Allocating transmissivities from constant head tests for the development of DFN models

Abstract

Constant head tests are commonly used for field measurements of fracture transmissivity. As a bulk transmissivity is measured for each test section, it is frequently unclear how this transmissivity relates to the hydraulic properties of individual fractures. The goal of this study is to determine if constant head tests conducted at scales larger than the average fracture spacing can be used to generate discrete fracture network (DFN) models that describe transport. The methodology involved generating DFNs using measurements from constant head tests conducted at lengths both above and below the average fracture spacing at a site in Ontario, Canada. Transport predictions from DFNs produced from different scales of hydraulic tests were compared to determine if a method for proportioning larger-scale test results to obtain a DFN representative of the smaller-scale tests could be established. The results of this study indicate that the choice of method used to apportion bulk transmissivities has a significant impact on transport simulations, with a difference in mass arrival of over a factor of two at travel distances less than 50 m. While the most appropriate method is case specific, the error resulting from the choice needs to be considered when using DFN models.