This study proposes to investigate anisotropic dispersion in fractured media and convective mixing associated with variable-density flow. It is generally not appreciated that the pattem of mixing in fractured systems can be anisotropic, as determined by the geometry of the fracture network, and can deviate from the classical isotropic theory. In terms of unstable mixed flows in fractured rocks, results are limited to a few studies. There are significant opportunities with our proposed study to contribute ftmdamental knowledge to physical mass transport in fractured rock systems. The goal is to elucidate processes of physical mass transport in fractured rock systems - how dispersion and convective dispersion are influenced by characteristics of the fracture geometry. To achieve this goal we propose to conduct both physical and numerical model studies of mass transport in 2-D and 3-D systems with complex fracture geometries. We are interested in exploring the conditions under which anisotropic dispersion develops and is manifest in variable-density systems. The study is organized around four different tasks. Task I involves a series of laboratory experiments with 2-D fractured systems. Task 2 extends the 3-D variable-density code MITSU3D to fracture networks with permeable matrix blocks. Task 3 is designed to study physical mass transport experimentally in a well-characterized 3-D fracture network. Task 4 is a complimentary model study designed to use the experimental results in conjunction with MITSU3D to develop appropriate methods to seed instabilities. The proposed study will contribute to knowledge by: (i) providing ftindainental new knowledge concerning the nature of dispersion and convective dispersion in fractured rock systems, (ii) developing and applying new capabilities for the physical modeling of mass transport in complex fractured rock systems, and for visualizing tracer distributions in three dimensions, and, (iii) developing and applying new modeling approaches for discrete-fracture systems, which can accommodate complexities in network geometry and in processes.
