Geospatial And Machine Learning-Based Assessment Of Flood Vulnerability In Lokoja, Kogi State, For Sustainable Infrastructure Planning

Abstract

Flooding remains one of the most devastating environmental hazards affecting rapidly urbanising cities in developing countries, particularly within riverine floodplains where urban expansion often occurs without adequate planning. In Nigeria, recurrent flooding continues to threaten urban infrastructure, livelihoods, and environmental sustainability. Lokoja, the capital of Kogi State, is particularly vulnerable due to its location at the confluence of Rivers Niger and Benue, making it one of the most flood-prone urban centres in the country. This study applies an integrated geospatial and machine learning approach to assess flood vulnerability in Lokoja and support sustainable infrastructure planning. Multi-source geospatial datasets, including Shuttle Radar Topography Mission Digital Elevation Model, Sentinel-2 satellite imagery, and Climate Hazards Group InfraRed Precipitation with Station data, were utilised. Flood conditioning factors, including elevation, slope, distance to drainage channels, land use/land cover, drainage density, rainfall intensity, flow accumulation, and vegetation condition, were derived and processed within a GIS environment. These variables were incorporated into a Random Forest machine learning model to generate a predictive flood vulnerability map for the study area. The model achieved moderate predictive performance with an overall validation accuracy of approximately 65%. Variable importance analysis indicated that distance to drainage and elevation were the dominant flood-conditioning factors. The results show that low-lying floodplain areas close to the Niger-Benue river channels exhibit the highest flood vulnerability. In contrast, several built-up areas and infrastructure corridors fall within moderate to very high vulnerability zones. The study demonstrates the value of integrating geospatial analysis and machine learning for flood-risk assessment in data-constrained urban environments. The findings provide spatial decision-support information for urban planners, policymakers, and disaster management agencies seeking to promote climate-resilient infrastructure planning and sustainable urban development in Lokoja.