Main Article Content
Evaluation of probabilistic models for characterizing design low-flows of River Ogun, Southwest Nigeria
Abstract
Information of low-flow is important for maintaining instream flow, conserving biodiversity, enhancing food production, industrial abstraction, tourism and dilution of effluents from industries and households. This study establishes suitable probabilistic models for characterizing different durations of design low-flows of Ogun River. The adequacy of fit of four probability distributions, namely Reversed Generalized Extreme Value (GEVR) distribution, Generalized Normal (GNO) distribution, Generalized Logistic (GLO) distribution and Pearson Type III (PE3) were evaluated using the Anderson-Darling (A2) goodness-of-fit statistic and the D-index diagnostic test. The study revealed that GLO is best suited for predicting the annual minimal, 3-day minima, 7-day minima, 10-day minima, 15-day minima and 30-day minima based on the A2 and D-index values. Six mathematical models derived from probability plots were established to relate the different low-flow series to their non-exceedance probability. The models could be used for characterizing low-flows and for water resources management of Ogun River Basin.