Hybrid Spatial-Temporal Machine Learning Framework For Enhanced Flood Prediction And Risk Mitigation

K.Lakshman Kumar; Dr. Ranga Swamy Sirisati

doi:10.29284/ece1jm33

Authors

K.Lakshman Kumar Research Scholar, Department of CSE, Bharatiya Engineering Science and Technology Innovation University(BESTIU), Gownivaripaill, Gorantla in Andhra Pradesh
Dr. Ranga Swamy Sirisati Associate Professor & HOD, Department of Information Technology, Vignan's institute of management and technology for women Kondapur Ghatkesar Hyderabad-501301

DOI:

https://doi.org/10.29284/ece1jm33

Keywords:

Flood Prediction, Machine Learning, Spatial-Temporal Model, Geographic Information Systems (GIS), Cascade Forest Model (CFM), Long Short-Term Memory (LSTM), Real-Time Data Integration, Susceptibility Mapping, Hybrid Model, Disaster Risk Management.

Abstract

Flood prediction and susceptibility mapping are vital for mitigating the impacts of one of the most destructive natural disasters. This study presents a comprehensive analysis of machine learning models for flood prediction, focusing on the comparative performance of the Cascade Forest Model (CFM), Long Short-Term Memory (LSTM) networks, Gradient Boosting, and a proposed Hybrid Spatial-Temporal Model. Using datasets comprising historical flood records, satellite imagery (Sentinel-1 and Sentinel-2), and IoT sensor data, we evaluated the models on metrics such as Overall Accuracy (OA), Area Under the Curve (AUC), Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE). The proposed Hybrid Spatial-Temporal Model achieved the highest performance, with an accuracy of 95% and an AUC of 0.96, surpassing CFM (93% accuracy, AUC =0.95 ) and Gradient Boosting ( 87.5% accuracy, AUC = 0.90). The LSTM model demonstrated exceptional temporal prediction capabilities, with an RMSE of 0.10 and an MAE of 0.06 . Comparative analysis with recent models, such as LightGBM (99.75% accuracy, AUC = 0.99) and ANN (90.7% accuracy, AUC = 0.871), underscores the Hybrid Model's superior adaptability across spatial and temporal domains. The integration of GIS-based spatial analysis and LSTM's temporal prediction significantly improved prediction robustness and generalizability. Feature selection revealed rainfall and elevation as critical factors influencing flood susceptibility, while IoT-enhanced real-time data integration boosted prediction accuracy by 15%. These findings highlight the importance of combining spatial and temporal insights for effective flood risk management. This study provides actionable insights for developing real-time flood prediction systems that prioritize accuracy, interpretability, and scalability.

References

[1]. Lakshman, K., and Ranga Swamy Sirisati. "Development and Comparative Analysis of Advanced Machine Learning Algorithms for Flood Prediction and Susceptibility Mapping." NeuroQuantology, vol. 22, no. 4, 2024, pp. 201-214, doi:10.48047/nq.2024.22.4.nq24025.

[2]. Rajab, Adel, et al. "Flood Forecasting by Using Machine Learning: A Study Leveraging Historic Climatic Records of Bangladesh." Water, vol. 15, no. 22, 2023, pp. 3970-3985.

[3]. Wu, Xuan, et al. "A Near-Real-Time Flood Detection Method Based on Deep Learning and SAR Images." Remote Sensing, vol. 12, no. 1, 2020, pp. 101-120.

[4]. Prakash, Indra, et al. "Application of GIS and Machine Learning to Predict Flood Areas in Nigeria." International Journal of Disaster Risk Reduction, vol. 46, 2021, pp. 101567-101580.

[5]. Zhang, Li, et al. "Improving Flood Prediction with Gated Recurrent Units and Hydrological Data." Journal of Hydrology, vol. 603, 2023, pp. 126-140.

[6]. Li, Xiang, et al. "Flood Susceptibility Mapping Using CatBoost and Geospatial Analysis." Environmental Modelling & Software, vol. 145, 2022, pp. 105-120.

[7]. Adel, Mohammad, et al. "Random Forest and SVM Models for Flood Forecasting: A Comparative Study." IEEE Access, vol. 10, 2023, pp. 57045-57058.

[8]. Hassan, Amir, et al. "A Hybrid Decision Tree Approach for Flood Risk Prediction in ResourceConstrained Environments." Water Resources Management, vol. 35, no. 4, 2021, pp. 1225-1240.

[9]. Fernandes, A. A. T., et al. "Logistic Regression for Flood Prediction in Low-Resource Regions." Revista de Sociologia e Política, vol. 28, no. 2, 2021, pp. 120-135.

[10]. Baalaji, Vignesh, and Sandhya A. "Flood Prediction System Using Multilayer Perceptron Classifier and Neural Networks." IRJET, vol. 7, no. 5, 2020, pp. 1000-1010.

[11]. Miah, Mohammad, et al. "Machine Learning Models for Flood Risk Prediction: A Case Study of Bangladesh." International Journal of Advanced Computer Science and Applications, vol. 11, no. 4, 2020, pp. 45-55.

[12]. Lin, Yan-Ting, et al. "Quantifying Flood Water Levels Using Image-Based Volunteered Geographic Information." Remote Sensing, vol. 12, no. 6, 2020, pp. 3200-3220.

[13]. Sankaranarayanan, S., et al. "Deep Learning Techniques for Weather-Based Flood Prediction." Journal of Water and Climate Change, vol. 11, no. 4, 2019, pp. 1766-1783.

[14]. Mohammed, Khalaf, et al. "Flood Severity Prediction Using Machine Learning Algorithms." IEEE Congress on Evolutionary Computation (CEC), vol. 10, no. 2, 2018, pp. 4050-4065.

[15]. Halgamuge, Malka, and Thas Nirmalathas. "Analysis of Large Flood Events in Australia and India (1985-2016)." International Journal of Disaster Risk Reduction, vol. 30, 2017, pp. 120-132.

[16]. Lin, Wen-Hao, et al. "Geospatial and Remote Sensing Techniques in Flood Mapping: A Machine Learning Perspective." Journal of Environmental Informatics, vol. 41, no. 1, 2022, pp. 24-40.

[17]. Kumar, Anil, et al. "Mapping a Novel Metric for Flash Flood Recovery Using Interpretable Machine Learning." International Journal of Hydrology Science and Technology, vol. 12, no. 2, 2020, pp. 145159.

[18]. Dile, Yihun Taddele, and Raghavan Srinivasan. "Evaluation of Climate Data for Hydrologic Prediction in Scarce Watersheds." Journal of the American Water Resources Association, vol. 50, no. 5, 2014, pp. 1225-1240.

[19]. Swamy, Sirisati Ranga, et al. "Dimensionality Reduction Using Machine Learning for Flood Susceptibility." International Journal of Innovative Technology and Exploring Engineering, vol. 9, no. 2, 2019, pp. 1740-1745.

[20]. Xuan, Tao, et al. "Positive-Unlabeled Learning for Urban Flood Susceptibility." Environmental Earth Sciences, vol. 79, no. 6, 2020, pp. 1-10.

[21]. Sirisati, Ranga Swamy, et al. "Cancer Sight: Illuminating the Hidden-Advancing Breast Cancer Detection with Machine Learning-Based Image Processing Techniques." 2023 International Conference on Sustainable Communication Networks and Application (ICSCNA). IEEE, 2023.

[22]. Swamy, Sirisati Ranga, et al. "Multi-Features Disease Analysis Based Smart Diagnosis for COVID-19." Comput. Syst. Sci. Eng. 45.1 (2023): 869-886.

[23]. Sirisati, Ranga Swamy, et al. "An Enhanced Multi Layer Neural Network to Detect Early Cardiac Arrests." 2021 5th International Conference on Electronics, Communication and Aerospace Technology (ICECA). IEEE, 2021.

[24]. Nagasri, Devara, et al. "Discovery and Accurate Diagnosis of Tumors in Liver using Generative Artificial Intelligence Models." Journal of Next Generation Technology (ISSN: 2583-021X) 4.2 (2024).

[25]. Srinivasa Rao, P. S. V., Mekala Srinivasa Rao, and Ranga Swamy Sirisati. "A Hybrid Clinical Data Predication Approach Using Modified PSO." Smart Computing Techniques and Applications: Proceedings of the Fourth International Conference on Smart Computing and Informatics, Volume 2. Springer Singapore, 2021.

[26]. SwamySirisati, Ranga, Mekala Srinivasa Rao, and Srinivasulu Thonukunuri. "Analysis of Hybrid Fusion-Neural Filter Approach to detect Brain Tumor." 2020 Sixth International Conference on Parallel, Distributed and Grid Computing (PDGC). IEEE, 2020.