Spatial and Seasonal Patterns of Flood Inundation in Lokoja, Kogi State, Nigeria

Abstract

The study examines spatial patterns of flood inundation in Lokoja, Kogi state, Nigeria. Maximum Likelihood Classifier algorithm of the supervised land use/cover classification technique was utilized. The results obtained from the analysis were used to estimate the magnitude and visualize the seasonal and spatial pattern of flood inundation event. Eight Landsat Images comprising of two sets for each year (dry and wet seasons) were acquired from the portal of United States Geological Survey (2018). The Landsat images were classified into land cover classes such as Built Up, Vegetation and Water Body. After completing the land cover classification, the area of each class was determined and converted to square kilometers and percentages for both wet and dry seasons. Based on the classification, the brown colour depicts the built-up areas, blue for water body, and green for vegetation. Finally, accuracy assessment was carried out using historical Google Earth images, informed knowledge of the area, and GPS coordinates. ArcMap 10.5 was used to produce land use/cover maps for the study period. The result overall, revealed the effect of flood inundation to be more intense on vegetation. 1.62%, 4.60%, 23.05% and 6.43% of vegetated land was lost in 1999, 2009, 2012 and 2018, respectively.  Therefore, efforts to improve resilience against variable weather, flood inundation and seasonal uncertainties should be encouraged.