H21A-1340
Growing eddies reduce tailing in rough-walled rock fractures

Tuesday, 15 December 2015
Poster Hall (Moscone South)
In Wook Yeo1, Seung Hyun Lee2, Kang Kun Lee2 and Russell L Detwiler3, (1)Chonnam National University, Dept. of Earth and Environmental Sciences, Gwangju, South Korea, (2)Seoul National University, School of Earth and Environmental Sciences, Seoul, South Korea, (3)University of California Irvine, Irvine, CA, United States
Abstract:
We present the first direct observation of fluid flow and solute transport in a microscopic rough-walled fracture using micro-PIV to assess the evolution of non-Fickian tailing as eddies developed at larger fluid velocities. These observations demonstrated a previously-unidentified, important phenomenon: normalized BTCs became highly skewed towards later times up until a limiting fluid velocity, beyond which tailing decreased and peak concentrations increased. Further microscopic observation of particle trajectories clarified the likely cause of the reduced tailing at higher velocities. Tailing increased until the onset of eddies in large-aperture regions. As eddies became fully developed, particles were initially entrained in the eddies, but then cast back into the main flow channel, which reduced tailing. Numerical studies designed to more clearly understand the 3D nature of flow and transport showed that streamlines swirled near the fracture walls in a large-aperture region and re-entered the main flow channel. These numerical results support our interpretation of the experimental observation. This study, based on combined direct observations and numerical simulations, clearly demonstrated that there was no well-defined separation stream surface delineating so-called immobile (recirculation) and mobile zones in the 3D nature of the flow and transport and the tails decreased with growing eddies due to mass transfer by advective paths from eddies into the main flow channels. These experimental and numerical results contradict results from numerous previous studies based upon simulations in 2D fracture geometries and highlight the need for caution when using 2D simulations to understand 3D transport processes.