Evaluating Flood Susceptibility in the Brahmaputra River Basin: An Insight into Asia's Eastern Himalayan Floodplains Using Machine Learning and Multi-Criteria Decision-Making
Authors:
- Jatan DebnathJDJatan Debnath
- Dhrubajyoti SahariahDSDhrubajyoti Sahariah
- Meghna MazumdarMMMeghna Mazumdar
- Durlov LahonDLDurlov Lahon
- Gowhar MerajGowhar Meraj
- Shizuka HashimotoShizuka Hashimoto
- Pankaj KumarPankaj Kumar
- Suraj Kumar SinghSuraj Kumar Singh
- Shruti KangaShruti Kanga
- Kesar ChandKCKesar Chand
- Anup SaikiaASAnup Saikia
Open Access
- 8 December 2023
- journal article
- research article
- Published by Springer Nature in Earth Systems and Environment
- Vol. 7 (4) , 733-760
Abstract
Floods represent a significant threat to human life, property, and agriculture, especially in low-lying floodplains. This study assesses flood susceptibility in the Brahmaputra River basin, which spans China, India, Bhutan, and Bangladesh—an area notorious for frequent flooding due to the saturation of river water intake capacity. We developed and evaluated several innovative models for predicting flood susceptibility by employing Multi-Criteria Decision Making (MCDM) and Machine Learning (ML) techniques. The models showed robust performance, evidenced by Area Under the Receiver Operating Characteristic Curve (AUC-ROC) scores exceeding 70% and Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE) scores below 30%. Our findings indicate that approximately one-third of the studied region is categorized as moderately to highly flood-prone, while over 40% is classified as low to very low flood-risk areas. Specific regions with high to very high flood susceptibility include Dhemaji, Dibrugarh, Lakhimpur, Majuli, Darrang, Nalbari, Barpeta, Bongaigaon, and Dhubri districts in Assam; Coochbihar and Jalpaiguri districts in West Bengal; and Kurigram, Gaibandha, Bogra, Sirajganj, Pabna, Jamalpur, and Manikganj districts in Bangladesh. Owing to their strong performance and the suitability of the training datasets, we recommend the application of the developed MCDM techniques and ML algorithms in geographically similar areas. This study holds significant implications for policymakers, regional administrators, environmentalists, and engineers by informing flood management and prevention strategies, serving as a climate change adaptive response within the Brahmaputra River basin.Keywords
Funding Information
- The University of Tokyo
This publication has 91 references indexed in Scilit:
- Response of River Flow Regime to Various Climate Change Scenarios in Ganges-Brahmaputra- Meghna BasinJournal of Water Resources and Ocean Science, 2013
- Classifiers for the detection of flood-prone areas using remote sensed elevation dataJournal of Hydrology, 2012
- Optimization of Weighted Aggregated Sum Product AssessmentElektronika ir Elektrotechnika, 2012
- A Review of Risk Perceptions and Other Factors that Influence Flood Mitigation BehaviorRisk Analysis, 2012
- Manifestation of remote sensing data and GIS on landslide hazard analysis using spatial-based statistical modelsArabian Journal of Geosciences, 2009
- Delineation of landslide hazard areas on Penang Island, Malaysia, by using frequency ratio, logistic regression, and artificial neural network modelsEnvironmental Earth Sciences, 2009
- Compromise solution by MCDM methods: A comparative analysis of VIKOR and TOPSISEuropean Journal of Operational Research, 2004
- Floods of the Brahmaputra River in IndiaWater International, 1997
- The Nature of Statistical Learning TheoryTechnometrics, 1996
- Multiobjective optimization in river basin developmentWater Resources Research, 1980