Remediation of DNAPL contaminated sites in fractured rock has proven to be very difficult. No current technology can be used to remediate such sites in a timely and economic manner due to the inherent heterogeneity of fractured rock and back diffusion of contaminants stored in the rock matrix. This study was conducted in order to evaluate the viability of biostimulation of native biofilm as a means to control flow and back diffusion at fractured rock sites. A field trial was conducted at an uncontaminated site in southern Ontario. The site is underlain by dolomites of the Lockport formation. Three major fracture zones have been identified in the study area. Two closely spaced (5.04 m) boreholes were used to isolate a fracture zone at a depth of 17 m with straddle packers. These boreholes were used to create an injection-withdrawal system with recirculation, which was used for tracer injection in order to load the rock matrix with a conservative dye tracer and to inject nutrients for 21 days in order to stimulate the growth of biofilm in the fracture. Evaluation of the ability of the biofilm to control flow through the fracture was conducted through pulse interference tests. Pulse interference tests were conducted before and after the injection of nutrients. The results from the pulse interference tests showed a maximum 65% reduction in transmissivity, which is equivalent to a 28% reduction in fracture aperture shortly after the cessation of biostimulation. In order to investigate the effect of the biofilm stimulation on matrix diffusion the rock matrix was loaded with Lissamine, a conservative fluorescent dye tracer prior to biostimulation and its concentration was monitored at injection and withdrawal wells. The effect that biostimulation had on matrix diffusion was determined by comparing field concentration measurements with a model that simulates a system unaffected by biofilm stimulation. The biostimulation lowered the concentration of tracer attributable to back diffusion at the withdrawal well by about 20% for approximately 30 days following the cessation of biostimulation. It is also thought that large amounts of tracer might have been trapped in the biofilm as it formed and was then released back into the fracture as the biofilm deteriorated. / Thesis (Master, Civil Engineering) -- Queen's University, 2009-08-11 19:27:44.232

An analytical model is presented for the analysis of pulse interference tests conducted in a fractured porous medium with connection to a free surface boundary at the water table. The solution is applicable to open borehole pulse interference tests due to the accommodation of multiple horizontal fractures intersecting each wellbore and a connection from the uppermost horizontal fracture to a free surface boundary. The solution is derived using the Laplace transform method and evaluation of the solution is performed by numerical inversion using the Talbot algorithm. Based on an informal sensitivity analysis, unique values for transmissivity, storativity, specific storage, vertical hydraulic conductivity and specific yield are predicted over a range of realistic values for these parameters. The analytical model is used to analyze slug tests and pulse interference tests conducted in a fractured gneiss formation. The results of these tests are compared to long-term pumping tests to identify the effect of measurement scale on transmissivity, storativity, vertical hydraulic conductivity and specific yield obtained in a fractured rock setting. Scale artefacts relating to measurement or analysis methods are minimized through the use of consistent well configurations in each of the applied hydraulic testing methods. The geometric mean estimates of transmissivity and storativity are shown to vary by less than an order of magnitude from local-scale tests to long-term pumping tests. The geometric mean specific yield result from a series of pulse interference tests that samples both highly fractured and poorly fractured portions of the rock formation approximates the long-term pumping test estimate of specific yield. The geometric mean result for vertical hydraulic conductivity decreases by approximately 1.5 orders of magnitude from the slug test to pumping test scale; however, pulse interference tests conducted on highly fractured portions of the formation produce vertical hydraulic conductivity estimates that are within a half order of magnitude of the long-term pumping test results. This suggests that the performance of pulse interference tests on a highly fractured portion of a rock formation may be a less time-intensive alternative to large-scale pumping tests in the determination of vertical hydraulic conductivity. / Thesis (Master, Civil Engineering) -- Queen's University, 2011-06-24 19:29:52.743