Abstract
Recent developments in the identification, estimation and diagnostic checking of deseasonalized autoregressive moving-average (ARMA) models and periodic autoregressive (PAR) models are reviewed. These techniques are then used for fitting PAR and deseasonalized ARMA models to three specific monthly riverflow time series. The Akaike information criterion (AIC) and Bayes information criterion (BIC) suggest the PAR model provides the best fit. Split-sample simulation experiments show that the PAR models preserve the critical drought statistics of the historical flow sequences.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have