The current understanding of contaminant migration in fractured sedimentary rock aquifers is inadequate due to the difficulty in describing the geologic and hydrogeologic controls on contaminant fate and transport with appropriate detail. To address contamination at fractured rock sites, multiple methods focusing on different aspects of the hydrologic system are required, and particular emphasis needs to be placed on the rock matrix. This thesis shows the further development and utility of the decade-old rock core VOC method (i.e. CORETM), a rock matrix method, when used in conjunction with multiple high resolution datasets as it applies to a 100 m thick highly productive dolostone aquifer in Guelph, Ontario. The research site and surrounding area, located in the northwestern quadrant of the municipality of Guelph, was a productive zone for water supply until the early 1990s when the two closest municipal supply wells (Sacco, Smallfield) were shut down (1991, 1993 respectively) due to volatile organic compounds (VOCs) in the groundwater. Trichloroethene (TCE), a VOC, was used as a degreaser at the Guelph site and likely entered the groundwater more than 20 years ago. The thin overburden, shallow water table, relatively constant dolostone mineralogy, proximity to the UW analytical laboratory, relatively simple plume composition showing minimal degradation, and local importance make this an excellent study site for TCE fate and migration in fractured sedimentary rocks. This thesis is composed of four chapters. Chapter 1 provides a brief background to the rock core VOC method and gives the conceptual framework for the investigation. Chapter 2 focuses on the further development of the rock core VOC method by providing the field validation of a recently adapted extraction method for VOCs in rock core using microwave assisted extraction (MAE), demonstrating the importance of rapid field preservation of samples, and comparing to the industry standard purge and trap method for VOCs on solid matrices. Results indicate that the microwave assisted extraction (MAE) method typically provides equivalent or higher concentrations when compared with the shake-flask and purge and trap extraction methods, indicating more complete extraction or less loss during transfer and/or storage. The purge and trap method provided false negatives (i.e. non-detects) due to inadequate preservation, incomplete extraction, and the elevated detection limit for TCE. The necessity for field preservation was examined by comparing crushed rock samples preserved in methanol in the field to samples unpreserved in the field with a laboratory addition of methanol less than 12 hours later. Chapter 3 creates high resolution porosity and bulk density logs by using selected geophysical logging tools in combination with core derived physical properties for the purpose of calculating porewater concentrations from total contaminant mass concentrations obtained from the rock core VOC method and sample specific rock properties relevant to the conversion. This is beneficial because total mass estimates obtained from the rock core VOC method are not necessarily indicative of the groundwater concentrations given the presence of solid organic carbon controlled sorption. Chapter 4 is a demonstration of the discrete fracture network approach (Parker 2007) applied to the Guelph field site with emphasis on the insights gained through high resolution contaminant profiles generated from cored holes in or near the source area and along a transect. Together, these four chapters present a framework for investigating VOC contamination in fractured sedimentary rocks and with emphasis on evaluating recent advances in the rock core VOC methodology in a field site context.

A study was carried out in a sedimentary fractured rock site located in south central Wisconsin, US, which was impacted by DNAPL releases estimated to occur in the 1950’s and 1960’s. The majority of the DNAPL has accumulated in the upper portion of the Lone Rock Formation at a depth between 140 and 180 ft bgs referred as Layer 5 in this study. A groundwater VOC plume of more than 3km long has formed in this Layer. The DNAPL is mainly composed of 1,1,1-TCA, PCE, TCE and BTEX, while large amounts of biodegradation products such as cis-DCE and 1,1-DCA are present in the plume. Long term VOC data have been collected at the site and diverse geological and hydrogeological techniques have been applied to have a better understanding of the DNAPL history and behavior of the VOC plume. Evidence of biodegradation was also documented near the DNAPL source in these studies. The thesis objectives of the present study aimed first to have a better understanding of the long term contaminant distribution and degradation history at the site. This objective was accomplished reviewing the VOC historical concentration data collected from 1992 to 2006 in the wells tapping the most contaminated. hydrogeological unit in the bedrock (Layer 5) and in the overburden aquifer (referred as Layer 2). The second objective aimed to evaluate the current degree or extent of biodegradation of chlorinated compounds, which was accomplished evaluating the current groundwater redox conditions and using a combined analysis of VOC concentration and carbon isotope data collected in groundwater in September 2007. The historical data collected between1992 to 2006 showed the degradation of the VOC plume in Layer 5 was controlled by the availability of electron acceptors and redox conditions in the fracture bedrock aquifer. This pattern and the extension of the VOC plume were linked to different DNALP pumping events in the source zone and the operation of a Hydraulic Barrier System. The current geochemical and isotope study showed a different pattern of biodegradation of chlorinated compounds in different parts of the plume. The cis-DCE tend to accumulate in the area from the source to the middle of the plume and around 80 % of biodegradation of 1,1,1-TCA to 1,1-DCA was observed in this area. The fringes of the plume were characterized by a dominant presence of TCE and 1,1,1 TCA. These patterns were linked to different redox conditions and amount of electron acceptors. The cis-DCE dominated area is characterized by anaerobic conditions and the presence of relative high amount of BTEX. The TCE-dominated area is under aerobic condition and no BTEX was found in this area. The operation of the Hydraulic Barrier System seems to have change redox condition which influenced the extent of degradation in the plume, especially in the area between the extraction wells. The formation of large amounts of VC in Layer 2 and the more reducing (at least sulfate reducing and maybe methanogenic conditions) of the groundwater in this Layer compared to Layer 5 confirmed the extent of VOC biodegradation is linked to the availability of electron donors. This study provides information about the current degree of the biodegradation of chlorinated compounds at a fracture rock site. This information is very valuable for the evaluation of natural attenuation as strategy for long term plume management or for future remediation strategies such as biostimulation or bioaugmentation at the site. This study also shows the present and long term behavior of the chlorinated compounds (degradation history) in the most contaminated hydrogeologic unit (Layer 5), has mainly been controlled by plume management strategies including DNAPL pumping in the source and the creation of a Hydraulic Barrier System. The ketones and BTEX, that acted as electron donors and carbon substrate for the microbial community responsible for the dechlorination of chlorinated compounds were shown to have controlled the past and current redox conditions and thus the degree and potential of biodegradation of chlorinated ethenes and chlorinated ethanes at the study site.

Much is known about natural attenuation of contaminants in granular aquifers because many contaminant plumes in these aquifers have been intensively monitored with detailed sampling along cross sections positioned across the plumes (i. e. transects). However, little is known about natural attenuation of contaminant plumes in fractured rock. In this thesis study, strong natural attenuation of a persistent co-mingled plume of trichloroethylene (TCE) and an herbicide (metolachlor) in a 100 <em>m</em> thick dolostone aquifer used for municipal water supply in Cambridge, Ontario is shown based on detailed delineation of groundwater contaminant concentrations along a single transect located 150 <em>m</em> downgradient from the area where the metolachlor entered the dolostone and 300 <em>m</em> downgradient from the TCE source area. This delineation was accomplished using depth-discrete, multilevel groundwater monitoring systems in five cored holes and detailed analyses of contaminant concentration in rock cores. The maximum metolachlor concentration on the transect is a factor of 20 below the maximum concentration in the metolachlor source area and the maximum TCE concentration on the transect is lower by a factor of 100 from the TCE source area. <br /><br /> Matrix diffusion and strong temporal variability of the groundwater flow system caused by pumping of nearby municipal wells have likely caused strong natural attenuation of metolachlor and TCE and degradation has likely contributed to even stronger TCE attenuation. The transect shows rock core concentrations much higher than the groundwater concentrations in the multilevel systems at the same locations and in the conventional monitoring wells, which indicates that plume persistence is likely maintained by back diffusion from the rock matrix, which has very low hydraulic conductivity but substantial porosity, into the active groundwater flow in the fractures. <br /><br /> Metolachlor has been observed at very low concentrations and has persisted at these concentrations in the nearest municipal pumping well located approximately 780 <em>m</em> downgradient of the transect, however this well shows no detectable TCE. The relatively low concentrations along the transect and the replenishment of the plume by back diffusion suggests that a substantial increase of metolachlor or TCE in the municipal well is unlikely.

The current understanding of contaminant migration in fractured sedimentary rock aquifers is inadequate due to the difficulty in describing the geologic and hydrogeologic controls on contaminant fate and transport with appropriate detail. To address contamination at fractured rock sites, multiple methods focusing on different aspects of the hydrologic system are required, and particular emphasis needs to be placed on the rock matrix. This thesis shows the further development and utility of the decade-old rock core VOC method (i.e. CORETM), a rock matrix method, when used in conjunction with multiple high resolution datasets as it applies to a 100 m thick highly productive dolostone aquifer in Guelph, Ontario. The research site and surrounding area, located in the northwestern quadrant of the municipality of Guelph, was a productive zone for water supply until the early 1990s when the two closest municipal supply wells (Sacco, Smallfield) were shut down (1991, 1993 respectively) due to volatile organic compounds (VOCs) in the groundwater. Trichloroethene (TCE), a VOC, was used as a degreaser at the Guelph site and likely entered the groundwater more than 20 years ago. The thin overburden, shallow water table, relatively constant dolostone mineralogy, proximity to the UW analytical laboratory, relatively simple plume composition showing minimal degradation, and local importance make this an excellent study site for TCE fate and migration in fractured sedimentary rocks. This thesis is composed of four chapters. Chapter 1 provides a brief background to the rock core VOC method and gives the conceptual framework for the investigation. Chapter 2 focuses on the further development of the rock core VOC method by providing the field validation of a recently adapted extraction method for VOCs in rock core using microwave assisted extraction (MAE), demonstrating the importance of rapid field preservation of samples, and comparing to the industry standard purge and trap method for VOCs on solid matrices. Results indicate that the microwave assisted extraction (MAE) method typically provides equivalent or higher concentrations when compared with the shake-flask and purge and trap extraction methods, indicating more complete extraction or less loss during transfer and/or storage. The purge and trap method provided false negatives (i.e. non-detects) due to inadequate preservation, incomplete extraction, and the elevated detection limit for TCE. The necessity for field preservation was examined by comparing crushed rock samples preserved in methanol in the field to samples unpreserved in the field with a laboratory addition of methanol less than 12 hours later. Chapter 3 creates high resolution porosity and bulk density logs by using selected geophysical logging tools in combination with core derived physical properties for the purpose of calculating porewater concentrations from total contaminant mass concentrations obtained from the rock core VOC method and sample specific rock properties relevant to the conversion. This is beneficial because total mass estimates obtained from the rock core VOC method are not necessarily indicative of the groundwater concentrations given the presence of solid organic carbon controlled sorption. Chapter 4 is a demonstration of the discrete fracture network approach (Parker 2007) applied to the Guelph field site with emphasis on the insights gained through high resolution contaminant profiles generated from cored holes in or near the source area and along a transect. Together, these four chapters present a framework for investigating VOC contamination in fractured sedimentary rocks and with emphasis on evaluating recent advances in the rock core VOC methodology in a field site context.

A study was carried out in a sedimentary fractured rock site located in south central Wisconsin, US, which was impacted by DNAPL releases estimated to occur in the 1950’s and 1960’s. The majority of the DNAPL has accumulated in the upper portion of the Lone Rock Formation at a depth between 140 and 180 ft bgs referred as Layer 5 in this study. A groundwater VOC plume of more than 3km long has formed in this Layer. The DNAPL is mainly composed of 1,1,1-TCA, PCE, TCE and BTEX, while large amounts of biodegradation products such as cis-DCE and 1,1-DCA are present in the plume. Long term VOC data have been collected at the site and diverse geological and hydrogeological techniques have been applied to have a better understanding of the DNAPL history and behavior of the VOC plume. Evidence of biodegradation was also documented near the DNAPL source in these studies. The thesis objectives of the present study aimed first to have a better understanding of the long term contaminant distribution and degradation history at the site. This objective was accomplished reviewing the VOC historical concentration data collected from 1992 to 2006 in the wells tapping the most contaminated. hydrogeological unit in the bedrock (Layer 5) and in the overburden aquifer (referred as Layer 2). The second objective aimed to evaluate the current degree or extent of biodegradation of chlorinated compounds, which was accomplished evaluating the current groundwater redox conditions and using a combined analysis of VOC concentration and carbon isotope data collected in groundwater in September 2007. The historical data collected between1992 to 2006 showed the degradation of the VOC plume in Layer 5 was controlled by the availability of electron acceptors and redox conditions in the fracture bedrock aquifer. This pattern and the extension of the VOC plume were linked to different DNALP pumping events in the source zone and the operation of a Hydraulic Barrier System. The current geochemical and isotope study showed a different pattern of biodegradation of chlorinated compounds in different parts of the plume. The cis-DCE tend to accumulate in the area from the source to the middle of the plume and around 80 % of biodegradation of 1,1,1-TCA to 1,1-DCA was observed in this area. The fringes of the plume were characterized by a dominant presence of TCE and 1,1,1 TCA. These patterns were linked to different redox conditions and amount of electron acceptors. The cis-DCE dominated area is characterized by anaerobic conditions and the presence of relative high amount of BTEX. The TCE-dominated area is under aerobic condition and no BTEX was found in this area. The operation of the Hydraulic Barrier System seems to have change redox condition which influenced the extent of degradation in the plume, especially in the area between the extraction wells. The formation of large amounts of VC in Layer 2 and the more reducing (at least sulfate reducing and maybe methanogenic conditions) of the groundwater in this Layer compared to Layer 5 confirmed the extent of VOC biodegradation is linked to the availability of electron donors. This study provides information about the current degree of the biodegradation of chlorinated compounds at a fracture rock site. This information is very valuable for the evaluation of natural attenuation as strategy for long term plume management or for future remediation strategies such as biostimulation or bioaugmentation at the site. This study also shows the present and long term behavior of the chlorinated compounds (degradation history) in the most contaminated hydrogeologic unit (Layer 5), has mainly been controlled by plume management strategies including DNAPL pumping in the source and the creation of a Hydraulic Barrier System. The ketones and BTEX, that acted as electron donors and carbon substrate for the microbial community responsible for the dechlorination of chlorinated compounds were shown to have controlled the past and current redox conditions and thus the degree and potential of biodegradation of chlorinated ethenes and chlorinated ethanes at the study site.

Isotopic methods have been developed over the last 10 years as a method for determining chemical interactions of chlorinated solvents. These methods are especially promising for. This study attempts to employ and develop compound specific isotopic analyses of TCE and cDCE, along with chemical data, to characterize the degradation of TCE in a fractured bedrock aquifers. The Smithville site is a contaminated field site with extremely high levels of TCE contamination that is currently undergoing monitored remediation. From December 2008 until April 2010 extended samples were collected from the site to provide additional data analyses including isotopic data. The redox conditions at the site are anoxic to reducing, with sulfate reduction and methanogenesis as dominant terminal electron accepting processes. Redox data indicates that well electrochemical conditions are highly variable within the site, including areas near the source zone that not very reducing. Documented changes in groundwater conditions to much more reducing environments indicate that oxidation of organic matter is occurring at the Smithville site in select wells. Chemical analyses of TCE, DCE, VC, ethene and ethane are employed determine whether reductive dechlorination was occurring at the site. Results of field testing indicate that many wells on site, especially in the proximity of the source zone, dechlorination products were found. The isotopic data had a high range in both carbon and chlorine isotopes. Chlorine isotopic data ranges from a δ37Cl(TCE) of 1.39 to 4.69, a δ37Cl(cDCE) of 3.57 to 13.86, a δ13C(TCE) of -28.9 to -20.7, and a δ13C(cDCE) of -26.5 to -11.82. The range in values indicate varying degrees of degradation throughout the site, with the same wells grouping together. Combined chemical, redox and isotopic data shows that degradation seems to be a removal process for TCE at the Smithville site. Concentrations of chemicals created as a result of TCE degradation verify degradation, especially in wells 15S9, R7 and 17S9. Historically production of DCE in significant amounts, above 1.0 ppb, was observed to only occur after 2003. In addition to this, DCE data shows that the percentage of DCE made up of cDCE is above 96%. This indicates that microbes most likely mediate the processes that formed DCE from TCE. The linear regression of the delta-delta plot for isotopic TCE data shows line that is likely a direct function of the carbon and chlorine isotopic fractionation imparted upon the original TCE released. The slope found is consistent with data collected from other studies though cannot be applied to determining the process directly given the range of variability in isotopic field data.

Fractures are the main flow channels in naturally fractured reservoirs. Therefore the fracture permeability is a critical parameter to production optimization and reservoir management. Fluid pressure reduction caused by production induces an increase in effective stress in naturally fractured reservoirs. The change of effective stress induces fracture deformation and changes fracture aperture and permeability, which in turn influences the production. Coupled interactions exist in the fractured reservoir: (i) fluid pressure change induces matrix deformation and stress change; (ii) matrix deformation induces fluid volume change and fluid pressure change; (iii) fracture deformation induces the change of pore pressure and stress in the whole field (the influence disappears at infinity); (iv) the change of pore pressure and stress at any point has an influence on the fracture and induces fracture deformation. To model accurately the influence of pressure reduction on the fracture permeability change in naturally fractured reservoirs, all of these coupled processes need to be considered. Therefore, in this dissertation a fully coupled approach is developed to model the influence of production on fracture aperture and permeability by combining a finite difference method to solve the fluid flow in fractures, a fully coupled displacement discontinuity method to build the global relation of fracture deformation, and the Barton-Bandis model of fracture deformation to build the local relation of fracture deformation. The fully coupled approach is applied to simulate the fracture permeability change in naturally fracture reservoir under isotropic in situ stress conditions and high anisotropic in situ stress conditions, respectively. Under isotropic stress conditions, the fracture aperture and permeability decrease with pressure reduction caused by production, and the magnitude of the decrease is dependent on the initial effective in situ stress. Under highly anisotropic stress, the fracture permeability can be enhanced by production because of shear dilation. The enhancement of fracture permeability will benefit to the production of oil and gas.

Horizontal wells have been used to increase reservoir recovery, especially in unconventional reservoirs, and hydraulic fracturing has been applied to further extend the contact with the reservoir to increase the efficiency of development. In the past, many models, analytical or numerical, were developed to describe the flow behavior in horizontal wells with fractures. Source solution is one of the analytical/semi-analytical approaches. To solve fractured well problems, source methods were advanced from point sources to volumetric source, and pressure change inside fractures was considered in the volumetric source method. This study aims at developing a method that can predict horizontal well performance and the model can also be applied to horizontal wells with multiple fractures in complex natural fracture networks. The method solves the problem by superposing a series of slab sources under transient or pseudosteady-state flow conditions. The principle of the method comprises the calculation of semi-analytical response of a rectilinear reservoir with closed outer boundaries. A statistically assigned fracture network is used in the study to represent natural fractures based on the spacing between fractures and fracture geometry. The multiple dominating hydraulic fractures are then added to the natural fracture system to build the physical model of the problem. Each of the hydraulic fractures is connected to the horizontal wellbore, and the natural fractures are connected to the hydraulic fractures through the network description. Each fracture, natural or hydraulically induced, is treated as a series of slab sources. The analytical solution of superposed slab sources provides the base of the approach, and the overall flow from each fracture and the effect between the fractures are modeled by applying superposition principle to all of the fractures. It is assumed that hydraulic fractures are the main fractures that connect with the wellbore and the natural fractures are branching fractures which only connect with the main fractures. The fluid inside of the branch fractures flows into the main fractures, and the fluid of the main fracture from both the reservoir and the branch fractures flows to the wellbore. Predicting well performance in a complex fracture network system is extremely challenged. The statistical nature of natural fracture networks changes the flow characteristic from that of a single linear fracture. Simply using the single fracture model for individual fracture, and then adding the flow from each fracture for the network could introduce significant error. This study provides a semi-analytical approach to estimate well performance in a complex fracture network system.

The application of carbon-dioxide injection for enhanced oil recovery and/or sequestration purposes has gained impetus in the last decade. It is known that well test analysis plays a crucial role on getting information about reservoir properties, boundary conditions, etc. Although there are some studies related to the well test analysis in the fractured reservoirs, most of them are not focused on the carbon dioxide injection into the reservoir. Naturally fractured reservoirs (NFR) represent an important percentage of the worldwide hydrocarbon reserves and current production. Reservoir simulation is a fundamental technique in characterizing this type of reservoirs. Fracture properties are often not clear due to difficulty to characterize the fracture systems. On the other hand, well test analysis is a well known and widely applied reservoir characterization technique. Well testing in NFR provides two significant characteristic parameters, storativity ratio (&omega / ) and interporosity flow coefficient (&lambda / ). The storativity ratio is related to fracture porosity. The interporosity flow coefficient can be linked to the shape factor which is a function of fracture spacing. In this study, the effects of fracture and fluid flow factors (geometry, orientation and flow properties) on pressure and pressure derivative behavior are studied by applying a reservoir simulation model. Model is utilized mainly for the observation of multiphase flow effects in CO2 flooded fractured reservoirs. Several runs are conducted for various ranges of the aforementioned properties in the CO2 flooded reservoir. Results of well test analysis are compared to the input data of simulation models on a parameter basis.

A naturally fractured reservoir (NFR) is a reservoir with a connected network of fractures created by natural processes such as diastrophism and volume shrinkage (Ordonez et al. 2001). There are two models to simulate this kind of reservoirs: the discrete fracture model and the dual porosity model. In the dual porosity model, the matrix blocks occupy the same physical space as the fracture network and are identical rectangular parallelepipeds with no direct communication between isotropic and homogeneous matrix blocks. However, each fracture and matrix property is defined separately in the discrete fracture model. Another feature of this thesis is tracer testing. In this process, a chemical or radioactive element is injected to the reservoirs, and then it can be traced using the devices, which are designed to detect the tracers. Tracer tests have several advantages such as determining residual oil saturation, identifying barriers or high permeability zones in reservoirs, and providing the information on flow patterns. Limited number of research studies has been done on performing tracer tests in naturally fractured reservoirs. Also because there is not enough information about the advantages and disadvantages of the discrete fracture and the dual porosity models, researchers and engineers lack the expertise to confidently select either the discrete fracture or the dual porosity models to simulate the different types of NFRs. In this thesis, we compared the oil and water productions, and tracer concentration curves in various reservoir conditions, using both the discrete fracture and the dual porosity models. We used the ECLIPSE, which is a commercial software package in the area of petroleum industry, to model a naturally fractured reservoir. We performed a simple waterflooding with two conservative tracers on the reservoirs. The results presented in each section include the graphs of the oil production rate, water production rate, and tracer concentration. In addition, we presented the oil saturation profiles of a cross-section, which includes the production and injection wells. The results illustrated that both the discrete fracture and the dual porosity models are in good agreement, except for a few special cases. Generally, the oil production using the dual porosity model is more than in the discrete fracture model. The major disadvantage of the dual porosity model is that the fluid distribution in the matrix blocks is changing homogenously during the waterflooding period. In other words, ECLIPSE shows a constant value of the oil and water saturations in each time step for the matrix blocks. However, the dual porosity model is 3 to 4 times faster than the discrete fracture model. In the discrete fracture model, the users have complete control in defining the reservoirs. For example, the fracture aperture, fracture spacing, and fracture porosities can be set by the user. The disadvantage of this model is that millions of grid blocks are needed to model a large reservoir with small fracture spacing. / text