ASSESSMENT OF SMEARED CONDUIT APPROACH FOR MODELING KARST CONDUIT FLOWS

Karst aquifers have been modeled in the past by various distributed parameter models, which include discrete-conduit models, single-continuum, smeared-conduit models and dual-continuum models. In the smeared-conduit approach (SCA), no attempt is made to delineate the actual geometry of the relatively small conduits. Instead, the approach tries to capture the effects of a conduit on much larger computational grid blocks by assigning effective properties to the grid blocks. Regional-scale numerical models constructed using MODFLOW (A. W. Harbaugh, E. R. Banta, M. C. Hill, and M. G. McDonald, USGS Open File Report 00-92, 2000), for example, often adopt the SCA. In the literature, few studies have been conducted to address the accuracy of the SCA, especially under the transient-flow conditions. A quantitative study is thus warranted to assess the performance of SCA for modeling karst conduit flows. For comparison purposes, a finite-element model using embedded line elements to represent discrete conduits was used to establish a synthetic ‘truth’ case. The capability of the smeared-conduit model to reproduce the results of the more complex finite-element model was then evaluated. We show the performance of SCA for a series of test problems with increasingly complex conduit geometry.

Acknowledgement The authors are grateful to the Awwa Research Foundation, Edwards Aquifer Authority, and Southwest Florida Water Management District for supporting this work.