The Effects of Calibration Parameters in Muskingum Models on Flood Prediction Accuracy

Abstract

Attenuation, time lag, outflow peak and storage are very essential factors required in flood risk prediction and flood pattern. However, the accurate prediction strongly depends on appropriate calibration of routine parameters of the model, such as weighting factor (x) and storage time constant (K). The weighting factor being used to determine storage time constant has not been given consideration in the previous studies and this could have led to inaccurate prediction in the past. In this work, a set of data obtained from an ungauged Awara river in Ondo State, Nigeria were used to test the effects of a weighting factor, x at levels ranging from 0.1-0.5 at interval of 0.1. The Muskingum model was used to obtain the storage and weighted discharge storage. It was observed that the correlation coefficient (R2) decreases with an increase in the weighting factor (x). This implies that there is a strong relationship between storage and weighted discharge storage at 0.1-0.3 levels of x while, the relationship is fair at 0.4-0.5 levels. It is therefore appropriate to choose a value of x ranging between 0.1 and 0.3 for attenuation prediction, while values of x ranging between 0.4 and 0.5 would be appropriate for accurate prediction of both outflow peak and storage.