Wednesday, December 5, 2007 : 10:40 a.m.
A Modflow Application to Simulate Ground Water Flow Through a Karst Site
A three-dimensional (3D) MODFLOW model was developed to simulate ground water flow through a site located in a valley/ridge physiographic province. The area is underlain by complexly folded and faulted bedrock. The site included two formations separated by a fault line. One formation includes predominantly shales, siltstones, and sandstones that have undergone weathering, forming a dual-porosity bedrock/matrix fracture system. The other formation is carbonate bedrock (limestone and dolostones) that underlie much of the site. In addition to the dual-porosity system, water seepage and weathering in these carbonate rocks has formed typical karst hydrogeologic conditions, including solution conduits in carbonate bedrock. A very large range of porosity and groundwater flow velocity is observed for this type of hydrogeologic system. Horizontally, the MODFLOW model included the two formations separated by the fault line. Vertically, the model included an unconsolidated layer on top a bedrock layer. Probabilistic distribution of hydraulic conductivity (K) for these formations and layers were derived using more than 150 K measurements. Realizations of the randomized K-fields in unconsolidated and bedrock zones were generated and included in the model. In addition, results from three tracer tests were available. These tests were used to estimate the porosities of the bedrock regions by seepage velocity calibration. The model was calibrated to observed heads and tracer velocities. Thereafter, a single alignment of high conductivity model cells in the upper bedrock formation was created to simulate the influence of a hypothetical karst conduit on flow and contaminant transport. Particle tracks were simulated using MODPATH and travel times within and outside the simulated conduit were determine. The tracks suggest that MODFLOW may be used to approximate flow condition in a karst formation based on knowledge of the heterogeneity obtained from probabilistic K-distribution in conjunction with seepage velocities and porosities obtained from tracer tests.
Jahangir Morshed, Ph.D., Science Applications International Corp. Dr. Morshed has over 10 years of research/consulting experience in water resources planning and management. His research focused on the application of advanced computational techniques for simulation/optimization of water resources system. His consulting experience includes surface/groundwater flow and contaminant transport modeling under various application settings. His experience further includes stochastic (probabilistic) modeling, and he has worked with Monte Carlo Sampling (MCS), Latin Hypercube Sampling (LHS), and other techniques. Over the years, he has acquired a strong computer background to develop numerical models and to support computational efforts.