Stochastic Five Daily Stream Flow Model

by Nath T. Kottegoda, Lectuer; Civ. Engrg. Dept., Univ. of Birmingham, Birmingham, England,

Serial Information: Journal of the Hydraulics Division, 1972, Vol. 98, Issue 9, Pg. 1469-1485

Document Type: Journal Paper


A stochastic data generating technique is applied to extend the records of British rivers. Because of the shortcomings in daily flow models, a five daily time unit is adopted. Analysis of the autocorrelations suggest that an autoregressive process is more representative of the data than other short memory types. This is confirmed by tests on the synthesized data. For the series examined, a fourth order process is adequate. The probability distributions of the independent residuals are confined to a limited range but different functions are necessary to fit individual series. The Pearson system and Johnson Type curves are examined and best fits are obtained by the Pearson Type III and Type VI and the lognormal distributions. The correlation structure is preserved in the generated data and other significant characteristics are similar to those in historical records.

Subject Headings: Streamflow | Stochastic processes | Water flow | Historic preservation | Hydrologic data | Rivers and streams | Flow duration | Autoregressive models

Services: Buy this book/Buy this article


Return to search