H53I-05
Comparison of Hydraulic Methods and Tracer Experiments as Applied to the Development of Conceptual Models for Discrete Fracture Networks

Friday, 18 December 2015: 14:40
3018 (Moscone West)
Kent S Novakowski, Queen's University, Kingston, ON, Canada
Abstract:
The development of conceptual models for solute migration in discrete fracture networks has typically been based on a combination of core logs, borehole geophysics, and some form of single-well hydraulic test using discrete zones. More rarely, interwell hydraulic tests and interwell tracer experiments are utilised to directly explore potential transport pathways. The latter methods are less widely employed simply due to potentially significant increases in the cost and effort in site characterization. To date however there is a paucity of literature comparing the efficacy of the standard procedure with what should be more definitive identification of transport pathways using interwell methods. In the present study, a detailed comparison is conducted by developing conceptual models from three separate data sets, the first based on core logs, geology and single-well hydraulic tests, the second based on a large suite of pulse interference tests, and the third based on a series of radially-divergent and injection-withdrawal tracer experiments. The study was conducted in an array of five HQ-sized wells, 28-32 m in depth and arranged in a five star pattern, 10 m on a side. The wells penetrate the contact between a Cambrian-aged limestone, and underlying Precambrian gneiss. The core was logged for potentially open fractures using a ranking system, and 87 contiguous hydraulic tests were conducted using a 0.85-m packer spacing. A total of 57 pulse interference tests were conducted using two wells as injection points, and 11 tracer experiments were conducted using either sample collection or in-situ detection via a submersible fluorometer. The results showed very distinct conceptual models depending on the data set, with the model based on the single-well testing significantly over-predicting the number and connection of solute transport pathways. The results of the pulse interference tests also over predict the transport pathways, but to a lesser degree. Quantification of the pore volume using the various data sets was also undertaken and shows agreement with the general observations.