Fractured rock is the host or foundation for innumerable engineered structures related to energy, water, waste, and transportation. Characterizing, modeling, and monitoring fractured rock sites is critical to the functioning of those infrastructures, as well as to optimizing resource recovery and contaminant management. Fractured rock is defined in this report as a mass of rock matrix broken up by fractures. The rock may be crystalline with nominal porosity (i.e., many igneous rocks) or granular with varying amounts of cementation or porosity (i.e., sedimentary rocks).
At the request of the National Aeronautics and Space Administration, the U.S. Nuclear Regulatory Commission, and the U.S. Department of Energy, the National Academies of Sciences, Engineering, and Medicine (the National Academies) conducted a study to address issues relevant to subsurface fluid flow and transport in fractured rock systems. The National Academies convened an expert committee of researchers and practitioners to examine the state of the practice and state of the art in characterization of fractured rock and of the mechanical, chemical, and biological processes related to subsurface contaminant fate and transport. The committee also considered conceptual modeling of fractured rock and fluid and contaminant transport within it, the detection and monitoring of fluid and contaminant pathways and travel times, and remediation of contaminated sites in the event of system failures. The committee’s Statement of Task is in Box S.1. The committee interpreted its task to include consideration of all types of naturally fractured rock systems beneath the vadose zone to depths of 3 to 5 kilometers.
This report examines new developments, knowledge, and approaches to engineering at fractured rock sites since publication of the 1996 National Research Council report Rock Fractures and Fluid Flow: Contemporary Understanding and Applications. Fundamental understanding of the physical nature of fractured rock has changed little since 1996, but many new characterization tools have been developed, and there is now greater appreciation for the importance of chemical and biological processes that can occur in the fractured rock environment. Findings in this report can be applied to all types of engineered infrastructure and engineered in situ processes, but especially to engineered repositories for buried or stored waste and to fractured rock sites that have been contaminated as a result of past disposal or other practices. Impacts from artificially induced fractures to enhance hydrocarbon recovery (i.e., hydraulic fracturing) are not part of this report.
This report describes how existing tools—some only recently developed—can be used to increase the accuracy and reliability of engineering design and management given the interacting forces of nature. Examples of science and engineering research that can advance those tools or develop new tools are provided. Recommendations are organized by theme and include recommendations for practice and research in a given thematic area. The recommendations are high level and intended to help the practitioner, researcher, and decision maker take a more interdisciplinary approach to engineering in the fractured rock environment. An integrated systems approach for engineering fractured rock sites is emphasized, and recommendations are presented in a broader systems context. The recommendations are not intended for specific groups of experts or professional sectors. Using interdisciplinary approaches makes it possible to conceptualize and model the fractured rock environment with acceptable levels of uncertainty and reliability. They make it
possible to design systems that maximize remediation effectiveness and long-term performance. The fractured rock environment can be complex, but it is subject to known laws of nature.
CONTAMINANTS IN THE SUBSURFACE
Groundwater and contaminant transport and storage take place within void spaces (i.e., porosity) in both the rock matrix and fractures. However, although flow is dominated generally by only a few of the rock fractures in a fractured rock system, storage takes place predominantly within the rock matrix. This fundamental distinction between transport and storage has profound implications to fractured rock characterization, modeling, monitoring, and remediation.
Contaminant fate and transport in the fractured rock environment are affected significantly by contaminant solubility. Minimally water-soluble contaminants such as non-aqueous phase liquids can flow at rates and directions different from those of groundwater. Those that are denser than water (dense non-aqueous phase liquids such as trichloroethene) migrate downward in fractured rock until a barrier is encountered, making them difficult to locate and characterize. Water soluble contaminants can travel with water great distances from their sources, react with surrounding geologic and organic materials, or precipitate from solution.
UNDERSTANDING THE FRACTURED ROCK ENVIRONMENT
Adequate characterization and modeling of the fractured rock environment and the potential for contaminant fate and transport is critical when siting, designing, and managing infrastructure that could release contaminants. Modeling is also necessary to assess and design remediation schemes for areas where contaminants have already entered the subsurface. First steps in characterization include understanding the geologic setting and history so that the genesis of fractures can be understood and the most important hydrogeologic features can be defined. Expertise from several fields is best integrated at the earliest stages of project development and during data collection, analyses, and decision making to account for fractured rock processes that affect contaminant transport, storage, and transfer between the rock matrix and fractures. Such an integrated approach may yield better recognition of, for example, the ways discrete fractures may dominate flow and transport, or the reliability of measurements across large areas from long, open boreholes.
Recommendation 1. Take an interdisciplinary approach to engineering in fractured rock and use site geologic, geophysical, geomechanical, hydrologic, and biogeochemical information to conceptualize transport pathways, storage porosities, fate-and-transport mechanisms, and the coupled processes that control rock fracture-matrix interactions.
Use available geologic, hydrogeologic, and geophysical information to conceptualize flow pathways, fluid and contaminant storage, alteration, or attenuation through geochemical reactions and transformations via biological processes. When there are little data, for example, during early stages of characterization, fractured rock hydrogeology can be conceptualized by first developing a site conceptual model that assumes fluid storage occurs primarily in the rock matrix, and flow occurs primarily within fractures (i.e., a dual porosity/single permeability environment). Simplify the conceptual model (i.e., to single effective porosity and permeability) or add further detail (i.e., to dual porosity/dual permeability or discrete fracture networks) only when justified by site-specific evidence. As flow and transport are evaluated, recognize the limitations of advection-dispersion-diffusion approaches to solute transport solutions, particularly with regard to scale. Parameters developed at the meter scale, for example, may not be applicable at the kilometer scale.
Recommendation 2. Estimate the potential for contaminant to be transported into, stored in, and transported back out of rock matrix over time.
Fractures and fracture networks can be discrete flow pathways, barriers, or provide void space for storage. Understanding fracture geometries, possible heterogeneities, and interactions between the rock matrix and fractures (e.g., advection, diffusion, sorption, biodegradation, filtration, and capillary processes) is critical to understanding fate and transport. The storage of contaminants in the rock matrix and their slow diffusive transfer from the matrix to fluid flowing in adjacent fractures can result in remediation time frames of decades or centuries. Failure to quantify and differentiate between stored and mobile contaminants and site-specific storage and release mechanisms can lead to gross miscalculation of subsurface contaminant distribution and the effectiveness of remedial measures (e.g., pump and treat). The effects of rock fracture and matrix void sizes on capillary and wetting processes, bioactivity, abiotic reactions, and multiphase flow should also be evaluated.
Recommendation 3. Characterize chemical, biological, thermal, mechanical, and hydraulic processes, their interactions, and the conditions that can lead to their coupling to better understand transport through discrete fractures and contaminant transfer between fractures and rock matrix.
The importance of coupled processes in fate and transport is not always recognized or addressed. Individual and combined processes can change how fluids containing contaminants, migrating fines, and colloids interact with the rock matrix and fracture infillings and coatings. It is necessary to take into account the possibility of flow localization within fractures, channels within fracture planes and intersections, and hydro-mechanical coupling that results in changes in fracture and channel geometries. The effects of these changes on advection, dispersion, and diffusion in the interpretation and prediction of transport rates need to be understood to characterize fate and transport. There is also the potential for mineral fines to clog pathways or actually enable contaminant transport (i.e., Pickering emulsions). Fluid properties (e.g., miscibility and mutual solubility, density, viscosity, and acidity) and reactivity with rock or infilling material can also be important. Immiscible fluids exist as a separate phase from water and may impede or block groundwater flow paths, forcing groundwater to flow around them. When there is the possibility of thermal changes, assess subsurface heat conduction and temperature changes under static- and advective-flow conditions, including occurrences of flow localization. It is also important to consider how all of these factors—particularly stress, temperature, and groundwater chemistry—change with depth at a site.
Characterization of microbial communities and activities in fractured rock allows better characterization and prediction of fluid and contaminant fate and transport, as well as more effective application of bioremediation technologies. Subsurface microbial communities can affect physical and geochemical characteristics and may be responsible for a variety of dynamic processes including mineral formation and dissolution, as well as changes in redox chemistry, fluid surface tension, and acidity. There are few comprehensive applications to understand microbial community structure and function in fractured rock environments, and, given their importance in the geochemical properties and potential for bioremediation in fractured rock, more research in this area represents a singularly important investment.
Recommendation 4. Expand research to define and quantify microbial influences on fluid and contaminant fate and transport in fractured rock over timescales relevant to contaminant remediation processes.
Reliable, reproducible, quantitative, and statistically valid experimental information on the spatial and temporal dynamics of biological communities in fractured rock is needed to understand the processes that take place, their effects, and their rates of occurrence. Knowledge of biological processes in fractured rock has advanced in the past decade because of improvements in molecular tools (e.g., genome sequencing at the individual cell level is now possible), but the study of in situ biological systems, especially for rock at great depth, is still impractical, and cultivating microorganisms from fractured rock environments remains difficult. However, by coupling advanced molecular tools with computational modeling, it should now be possible to systematically address fundamental questions.
Several areas of research could be particularly useful, such as identification of parameters that determine the extents of phylogenetic, genetic, and functional diversity of microbial communities
in fractured rock environments; the influence of microbial populations on hydraulic and hydrogeochemical characteristics; the changes of phylogenetic and functional structures in microbial communities across various spatial and temporal scales (including associated relationships with hydrogeochemistry and the anoxic, high-salinity, high-temperature, and high-stress conditions of the deep environment); and the microbial community response to environmental perturbations in fractured rock environments.
CHARACTERIZATION TECHNIQUES AND TOOLS
Site characterization and monitoring tools and techniques for geologic, geophysical, geochemical, hydraulic, biological, and geomechanical observation, testing, and sampling have improved considerably since 1996. For example, remote sensing tools such as light detection and ranging permit large-scale fracture mapping at a high resolution, and borehole geophysical and imaging methods, pumping tests, and flow meter measurements help to identify zones and rates of flow. Advanced numerical algorithms now allow more information to be extracted from these and other test data, essentially creating opportunities for three-dimensional tomographic representations.
Most of these methods, however, are limited in their ability to allow the mapping of fracture characteristics at the needed scales and resolution. Borehole sampling and testing, for example, can be costly and provide limited information on a relatively small scale. Compressive stresses can complicate their use at the kilometer-scale depth. Cross-hole tracer or other hydraulic testing used to understand larger-scale issues such as fracture connectivity and flow dimension require multiple potentially costly boreholes. Remote methods such as seismic imaging provide only the means to infer rather than observe subsurface characteristics. Cross-hole radar and micro-seismic tomography are well suited for sites at depths, within their respective ranges.
Recommendation 5. Improve characterization and monitoring through new and expanded research in surface- and borehole-based geologic, geophysical, geochemical, hydraulic, biologic, and geomechanical technologies.
There is a need to advance research on technologies that use in situ measurements (e.g., innovative tracer tests and flexible liners) to characterize and monitor the volumes, spatial distribution, and transfer of contaminants between rock matrix and fractures. Fracture mapping techniques (e.g., thermal, electrical resistivity, and geochemical) could be improved to identify conductivity and flow in fractures, calculate in situ gradients, and determine flow directions. Geophysical techniques such as micro-seismic, electrical, nuclear magnetic resonance, and radar could be advanced to track tracers or contaminants remotely, including between injection and withdrawal locations. Research is needed on how to extend use of geophysical tools commonly used in porous media to fractured rock applications. Techniques such as seismic, microseismic, and hydraulic tomography, hydraulic interference, and tracer testing techniques need to be developed that allow for characterization of flow paths at different scales, and that advance joint inversion methodologies for fracture process parameterization. An important advance in characterization would be to determine ways that geophysical responses to fracture locations and geometries could be incorporated into fluid flow models.
Without a conceptual framework, numerical modeling will likely misrepresent hydrologic behavior at even the smallest project site. Conceptual model templates are available for different geologic environments and can be used to develop suitable conceptual models that incorporate the rock and hydrologic structures in a systematic way. Such conceptual models are called hydrostructural models. This kind of conceptual modeling promotes interdisciplinary approaches to decision making.
Because fractures can occur at the micron to kilometer scale, the full range of fluid fate and transport possibilities needs to be considered to determine which fractures and physical characteristics—and at what scales—are of hydrologic importance. However, current numerical modeling practice often relies on single-porosity, equivalent-continuum numerical models that are unable to account for discrete fracture flow pathways and fracture-matrix interactions more suitable to a homogenous porous media. In practice, budget, data, time, personnel skill sets, and regulatory expectations frequently drive the choice of modeling methods.
Recommendation 6. Develop appropriate hydrostructural conceptual models for fracture and rock matrix geometries and properties, and perform preliminary calculations (e.g., analytic or simple numerical) to better inform and allocate resources for site characterization, modeling, and remediation.
It is good practice to follow a systematic and well-documented approach to hydrostructural conceptual model development to understand the nature of underlying assumptions and simplifications, especially in the absence of supporting data. This can begin with generic geologic and geomechanical conceptual models and parameterizations that are informed but not limited by site-specific direct measurements. Next, develop a broad, semi-quantitative understanding of the important processes and parameters, and refine that understanding with simplified calculations that scope, define, and assess alternative models, their uncertainties, and the appropriateness of simplified versus more sophisticated modeling approaches. It is important to take advantage of such scoping calculations when making engineering decisions intended to reduce uncertainties in analysis—such as decisions related to site characterization activities and infrastructure or remediation design. Qualitative field data can be integrated into models through analysis tools such as parameter estimation, sensitivity analysis, and uncertainty quantification.
Commercial and publicly available numerical codes can facilitate fractured rock modeling for site-specific applications. To simplify calculations and make computations less resource intensive, these models often must be processed (upscaled) to average, narrow, or focus the ranges of specific hydrostructurally important properties and features.
Recommendation 7. Base numerical models on an appropriate hydrostructural model, ensuring that simplification and upscaling of the hydrostructural model maintain those features, properties, and processes that dominate contaminant transport, fate, and storage. Evaluate impacts of uncertainties introduced by simplification and upscaling.
Modelers need to consider carefully the appropriate applications and limitations of numerical models and analysis tools when selecting among them. The best models are those that synthesize data, confirm or refute the validity of the conceptual model, and quantify the range and uncertainties of expected system behaviors at the multiple scales appropriate to the problem of interest, using an approach consummate with the level of detail of the available data. It is important to apply equivalent continuum models only when the problem scale is much greater than the scale of the fractured rock mass. It is also important to analyze the need for risk- and uncertainty-based modeling approaches, even where calibrated and conditioned models match available measurements. Use non-deterministic modeling techniques (e.g., probabilistic and stochastic methods) when groundwater paths and response functions cannot otherwise be identified.
REMEDIATION AND MONITORING
The lack of a common framework, understanding, or expectations regarding remediation objectives, assessment, and realistic end points hinder effective engineering. Long-term remediation goals for most U.S. hazardous waste sites are based on drinking water Maximum Contaminant Levels (MCLs). Remediation can take decades or centuries to reach those MCLs. Regulatory agencies are beginning to accept long remediation time frames, and alternative approaches to remediation are becoming more acceptable.
Traditional pump-and-treat remediation methods draw water from fractures, remove contaminant, and return the water to the ground. Because contaminant is generally stored in the rock matrix of fractured rock, remediation focused solely on treating water in fractures is often futile. However, other remediation methods may be viable. Bioremediation, for example, relies on in situ biotransformation of contaminants by microorganisms. Thermal remediation promotes volatilization of contaminants, which are then removed via vapor extraction. However, better control of heating and extraction processes and more accurate prediction of results are needed to move this approach beyond the experimental stage.
Information regarding innovative underground remediation approaches is publicly available (e.g., through the U.S. Environmental Protection Agency). Less available is information specific to fractured rock site remediation. Lessons learned from remediation and monitoring in this setting are poorly documented, resulting in duplicated efforts and wasted resources in practice.
Recommendation 8. Develop and communicate realistic expectations related to remediation effectiveness through realistic goal setting and through explicit consideration of uncertainties in design, realistic use of natural attenuation, comprehensive monitoring programs, and dissemination of performance data to the technical community.
Best practices for remediation can be advanced through publicly accessible practitioner-driven and government-facilitated research-level documentation that details the remediation technologies applied across a variety of fractured rock settings. Expectations need to include the assumption that remediation of both rock fractures and matrix will likely be necessary. Estimates of plume longevity based on sound characterization need to be incorporated into remedial action plans that include natural attenuation as part of remediation. Regulatory frameworks that set realistic remedial objectives and formalize the transition from active remediation to long-term monitoring are needed, as are better information transfer mechanisms within and between agencies responsible for addressing fractured rock issues.
Recommendation 9. Incorporate long-term performance into monitoring system design.
Monitoring systems need to be designed for durability and to accommodate long-term performance and data needs. Accommodations include those for the expected operational and maintenance requirements for the duration of the infrastructure being monitored; for new technologies that might be developed during performance monitoring; and for potential variations in climate, water levels, temperatures, and other site conditions. An effective monitoring system includes meaningful sampling frequencies to monitor trends, and allows feedback that informs monitoring strategies in response to new trends or findings. Efficient monitoring system design is site specific and designed to require the minimal amounts of analytes to quantify performance effectiveness given localized discrete pathways, contaminant storage in the rock matrix, and geologic heterogeneity and anisotropy.
Monitoring systems need to be able to store, manage, and allow access to data in meaningful ways. Research on automated data collection, archiving, and retrieval and on the triggering of alarms when data values surpass tolerance levels is recommended. Research is also recommended on alternative approaches for remote monitoring of field-assessable parameters that trigger additional sampling and analysis. Best-practice protocols need to be established and communicated to future practitioners responsible for monitoring systems decades or centuries after monitoring systems are put in place. Advances in science and technologies need to be incorporated—as they develop—into engineering practice.
Engineering processes often involve the selection and implementation of a design early in the project development stage before an adequate conceptual site model is formed. Legal requirements and fixed-cost approaches favor linear project development in which the site is characterized, the engineering design is selected, and the project is implemented, all without the benefit of any new information gathered during successive steps. By using observational methods and adaptive approaches, new information could result in altered thinking about the site hydrostructural model and realistic engineered outcomes.
Recommendation 10. Use observational methods and adaptive approaches to inform engineering decisions made for fractured rock sites.
Adaptive and observational approaches to characterize and manage fractured rock sites are more effective than prescriptive, linear approaches. Risk-based and performance-based criteria are useful, where appropriate, in prioritizing the components of contaminated site management. Model assumptions concerning transport pathways can be tested and refined with characterization and monitoring data, as can chemical and biological processes that affect remediation. Systems analysis approaches can be used to integrate the value of information and to support decisions made under uncertainty.