Flood Hazard Zonation and Prediction using Deep Learning Models in the Qarasu Watershed, Golestan Province

Document Type : Research

Authors

1 Associate Professor, Soil Conservation and Watershed Management Research Institute, Agricultural Research, Education and Extension Organization (AREEO), Tehran, Iran

2 M.Sc., Islamic Azad University, Lenjan, Isfahan, Iran

10.22092/wmrj.2026.371561.1648

Abstract

Introduction and Goal
Flooding is one of the most destructive natural disasters with social, economic and environmental consequences. Therefore, machine learning methods have been developed to model and predict it. The purpose of this research was to zone and predict flood hazard using three deep learning models, including Gated Recurrent Unit (GRU), Long Short-Term Memory (LSTM) and Temporal Convolutional Network (TCN) models, and to introduce the most suitable ones in the Qarasu watershed of Golestan Province. The most important innovation of this research was comparing its results with the results of other research related to susceptibility zonation within the country, using new deep learning models, and considering the most flood factors to determine the most efficient model and increase the accuracy of flood prediction maps.
Materials and Methods
The research area of ​​the Qarasu watershed in Gorganwas was selected. Then maps of the effective factors were collected and prepared. In this study, 16 factors were selected as independent variables and information received from the Ministry of Energy and field visits were selected as dependent variable. Then, a flood event distribution map was prepared. To model and predict flood susceptibility, the flood event distribution map was divided into two groups of experimental points (30%) and training points (70%). After running the TCN, LSTM and GRU models, the maps obtained from flood hazard zonation maps were classified into five classes: very low, low, moderate, high and very high. Finally, the classification accuracy and validation of the zonation and prediction maps were evaluated the flood susceptibility was predicted, and the most appropriate model was selected.
Results and Discussion
To prepare a flood event zonation map, spatial data of 59 flood occurrence points (dependent variable) were collected using available information and by analyzing radar images acquired before and after flood events in Google Earth. Each of these locations experienced 10 recorded and documented flood events during the 20- year study period (2004–2024). Of the total dataset, 70% were randomly selected and divided as training data and 30% as test data for model implementation and validation, respectively. In the next step, 16 factors in the form of environmental groups, hydrology, watershed topography along with climatic data (as independent variables) for zonation and modeling. The classification accuracy of the models was evaluated using two indices, frequency ratio (FR) and kernel cell area (SCAI). The results showed that most flood areas were in the high-hazard class (high and very high) and the largest share of the area of ​​the region was also related to these areas. In addition, based on the area under the relative operator curves and cost, it was determined that the most efficient, appropriate, and cost-effective model was the TCN model compared to the LSTM and GRU models. Among the models studied, the highest and lowest areas under the relative operator and cost curves of the TCN model were 0.92 and 0.08, respectively. The results of this study showed that, based on the TCN model, 45% of the area of ​​the region was in the very high susceptibility class, containing 83% oflood events. Also, based on the LSTM model, 23% of the region contained 60% of the flood events, and based on the GRU model, 10% of the region contained 4% of the flood events. The flood hazard zonation maps obtained in this study can be the basis for planning and crisis management caused by flood events, reducing flood hazard, and managing it. Furthermore, based on the excellent performance of all three TCN, LSTM, and GRU models, the results of this research are necessary and essential for the development of future projects of various organizations operating in many developing countries.
Conclusion and Suggestions
After running all three TCN, LSTM, and GRU models, the AUC-ROC index results in both training and testing step showed excellent performance of the mentioned models. The maps prepared from flood hazard zonation using all three models indicated high accuracy in zone classification and appropriate distribution of flood points in high and very high class. The fit of the TCN model in predicting flood occurrence in the test phase was better compared to the other two models. The evaluation of classification accuracy and validation of the models showed that two TCN and LSTM models, having the appropriate thresholds for classification, are the priority of their application for zonation and predicting the hazard of flooding. Finally, in this study, the area under the value or cost curve index values ​​for TCN, LSTM, and GRU models were obtained as 0.08, 0.10, and 0.11, respectively. Based on the importance of morphometric features in the zoning of flood hazard maps, it is suggested that future research should use these metrics along with other common metrics in the preparation of zoning maps. Furthermore, based on the findings of this study, the application of deep learning methods and their combinations is recommended. It is also suggested that novel hybrid approaches and suitable optimization algorithms be investigated and evaluated in other watersheds of Golestan Province.

Keywords

Main Subjects

References

Aboah A, Adu-Gyamfi Y, Gursoy SV, Merickel J, Rizzo M, Sharma A. 2023. Driver maneuver detection and analysis using time series segmentation and classification. Journal of Transportation Engineering. Part A, Systems. 149(3): 1–15. https://doi.org/10.1061/JTEPBS.TEENG-7312
GDNRWM- General Department of Natural Resources and Watershed Management of Golestan Province, Iran. 2018.
Behera I, Jain N, Martha TR, Bose M. 2026. Optimizing sampling strategies for AI-based landslide susceptibility mapping: a case study of Idukki, Kerala. Nat Hazards. 122(1). https://doi.org/10.1007/s11069-025-07842-8
Bentivoglio R, Isufi E, Jonkman SN, Taormina R. 2022. Deep learning methods for flood mapping: a review of existing applications and future research directions. Hydrology and Earth System Sciences. 26(16): 4345–4378. https://doi.org/10.5194/hess
Chen H, Lu T, Huang J, He X, Yu K. 2023. An improved VMD-LSTM model for time-varying GNSS time series prediction with temporally correlated noise. Remote Sensing. 15: 3694. https://doi.org/10.3390/ rs15143694
Chen J, Li Q, Wang H, Deng M. 2020. A machine learning ensemble approach based on random forest and radial basis function neural network for risk evaluation of regional flood disaster: a case study of the Yangtze River Delta, China. International Journal of Environmental Research Public Health. 17(1): 49. https://doi.org/10.3390/ijerph17010049
Chen W, Panahi M, Tsangaratos P, Shahabi H, Ilia I, Panahi S, Ahmad BB. 2019. Applying population-based evolutionary algorithms and a neuro-fuzzy system for modeling landslide susceptibility. Catena. 172: 212–231. https://doi.org/10.1016/j.catena.2018.08.025
Chen Y, He Y, Zhang LF. 2021. Prediction of InSAR deformation time-series using a long short-term memory neural network. International Journal of Remote Sensing. 42(18): 6921–6944. https://doi.org/10.1080/01431161.2021.1947540
Choubin B, Moradi E, Golshan M, Adamowski J, Sajedi-Hosseini F, Mosavi A. 2019. An ensemble prediction of flood susceptibility using multivariate discriminant analysis, classification and regression trees, and support vector machines. Science of the Total Environment. 651: 2087–2096. https://doi.org/10.1016/j.scitotenv.2018.10.064
Convertino M, Annis A, Nardi F. 2019. Information-theoretic portfolio decision model for optimal flood management. Environmental Modelling and Software 119: 258–274. https://doi.org/10.1016/j.envsoft.2019.06.013
Costache R, Zaharia L. 2017. Flash-flood potential assessment and mapping by integrating the weights-of-evidence and frequency ratio statistical methods in GIS environment–case study: Bâsca Chiojdului River catchment (Romania). Journal of Earth System Science. 126(4): 1–19.‏ https://doi.org/10.1007/s12040-017-0828-9
Dalu MT, Shackleton CM, Dalu T. 2018. Influence of land cover, proximity to streams and household topographical location on flooding impact in informal settlements in the Eastern Cape, South Africa. International Journal of Disaster Risk Reduction 28: 481–490. https://doi.org/10.1016/j.ijdrr.2017.12.009
Dano UL, Balogun AL, Matori AN, Wan Yusouf K, Abubakar IR, Said Mohamed MA, Aina YA, Pradhan B. 2019. Flood susceptibility mapping using GIS-based analytic network process: A case study of Perlis, Malaysia. Water. 11(3): 615. https://doi.org/10.3390/w11030615
Dodangeh E, Choubin B, Eigdir AN, Nabipour N, Panahi M, Shamshirband S, Mosavi A. 2020a. Integrated machine learning methods with resampling algorithms for flood susceptibility prediction. Science of the Total Environment. 705: 135983. https://doi.org/10.1016/j.scitotenv.2019.135983
Dodangeh E, Panahi M, Rezaie F, Lee S, Bui DT, Lee CW, Pradhan B. 2020b. Novel hybrid intelligence models for flood-susceptibility prediction: Meta optimization of the GMDH and SVR models with the genetic algorithm and harmony search. Journal of Hydrology. 590: 125423. https://doi.org/10.1016/j.jhydrol.2020.125423
Drummond C, Holte RC. 2006. Cost curves: An improved method for visualizing classifier performance. Machine Learning. 65: 95–130. https://doi.org/10.1007/s10994-006-8199-5
Eslaminezhad SA, Eftekhari M, Akbari M. 2020. GIS-based flood risk zoning based on data-driven models. Journal of Hydraulic Structures. 6(4): 75–98.‏ https://doi.org/10.22055/jhs.2021.36629.1163
Gebrehiwot A, Hashemi-Beni L, Thompson G, Kordjamshidi P, Langan T. 2019. Deep convolutional neural nNetwork for flood extent mapping using unmanned aerial vehicles data. Sensors. 19(7): 1486. https://doi.org/10.3390/s19071486
Islam ARMT, Talukdar S, Mahato S, Kundu S, Eibek KU, Pham QB, Linh NTT. 2021. Flood susceptibility modelling using advanced ensemble machine learning models. Geoscience Frontiers. 12(3): 101075.‏ https://doi.org/10.1016/j.gsf.2020.09.006
Kaur S, Kaur A, Kumar M. 2026. Recent Advances in Computational Methods in Science and Technology: Volume 1 (1st ed.). CRC Chemical Rubber Company Press. https://doi.org/10.1201/9781003662839
Khosravi K, Shahabi H, Pham BT, Adamowski J, Shirzadi A, Pradhan B, Dou J, Ly HB, Gróf G, Ho HL. 2019. A comparative assessment of flood susceptibility modeling using Multi-Criteria Decision-Making Analysis and Machine Learning Methods. J. Hydrol Journal of Hydrology. 573: 311–323. https://doi.org/10.1016/j.jhydrol.2019.03.073
Kia MB, Pirasteh S, Pradhan B, Mahmud A R, Sulaiman WNA, Moradi A. 2012. An artificial neural network model for flood simulation using GIS: Johor River Basin, Malaysia. Environ. Earth Science. 67: 251–264. https://doi.org/10.1007/s12665-011-1504-z
Kratzert F, Herrnegger M, Klotz D, Hochreiter S, Klambauer G. 2019. Neural hydrology–interpreting LSTMs in hydrology. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). 11700 Lecture Notes in Computer Science. pp. 347–362. https://doi.org/10.1007/978-3-030-28954-6
LeCun Y, Bengio Y, Hinton G. 2015. Deep learning. Nature. 521(7553): 436–444. https://doi.org/10.1038/nature14539
Li J, Fan C, Zhao K, Zhang Z, Duan P. 2025. Landslide displacement prediction using time series InSAR with combined LSTM and TCN. Application to the Xiao Andong landslide, Yunnan Province, China. Natural Hazards. 121(4): 3857–3884.‏ https://doi.org/10.1007/s11069-024-06937-y
Maier HR, Jain A, Dandy GC, Sudheer KP. 2010. Methods used for the development of neural networks for the prediction of water resource variables in river systems: Current status and future directions. Environmental Modelling and Software. 25: 891–909. https://doi.org/10.1016/j.envsoft.2010.02.003
Malekian A, Khozani AO, Ashoornejad Gh. 2012. Zoning of flooding potential of Akhtarabad watershed using fuzzy hierarchical analysis method. Natural Geography Research. 4(44):162–191. (In Persian). https://doi.org/10.22059/jphgr.2012.30246
Mosavi A, Ozturk P, Chau KW. 2018. Flood prediction using machine learning models: Literature Review. Water. 10(11): 1536.‏ https://doi.org/10.3390/w10111536
Nachappa TG, Piralilou ST, Gholamnia K, Ghorbanzadeh O, Rahmati O, Blaschke T. 2020. Flood Susceptibility Mapping with Machine Learning, Multi-Criteria Decision Analysis and Ensemble Using Dempster Shafer Theory. Journal of Hydrology. 590: 125275. https://doi.org/10.1016/j.jhydrol.2020.125275
Nhu VH, Mohammadi A, Shahabi H, Ahmad BB, Al-Ansari N, Shirzadi A, Clague JJ, Jafari A, Chen W, Nguyen H. 2020. Landslide susceptibility mapping using machine learning algorithms and remote sensing data in a tropical environment. International Journal of Environmental Research and Public Health 17(14): 4933. https://doi.org/10.3390/ijerph17144933
Pham BT, Avand M, Janizadeh S, Phong T V, Al-Ansari N, Ho LS, Das S, Le HV, Amini A, Bozchaloei SK, Jafari F, Prakash, I. 2020. GIS based hybrid computational approaches for flash flood susceptibility assessment. Water. 12(3):683.‏ https://doi.org/10.3390/w12030683
Pourghasemi HR, Moradi HR, Aghda SMF, Gokceoglu C, Pradhan B. 2014. GIS-based landslide susceptibility mapping with probabilistic likelihood ratio and spatial multi-criteria evaluation models (North of Tehran, Iran). Arabian Journal of Geosciences. 7: 1857–1878. https://doi.org/10.1007/s12517-012-0825-x
Pourtaghi ZS, Pourghasemi HR. 2015. GIS-based groundwater spring potential assessment and mapping in the Birjand Township. Southern Khorasan Province, Iran. Hydrogeology Journal. 22: 643–662. https://doi.org/10.1007/s10040-013-1089-6
Pradhan B. 2009. Flood susceptible mapping and risk area delineation using logistic regression, GIS and remote sensing. Journal of Spatial Hydrology. 9(2): 1–18. https://scholarsarchive.byu.edu/josh/vol9/iss2/4
Rahmati O, Darabi H, Panahi M, Kalantari Z, Naghibi SA, Ferreira CS, Kornejady A, Karimidastenaei Z, Mohammadi F, Stefanidis S, Tien Bui D. 2020. Development of novel hybridized models for urban flood susceptibility mapping. Scientific Reports. Jul 31. 10(1): 12937. https://doi.org/10.1038/s41598-020-69703-7
Rahmati O, Haghizadeh A, Stefanidis S. 2016a. Assessing the accuracy of GIS-based analytical hierarchy process for watershed prioritization. Gorganrood River Basin, Iran. Water Resources Management. 30(3): 1131–1150. https://doi.org/10.1007/s11269-015-1215-4
Rahmati O, Pourghasemi HR, Zeinivand H. 2016b. Flood susceptibility mapping using frequency ratio and weights-of-evidence models in the Golastan Province, Iran. Geocarto International. 31(1): 42–70.‏ https://doi.org/10.1080/10106049.2015.1041559
Rozalis S, Morin E, Yair Y, Price C. 2010. Flash flood prediction using an uncalibrated hydrological model and radar rainfall data in a Mediterranean watershed under changing hydrological conditions. Journal of Hydrology. 394: 245–255. https://doi.org/10.1016/j.jhydrol.2010.03.021
Sajedi-Hosseini F, Malekian A, Choubin B, Rahmati O, Cipullo S, Coulon F, Pradhan B. 2018. A novel machine learning-based approach for the risk assessment of nitrate groundwater contamination. Science of the Total Environment. 644: 954–962.‏ https://doi.org/10.1016/j.scitotenv.2018.07.054
Saleh A, Tan ML, Nazri FM, Yaseen ZM, Zhang F. 2025. Advancing tropical flood susceptibility mapping with multi-head attention mechanism and deep learning models. Advances in Space Research. https://doi.org/10.1016/j.asr.2025.07.022
Shahabi H, Shirzadi A, Ghaderi K, Omidvar E, Al-Ansari N, Clague JJ, Geertsema M, Khosravi K, Amini A, Bahrami S, Rahmati O. 2020. Flood detection and susceptibility mapping using sentinel-1 remote sensing data and a machine learning approach: Hybrid intelligence of bagging ensemble based on k-nearest neighbor classifier.  Remote Sensing. 12 (2): 266. https://doi.org/10.3390/rs12020266
Shirani K, Sadri MA. 2023. Prioritizing effective factors and zoning of flood events using probabilistic models and GIS in the Gorganrood watershed of Golestan province. Watershed Management Research. 36(1): 111–131. (In Persian). https://doi.10.22092/wmrj.2022.358648.1473
Shirani K, Pasandi M, Arabameri A. 2018. Landslide susceptibility assessment by dempster–shafer and index of entropy models, Sarkhoun basin, southwestern Iran. Natural Hazards. 93: 1379–1418. https://doi.org/10.1007/s11069-018-3356-2
Shirani K, Peyrowan HR, Naderi Samani R. 2024. Applied principles of zonation and prediction landslide susceptibility (Machine Learning and Programming in MATLAB). Soil Conservation and Watershed Management Institute. 259 p. (In Persian).
Tehrany MS, Pradhan B, Jebur MN. 2014. Flood susceptibility mapping using a novel ensemble weights-of-evidence and support vector machine models in GIS. Journal of Hydrology. 512: 332–343. https://doi.org/10.1016/j.jhydrol.03.008
Vafakhah M, Loor SMH, Pourghasemi HR, Katebikord A. 2020. Comparing performance of random forest and adaptive neuro-fuzzy inference system data mining models for flood susceptibility mapping. Arabian Journal of Geosciences. 13: 417. https://doi.org/10.1007/s12517-020-05363-1
Wang Y, Fang Z, Wang M. 2020. Comparative study of landslide susceptibility mapping with different recurrent neural networks. Computers and Geosciences. 138: 104445. https://doi.org/10.1016/j.cageo.2020.104445
Zhang D, Yang J, Li F, Han S, Qin L. 2022. Landslide risk prediction model using an attention-based temporal convolutional network connected to a recurrent neural network. IEEE Access. 10: 37635–37645. https://doi.org/10.1109/ACCESS.2022.3165051
Zhang YG, Tang J, He ZY. 2021. A novel displacement prediction method using gated recurrent unit model with time series analysis in the Erdaohe landslide. Nat Hazards. 105: 783–813. https://doi.org/10.1007/s11069-020-04337-6
Zhao ZA, He Y, Yao S, Yang W, Wang W, Zhang L, Sun Q. 2022. A comparative study of different neural network models for landslide susceptibility mapping. Advances in Space Research. 70(2): 383–401.‏ https://doi.org/10.1016/j.asr.2022.04.055
Watershed Management Research
Volume 39, Issue 3 - Serial Number 152
September 2026
Pages 22-44

Files

History

  • Receive Date: 09 December 2025
  • Revise Date: 18 February 2026
  • Accept Date: 21 June 2026

Share

How to cite

Statistics