Premium
Estimating correlations in multivariate streamflow models
Author(s) -
Stedinger Jery R.
Publication year - 1981
Publication title -
water resources research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.863
H-Index - 217
eISSN - 1944-7973
pISSN - 0043-1397
DOI - 10.1029/wr017i001p00200
Subject(s) - autocorrelation , streamflow , multivariate statistics , mathematics , estimator , series (stratigraphy) , statistics , logarithm , monte carlo method , multivariate normal distribution , geology , geography , drainage basin , cartography , paleontology , mathematical analysis
Multivariate log normal distributions are often used to model and generate multisite and multiseason streamflow sequences. A Monte Carlo study evaluated alternative estimators of the cross correlation between autocorrelated streamflow series and the lag 1 autocorrelation of a single series when some or all of the flows have a log normal distribution. Generally, smaller mean square error estimates of the correlations of flows are obtained by using the variances and covariances of the flow's logarithms to estimate a streamflow model's parameters. Sometimes it is advantageous to prewhiten two series before calculating their cross correlation and to unbias the estimated autocorrelation of the streamflow's logarithms.