Classification and Prediction of Channel Morphology Within Selected Third-Order Basins (Southwestern Nigeria)

Abstract

Process-form interactions as a means through which classification and prediction are better presented across river systems have received attention in mountainous regions and extra-tropical areas with (non) glacial history. However, in the low-relief areas without glacial history, detailed descriptions of basins are not as popular and widely presented as their counterparts in other climes. This research seeks to classify and predict river systems by using the Rosgen and Montgomery and Buffington framework. Seventeen third-order basins were extracted from digital elevation models (DEMs). They were validated by using topographical maps. Ten basins were picked systematically. Six were randomly selected out of them. Eighteen morphological variables using standard procedures and hydraulic equations were measured and derived at bankfull. Alongside, domains were identified and described within each basin. Step-wise discriminant analysis was performed on the entire dataset in order to identify the variables that can successfully classify the channel types. The result produced three dominant explanatory variables which are cross-sectional area, hydraulic radius and width (CA–W–HR). These variables successfully predicted 53% of the domains across the dataset. Hierarchical cluster analysis was performed by using CA–W–HR. The iteration produced five distinct groups: plane-bed-bedrock channels, partly-confined mixed channels, partly-confined bedrock channel, unconfined alluvial channels and partly-confined mixed channels void of step-pools.