Applying Parameter-Estimation Methods to Recovery-Test and Slug-Test Analyses

Abstract

Parameter-estimation methods, including an exhaustive-search method and PEST (Parameter ESTimation) software, were applied to recovery-test data and slug-test data to obtain best estimates of transmissivity (T) by minimizing the sums of residuals. Each residual represents the difference between the field-measured water-level value and the value calculated by the appropriate non-linear equation. The exhaustive-search method in both cases involves computing the sums of residuals for an array of transmissivity and storativity values selected by the user for testing. Two new Fortran programs are presented that employ the exhaustive-search method. They utilize Picking's method for analyzing recovery-test data and the analytical equation for analyzing slug-test data derived by Cooper, Bredehoeft, and Papadopulos. Picking's method involves application of the Papadopulos and Cooper's equation for drawdown in finite-diameter wells. Utilizing field data reported in the literature, the estimated transmissivity values from the exhaustive-search methods were compared to the literature values obtained by type-curve matching techniques. The exhaustive-search values corresponded closely to the curve-matching values. Estimates for T were also obtained from recovery-test and slug-test data from two sites in southeastern Pennsylvania. For these sites, the PEST program was also applied to the data to evaluate the accuracy of the exhaustive-search methods. The results from the two methods were generally in good agreement. The two new Fortran programs are practical tools for the hydrogeologist, as they require less time compared to type-curve matching and the PEST method, and they yield accurate estimates of transmissivity.