Generation of Ungaged Streamflow Data

Publisher Website
Google Scholar
Abstract
A least-squares model for estimating monthly and daily mean streamflow data at ungaged sites has been developed. The predictor variables are streamflow data recorded at nearby hydrometric sites. The parameters required by the model are estimates of the mean and standard deviation of flow at the ungaged site and estimates of the correlation coefficients between flows at the ungaged site and nearby hydrometric sites. These parameters were obtained from least squares models using physiographic characteristics as predictor variables. Computational problems were experienced when the estimated correlation matrix was not positive semidefinite. These problems were overcome by adjusting the estimated correlation coefficients using Rosenbrock's hill climbing method to make the matrix positive semidefinite. The model was tested by estimating streamflow records in central Canada that were assumed to be nonexistent. The reliability of the estimated data depended primarily upon the accuracy with which the model parameters were determined.

This publication has 0 references indexed in Scilit: