Bibliography - Michael Celia

1. Bandilla, K. W., Benjamin Court, Thomas R. Elliot, and Michael Celia, February 2012: Comparison of Brine Production Scenarios for Geologic Carbon Sequestration Operations. Carbon Management Technology Conference, doi:10.7122/151250-MS
   [ Abstract ]

   Large volumes of CO₂ will have to be stored in the subsurface for carbon capture and geological sequestration to have a significant impact on the reduction of carbon emissions. Injection of large volumes of CO₂ into deep saline formations can lead to significant pressure increases within that formation. The increased pressure can be a limiting factor for injection rates; it can also drive vertical brine migration through leakage pathways (e.g., abandoned wells) that could contaminate sources of drinking water. Production of brine from the injection formation can reduce the pressure increase while also limiting the spatial extent of the pressure increase. The impact of brine extraction is investigated using a hypothetical injection domain conditioned by parameters from the Illinois Basin. The domain contains one injection well and encompasses several aquifers connected through diffusive brine leakage. A vertically-integrated approach is used to model the injection formation and overlying aquifers. A set of production scenarios illustrates the impact of brine production on injection rates and vertical brine movement. The scenarios include production with surface disposal and production with reinjection into overlying formations (with and without desalinization). The results show that brine production can reduce the pressure buildup in the injection formation, leading to an increase in injectivity and a concomitant reduction in fresh water contamination risk by reducing the area of potential impact. While reinjection of brine into an overlying aquifer solves the disposal problem, it also reduces the effectiveness of brine production by increasing the pressure. Injection of a smaller amount of more concentrated brine resulting from desalinization reduces the impact of reinjection and acts as an additional source of fresh water, but increases the cost of the injection operation. Based on the results from these numerical experiments pressure management through brine production should be considered for industrial-scale CO₂ injection operations, as it increases injectivity and reduces the size of the area of potential impact. However, the brine disposal problem needs to be solved for brine production to be a useful endeavor.

2. Buscheck, Thomas A., Yunwei Sun, Mingjie Chen, Yue Hao, Thomas J. Wolery, William L. Bourcier, Benjamin Court, Michael Celia, S. Julio Friedmann, and Roger D. Aines, 2012: Active CO₂ reservoir management for carbon storage: Analysis of operational strategies to relieve pressure buildup and improve injectivity. International Journal of Greenhouse Gas Control, Elsevier, 6, doi:10.1016/j.ijggc.2011.11.007 230-245
   [ Abstract ]

   For industrial-scale CO₂ injection in saline formations, pressure buildup can limit storage capacity and security. Active CO₂ Reservoir Management (ACRM) combines brine production with CO₂ injection to relieve pressure buildup, increase injectivity, manipulate CO₂ migration, and constrain brine leakage. By limiting pressure buildup, in magnitude, spatial extent, and duration, ACRM can reduce CO₂ and brine leakage, minimize interactions with neighboring subsurface activities, allowing independent assessment and permitting, reduce the Area of Review and required duration of post-injection monitoring, and reduce cost and risk. ACRM provides benefits to reservoir management at the cost of extracting brine. The added cost must be offset by the added benefits to the storage operation and/or by creating new, valuable uses that can reduce the total added cost. Actual net cost is expected to be site specific, requiring detailed analysis that is beyond the scope of this paper, which focuses on the benefits to reservoir management. We investigate operational strategies for achieving an effective tradeoff between pressure relief/improvedinjectivity and delayed CO₂ breakthrough at brine producers. For vertical wells, an injection-only strategy is compared to a pressure-management strategy with brine production from a double-ring 9-spot pattern. Brine production allows injection to be steadily ramped up while staying within the pressure-buildup target, while injection-only requires a gradual ramp-down. Injector/producer horizontal/well pairs were analyzed for a range of well spacings, storage-formation thickness and area, level and dipping formations, and for homogeneous and heterogeneous permeability. When the producer is downdip of the injector, the combined influence of buoyancy and heterogeneity can delay CO₂ breakthrough. Both vertical and horizontal wells can achieve pressure relief and improved CO₂ injectivity, while delaying CO₂ breakthrough. Pressure buildup and CO₂ breakthrough are sensitive to storage-formation permeability and insensitive to all other hydrologic parameters except caprock-seal permeability, which only affects pressure buildup for injection-only cases.

3. Celia, Michael, and Jan M. Nordbotten, 2012: Geological Storage of CO₂: Modeling Approaches for Large-Scale Simulation , John Wiley & Sons, Inc., 264pp.
4. Court, Benjamin, K. W. Bandilla, Michael Celia, Thomas A. Buscheck, Jan M. Nordbotten, M. Dobossy, and Adam Janzen, 2012: Initial evaluation of advantageous synergies associated with simultaneous brine production and CO₂ geological sequestration. International Journal of Greenhouse Gas Control, Elsevier, 8, doi:10.1016/j.ijggc.2011.12.009 90-100
   [ Abstract ]

   Mitigation of global atmospheric carbon emissions requires a worldwide ramping up of CO₂ capture and sequestration (CCS) implementation in the next decades. While CCS could be deployed in isolation, there is also the possibility to consider CO₂ injection within a much broader framework of reservoir and resource management including active water (brine) management. The goal of this study is to provide an initial analysis of three identified synergies related to active brine management in CCS operations. The potential advantages of coupling simultaneous brine production to a large-scale CO₂ geological sequestration operation are explored through three separate modeling studies. Our results demonstrate that brine production can provide important pressure-control benefits, including increased injectivity potential through reduction of the injection well pressure, significant reduction of the extent of the Area of Review, within which operators must procure property rights and monitor and remediate potential leakage pathways, and reduction in the risk of CO₂ and brine leakage. The latter is especially important in reservoirs, like many in North America, where a significant number of potential leakage pathways, particularly abandoned wells, may exist within the Area of Review. We also observe that brine production has minimal impact on the overall shape of the CO₂ plume, with plume shape and extent strongly governed by formation parameters.

5. Elliot, Thomas R., and Michael Celia, 2012: Potential restrictions for CO₂ sequestration sites due to shale and tight gas production. Environmental Science and Technology, Washington, DC, American Chemical Society, (February 21, 2012), doi:10.1021/es2040015 1-16
   [ Abstract ]

   Carbon Capture and Geological Sequestration is the only available technology that both allows continued use of fossil fuels in the power sector and reduces significantly the associated CO₂ emissions. Geological sequestration requires a deep permeable geological formation into which captured CO₂ can be injected, and an overlying impermeable formation, called a caprock, that keeps the buoyant CO₂ within the injection formation. Shale formations typically have very low permeability and are considered to be good caprock formations. Production of natural gas from shale and other tight formations involves fracturing the shale with the explicit objective to greatly increase the permeability of the shale. As such, shale gas production is in direct conflict with the use of shale formations as a caprock barrier to CO₂ migration. We have examined the locations in the United States where deep saline aquifers, suitable for CO₂ sequestration, exist, as well as the locations of gas production from shale and other tight formations. While estimated sequestration capacity for CO₂ sequestration in deep saline aquifers is large, up to 80% of that capacity has areal overlap with potential shale-gas production regions and, therefore, could be adversely affected by shale and tight gas production. Analysis of stationary sources of CO₂ shows a similar effect: about two-thirds of the total emissions from these sources are located within 20 miles of a deep saline aquifer, but shale and tight gas production could affect up to 85% of these sources. These analyses indicate that co-location of deep saline aquifers with shale and tight gas production could significantly affect the sequestration capacity for CCS operations. This suggests that a more comprehensive management strategy for subsurface resource utilization should be developed.

6. Gasda, S., Jan M. Nordbotten, and Michael Celia, 2012: Application of simplified models to CO₂ migration and immobilization in large-scale geological systems. International Journal of Greenhouse Gas Control, Elsevier, 9, doi:10.1016/j.ijggc.2012.03.001 72-84
   [ Abstract ]

   Long-term stabilization of injected carbon dioxide (CO₂) is an essential component of risk management for geological carbon sequestration operations. However, migration and trapping phenomena are inherently complex, involving processes that act over multiple spatial and temporal scales. One example involves centimeter-scale density instabilities in the dissolved CO₂ region leading to large-scale convective mixing that can be a significant driver for CO₂ dissolution. Another example is the potentially important effect of capillary forces, in addition to buoyancy and viscous forces, on the evolution of mobile CO₂. Local capillary effects lead to a capillary transition zone, or capillary fringe, where both fluids are present in the mobile state. This small-scale effect may have a significant impact on large-scale plume migration as well as long-term residual and dissolution trapping. Computational models that can capture both large and small-scale effects are essential to predict the role of these processes on the long-term storage security of CO₂ sequestration operations. Conventional modeling tools are unable to resolve sufficiently all of these relevant processes when modeling CO₂ migration in large-scale geological systems. Herein, we present a vertically-integrated approach to CO₂ modeling that employs upscaled representations of these subgrid processes. We apply the model to the Johansen formation, a prospective site for sequestration of Norwegian CO₂ emissions, and explore the sensitivity of CO₂ migration and trapping to subscale physics. Model results show the relative importance of different physical processes in large-scale simulations. The ability of models such as this to capture the relevant physical processes at large spatial and temporal scales is important for prediction and analysis of CO₂ storage sites.

7. Nogues, J. P., Benjamin Court, M. Dobossy, Jan M. Nordbotten, and Michael Celia, 2012: A methodology to estimate maximum probable leakage along old wells in a geological sequestration operation. International Journal of Greenhouse Gas Control, Elsevier, 7, doi:10.1016/j.ijggc.2011.12.003 39-47
   [ Abstract ]

   This study presents a computational methodology to estimate the maximum probable leakage of CO₂ along old wells in a geological sequestration operation. The methodology quantifies the maximum probable CO₂ leakage as a function of the statistical characterization of existing wells. We use a Monte Carlo approach based on a computationally efficient simulator to run many thousands of realizations. Results from the Monte Carlo simulations are used to determine maximum leakage rates at 95% confidence. Uncertainty in the analysis is due to leaky well parameters, which are known to be highly uncertain. We consider a wide range of parameter values, with our focus on assignment of effective well permeability values and the correlation of those values along individual wells. We use a specific location in Alberta, Canada, to demonstrate the methodology using a hypothetical injection and an assumed probability structure for the well permeabilities. We show that for a wide range of parameter values, the amount of leakage is within the bounds suggested as acceptable for climate change mitigation.

8. Celia, Michael, and Jan M. Nordbotten, 2011: How Simple Can We Make Models for CO₂ Injection, Migration, and Leakage? Energy Procedia, Elsevier, 4, doi:10.1016/j.egypro.2011.02.322 3857-3864
   [ Abstract ]

   Analysis of geological storage of CO₂ almost always involves some set of computational models that provide a mathematical description of the problem. These models can have many purposes, but ultimately they should be able to answer practical questions about the system. These questions usually involve the spatial extent of the CO₂ plume, the spatial extent of pressure perturbations, the spatial and temporal dynamics of leakage out of the injection formation, and the spatial temporal evolution of different trapping mechanisms. Answers to these questions require models that apply to large spatial and temporal scales while including certain small-scale features like leakage pathways. Development of computationally efficient models that can span the appropriate scales may be achieved by analyzing the length and time scales associated with the important processes in the system, and incorporating those scales into a systematic model development. Such a procedure can be described as multi-part on scaling arguments for the physical processes involved, to produce a sequence of successively simpler models. Through this approach, the assumptions in all of the simplified models are made transparent, and the length and time scales appropriate for different models can be identified. In addition, by associating length and time scales to the questions being asked, models can be developed that are consistent with those scales and therefore are appropriate to answer the questions.

9. Celia, Michael, Jan M. Nordbotten, Benjamin Court, M. Dobossy, and S. Bachu, 2011: Field-scale application of a semi-analytical model for estimation of CO₂ and brine leakage along old wells. International Journal of Greenhouse Gas Control, Elsevier, (5), doi:10.1016/j.ijggc.2010.10.005
   [ Abstract ]

   Carbon capture and geological storage (CCS) operations will require an environmental risk analysis to determine, among other things, the risk that injected CO₂ or displaced brine will leak from the injection formation into other parts of the subsurface or surface environments. Such an analysis requires site characterization including identification of potential leakage pathways. In North America, the century-long legacy of oil and gas exploration and production has left millions of oil and gas wells, many of which are co-located with otherwise good geological storage sites. Potential leakage along existing wells, coupled with layered stratigraphic sequences and highly uncertain parameters, makes quantitative analysis of leakage risk a significant computational challenge. However, new approaches to modeling CO₂ injection, migration, and leakage allow for realistic scenarios to be simulated within a probabilistic framework. Using a specific field site in Alberta, Canada, we perform a range of computational studies aimed at risk analysis with a focus on CO₂ and brine leakage along old wells. The specific calculations focus on the injection period, when risk of leakage is expected to be largest. Specifically, we simulate 50 years of injection of supercritical CO₂ and use a Monte Carlo framework to analyze the overall system behavior. The simulations involve injection, migration, and leakage over the 50-year time horizon for domains of several thousand square kilometers having multiple layers in the sedimentary succession and several thousand old wells within the domain. Because we can perform each simulation in a few minutes of computer time, we can run tens of thousands of simulations and analyze the outputs in a probabilistic framework. We use these kinds of simulations to demonstrate the importance of residual brine saturations, the range of current options to quantify leaky well properties, and the impact of depth of injection and how it relates to leakage risk.

10. Court, Benjamin, Thomas R. Elliot, J.A. Dammel, Thomas A. Buscheck, J. Rohmer, and Michael Celia, 2011: Promising synergies to address water, sequestration, legal, and public acceptance issues associated with large-scale implementation of CO₂ sequestration. The Journal for Mitigation and Adaption Strategies for Global Change, Springer, doi:10.1007/s11027-011-9314-x
    [ Abstract ]

    Stabilization of CO₂ atmospheric concentrations requires practical strategies to address the challenges posed by the continued use of coal for baseload-electricity production. Over the next two decades, CO₂ capture and sequestration (CCS) demonstration projects would need to increase several orders of magnitude across the globe in both size and scale. This task has several potential barriers which will have to be accounted for. These barriers include those that have been known for a number of years including safety of subsurface sequestration, pore-space competition with emerging activities like shale gas production, legal and regulatory frameworks, and public acceptance and technical communication. In addition water management is a new challenge that should be actively and carefully considered across all CCS operations. A review of the new insights gained on these previously and newly identified challenges, since the IPCC special report on CCS, is presented in this paper. While somewhat daunting in scope, some of these challenges can be addressed more easily by recognizing the potential advantageous synergies that can be exploited when these challenges are dealt with in combination. For example, active management of water resources, including brine in deep subsurface formations, can provide the additional cooling-water required by the CO₂ capture retrofitting process while simultaneously reducing sequestration leakage risk and furthering efforts toward public acceptance. This comprehensive assessment indicates that water, sequestration, legal, and public acceptance challenges ought to be researched individually, but must also be examined collectively to exploit the promising synergies identified herein. Exploitation of these synergies provides the best possibilities for successful large-scale implementation of CCS.

11. Dobossy, M., Michael Celia, and Jan M. Nordbotten, 2011: An Efficient Software Framework for Performing Industrial Risk Assessment of Leakage for Geological Storage of CO₂. International Conference on Greenhouse Gas Technologies (GHGT 10), Elsevier/Energy Procedia, doi:10.1016/j.egypro.2011.02.368 4207-4214
    [ Abstract ]

    In response to anthropogenic CO₂ emissions, geological storage has emerged as a practical and scalable bridge technology while renewables and other environmentally friendly energy production methods mature. While an attractive solution, geological storage of CO₂ has inherent risk. Two primary concerns are recognized: 1) leakage of CO₂through caprock imperfections, and 2) brine displacement resulting in contamination of drinking water sources. Three mechanisms for both CO₂ and brine leakage have been identified: diffuse leakage through the caprock, leakage through faults and fractures in the caprock, and finally, leakage through man-made pathways such as abandoned wells from oil and gas exploration. While the first two leakage mechanisms are important, we emphasize the risks associated with the presence of abandoned wells. This is due to the large number and density of wells from a history of oil and gas exploration around the world, and the high degree of uncertainty surrounding the properties of these abandoned wells. With current proposed legislation in both the United States and Europe, a need is emerging for practical assessment of leakage risk. In order to accurately predict leakage of brine and CO₂ from the injection layer, the geological information for the injection site and the location and makeup of the man-made leakage pathways previously alluded to must be taken into account. Unfortunately, both the geology and abandoned well metadata are typically high in uncertainty, which must be accounted for. With such a high number of random variables, the current state of the art is running many realizations of a system, using a Monte Carlo approach. This requires that the underlying solution algorithms be accurate, and efficient. In the past, many researchers in both academia and industry have turned to robust numerical analysis packages used in the oil industry. However, due to the large range of scales important to this problem (domains of tens of kilometers on a side affected by leakage pathways with diameters of tens of centimeters) such modeling techniques become computationally expensive for all but the most basic analysis. A computational model developed at Princeton University, and currently being commercialized by Geological Storage Consultants, LLC has been shown to be efficient with sufficient accuracy to allow for comprehensive risk assessment of CO₂ injection projects. The model allows for mixing solution methods- using computationally expensive algorithms for formations of greater importance (e.g.- the injection formation) and more efficient, simplified algorithms in other areas of the domain. This ability to arbitrarily mix solution methods offers significant flexibility in the design and execution of models. This paper addresses the framework and algorithms used, and illustrates the importance of efficiency and parallelism using the case study of an injection site in Alberta, Canada. We show how the framework can be used for project planning, for risk mitigation (insurance), and for regulatory groups. Finally, the importance of flexible analysis tools that allow for efficient and effective management of computational resources is discussed.

12. Gasda, S., Jan M. Nordbotten, and Michael Celia, 2011: The impact of local-scale processes on large-scale CO₂ migration and immobilization. International Conference on Greenhouse Gas Technologies (GHGT 10), Elsevier/Energy Procedia, doi:10.1016/j.egypro.2011.02.327 3896-3903
    [ Abstract ]

    Storage security of injected carbon dioxide (CO₂) is an essential component of risk management for geological carbon sequestration operations. During the injection and early post-injection periods, CO₂ leakage may occur along faults and leaky wells, but this risk may be partly managed by proper site selection and sensible deployment of monitoring and remediation technologies. On the other hand, long-term storage security is an entirely different risk management problem—one that is dominated by a mobile CO₂ plume that may travel over very large spatial distances, over long time periods, before it is trapped by a variety of different physical and chemical processes. In the post-injection phase, the mobile CO₂ plume migrates in large part due to buoyancy forces, following the natural topography of the geological formation. The primary trapping mechanisms are capillary and solubility trapping, which evolve over thousands to tens of thousands of years and can immobilize a significant portion of the mobile, free-phase CO₂ plume. However, both the migration and trapping processes are inherently complex, involving a combination of small and large spatial scales and acting over a range of time scales. Solubility trapping is a prime example of this complexity, where small-scale density instabilities in the dissolved CO₂ region leads to convective mixing that has that has a significant effect on the large-scale dissolution process over very long time scales. Another example is the effect of capillary forces on the evolution of mobile CO₂, an often-neglected process except with regard to residual trapping. As the plume migrates due to buoyancy and viscous forces, local capillary effects acting at the CO₂-brine interface lead to a transition zone where both fluids are present in the mobile state. This small-scale effect may have a significant impact on large-scale plume migration as well as long-term residual and dissolution trapping. Using appropriate models that can capture both large and small-scale effects is essential for understanding the role of these processes on the long-term storage security of CO₂ sequestration operations. There are several approaches to modeling long-term CO₂ trapping mechanisms. One modeling option is the use of traditional numerical methods, which are often highly sophisticated models that can handle multiple complex phenomena with high levels of accuracy. However, these complex models quickly become prohibitively expensive for the type of large-scale, long-term modeling that is necessary for risk assessment applications such as the late post-injection period. We present an alternative modeling option that combines vertically-averaged governing equations with an upscaled representation of the dissolutionconvective mixing process and the local capillary transition zone at the CO₂-brine interface. CO₂ injection is solved numerically on a coarse grid, capturing the large-scale injection problem and the post-injection capillary trapping, while the upscaled dissolution and capillary fringe models capture these subscale effects and eliminate the need for expensive grid refinement to capture the subscale instabilities associated with convective mixing or the details of the capillary transition zone. With thismodeling approach, we demonstrate the effect of different modeling choices associated with dissolution and capillary processes for typical large-scale geological systems.

13. Gasda, S., Jan M. Nordbotten, and Michael Celia, 2011: Vertically averaged approaches for CO₂ migration with solubility trapping. Water Resources Research, American Geophysical Union, 47(W05528), doi:10.1029/2010WR009075 1-14
    [ Abstract ]

    The long-term storage security of injected carbon dioxide (CO₂) is an essential component of geological carbon sequestration operations. In the postinjection phase, the mobile CO₂ plume migrates in large part because of buoyancy forces, following the natural topography of the geological formation. The primary trapping mechanisms are capillary and solubility trapping, which evolve over hundreds to thousands of years and can immobilize a significant portion of the mobile CO₂ plume. However, both the migration and trapping processes are inherently complex, spanning multiple spatial and temporal scales. Using an appropriate model that can capture both large- and small-scale effects is essential for understanding the role of these processes on the long-term storage security of CO₂ sequestration operations. Traditional numerical models quickly become prohibitively expensive for the type of large-scale, long-term modeling that is necessary for characterizing the migration and immobilization of CO₂ during the postinjection period. We present an alternative modeling option that combines vertically integrated governing equations with an upscaled representation of the dissolution-convection process. With this approach, we demonstrate the effect of different modeling choices for typical large-scale geological systems and show that practical calculations can be performed at the temporal and spatial scales of interest.

14. Gasda, S., James Z. Wang, and Michael Celia, 2011: Analysis of in-situ wellbore integrity data for existing wells with long-term exposure to CO₂. Energy Procedia, Elsevier, 4, doi:10.1016/j.egypro.2011.02.525 5406-5413
    [ Abstract ]

    An important aspect of the risk associated with geological carbon dioxide sequestration is the integrity of existing wellbores that penetrate geological layers targeted for CO₂ injection. CO₂ leakage may occur through multiple pathways along a wellbore within the 'disturbed zone' surrounding the well casing. The disturbed zone is defined as the annular region along the exterior of the steel wellbore casing that includes the Portland cement sheath, the damage zone of the host rock and the casing-cement-rock interfaces. The effective permeability of this zone is a key parameter of wellbore integrity required for validation of numerical models. Effective permeability may depend on a number of complex factors, including long-term attack by aggressive fluids, poor well completion or actions related to production of fluids through the wellbore. Field tests are essential to understanding the in situ leakage properties of the millions of wells that exist in mature sedimentary basins in North America. We present results from recent field studies of different CO₂ producing wells from both natural CO₂ reservoirs and enhanced oil recovery (EOR) operations. These surveys have included a particular downhole pressure test, the vertical interference test (VIT), designed to determine the extent of hydraulic communication along the exterior of the well casing. The VIT test involves perforating the well casing in two separate intervals, both of which are located within the shale caprock and bracket a zone of cement identified to have a lower quality bond. Once the intervals are isolated with an inflatable packer, the system is pressurized from surface and held at a constant pressure, while simultaneously, the transient pressure response is measured in the lower isolated interval. The pressure transient data is an indicator of the extent of hydraulic communication and is the focus of subsequent analysis. The effective wellbore permeability can be determined through numerical analysis of the VIT data. Our objective is to identify to most effective method of analysis for estimating wellbore permeability. We evaluate two different automated parameter estimation methods, nonlinear regression and shuffled complex evolution metropolis methods. Within this study, we also estimated parameters such as permeability and compressibility of the low permeability shale zone to determine their effect on the resulting estimate of wellbore permeability. The results of this work demonstrate that parameter estimation can be effective at identifying the key parameters associated with wellbore integrity from VIT field tests, and ultimately reducing the uncertainty regarding the integrity of existing wellbores.

15. Nogues, J. P., Jan M. Nordbotten, and Michael Celia, 2011: Detecting leakage of brine or CO₂ through abandoned wells in a geological sequestration operation using pressure monitoring wells. Energy Procedia, 4, doi:10.1016/j.egypro.2011.02.292 3620-3627
    [ Abstract ]

    For risk assessment, policy design and GHG emission accounting it is extremely important to know if any CO₂ or brine has leaked from a geological sequestration (GS) operation. As such, it is important to understand if it is possible to use certain technologies to detect it. This detection of leakage is one of the most challenging problems associated with GS due to the high uncertainty in the nature and location of leakage pathways. In North America for example millions of legacy oil and gas wells present the possibility of CO₂ and brine to leak out of the injection formation. The available information for these potential leaky wells is very limited and the main parameters that control leakage, like permeability of the sealing material are not known. Here we propose to explore the possibility of detecting such leakage by the use of pressure-monitoring wells located in a formation overlying the injection formation. The detection analysis is based on a system of equations that solve for the propagation of a pressure pulse using the superposition principle and an approximation to the well function. We explore the questions of what can be gained by using pressure-monitoring wells and what are the limitations given a specific accuracy threshold of the measuring device. We also try to answer the question of where these monitoring wells should be placed to optimize the objective of a monitoring scheme. We believe these results can ultimately lead to practical design strategies for monitoring schemes, including quantitative estimation of increased probability of leak detection per added observation well.

16. Court, Benjamin, Michael Celia, Jan M. Nordbotten, and Thomas R. Elliot, 2010: Active and Integrated Management of Water Resources Throughout CO₂ Capture and Sequestration Operations. International Conference on Greenhouse Gas Technologies (GHGT 10), Elsevier/Energy Procedia,
    [ Abstract ]

    Most projected climate change mitigation strategies will require a significant expansion of CO₂ Capture and Sequestration (CCS) in the next two decades. Four major categories of challenges are being actively researched: CO₂ capture cost, geological sequestration safety, legal and regulatory barriers, and public acceptance. Herein we propose an additional major challenge category across all CCS operations: water management. For example a coal-fired power plant retrofitted for CCS requires twice as much cooling water as the original plant. This increased demand may be accommodated by brine extraction and treatment, which would concurrently function as large-scale pressure management and a potential source of freshwater. At present the interactions among freshwater extraction, CO₂ injection, and brine management are being considered too narrowly -in the case of freshwater almost completely overlooked- in the technical and regulatory CCS community. This paper presents an overview of each of these challenges and potential integration opportunities. Active management of CCS operations through an integrated approach -including brine production, treatment, use for cooling, and partial reinjection- can address challenges simultaneously with several synergistic advantages. The paper also considers the related potential impacts of pore space competition (with future groundwater use, gas storage and shale gas) on CCS expansion. Freshwater and brine must become key decision making inputs throughout CCS operations, building on existing successful industrial-scale integrations.

17. Crow, W., J. W. Carey, S. Gasda, D. B. Williams, and Michael Celia, 2010: Wellbore integrity analysis of a natural CO₂ producer. International Journal of Greenhouse Gas Control, Elsevier, (4), doi:10.1016/j.ijggc.2009.10.010 186-197
    [ Abstract ]

    Long-term integrity of existing wells in a CO₂-rich environment is essential for ensuring that geological sequestration of CO₂ will be an effective technology for mitigating greenhouse gas-induced climate change. The potential for wellbore leakage depends in part on the quality of the original construction as well as geochemical and geomechanical stresses that occur over its life-cycle. Field data are essential for assessing the integrated effect of these factors and their impact on wellbore integrity, defined as the maintenance of isolation between subsurface intervals. In this report, we investigate a 30-year-old well from a natural CO₂ production reservoir using a suite of downhole and laboratory tests to characterize isolation performance. These tests included mineralogical and hydrological characterization of 10 core samples of casing/cement/formation, wireline surveys to evaluate well conditions, fluid samples and an in situ permeability test. We find evidence for CO₂ migration in the occurrence of carbonated cement and calculate that the effective permeability of an 11'-region of the wellbore barrier system was between 0.5 and 1 milliDarcy. Despite these observations, we find that the amount of fluid migration along the wellbore was probably small because of several factors: the amount of carbonation decreased with distance from the reservoir, cement permeability was low (0.3-30 microDarcy), the cement-casing and cement-formation interfaces were tight, the casing was not corroded, fluid samples lacked CO₂, and the pressure gradient between reservoir and caprock was maintained. We conclude that the barrier system has ultimately performed well over the last 3 decades. These results will be used as part of a broader effort to develop a long-term predictive simulation tool to assess wellbore integrity performance in CO₂ storage sites.

18. Nordbotten, Jan M., J. P. Nogues, and Michael Celia, March 2010: Appropriate Choice of Average Pressure for Upscaling Relative Permeability in Dynamic Flow Conditions. SPE Journal, 15(1), doi:10.2118/113558-PA 228-237
    [ Abstract ]

    A new macroscale pressure definition is investigated through a series of upscaling calculations for two-phase flow in porous media. This definition is taken from previous theoretical work by the authors and aims at correcting for systematic subscale heterogeneities including those generated by nonlinear dependencies on (heterogeneous) saturation distributions. Traditional intrinsic phase-averaged pressure leads to nonmonotone and multivalued upscaled constitutive functions (e.g., discontinuous upscaled-relative-permeabilities exceeding unity). Using both analytical and numerical calculations, the new macroscale pressure definition is shown to lead to better-behaved upscaled functions.

19. Person, M., A. Banerjee, J. Rupp, Cristian Medina, P. Lichtner, Carl Gable, R. Pawar, Michael Celia, Jennifer McIntosh, and Victor Bense, 2010: Assessment of Basin-scale Hydrologic Impacts of CO₂ Sequestration, Illinois Basin. International Journal of Greenhouse Gas Control, (4), doi:10.1016/j.ijggc.2010.04.004 840-854
    [ Abstract ]

    Idealized, basin-scale sharp-interface models of CO₂ injection were constructed for the Illinois basin. Porosity and permeability were decreased with depth within the Mount Simon Formation. Eau Claire confining unit porosity and permeability were kept fixed. We used 726 injection wells located near 42 power plants to deliver 80 million metric tons of CO₂/year. After 100 years of continuous injection, deviatoric fluid pressures varied between 5.6 and 18 MPa across central and southern part of the Illinois basin. Maximum deviatoric pressure reached about 50% of lithostatic levels to the south. The pressure disturbance (>0.03 MPa) propagated 10–25 km away from the injection wells resulting in significant well–well pressure interference. These findings are consistent with single-phase analytical solutions of injection. The radial footprint of the CO₂ plume at each well was only 0.5–2 km after 100 years of injection. Net lateral brine displacement was insignificant due to increasing radial distance from injection well and leakage across the Eau Claire confining unit. On geologic time scales CO2 would migrate northward at a rate of about 6 m/1000 years. Because of paleo-seismic events in this region (M5.5–M7.5), care should be taken to avoid high pore pressures in the southern Illinois basin.

20. Bachu, S., and Michael Celia, 2009: Assessing the Potential for CO2 Leakage, Particularly through Wells, from CO2 Storage Sites. The Science and Technology of Carbon Sequestration, http://cat.inist.fr/?aModele=afficheN&cpsidt=22364988, 183, 203-216
    [ Abstract ]

    Assessment of the potential for CO2 leakage from geological storage sites is essential for the implementation of CO2 capture and storage in geological media. Possible pathways for CO2 leakage from a storage site include natural interruptions and breaches through the confining strata, faults and fractures, and degraded wells. Knowledge of the geology and stress regime is essential in assessing the potential for CO2 leakage through natural features and induced fractures. Assessment of the potential for leakage through degraded wells is much more difficult because of the large number of wells, the lack of knowledge about their condition, and the computational difficulties relating to the simulation of CO2 leakage through many wells across a multilayered succession of aquifers and aquitards. The large number of wells and the variability in their present and future conditions require a stochastic approach by which a large number of statistical realizations provides a probability distribution for CO2 leakage. The large disparity between the length scales associated with injected plumes and those associated with leakage pathways along wells leads to numerical intractability for statistical simulations. Semianalytical models, although constrained by assumptions needed to solve the mathematical system of equations, provide a framework for estimating the potential for and rates of CO2 leakage through degraded wells. An example from the Alberta Basin in Canada provides an illustration of the types of information these models can generate. The models must be coupled to specific field observational and measurement programs to support full implementation of CO2 geological storage.

21. Celia, Michael, Jan M. Nordbotten, , M. Dobossy, and Benjamin Court, 2009: Risk of Leakage versus Depth of Injection in Geological Storage. Energy Procedia, 1(1), doi:10.1016/j.egypro.2009.02.022 2573-2580
    [ Abstract ]

    One of the outstanding challenges for large-scale CCS operations is to develop reliable quantitative risk assessments with a focus on leakage of both injected CO₂ and displaced brine. A critical leakage pathway is associated with the century-long legacy of oil and gas exploration and production, which has led to many millions of wells being drilled. Many of those wells are in locations that would otherwise be excellent candidates for CCS operations, especially across many parts of North America. Quantitative analysis of the problem requires special computational techniques because of the unique challenges associated with simulation of injection and leakage in systems that include hundreds or thousands of existing wells over domains characterized by layered structures in the vertical direction and very large horizontal extent. An important feature of these kinds of systems is the depth of each well, and the fact that the number of wells penetrating different formations decreases as a function of depth. As such, one might reasonably expect the risk of leakage to decrease with depth of injection. With the special computational models developed to simulate injection and leakage along multiple wells, in layered systems with multiple formations, quantitative assessment of risk reduction as a function of injection depth can be made. An example of such a system corresponds to the Wabamun Lake area southwest of Edmonton, Alberta, Canada, where several large coal-fired power plants are located. Use of information about both the existing wells and the local stratigraphy allows a realistic model to be constructed. Leakage along existing wells is assumed to follow Darcy’s Law, and is characterized by a set of effective permeability values. These values are assigned stochastically, using several different methods, within a Monte Carlo simulation framework. Computational results show the clear trade-off between depth of injection and risk of leakage. The results also show how properties within the different formations affect the risk profiles. In the Wabamun Lake area, one of the formations has the highest injectivity, by far, while having a moderate number of existing wells. Its moderate risk of leakage, as compared to injections in formations above and below, shows some of the key factors that are likely to influence injection design for large-scale CCS operations.

22. Celia, Michael, and Jan M. Nordbotten, 2009: Practical Modeling Approaches for Geological Storage of Carbon Dioxide. Ground Water, 47(5), doi:10.1111/j.1745-6584.2009.00590.x 627–638
    [ Abstract ]

    The relentless increase of anthropogenic carbon dioxide emissions and the associated concerns about climate change have motivated new ideas about carbon-constrained energy production. One technological approach to control carbon dioxide emissions is carbon capture and storage, or CCS. The underlying idea of CCS is to capture the carbon before it emitted to the atmosphere and store it somewhere other than the atmosphere. Currently, the most attractive option for large-scale storage is in deep geological formations, including deep saline aquifers. Many physical and chemical processes can affect the fate of the injected CO2, with the overall mathematical description of the complete system becoming very complex. Our approach to the problem has been to reduce complexity as much as possible, so that we can focus on the few truly important questions about the injected CO2, most of which involve leakage out of the injection formation. Toward this end, we have established a set of simplifying assumptions that allow us to derive simplified models, which can be solved numerically or, for the most simplified cases, analytically. These simplified models allow calculation of solutions to large-scale injection and leakage problems in ways that traditional multicomponent multiphase simulators cannot. Such simplified models provide important tools for system analysis, screening calculations, and overall risk-assessment calculations. We believe this is a practical and important approach to model geological storage of carbon dioxide. It also serves as an example of how complex systems can be simplified while retaining the essential physics of the problem.

23. Class, H., R. Ebigbo, R. Helmig, H. Dahle, Jan M. Nordbotten, Michael Celia, P. Audigane, M. Darcis, J. Ennis-King, Y. Fan, B. Flemisch, and S. Gasda, et al., 2009: A Benchmark Study on Problems Related to CO₂ Storage in Geological Formations: Summary and Discussion of the Results. Computational Geosciences, doi:10.1007/s10596-009-9146-x
    [ Abstract ]

    This paper summarises the results of a benchmark study that compares a number of mathematical and numerical models applied to specific problems in the context of carbon dioxide (CO₂) storage in geologic formations. The processes modeled comprise advective multiphase flow, compositional effects due to dissolution of CO₂ into the ambient brine, and non-isothermal effects due to temperature gradients and the Joule-Thompson effect. The problems deal with leakage through a leaky well, methane recovery enhanced by CO₂ injection, and a reservoir-scale injection scenario into a heterogeneous formation. We give a description of the benchmark problems, then briefly introduce the participating codes, and finally present and discuss the results of the benchmark study.

24. Crow, W., D. B. Williams, J. W. Carey, Michael Celia, and S. Gasda, 2009: Solubility and Diffusivity of SO₂ for Co-injection with CO₂ in Geological Sequestration. EOS Trans. AGU, 89(53), Fall Meet. S,
    [ Abstract ]

    There are potential economic benefits to the co-injection of SO₂ with CO₂ in the context of geological sequestration, but the impact of this co-injection on the fate and migration of SO₂ and CO₂ is poorly understood. Previous modeling studies have shown that injection of SO₂ with CO₂ would create highly acidic conditions due to formation of sulfuric acid. However, little is known regarding the solubility of SO₂ under high pressure, high salinity conditions, and the kinetic limitations of SO₂ diffusion in a CO₂ phase. A method to estimate the phase partitioning of SO₂ under geological storage conditions was developed in this study. The method uses the Krichevsky-Ilinskaya equation to correct for high pressures and the Schumpe model for mixed electrolyte solutions. Henry's constants for a broad range of brine solutions were calculated at storage conditions of 100 bar pressure. The Henry's constant for SO₂ is 1.5 M/atm at 40°C and is 0.86 M/atm at 60°C. Under these same conditions, the Henry's constant for CO₂ is much smaller, roughly 0.01 M/atm (40°C to 60°C). Henry's constants increase with increasing pressure but decrease with increasing temperature. These effects can be observed by comparing the SO₂ Henry's constants under storage conditions with the value under ambient temperature and pressure conditions in pure water, 1.2 M/atm. To simulate diffusion through stationary CO₂, a nonsteady state two-dimensional model of SO2 diffusion through supercritical CO₂ was also created. A binary diffusion coefficient of 5×10-8 m²/sec was estimated based on the Takahashi correlation to account for high pressures, where a low pressure coefficient was determined using the Fuller estimation. Binary diffusion coefficients for polar compounds in supercritical CO₂ have been previously studied and are on the same order of magnitude as the binary diffusion coefficient estimated in this study. The system that was modeled is a cone-shaped system representing separate-phase CO₂ confined in a formation after injection. Boundary conditions consisted of a no-flux boundary at the top of the cone to account for the impermeable confining caprock, and a zero concentration boundary at the cone edge to simulate a worst case scenario for dissolution. The initial conditions considered a uniform concentration of one percent SO₂ everywhere in the cone. To numerically simulate the concentration profile throughout the cone, a time-split explicit difference method was applied. The diffusion modeling results show that contact between SO₂ and formation brine will be diffusion limited; after 3000 years pproximately 75% of sulfur remains in the cone. In summary, while SO₂ is highly soluble in water, its slow diffusion through a supercritical CO2 phase will likely inhibit its mass transfer.

25. Crow, W., D. B. Williams, J. W. Carey, Michael Celia, and S. Gasda, 2009: Wellbore Integrity Analysis of a Natural CO₂ Producer. Energy Procedia, 1, doi:10.1016/j.egypro.2009.02.150 3561-3569
    [ Abstract ]

    The long-term integrity of wellbores in a CO₂-rich environment is a complex function of material properties and reservoir conditions including brine and rock compositions, CO₂- pressure, and formation pressure and temperature gradients. Laboratory experiments can provide essential information on rates of material reaction with CO₂-. However, field data are essential for assessing the integrated effect of these factors in subsurface conditions to provide a basis for validation of numerical models of wellbore behavior. We present a comprehensive study and conclusions from an investigation of a 30-year old well from a natural CO₂- production reservoir. The wellbore was exposed to a 96% CO₂- fluid from the time of cement placement. This site is unique for two reasons: it represents a higher, sustained concentration of CO₂- compared to enhanced oil recovery fields and both the reservoir and caprock are clastic rocks that may possess less buffering capacity than carbonate reservoirs. A sampling program resulted in the recovery of 10 side-wall cement cores extending from the reservoir through the caprock. The hydrologic, mineralogical and mechanical properties of these samples were measured and those results were combined with an in-situ pressure-response test to investigate cement integrity over a range of length scales. Fluid sampling was conducted with pressure and temperature measurements for geochemical analysis of the cemented annulus and the adjacent formation. These combined data sets provide an assessment of well integrity including original cement seal and the impacts of CO₂-. Cement evaluation wireline surveys indicate good coverage and bonding, consistent with observations from sidewall cement core samples that have tight interfaces with the casing and formation. Although alteration of the cement samples is present in all cores in varying degrees, hydraulic isolation has prevented leakage based on the pressure gradient measured between the caprock and CO₂- formation as well as lack of corrosion and no casing pressure history. Simulation of a hydraulic isolation test (Vertical Interference Test) indicates the best match for effective permeability of the wellbore system is approximately 1-10 millidarcies which suggests cement interfaces are a more significant potential migration pathway as compared with the cement matrix. Effective placement of the Portland-fly ash cement system was a key element in the observed performance of the barrier system that provides hydraulic isolation. The types of information collected in this survey permit analysis of individual components (casing, cement and reservoir fluid and pressure measurements) for comparison to the larger scale system including the interfaces. The results will be used as part of the CO₂- Capture Project’s effort to develop a long-term predictive simulation tool to assess wellbore integrity performance in CO₂- storage sites.

26. Gasda, S., Michael Celia, and Jan M. Nordbotten, 2009: Vertical Equilibrium with Subscale Analytical Methods for Geological CO₂ Sequestration. Computational Geosciences, doi:10.1007/S10596-009-9138-X
    [ Abstract ]

    Large-scale implementation of geological CO₂ sequestration requires quantification of risk and leakage potential. One potentially important leakage pathway for the injected CO₂ involves existing oil and gas wells. Wells are particularly important in North America, where more than a century of drilling has created millions of oil and gas wells. Models of CO₂ injection and leakage will involve large uncertainties in parameters associated with wells, and therefore a probabilistic framework is required. These models must be able to capture both the largescale CO₂ plume associated with the injection and the small-scale leakage problem associated with localized flow along wells. Within a typical simulation domain, many hundreds of wells may exist. One effective modeling strategy combines both numerical and analytical models with a specific set of simplifying assumptions to produce an efficient numerical–analytical hybrid model. The model solves a set of governing equations derived by vertical averaging with assumptions of a macroscopic sharp interface and vertical equilibrium. These equations are solved numerically on a relatively coarse grid, with an analytical model embedded to solve for wellbore flow occurring at the sub-gridblock scale. This vertical equilibrium with sub-scale analytical method (VESA) combines the flexibility of a numerical method, allowing for heterogeneous and geologically complex systems, with the efficiency and accuracy of an analytical method, thereby eliminating expensive grid refinement for sub-scale features. Through a series of benchmark problems, we show that VESA compares well with traditional numerical simulations and to a semi-analytical model which applies to appropriately simple systems. We believe that the VESA model provides the necessary accuracy and efficiency for applications of risk analysis in many CO₂ sequestration problems.

27. Nordbotten, Jan M., D. Kavetski, Michael Celia, and , 2009: Model for CO₂ Leakage Including Multiple Geological Layers and Multiple Leaky Wells. Environmental Science and Technology, 43(3), doi:10.1021/es801135v 743-749
    [ Abstract ]

    Geological storage of carbon dioxide (CO₂) is likely to be an integral component of any realistic plan to reduce anthropogenic greenhouse gas emissions. In conjunction with large-scale deployment of carbon storage as a technology, there is an urgent need for tools which provide reliable and quick assessments of aquifer storage performance. Previously, abandoned wells from over a century of oil and gas exploration and production have been identified as critical potential leakage paths. The practical importance of abandoned wells is emphasized by the correlation of heavy CO₂) emitters (typically associated with industrialized areas) to oil and gas producing regions in North America. Herein, we describe a novel framework for predicting the leakage from large numbers of abandoned wells, forming leakage paths connecting multiple subsurface permeable formations. The framework is designed to exploit analytical solutions to various components of the problem and, ultimately, leads to a grid-free approximation to CO₂) and brine leakage rates, as well as fluid distributions. We apply our model in a comparison to an established numerical solver for the underlying governing equations. Thereafter, we demonstrate the capabilities of the model on typical field data taken from the vicinity of Edmonton, Alberta. This data set consists of over 500 wells and 7 permeable formations. Results show the flexibility and utility of the solution methods, and highlight the role that analytical and semi-analytical solutions can play in this important problem.

28. Peters, Catherine A., George Scherer, Michael Celia, Jean Hervé Prévost, T. C. Onstott, P. F. Dobson, C. M Oldenburg, B. Freifeld, J. Birkholzer, J. Wang, S. Benson, and T. J. Phelps, et al., in press: Collaborative Research: DUSEL CO₂, A Deep Underground Laboratory for Geologic CO₂ Sequestration Studies: A proposal for the conceptual design of the facility and experiments. NSF. 0/09.
    [ Abstract ]

    Princeton University and Lawrence Berkeley National Laboratory have forged a new collaboration to examine the feasibility and risks of carbon sequestration, a method of countering global warming by storing greenhouse gases deep underground. To develop a sound understanding of carbon sequestration, we will build a deep underground laboratory to study the processes of trapping and storing CO₂, including the risks of unintended leakage. It will be part of the new DUSEL facility at the Homestake mine in South Dakota. The “DUSEL CO₂, facility will make the United States the only country with a deep underground laboratory for controlled study of geologic carbon sequestration, providing a unique opportunity for global leadership. The findings from these unique experiments will advance carbon management technology worldwide and help reduce global greenhouse gas emissions. The features and capabilities of the planned facility are unprecedented. The experimental design exploits the nearly half-kilometer vertical extent of existing “sandline” borings at Homestake. Pipes will be installed within the sandlines to serve as long flow columns. These columns will contain the CO₂, and allow experimentation at the same pressure and temperature conditions as in deep subsurface reservoirs. Fill materials will mimic sedimentary layering, as well as cements in plugged wells. Instrumentation will enable detailed monitoring of flow, pressure, temperature, brine composition, geomechanics, and microbial activity. As part of the initial suite of experiments, we plan to simulate a leak in which CO₂, changes from a supercritical fluid to a subcritical gas as the pressure drops during upflow over tens to hundreds of meters. We will test for possible acceleration in CO₂, flow due to increasing buoyancy. Also, we will examine the interactions of CO₂, with cap-rocks and well cements, and determine whether CO₂, will enlarge flow pathways or cause selfsealing. Finally, we will investigate the effects of anaerobic, thermophilic bacteria on CO₂, conversion to methane and carbonate. This project is being led by researchers at Princeton and LBNL, and involves no-cost collaboration with individuals at ORNL, Stanford University, Schlumberger and the U.S. DOE NETL. During this three-year project, the team is working to (i) prioritize future experiments that will be conducted at DUSEL CO₂, (ii) build models that simulate experimental conditions and predict process dynamics, and (iii) develop a Work-Breakdown Structure (WBS) schedule for design, procurement, construction, operation and deconstruction of the facility over the facility lifetime. International awareness about DUSEL CO₂, is being fostered through international workshops and formation of an International Advisory Committee. Also, we are collaborating with other DUSEL scientists on education and outreach about “deep science,” with particular focus on climate change and energy solutions. DUSEL education and outreach activities are focused on Native American communities in South Dakota and operation of the Visitor Center at the Sanford Lab at Homestake. To inspire and educate the next generation of leaders, we are involving undergraduate and graduate students in DUSEL CO₂, research at Princeton University.

29. Binning, P. J., and Michael Celia, 2008: Pseudokinetics arising from the Upscaling of Geochemical Equilibrium. Water Resources Research, 44, doi:10.1029/2007WR006147
    [ Abstract ]

    Multicomponent contaminant transport models in groundwater are typically based on assumptions of local geochemical equilibrium on the grid scale. However, in heterogenous systems there may be significant coupling between transport processes and geochemical equilibrium at smaller than grid block scale. Here various pore- and fieldscale examples are considered to illustrate the impact of transport processes on assumptions of geochemical equilibrium. In each example the flow length scales required to reach equilibrium are calculated. It is shown that these can be as large as many meters at the pore scale and kilometers at field scales. The influence of heterogeneity in the distribution of the reactive zones is assessed for the pore-scale example, and it is shown that patchiness of reactive zones within a pore increases equilibration length, with the length and density of reactive zones, pore radius, and diffusion coefficient all playing a role in the equilibration length. When constructing models of field-scale problems it may not be reasonable to apply geochemical equilibrium, and it may be necessary to explicitly couple pore-scale and field-scale models in a multiscale simulation. A field-scale example is also shown to illustrate that the upscaling of geochemical equilibrium poses a significant practical problem because we usually do not know the spatial location and distribution of geochemically active sites, and this information is essential input to geochemical transport models.

30. Gasda, S., Jan M. Nordbotten, and Michael Celia, 2008: Determining Effective Wellbore Permeability from a Field Pressure Test: A Numerical Analysis of Detection Limits. Environmental Geology, 54(6), doi:10.1007/S00254-007-0903-7 1207-1215
    [ Abstract ]

    We propose a simple pressure test that can be used in the field to determine the effective permeability of existing wellbores. Such tests are motivated by the need to understand and quantify leakage risks associated with geological storage of CO₂ in mature sedimentary basins. If CO₂ is injected into a deep geological formation, and the resulting CO₂ plume encounters a wellbore, leakage may occur through various pathways in the ‘‘disturbed zone’’ surrounding the well casing. The effective permeability of this composite zone, on the outside of the well casing, is an important parameter for models of leakage. However, the data that exist on this key parameter do not exist in the open literature, and therefore specific field tests need to be done in order to reduce the uncertainty inherent in the leakage estimates. The test designed and analyzed herein is designed to measure effective wellbore permeability within a lowpermeability caprock, bounded above and below by permeable reservoirs, by pressurizing the reservoir below and measuring the response in the reservoir above. Alternatively, a modified test can be performed within the caprock without directly contacting the reservoirs above and below. We use numerical simulation to relate pressure response to effective well permeability and then evaluate the range of detection of the effective permeability based on instrument measurement error and limits on fracture pressure. These results can guide field experiments associated with site characterization and leakage analysis.

31. Gasda, S., Michael Celia, and Jan M. Nordbotten, 2008: Upslope Plume Migration and Implications for Geological CO₂ Sequestration in Deep Saline Aquifers. IES Journal A: Civil and Structural Engineering, 1(1), doi:10.1080/19373260701620154
    [ Abstract ]

    Recent investigations regarding CO₂ sequestration in deep saline aquifers have focused on characterization of the injected plume, its migration within the aquifer over time, and possible leakage out of the aquifer. To study these complex flow systems, simplified models are sometimes used to describe both plume evolution and the amount of leakage. Simplifications may include an assumption of perfectly horizontal geological formations, negligible capillary pressure, and symmetry of the injection plume. In this study, we explicitly test the limits of the assumption of a horizontal aquifer through numerical simulation of typical injection scenarios in continental sedimentary basins. Our approach is to simulate injection of CO₂ into a confined saline aquifer for an extended period (we have used 15 years) and examine the effect of different degrees of slope, as well as other system parameters, on plume asymmetry using measures such as the location of the centroid of the CO₂ plume. Dimensional analysis of this system shows that the centroid migrates upslope in proportion with buoyancy, aquifer permeability, and slope, whereas increased porosity and CO₂ viscosity mitigate upslope migration of the centroid. The results of this study show that the effect of slope can be ignored for many aquifers likely to become CO₂ sequestration sites in North America. However, slope will be more important for higher permeability aquifers, such as the site used in the Sleipner sequestration project in the North Sea.

32. Nordbotten, Jan M., Michael Celia, H. Dahle, and S. M. Hassanizadeh, 2008: On the Definition of Macroscale Pressure for Multiphase Flow in Porous Media. Water Resources Research, 44(W06S02), doi:10.1029/2006WR005715
    [ Abstract ]

    We consider immiscible two-phase flow in porous media, starting with the Stokes equations. Our analysis leads to Darcy’s law but with notable differences from the usual interpretation. The most immediate difference is the interpretation of macroscale pressure, which, contrary to previous derivations, does not equal the intrinsic phase average pressure. We recover the intrinsic average only when systematic subscale heterogeneities, in material properties or fluid distribution, are absent. Examples using capillary tube and dynamic pore network models are given. These results impact our understanding of multiphase flow and have a direct effect on numerical upscaling efforts, including calculations of continuum-scale flow parameters from pore-scale network models.

33. Person, M., A. Banerjee, J. Rupp, P. Lichtner, R. Pawar, and Michael Celia, 2008: Basin-scale Hydrologic Impacts of CO₂ Sequestration within the Mt. Simojn Formation, Illinois Basin: Scaling Calculations using Sharp-Interface Theory. American Geophysical Union, Fall Meeting,
    [ Abstract ]

    The Illinois Basin hosts dozens of coal fired power plants generating more than 80 million metric tons of CO₂ annually. Here we present a suite of basin-scale, hydrologic models of the Mt Simon formation, Illinois Basin using sharp interface theory. The goal of these models is to determine what are the basin-scale hydrologic consequences of CO₂2 injection and whether some regions of the Illinois Basin would represent a better venue for carbon sequestration than others. While this approach makes some restrictive simplifying assumptions, it allows us to assess the problem at the sedimentary basin scale. Our solution domain spans the northern two thirds of the Illinois Basin (about 230,000 km²). We allowed porosity and permeability to decrease with depth from 0.2 to 0.05 and 400 to 2 mD, respectively. We injected CO₂ using 727, 10 inch diameter injection wells delivering about 210 kg/minute/well. The wells were positioned about 2 km apart in a radial pattern around known power plant locations. We ran the injection wells for 100 years. The wells were then shut in for an additional 900 years. Results indicate that after 100 years of continuous injection, deviatoric fluid pressures varied between 9.2 to 0.5 MPa between the deepest and shallowest injection wells. For the deepest portion of the basin (~ 3.1 km), deviatoric pressures reach about 22 percent of lithostatic levels. Owing to the rather subtle regional hydraulic gradient (200m/500km), long-range separate-phase CO₂-migration is driven by buoyancy at a rate of only 2 m/year. If CO₂ remained as a separate phase on time scales of 100,000 years, the injected CO₂ would migrate about 200 km to the north before charging gentle structural traps. Owing to the radial, bowl-shaped geometry of the Illinois Basin, net brine displacement to the north would be small, probably less than 100 m. Our analysis suggest that the Mt. Simon formation represents a good venue for CO₂ sequestration although shallower regions ( ~ 2 km depth) would pose less risk of catastrophic breaching due to high deviatoric fluid pressures. Fluid pressures do not return to hydrostatic conditions after 1000 years due to buoyant forces resulting from the presence of a separate CO₂ phase.

34. Peters, Catherine A., W. B. Lindquist, and Michael Celia, March 2008: Up-Scaling Mineral Accessibility and Pore Networks for CO₂ Reactive Transport in Sandstones. Global Change Biology,
    [ Abstract ]

    Widespread implementation of geological storage of CO₂ will require an understanding of acid-driven reactions with formation minerals. Predicting these reactions and their time scales requires rate laws that are appropriate for sedimentary rocks and estimates of accessible surface areas of reactive minerals. This project addresses these needs through a study that combines imaging of sandstone pore structure and minerals, and network-modeling of reaction rates in porous media. Rock specimens come from the Viking formation in the Alberta Sedimentary Basin. Imaging methods include X-ray computed microtomography (CT), backscatter electron microscopy (BSE) and energy dispersive X-ray (EDX) spectroscopy.
    One important goal is to characterize pore contact with individual minerals thereby quantifying meaningful surface areas for use in reactive transport models. The suite of techniques employed and the innovative means by which the images are collectively interpreted provides a wealth of information to address this goal. For example, a novel method of interpreting BSE images (which are high resolution) combined with EDX images (which can generate mineral maps) leads to 2D images that provide detailed characterization of the proximity of reactive minerals to pore space. Extension of this image processing approach to 3D, using CT images to broadly classify mineral categories, allows us to relate detailed information about pore structure with mineral accessibility, albeit with coarser resolution.
    All the specimens are sandstones of comparable porosity and grain diameter, and yet order of magnitude variation is found in pore structure and reactive mineral properties across them. In general, we have found that mineral volumetric content is a poor indicator of proportionate poreto- mineral surface area due to the means by which minerals are obscured in consolidated media. For example, kaolinite and other authigenic clay minerals that coat grains and fill primary pore space account for only 5% to 30% of mineral content, but 65% to 90% of pore-mineral contact boundaries. Minerals that would react under acidic conditions may account for 5% to 10% (vol.) of mineral matter, but if these percentages are used to apportion surface area, they would overestimate reaction rates by three to five times.
    These detailed characterizations of pore structure and mineral spatial patterning are being used to develop pore-network models that simulate reactive transport. Simulations of conditions representative of COWidespread implementation of geological storage of CO2 will require an understanding of acid-driven reactions with formation minerals. Predicting these reactions and their time scales requires rate laws that are appropriate for sedimentary rocks and estimates of accessible surface areas of reactive minerals. This project addresses these needs through a study that combines imaging of sandstone pore structure and minerals, and network-modeling of reaction rates in porous media. Rock specimens come from the Viking formation in the Alberta Sedimentary Basin. Imaging methods include X-ray computed microtomography (CT), backscatter electron microscopy (BSE) and energy dispersive X-ray (EDX) spectroscopy. One important goal is to characterize pore contact with individual minerals thereby quantifying meaningful surface areas for use in reactive transport models. The suite of techniques employed and the innovative means by which the images are collectively interpreted provides a wealth of information to address this goal. For example, a novel method of interpreting BSE images (which are high resolution) combined with EDX images (which can generate mineral maps) leads to 2D images that provide detailed characterization of the proximity of reactive minerals to pore space. Extension of this image processing approach to 3D, using CT images to broadly classify mineral categories, allows us to relate detailed information about pore structure with mineral accessibility, albeit with coarser resolution. All the specimens are sandstones of comparable porosity and grain diameter, and yet order of magnitude variation is found in pore structure and reactive mineral properties across them. In general, we have found that mineral volumetric content is a poor indicator of proportionate poreto- mineral surface area due to the means by which minerals are obscured in consolidated media. For example, kaolinite and other authigenic clay minerals that coat grains and fill primary pore space account for only 5% to 30% of mineral content, but 65% to 90% of pore-mineral contact boundaries. Minerals that would react under acidic conditions may account for 5% to 10% (vol.) of mineral matter, but if these percentages are used to apportion surface area, they would overestimate reaction rates by three to five times. These detailed characterizations of pore structure and mineral spatial patterning are being used to develop pore-network models that simulate reactive transport. Simulations of conditions representative of CO₂ injection for geological storage are being used to examine up-scaling of reaction rates from the pore-scale to the core-scale.

35. DePaolo, D. J., F. M. Orr, Jr., S. Benson, and Michael Celia, et al., June 2007: Basic Research Needs for Geosciences: Facilitating 21st Century Energy Systems. Office of Basic Energy Sciences, U.S. DOE, Report from the Workshop Held February 21-23, 2007, http://www.science.doe.gov/bes/reports/files/GEO_rpt.pdf,
    [ Abstract ]

    Serious challenges must be faced in this century as the world seeks to meet global energy needs and at the same time reduce emissions of greenhouse gases to the atmosphere. Even with a growing energy supply from alternative sources, fossil carbon resources will remain in heavy use and will generate large volumes of carbon dioxide (CO₂). To reduce the atmospheric impact of this fossil energy use, it is necessary to capture and sequester a substantial fraction of the produced CO₂. Subsurface geologic formations offer a potential location for long-term storage of the requisite large volumes of CO₂. Nuclear energy resources could also reduce use of carbon based fuels and CO₂ generation, especially if nuclear energy capacity is greatly increased. Nuclear power generation results in spent nuclear fuel and other radioactive materials that also must be sequestered underground. Hence, regardless of technology choices, there will be major increases in the demand to store materials underground in large quantities, for long times, and with increasing efficiency and safety margins.
    Rock formations are composed of complex natural materials and were not designed by nature as storage vaults. If new energy technologies are to be developed in a timely fashion while ensuring public safety, fundamental improvements are needed in our understanding of how these rock formations will perform as storage systems.
    This report describes the scientific challenges associated with geologic sequestration of large volumes of carbon dioxide for hundreds of years, and also addresses the geoscientific aspects of safely storing nuclear waste materials for thousands to hundreds of thousands of years. The fundamental crosscutting challenge is to understand the properties and processes associated with complex and heterogeneous subsurface mineral assemblages comprising porous rock formations, and the equally complex fluids that may reside within and flow through those formations. The relevant physical and chemical interactions occur on spatial scales that range from those of atoms, molecules, and mineral surfaces, up to tens of kilometers, and time scales that range from picoseconds to millennia and longer. To predict with confidence the transport and fate of either CO₂ or the various components of stored nuclear materials, we need to learn to better describe fundamental atomic, molecular, and biological processes, and to translate those microscale descriptions into macroscopic properties of materials and fluids. We also need fundamental advances in the ability to simulate multiscale systems as they are perturbed during sequestration activities and for very long times afterward, and to monitor those systems in real time with increasing spatial and temporal resolution. The ultimate objective is to predict accurately the performance of the subsurface fluid-rock storage systems, and to verify enough of the predicted performance with direct observations to build confidence that the systems will meet their design targets as well as environmental protection goals.
    The report summarizes the results and conclusions of a Workshop on Basic Research Needs for Geosciences held in February 2007. Five panels met, resulting in four Panel Reports, three Grand Challenges, six Priority Research Directions, and three Crosscutting Research Issues. The Grand Challenges differ from the Priority Research Directions in that the former describe broader, longterm objectives while the latter are more focused.

36. Meng, K., Robert H. Williams, and Michael Celia, 2007: Opportunities for low-cost CO₂ storage demonstration projects in China. Energy Policy, 35(4), doi:10.1016/j.enpol.2006.08.016 2368-2378
    [ Abstract ]

    Several CO₂ storage demonstration projects are needed in a variety of geological formations worldwide to prove the viability of CO₂ capture and storage as a major option for climate change mitigation. China has several low-cost CO₂ sources at sites that produce NH₃ from coal via gasification. At these plants, CO₂ generated in excess of the amount needed for other purposes (e.g., urea synthesis) is vented as a relatively pure stream. These CO₂ sources would potentially be economically interesting candidates for storage demonstration projects if there are suitable storage sites nearby. In this study a survey was conducted to estimate CO₂ availability at modern Chinese coal-fed ammonia plants. Results indicate that annual quantities of available, relatively pure CO₂ per site range from 0.6 to 1.1 million tonnes. The CO₂ source assessment was complemented by analysis of possible nearby opportunities for CO₂ storage. CO₂ sources were mapped in relation to China’s petroliferous sedimentary basins where prospective CO₂ storage reservoirs possibly exist. Four promising pairs of sources and sinks were identified. Project costs for storage in deep saline aquifers were estimated for each pairing ranging from $15–21/t of CO₂. Potential enhanced oil recovery and enhanced coal bed methane recovery opportunities near each prospective source were also considered.

37. Nordbotten, Jan M., Michael Celia, H. Dahle, and S. M. Hassanizadeh, 2007: Interpretation of Macro-Scale Variables in Darcy's Law. Water Resources Research, 43(W08430), doi:10.1029/2006WR005018
    [ Abstract ]

    The pursuit of a theoretical foundation of Darcy’s law based on volume averaging of equations at the scale of flow in pores has a long history. While theories are well established for homogeneous systems, more complex systems exhibit inconsistencies in the resulting equations. The difficulties often lie in the treatment of surface integral terms arising from the classical averaging theorems used to transform averages of derivatives into derivatives of averages. In this work we extend the intrinsic phase average as a macroscale variable to a family of more general macroscale variables, which take into account systematic dependencies of averaging volume size on the macroscale. Comparison to Darcy’s law gives new insight into the relationship between variables at the microscale and macroscale.

38. Nordbotten, Jan M., I. Rodriguez-Iturbe, and Michael Celia, 2007: Stochastic Coupling of Rainfall and Biomass Dynamics In , 43(W01408), doi:10.1029/2006WR005068
    [ Abstract ]

    A modeling scheme is presented to derive the probabilistic structure of plant biomass when subjected to stochastic precipitation conditions. Using the fact that soil moisture varies on a much shorter scale than plant biomass, analytical expressions are derived for the steady state probability distribution of the average plant biomass over a period T, which is expected to be of the order of 2 or 3 months. Analytical expressions are also given for the time-dependent mean and variance of the biomass over a period T. The analysis is based on a simple model linking the daily dynamics of plant growth and soil moisture. The derived analytical expressions reproduce the results obtained from a full simulation of the underlying model very well. The results allow the study of the impact of different climatic scenarios regarding changes in frequency and intensity of rainfall, as well as changes in the mean seasonal temperature, on the expected value and variance of plant biomass throughout time.

39. Puma, M. J., I. Rodriguez-Iturbe, Michael Celia, and A. J. Guswa, 2007: Implications of Rainfall Temporal Resolution for Soil-moisture and Transpiration Modeling. Transport in Porous Media, 68, doi:10.1007/S11242-006-9057-4 37-67
    [ Abstract ]

    Dimensionless groups of parameters characterizing an ecosystem are valuable indicators for the a priori assessment of the effect of rainfall data resolution on predictions of soil moisture and transpiration. Knowledge of these dimensionless groups enables identification of appropriate levels of rainfall data resolution, when using historical rainfall directly or when using it to derive rainfall model parameters for use in models of soil–plant–climate systems.Detailed simulation studies of the soil, plant, and climate systems in Colorado and Texas, highly resolved in time and vertical space, show that historical rainfall data resolved at the daily level allow accurate prediction of soil-moisture and transpiration dynamics for smaller time resolutions. These results support inferences based on the dimensionless groups. Furthermore, no significant improvement in the prediction of soil-moisture and transpiration dynamics is attained, when representing rainfall through a more complex Neyman–Scott model rather than the simple rectangular pulses Poisson model.

40. Celia, Michael, , Jan M. Nordbotten, D. Kavetski, and S. Gasda, 2006: A Risk Assessment Modeling Tool to Quantify Leakage Potential through Wells in Mature Sedimentary Basins. Proceedings of the 8th International Conference on Greenhouse Gas Control Technologies (GHGT-8),
    [ Abstract ]

    The mature sedimentary basins of North America have a long history of oil and gas exploration and production. This has resulted in many wells being drilled, with a substantial number of them now abandoned. Therefore, injection and storage of CO₂ in these basins requires analysis of possible leakage along those wells. A computationally fast semianalytical model of CO₂ injection and potential leakage along wells has been developed, capturing many of the dominant physical characteristics of largescale injection systems. This paper illustrates the capabilities of the model using a case study based on a potential CO₂ sequestration site in Alberta, Canada. The selection of model inputs reflecting the uncertainty in the condition of abandoned wells is considered. The use of the semianalytical model in a probabilistic risk assessment framework is discussed, outlining avenues for systematic regulatory analysis of injection scenarios and systems.

41. Duguid, A., and Michael Celia, May 2006: Geologic CO₂ sequestration in abandoned oil and gas fields and human health risk assessment. Proceedings of the 5th Annual Conference on Carbon Capture and Sequestration,
    [ Abstract ]

    Sequestration in abandoned petroleum fields has the potential to reduce atmospheric emissions of CO₂ if it is adopted on a large scale. However, sequestration sites may pose risks to people who live in their vicinity. CO₂ release from the sequestration formation through abandoned wells to the vadose zone and then from the vadose zone into people’s houses could cause exposure to high levels of CO₂. CO₂ is different from many other chemicals that may be released into the environment because its effects are acute instead of chronic. Existing literature on the health effects of CO₂ in humans was surveyed to establish risk-based screening levels that could be used near a sequestration site. Two potential screening levels were identified: (1) one person in a million becomes dizzy from inhalation of CO₂ in the basement of a house (3.7780% CO₂), and (2) one person in a million loses consciousness from exposure to CO₂ (6.6744% CO₂). A hypothetical risk assessment was conducted using a semianalytical wellfield model developed at Princeton University coupled with analytical models of diffusion through the vadose zone and foundation walls. The assessment assumed that a wellfield in Alberta, Canada, was transformed into a sequestration site with an injection rate of 43,200 t-CO₂/day and that a subdivision has been built near the site. The results showed that CO₂ levels on the site will not reach either of the identified screening levels unless the value used for the exchange rate for air in the houses is very small.

42. Gasda, S., Jan M. Nordbotten, and Michael Celia, June 2006: Significance of Dipping Angle on CO₂ Plume Migration in Deep Saline Aquifers. Proceedings of the XVI Intl Conf on Computational Methods in Water Resources, Copenhagen, http://proceedings.cmwr-xvi.org/getFile.py/access?contribId=63,
    [ Abstract ]

    Recent investigations regarding CO₂ sequestration in deep, saline aquifers have focused on characterization of the injected plume, its migration within the aquifer over time, and possible leakage out of the aquifer. As part of our efforts to understand and quantify leakage potential in CO₂ storage systems, a semi-analytical solution has been developed that describes the plume shape evolution as well the amount of leakage, with a focus on leakage along abandoned wells. The semi-analytical solutions require a number of simplifying assumptions, including a perfectly horizontal aquifer, negligible capillary pressure, and symmetry of the injection plume. Each of these assumptions can be tested systematically through application of more general numerical simulators. In typical sedimentary basins, it is common to have sloping aquifers with a vertical rise of up to 3-4 km over the total horizontal length of the basin (hundreds of kilometers). In this study, we use a general two-phase numerical simulator to assess the limitations of the assumptions required to derive semi-analytical solutions to these systems. In this presentation we will present results from these simulations and discuss their implications regarding the extent to which CO₂ injection systems can be simplified.

43. Haegland, H., H. Dahle, G. T. Eigestad, Jan M. Nordbotten, Michael Celia, and A. Assteerawatt, June 2006: Streamline Methods on Fault Adapted Grids for Risk Assessment of Storage of CO₂ in Geological Formations. Proceedings of the XVI Intl Conf on Computational Methods in Water Resources, Copenhagen, http://proceedings.cmwr-xvi.org/getFile.py/access?contribId=68,
    [ Abstract ]

    Geological storage of CO₂ in possibly fractured and faulted media, involves the risk of leakage. The extent of leakage may be assessed with statistical methods through analysis of simulations of multiple realizations of a stochastic model. Numerical simulation of numerous such realizations typically requires considerable computational cost, motivating the use of fast numerical methods, such as streamline simulation, for screening. Streamline methods have shown to be eective for reservoir characterization and simulation. In this work we will develop methodology which allows for tracing of streamlines in fractured or faulted media. The work is motivated in part by the need to assess potential of geological storage of CO₂ and is also highly relevant for reservoir simulation.

44. Kavetski, D., Jan M. Nordbotten, and Michael Celia, June 2006: Analysis of Potential CO₂ Leakage through Abandoned Wells using a Semi-analytical Model. Proceedings of the XVI Intl Conf on Computational Methods in Water Resources, Copenhagen, http://proceedings.cmwr-xvi.org/getFile.py/access?contribId=1,
    [ Abstract ]

    Potential injection sites for geological CO₂ storage include deep formations in mature sedimentary basins. Many of these basins have a long history of oil and gas exploration and production and the vicinity of the injection site may therefore be perforated by hundreds of wells, potentially penetrating into the injection formation. Since typical injection operations may lead to CO₂ plums extending tens of kilometres from the injection site, geosequestration models must be able to simulate large spatial domains (up to and above 1,000 km²), while resolving the local dynamics in all the wells. Furthermore, many of these wells are abandoned and their locations and hydraulic properties might be uncertain or unknown. Therefore, risk assessment based on Monte Carlo simulations may be necessary to estimate the resulting uncertainty in the leakage. In this paper, we present a semi-analytical model that simulates the evolution of CO₂ plumes and leakage in multiple brine aquifers penetrated by multiple wells over decadal to century time scales. The model equations and state variables are obtained from the self-similarity of the plume shapes and are defined solely at well locations. Since the model does not require domain discretisation in the traditional numerical sense, it is highly computationally efficient, potentially thousands of times faster than existing numerical multiphase simulators. This paper demonstrates the insights gained by applying this model to a potential injection site in the Alberta Basin, Canada, with more than 500 existing wells over a domain of 900 km². Several leakage measures and statistics are presented and discussed.

45. Nordbotten, Jan M., and Michael Celia, June 2006: Analysis of Plume Extent using Analytical Solutions for CO₂ Storage. Proceedings of the XVI Intl Conf on Computational Methods in Water Resources, Copenhagen, http://proceedings.cmwr-xvi.org/getFile.py/access?contribId=50,
    [ Abstract ]

    The evaluation of possible leakage pathways from CO₂ storage operations requires attention to the magnitude, concentration and timescales involved. Herein we discuss a problem related to CO₂ storage, the migration of CO₂ as it is injected. This is accomplished through the application of analytical solutions. In particular, we derive a new analytical insight into the problem of fluid injection into a confined aquifer, which gives us analytically the furthest extent of the injected fluid (CO₂), as well as the extent of the region in which the formation fluid (brine) has evaporated into the injected fluid. We apply these new analytical solutions to a hypothetical injection case based on data from Alberta, Canada, and discuss the results in terms of the surprising variability in observed system responses. We conclude this paper by emphasizing the value of analytical solutions both in semi-analytical frameworks and as benchmarks for numerical simulations.

46. Nordbotten, Jan M., and Michael Celia, 2006: An Improved Analytical Solution for Interface Upconing around a Well. Water Resources Research, 42(W08433), doi:10.1029/2005WR004738
    [ Abstract ]

    Interface models of two-phase flow to a well usually apply vertical equilibrium assumptions to capture the vertical pressure variation in the two fluids. When large gradients in interface height occur, this leads to poor approximations. To deal with largegradient interfaces, a variety of correction factors have been proposed on the basis of single-phase flow analogies. Herein we propose to include the effects of vertical flow directly through a nonlinear vertical pressure variation. This more complex formulation captures significantly more of the flow dynamics in the near-well region.

47. Nordbotten, Jan M., and Michael Celia, 2006: Similarity Solutions for Fluid Injection into Confined Aquifers. Journal of Fluid Mechanics, 561, doi:10.1017/S0022112006000802 307-327
    [ Abstract ]

    Fluid injection into the deep subsurface, such as injection of carbon dioxide (CO₂) into deep saline aquifers, often involves two-fluid flow in confined geological formations. Similarity solutions may be derived for these problems by assuming that a sharp interface separates the two fluids, by imposing a suitable no-flow condition along both the top and bottom boundaries, and by including an explicit solution for the pressure distribution in both fluids. When the injected fluid is less dense and less viscous than the resident fluid, as is the case for CO₂ injection into a resident brine, gravity override produces a fluid flow system that is captured well by the similarity solutions. The similarity solutions may be extended to include slight miscibility between the two fluids, as well as compressibility in both of the fluid phases. The solutions provide the location of the interface between the two fluids, as well as drying fronts that develop within the injected fluid. Applications to cases of supercritical CO₂ injection into deep saline aquifers demonstrate the utility of the solutions, and comparisons to solutions from full numerical simulations show the ability to predict the system behaviour.

48. Nordbotten, Jan M., I. Rodriguez-Iturbe, and Michael Celia, 2006: Non-uniqueness of Evapotranspiration due to Spatial heterogeneity of Plant Species. Proceedings of The Royal Society of London, 462(2072), doi:10.1098/rspa.2005.1641 2359-2371
    [ Abstract ]

    Spatially averaged soil moisture dynamics are studied under seasonally fixed conditions. We consider rainfall as a marked Poisson process, uniformly covering a spatial domain consisting of multiple plant types. Each plant type is considered to have different characteristics in terms of evapotranspiration functions, root-zone depth and rainfall interception. Equations for the evolution of joint probability density functions for individual soil moistures associated with different plant types are developed, and the non-uniqueness of the spatially averaged evapotranspiration function as a function of the average soil moisture is demonstrated and quantified in an example.

49. Peters, Catherine A., J. A. Lewandowski, M. L. Maier, Michael Celia, and , 2006: Mineral Grain Spatial Patterns and Reaction Rate Up-Scaling. Proceedings of the XVI Intl Conf on Computational Methods in Water Resources, Copenhagen, http://esd.lbl.gov/ESD_staff/li_li/lilicmwrxviCAP.pdf,
    [ Abstract ]

    Reactive transport models that describe mineral reactions in porous media rely on laboratory measurements of rate parameters that may fail to represent reactions defined at larger averaging scales. In recently completed work, we used pore-scale network models to investigate the effects of heterogeneities in pore structure and mineral distribution on geochemical reaction rates in porous media. Our findings revealed significant scaling effects from variations in reactive mineral distribution, especially for the highly acidic conditions encountered in geological sequestration of carbon dioxide. In this paper we present preliminary findings from electron scanning BSE maps, to analyze spatial patterns of minerals in sedimentary rocks. Samples include sandstones from the Viking formation in the Alberta basin in western Canada. Image analysis was used to quantify pore space and examine reactive minerals in relation to pore locations. Typically, reactive minerals occur as distinct grains and inclusions, and their percent abundance is larger than the extent of their contact with pore fluids.

50. Celia, Michael, , Jan M. Nordbotten, D. Kavetski, and S. Gasda, May 2005: Modeling Critical Leakage Pathways in a Risk Assessment Framework: Representation of Abandoned Wells. Proceedings of the 4th Annual Conference on Carbon Capture and Sequestration, http://www.netl.doe.gov/publications/proceedings/05/carbon-seq/Tech%20Session%20Paper%20115.pd,
    [ Abstract ]

    In many locations in North America, likely injection sites for CO₂ storage in deep geological formation are located in mature sedimentary basins. These basins have a century-long history of oil and gas exploration and production, which has led to hundreds of thousands of wells (the Alberta Basin) to more than a million wells (Texas) being drilled. The spatial density of these wells is on the order of 0.5 to 5 wells per square kilometer. Therefore, a typical injection will produce a CO₂ plume that intersects hundreds of existing wells, many of which are abandoned and some of which have uncertain or unknown locations. In order to analyze the leakage potential in such situations, computational models must be developed that can cover large spatial areas (of order 1,000 km²) while resolving the local flow dynamics in all of the hundreds of wells. In addition, both the layered structure of the subsurface, and possible leakage along wells and into successive overlying permeable layers in the subsurface, also need to be represented. We have developed a semi-analytical model that can simulate all of these attributes, over decadal to century time scales, while running quickly on a laptop computer. With this tool, risk assessment based on Monte Carlo analysis can be carried out, and a quantitative analysis of leakage potential can be performed.

51. Dahle, H., Michael Celia, and S. M. Hassanizadeh, 2005: Bundle-of-Tubes Model for Calculating Dynamic Effects in the Capillary Pressure – Saturation Relationship. Transport in Porous Media, 58(1-2), doi:10.1007/1-4020-3604-3 5-22
    [ Abstract ]

    Traditional two-phase flow models use an algebraic relationship between capillary pressure and saturation. This relationship is based on measurements made under static conditions. However, this static relationship is then used to model dynamic conditions, and evidence suggests that the assumption of equilibrium between capillary pressure and saturation may not be be justified. Extended capillary pressure–saturation relationships have been proposed that include an additional term accounting for dynamic effects. In the present work we study some of the underlying pore-scale physical mechanisms that give rise to this so-called dynamic effect. The study is carried out with the aid of a simple bundle-of-tubes model wherein the pore space of a porous medium is represented by a set of parallel tubes. We perform virtual two-phase flow experiments in which a wetting fluid is displaced by a non-wetting fluid. The dynamics of fluid–fluid interfaces are taken into account. From these experiments, we extract information about the overall system dynamics, and determine coefficients that are relevant to the dynamic capillary pressure description. We find dynamic coefficients in the range of 10²−10³ kgm⁻¹ s⁻¹, which is in the lower range of experimental observations. We then analyze certain behavior of the system in terms of dimensionless groups, and we observe scale dependency in the dynamic coefficient. Based on these results, we then speculate about possible scale effects and the significance of the dynamic term.

52. Gasda, S., and Michael Celia, 2005: Upscaling Relative Permeabilities in a Structured Porous Medium. Advances in Water Resources, 28(5), doi:10.1016/j.advwatres.2004.11.009
    [ Abstract ]

    Upscaling of multi-phase flow problems for a heterogeneous porous medium requires modification of constitutive functions at the grid-block scale. A particular type of heterogeneity that has important environmental consequences involves thin, continuous streaks of high permeability through lower-permeability background rocks. These streaks, which may correspond to features like abandoned wells in mature sedimentary basins, can become preferential flow paths for an invading fluid. Quantification of flow through these types of heterogeneities in deep, geological formations is necessary for estimates of migration and possible leakage of injected fluids such as hazardous liquid wastes, municipal liquid wastes, and, possibly, carbon dioxide. One of the important constitutive functions for proper estimation of flow through these flow paths is the relative permeability function. In the simple case of a single high-permeability streak in a uniform rock matrix, with both materials having identical (local) relative permeability functions, the upscaled relative permeability must be changed significantly to capture the proper leakage. Standard petroleum reservoir pseudo-functions for relative permeability capture the general features of the upscaled function, but they still produce errors of several hundred percent in the leakage estimation. Detailed three-dimensional numerical simulations and associated upscaled calculations demonstrate the proper form for the upscaled relative permeability, and provide a modified derivation of pseudo-functions to capture the leakage behavior in upscaled models.

53. Nordbotten, Jan M., Michael Celia, , and H. Dahle, 2005: Semi-Analytical Solution for CO₂ Leakage Through An Abandoned Well. Environmental Science and Technology, 39(2), doi:10.1021/es035338i 602-611
    [ Abstract ]

    Capture and subsequent injection of carbon dioxide into deep geological formations is being considered as a means to reduce anthropogenic emissions of CO₂ to the atmosphere. If such a strategy is to be successful, the injected CO₂ must remain within the injection formation for long periods of time, at least several hundred years. Because mature continental sedimentary basins have a century-long history of oil and gas exploration and production, they are characterized by large numbers of existing oil and gas wells. For example, more than 1 million such wells have been drilled in the state of Texas in the United States. These existing wells represent potential leakage pathways for injected CO₂. To analyze leakage potential, modeling tools are needed that predict leakage rates and patterns in systems with injection and potentially leaky wells. A new semianalytical solution framework allows simple and efficient prediction of leakage rates for the case of injection of supercritical CO₂ into a brine-saturated deep aquifer. The solution predicts the extent of the injected CO₂ plume, provides leakage rates through an abandoned well located at an arbitrary distance from the injection well, and estimates the CO₂ plume extent in the overlying aquifer into which the fluid leaks. Comparison to results from a numerical multiphase flow simulator show excellent agreement. Example calculations show the importance of outer boundary conditions, the influence of both density and viscosity contrasts in the resulting solutions, and the potential importance of local upconing around the leaky well. While several important limiting assumptions are required, the new semianalytical solution provides a simple and efficient procedure for estimation of CO₂ leakage for problems involving one injection well, one leaky well, and multiple aquifers separated by impermeable aquitards.

54. Nordbotten, Jan M., Michael Celia, and , 2005: Injection and Storage of CO₂ in Deep Saline Aquifers: Analytical Solution for CO₂ Plume Evolution during Injection. Transport in Porous Media, 58(3), doi:10.1007/s11242-004-0670-9 339-360
    [ Abstract ]

    Injection of fluids into deep saline aquifers is practiced in several industrial activities, and is being considered as part of a possible mitigation strategy to reduce anthropogenic emissions of carbon dioxide into the atmosphere. Injection of CO₂ into deep saline aquifers involves CO₂ as a supercritical fluid that is less dense and less viscous than the resident formation water. These fluid properties lead to gravity override and possible viscous fingering. With relatively mild assumptions regarding fluid properties and displacement patterns, an analytical solution may be derived to describe the space–time evolution of the CO₂ plume. The solution uses arguments of energy minimization, and reduces to a simple radial form of the Buckley–Leverett solution for conditions of viscous domination. In order to test the applicability of the analytical solution to the CO₂ injection problem, we consider a wide range of subsurface conditions, characteristic of sedimentary basins around the world, that are expected to apply to possible CO₂ injection scenarios. For comparison, we run numerical simulations with an industry standard simulator, and show that the new analytical solution matches a full numerical solution for the entire range of CO₂ injection scenarios considered. The analytical solution provides a tool to estimate practical quantities associated with CO₂ injection, including maximum spatial extent of a plume and the shape of the overriding less-dense CO₂ front.

55. Puma, M. J., Michael Celia, I. Rodriguez-Iturbe, and A. J. Guswa, 2005: Functional Relationship to Describe Temporal Statistics of Soil Moisture averaged over Different Depths. Advances in Water Resources, 28, doi:10.1016/j.advwatres.2004.08.015 553-566
    [ Abstract ]

    Detailed simulation studies, highly resolved in space and time, show that a physical relationship exists among instantaneous soilmoisture values integrated over different soil depths. This dynamic relationship evolves in time as a function of the hydrologic inputs and soil and vegetation characteristics. When depth-averaged soil moisture is sampled at a low temporal frequency, the structure of the relationship breaks down and becomes undetectable. Statistical measures can overcome the limitation of sampling frequency, and predictions of mean and variance for soil moisture can be defined over any soil averaging depth d. For a water-limited ecosystem, a detailed simulation model is used to compute the mean and variance of soil moisture for different averaging depths over a number of growing seasons. We present a framework that predicts the mean of soil moisture as a function of averaging depth given soil moisture over a shallow d and the average daily rainfall reaching the soil.

56. Scherer, George, Michael Celia, Jean Hervé Prévost, , R. Bruant, A. Duguid, R. Fuller, S. Gasda, M. Radonjic, and W. Vichit-Vadakan, 2005: Leakage of CO₂ through Abandoned Wells: Role of Corrosion of Cement. The CO₂ Capture and Storage Project (CCP), Volume II, Chapter 10, 823-844
    [ Abstract ]

    The potential leakage of CO₂ from a geological storage site through existing wells represents a major concern. An analysis of well distribution in the Viking Formation in the Alberta basin, a mature sedimentary basin representative of North American basins, shows that a CO₂ plume and/or acidified brine may encounter up to several hundred wells. A review of the literature indicates that cement is not resistant to attack by acid, but little work has been reported for temperatures and pressures comparable to storage conditions. Therefore, an experimental program has been undertaken to determine the rate of corrosion and the changes in properties of oil well cements exposed to carbonated brine. Preliminary results indicate a very high rate of attack, so it is essential to have accurate models of the composition and pH of the brine, and the time that it will remain in contact with cement in abandoned wells. A model has been developed that incorporates a flash calculation of the phase distribution, along with analysis of the fluxes and pressure of the liquid, solid and vapor phases. A sample calculation indicates that wells surrounding the injection site may be in contact with the acidified brine for years.

57. Altevogt, A., and Michael Celia, 2004: Numerical Modeling of Carbon Dioxide in Unsaturated Soils Due to Deep Subsurface Leakage In , 40(W03509), doi:10.1029/2003WR002848
    [ Abstract ]

    A two-dimensional numerical model was utilized to explore the flux mechanisms governing CO₂ transport in the vadose zone. The simulations were set up to approximately correspond to a site of natural CO₂ leakage at Mammoth Mountain, California. The mass fraction gradient driving force, responsible for diffusive and slip fluxes, was determined to lead to less plume spreading than advection alone. Density-driven flow of CO₂ led to significantly greater spreading of the plume and greater storage of CO₂ within the vadose zone than if density contrasts were not accounted for. Exposure assessment simulations indicate that for the conditions of interest there may be no physically realistic domain that would lead to CO₂ levels below the criteria for human health impacts (sub 10%) in surface soils for the leakage rate present at Mammoth Mountain.

58. Celia, Michael, , Jan M. Nordbotten, S. Gasda, and H. Dahle, September 2004: Quantitative Estimation of CO₂ Leakage from Geological Storage: Analytical Models, Numerical Models, and Data Needs. Proceedings of the 7th International Conference on Greenhouse Gas Control Technologies, (GHGT-7), http://uregina.ca/ghgt7/PDF/papers/peer/228.pdf,
    [ Abstract ]

    Geological storage of CO₂ in mature sedimentary basins of North America requires special consideration of the large number of existing wells. Those wells represent potential leakage pathways for the stored CO₂, and must be analyzed in the context of an overall environmental risk assessment. Analysis of well patterns in the Alberta basin, Canada, indicates that injected CO₂ plumes are expected to contact from several tens to several hundreds of existing wells, depending on the local density of wells in the vicinity of the injection. Quantitative analysis of the impact of these wells requires an extensive data collection effort, analysis of materials used in well construction and abandonment, and different levels of computational modeling to ascertain the risk associated with these wells. New analytical solutions provide a promising avenue for leakage analysis at the large scale. Results from these models show accuracy comparable to more complex numerical simulators at a small fraction of the computational time. This allows many simulations to be run so that different parameters values can be explored. These large-scale models need to be coupled to smaller-scale detailed models of material behavior within the leaky well to provide a complete analysis of the problem.

59. Celia, Michael, H. Dahle, and S. M. Hassanizadeh, June 2004: Dynamic Effects in Capillary Pressure Relationships for Two-phase Flow in Porous Media: Insights from Bundle-of-Tubes Models and their Implications. Proceedings of the Computational Methods in Water Resources 2004 Conference, 1, 127-138
    [ Abstract ]

    Traditional multi-phase flow models use an algebraic relationship between capillary pressure and saturation. This relationship is based on measurements made under static conditions. However, this static relationship is then used to model dynamic conditions and evidence suggest that the assumption of equilibrium between capillary pressure and saturation may not be justified. Extended capillary pressure-saturation relationships have been proposed that include an additional term accounting for dynamic effects. In the present work, we study the underlying pore-scale physical mechanisms that give rise to this so-called dynamic effect. The study is carried out with the aid of a simple bundle-of-tubes model wherein the pore space of a porous medium is represented by a set of parallel tubes. We perform virtual two-phase flow experiments in which a wetting fluid is displaced by a non-wetting fluid. The dynamics of fluid-fluid interfaces are taken into account, and we consider systems in which viscosity differences influence the displacement process. From these experiments, we extract information about overall system dynamics, determine large-scale effects that are associated with viscosity differences between the two fluids. Based on these results, we then speculate about possible scale effects and the significance of the dynamic term.

60. Duguid, A., M. Radonjic, R. Bruant, T. Mandecki, George Scherer, and Michael Celia, 2004: The Effect of CO₂ Sequestration on Oil Well Cements. Proceedings of the 7th International Conference on Greenhouse Gas Control Technologies, (GHGT-7), http://www.princeton.edu/~cmi/research/Vancouver04/GHGT7Duguid.pdf,
    [ Abstract ]

    Experiments were conducted to examine the effects of CO₂ sequestration conditions on cements used to construct and abandon oil and gas wells. The results showed that significant damage, complete loss of the calcium hydroxide phase, can take place over a time span as short as seven days.

61. Gasda, S., , and Michael Celia, 2004: Spatial characterization of the location of potentially leaky wells penetrating a deep saline aquifer in a mature sedimentary basin. Environmental Geology, 46(6-7), doi:10.1007/s00254-004-1073-5 707-720
    [ Abstract ]

    This work was motivated by considerations of potential leakage pathways for CO₂ injected into deep geological formations for the purpose of carbon sequestration. Because existing wells represent a potentially important leakage pathway, a spatial analysis of wells that penetrate a deep aquifer in the Alberta Basin was performed and various statistical measures to quantify the spatial distribution of these wells were presented. The data indicate spatial clustering of wells, due to oil and gas production activities. The data also indicate that the number of wells that could be impacted by CO₂ injection, as defined by the spread of an injected CO₂ plume, varies from several hundred in high welldensity areas to about 20 in low-density areas. These results may be applied to other mature continental sedimentary basins in North America and elsewhere, where detailed information on well location and status may not be available.

62. Guswa, A. J., Michael Celia, and I. Rodriguez-Iturbe, 2004: Effect of Vertical Resolution on Predictions of Transpiration in Water-limited Eosystems. Advances in Water Resources, 27, doi:10.1016/j.advwatres.2004.03.001 467-480
    [ Abstract ]

    Water-limited ecosystems are characterized by precipitation with low annual totals and significant temporal variability, transpiration that is limited by soil-moisture availability, and infiltration events that may only partially rewet the vegetation root zone. Average transpiration in such environments is controlled by precipitation, and accurate predictions of vegetation health require adequate representation of temporal variation in the timing and intensity of plant uptake. Complexities introduced by variability in depth of infiltration, distribution of roots, and a plant’s ability to compensate for spatially heterogeneous soil moisture suggest a minimum vertical resolution required for satisfactory representation of plant behavior. To explore the effect of vertical resolution on predictions of transpiration, we conduct a series of numerical experiments, comparing the results from models of varying resolution for a range of plant and climate conditions. From temporal and spatial scales of the underlying processes and desired output, we develop dimensionless parameters that indicate the adequacy of a finite-resolution model with respect to reproducing characteristics of plant transpiration over multiple growing seasons. These parameters may be used to determine the spatial resolution required to predict vegetation health in water-limited ecosystems.

63. Nordbotten, Jan M., Michael Celia, and , 2004: Analytical Solutions for Leakage Rates through Abandoned Wells. Water Resources Research, 40(W04204), doi:10.1029/2003WR002997
    [ Abstract ]

    Disposal of waste fluids via injection into deep saline aquifers is practiced in a variety of industries. Injection takes place in sedimentary basins that often have a history of oil and gas exploration and production, which means that wells other than those used for waste disposal may exist in the vicinity of the injection site. These existing wells provide possible pathways for leakage of waste fluids toward the shallow subsurface and the land surface. For single-phase flows of liquids with essentially constant properties, the equations governing the system are linear, and solutions may be written using the superposition principle. Because leakage through existing wells produces a time-varying flux rate, the solution of the governing equations involves convolution integrals. Previous solutions have addressed the problem of one injection well, one existing (passive) well, and a simple geometry of two aquifers separated by an aquitard by use of Laplace transforms. Even for this simple case, inversion of the transform is difficult. Solutions involving more than one passive well have not been developed. Nor have solutions been developed for more than two aquifers and one aquitard. Realistic injection cases often involve layered systems with multiple aquifers and aquitards, as well as multiple passive wells, sometimes numbering in the hundreds. Solutions for the general case of multiple aquifers and wells may be developed through introduction of approximations to the well function and appropriate simplification of the convolution integral. Such a solution is computationally simple. Comparison to solutions using the full (Laplace transform) solution indicates that the new solution procedure produces excellent results. Application of the new solution to a case of multiple passive wells shows that the cumulative leakage flux in the passive wells is not a simple sum of the single-well case, owing to leakage induced drawdown around the passive wells. In addition, application to the case of multiple aquifers and aquitards demonstrates the importance of leakage into intervening aquifers as a mechanism to mitigate leakage into shallow zones, a process referred to as the ‘‘elevator model.’’ The new analytical solution provides a tool to analyze practical injection problems and forms a foundation on which more complex solutions, such as those involving injection of a nonaqueous fluid into a deep brine formation, may be based.

64. Zaman, M.S.U., L. A. Ferrand, and Michael Celia, 2004: Type Curves and Effective Parameters for Unsaturated Flow Systems with Structured Heterogeneities. Advances in Water Resources, 27, doi:10.1016/j.advwatres.2004.02.005 399-410
    [ Abstract ]

    Unsaturated flow simulation requires identification of soil parameters at length scales that usually subsume smaller-scale heterogeneities. The standard two-parameter constitutive model performs well for predictions of moisture plume evolution in homogeneous soils but may not be as successful in capturing the flow and dispersion of soil water in heterogeneous domains. The first step in evaluating the limitations of this model is to develop a clear understanding of the effects of the constitutive parameters on moisture plume evolution. One approach is to define type curves derived from multiple homogeneous simulations. These type curves are based on bulk measures of system behavior, suitable for comparison to responses of heterogeneous systems. Simulation results for a range of heterogeneities defined on a specific test system indicate that some heterogeneity patterns allow definition of effective parameters for use in the two-parameter constitutive model, while others do not. For those heterogeneity patterns, the mathematical structure of the governing equations applied at the large scale must have a form that is different from the equation that underlies the type curves.

65. Celia, Michael, and , 2003: Geological Sequestration of CO₂: Is Leakage Unavoidable and Acceptable? Proceedings of the 6th International Conference on Greenhouse Gas Control Technologies (GHGT-6), http://www.princeton.edu/~cmi/research/kyoto02/celia&bachu.kyoto%2002.pdf, 1, 477-482
    [ Abstract ]

    Geological sequestration of CO₂ requires careful risk analysis to avoid unintended consequences of the subsurface injection. One potentially serious problem associated with injection into mature sedimentary basins is the possible leakage of injected CO₂ through or along existing wells. Over long time scales, these wells may serve as short-circuit pathways for leakage, with possible contamination of shallow subsurface zones, and ultimate leakage back into the atmosphere. Transport models for leakage analysis and overall risk assessment must include proper representation of the effects of existing wells. A multi-scale framework offers a guide for inclusion of existing wells into practical simulators.

66. Nordhaug, H., Michael Celia, and H. Dahle, 2003: A Pore Network Model for Calculation of Interfacial Velocities. Advances in Water Resources, 26(10), doi:10.1016/S0309-1708(03)00100-3 1061-1074
    [ Abstract ]

    Two-phase flow in porous media is characterized by fluid–fluid interfaces that separate fluid phases at the pore scale. These interfaces support pressure differences between phases, and their dynamics lead to changes in phase saturation within the porous medium. Dynamic pore-scale network models mathematically track the dynamic position of each fluid–fluid interface through a pore network, based on imposed boundary conditions, fluid and solid properties, and geometric characteristics of the network. Because these models produce a detailed description of both phase and interface dynamics, results from these models can be volume averaged to provide values for many upscaled variables. These include traditional variables such as saturation and macroscopic capillary pressure, as well as non-traditional variables such as amount of interfacial area in the averaging volume. With appropriate geometric definitions in the dynamic pore-scale model, a new algorithm may be included in the pore-scale network model to calculate a new variable: average interfacial velocity. This algorithm uses local information in any pore that contains a fluid–fluid interface to estimate the velocity of that interface over a time step. Summation over all interfaces in the network provides a measure of average velocity. Computations for dynamic drainage experiments indicate that this average interfacial velocity is well defined and exhibits distinct behavior for stable and unstable displacements. Comparison of calculated interfacial velocities to a theoretical conjecture on the functional dependence of this macroscopic variable demonstrates another important use of pore-scale model, namely testing of new theories involving non-traditional variables.

67. Ataie-Ashtiani, B., S. M. Hassanizadeh, and Michael Celia, 2002: Effects of Heterogeneities on Capillary Pressure - Saturation - Relative Permeability Relationships. Journal of Contaminant Hydrology, 56(3-4), doi:10.1016/S0169-7722(01)00208-X 175-192
    [ Abstract ]

    In theories of multiphase flow through porous media, capillary pressure–saturation and relative permeability–saturation curves are assumed to be intrinsic properties of the medium. Moreover, relative permeability is assumed to be a scalar property. However, numerous theoretical and experimental works have shown that these basic assumptions may not be valid. For example, relative permeability is known to be affected by the flow velocity (or pressure gradient) at which the measurements are carried out. In this article, it is suggested that the nonuniqueness of capillary pressure–relative permeability–saturation relationships is due to the presence of microheterogeneities within a laboratory sample. In order to investigate this hypothesis, a large number of ‘‘numerical experiments’’ are carried out. A numerical multiphase flow model is used to simulate the procedures that are commonly used in the laboratory for the measurement of capillary pressure and relative permeability curves. The dimensions of the simulation domain are similar to those of a typical laboratory sample (a few centimeters in each direction). Various combinations of boundary conditions and soil heterogeneity are simulated and average capillary pressure, saturation, and relative permeability for the ‘‘soil sample’’ are obtained. It is found that the irreducible water saturation is a function of the capillary number; the smaller the capillary number, the larger the irreducible water saturation. Both drainage and imbibition capillary pressure curves are found to be strongly affected by heterogeneities and boundary conditions. Relative permeability is also found to be affected by the boundary conditions; this is especially true about the nonaqueous phase permeability. Our results reveal that there is much need for laboratory experiments aimed at investigating the interplay of boundary conditions and microheterogeneities and their effect on capillary pressure and relative permeability.

68. Binning, P. J., and Michael Celia, 2002: A Forward Particle Tracking Eulerian Lagrangian Localized Adjoint Method for Solution of the Contaminant Transport Equation in Three Dimensions. Advances in Water Resources, 25(2), doi:10.1016/S0309-1708(01)00051-3 147-157
    [ Abstract ]

    The contaminant transport equation is solved in three dimensions using the Eulerian–Lagrangian Localized Adjoint Method (ELLAM). Trilinear and finite volume test functions defined by the characteristics of the governing equation are employed and compared. Integrations are simplified by forward tracking of integration points along the characteristics. The resulting equations are solved using a preconditioned conjugate gradient method. The algorithm is coupled to a block-centered finite difference approximation of the groundwater flow equation similar to that used in the popular MODFLOW code. The ELLAM is tested by comparison with 1D and 3D analytic solutions. The method is then applied with random, spatially correlated hydraulic conductivities in a simulation of a tracer experiment performed on Cape Cod, Massachusetts. The linear test function ELLAM was found to perform better than the finite volume ELLAM. Both ELLAM formulations were found to be robust, computationally efficient and relatively straightforward to implement. When compared to traditional particle tracking and characteristics codes commonly used with MODFLOW, the ELLAM retains the computational advantages of traditional characteristic methods with the added advantage of good mass conservation.

69. Bruant, R., A. J. Guswa, Michael Celia, and Catherine A. Peters, 2002: Safe Storage of Carbon Dioxide in Deep Saline Aquifers. Environmental Science and Technology, 36(11), doi:10.1021/es0223325 240A-245A
    [ Abstract ]

    Over the past 420,000 years, global average atmospheric CO₂ concentrations have fluctuated narrowly between 180 and 280 parts per million by volume (ppmv), but since the Industrial Revolution, CO₂ concentrations have increased to ~370 ppmv. This increase is believed to be contributing to risingmean global temperatures (1, 2). Average annual global anthropogenic CO₂ emissions during the 1990s were ~27 GtCO₂/yr (1 GtCO₂ = 109 metric tons of CO₂ = 10¹² kg of CO₂ = 0.27 GtC). The Intergovernmental Panel on Climate Change estimates that under a “business-as- usual” energy scenario, global emissions will reach ~77 GtCO₂/yr by 2100, and the average atmospheric CO₂ concentration will reach ~750 ppmv (2). To stabilize atmospheric CO₂ concentrations at 550 ppmv, which is approximately twice preindustrial concentrations, global emissions must be continuously reduced so that by 2050, global emissions are 15 GtCO₂/yr less than the business-as-usual projection, and by 2100, emissions are 50 GtCO₂/yr less (2, 3).

70. Celia, Michael, 2002: How Hydrogeology Can Save the World (An Editorial). Ground Water, 40(2), doi:10.1111/j.1745-6584.2002.tb02495.x 113
    [ Abstract ]

    I have read with interest recent editorials and articles in Ground Water discussing the future of hydrogeological research. Is it the beginning of the end, or is it simply time for a shift in focus? What can we infer from citation trends of from a move toward more practical applications? These questions lead to interesting debates. Definitive resolution of these debates is unlikely. However, there does appear to be a sense in the community that new directions, with applications that address truly important problems, are needed to revitalize hydrogeological research.
    While ground water contamination was a highly visible and important environmental concern for much of the 1980s and early 1990s, this concern has faded considerably, and the current problems that dominate the broad environmental landscape are carbon dioxide, greenhouse gas emissions, and global warming. A recent editorial in Ground Water discussed how ground water resources, and the field of hydrogeology, might be impacted by global warming. While this type of passive role needs to be planned for, I submit that hyddrogeology may have a central, active role to play in solving the CO₂ problem, and that CO₂ sequestration represents a new and potentially important problem area for hydrogeologic research.

71. Celia, Michael, and A. J. Guswa, 2002: Hysteresis and Upscaling in Two-Phase Flow through Porous Media. Proceedings of the Joint Summer Research Conference on Fluid Flow & Transport in Porous Media, American Mathematical Society, 295, 93-104
    [ Abstract ]

    Modeling two-phase flow in porous media is difficult because of the strong nonlinearities in the governing equations and the presence of hysteresis in some of the constitutive relationships. Further difficulties are introduced by the need to upscale many of the parameters to achieve scale consistency in a model. Computational studies provide insights into the behavior of these systems under the process of upscaling. In this work, we focus on behavior of water flow in unsaturated soils, including the dominant effects of evaporation and transpiration. Numerical simulations are used to upscale constitutive relationships. We demonstrate how upscaling can lead to hysteresis in constitutive relationships for cases when no small-scale hysteresis is included. This shows how upscaling and hysteresis are linked for nonlinear systems such as two-phase flow in porous media.

72. Dahle, H., Michael Celia, S. M. Hassanizadeh, and K. H. Karlsen, 2002: A Total Pressure - Saturation Formulation of Two-Phase Flow incorporating Dynamic Effects in the Capillary Pressure - Saturation Relationship. Proceedings of the XIV Intl Conf on Computational Methods in Water Resources, Delft, 1067-1074
    [ Abstract ]

    New theories suggest hat the relationship between capillary pressure and saturation should be enhanced by a dynamic term that is proportional to the time rate of change of saturation. This so-called dynamic capillary pressure formulation is supported by laboratory experiments, and can be included in various forms of the governing equations for the two-phase flow in porous media. An extended model of two-phase flow in porous media may be developed based on fractional flow curves and a total pressure - saturation description that includes the dynamic capillary pressure terms. A dimensionless form of the resulting equation set provides an ideal tool to study the relative importance of the dynamic capillary pressure effect. This equation provides rich set of mathematical research questions, and numerical solutions to the equation provide insights into the behavior of two-phase immiscible flow. For typical two-phase flow systems, dynamic capillary pressure acts to retard infiltration fronts, with responses dependent on system parameters including boundary conditions.

73. Gielen, T., S. M. Hassanizadeh, Michael Celia, and H. Dahle, 2002: Study of Pc-Sw Relationship using a Dynamic Pore-Scale Network Model. Proceedings of the XIV Intl Conf on Computational Methods in Water Resources, Delft, 1099-1100
    [ Abstract ]

    Current theories of multiphase flow rely on capillary pressuer and saturation relationships that are commonly measured under static conditions. Recently, new multiphase flow theories have been proposed that include an extended capillary pressure-saturation relationship that is valid under dynamic conditions. In this relationship, the difference between the two fluid pressures is called dynamic capillary pressure, and is assumed to be a function of the saturation and its time rate of change.
    In this work, we test this relationship using a pore-scale network model. Our model consists of a three-dimensional network of tubes (pore throats) connected to each other by pore bodies. Both pore bodies and pore throats are assumed to have square cross sections. We perform numerical experiments wherein typical experimental procedures for both static and dynamic measurements of capillary pressure-saturation curves are simulated. We determine the value of the dynamic coefficient τ for our network model.

74. Guswa, A. J., Michael Celia, and I. Rodriguez-Iturbe, 2002: Models of Soil-Moisture Dynamics in Ecohydrology: A Comparative Study. Water Resources Research, 38(9), doi:10.1029/2001WR000826
    [ Abstract ]

    An accurate description of plant ecology requires an understanding of the interplay between precipitation, infiltration, and evapotranspiration. A simple model for soil moisture dynamics, which does not resolve spatial variations in saturation, facilitates analytical expressions of soil and plant behavior as functions of climate, soil, and vegetation characteristics. Proper application of such a model requires knowledge of the conditions under which the underlying simplifications are appropriate. To address this issue, we compare predictions of evapotranspiration and root zone saturation over a growing season from a simple bucket-filling model to those from a more complex, vertically resolved model. Dimensionless groups of key parameters measure the quality of the match between the models. For a climate, soil, and woody plant characteristic of an African savanna the predictions of the two models are quite similar if the plant can extract water from locally wet regions to make up for roots in dry portions of the soil column; if not, the match is poor.

75. Hassanizadeh, S. M., Michael Celia, and H. Dahle, 2002: Dynamic Effects in the Capillary Pressure - Saturation Relationship and their Impacts on Unsaturated Flow. Vadose Zone Hydrology, 1, 38-57
    [ Abstract ]

    Capillary pressure plays a central role in the description of water flow in unsaturated soils. While capillarity is ubiquitous in unsaturated analyses, the theoretical basis and practical implications of capillarity in soils remain poorly understood. In most traditional treatments of capillary pressure, it is defined as the difference between pressures of phases, in this case air and water, and is assumed to be a function of saturation. Recent theories have indicated that capillary pressure should be given a more general thermodynamic definition, and its functional dependence should be generalized to include dynamic effects. Experimental evidence has slowly accumulated in the past decades to support a more general description of capillary pressure that includes dynamic effects. A review of these experiments shows that the coefficient arising in the theoretical analysis can be estimated from the reported data. The calculated values range from 104 to 107 kg (m s)-1. In addition, recently developed pore-scale models that simulate interface dynamics within a network of pores can also be used to estimate the appropriate dynamic coefficients. Analyses of experiments reported in the literature, and of simulations based on pore- scale models, indicate a range of dynamic coefficients that spans about three orders of magnitude. To examine whether these coefficients have any practical effects on larger scale problems, continuum-scale simulators may be constructed in which the dynamic effects are included. These simulators may then be run to determine the range of coefficients for which discernable effects occur. Results from such simulations indicate that measured values of dynamic coefficients are within one order of magnitude of those values that produce significant effects in field simulations. This indicates that dynamic effects may be important for some field situations, and numerical simulators for unsaturated flow should generally include the additional term(s) associated with dynamic capillary pressure.

76. Held, R. J., W. Kinzelbach, and Michael Celia, 2002: Characterization of Stable and Unstable Flow Regimes via Dynamic Pore-Scale Network Simulation. Proceedings of the XIV Intl Conf on Computational Methods in Water Resources, Delft, 1059-1066
    [ Abstract ]

    Dynamic network models for two-phase flow in porous media describe the physics of fluid motion and propagation of interfaces. The resolution is at the scale of single pores. A network structure representation in the model allows for simulation of macroscopic flow phenomena under variable influence of gravitational, viscous and capillary forces. Macroscopic averages such as specific interfacial areas and macroscopic interfacial velocities are investigated and effective interfacial tensions according to Chuoke et al. may be computed from the simulations. Quantification of such variables affords us the exploration of a system characteristic length scale for displacement stability. An example of viscous instability is presented.

77. Russell, T. F., and Michael Celia, 2002: An Overview of Research on Eulerian-Lagrangian Localized Adjoint Methods (ELLAM). Advances in Water Resources, 25(8-12), doi:10.1016/S0309-1708(02)00104-5 1215-1231
    [ Abstract ]

    For problems of convection–diffusion type,Eulerian–Lagrangian localized adjoint methods provide a methodology that maintains the accuracy and efficiency of Eulerian–Lagrangian methods,while also conserving mass and systematically treating any type of boundary condition. In groundwater hydrology,this framework is useful for solute transport,as well as vadose-zone transport,multiphase transport,and reactive flows. The formulation was originated around 1990 by the authors,Herrera and Ewing,in a paper that appeared in Advances in Water Resources [Adv. Water Resour. 13 (1990) 187]. This paper reviews the progress in the development,analysis,and application of these methods since 1990,and suggests topics for future work.

78. Ataie-Ashtiani, B., S. M. Hassanizadeh, M. Oostrom, Michael Celia, and M. D. White, 2001: Effective Parameters for Two-Phase Flow in a Porous Medium with Periodic Heterogeneities. Journal of Contaminant Hydrology, 49(1-2), doi:10.1016/S0169-7722(00)00190-X 87-109
    [ Abstract ]

    Computational simulations of two-phase flow in porous media are used to investigate the feasibility of replacing a porous medium containing heterogeneities with an equivalent homogeneous medium. Simulations are performed for the case of infiltration of a dense nonaqueous phase liquid (DNAPL). in a water-saturated, heterogeneous porous medium. For two specific porous media, with periodic and rather simple heterogeneity patterns, the existence of a representative elementary volume (REV). is studied. Upscaled intrinsic permeabilities and upscaled nonlinear constitutive relationships for two-phase flow systems are numerically calculated and the effects of heterogeneities are evaluated. Upscaled capillary pressure–saturation curves for drainage are found to be distinctly different from the lower-scale curves for individual regions of heterogeneity. Irreducible water saturation for the homogenized medium is found to be much larger than the corresponding lower-scale values. Numerical simulations for both heterogeneous and homogeneous representations of the considered porous media are carried out. Although the homogenized model simulates the spreading behavior of DNAPL reasonably well, it still fails to match completely the results form the heterogeneous simulations. This seems to be due, in part, to the nonlinearities inherent to multiphase flow systems. Although we have focused on a periodic heterogeneous medium in this study, our methodology is applicable to other forms of heterogeneous media. In particular, the procedure for identification of a REV, and associated upscaled constitutive relations, can be used for randomly heterogeneous or layered media as well.

79. Bruant, R., R. J. Held, Catherine A. Peters, and Michael Celia, 2001: Pore Scale Network Simulation of Single and Multiple Component Non-Aqueous Phase Luquid (NAPL) Dissolution. American Geophysical Union,
    [ Abstract ]

    A computational three-dimensional pore-scale network model was used to quantify residual single- and multi-component non-aqueous phase liquid (NAPL) dissolution driven by aqueous-phase advection. The pore network was discretized into spherical pore bodies and biconical pore throats to represent the effective void space and void distribution of a fine-grained Ottawa sand. Fluid saturations, positions, and interfacial areas, in addition to aqueous-phase flow, were established by externally applied pressure gradients. Mass transfer from the NAPL to the aqueous phase was computed as a local flux across each interface using a stagnant boundary layer Fickian diffusion model. Subsequent mass transport in the aqueous phase was simulated by a volume-conserving characteristic method along streamlines. The model dynamically calculated interface retraction resulting from mass transfer between the non-aqueous and aqueous phases, and concurrently tracked physical changes in NAPL saturation, NAPL composition, and interfacial geometry. The model avoids scale inconsistencies, allowing pore-scale through continuum-scale description of NAPL dissolution. In this presentation, results from NAPL dissolution simulations will be compared (as a function of saturation and location) to laboratory experiments and implications for up-scaling mass transfer coefficients will be discussed. Dependence of multi-component NAPL composition on mass transfer phenomena and differences between single- and multi-component systems also will be highlighted.

80. Held, R. J., and Michael Celia, 2001: Pore-Scale Modeling Extension of Constitutive Relationships in the Range of Residual Saturations. Water Resources Research, http://www.agu.org/journals/wr/v037/i001/2000WR900234/2000WR900234.pdf, 37(1), 165-170
    [ Abstract ]

    A pore network model is used to describe constitutive relationships between saturation, capillary pressure, relative permeabilities, and interfacial areas over the full range of saturations. We focus on residual nonwetting-phase saturations, that is, the saturation range between main drainage and primary drainage. Interphase mass transfer and dynamic miscible transport are modeled to generate the range of saturations over which the extended constitutive relationships are calculated. To accommodate all saturations in a consistent way, we define capillary pressure as the areal average of local capillary pressures associated with each fluid-fluid interface. The extended constitutive relationships provide input for continuum-scale equations.

81. Held, R. J., and Michael Celia, 2001: Modeling Support of Functional Relationships between Capillary Pressure, Saturation, Interfacial Area, and Common Lines. Advances in Water Resources, 24(3-4), doi:10.1016/S0309-1708(00)00060-9 325-343
    [ Abstract ]

    Computational pore-scale network models describe two-phase porous media flow systems by resolving individual interfaces at the pore scale, and tracking these interfaces through the pore network. Coupled with volume averaging techniques, these models can reproduce relationships between measured variables like capillary pressure, saturation, and relative permeability. In addition, these models allow nontraditional porous media variables to be quantifed, such as interfacial areas and common line lengths. They also allow explorations of possible relationships between these variables, as well as testing of new theoretical conjectures. Herein we compute relationships between capillary pressure, saturation, interfacial areas, and common line lengths using a pore-scale network model. We then consider a conjecture that definition of an extended constitutive relationship between capillary pressure, saturation, and interfacial area eliminates hysteresis between drainage and imbibition; such hysteresis is commonly seen in the traditional relationship between capillary pressure and saturation. For the sample pore network under consideration, we find that hysteresis can essentially be eliminated using a specific choice of displacement rules; these rules are within the range of experimental observations for interface displacements and therefore are considered to be physically plausible. We find that macroscopic measures of common line lengths behave similarly to fluid-fluid interfacial areas, although the functional dependencies on capillary pressure and saturation differ to some extent.

Direct link to page: http://cmi.princeton.edu/bibliography/results.php?author=3468