Author(s): Jose Nilson Bezerra Campos; Francisco De Assis De Souza Filho; Jose Carlos De Araujo
Linked Author(s):
Keywords: No Keywords
Abstract: An analysis of the precision of reservoir yield evaluation as a function of the coefficient of variation of annual inflows (Cv) is performed. Two aspects are involved: the size of inflow time series required for a given precision in the estimation of the population mean of annual inflow; and the variance of reservoir yield as a function of Cv and of the size of the time series. The value of Cv analyzed range from 0.2-characteristic of Northeast of the USA rivers-to 1,4-characteristic of some rivers in Brazil's semi-arid Northeast. The results showed that uncertainty and errors affect markedly rivers with high coefficient of variation. For instance, if a 30-year time series of annual inflow is good in estimating the mean of the population of annual inflows for Northeast of USA, to reach the same precision in the Northeast of Brazil it would be necessary about 690 years of data. Besides, assuming a known population mean value, the reservoir yield for Cv=0.2 is about 2.8 times greater than that for Cv=1.4. Thus, the higher the Cv, the lower the reservoir yield and the less accurate its preview.
Year: 1997