Can we simulate regional groundwater flow in a karst system using equivalent porous media models? Case study, Barton Springs Edwards aquifer, USA

Bridget R. Scanlon
Robert E. Mace
Brian Smith

Please visit https://digitalcommons.usf.edu/kip_articles/695 to view this article.

Abstract

Various approaches can be used to simulate groundwater flow in karst systems, including equivalent porous media distributed parameter, lumped parameter, and dual porosity approaches, as well as discrete fracture or conduit approaches. The purpose of this study was to evaluate two different equivalent porous media approaches: lumped and distributed parameter, for simulating regional groundwater flow in a karst aquifer and to evaluate the adequacy of these approaches. The models were applied to the Barton Springs Edwards aquifer, Texas. Unique aspects of this study include availability of detailed information on recharge from stream-loss studies and on synoptic water levels, long-term continuous water level monitoring in wells throughout the aquifer, and spring discharge data to compare with simulation results. The MODFLOW code was used for the distributed parameter model. Estimation of hydraulic conductivity distribution was optimized by using a combination of trial and error and automated inverse methods. The lumped parameter model consists of five cells representing each of the watersheds contributing recharge to the aquifer. Transient simulations were conducted using both distributed and lumped parameter models for a 10-yr period (1989–1998). Both distributed and lumped parameter models fairly accurately simulated the temporal variability in spring discharge; therefore, if the objective of the model is to simulate spring discharge, either distributed or lumped parameter approaches can be used. The distributed parameter model generally reproduced the potentiometric surface at different times. The impact of the amount of pumping on a regional scale on spring discharge can be evaluated using a lumped parameter model; however, more detailed evaluation of the effect of pumping on groundwater levels and spring discharge requires a distributed parameter modeling approach. Sensitivity analyses indicated that spring discharge was much more sensitive to variations in recharge than pumpage, indicating that aquifer managem