Flood Susceptibility Mapping in Kolhapur City

Strategic mitigation of monsoon-induced inundation re-quires predictive models capable of processing massive, non-linear geographical matrices. This research engineered a rig-orous, standardized architecture to test statistical and machine learning paradigms within Kolhapur city. By strictly control-ling the preprocessing pipeline and validation constraints, the study conclusively proves that ensemble algorithmic models provide superior spatial intelligence compared to traditional methodologies. While the Frequency Ratio Model generated a mathemat-ically functional baseline (89.20%), it proved incapable of mapping heterogeneous local hazards. Support Vector Machine mapping (79.17%) suffered from hyper-dimensional scaling issues, producing fragmented risk zones. The Random For-est architecture, however, demonstrated exceptional predictive resilience, registering the highest accuracy (90.28%) and an ROC-AUC of 0.93. The resulting susceptibility map provides local government authorities with verified, actionable spatial data required to regulate future urban zoning and engineer targeted drainage infrastructure. To advance regional hazard resilience, future modeling efforts must transition from static topological susceptibility to-ward dynamic hydro-meteorological forecasting [8]. Integrat-ing the optimized Random Forest architecture with real-time stream gauge telemetry and dynamic satellite soil-moisture indices would facilitate a live early-warning system, drastically reducing the latency of municipal disaster response protocols during peak monsoon periods.

[1]     S. K. Powar and S. S. Panhalkar, “Flood-susceptibility analysis of Kolha-pur City using Frequency Ratio Model,” Geomatics and Environmental Engineering, vol. 18, no. 6, pp. 23-45, 2024. [2]     S. S. Panhalkar and S. K. Powar, “GIS-based flood susceptibility mapping in Kolhapur region,” International Journal of Advanced Remote Sensing and GIS, 2022. [3]     H. Arabameri, B. Pradhan, A. Rezaei, K. Lee, and N. Saro, “GIS-based flood susceptibility mapping using frequency ratio and evidential belief function models,” Environmental Earth Sciences, vol. 78, no. 21, 2019. [4]     T. M. Asrade, S. A. Abebe, K. B. Tadesse, M. S. Kerebih, and T. M. Meshesha, “Flood susceptibility assessment using three machine learning techniques and comparison of their performance,” Scientific Reports, vol. 16, no. 1, art. no. 8099, 2026. [5]     K. Khosravi, B. T. Pham, I. Prakash, and D. T. Bui, “A comparative assessment of decision trees algorithms for flash flood susceptibility modeling at Haraz watershed, Iran,” Science of the Total Environment, vol. 627, pp. 744-755, 2018. [6]     D. T. Bui, P. Tsangaratos, V. T. Nguyen, and K. Khosravi, “Comparative analysis of machine learning methods for spatial prediction of flood-prone areas,” ISPRS International Journal of Geo-Information, vol. 9, no. 3, 2020. [7]     A. Riche, A. Drias, M. Guermoui et al., “A novel hybrid deep-learning approach for flood-susceptibility mapping,” Remote Sensing, vol. 16, no. 19, art. no. 3673, 2024. [8]     O. Rahmati, H. R. Pourghasemi, and H. Zeinivand, “Flood susceptibility mapping using frequency ratio and weights-of-evidence models,” Geo-matics, Natural Hazards and Risk, vol. 7, no. 5, pp. 1639-1661, 2016. [9]     M. S. Tehrany, B. Pradhan, and M. N. Jebur, “Flood susceptibility mapping using a novel ensemble weights-of-evidence and support vector machine models,” Journal of Hydrology, vol. 512, pp. 332-343, 2014. [10] A. Mosavi, P. Ozturk, and K. W. Chau, “Flood prediction using machine learning models: Literature review,” Water, vol. 10, no. 11, art. no. 1536, 2018.