Statistical error analysis of ground water systems
Date
1971-03
Authors
Bibby, Robert, author
Sunada, Daniel K., advisor
Waltz, James P., committee member
Breitenbach, E. A, committee member
Boes, Duane C., committee member
Journal Title
Journal ISSN
Volume Title
Abstract
A method is developed, which, by considering the input variables to a numerical model of flow in porous media as random variables, enables the accuracy of these input variables to be related to the accuracy of the output. The input variables considered are initial head, permeability, discharge, storage coefficient and saturated thickness and the output variable is head after a period of time. The method involves the use of the Monte Carlo technique to generate a random sample of the final head, the computation of a tolerance limit width and a coefficient of variation on the final head which are used as measures of its accuracy, and a regression analysis to determine a predictive relation between the accuracy of the input variables and the accuracy of the final head. The results indicate that if only one of the input variables contains error then this error is linearly related to the error in final head. If all input variables contain error, then only the error on initial head is significant in predicting the error in final head. In addition, a method of estimating the parameters of the probability density functions of the input variables from available field data is described and the relation is determined between the accuracy of these estimates and the number of data points used to make the estimate. The significance and application of the results in ground water system management is discussed.
Description
Rights Access
Subject
Groundwater -- Colorado