Aspects of Aquifer Test Error Analysis

Abstract

Errors in the estimation of the aquifer parameters T and S derived from aquifer test data are examined as to their cause and effects. The analysis is based on the Theis equation. The basic causes of error are in the measurements of drawdown and pumping rate, in fitting the model to the data and in violations of model assumptions. Measurement errors were studied experimentally. Curve fittings by hydrologists were compared to "automatic" curve fittings obtained by nonlinear regression. The covariance matrix of T and S obtained in this manner was used, in conjunction with sensitivity analysis, to estimate the error in prediction of future drawdown. While automatic fitting is not a perfect substitute for graphical fitting, there is a definite relation between the two methods which allows the use of the statistics developed by nonlinear regression theory to be used to study the cause, effects and risks inherent in aquifer analysis.