Please use this identifier to cite or link to this item:
Title: Analysis of Stochastic characteristics of the Benue River flow process
Other Titles: NONE
Authors: Otache, Martins Yusuf
Keywords: trend
over time
over realisation
Issue Date: 2008
Publisher: CHIN. J. Oceanol. & Limnol.
Citation: Martins Y. Otache., Bakir, M., and Zhijia, L. (2008a). Analysis of Stochastic characteristics of the Benue River flow process, CHIN. J. Oceanol. & Limnol. Vol. 26(2), pp: 142-151.
Abstract: Stochastic characteristics of the Benue River streamflow process are examined under conditions of data austerity. The streamflow process is investigated for trend, non-stationarity and seasonality for a time period of 26 years. Results of trend analyses with Mann-Kendall test show that there is no trend in the annual mean discharges. Monthly flow series examined with seasonal Kendall test indicate the presence of positive change in the trend for some months, especially the months of August, January, and February. For the stationarity test, daily and monthly flow series appear to be stationary whereas at 1%, 5%, and 10% significant levels, the stationarity alternative hypothesis is rejected for the annual flow series. Though monthly flow appears to be stationary going by this test, because of high seasonality, it could be said to exhibit periodic stationarity based on the seasonality analysis. The following conclusions are drawn: (1) There is seasonality in both the mean and variance with unimodal distribution. (2) Days with high mean also have high variance. (3) Skewness coefficients for the months within the dry season period are greater than those of the wet season period, and seasonal autocorrelations for streamflow during dry season are generally larger than those of the wet season. Precisely, they are significantly different for most of the months. (4) The autocorrelation functions estimated “over time” are greater in the absolute value for data that have not been deseasonalised but were initially normalised by logarithmic transformation only, while autocorrelation functions for i = 1, 2, …, 365 estimated “over realisations” have their coefficients significantly different from other coefficients.
Description: NONE
Appears in Collections:Agric. and Bioresources Engineering

Files in This Item:
File Description SizeFormat 
s00343-008-0142-0.pdf298.11 kBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.