Parameter Estimation for First-Order Autoregressive Model

by Luis E. Garcia-Martinez, Head; Div. Hidrografica, Instituto Geografico Nacional and Prof., Escuela Regional de Ingenieria Sanitaria, Guatemala City, Guatemala,

Serial Information: Journal of the Hydraulics Division, 1972, Vol. 98, Issue 8, Pg. 1343-1349

Document Type: Journal Paper


Stochastic simulation of streamflow depends upon estimation of the correlation structure of the series. Several estimators of the autocorrelation coefficient are investigated and all are shown to be biased for sample sizes beyond usually encountered in hydrology. That estimator with least bias is shown to have greatest variance and that with least variance to have greatest bias. The estimator chosen as best must be judged on the relative effects of bias and variance.

Subject Headings: Parameters (statistics) | Autoregressive models | Stochastic processes | Streamflow | Correlation | Hydrology

Services: Buy this book/Buy this article


Return to search