DETERMINATION OF SUITABLE PROBABILITY DISTRIBUTION MODELS FOR PEAK FLOOD LEVEL SERIES OF SELECTED GAUGING STATIONS IN THE NIGER AND BENUE RIVERS

Publication Date : 01/02/2008


Author(s) :

W. A Salami, N. A Egharevba.


Volume/Issue :
Volume 3
,
Issue 1
(02 - 2008)



Abstract :

Probability distribution models were developed for forecasting of flood level along Niger River and its major tributary, Benue River in Nigeria. Nine gauging stations were selected along the river channels; annual peak flood levels were selected and subjected to five different probability distribution analyses to determine the best fit probability functions for each gauging station. The probability distribution models adopted are Gumbel, Log-Gumbel, Log-Normal, Log-Pearson and Exponential distribution functions. Mathematical equations were established and used to predict values of flood levels. Statistical tests comprising of chi-square, student’s t-distribution, correlation coefficient, coefficient of determination, and standard error of estimate were carried out to determine the reliability of the predicted values. The model that satisfies the statistical tests mostly was selected as the best fit model. The study revealed that the peak flood levels at Idah, Yola, Garua and Dadinkowa are best fitted with Gumbel probability distribution model, while at Jebba and Lokoja the best fit probability model is Log-Gumbel. At Ibi the best fit probability model is Log-Normal, while at Makurdi and Onitsha the best fit probability model is Log-Pearson.


No. of Downloads :

0


About BayeroJet

The new Bayerojet Journal is designed to be able to manage the increasing number of published articles. The new system will allow the publishers as well as the Bayerojet team to make publishing more efficient. If you wish to publish an article it will be very easy. All you have to do is to submit your paper online and wait to the review before it will be finally published. You can manage your articles and send new versions at any time. Browse through our page to find out more.