2019· International Journal of Modern Research in Science & Engineering· Vol 2, pp. 1-14· 0 citations
TL;DR
A flood prediction framework that integrates Geographic Information Systems (GIS) with Machine Learning (ML) and results indicate that hybrid GIS-ML models outperform traditional approaches by effectively capturing non-linear relationships between hydrometeorological and urban factors.
Abstract
Cities flooding is a major hydrological risk in rapidly urbanizing regions, intensified by climate change, extreme rainfall, and inadequate drainage systems. Traditional hydrological models require extensive calibration and high-resolution data, which are often unavailable for many cities. To address this issue, this study proposes a flood prediction framework that integrates Geographic Information Systems (GIS) with Machine Learning (ML). Multi-source geospatial data such as digital elevation models (DEM), land use–land cover (LULC), soil properties, drainage density, rainfall intensity, and historical flood records are used to derive flood conditioning factors in a GIS environment. These factors are then applied as inputs for supervised ML algorithms including Artificial Neural Networks (ANN), Random Forest (RF), Support Vector Machine (SVM), and Gradient Boosting. Model performance is evaluated using RMSE, MAE, R², and ROC–AUC metrics. Results indicate that hybrid GIS-ML models outperform traditional approaches by effectively capturing non-linear relationships between hydrometeorological and urban factors. The framework provides a reliable decision-support tool for urban planners and disaster management authorities to identify flood-prone areas, improve drainage planning, and enhance early warning systems.
Flood occurrence in tropical regions is intensifying due to climate variability and land-use change, increasing the need for reliable flood response time estimation. Accurate prediction of flood lag time (TL)—the interval between the centroid of excess rainfall and peak runoff—is critical for flood early warning and water resource planning. However, TL estimation remains challenging in data-scarce regions because of complex interactions among watershed morphology, rainfall characteristics, and runoff generation processes. This study evaluates four machine learning (ML) algorithms—Random Forest (RF), Artificial Neural Network (ANN), Extreme Gradient Boosting (XGBoost), and Support Vector Machine (SVM)—for predicting TL across twenty gauged watersheds in the Blue Nile Basin of Ethiopia. Fourteen physiographic and hydro-climatic watershed characteristics were used as predictors. Among the tested models, XGBoost achieved the highest training performance (R2 = 0.98, NSE = 0.96), while RF showed better generalization in the test dataset (R2 = 0.77, NSE = 0.70, KGE = 0.71). SVM produced the lowest prediction errors (MAE = 0.95; RMSE = 2.25) but had lower explanatory power (R2 = 0.49). To enhance interpretability and practical applicability, ML-based feature importance was used to develop a parsimonious empirical model: TL = 0.8 + 0.011A − 0.023RI, where A is watershed area and RI is rainfall intensity. This model explained 51% of TL variability and retained much of the predictive skill of more complex ML models. The proposed hybrid ML–empirical framework provides a transparent and operational approach for flood response time estimation in tropical highland watersheds. Its broader applicability remains subject to additional watershed-level validation and regional calibration.
Dagnenet Sultan, N. Haregeweyn, M. Tsubo et al.· Water· 0 citations
Urban flooding has become a growing concern due to rapid land development and increasingly unpredictable climate patterns. Effective emergency response—particularly the strategic placement of temporary shelters—is crucial for minimizing the impacts of disasters. This study presents a data-driven framework for identifying and spatially optimizing emergency shelters in flood-prone urban environments. The methodology was implemented in a major metropolitan area in southern Iran, which was selected as a representative case due to its high flood risk and urban complexity. Flood susceptibility was modeled using 16 explanatory variables encompassing climatic, geographical, urban development, and urban infrastructure factors. Six machine learning algorithms were tested and compared: Random Forest, K-Nearest Neighbors (KNN), Gradient Boosting Machine (GBM), eXtreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), and Categorical Boosting (CatBoost). Among them, CatBoost achieved the strongest test performance, with an F1-score of 0.8627 and an AUC of 0.9218, and was used to generate the final flood-susceptibility map. Buildings with below-median building-level susceptibility scores were retained and spatially allocated using K-Means clustering. The selected k = 28 was supported by cluster-validity and stability analyses, with a Silhouette Score of 0.5991, a Davies–Bouldin Index of 0.4578, and a mean adjusted Rand index of 0.8938. A supplementary route-exposure assessment further identified access-route segments intersecting very high flood-susceptibility zones. The proposed framework provides a transferable decision-support approach for prioritizing preliminary flood-emergency shelter.
Zahra Babaei, M. S. Mesgari, Hamed Zibaei et al.· Natural Hazards· 0 citations
The southern part of India, Kochi, is more vulnerable to flooding due to changing weather patterns, low-lying terrain, and fast urban growth. In this research, the likelihood of a flood occurrence using GIS and a hybrid forecasting approach that combines AHP and ANN were evaluated. Elevation, slope, stream density, Land Use and Land Cover (LULC), Topographic Wetness Index (TWI), and rainfall were all included in the weighted overlay approach used to calculate the Flood Vulnerability Index (FVI). To understand the blended risk assessments, an artificial neural network was trained using the Levenberg-Marquardt approach, AHP-derived weights, and the same physical elements. The ANN findings are dependable since the model generated accurate predictions and had a low mean squared error (R = 0.989). The cities, including Kaloor, Fort Kochi, and Palarivattom, are situated in all low-lying regions with a large population and inadequate sanitary facilities, putting the residents at serious danger. Combining the two results in a risk map that is more accurate than either AHP or ANN alone. Future attempts to strengthen communities, lessen the threat of storms, and improve the environmental quality of coastal regions may benefit from this.
Ankita Saxena, Yogesh M. Keskar, P. Raju et al.· International Journal of Civ...· 0 citations
Floods represent a major hazard in India’s sub-Himalayan regions, with recurrent events in Tripura causing significant socio-economic and infrastructural losses. The August 2024 flood in Gomati district highlighted the urgent need for accurate flood susceptibility mapping to guide risk mitigation and disaster preparedness. This study integrates remote sensing (RS), geographic information system (GIS) and machine learning (ML) approaches to assess flood susceptibility. Sentinel-1 synthetic aperture radar (SAR) data were used to map inundated areas, overcoming limitations of cloud cover in optical imagery. A comprehensive geospatial database was compiled, incorporating topographic, hydrological, climatic, land-use, soil and geological factors. Multicollinearity testing and feature selection ensured robust and independent predictors for model training. The random forest (RF) algorithm was applied to generate flood susceptibility maps, categorizing the district into very low, low, moderate, high and very high flood-prone zones. Model validation using the receiver operating characteristic (ROC) curve yielded a high area under the curve (AUC = 0.929), indicating strong predictive capability. Results identified the highest-risk areas along the Gomati and San Ganga river plains, driven by low elevation, gentle slopes, high drainage density and unconsolidated sediments. The study offers actionable insights for policymakers, enabling targeted flood mitigation, improved land-use planning, early warning systems and community-based disaster management strategies. The integrated methodology demonstrates the potential of combining SAR data and ML techniques for reliable flood susceptibility assessment, providing a replicable framework for other flood-prone regions.
Sah Kausar Reza, J. Chakraborty, S. Chattaraj et al.· Discover Environment· 0 citations
Flooding poses an escalating threat to Kisumu County, Kenya, driven by climate change and rapid urbanization.
This study develops an integrated framework combining remote sensing, GIS, and machine learning for flood risk mapping and forecasting using multi‐source geospatial data (Sentinel‐1 SAR, Landsat 8, SRTM DEM, SMAP soil moisture) and meteorological data (2014‐2025). Four machine learning models random forest (RF), artificial neural network (ANN), deep neural network (DNN), and convolutional neural network (CNN) were developed and validated.
The ANN achieved the highest accuracy (R
2
= 0.910, RMSE = 0.038). Excluding RF (R
2
= 0.520) from the ensemble improved performance to R
2
= 0.916 (RMSE = 0.035), a 22% error reduction. Flood risk mapping classified Kisumu County into five categories, revealing that 40.9% (857.8 km
2
) faces moderate to very high risk, with critical hotspots along the Lake Victoria shoreline and in Kisumu City and Ahero. Slope (20%), NDBI (15%), soil moisture (15%), and stream proximity (15%) were identified as dominant flood drivers. Uncertainty quantification revealed low model variance (R
2
σ = 0.006) and demonstrated that aleatoric uncertainty (62%) dominates epistemic uncertainty (38%).
This framework provides a replicable, data‐driven methodology for flood risk assessment in data‐scarce regions globally.
Herine Auma, M. Gebreslasie, A. Osio· Frontiers in Environmental S...· 0 citations
Flood susceptibility mapping that integrates hydrological and geospatial information remains limited at the sub-watershed scale, particularly in rapidly developing coastal urban areas. This study maps flood susceptibility in the Batang Kandis Sub-watershed, Padang City, by integrating hydrological data through Google Earth Engine (GEE). A quantitative spatial approach was employed using six parameters: rainfall, elevation, slope, soil texture, land cover, and proximity to rivers. Each parameter was classified, scored, and weighted using the Analytic Hierarchy Process (AHP), after which a weighted overlay was applied to generate the Flood Susceptibility Index (FSI). The results indicate that 58.77% of the study area is classified as having very high flood susceptibility, while 21.39% is classified as highly susceptible. These zones are concentrated primarily in downstream areas characterized by low elevations, gentle slopes, extensive built-up surfaces, and close proximity to river networks. Validation against historical inundation data from Ina-Geoportal demonstrated 97.10% spatial correspondence with areas classified as having high or very high susceptibility. These findings confirm that integrating hydrological and geospatial parameters through GEE provides a reliable approach to identifying flood-prone areas at the sub-watershed scale. The resulting susceptibility map provides an evidence-based foundation for prioritizing flood mitigation, strengthening watershed management, and supporting adaptive spatial planning in rapidly developing coastal urban areas.
