In hydrologic time series analysis, the seasonal component is estimated by using averages of monthly values or by Fourier analysis. Useful as these methods have been, they have not been investigated in depth. Estimation of seasonal components in econometrics, on the other hand, has received considerable attention. The main reason for this increased attention is that econometric time series are nonstationary and are often more variable and complex than hydrologic time series. Recently, Jaynes (1985) has argued that informative priors can have great effect on the estimation of seasonal components and of noise. Jaynes'(1985) method has been analyzed in this study and several properties of his estimates are discussed. The method has been applied to estimate Seasonal Components of monthly river flow series.
Relation:
Proceedings of the Sixth IAHR International Symposium on Stochastic Hydr=aulics，頁767-774
