Analysis of pumping tests of partially penetrating wells in an unconfined aquifer using inverse numerical optimization

Inverse numerical modeling was applied to analyze pumping tests of partiall y penetrating wells carried out in three wells established in an unconfined aquifer in Vejen, Denmark, where extensive field investigations had previo usly been carried out, including tracer tests, mini-slug tests, and other h ydraulic tests. Drawdown data from multiple piezometers located at various horizontal and vertical distances from the pumping well were included in th e optimization. Horizontal and vertical hydraulic conductivities, specific storage, and specific yield were estimated, assuming that the aquifer was e ither a homogeneous system with vertical anisotropy or composed of two or t hree layers of different hydraulic properties. In two out of three cases, a more accurate interpretation was obtained for a multi-layer model defined on the basis of lithostratigraphic information obtained from geological des criptions of sediment samples, gamma-logs, and flow-meter tests. Analysis o f the pumping tests resulted in values for horizontal hydraulic conductivit ies that are in good accordance with those obtained from slug tests and min i-slug tests. Besides the horizontal hydraulic conductivity, it is possible to determine the vertical hydraulic conductivity, specific yield, and spec ific storage based on a pumping test of a partially penetrating well. The s tudy demonstrates that pumping tests of partially penetrating wells can be analyzed using inverse numerical models. The model used in the study was a finite-element flow model combined with a non-linear regression model. Such a model can accommodate more geological information and complex boundary c onditions, and the parameter-estimation procedure can be formalized to obta in optimum estimates of hydraulic parameters and their standard deviations.