<p>The presence of fractures in consolidated media allows for rapid transport of aqueous contaminants through convoluted pathways and for diffusion into the rock matrix adjacent to the fracture, which significantly complicates our ability to make transport predictions. Despite the need to predict transport in fractures over substantive distances, very few tracer experiments have been conducted at large scale (>50m) due to experimental difficulty and cost associated with such experiments. &#160;Even where these studies have been conducted, the results have often been difficult to model accurately without the use of extra fitting parameters. The objective of this study is to improve our understanding of key transport processes in complex large-scale fracture networks in carbonate rock by simulating the results of a tracer experiment conducted at a network scale. The tracer experiment used for this study was conducted previously by injecting a conservative dye tracer into an isolated 10 m section of a well and with breakthrough in six downstream observation wells open over a similar depth range. These observation points were located at distances of up to 245 m from the injection well. Measurement of the tracer breakthrough was conducted using a downhole fluorometer, allowing for observation of the full concentration profile in each well over time. To simulate the results, a DFN approach with a control-volume finite element model is used, which allows for irregular grid blocks and maintenance of the mass balance within the simulation domain. Because of the measurement of full concentration profiles, simulating transport inside the observation wells is also a focus of this study. In order to achieve a fit between the simulated and measured data, combinations of various fracture network geometries with aperture and matrix porosity heterogeneity are examined.</p>
