumar, S. & Cleary, B. (2024). Deep Convolutional LSTM for improved flash flood prediction. Water, 6. https://doi.org/10.3389/frwa.2024.1346104
12. Ebrahimi, A., Roozbahani, A. & Bani habib, M.A. (2018). Groundwater Level Prediction Using Dynamic Bayesian Networks Based on Sensitivity Analysis (Case Study: Birjand Plain). Iranian Water, 12(2), 91-100
13. Lindemann, B., Muller, T., Vietz, H., Jazdi, N.& Weyrich, M.(2021). A survey on long short-term memory networks for time series prediction. 14th CIRP Conference on Intelligent Computation in Manufacturing Engineering, CIRP ICME ’20. 650-655. https://doi.org/10.1016/j.procir.2021.03.088
14. Liao, Y., Wang, Z., Yu, H., Gao, W., Zeng, Z., Li, X., & Lai, C. (2025). Accelerating urban flood inundation simulation under spatio-temporally varying rainstorms using ConvLSTM deep learning model. Water Resources Research, 61, e2025WR040433. https://doi.org/10.1029/2025WR040433
15. Liu, W., Wang, Y., Zhong, D., Xie, S., & Xu, J. (2022). ConvLSTM Network-Based Rainfall Nowcasting Method with Combined Reflectance and Radar-Retrieved Wind Field as Inputs. Atmosphere, 13(3), 411. https://doi.org/10.3390/atmos13030411
16. Lingling Xi and Dandan Gong 2024 J. Phys.: Conf. Ser. 2731 012038. DOI 10.1088/1742-6596/2731/1/012038
17. M. Fathi, S. (2018). SLIM LSTMs. https://doi.org/10.48550/arXiv.1812.11391
18. Rugina, A.M. (2023). The Use of Recurrent Neural Networks (S-RNN, LSTM, GRU) For Flood Forecasting Based on Data Extracted from Classical Hydraulic Modeling. Modelling in Civil Environmental Engineering, 18(3). DOI: https://doi.org/10.2478/mmce-2023-0011
19. Shirazi, F., Zahiri, A., Piri, J. & Dehghani, A.A. (2025). Development a new hydraulic method for prediction of river flood discharge. Journal of Watershed Management Research, 14(28), 110-123. doi.org/10.61186/jwmr.14.28.110. (In Persian).
20. Shi, X., Chen, Z., Wang, H., Yeong, D.T., Wong, W.K. & Woo, W.C. (2015). Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting. https://doi.org/10.48550/arXiv.1506.04214
21. Wang, Y., Wang, W., Zang, H., & Xu, D. (2023). Is the LSTM Model Better than RNN for Flood Forecasting Tasks? A Case Study of HuaYuankou Station and LouDe Station in the Lower Yellow River Basin. Water, 15(22), 3928. https://doi.org/10.3390/w15223928
22. X. Dai, H. Yin and N. K. Jha.(2019). Grow and Prune Compact, Fast, and Accurate LSTMs. in IEEE Transactions on Computers, 69(3), 441-452. doi: 10.1109/TC.2019.2954495.
23. Zhang, Y., Zhou, Z., Van Griensven Thé, J., Yang, S.X., & Gharabaghi, B. (2023). Flood Forecasting Using Hybrid LSTM and GRU Models with Lag Time Preprocessing. Water. DOI:10.3390/w15223982
24. Zhang, Y., Liu, Y., Kong, Y., & Hu, P. (2024). An Improved Method for Retrieving Subsurface Temperature Using the ConvLSTM Model in the Western Pacific Ocean. Journal of Marine Science and Engineering, 12(4), 620. https://doi.org/10.3390/jmse12040620