Uncertainty Analysis for the Estimated Hydraulic Conductivity from the Reduced Duration Pumping Test

Abstract

Use of a homogeneous aquifer method to represent a heterogeneous aquifer for parameter estimation will inevitably result in uncertainties in the estimated parameter values. The objective of this paper is to evaluate the impacts of the reduced duration of the pumping test on the estimated hydraulic conductivity using the observations from pumping tests. Firstly, a heterogeneous hydraulic conductivity K field was generated using the Monte Carlo method. A MODFLOW groundwater flow model was then constructed to perform numerical pumping tests. The K value was estimated by the inverse method based on the Theis solution for an unconfined aquifer. The K fields were generated based on the hydraulic conductivity data sets collected for a karst aquifer in southwest of China. The effective hydraulic conductivity (Ke) was calculated to represent the hydraulic conductivity of a heterogeneous aquifer. For a single test in a heterogeneous aquifer, while the duration of the test is less than a critical value, the reduced duration will lead to the big uncertainty of the estimated hydraulic conductivity. However, Ke can achieve a relative stable value while the number of reduced duration tests is big enough to eliminate the impacts of the uncertainty of the separate K values.