Geologic Sequestration of Carbon
United States
Environmental Protection
Agency
Underground Injection Control (UIC)
Program Class VI Well Area of Review
Evaluation and Corrective Action
Guidance
Office of Water (4606M) EPA 816-R-13-005 May 2013
-------
Disclaimer
The Federal Requirements under the Underground Injection Control Program for Carbon
Dioxide Geologic Sequestration Wells (75 FR 77230, December 10, 2010), known as the Class
VI Rule, establishes a new class of injection well (Class VI).
The Safe Drinking Water Act (SDWA) provisions and U.S. Environmental Protection Agency
(EPA) regulations cited in this document contain legally-binding requirements. In several
chapters this guidance document makes suggestions and offers alternatives that go beyond the
minimum requirements indicated by the Class VI Rule. This is intended to provide information
and suggestions that may be helpful for implementation efforts. Such suggestions are prefaced by
"may" or "should" and are to be considered advisory. They are not required elements of the rule.
Therefore, this document does not substitute for those provisions or regulations, nor is it a
regulation itself, so it does not impose legally-binding requirements on EPA, states, or the
regulated community. The recommendations herein may not be applicable to each and every
situation.
EPA and state decision makers retain the discretion to adopt approaches on a case-by-case basis
that differ from this guidance where appropriate. Any decisions regarding a particular facility
will be made based on the applicable statutes and regulations. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use. EPA is taking
an adaptive rulemaking approach to regulating Class VI injection wells, and the agency will
continue to evaluate ongoing research and demonstration projects and gather other relevant
information as needed to refine the rule. Consequently, this guidance may change in the future
without a formal notice and comment period.
While EPA has made every effort to ensure the accuracy of the discussion in this document, the
obligations of the regulated community are determined by statutes, regulations or other legally
binding requirements. In the event of a conflict between the discussion in this document and any
statute or regulation, this document would not be controlling.
Note that this document only addresses issues covered by EPA's authorities under the SDWA.
Other EPA authorities, such as Clean Air Act (CAA) requirements to report carbon dioxide
injection activities under the Greenhouse Gas Mandatory Reporting Rule (GHG MRR), are not
within the scope of this document.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
-------
Executive Summary
EP'A''s Federal Requirements Under the Underground Injection Control (UIC) Program for
Carbon Dioxide Geologic Sequestration Wells has been codified in the U.S. Code of Federal
Regulations (40 CFR 146.81 et seq.), and is referred to as the Class VI Rule. This rule
establishes a new class of injection well (Class VI) and sets minimum federal technical criteria
for Class VI injection wells for the purposes of protecting underground sources of drinking water
(USDWs). This guidance is part of a series of technical guidance documents that EPA is
developing to support owners or operators of Class VI wells and UIC Program permitting
authorities in the implementation of the Class VI Rule. The Class VI Rule and related documents
are available at http://water.epa.gov/type/groundwater/uic/wells_sequestration.cfm.
The Class VI Rule requires owners or operators of Class VI injection wells to delineate the area
of review (AoR) for the proposed Class VI well, which is the region surrounding the proposed
well where USDWs may be endangered by the injection activity [40 CFR 146.84]. The Class VI
Rule requires that the AoR be delineated using computational modeling and the AoR must be
reevaluated periodically during the lifetime of the geologic sequestration (GS) project [40 CFR
146.84]. Within the AoR, the owners or operators must identify all potential conduits for fluid
movement out of the injection zone, including both geologic features and artificial penetrations
[40 CFR 146.84(c)(l)(iii)]. The owner or operator must then evaluate those artificial penetrations
that may penetrate the confining layer(s) of the injection project for the quality of casing and
cementing, and in the case of abandoned wells, for the quality of plugging and abandonment, and
perform corrective action on any identified artificial penetrations that could serve as a conduit for
fluid movement [40 CFR 146.84(c)(2), 146.84(c)(3), and 146.84(d)]. The Class VI Rule allows,
at the discretion of the UIC Program Director, the use of "phased" corrective action, where
certain regions of the AoR are addressed prior to injection and other regions of the AoR are
addressed during the injection phase of the project [40 CFR 146.84(b)(2)(iv)].
This guidance provides information regarding modeling requirements and recommendations for
delineating the AoR, describes the circumstances under which the AoR is to be reevaluated, and
describes how to perform an AoR reevaluation. In addition, the guidance presents information on
how to identify, evaluate, and perform corrective action on artificial penetrations located within
the AoR.
The introductory section reviews the definition of the AoR and regulations pertaining to AoR
and corrective action in the Class VI Rule. Following that section:
Section 2 addresses computational modeling of GS;
Section 3 addresses AoR delineation using computational models;
Section 4 addresses identification, evaluation, and performing corrective action on
artificial penetrations within the AoR; and
Section 5 addresses AoR reevaluation.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
-------
For each section, the guidance:
Explains how to perform activities necessary to comply with AoR and corrective
action requirements (e.g., performing computational modeling). Illustrative examples
are provided in several cases;
Provides references to comprehensive reference documents and the scientific
literature for additional information; and
Explains how to report to the UIC Program Director the results of activities related to
AoR and corrective action.
UIC Program Class VI Well Area of Review in
Evaluation and Corrective Action Guidance
-------
Table of Contents
Disclaimer i
Executive Summary ii
Table of Contents iv
List of Figures vi
List of Tables vii
Acronyms and Abbreviations viii
Definitions ix
Unit Conversions xii
1. Introduction 1
1.1. Overview of the Class VI Rule AoR and Corrective Action Requirements 2
1.2. Organization of this Guidance 5
2. Computational Modeling for Geologic Sequestration 7
2.1. Modeled Processes 10
2.2. Model Parameters 11
2.2.1. Intrinsic Permeability 14
2.2.2. Relative Permeability and Capillary Pressure 16
2.2.3. Injection Rate 19
2.2.4. Fluid Properties and Equations of State 19
2.2.5. Mass-Transfer Coefficients 19
2.2.6. Mineral Precipitation Kinetic Parameters 20
2.2.7. Model Orientation and Gridding Parameters 20
2.3. Computational Approaches 21
2.3.1. Numerical Approaches 21
2.3.2. Analytical, Semi-Analytical, and Hybrid Approaches 22
2.4. Model Uncertainty and Sensitivity Analyses 22
2.5. Model Calibration 23
2.6. Existing Codes used for Development of GS Models 28
3. AoR Delineation Using Computational Models 30
3.1. AoR Delineation Class VI Rule Requirements 30
3.2. Data Collection and Compilation 31
3.2.1. Site Hydrogeology 31
3.2.2. Operational Data 32
3.3. Model Development 33
3.3.1. Conceptual Model of the Proposed Injection Site 33
3.3.2. Determination of Physical Processes to be Included in the Computational
Model 36
3.3.3. Computational Model Design 36
3.3.3.1. Computational Code Determination 36
3.3.3.2. Model Spatial Extent, Discretization, and Boundary Conditions 37
UIC Program Class VI Well Area of Review iv
Evaluation and Corrective Action Guidance
-------
3.3.3.3. Model Timeframe 37
3.3.3.4. Parameterization 37
3.3.4. Executing the Computational Model 38
3.4. AoR Delineation Based on Model Results 38
3.4.1. Determination of Threshold Pressure Front 39
3.5. Reporting AoR Delineation Results to the UIC Program Director 48
4. Identifying Artificial Penetrations and Performing Corrective Action 50
4.1. Rule Requirements 50
4.2. Identifying Artificial Penetrations within the AoR 51
4.2.1. Historical Research 52
4.2.2. Site Reconnaissance 53
4.2.3. Aerial and Satellite Imagery Review 53
4.2.4. Geophysical Surveys 53
4.2.4.1. Magnetic Methods 54
4.2.4.3. Electromagnetic Methods 56
4.2.4.4. Ground Penetrating Radar 56
4.3. Assessing Identified Abandoned Wells 56
4.3.1. Abandoned Well Plugging Records Review 57
4.3.2. Abandoned Well Field Testing 60
4.4. Performing Corrective Action on Wells Within the AoR 62
4.4.1. Plugging of Wells within the AoR 65
4.4.2. Remedial Cementing 66
4.5. Reporting Well Identification, Assessment, and Corrective Action to the UIC
Program Director 66
5. AoR Reevaluation 68
5.1. Class VI Rule Requirements Related to AoR Reevaluation 68
5.2. Conditions Warranting an AoR Reevaluation 69
5.2.1. Minimum Fixed Frequency 69
5.2.2. Significant Changes in Operations 69
5.2.3. Results from Site Monitoring that Differ From Model Predictions 70
5.2.4. Ongoing Site Characterization 71
5.3. Performing an AoR Reevaluation 71
5.3.1. Demonstrating Adequate Existing AoR Delineation 75
5.3.2. Modifying the Existing AoR Delineation 75
References 78
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
-------
List of Figures
Figure 1-1: Flow Chart of Monitoring and Modeling at a GS Project 3
Figure 2-1: Equations of State for Carbon Dioxide 9
Figure 2-2: Example Relative Permeability-Saturation and Capillary Pressure-Saturation
Relationships for Water and Carbon Dioxide 17
Figure 2-3: Geologic Schematic of Frio Brine Pilot Project 25
Figure 2-4: Observed and Modeled Carbon Dioxide Arrival at the Observation Well Based
on Change in Fluid Density 26
Figure 2-5: Observed and Modeled Pressure Increase at (a) the Injection Well and (b) the
Monitoring Well 26
Figure 2-6: Comparison of (a) Initial and (b) Post-Calibration Model Predictions of Carbon
Dioxide Plume Evolution 27
Figure 3-1: Hypothetical Conceptual Site Model for Geologic Sequestration 34
Figure 3-2: Fluid Density Functions for Varying Salinities 40
Figure 3-3: Hypothetical Geologic Sequestration Site: Cross Sectional Schematic and
Calculations 43
Figure 3-4: Hypothetical Geologic Sequestration Site: Model Predicted Maximum Pressure
Within the Injection Zone 45
Figure 3-5: Hypothetical Geologic Sequestration Site: Model Predicted Extent of
Supercritical Carbon Dioxide Plume Over Time 46
Figure 3-6: Hypothetical Geologic Sequestration Site: Initial Area of Review Based on
Model Results 47
Figure 4-1: Total Field Aeromagnetic Map, Cook Creek Oil Field, Arcadia, Oklahoma 55
Figure 4-2: Examples of Carbon Dioxide Leakage Through Improperly Abandoned Wells 58
Figure 4-3: Well Evaluation Decision Tree 64
Figure 5-1: Hypothetical Geologic Sequestration Site: Comparison of Model Predictions and
Plume Monitoring Results at 20 Years of Injection 72
Figure 5-2: Hypothetical Geologic Sequestration Site: Comparison of Model Predictions and
Pressure Monitoring Results at 20 Years of Injection 73
Figure 5-3: Hypothetical Geologic Sequestration Site: Initial AoR Delineation and
Delineation after Reevaluation 77
UIC Program Class VI Well Area of Review vi
Evaluation and Corrective Action Guidance
-------
List of Tables
Table 2-1: Model Parameters for Multiphase Fluid Modeling of Geologic Sequestration 12
Table 4-1: Tools for Assessment of the Integrity of Abandoned Wells 62
UIC Program Class VI Well Area of Review vii
Evaluation and Corrective Action Guidance
-------
Acronyms and Abbreviations
AoR Area of review
API American Petroleum Institute
CAA Clean Air Act
CFR Code of Federal Regulations
DOE United States Department of Energy
EPA United States Environmental Protection Agency
GHG MRR Greenhouse Gas Mandatory Reporting Rule
GPR Ground penetrating radar
GS Geologic sequestration
LBNL Lawrence Berkeley National Laboratory
MAE Mean-absolute error
mD Millidarcies
ME Mean error
MIT Mechanical integrity test
MPa Megapascals
PISC Post-injection site care
RCSP Regional Carbon Sequestration Partnership
RMSE Root-mean squared error
SDWA Safe Drinking Water Act
UIC Underground Injection Control
USDW Underground source of drinking water
USGS United States Department of the Interior, United States Geological Survey
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
vm
-------
Definitions
Key to definition sources:
l:40CFR146.81(d).
2: Definition drafted for the purposes of this document.
3: Class VI Rule Preamble.
4:40CFR144.3.
Area of Review (AoR) means the region surrounding the geologic sequestration project where
USDWs may be endangered by the injection activity. The area of review is delineated using
computational modeling that accounts for the physical and chemical properties of all phases of
the injected carbon dioxide stream and displaced fluids, and is based on available site
characterization, monitoring, and operational data as set forth in 40 CFR 146.84.l
Boundary condition parameters refers to the parameters that describe fluid flow rates and/or
pressures at the edges of the model domain and in the location of injection/extraction wells.
Capillary Pressure refers to the difference of pressures between two phases existing in a system
of interconnecting pores or capillaries. The difference in pressure is due to the combination of
surface tension and curvature in the capillaries.2
Computational code refers to a series of interrelated mathematical equations solved by
computer to represent the behavior of a complex system. For the purposes of GS, computational
models represent, at a minimum, the flow and transport of multiple fluids and components in
varying phases through porous media. Computational codes offer the ability to predict fluid flow
in the subsurface using scientifically accepted mathematical approximations and theory. The use
of computational codes is necessary because the mathematical formulations describing fluid flow
are complicated and in many cases, non-linear. Several codes have been specifically developed
or tailored for injection activities similar to GS, and can be used for this purpose.2
Computational model means a mathematical representation of the injection project and relevant
features, including injection wells, site geology, and fluids present. For a GS project, site specific
geologic information is used as input to a computational code, creating a computational model
that provides predictions of subsurface conditions, fluid flow, and carbon dioxide plume and
pressure front movement at that site. The computational model includes all model input and
predictions (i.e., outputs).
Confining zone means a geologic formation, group of formations, or part of a formation
strati graphically overlying the injection zone(s) that acts as barrier to fluid movement. For Class
VI wells operating under an injection depth waiver, confining zone means a geologic formation,
group of formations, or part of a formation strati graphically overlying and underlying the
injection zone.1
UIC Program Class VI Well Area of Review ix
Evaluation and Corrective Action Guidance
-------
Constitutive relationship typically means, empirically based approximations used to simplify
the system and estimate unknowns in cases where the parameters of the governing equations are
not readily available for use in the equation because necessary information is not typically
measurable, and thus not directly input into the model. An example of a constitutive relationship
is relative permeability-saturation functions. These functions estimate the relative permeability
of a particular fluid in a porous media as a function of its saturation at a given location and time.
This permeability is then used in the governing equation to predict flow.2
Equation of state refers to an equation that expresses the equilibrium phase relationship
between pressure, volume and temperature for a particular chemical species.2
Geologic sequestration (GS) means the long-term containment of a gaseous, liquid or
supercritical carbon dioxide stream in subsurface geologic formations. This term does not apply
to carbon dioxide capture or transport.1
Geophysical surveys refers to the use of geophysical techniques (e.g., seismic, electrical,
gravity, or electromagnetic surveys or well logging methods such as gamma ray and spontaneous
potential) to characterize subsurface rock formations.3
Governing equation refers to the mathematical formulae that form the basis of the
computational code are termed governing equations. For GS modeling, they govern the predicted
behavior of fluids in the subsurface provided by the code. Governing equations are mathematical
approximations for describing flow and transport of fluids and their components in the
environment.2
Ground Penetrating Radar (GPR) refers to a geophysical method that utilizes microwave
technology in order to characterize features found in the subsurface.2
Heterogeneity refers to the spatial variability in the geologic structure and/or physical properties
of the site.
Hysteresis means the phenomenon where the response of a system depends not only on the
present stimulus, but also on the previous history of the medium. For example, in a GS project,
relative permeability, capillary pressure, and residual trapping will depend upon the saturation
history of the formation.
Immiscible refers to the property wherein two or more liquids or phases do not readily dissolve
in one another.2
Initial conditions refers to parameter values at the start of the model simulation.2
Intrinsic permeability refers to a parameter that describes properties of the subsurface that
impact the rate of fluid flow. Larger intrinsic permeability values correspond to greater fluid
flow rates. Intrinsic permeability has units of area (distance squared).2
Model calibration means adjusting model parameters in order to minimize the difference
between model predictions and monitoring data at the site.2
UIC Program Class VI Well Area of Review x
Evaluation and Corrective Action Guidance
-------
Multiphase flow refers to flow in which two or more distinct phases are present (e.g., liquid,
gas, supercritical fluid).2
Numerical Artifacts refers to model results that are created erroneously based on computational
limitations of the model, which may result from improper model development.2
Parameter means a mathematical variable used in governing equations, equations of state, and
constitutive relationships. Parameters describe properties of the fluids present, porous media, and
fluid sources and sinks (e.g., injection well). Examples of model parameters include intrinsic
permeability, fluid viscosity, and fluid injection rate.
Relative permeability refers to a factor, between 0 and 1, that is multiplied by the intrinsic
permeability of a formation to compute the effective permeability for a fluid in a particular pore
space. When immiscible fluids (e.g., carbon dioxide, water) are present within the pore space of
a formation, the ability for flow of those fluids is reduced, due to the blocking effect of the
presence of the other fluid. This reduction is represented by relative permeability.
Sensitivity Analyses refers to the study of how the output of a model varies based in changes to
an input variable or model parameter over a specified range of values. The results of a sensitivity
analysis determine the which input variable and model parameter variability have the greatest
effect on the model results.2
Stochastic Methods means the use of probability statistical methods in development of one or
more possible realizations of the spatial patterns of the value(s) of a given set of model
parameters.
Underground Injection Control Program refers to the program EPA, or an approved state, is
authorized to implement under the Safe Drinking Water Act (SDWA) that is responsible for
regulating the underground injection of fluids by wells injection. This includes setting the federal
minimum requirements for construction, operation, permitting, and closure of underground
injection wells.
Underground Source of Drinking Water (USDW) means an aquifer or its portion which
supplies any public water system; or which contains a sufficient quantity of ground water to
supply a public water system; and currently supplies drinking water for human consumption; or
contains fewer than 10,000 mg/1 total dissolved solids; and which is not an exempted aquifer.4
UIC Program Class VI Well Area of Review xi
Evaluation and Corrective Action Guidance
-------
Unit Conversions
Imperial/Non-Metric Unit
Metric Unit
IFoot
IMile
1 Pound per Square Inch (psi)
Temperature in Degrees Fahrenheit (°F)
1 Pound (Ib)
1 Megatonne (Mt)
1 Metric Ton (tonne; t)
1 Cubic Foot
0.3048 Meters
1.609 Kilometers
0.006895 Megapascals (MPa)
Temperature in Degrees Celsius
(°F - 32) x 0.56
0.4536 Kilograms
1 x 106 Tonnes
1,000 kg
0.0283 Cubic Meters
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
xi i
-------
1. Introduction
Area of review (AoR) evaluations and corrective action are long-standing permit requirements of
the Underground Injection Control (UIC) Program of the U.S. Environmental Protection Agency
(EPA). The AoR refers to the delineated region surrounding the injection well(s) wherein the
potential exists for underground sources of drinking water (USDWs) to be endangered by the
leakage of injectate and/or formation fluids. Typically, for injection well classes other than Class
VI, the AoR is defined either as a fixed radius around the injection well or by a relatively simple
radial calculation. Owners or operators of injection wells are required to identify any potential
conduits for fluid movement, including artificial penetrations (e.g., abandoned well bores) within
the AoR, assess the integrity of any artificial penetrations, and perform corrective action where
necessary to prevent fluid movement into a USDW [40 CFR 144.55, 146.84(d)].
The Class VI Rule introduces enhanced AoR and corrective action requirements for Class VI
injection wells that are tailored to the unique circumstances of geologic sequestration (GS) of
carbon dioxide projects [40 CFR 146.84]. The purpose of this guidance is to identify appropriate
methods for delineating the AoR and performing corrective action for Class VI injection wells.
The intended primary audiences of this guidance document are Class VI injection well owners or
operators and their representatives conducting AoR delineation modeling or performing artificial
penetration identification, assessment, and corrective action activities. The UIC Program staff
who are responsible for reviewing and approving Class VI injection well permit applications and
related reports concerning AoR delineation and corrective action are another intended audience
of this guidance document.
This document is one of a series of technical guidance documents intended to provide
information and possible approaches for addressing various aspects of permitting and operating a
Class VI injection well. Three of these companion guidance documents focus on site
characterization, well construction, and testing and monitoring:
The UIC Program Class VI Well Site Characterization Guidance;
The UIC Program Class VI Well Construction Guidance; and
The UIC Program Class VI Well Testing and Monitoring Guidance.
These guidance documents are intended to complement each other and to assist owners or
operators in preparing permit applications that satisfy the requirements of the Class VI Rule.
Class VI injection well regulations are tailored to the characteristics of individual sites. For
example, the required site characterization data collected will inform the model development for
AoR delineation, and AoR models will be reevaluated, and perhaps change, based on the results
of site testing and monitoring data (Figure 1-1). Cross-linkages between guidance documents are
noted in the text where appropriate. Additional guidance on developing, presenting, and using
the required Class VI project plan information as part of a Class VI injection well permit
application is provided in the UIC Program Class VI Well Project Plan Development Guidance.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
-------
1.1. Overview of the Class VI Rule AoR and Corrective Action Requirements
The Class VI Rule defines the AoR as the region surrounding the GS project where USDWs may
be endangered by the injection activity [40 CFR 146.84(a)]. The purpose of the AoR and
corrective action requirements of the Class VI Rule is to ensure that the areas potentially
impacted by a proposed GS operation are delineated, all wells that need corrective action receive
it, and that this process is updated throughout the injection project. While the details of all of the
requirements are presented in later sections of this guidance, the basic requirements that owners
or operators of GS projects must meet include:
Prepare, maintain, and comply with an AoR and Corrective Action Plan that includes all
of the required elements of the plan [40 CFR 146.84(b)];
Delineate the AoR using computational modeling and identify all wells that require
corrective action [40 CFR 146.84(c)];
Perform all required corrective action on wells in the AoR [40 CFR 146.84(d)];
Reevaluate the AoR throughout the life of the project [40 CFR 146.84(e)];
Ensure that the Emergency and Remedial Response Plan and financial responsibility
demonstration account for the most recently approved AoR [40 CFR 146.84(f)]; and
Retain modeling inputs and data used to support AoR reevaluations for 10 years [40 CFR
146.84(g)].
If a vertical leakage pathway out of the injection zone is present, USDWs in the vicinity of a
proposed Class VI injection well may be endangered by (1) movement of carbon dioxide into the
USDW, impairing drinking water quality through changes in pH, contamination by trace
impurities in the injectate (e.g., mercury, hydrogen sulfide), and leaching of metals and/or
organics; and (2) movement of non-potable water (e.g., brine) out of the injection formation into
a USDW as caused by elevated formation pressures induced by injection. Therefore, the AoR
encompasses the region overlying the separate-phase (e.g., supercritical, liquid, or gaseous)
carbon dioxide plume and the region overlying the pressure front where fluid pressures are
sufficient to force fluids into a USDW. While it may often be the case that the AoR will
encompass the boundary of the GS project, within which all project activities will occur, the
Class VI Rule does not require that the AoR and overall project boundary be equivalent in all
cases.
The Class VI Rule requires that "the AoR is delineated using computational modeling that
accounts for the physical and chemical properties of all phases of the injected carbon dioxide
stream and is based on available site characterization, monitoring, and operational data" [40 CFR
146.84(a)]. As discussed below, GS computational modeling for Class VI injection wells is
complex and requires advanced methods. Additionally, the AoR must be reevaluated at a
minimum fixed frequency not to exceed five years, or when monitoring and operational
conditions warrant [40 CFR 146.84(e)]. The purpose of Class VI well AoR reevaluations is to
UIC Program Class VI Well Area of Review 2
Evaluation and Corrective Action Guidance
-------
ensure that site monitoring data are used to update modeling results, and that the AoR delineation
reflects any changes in operational conditions. The general relationship between site
characterization, modeling, and monitoring activities at a GS project is shown in Figure 1-1.
Site Characterization
Proposed Operating
Data
Computational Modeling/
AoR Delineation
Model Calibration
Monitoring System
Design
Monitoring Data
Collection and Interpretation
Figure 1-1: Flow Chart of Monitoring and Modeling at a GS Project.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
-------
An individual Class VI permit must be obtained separately for each injection well, as area
permits are not allowed under the Class VI Rule [40 CFR 144.33]. However, EPA anticipates
that, in many cases, multiple injection wells will be operated within a single GS project. If
approved by the UIC Program Director, AoR delineation and corrective action activities may be
performed collectively for all wells included within a single project. All required submittals (e.g.,
maps of the delineated the AoR and the AoR and Corrective Action Plan) must be submitted
separately for each well, however, so that they may be incorporated into each well's Class VI
permit. In all cases, EPA recommends that AoR delineation models account for all wells
injecting into (including any injection wells associated with other UIC well classes or other Class
VI operations) or pumping from the injection zone or any other zones that are hydraulically
connected to the injection zone.
The corrective action requirements for Class VI wells are generally similar to those for other
injection well classes. However, due to the potentially large AoR of GS projects, EPA has
allowed the use of phased corrective action, if approved by the UIC Program Director [40 CFR
146.84(b)(2)(iv)]. If phased corrective action is approved by the UIC Program Director, owners
or operators would be allowed to perform corrective action only on the subset of artificial
penetrations located within the AoR prior to injection that are located in regions nearest the
injection well(s). Corrective action would continue during injection in the remaining regions of
the AoR prior to carbon dioxide migration or pressure elevation in that area. EPA encourages
owners or operators to perform all necessary corrective action on deficient wells identified
during the initial AoR delineation or AoR reevaluations before the end of the injection phase.
As a part of a Class VI permit application, the owner or operator must submit an AoR and
Corrective Action Plan that describes the anticipated activities that will be performed to comply
with these requirements [40 CFR 146.84(b)]. The AoR and Corrective Action Plan must be
approved by the UIC Program Director prior to submittal of the initial AoR delineation and
issuance of a permit [40 CFR 146.84(b)]. This plan will facilitate dialogue between the owners
or operators and the UIC Program Director to ensure that the UIC Program Director understands
and agrees early in the project lifetime with the methods necessary to delineate the AoR and
complete all required corrective action. A Class VI AoR and Corrective Action Plan must
include the following information [40 CFR 146.84(b)]:
1. The method for delineating the AoR, including the model to be used, assumptions that
will be made, and the site characterization data on which the model will be based;
2. The minimum fixed frequency, at least once every five (5) years, that the owner or
operator proposes to reevaluate the AoR;
3. The site- and project-specific monitoring and operational conditions that would warrant a
reevaluation of the AoR prior to the next routinely scheduled reevaluation;
4. How specific monitoring and operational data (e.g., injection rate and pressure) will be
used to inform an AoR reevaluation;
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
-------
5. How corrective action will be conducted, including what corrective action will be
performed prior to injection and what, if any, portions of the AoR will have corrective
action addressed on a phased basis and how the phasing will be determined;
6. How corrective action will be adjusted if there are changes in the AoR; and
7. How site access will be guaranteed for future corrective action.
The requirements related to the AoR and Corrective Action Plan are discussed in depth in the
UIC Program Class VI Well Project Plan Development Guidance.
1.2. Organization of this Guidance
This guidance document is organized to generally follow the sequence of AoR delineation and
corrective action activities that an owner or operator will perform over the life of a proposed and
later permitted Class VI injection project. These activities will generally proceed as described
below.
Prior to the issuance of a permit for the construction of a new Class VI well (or the conversion of
an existing well):
1. Collection of relevant site characterization and operational data [40 CFR 146.82(a)(3),
146.82(a)(5), 146.82(a)(6), and 146.83];
2. Determination of relevant operational data that will inform the AoR modeling [40 CFR
146.82(a)(7), and 146.82(a)(10)-(ll)];
3. Development of an AoR and Corrective Action Plan [40 CFR 146.82(a)(13) and
146.84(b)];
4. Performing AoR modeling and delineation [40 CFR 146.82(a)(2)]; and
5. Identification and assessment of artificial penetrations within the AoR [40 CFR
146.82(a)(4)].
Prior to granting approval for injection:
6. Collection and/or updating of relevant site characterization and operational data that will
inform AoR modeling [40 CFR 146.82(c)(2)-(5), 146.82(c)(7), and 146.83];
7. Identification of any needed updates to the AoR and Corrective Action Plan [40 CFR
146.82(c)(9)];
8. Finalizing AoR modeling and delineation [40 CFR 146.82(c)(l)]; and
9. Performing corrective action on those penetrations that may serve as a conduit for fluid
movement [40 CFR 146.82(c)(6) and 146.84(d)].
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
-------
During injection and post-injection site care (PISC):
10. Reevaluation of the AoR periodically, at least once every five (5) years [40 CFR
146.82(c)(9) and 146.84(e)], and updating the AoR and Corrective Action Plan; and
11. If phased corrective action is approved or when additional corrective action is warranted
based on AoR reevaluations, performing corrective action [40 CFR 146.82(c)(6) and
146.84(d)].
Activities (1) through (5) must be performed prior to receiving approval to construct a Class VI
injection well and activities (6) through (9) must be performed prior to receiving approval to
inject carbon dioxide, and their results must be submitted to the UIC Program Director as part of
the Class VI injection well permit application [40 CFR 146.82(a)]. The remaining activities will
be performed after a permit application has been approved by the UIC Program Director and the
Class VI injection well is actively operating.
This guidance document generally focuses on activities (4), (5), (8), (9), (10), and (11), as
follows:
Section 2 provides background on computational modeling of GS activities;
Section 3 discusses performing computational modeling to delineate the AoR and comply
with permit requirements (activities 4 and 8);
Section 4 focuses on abandoned well identification, assessment, and corrective action on
all artificial penetrations within the AoR (activities 5, 9, and 11); and
Section 5 focuses on reevaluation of the AoR (activity 10).
Site characterization activities (activities 1 and 6) are discussed briefly in this guidance (Section
3.2), and are covered in more detail in the UIC Program Class VI Well Site Characterization
Guidance. Preparation of the AoR and Corrective Action Plan (activity 3) and identification of
any updates (activity 7) are also discussed briefly herein, and in more detail in the UIC Program
Class VI Well Project Plan Development Guidance.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
-------
2. Computational Modeling for Geologic Sequestration
This section discusses the fundamentals of computational modeling for GS to provide the
necessary background for owners or operators, and to assist in understanding and complying
with the Class VI Rule. While computational modeling for GS may entail a vast amount of
complex information, the purpose of this section is to provide a brief introduction to the
modeling techniques and fundamentals that may be necessary or significant for satisfying the
specific rule requirements. Following an introduction that explains the use of modeling in the
context of meeting the requirements of the Class VI Rule:
Section 2.1 discusses the processes that can be modeled;
Section 2.2 presents a discussion of the parameters that are included in AoR models;
Section 2.3 presents available computational approaches, including numerical, analytical,
semi-analytical, and hybrid approaches;
Section 2.4 discusses model uncertainty and sensitivity analyses;
Section 2.5 discusses model calibration; and
Section 2.6 provides an overview of existing codes used for development of GS models.
The AoR for a Class VI injection project must be delineated using a computational model that
accounts for the physical and chemical properties of all phases of the injected carbon dioxide [40
CFR 146.84(a)]. A computational model is based on available data, and it is a mathematical
representation of the GS project and relevant features, including injection wells, site geology,
and fluids present. As described below, a site-specific computational modelis designed by
incorporating the GS site and operational characteristics into a computational code, which is a
computer program that has been designed to simulate multiphase flow and other pertinent
processes in geologic media based on scientific principles and accepted mathematical (i.e.,
governing) equations.
Computational codes that may be used for modeling of GS are necessarily more technically
complex than commonly used ground water flow codes because GS modeling considers
multiphase (e.g., gas, liquid, supercritical fluid) flow of several fluids (i.e., ground water, carbon
dioxide, hydrocarbons), phase changes of carbon dioxide, heat flow, and significant pressure
changes. Furthermore, in some cases models consider reactive transport (e.g., chemical reactions
between constituents) and geomechanical processes (e.g., induced fault activation). As discussed
below, the Class VI Rule requires that the AoR be delineated using models that include
multiphase flow [40 CFR 146.84(a)], but not necessarily reactive transport or geomechanical
processes. However, inclusion of these processes in the AoR delineation model may be
important in some cases and may be required by the UIC Program Director.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
-------
Several codes are available that are capable for use in development of adequate models for
delineation of the AoR at a GS site and for complying with Class VI injection well permit
requirements (Section 2.6). Although available codes are sophisticated and based on the best-
available scientific understanding, computational models are approximations and are never
perfect representations of reality, and they cannot provide a completely accurate prediction of
fluid movement at a GS site. Therefore, the characterization of model uncertainty, using
sensitivity analyses, and conducting model calibrations are very important parts of many
computational approaches.
Research studies have provided valuable information on the capabilities of available models,
what information may be collected in order to properly inform model development, and how the
model results may be presented. Importantly, the information and guidance provided here is
based on currently available data, scientific research, and project experience. EPA recognizes
that data collected from future GS projects may advance GS computational modeling and AoR
delineation. EPA is committed to an adaptive approach for regulation of GS projects and may
revisit aspects of Class VI federal regulations, and guidance, as new data becomes available.
There is a long history of simulating multiphase flow and transport in porous media using
computational models. Comprehensive reviews of multiphase flow in porous media and
modeling are provided elsewhere (e.g., Finder and Gray, 2008; Miller et al., 1998; Gerritsen and
Durlofsky, 2005; Finsterle, 2004). These models solve a series of governing equations to predict
the composition and volumetric fraction of each phase state (e.g., liquid, gas, supercritical fluid)
as a function of space and time for a particular set of circumstances. Governing equations are
formulated to describe the flow and transport of several chemical species in several phases, in
which interphase mass transfer may be important. Typically, flow equations are derived by
substituting a multiphase form of Darcy's Law into continuum balance expressions.
The solution of the continuum balance equations requires that they be supplemented with closure
relations that express unknowns in terms of accessible parameters. These include equations of
state and constitutive relationships. Equations of state express the equilibrium phase relationship
between pressure, volume, and temperature for a particular chemical. Accepted equations of state
for carbon dioxide are presented in Figure 2-1, and are discussed in Section 2.2.4. Constitutive
relationships are typically empirically based approximations used to simplify the system and
estimate unknowns. Examples of constitutive relationships are saturation-relative permeability
relationships, interphase mass transfer relations, and solution reaction relations.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
-------
1200
1000-
800
5)
£
c
w
Q 400
200
0
Liquid i
Supercritical Region
o
30
60
180
210
90 120 150
Temperature (°C)
(a) Equations of State for CO, Giving the Phase State as a Function
of Temperature, Pressure, and Density
0.18
0.16-
0.14-
"3T
« 0.12-
E,
g 0.08
tn
> 0.06-
0.04
0.02-I
0
Supercritical Region
Gas
o
30
60
180
210
90 120 150
Temperature (°C)
(b) Equations of State for CO, Giving the Phase State as a Function
of Temperature, Pressure, and Viscosity
Explanation
| ^/ | Vaporization curve
( j Critical point
|- ~| SLpercritical boundary Source: After Nordbotten et al (2005), Fig. 2
Figure 2-1: Equations of State for Carbon Dioxide.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
-------
2.1. Modeled Processes
Computational codes used for GS vary in complexity and may include routines for multiphase
flow, reactive transport, and geomechanical processes. Traditionally, codes have been developed
as separate entities to simulate these processes. Present-day simulators typically address and
couple multiphase flow with geomechanical processes or geochemical processes. However,
depending on site-specific characteristics, where the geochemical and geomechanical processes
are significant, robust simulation of GS may require interactive coupling of all three processes.
The Class VI Rule only requires that multiphase flow be included in computational modeling.
However, the owner or operator, or the UIC Program Director, may determine that reactive
transport and/or geomechanical modeling should also be included for a particular proposed
project. For example, reactive transport could be relevant if permeability and/or porosity are
predicted, based on previous testing, to change as a result of precipitation/dissolution reactions.
Geomechanical processes could be relevant if previous testing has indicated that pressure and
stress changes hydrogeologic properties and/or affects confining unit integrity.
Codes used to simulate multiphase flow generally incorporate some or all of the following
processes: phase transition behavior of carbon dioxide (gas, liquid, supercritical fluid) and
associated buoyancy; dissolution of carbon dioxide in brine and oil and associated increased
density; dissolution of water in carbon dioxide; variable viscosity and density of brine and
carbon dioxide phases; thermal effects such as cooling or freezing due to carbon dioxide
expansion from supercritical and liquid phases; and reduced fluid permeability due to the
presence of several immiscible fluids within a pore space.
Codes used to simulate reactive transport generally incorporate rate-limited intra-aqueous
reactions, mineral dissolution and precipitation, changes in porosity and permeability due to
these reactions, and multi-component gas mixtures. Reactive transport models can be used to
determine the impact of carbon dioxide and its co-injectates (e.g., hydrogen sulfide, sulfur
dioxide) on aquifer acidification, the concomitant mobilization of metals, and any mineral
trapping of carbon dioxide (e.g., precipitation of carbonate minerals). Reactive transport models
can also be used to assess corrosion of well construction materials as influenced by carbon
dioxide.
The length scales associated with interfacial geochemistry are very small (e.g., micrometers to
millimeters) compared to multiphase flow simulation (meters to kilometers). Small grid spacing
around these regions may imply associated small time steps, so that the overall problem becomes
computationally demanding when trying to couple these reactions to multiphase flow. Data
related to geochemical rate parameters are generally lacking (e.g., Knauss et al., 2005; Xu et al.,
2006), and have to be estimated for a wide range of possible environmental conditions and
mineralogical interfaces. Several common codes that may be used for AoR delineation, such as
ECLIPSE, normally do not include routines for reactive transport.
Geomechanical codes can be used to evaluate the effect of reservoir pressurization and buoyancy
on the integrity of geologic confining units, reactivation of existing fractures and faults, and rock
properties such as porosity and permeability. The amount and spatial distribution of pressure
UIC Program Class VI Well Area of Review 10
Evaluation and Corrective Action Guidance
-------
buildup in a geologic formation will depend on the rate of injection, the permeability and
thickness of the injection formation, the mechanistic properties of the rock matrix, the
permeability of the confining units, and the presence or absence of permeability barriers, and
boundary conditions of the system. Models used to simulate geomechanical processes generally
incorporate effective stress/strain relationships, aperture stiffness and associated closing and
widening, and variation in porosity and permeability. Geomechanical modeling may require
simulation on both a large and small scale (individual fractures), which can be computationally
challenging (i.e., require long model processing times on the order of days). When individual
fractures are considered, the spatial grid resolution may need to be on the order of meters or less.
Therefore, smaller-domain models may be necessary to investigate migration through individual
fractures. Generally, simulation of flow through a fractured reservoir requires codes that have
been designed for this purpose.
2.2. Model Parameters
A parameter is a variable in the governing equations of the model that may be of uniform value
throughout the domain, or may vary in space and time. While maintaining salient features of the
hydrogeologic system, some system aspects are often lumped together in simulation models and
described by effective parameters that are estimated or averaged from several data sources.
Relevant parameters for multiphase flow modeling of GS are summarized in Table 2-1, and
include hydrogeologic characteristics, fluid properties, chemical properties, fluid injection and
withdrawal rates, initial and boundary conditions, system orientation (e.g., model domain, grid
cell size), and simulation control parameters. Initial conditions describe parameter values at the
start of the model run. Boundary condition parameters describe conditions of the system (e.g.,
fluid flow rates and/or pressures) at the edges of the model domain and at the location of
injection and/or extraction wells. While fluid injection and/or withdrawal rates and simulation
control parameters are project specific, other particularly important site- and project-specific
parameters for GS include formation intrinsic permeability, porosity, relative permeability,
compressibility, fluid viscosity, and fluid density. These parameters are discussed in the
following subsections.
Parameter values are to be based on site data to the extent possible. However, as discussed
below, in cases where detailed site geologic characterization data are unavailable (i.e., if
formation testing is not complete or core samples are not available), parameter values may be
estimated from standard values or relationships in the scientific literature. Model calibration,
which may occur during AoR reevaluation, consists of adjusting a subset of the estimated
parameter values to minimize the difference between model simulations and observed
monitoring data. Model parameters may also be adjusted based on newly acquired site
characterization data. For example, data gathered during well logging may inform updates to
parameter values. See the UIC Program Class VI Well Site Characterization Guidance for more
information.
UIC Program Class VI Well Area of Review 11
Evaluation and Corrective Action Guidance
-------
Table 2-1: Model Parameters for Multiphase Fluid Modeling of Geologic Sequestration.
Parameter
Description
Estimation Methods
Hydrogeologic Properties
Intrinsic Permeability
Porosity
Capillary Pressure
Relative Permeability
Fluid Pressure
Temperature
Formation Compressibility
Water Saturation
Carbon Dioxide Saturation
Storativity
Represents properties of the
subsurface that impact the rate
of fluid flow.
The relative volume of void
space within a formation.
Controls the volume of carbon
dioxide that may be stored.
The pressure difference across
the interface of two immiscible
fluids (e.g., carbon dioxide and
water)
Factor that determines the
decrease in permeability for a
fluid due to the presence of
other immiscible fluids
Force acting on a unit area,
measure of the potential energy
per volume of fluid
Measure of the internal energy
of a fluid
Measure of change in aquifer
volume with a change in fluid
pressure
The percent of system void
space occupied by aqueous
fluids
The percent of system void
space occupied by carbon
dioxide
The volume of fluid released
from storage per unit decline in
head per unit area of the
formation
See the UIC Program Class
VI Well Site Characterization
Guidance, and Section 2.2.1
of this guidance
See the UIC Program Class
VI Well Site Characterization
Guidance
Calculated based on fluid
saturations (see Section
2.2.2 of this guidance)
Calculated based on fluid
saturations (see Section
2.2.2 of this guidance)
See the UIC Program Class
VI Well Site Characterization
Guidance
See the UIC Program Class
VI Well Site Characterization
Guidance
See the UIC Program Class
VI Well Site Characterization
Guidance
See the UIC Program Class
VI Well Site Characterization
Guidance
Calculated by the
computational model
See Standard References,
e.g., Fetter, 2001
Fluid Properties
Viscosity
Measure of the internal
resistance to flow
Calculated based on
equations of state, also
influenced by fluid
composition (see Section
2.2.4 of this guidance)
Dimensions
L2
Dimensionless
M/LT2
Dimensionless
M/LT2
Temperature
LT2/M
Dimensionless
Dimensionless
Dimensionless
M/LT
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
-------
Parameter
Density
Composition
Fluid Compressibility
Description
The mass of a fluid per unit
volume
Molecular makeup, by volume
or mass, of a fluid.
Measurement of salinity,
concentration of trace
compounds
The change in volume of a fluid
from a unit change in pressure
Estimation Methods
Calculated based on
equations of state, also
influenced by fluid
composition (see Section
2.2.4 of this guidance)
See the UIC Program Class
VI Well Site Characterization
Guidance
See Standard References,
e.g., Perry and Green, 1984
Chemical Properties
Aqueous Diffusion
Coefficient
Aqueous Solubility
Solubility in Carbon Dioxide
The rate of chemical transport
due to a concentration gradient
The maximum concentration of
a chemical (e.g., carbon dioxide)
dissolved in the aqueous phase
The maximum concentration of
a chemical (e.g., water)
dissolved in separate-phase
carbon dioxide.
See Standard References,
e.g., Tamimi etal., 1994
Salinity, temperature and
pressure dependent (see
Spycher et al., 2003;
Spycher and Pruess, 2005)
Temperature and pressure
dependent (see Spycher et
al., 2003; Spycher and
Pruess, 2005)
Fluid injection and withdrawal rates
Injection Rates
Withdrawal Rates
Boundary Conditions
Injection rates at each well
Any fluid withdrawal rates
within model domain
Fluid pressures and/or flow
rates at the edges of the model
domain
Planned site
operational data
Measure rates for wells
conducting pumping
within the AoR
Tested in conjunction with
model extent, to ensure no
artificial influence on model
results
Fluid injection and withdrawal rates (Continued)
Initial Conditions
Fluid pressures and/or flow
rates within the domain at the
beginning of the model run
Based on pre-injection site
characterization data, see
the UIC Program Class VI
Well Site Characterization
Guidance
Dimensions
M/L3
M/L3
LT2/M
L2/T
M/L3
M/L3
L3/T
L3/T
Varies
Varies
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
13
-------
Parameter
Description
Estimation Methods
System Orientation and Simulation Controls
Model Extent (domain)
Number of Model Layers
Layer Thickness
Grid Cell Size
Model Timeframe
Time Step Size
The lateral extent of the model
in all directions
Model vertical discretization
Vertical extent of each model
layer
Lateral size of each model cell
The complete duration of the
model run
The duration of each temporal
interval during the model
timeframe
Tested in conjunction with
boundary conditions, to
ensure no artificial influence
on model results
Based on conceptual site
model of site stratigraphy,
see the UIC Program Class
VI Well Site Characterization
Guidance
See the UIC Program Class
VI Well Site Characterization
Guidance
May vary throughout
domain, as dictated by
conceptual model and
computational necessities
Tested to ensure long
enough to allow for
pressure decline to pre-
injection conditions
Often controlled by code,
tested to ensure small
enough to not artificially
influence results
Dimensions
L
Dimensionless
L
L2
T
T
L = Length; M = Mass; T = Time
2.2.1. Intrinsic Permeability
Intrinsic permeability is a property of the solid phase (i.e., porous medium) in the subsurface that
impacts the rate of fluid flow. Larger intrinsic permeability values correspond to greater fluid
flow rates. Intrinsic permeability has units of length squared and is often reported in the units of
millidarcies (mD); one mD is equal to 9.9-10~16 square meters (m2). Typical permeability values
for an injection zone at a GS project range from 1 to 102 mD (e.g., Sorensen et al., 2005; Fischer
et al., 2005; MGSC, undated; ISGS, 2009). Typical permeability values for a confining unit (e.g.,
shale) range from 10"7 to 10"4 mD (e.g., Soeder, 1988).
Intrinsic permeability incorporates the effects of formation porosity, pore structure, such as pore-
size distribution and connectivity, and the presence of fractures or faults. The spatially
heterogeneous nature of subsurface materials results in a heterogeneous intrinsic permeability
distribution in most formations. Additionally, intrinsic permeability is an anisotropic parameter,
in that lateral intrinsic permeability is often significantly larger than vertical intrinsic
permeability due to depositional layering. Anisotropy in intrinsic permeability, both vertical and
horizontal, may also be an effective property of fractured rock media. Intrinsic permeability is
typically estimated from a combination of hydrogeologic field tests (e.g., aquifer tests, pressure
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
14
-------
fall-off tests), laboratory core analysis, and geophysical well logging. The UIC Program Class
VI Well Site Characterization Guidance provides details regarding estimation of formation
intrinsic permeability. Intrinsic permeability values are often adjusted during model calibration.
See Box 2-1 for more information.
During the development of the computational model, the model developer determines how to
estimate values of intrinsic permeability within the entire model domain based on results of site
characterization activities at discrete locations. For modeling purposes, the simplest description
of intrinsic permeability is a homogenous distribution, which incorporates a single value for the
entire subsurface domain based on an average of available data. A model that assumes a
homogeneous permeability distribution, however, will not account for heterogeneity (if it exists)
that causes preferential flow paths or confining strata, or for the depth dependence of
permeability in an updipping formation. Another option is to incorporate a layered distribution,
which incorporates a single permeability value for each geologic stratum in the domain, and can
be constructed by using geologic maps and cross-sections of the proposed project site.
Alternatively, geostatistical and stochastic methods are available to create a statistical ensemble
of possible permeability distributions that incorporate both lateral and vertical heterogeneity
based on available site characterization data. Spatial variability of permeability is thus described
by a relatively small number of geostatistical parameters. Considering the large areas that are
anticipated to be modeled for AoR delineations of proposed Class VI injection well project sites,
techniques are among the best methods for incorporating realistic heterogeneity distributions into
the computational model with limited data (see inset, Box 3-1).
Compared to homogeneous or layered permeability distributions, depending on the available
data, intrinsic permeability fields developed with geostatistical techniques may provide a more
realistic representation of conditions within the formation, and resulting models may better
represent carbon dioxide migration through high-permeability channels. Commercial software
packages are available for use in the development of heterogeneous intrinsic permeability
distributions based on available site data (e.g., T-PROGS; Carle, 1999). See Doughty and Pruess
(2004), Juanes et al. (2006), Obi and Blunt (2006), and Flett et al. (2007) for examples of the
development of heterogeneous profiles based on geostatistical techniques. For more information
regarding the use of geostatistical methods, see the UIC Program Class VI Site Characterization
Guidance.
Several previous studies have evaluated the impact of permeability values on computational
modeling results, through the use of parameter sensitivity analyses. Law and Bachu (1996) and
Lindeburg (1997) demonstrated that for a homogeneous system, carbon dioxide mobility
increases with increased formation permeability. Comparing homogeneous formations and those
with layered heterogeneities, Lindeberg (1997) showed that the presence of thin shale layers
increases sweep and thus carbon dioxide dissolution. For systems that are heterogeneous in three
dimensions, Flett et al. (2007) illustrated that increased heterogeneity resulted in increased lateral
migration and therefore dissolution. However, increasing heterogeneity also decreased the rate of
residual phase trapping by delaying water imbibition into previously carbon dioxide-filled pore
space. Overall, increased heterogeneity resulted in slower carbon dioxide migration and
UIC Program Class VI Well Area of Review 15
Evaluation and Corrective Action Guidance
-------
decreased accumulation at the confining layer compared to a homogeneous case. Pruess (2008)
showed that for discharge through a fault, decreased fault permeability resulted in delayed
leakage to the surface and an increased maximum leakage rate.
Simulations by Zhou et al. (2008) indicate that patterns of formation pressure increase induced
by carbon dioxide injection are sensitive to permeability. Larger formation permeability values
resulted in less localized pressure increase surrounding the injection well. In addition, larger
confining layer permeability resulted in less pressure buildup throughout the formation due to
pressure dissipation and associated brine leakage.
2.2.2. Relative Permeability and Capillary Pressure
When immiscible fluids (e.g., carbon dioxide, water) are present within the pore spaces of a
geologic formation, the ability for flow of one of those fluids is reduced, due to the blocking
effect of the presence of the other fluid. Note that under any reservoir conditions, ground water
and carbon dioxide will be immiscible, and ground water and hydrocarbons will be immiscible.
Carbon dioxide and hydrocarbons, however, may be miscible or immiscible based on reservoir
conditions. The relative permeability is a scaling factor that represents the reduction in the
capacity for fluid flow due to the presence of other phases in porous media, with a value between
0 and 1. This value is multiplied by the intrinsic permeability of a geologic formation in order to
compute the effective permeability for a fluid in a particular pore space. The relative
permeability of a fluid is based on the properties and amounts of all fluids present within the
system. The greater the amount of pore space occupied by a particular fluid or phase (measured
as fluid saturation), the greater the relative permeability will be for that fluid. Because fluid
saturations change over time and location, relative permeability values typically vary during
model simulations.
The relative permeability for each fluid is typically calculated as a function of fluid saturations at
each location and time within a model. This is achieved via a relative permeability-saturation
function. The relative permeability-saturation function shape is based on properties of the porous
media and fluids present at a particular site. Residual fluid saturation also impacts the shape of
the relative permeability function, and describes the minimum fluid saturation within the porous
medium following immiscible fluid displacement. An example relative permeability-saturation
function is given in Figure 2-2. Note that this example function has been developed for a specific
site (Doughty, 2007) and may not be applicable to other GS sites. Capillary pressure-saturation
relationships (also known as characteristic curves) are also of importance because capillary
pressure gradients provide a driving force for fluid movement under unsaturated conditions.
UIC Program Class VI Well Area of Review 16
Evaluation and Corrective Action Guidance
-------
500
~ 400
n
D.
CO, Saturation (Sg)
1 0.8 0.6 0.4 S 0.2 0
3
us
1
D.
300
200
.
TO
100
Wetting
SBr>0
0 Slmin0.2 0.4 0.6 0.8
Aqueous Saturation (S,)
(a) Typical Capillary Pressure-Saturation Relationships
lU
CO, Saturation (Sg)
0.8 0.6 0.4 S 0.2 0
Aqueous Saturation (S,)
(b) Typical Relative Permeability-Saturation Relationships
for Water and CO,
Note: Dots show values of residual saturation for each curve.
Source: After Doughty (2007), Fig. 2
K - - liquid relative permeability
k :gas relative permeability
SBr residual gas saturation
Figure 2-2: Example Relative Permeability-Saturation and Capillary Pressure-
Saturation Relationships for Water and Carbon Dioxide.
Reproduced with permission of Springer Science + Business Media.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
17
-------
Previous research has shown that model predictions are very sensitive to the shape of the relative
permeability-saturation functions used. The UIC Program Class VI Well Site Characterization
Guidance provides details regarding measurement of relative permeability. Ideally, laboratory
core-analysis techniques will be used for experimental measurement of the relative permeability-
saturation and capillary pressure-saturation functions for a particular site at reservoir conditions,
with carbon dioxide and representative native fluids (e.g., Perrin et al., 2008; Bachu and
Bennion, 2008; Plug and Bruining, 2007). If this is not feasible, relative permeability-saturation
relationships may be estimated from core analysis using other immiscible fluids (e.g., Doughty et
al., 2007). Alternatively, previously reported functions may be used, such as those presented in
Figure 2-2, if the experimental system was very similar to the site conditions for which the
model will be applied. Relative permeability-saturation relationships are also commonly adjusted
during model calibration.
Doughty and Pruess (2004) compared site-specific characteristic curves to "generic" curves at
the Frio, TX GS pilot project site and found that the choice of characteristic curves had a
significant impact on plume size, shape, and mobility. The authors point out that the differences
in plume behavior for different sets of characteristic curves had important implications for
operation and monitoring of the pilot test. Similarly, Doughty et al. (2007) found that model
results were very sensitive to characteristic curve parameters. The authors constrained the value
of characteristic curve parameters by calibration to monitoring data.
Pruess (2008) compared the effect of using three-phase characteristic curves developed for
organic liquid-water-air systems (Stone, 1970) and simple linear characteristic curves. The
choice of characteristic curves was found to have a significant impact on the observed leakage
rate of carbon dioxide through a fault system. The linear characteristic curves resulted in
simulated earlier leakage of carbon dioxide to the surface and lower leakage rates. Use of three-
phase relationships resulted in small fluid permeability at intermediate saturations due to phase-
interference effects.
The impact of using hysteretic versus non-hysteretic characteristic curves has also been
compared. Hysteresis refers to the dependence of the shape of the characteristic curve on the
history of fluid flow within the formation. For example, characteristic curves are often observed
to have a different shape when non-wetting fluids (e.g., supercritical carbon dioxide) are
displacing wetting fluids (e.g., formation water), than when wetting fluids are displacing non-
wetting fluids. Juanes et al. (2006) showed that consideration of hysteresis and capillary trapping
resulted in the carbon dioxide being spatially distributed over a larger area with less
accumulation at the confining layer. Doughty (2007) found that results from simulations with
non-hysteretic curves did a poor job of matching simulations with hysteretic curves in
homogeneous and heterogeneous media. Relative to non-hysteretic cases, simulations including
hysteresis exhibited a more mobile plume leading edge (where there is no water imbibition) and
a slower trailing edge with a significant amount of residual trapping (due to water imbibition).
UIC Program Class VI Well Area of Review 18
Evaluation and Corrective Action Guidance
-------
2.2.3. Injection Rate
The carbon dioxide injection rate at proposed Class VI injection wells is incorporated into the
model by assigning the injection rate parameter at a constant or variable-rate boundary condition,
or by defining various source terms for specific nodes that correspond to injection location.
However, it is important to note that calculated pressure values where injection rates are applied
may need to be monitored to identify flow-controlled or pressure-limited cases. Several
researchers have reported that increasing the carbon dioxide injection rate results in increased
migration rates (e.g., Law and Bachu, 1996; Saripalli and McGrail, 2002; Juanes et al., 2006).
Juanes et al. (2006) considered capillary trapping in highly heterogeneous media, and found that
increased injection rate resulted in more simulated residual trapping due to invasion of carbon
dioxide into a wider range of pore sizes. Therefore, in the long term, increased injection rates
actually decreased the final simulated extent of carbon dioxide migration, as more mass was
immobilized through capillary forces. Pruess (2008) modeled leakage to the ground surface
through a fault system, and simulations indicated that larger injection rates resulted in increased
maximum surface discharge rates relative to injection rates.
2.2.4. Fluid Properties and Equations of State
The density, viscosity, and phase-state of the carbon dioxide injectate, ground water, and any
other fluids that may be present (e.g., hydrocarbons), are important model input parameters.
However, these properties change significantly across the temperature and pressure range that
will be encountered at GS projects, and they are also affected by salinity. The equations of state
describe these fluid properties and the existence of phases as a function of pressure and
temperature; they are used by the model to calculate properties at conditions encountered in the
simulation as they change with location and time. Graphs developed from accepted equations of
state for carbon dioxide are depicted in Figure 2-1. Previous studies have shown that model
results are sensitive to the equations of state used (Pruess et al., 2004; Han and McPherson,
2008).
The composition of the injectate will be reflected in several chemical and physical parameters
assigned to the carbon dioxide fluid in the model simulations. Several studies have evaluated the
impact of common carbon dioxide stream impurities hydrogen sulfide and sulfur dioxide on
geochemical reactions and mineral trapping. Both Knauss et al. (2005) and Xu et al. (2007)
showed that the addition of hydrogen sulfide had little impact, whereas the addition of sulfur
dioxide resulted in a lower pH in the injection zone, less carbon-bearing mineral precipitation,
and more formation-mineral dissolution.
2.2.5. Mass-Transfer Coefficients
Mass transfer coefficients describe the equilibrium concentration of chemical constituents (e.g.,
water, carbon dioxide) between separate phases. For example, the equilibrium aqueous
concentration of carbon dioxide dissolved in ground water in contact with separate-phase (e.g.,
supercritical) carbon dioxide is described by a partitioning coefficient. Other mass-transfer
UIC Program Class VI Well Area of Review 19
Evaluation and Corrective Action Guidance
-------
coefficients describe the distribution of constituents between the gaseous, aqueous, separate-
phase carbon dioxide, and solid phases. For the case of reactive transport modeling, mass-
transfer coefficients describe equilibrium concentration of constituents between mineral and
dissolved phases. Similar to fluid properties, mass-transfer coefficients are in many cases
temperature and pressure dependent. Mass-transfer coefficients may also be dependent on
properties of the formation and fluids present, such as ground water salinity. Reference
documents are available that provide many necessary mass-transfer coefficients (e.g., Green and
Perry, 2008), and several commonly used codes include necessary mass-transfer coefficients
(e.g., TOUGH2-ECO2N; Pruess and Spycher, 2007).
2.2.6. Mineral Precipitation Kinetic Parameters
Mineral precipitation is a subset of reactive transport problems and represents a trapping
mechanism for carbon dioxide as well as a mechanism for permeability modification. As noted
above, the Class VI Rule does not stipulate that reactive transport be considered in AoR
delineation modeling. However, the owner or operator, or the UIC Program Director, may
determine that reactive transport modeling should be considered for a particular project.
Studies accounting for mineral precipitation typically include precipitation kinetic (i.e., rate)
parameters. Although precipitation rates have a large impact on mineral trapping, there is a great
deal of uncertainty related to these parameters (Knauss et al., 2005; Xu et al., 2006).
Furthermore, complex interrelationships exist between the rates of separate mineral species in a
formation. For example, a sensitivity analysis for trapping through dawsonite [NaAl(CO3)(OH)2]
precipitation showed that decreasing dawsonite kinetics resulted in increased formation of other
trapping minerals calcite [CaCOs] and magnesite [MgCOs] (Knauss et al., 2005). Izgec et al.
(2008) showed that changes in formation permeability resulting from mineralization reactions
were very sensitive to kinetic rate parameters. Several modeling studies have indicated that
geochemical equilibrium following injection may not occur for thousands of years (e.g., Xu et
al., 2006; Gaus et al., 2005).
2.2.7. Model Orientation and Gridding Parameters
Numerical modeling requires the developer to define the spatial and temporal domains, grid
spacing and gridding routine, and domain boundaries. These features of the model are typically
designed with an effort to minimize computational demand and therefore processing time.
However, there is potential for erroneous results based on grid features of the model (i.e.,
numerical artifacts), which can mask or enhance the effects of physical processes. A few studies
have focused on evaluating the impacts of numerical artifacts for models of GS.
Doughty and Pruess (2004) tested the impact of varying grid cell sizes for a model of the Frio
formation pilot GS project site in Texas. They found that the overall pattern of plume movement
was similar for different grid sizes, but overly coarse grids were not able to simulate buoyancy-
driven flow within individual sand channels. The authors also observed that the choice of grid
block sizes and gridding routine could result in preferential flow in the grid axis direction and
UIC Program Class VI Well Area of Review 20
Evaluation and Corrective Action Guidance
-------
numerical dispersion. Similarly, Juanes et al. (2006) observed that overly coarse grid block sizes
that did not capture specific migration pathways overestimated carbon dioxide movement and the
amount of capillary trapping. Doughty et al. (2007) note that higher-resolution models are
needed for understanding of near well-bore effects. Yamamoto and Doughty (2009) demonstrate
that grid refinement may have a substantial effect on overall simulated plume extent. Methods
have been developed to establish numerical grids with high resolution in areas of interest (e.g.,
near well bores and fractures), and lower resolution in other areas, such as near the model area
boundaries.
2.3. Computational Approaches
Computational codes consist of the set of interrelated mathematical equations (i.e., governing
equations, constitutive relationships, and equations of state) that are solved simultaneously in
order to predict fluid movement, pressure changes, and other changes, as a function of both
location and time. These equations include complex partial differential equations that cannot be
easily solved, and require complex estimation techniques. In most cases, numerical
approximation methods, discussed below, will be needed to adequately represent the several
physical processes necessary to delineate the AoR and comply with the Class VI Rule.
In certain circumstances, simpler analytic and semi-analytic approaches may be used to
complement numerical efforts in delineating the AoR. As discussed below, analytic and semi-
analytic approaches are not capable of representing several processes and features that are
important for predictions of fluid movement, and they often assume simple geometry and
homogeneity.
2.3.1. Numerical Approaches
Computational models used for practical applications typically consist of a numerical
formulation of the governing equations applied over a spatially discretized model domain that
defines the spatial extent and resolution of the problem (i.e., the model grid). This formulation is
solved by a numerical method, such as finite element or finite difference approximation. The
model grid is partitioned into grid cells, smaller spatial sub-units within the model. Fluid and
heat flow is then solved between adjoining grid cells, while maintaining a mass and energy
balance within the model. Phase changes, mass transfer, and chemical reactions can also be
calculated for phases and constituents within a cell. Each cell can be assigned unique parameter
values for physical properties (i.e., intrinsic permeability, porosity), allowing for three-
dimensional, detailed representations of physical heterogeneity. Numerical models may be used
for steady-state problems (in which injection and withdrawal rates are constant and the solution
is obtained only for infinite time when system variables become constant and the solution
becomes independent of the initial condition) and transient problems (in which injection and
withdrawal rates may vary in time, and the solution is obtained at several discrete times during
the model timeframe while the system variables exhibit change in time and depend upon the
initial condition).
UIC Program Class VI Well Area of Review 21
Evaluation and Corrective Action Guidance
-------
In addition to detailed geologic heterogeneity, numerical models are typically capable of
representing density-driven fluid flow (e.g., the buoyancy of carbon dioxide) and the dissolution
of carbon dioxide into ground water. Numerical models can also represent irreducible (i.e.,
residual) fluid saturations (i.e., the amount of fluid being "trapped" in geologic formation pore
space even after another immiscible fluid has passed through that area), multiphase flow effects,
and the concomitant reduced permeability.
The scale of spatial and temporal discretization of the model affects the accuracy of the solutions
to these numerical formulations. Finer scales of time and space reduce numerical solution error.
However, computational demand increases as the length scale (e.g., grid cell size) and time scale
(e.g., time-step size) decrease, and as additional processes are simulated. Methods have been
developed to mitigate increases in computational demand, while focusing on regions and times
of interest, such as adaptive grid block size (i.e., mesh) refinement. Another possibility is the use
of parallel computing, in which a single problem is broken up and distributed among many
processors (e.g., Zhang et al., 2007).
2.3.2. Analytical, Semi-Analytical, and Hybrid Approaches
Analytical and semi-analytical models may be used to complement numerical modeling efforts in
AoR delineation for Class VI wells. Compared to numerical models, analytical models have
much lower computational requirements and therefore lower processing times. Analytical and
semi-analytical codes may also be particularly useful for assessing the transport of carbon
dioxide through abandoned well bores, which is difficult in numerical models due to the
disparity in spatial scales. Analytical and semi-analytical models also may be used as screening
tools to quickly assess potential storage sites, or as a relatively simple comparative check on
numerical modeling results. Celia and Nordbotten (2009) suggest the use of hybrid numerical-
analytical models for cases where a large-scale numerical model could be combined with local
analytical models (e.g., describing wells), or the use of semi-analytical solutions where analytical
solution is used in the spatial dimension and finite-stepping is used for temporal changes.
However, strictly analytical and semi-analytical models are not able to explicitly account for
detailed physical and chemical characteristics of carbon dioxide injected as required under the
Class VI Rule. They are also not able to simulate other important processes, such as capillary
trapping, or account for varying injection rates or formation heterogeneity. The applicability of
these models is limited to simplified cases where an exact function can be found to satisfy the
governing equation and boundary and initial conditions. For example, in most cases, these
models assume homogenous aquifers (i.e., no variability in physical structure, porosity, or
intrinsic permeability). For most formations this is an unrealistic assumption, and it neglects
preferential fluid movement through heterogeneous channels within geologic formations.
2.4. Model Uncertainty and Sensitivity Analyses
As discussed above, computational models are an approximate representation of reality, and thus
predictions exhibit some degree of uncertainty. Model uncertainty is a result of the uncertainties
UIC Program Class VI Well Area of Review 22
Evaluation and Corrective Action Guidance
-------
related to the underlying science of the governing equations and the uncertainty in the parameter
values input to represent the actual system (USEPA, 2003). Uncertainty in governing equations
and model framework may arise from incomplete scientific data or lack of knowledge, as well as
the necessary simplifications that translate scientific concepts into mathematical equations.
Parameter uncertainty results from poor data quality (e.g., measurement errors, analytical
imprecision, limited sample size), lack of data, and the inherent variability in natural systems.
Model predictions depend largely on the values input for a number of key parameters and thus
may be significantly impacted by incomplete knowledge, or they may be process and scale
dependent. The predictive accuracy of a model improves with improved data quality, increased
data quantity, realistic assumptions that reflect observed conditions and scientific knowledge,
and a modeling domain (extent and resolution) that sufficiently and accurately represents the GS
project.
Significant uncertainty exists in modeling predictions of GS due to the difficulty in determining
the geological formation structure and permeability field throughout the extensive area likely to
be impacted by proposed large injection volumes, a relative lack of data on the behavior of
supercritical carbon dioxide in the subsurface, the drastic changes in transport behavior of carbon
dioxide caused by changes in pressure and/or temperature, and the buoyant nature of carbon
dioxide relative to native formation ground water.
The impact of parameter uncertainty on modeling results can be characterized through a model
sensitivity analysis, which consists of sequentially varying a single parameter in successive
model simulations while keeping all other model features constant. Sensitivity analyses provide
an indication of those modeling parameters that are most sensitive (i.e., that most impact
predictions of carbon dioxide migration, trapping, and pressure changes), and provide guidance
for what parameters to focus on during data collection, parameter estimation, and model
calibration. Accepted guides to environmental modeling (e.g., NRC, 2007) recommend the use
of sensitivity analyses in submission of modeling results. Parameters that are sensitive for a
particular model will be based on case-specific circumstances, and will be identified via
sensitivity analyses.
2.5. Model Calibration
Model calibration consists of using the computational model to simulate a past time period for
which monitoring data are available and adjusting relevant model parameters to reduce
differences between model results and the observed monitoring data. For example, during initial
model development, in-situ pressure data of the injection zone may be available for comparison
to model predictions of pre-injection fluid pressures. After the initiation of carbon dioxide
injection, monitoring data may be available regarding changes in reservoir pressure and fluid
properties that may be used to calibrate the model. It is generally understood that model
calibration reduces model error in prediction of future conditions.
Examples of observed data that may be used for model calibration include carbon dioxide
saturation values and fluid pressures. Model calibration involves adjustment of model
parameters, termed "calibration parameters," in order to minimize the difference between
UIC Program Class VI Well Area of Review 23
Evaluation and Corrective Action Guidance
-------
simulated and observed data values. Calibration parameters for a particular model will be site
specific, but they will commonly include intrinsic permeability and relative permeability-
saturation function parameters. Calibration may involve incorporating additional heterogeneities
or highly-permeable pathways. A case study of model calibration to monitoring data at an early
GS research site, the Frio Brine Pilot in Texas, is provided in Box 2-1 of this guidance document.
Model calibration is never perfect, in that simulated and observed values for a computational
model will not agree exactly. Calibration statistics are often used to characterize the error
difference between model simulations and observed data values. The objective of model
calibration is to minimize the value of the calibration statistics to the extent possible, using
model parameters values consistent with site data or realistic estimates.
Common calibration statistics include the mean error (ME), mean-absolute error (MAE), and the
root-mean squared error (RMSE). The ME is a simple average of the residual error between
observed and simulated values and, therefore, positive values will offset negative values. The
ME therefore provides an indication of the net bias (positive or negative) of the model simulated
values. MAE is similar to the ME, with the important distinction that the sum of the absolute
values of the residuals is calculated, thereby eliminating the offset that occurs by adding positive
and negative values. The MAE, therefore, is always positive and represents the average
difference between observed and simulated values. The RMSE is similar to the MAE, although
negative values of the residual between observed and simulated values are eliminated by
squaring the difference, and then the square root of the sum is determined prior to computing the
average.
Model calibration may be conducted by use of computer programs designed for this purpose
(i.e., automated calibration, see Finsterle, 2004), and/or adjusted manually based on best
professional judgment. In practice, the automated programs can be cumbersome; therefore,
manual parameter adjustment is a more standard practice in the calibration of complex models.
UIC Program Class VI Well Area of Review 24
Evaluation and Corrective Action Guidance
-------
Box 2-1. Model Calibration Case Study: Frio Brine Pilot Project
Pilot projects of GS can provide valuable insight into modeling predictions and monitoring
results comparison. The Frio Brine Pilot Project, in Dayton, TX, is an early experimental
project conducted primarily by researchers at the Texas Bureau of Economic Geology and
Lawrence Berkeley National Laboratory (LBNL). Two carbon dioxide injection and
monitoring experiments (Frio I and Frio II) have been conducted at Frio, supplemented by
numerical modeling. In this text box, separate-phase carbon dioxide data from monitoring
wells, pressure monitoring data, and geophysical monitoring data are presented. These figures
and discussion are taken from Doughty et al. (2007) and Ajo-Franklin et al. (2008).
A geologic schematic of the Frio pilot site is shown in Figure 2-3. For the Frio I pilot, 1,600
metric tons of carbon dioxide were injected over 10 days into a steeply dipping brine-
saturated later at a depth of 1,500 m. For the Frio II pilot, approximately 350 metric tons of
carbon dioxide was injected at a depth of 1,600 m. A number of pre-injection site
characterization, and operational and post-injection monitoring activities were conducted
along with both injections.
Frio I Injection Zone
Frio II Injection Zone
Figure 2-3: Geologic Schematic of Frio Brine Pilot Project. The arrow at top indicates the north direction.
From: Doughty et al. (2007).Reproduced with permission of Springer Science + Business Media.
For the Frio I pilot, a numerical model was calibrated by constraining the value of several
parameters to a variety of monitoring data. Key calibration parameters were determined to be
multi-phase flow parameters that describe the relative permeability-saturation relationship
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
25
-------
Box 2-1. Model Calibration Case Study: Frio Brine Pilot Project, continued
(referred to in the study as the irreducible liquid saturation, Si,) and the van-Genuchten (i.e.,
characteristic curve) parameter (rri). The value of these parameters was constrained by several
types of monitoring data (see Doughty et al., 2007). The researchers focused on calibration to
the arrival time of carbon dioxide at the monitoring well, and pressure monitoring at the
injection and monitoring wells. The arrival time of carbon dioxide at the injection well was
determined based on a reduction of fluid density collected at the observation well using a U-
tube sampling apparatus. The observed arrival time was compared to a series of model runs,
varying Sir and m (Figure 2-4). In addition, the observed pressure increase at both the
monitoring and the injection wells were compared to model predictions (Figure 2-5). Based
on these results, the value of the parameter Sir was constrained to a range of 0.15 to 0.30, and
the value of m was constrained to 0.9.
1200
1100
1003
u
Q
S
il 900
V
-'
st:
8,=
= 0.00, m
= 0.15, m
= 0.15, m
= 0.15, m
.04
= 0.9
= 0.7
= 0.5
* IP-Tube
Time (days)
Figure 2-4: Observed and Modeled Carbon Dioxide Arrival at the Observation Well Based on Change in
Fluid Density. From: Doughty et al. (2007). Reproduced with permission of Springer Science + Business
Media.
1.5
\
I
I "I
e 10 11 12
Figure 2-5: Observed and Modeled Pressure Increase at (a) the Injection Well and (b) the Monitoring
Well (from Doughty et al., 2007). Reproduced with permission of Springer Science + Business Media.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
26
-------
Box 2-1. Model Calibration Case Study: Frio Brine Pilot Project, continued
Frio II used an initial numerical model to predict the evolution of the carbon dioxide plume
overtime. Observed seismic geophysical data of plume migration showed that a thin finger of
carbon dioxide moved further up-dip than initially predicted by the model. The model was
calibrated to the seismic monitoring results by, among other changes, increasing the value of
the intrinsic permeability throughout the model, and increasing the thickness of a high-
permeability channel at the confining zone-injection zone interface. The initial and data-
calibrated model results are shown in Figure 2-6.
Figure 2-6: Comparison of (a) Initial and (b) Post-Calibration Model Predictions of Carbon Dioxide
Plume Evolution. From: Ajo-Franklin et al. (2008).
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
27
-------
2.6. Existing Codes used for Development of GS Models
A wide variety of modeling exercises have been reported in the peer-reviewed literature for GS
and have been reviewed previously (Schnaar and Digiulio, 2009). Several computational codes
have been developed for multiphase flow and transport problems, and a number of these codes
are publicly or commercially available for the owners or operators of a GS project to use in AoR
delineation modeling. Codes reported in the literature used for modeling of GS include
petroleum reservoir codes (STARS, Law and Bachu, 1996; GEM, Kumar et al., 2004; ECLIPSE,
Zhou et al., 2004; Juanes et al., 2006; CHEARS, Flett et al., 2007) and codes that have been
developed at U.S. Department of Energy (DOE) national laboratories for a range of multiphase
flow and transport problems (STOMP, CRUNCH, Knauss et al., 2005; TOUGH-series, Finsterle,
2004; Xu et al., 2006; Doughty and Pruess, 2004; Doughty, 2007). Additionally, DOE provides a
summary of available models that have been used to model processes associated with injection
for GS at the Regional Carbon Sequestration Partnership (RCSP) project sites. The document
presents the types of data that are needed for various models and how to obtain such information
(NETL, 2011). These codes vary not only in the physical processes considered, but also in
numerical techniques such as the spatial discretization method, iteration approach, and gridding
routines. The codes mentioned above are provided as examples that may be used for GS
modeling, and the lists given are not meant to include all available codes, or to suggest
preference for certain codes over others.
Codes used for modeling GS consider multiphase flow of carbon dioxide in supercritical, liquid,
and gaseous phases including miscible and immiscible displacement, dissolution of carbon
dioxide in ground water, density-driven flow, and flow of ground water as impacted by injection.
Available codes may also be further categorized based on their ability to consider, or to be
adjusted to consider, complex three-dimensionally heterogeneous formations, residual phase
trapping and characteristic-curve hysteresis, mineral precipitation/dissolution reactions and
subsequent mineral phase trapping and leaching of heavy metals, carbon dioxide sorption in
coal-bed methane problems, and leakage through abandoned well bores. Models based on the
TOUGH-series codes have been widely reported in the literature and are capable of considering
three-dimensional heterogeneous formations, carbon dioxide dissolution, residual phase trapping
and characteristic-curve hysteresis, coupled fluid flow and geomechanical processes, and mineral
precipitation (e.g., Finsterle, 2004; Doughty, 2007; Xu et al., 2006; Rutqvist et al., 2008).
Several codes were compared for identical GS problems in an LBNL study (Pruess et al., 2004)
in order to evaluate code comparability. Ten research groups representing six countries
participated in the study. The codes evaluated included TOUGH-series codes (LBNL, CSIRO
Petroleum, Industrial Research Limited), ECLIPSE 300 (Los Alamos National Laboratory), and
STOMP (Pacific Northwest National Laboratory), among others. The problems considered
varied in complexity and included mixture of gases in an open system, radial flow from an
injection well, discharge along a fault zone, injection with mineral trapping, and injection with
enhanced-oil recovery. For the most part, model results for the different codes were found to be
in good agreement. Most discrepancies were traced to differences in the calculation of fluid
properties (e.g., viscosity). These results emphasize the need for accurate descriptors of carbon
dioxide transport properties and equations of state.
UIC Program Class VI Well Area of Review 28
Evaluation and Corrective Action Guidance
-------
The use of proprietary codes (i.e., codes not available for free to the general public) may prevent
full evaluation of model results (e.g., NRC, 2007). There are several aspects of a model that can
be proprietary, and some may be more important than others for computational model
evaluation. For example, use of a proprietary user interface with a publicly available code may
not present a significant problem. Several popular codes in the petroleum-reservoir engineering
discipline are proprietary (e.g., ECLIPSE). However, these codes have been used in peer-
reviewed studies to model GS, and operators of particular GS sites may prefer to use these codes
as they have previous experience with them. As discussed below, when using a proprietary
model for AoR delineation, site operators of GS projects are encouraged to clearly disclose to the
UIC Program Director the code assumptions and, if necessary, governing equations and
equations of state with the permit application.
UIC Program Class VI Well Area of Review 29
Evaluation and Corrective Action Guidance
-------
3. AoR Delineation Using Computational Models
Determination of the AoR for proposed Class VI wells will consist of data collection and
compilation, development of the site computational model, delineation of the AoR based on
model results, and submission of the model results and AoR delineation to the UIC Program
Director with the Class VI permit application. The AoR and Corrective Action Plan must
describe how the owner or operator plans to conduct these activities and is subject to UIC
Program Director approval [40 CFR 146.84(b)].
The AoR delineation model must be submitted with the Class VI permit applicationi.e., with
the proposed AoR and Corrective Action Plan, as required at 40 CFR 146.82(a)(13)and the
modeled AoR will be finalized after all site data are collected and pre-injection testing is
complete, per 40 CFR 146.82(c)(l). Therefore, the submittal, evaluation, and approval of the
AoR, as part of the AoR and Corrective Action Plan, may be an iterative process, involving
multiple drafts, until all the information required is submitted at the appropriate level of detail as
determined by the UIC Program Director. See the UIC Program Class VI Well Project Plan
Development Guidance for more information on the AoR and Corrective Action Plan. Also see
the UIC Program Class VI Implementation Manual for additional information on how permitting
authorities will review the information submitted by owners or operators.
This section describes the AoR delineation process and provides several quantitative examples
revolving around a hypothetical GS site. EPA recommends that model development in all cases
be conducted by a professional expert with the understanding of multiphase flow processes and
experience with application of sophisticated computational models.
3.1. AoR Delineation Class VI Rule Requirements
The following Class VI Rule requirements pertain to AoR delineation:
40 CFR 146.84(a): The AoR is the region surrounding the GS project where USDWs
may be endangered by the injection activity. The AoR is delineated using computational
modeling that accounts for the physical and chemical properties of all phases of the
injected carbon dioxide stream and is based on available site characterization,
monitoring, and operational data.
40 CFR 146.84(c)(l): Owners or operators of Class VI wells must predict, using existing
site characterization, monitoring and operational data, and computational modeling, the
projected lateral and vertical migration of the carbon dioxide plume and formation fluids
in the subsurface from the commencement of injection activities until the plume
movement ceases, until pressure differentials sufficient to cause the movement of injected
fluids or formation fluids into a USDW are no longer present, or until the end of a fixed
time period as determined by the UIC Program Director. The model must:
(i) Be based on detailed geologic data collected to characterize the injection zone(s),
confining zone(s), and any additional zones; and anticipated operating data,
UIC Program Class VI Well Area of Review 30
Evaluation and Corrective Action Guidance
-------
including injection pressures, rates, and total volumes over the proposed life of
the GS project;
(ii) Take into account any geologic heterogeneities, other discontinuities, data quality,
and their possible impact on model predictions; and
(iii) Consider potential migration through faults, fractures, and artificial penetrations.
3.2. Data Collection and Compilation
Computational modeling utilizes the required site characterization data for a proposed Class VI
injection well site and applies scientifically accepted principles to estimate the carbon dioxide
plume and pressure front migration. The extent to which site and operational conditions are
realistically represented determines the validity of the resulting model predictions. Site
characterization data inform model parameterization and the development of the model and,
therefore, adequate data collection, analysis, and compilation are integral components of model
development. Table 2-1 of this guidance provides a summary of important model parameters,
many of which are determined based on site characterization data.
A variety of site characterization data are required to be collected for proposed GS projects [40
CFR 146.82 and 146.83]. These data are required to verify that the proposed injection zone at the
characterized site has adequate injectivity to accept the injected carbon dioxide at the proposed
rate and adequate volume to store the injectate over the lifetime of the project. Furthermore, site
characterization data verify that suitable confining zone(s) are present to restrict the upwards
movement of carbon dioxide. Additional features of the site, such as baseline geochemistry and
pre-injection fluid pressures, inform the interpretation of future monitoring results and support
reactive transport modeling if it is chosen to be used. As discussed below, much of the site
characterization data collected at the proposed Class VI injection well site are also necessary to
inform computational model development and AoR delineation. Site characterization
requirements and methods are discussed in more detail in the UIC Program Class VI Well Site
Characterization Guidance.
3.2.1. Site Hydrogeology
Regional and site-specific geology provide the foundations of the computational model used to
delineate the AoR. This includes site stratigraphy, including formation elevation and thickness,
as presented in cross sections and/or topographic maps. It is recommended that any data
regarding structural geology (including folding, and fracture and fault systems) be identified and
used when creating the computational model. For each geologic formation at the proposed
injection site, hydrogeologic information, including initial fluid pressure, horizontal and vertical
gradients, and ground water flow direction and velocity, should be considered. Other important
characteristics include intrinsic permeability and porosity of all formations ranging from the
uppermost USDW to beneath the injection zone. Where injection depth waiver applications are
considered, EPA recommends that these parameters be determined for all formations down to
and including the first USDW below the injection zone. See the UIC Program Class VI Well
Injection Depth Waivers Guidance for additional information. EPA recommends that the
UIC Program Class VI Well Area of Review 31
Evaluation and Corrective Action Guidance
-------
heterogeneity of these characteristics within each formation also be evaluated. Data regarding the
heterogeneity of these parameters are of particular importance in representing the injection and
confining zone(s). The Class VI Rule requires that AoR computational modeling take into
account any geologic heterogeneities and other discontinuities [40 CFR 146.84(c)(l)(ii)].
Thorough characterization of multiphase flow parameters is also recommended to properly
inform the computational modeling. These include parameters describing the capillary pressure-
saturation and relative permeability-saturation relationships of each formation, with the injection
and confining zones being of particular importance. See Figure 2-2 of this guidance for more
information. EPA recommends that accepted formulations of these relationships be defined that
are as specific to the site and fluids of interest (e.g., brine, carbon dioxide) as possible.
The quantity of data used to inform model development is recommended to be, at least, based on
the Class VI Rule site characterization requirements, as discussed in the UIC Program Class VI
Well Site Characterization Guidance. For pertinent data types, as discussed above, all data
collected to comply with site characterization requirements may be considered in the AoR
delineation. Furthermore, EPA recommends that any additional pertinent data available in the
vicinity of the site, for example from the U.S. Geological Survey (USGS) or other sources, also
be included in model development.
Additionally, EPA recommends that the lateral and vertical extents of all formations predicted to
exhibit contact with supercritical carbon dioxide or elevated pressure over the lifetime of the
proposed GS project be characterized for hydrogeologic properties. This may be an iterative
process because initial model estimates of plume and pressure front migration may indicate
further migration than previously assumed. In these cases, some additional site characterization
in these regions may be requested by the UIC Program Director before a permit is approved.
EPA recommends that adequate data be collected to reasonably estimate site heterogeneity.
Collection of sufficient data is always a challenge in geologic studies, and this is compounded by
the large areas that may be impacted by GS projects. Use of geophysical site characterization
techniques may reduce the burden of site characterization over large areas. See the UIC Program
Class VI Well Site Characterization Guidance for more information on using geophysical
methods to assist with collecting the required site characterization data for a Class VI injection
well permit application.
3.2.2. Operational Data
The Class VI Rule requires that the AoR computational modeling for a Class VI injection well be
based on existing or proposed operational data including injection pressures, rates, and total
volumes over the lifetime of the GS project [40 CFR 146.84(c)(l)(i)]. EPA recommends that
operational data also include the location and number of injection wells, and the injection well
construction details (e.g., total depth, perforated interval). In the case of GS projects with
multiple Class VI injection wells, it is important to note that each Class VI well is required to be
permitted separately, as area permits are not allowed [40 CFR 144.33(a)(5)]. However, EPA
strongly encourages potential Class VI injection well owners or operators to account for all
UIC Program Class VI Well Area of Review 32
Evaluation and Corrective Action Guidance
-------
injection wells associated with the proposed project, or any other injection or extraction wells in
the area, when developing the AoR model. EPA recommends that a single AoR delineation
model be used for all Class VI injection wells for a single GS project, and that the model include
the influences of all relevant wells. EPA also recommends that overlapping pressure
perturbations be evaluated for a given basin or hydraulically connected formations to determine
any combined risk to USDWs. The owner or operator may consult the UIC Program Director
regarding any existing or planned projects in the vicinity of the proposed well.
3.3. Model Development
Once adequate data are collected, model development consists of the formation of a conceptual
site model, design of the mathematical framework and grid, and parameterization (i.e.,
determination of input parameter values) (USEPA, 2003). The model is then executed to provide
predictions of fluid movement and pressure perturbations during the lifetime of the project.
A Note Regarding Hypothetical Examples
Several informative boxes are included within this guidance that provide examples of the
AoR delineation and reevaluation process based on a hypothetical site (Box 3-1, Box 3-2,
Box 5-1, and Box 5-2). The hypothetical site presented is not intended to be representative of
all GS projects. Assumptions and methods used in hypothetical examples may not be valid in
all cases. A length scale has not been included on hypothetical site figures, such that an
allowable size of the AoR, or distance between wells, is not unintentionally inferred from the
figures.
3.3.1. Conceptual Model of the Proposed Injection Site
A conceptual site model is a schematic representation of the proposed GS project, including all
major geologic elements present in the flow system and any relevant physical processes. In the
delineation process, the conceptual model is translated mathematically into a numerical model to
be solved for pressure and saturation. The conceptual site model is informed primarily by the
collected site characterization data and the proposed operational conditions, such as well-field
configuration and injection rates. EPA recommends that descriptions of the conceptual site
model present a clear statement and description of each element of the site, as well as any
assumptions and hypotheses related to the proposed injection site and the reasoning behind them
(e.g., lab experiments, empirical data, or peer-reviewed literature). The conceptual site model
also identifies the modeling region in three dimensions. Geologic stratigraphy, any other relevant
geologic features, all physical processes that will impact migration of carbon dioxide and ground
water, chemical species of interest, location of USDWs and potential conduits, conditions at site
boundaries that may inform model boundary conditions, and areas of sparse site characterization
data are also identified in the conceptual site model. See Box 3-1 for more information about the
conceptual site model.
UIC Program Class VI Well Area of Review 33
Evaluation and Corrective Action Guidance
-------
Box 3-1. Hypothetical Example of a Conceptual Site Model
A conceptual site model describes the general features of the anticipated Class VI project,
using one or several schematics and diagrams. EPA recommends that schematics be used to
show the general project orientation, both at the surface and at depth, important site features,
and known processes that will impact plume and pressure front evolution at the site. Report
text accompanying the conceptual site model schematic describes the relevant features at the
site. A hypothetical example conceptual site model schematic is shown in Figure 3-1, and the
example accompanying text is below.
For this hypothetical project, three injection wells are planned to inject a total of two million
tons of carbon dioxide per year for 30 years. The source of carbon dioxide is a coal-fired
power plant located approximately 200 miles to the north of the injection site. The injectate
will be supplied via pipeline to the site, delivered to a surface facility, and then supplied
separately to each of the injection wells. The injectate will be greater than 99% pure carbon
dioxide at all times, containing trace amounts of sulfur dioxide and nitrogen oxides.
" "' "" v VA
> x^-*1-*-
Schematic of Example
Fracture System
Pore-scale
Schematic of
Trapping Mechanisms
Wofe: Figure nor to scsto
Source: Daniel B. Steptiens & Associates, toe
Figure 3-1: Hypothetical Conceptual Site Model for Geologic Sequestration.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
34
-------
Box 3-1. Hypothetical Example of a Conceptual Site Model, continued
Injection will occur into a saline formation (Unit N), with a measured salinity of 50,000 mg/L,
at a depth of approximately 1,800 meters below ground surface. The formation dips slightly,
and carbon dioxide and pressure front movement are expected to be generally greater in the
up-dip direction. The permeability of the injection zone has been measured to range from 1 to
50 mD, with lower permeabilities generally at higher elevations and at the contact between
the confining and injection zones.
A shale unit, at least 20 meters thick throughout the vicinity, serves as the primary confining
unit (Unit M). The depth of the lowermost USDW (Unit C) varies somewhat throughout the
vicinity, but it is generally from 200 to 500 meters below ground surface. Intervening layers
of sand, shale, and clay units exist between the confining layer and lowermost USDW. A
secondary confining zone (Unit K) has been identified.
The majority of carbon dioxide is expected to migrate upwards through the zones of higher
permeability until encountering lower permeability zones within the injection zone, or the
injection zone/confining zone contact, and be physically trapped. Capillary trapping, mineral
trapping, and dissolution of carbon dioxide into ground water will also occur; however, at this
point the rate and total amount expected to be sequestered via the different mechanisms has
not been quantified. Currently, ground water in all subsurface formations flows generally to
the west. It is expected that pressure increases within the injection zone induced by the project
will cause ground water to generally flow radially away from the injection wells.
Two relevant geologic zones with a concentration of fractures are located in the vicinity of the
project, as shown on Figure 3-1. Fractures exist primarily in Unit K, the secondary confining
unit, but also are potentially identified in the primary confining zone, Unit M. Geologic
studies of the fractures and preliminary modeling have indicated that due to the orientation
and fracture widths, they will not serve as a leakage pathway during carbon dioxide injection.
However, these two relevant geologic zones will likely be locations for enhanced monitoring
during the lifetime of the project, based on consultation with the UIC Program Director.
A former oil and gas field is located to the northeast of the project. Further analysis, including
modeling, is used to determine if carbon dioxide may migrate into this area. If migration is
detected, enhanced corrective action and monitoring will occur within the area of the former
oil and gas field in consultation with the UIC Program Director. A fault zone exists far to the
east of the proposed site. Carbon dioxide is not expected to migrate as far as the fault zone.
However, this feature may also be further evaluated over the course of project, along with the
potential for brine migration through the fault zone as a result of pressure buildup in the
formation.
UIC Program Class VI Well Area of Review 35
Evaluation and Corrective Action Guidance
-------
3.3.2. Determination of Physical Processes to be Included in the Computational
Model
Prior to developing the computational model for a proposed Class VI injection well AoR
delineation, the owner or operator will need to determine what physical processes will be
considered in the computational model. This determination is based on the most significant
processes identified in the conceptual model, as well as those processes that can be realistically
included in the computational model. At a minimum, the Class VI Rule requires that the model
include multiphase flow of carbon dioxide and formation fluids [40 CFR 146.84(a) and (c)(l)].
Additional processes may be necessary for certain projects. For example, reactive transport could
be relevant if permeability and/or porosity are predicted, based on previous testing, to change as
a result of precipitation/dissolution reactions. In addition, geomechanical processes could be
relevant if pressure and stress may change hydrogeologic properties. If the aqueous carbon
dioxide plume is a potential risk factor, carbon dioxide dissolution into ground water may also be
considered in the AoR delineation model.
For some model applications, including reactive transport and geomechanical processes may be
impractical. Complications can arise from increases in computational demands (i.e., extremely
long computer processing times), lack of meaningful data on mineral precipitation/dissolution
kinetics, or the inability of preferred computational code. Furthermore, including these processes
may be unnecessary in many cases because the impact on plume and pressure front migration
may be relatively minor. The Class VI Rule does not require including reactive transport and
geomechanical processes in the AoR delineation modeling. However, the UIC Program Director
may request that the owner or operator include these additional processes in AoR delineation
modeling in cases where doing so would improve the understanding of plume and pressure
migration for the project.
3.3.3. Computational Model Design
After a conceptual site model has been developed, and the processes that will be considered have
been determined, the next step is to develop the site computational model. This includes the
determination of an appropriate computational code, and parameterization (i.e., populating the
code with the selected site-specific parameters) in order to develop the model.
3.3.3.1. Computational Code Determination
To create the computational model, EPA recommends that a code be used that includes routines
for the relevant physical processes at the site based on peer-reviewed theory and equations,
including equations of state for carbon dioxide and other chemical species of interest. EPA
recommends that the code also include accurate mass-transfer coefficients, including solubility
of carbon dioxide, as a function of primary thermodynamic variables (e.g., temperature, pressure,
phase saturations). If using an independently developed or untested code, EPA also recommends
that the developer model test cases found in the literature to verify the accuracy of the model
before submitting the Class VI injection well permit application to the UIC Program Director
(e.g., see Pruess et al., 2004).
UIC Program Class VI Well Area of Review 36
Evaluation and Corrective Action Guidance
-------
3.3.3.2. Model Spatial Extent, Discretization, and Boundary
Conditions
The computational model will be designed by determining the spatial boundaries of the problem
and spatial discretization. It is recommended that lateral grid spacing be fine enough to resolve
heterogeneities, as discussed above (e.g., Doughty and Pruess, 2004). Vertically, the model is
recommended to include the injection zone, with sufficient vertical resolution to properly
account for buoyant plume migration, and a sufficient section of the primary confining zone to
demonstrate to the satisfaction of the UIC Program Director that no leakage is expected to occur
through the confining zone. In some cases, the UIC Program Director may require that additional
zones be included in the computational model, such as overlying USDWs. Inclusion of
additional zones may allow for estimation of vertical leakage rates and pressure changes above
the primary confining zone, which may aid in performing risk assessments and designing the
Class VI testing and monitoring program.
Boundary conditions are typically based on hydrogeologic conditions in locations corresponding
to the edges of the model domain where the domain extends beyond the pressure front and/or
plume. Model testing may be conducted to ensure that grid spacing, gridding routine, and
boundary conditions do not result in numerical artifacts that impact the model results. If results
of such testing indicate artificial impacts, then adjustment of the model may be necessary prior to
running the model for a proposed Class VI injection well AoR delineation.
3.3.3.3. Model Timeframe
The Class VI Rule requires that the model used to delineate the AoR for a proposed Class VI
injection well be run from the commencement of injection activities until the plume movement
ceases, until pressure differentials sufficient to cause the movement of injected fluids or
formation fluids into a USDW are no longer present, or until the end of a fixed time period as
determined by the UIC Program Director [40 CFR 146.84(c)(l)]. In order to meet these
conditions, it may be necessary for the model simulation of the GS project to extend for several
hundred or thousands of years (e.g., Flett et al., 2007).
3.3.3.4. Parameterization
Parameterization is the final step in the initial development of the computational model, and it
consists of populating the computational code with the selected site-specific parameters. Key
parameters include formation intrinsic permeability, porosity, phase-partitioning coefficients,
and relative permeability-saturation parameters. Parameter values are based on the site-specific
data as much as possible, but may also be based on values and relationships from the scientific
literature. Geostatistical techniques can also be used to create a representation of realistic, three-
dimensionally heterogeneous conditions in the subsurface. See Section 2.2 of this guidance for
more information on model parameters. In some cases, a reasonable range of parameter values
may be identified for the purposes of later sensitivity analyses.
UIC Program Class VI Well Area of Review 37
Evaluation and Corrective Action Guidance
-------
3.3.4. Executing the Computational Model
The computational model is executed (i.e., solved) after parameterization, and this consists of
using the code to calculate phase saturations and composition, fluid pressures, and other system
aspects within the model domain for each point in time and space separated by specified
intervals (i.e., time step, grid spacing). Model results are typically text files that contain modeled
data for each grid cell, during each time step. In some cases, the model results will need to be
post-processed following execution of the model before they can be easily visualized and
interpreted. For example, model results may need to be transformed to produce site coordinates.
Model results of particular interest for Class VI injection well AoR delineation include
estimation of the extent of the separate-phase carbon dioxide plume migration and changes in
fluid pressures within the injection zone over time. See Section 3.4 for more information on AoR
delineation.
The use of an a priori AoR delineation based on computational modeling predictions highlights
the need for uncertainty and sensitivity analyses for the initial prediction. Conservative
predictions will be needed prior to the commencement of injection and the availability of any
site-specific data on carbon dioxide migration paths and rates. EPA recommends conducting
sensitivity analyses as the principal evaluation tool for characterizing the most and least
important sources of error in computational models (USEPA, 2003). Based on these results,
maximum-risk scenario simulations can be conducted considering plume extent and pressure
perturbation predictions that account for uncertainties in the model.
3.4. AoR Delineation Based on Model Results
The planned or predicted AoR submitted with the permit application for a proposed Class VI
injection well is required to be based on a delineation of the area where the GS project may cause
endangerment of USDWs, which in turn is required to be based on the results of computational
modeling [40 CFR 146.84(a) and 40 CFR 146.84(c)(l)]. The boundaries of the AoR are based on
simulated predictions of the extent of the separate-phase (i.e., supercritical, liquid, or gaseous)
plume and pressure front. As such, EPA recommends that the AoR encompass the maximum
extent of the separate-phase plume or pressure front over the lifetime of the project and entire
timeframe of the model simulations. The pressure front, as described below, is the extent of
pressure increase of sufficient magnitude to force fluids from the injection zone into the
formation matrix of a USDW.
In the case of GS projects with multiple Class VI injection wells, the owner or operator must
apply for and obtain a Class VI injection well permit for each individual well [40 CFR
144.33(a)(5)]. However, as discussed previously, a single AoR modeling exercise may be
conducted for all wells within a single project at the discretion of the UIC Program Director. In
all cases, EPA recommends that the AoR delineation boundaries for the cumulative GS project
be based on modeling that accounts for the anticipated injection rates from all planned Class VI
injection wells.
UIC Program Class VI Well Area of Review 38
Evaluation and Corrective Action Guidance
-------
Box 3-2 provides a hypothetical example of an AoR delineation based on computational
modeling results, including the calculation of the threshold pressure that defines the pressure
front. The determination of the pressure front in Box 3-2 (Step 2) is consistent with existing
standard practices for other well classes of the UIC Program (e.g., Thornhill et al., 1982) and is
applicable to any Class VI injection well for which, prior to injection, the injection zone is under-
pressurized compared to the lowermost USDW (i.e., Method 1, Section 3.4.1). Determination of
the pressure front is discussed in more detail in Section 3.4.1.
3.4.1. Determination of Threshold Pressure Front
The pressure front may be defined as the minimum pressure within the injection zone necessary
to cause fluid flow from the injection zone into the formation matrix of the USDW through a
hypothetical conduit (i.e., artificial penetration) that is perforated in both intervals. Several
methods, as described below, are available to estimate the value of the pressure front, and are
based on various assumptions. The owner or operator is encouraged to consult with the UIC
Program Director to determine which method is appropriate for the proposed GS project. For
instance, if an existing aquifer exemption is proposed to be expanded for a GS project, the
pressure front may be determined based upon the pressure increase necessary for formation
fluids to be displaced into portions of aquifer that are not exempted.
Method 1. Pressure front based on bringing injection zone and USDW to equivalent hydraulic
heads (applicable to under -pressurized case only).
As stated by Thornhill et al. (1982), the pressure-front component of the AoR is "the area around
an injection well where, during injection, the [hydraulic] head of the formation fluid in the
injection zone is equal to or greater than the [hydraulic] head of USDWs." Defined this way, the
pressure-front (P;;f) may be calculated by the following equation:
where Pu is the initial fluid pressure in the USDW, p; is the injection-zone fluid density, g is the
acceleration due to gravity, zu is the representative elevation of the USDW, and z; is the
representative elevation of the injection zone. Similarly, the increase in pressure that may be
sustained in the injection zone (AP;;f) is given by:
APif = PU + Pig (zu - Zi) - Pi [Eq-2]
where P; is the initial pressure in the injection zone. Eq-1 and Eq-2 are subject to the assumption
that the hypothetical open borehole is perforated exclusively within the injection zone and
USDW.
A positive value of AP; f (Eq-2) corresponds to an injection reservoir that is under-pressurized
relative to the USDW and can accommodate an increase in pressure equal to AP;^ prior to
potential fluid migration into the drinking water reservoir. A AP;;f value of zero corresponds to
the hydrostatic case, and a negative value of AP;,f relates to a situation where the injection zone is
UIC Program Class VI Well Area of Review 39
Evaluation and Corrective Action Guidance
-------
already over-pressurized and thus subject to potential fluid leakage from the injection reservoir
to the drinking water aquifer even prior to the planned GS project. Eq-1 and Eq-2 are only
applicable for calculating the allowable pressure increase for the under-pressurized case (i.e.,
positive value of AP;;f). Alternative methods may be applicable for the hydrostatic case or over-
pressurized cases (see below).
EPA recommends that Eq-1 and Eq-2 be applied using values of pressure and fluid density (i.e.,
Pu, P;, and pi) based on direct measurement of fluid properties in the direct vicinity of the
proposed project (i.e., see the UIC Program Class VI Well Site Characterization Guidance}.
Notably, the results of Eq-1 and Eq-2 are sensitive to the injection-zone fluid density (p;), which
is influenced by the pressure, temperature, and salinity of the injection zone. Salinity, pressure,
and temperature tend to increase with depth below the ground surface. If site-specific fluid
density values at reservoir conditions are not available, injection zone fluid density may be
estimated based on measured salinity, temperature, and pressure. Figure 3-2 presents fluid
density as a function of depth below ground surface for several different salinities, based on the
method of Peng and Robinson (1976).
Fluid Density (kg/m3)
960 970 980 990 1000 1010 1020 1030 1040 1050 1060 1070 1080 1090 1100
6,000 mg/L
10,000 mg/L
60,000 mg/L
100,000 mg/L
120,000 mg/L
12000
Figure 3-2: Fluid Density Functions for Varying Salinities.
Note also that, in using this method, P;,f is a function of the fluid density of the injection zone,
the elevation of both formations, and the fluid pressure within the USDW. To the extent that
these parameters vary spatially in the vicinity of the project, the value of P;;f may also vary
throughout the region of the AoR.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
40
-------
Method 2. Pressure front based on displacing fluid initially present in the borehole (applicable
to hydrostatic case only).
Under hydrostatic conditions, a pressure increase within the injection zone may be allowable due
to the fact that water entering a hypothetical borehole from the injection zone will be more dense
than the fluid initially present in that borehole. Fluid from the injection zone will displace the
fluid in the borehole, which will flow into the USDW. However, below a calculated threshold
pressure, a new pressure equilibrium will be established, and fluid from the injection zone will
not intrude into the USDW.
As given by Nicot et al. (2008) and Bandilla et al. (2012), assuming (1) hydrostatic conditions
and (2) initially linearly varying densities in the borehole and constant density once the injection-
zone fluid is lifted to the top of the borehole (i.e., uniform density approach), the threshold
pressure increase (APC) may be calculated:
- - a F (7 7-~\2 TFn ^1
- -, y s vzu zij i^q jj
where ^ is a linear coefficient defined by:
r _ Pi-Pu
and where pu is the fluid density of the USDW.
Nicot et al. (2008) also present a solution subject to the assumption of linearly varying densities
in the borehole both initially and when the injection zone fluid is lifted to the top of the borehole
(i.e., equilibrium approach), rather than the uniform density approach (assumption #2) used to
derive Eq-3 (see Nicot, 2008 and Bandilla, 2012). Birkholzer et al. (2011) state that the value of
APC calculated using the uniform density approach (Eq-3) may be less precise than the
equilibrium approach, but it is easier to apply and also more conservative for protection of the
USDW.
At pressure increase less than APC, the fluid originally present in the borehole will leak into the
USDW. This fluid leakage from the borehole may be acceptable and not cause degradation of
water quality within the USDW, as the volume of water in the borehole may be minor and
quickly diluted by water in the surrounding aquifer.
Calculation of the allowable threshold pressure increase using these methods (Eq-3 and Eq-4) is
applicable only to the hydrostatic case. In some instances, site-specific fluid pressure and density
measurements may not be available at the time of preparing the permit application and initial
AoR delineation in order to evaluate if the injection reservoir is over- or under-pressurized. If
warranted by previous site knowledge (i.e., no previous large-scale fluid withdrawal or injection
from the injection zone), it may be acceptable to initially assume hydrostatic conditions. The
owner or operator may choose to use these methods (Eq-3 and Eq-4) for an initial estimate of the
UIC Program Class VI Well Area of Review 41
Evaluation and Corrective Action Guidance
-------
threshold pressure allowable for delineation of the AoR. However, once site-specific data are
available for the proposed injection zone following construction of injection or monitoring wells,
EPA recommends reevaluation of the hydrostatic assumption. If the reservoir is under-
pressurized, the allowable threshold pressure increase may be greater than calculated using Eq-3
(see Method 1, above). However, if the injection reservoir is actually over-pressurized, the
allowable pressure increase will be less than calculated using Eq-3.
Methods for over-pressurized cases.
In some instances, the desired injection zone may already be over-pressurized relative to the
USDW prior to the injection project (i.e., AP;,f value is negative using Eq-2). In this situation,
fluid leakage would occur from the injection zone to the USDW through a borehole perforated
within both zones even prior to commencing injection. Additional pressure increase within the
injection zone owing to the injection associated with the GS project may increase fluid leakage
rates. Determination of the allowable pressure increase to be used in AoR delineation for the
over-pressurized case may require more sophisticated methods than the analytical equations
described above for Methods 1 and 2.
Possible methods to estimate an acceptable pressure increase for over-pressurized reservoirs
include:
1. Using similar methods as described in Nicot et al. (2008), some over-pressurization
within the injection zone may be allowable without causing sustained fluid leakage,
owing to the density differential between the injection zone and USDW. If the value of
APC using Eq-3 is greater than the absolute value of AP;;f using Eq-2, the difference in
magnitude between the two may be used as an estimate of the allowable pressure
increase, subject to the assumptions used to derive Eq-3 (see Method 2, above). To date,
peer-reviewed research papers on this topic have developed analytical solutions only for
the hydrostatic case. Future publications may address the initially over-pressurized case
and, if so, these methods may be used to calculate an allowable pressure increase in an
over-pressurized reservoir.
2. A multiphase numerical model may be designed to model leakage through a single well
bore, or through multiple well bores in the formation (see e.g., Birkholzer et al., 2011).
Additional pressure increases up to a certain point within an already over-pressurized
injection zone may not cause an appreciable increase in fluid leakage rates through a
hypothetical borehole. A sensitivity analysis may be conducted to bound the modeled
leakage rates.
3. In conjunction with item #2 above, numerical or analytic ground water modeling may be
conducted for the USDW to estimate how additional fluid leakage caused by the injection
project is diluted within the USDW and attenuated. Dilution of fluid leakage from a
borehole is impacted by the natural background flow rate of water within the USDW,
which is turn a function of the hydraulic gradient, aquifer thickness, and hydraulic
conductivity. An additional pressure increase may be allowable if it can be demonstrated
UIC Program Class VI Well Area of Review 42
Evaluation and Corrective Action Guidance
-------
to the UIC Program Director that negligible degradation of the USDW would result from
increased fluid leakage rates.
Box 3-2. Hypothetical Example of an AoR Delineation
The AoR is based on the results of computational modeling and encompasses the predicted
maximum extent of the separate-phase plume or pressure front over the lifetime of the project.
The pressure front is defined as the pressure, within
the injection zone, great enough to force
fluids from within the injection zone through a hypothetical open conduit into any overlying
USDW. This box provides a hypothetical example of an AoR delineation using a stepwise
approach. The example scenario is based on the conceptual site model described above (see
Box 3-1). First, the threshold pressure that defines the pressure front is determined. Next,
maps showing the maximum extent of the plume and pressure front are overlaid and the AoR
is delineated.
Step 1 . Determine the Pressure Front
A cross-sectional schematic of the hypothetical scenario is shown in Figure 3-3, which also
presents values for fluid density, and pressure (units
for each formation.
Elevation Sjaja
(m amsl)
Hydraulic head, USDW (hj .1
1690-
Lowermost USDW (Unit C) e=
1B35m
>-~z.:1615m i
1540-
^_Uj_
^HH Primary Confining Zone (Unit M)
^^V^^^l
ftfV
injection Zone (Unit N) N"*"*F*W |
100-
Note: Figure not fo scate
Source; Daniel B. Stephens & Associates. Inc.
of megapascals, MPa, equal to 1-10 Pa)
I p, = 1000 kg/m' P. -2.16 MPa
Pressure front (P ,) given by
= 14.6 MPa
^^I^HH
P,. = 13.4 MPa
g - 362m TDS = 50,000 mg/L
p, = 1,012kg/m'
T = 75°C
Figure 3-3: Hypothetical Geologic Sequestration Site: Cross Sectional Schematic and Calculations.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
43
-------
Box 3-2. Hypothetical Example of an AoR Delineation, continued
The methodology used here is consistent with the determination of the pressure front for other
well classes within the UIC Program (e.g., USEPA, 2002). As explained above, in Section 3.4
of this guidance, this methodology is applicable to any proposed Class VI injection well for
which, prior to injection, the injection zone is not over-pressurized compared to the
lowermost USDW (i.e., the injection zone has a lower or equal hydraulic head as compared to
the lowermost USDW).
The pressure front is determined by calculating the minimum pressure within the injection
zone (Pi,f) necessary to cause fluid flow from the injection zone into the formation matrix of
the USDW through a hypothetical conduit (i.e., artificial penetration) that is perforated in
both intervals. Pi/is calculated using Eq-1 (Section 3.4.1). In this example, Pi/is 14.6 MPa.
Step 2. Inspect Model Results to Determine the Maximum Extent of the Pressure Front
The computational model will provide a prediction of the pressures within the injection zone
over time. For the purpose of AoR delineation, EPA recommends using the pressure
distribution corresponding to the time of maximum lateral extent of the pressure front (P,,/).
This will likely correspond to a time of maximum injection rates during the operational phase
of the project or to the end of a long injection period.
EPA recommends contouring these predictions of pressure increase and providing the
predictions on a base map of the proposed project area (Figure 3-4). In this recommended
contour map, EPA also recommends highlighting the pressure equivalent to Pi:f. In the
hypothetical example provided here, the region encompassed by Ptf includes the three
planned Class VI injection well locations and a significant distance surrounding the area of
the proposed injection wells.
UIC Program Class VI Well Area of Review 44
Evaluation and Corrective Action Guidance
-------
Box 3-2. Hypothetical Example of an AoR Delineation, continued
Explanation
Injection
walltfl
O Infection well
18.4 Pressure atwell IMPn)
1C Pressure concur tMPa
_1 4.6 Pressure contour (MPa) (pressure great enough
10 ctrtjjxj Nuid tnovfirwril into USDWJ
Source: Daniel B Stephens & Associates, tnc
Figure 3-4: Hypothetical Geologic Sequestration Site: Model Predicted Maximum Pressure Within the
Injection Zone.
Step 3. Inspect Model Results to Determine the Maximum Extent of the Separate-Phase
Plume
The computational model will also provide a prediction of the extent of the separate-phase
plume as it evolves over time. EPA recommends that these data be also contoured and
provided on a base map (Figure 3-5). In the example provided here, the maximum extent of
the supercritical plume, as predicted by the model, exists at 50 years after carbon dioxide
injection commences.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
45
-------
Box 3-2. Hypothetical Example of an AoR Delineation, continued
Explanation
Injection
welltt
O Injection well _*.
»10 Supercritical plume extent at given lime (year?)
_ SQ_ Maximum predated extent of supercritical carton dioxide plume
Source; D*r,&t B. Stofit^m & Axoitiulus. tnc.
Figure 3-5: Hypothetical Geologic Sequestration Site: Model Predicted Extent of Supercritical Carbon
Dioxide Plume Over Time.
Step 4. Delineate the AoR
Lastly, the maximum extent of the separate-phase plume and pressure front is compared and
overlaid on the base map (see Figure 3-6). The AoR is delineated by drawing the contour line
that encompasses the maximum extent of the separate-phase plume or pressure front (Figure
3-6).
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
46
-------
Box 3-2. Hypothetical Example of an AoR Delineation, continued
It is important to note that the region encompassed by the pressure front will not in all cases
be larger in all directions than the extent of the separate-phase plume. This is because the
pressure front does not include all areas exhibiting any increase in pressure, only pressure
great enough to cause fluid movement into a USDW. Therefore, pressure differentials may
still exist outside of the pressure front, and separate-phase fluids may migrate beyond the
extent of the pressure front. For this reason, it is necessary to calculate the extent of both the
plume and pressure front to delineate the AoR for a proposed Class VI injection well and to
submit these separate delineation results to the UIC Program Director with the permit
application.
X
Explanation
Injection
O Infection well P* jJGfl^HH^B ^H^H
Pressure front ipressure great snojgh to cause fud movement into JSOViT
MaxfTrum predicted extern of supercritical carbon dlox«de plume
i Delineated area of review (maximum extent of supetcntical
diowde or o^essura front)
Source. OaafefS Stepfl0nt£A880eb&a
Figure 3-6: Hypothetical Geologic Sequestration Site: Initial Area of Review Based on Model Results.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
47
-------
3.5. Reporting AoR Delineation Results to the UIC Program Director
The owner or operator is required to submit the AoR and Corrective Action Plan with the initial
permit application [40 CFR 146.82(a)(13)]. Information pertaining to how this plan should be
submitted is provided in the UIC Program Class VI Well Project Plan Development Guidance.
The final delineated AoR based on computational modeling is submitted to the UIC Program
Director prior to receiving authorization to inject [40 CFR 146.82(c)(l)].
EPA recommends that this permit application submittal include all necessary information for the
UIC Program Director to evaluate the AoR delineation results and replicate the computational
modeling exercise if he or she elects to do so, as well as all model input and output data and files.
This may include providing inputs for the UIC Program Director to use in their verification
modeling effort. The owner or operator and the UIC Program Director should discuss the
specific needs as the permit application is submitted. For additional information on submitting
information to support Class VI permit applications, please see the UIC Program Class VI Well
Recordkeeping, Reporting, and Data Management Guidance for Owners and Operators.
EPA recommends that the permit application submittal include the following in support of the
AoR delineation:
The conceptual site model and all supporting data on which the model is based, including
the description of geologic stratigraphy and any relevant geologic features. See Box 3-1
of this guidance document for more information;
Attributes of the code used to create the computational model, including the code name,
name of developing organization, a full accounting of or reference to the model
governing equations, scientific basis, and any simplifying assumptions;
A description of the model domain, i.e., the model's lateral and vertical extents, geologic
layer thickness, and grid cell sizes, as presented on maps and cross-sections;
An accounting of all equations of state used for all fluids modeled (e.g. ground water,
carbon dioxide);
Any constitutive relationships, such as relative-permeability saturation relationships, and
how they were determined;
Values of all model parameters, as detailed in Table 2-1 of this guidance document,
throughout the entire model domain, as a function of time if necessary, including initial
conditions and boundary conditions, and a description of how model parameters were
determined based on site characterization data. This information may be submitted in
tabular or graphical/map formats;
If required by the UIC Program Director, the owner or operator must also include raw
model input and output files [40 CFR 146.82(a)(21)]. These files may be useful in model
UIC Program Class VI Well Area of Review 48
Evaluation and Corrective Action Guidance
-------
verification, or if the UIC Program Director wishes to run alternative
simulations/scenarios with the model;
Model results, including predictions of carbon dioxide and pressure-front migration over
the lifetime of the project. EPA recommends that the model results be presented in the
form of contour maps, cross sections, and/or graphs showing plume and pressure front
migration as a function of time, and that the permit application submittal include the
outcome of parameter sensitivity analyses;
A description of pressure front calculation and delineation of the AoR; and
If required by the UIC Program Director, the relevant qualifications and professional
experience of any individuals and/or consulting firms responsible for model
development, AoR delineation, and reevaluation, including examples of previous
multiphase modeling studies conducted.
UIC Program Class VI Well Area of Review 49
Evaluation and Corrective Action Guidance
-------
4. Identifying Artificial Penetrations and Performing Corrective
Action
The purpose of AoR delineation for a proposed GS project is to determine the area where any
geologic features or artificial penetrations (e.g., wells) may become conduits for fluid movement
out of the injection zone, or additional zones, and potentially cause endangerment to a USDW.
Artificial penetrations include any man-made structures, such as wells or mines, which provide a
flow path out of the injection zone. The Class VI Rule requires that the owner or operator
prepare, maintain, and implement a AoR and Corrective Action Plan that includes a description
of how corrective action will be performed on any artificial penetrations through the confining
zone and whether such action will be phased [40 CFR 146.84(b)(2)(iv)].
This section discusses the identification and evaluation of artificial penetrations and the
performance of required corrective action (if necessary). Monitoring activities necessary for
detection of fluid leakage into USDWs are discussed in the UIC Program Class VI Well Testing
and Monitoring Guidance.
4.1. Rule Requirements
The following rule requirements pertain to corrective action within the AoR:
40 CFR 146.84(c)(2): Using methods approved by the UIC Program Director, identify all
penetrations, including active and abandoned wells and underground mines, in the AoR
that may penetrate the confining zone(s). Provide a description of each well's type,
construction, date drilled, location, depth, record of plugging and/ or completion, and any
additional information the UIC Program Director may require;
40 CFR 146.84(c)(3): Determine which abandoned wells in the AoR have been plugged
in a manner that prevents the movement of carbon dioxide or other fluids that may
endanger USDWs, including use of materials compatible with the carbon dioxide stream;
40 CFR 146.84(d): Perform corrective action on all wells in the AoR that are determined
to need corrective action, using methods designed to prevent the movement of fluid into
or between USDWs, including use of materials compatible with the carbon dioxide
stream, where appropriate;
40 CFR 146.84(e)(2): During the AoR reevaluation process, identify all wells in the
reevaluated AoR that require corrective action in the same manner specified in 40 CFR
146.84(c);
40 CFR 146.84(e)(3): Perform corrective action on wells requiring corrective action in
the reevaluated AoR in the same manner specified in 40 CFR 146.84(d);
40 CFR 146.84(e)(4): Revise the AoR and Corrective Action Plan as necessary whenever
the AoR is reevaluated.
UIC Program Class VI Well Area of Review 50
Evaluation and Corrective Action Guidance
-------
4.2. Identifying Artificial Penetrations within the AoR
The Class VI Rule requires potential Class VI injection well owners or operators to identify all
artificial penetrations located within the delineated AoR, including active and abandoned wells
and underground mines, that may penetrate the confining zone, and provide a description of each
well's type, construction, date drilled, location, depth, and if applicable, the record of plugging
and/or completion, and any additional information the UIC Program Director may require [40
CFR 146.84(c)(2)]. If the identified abandoned wells have been improperly plugged or not
plugged at all, such penetrations can provide unimpeded flow conduits out of the injection zone.
As such, they must be properly plugged in order to prevent endangerment of USDWs [40 CFR
146.84(d)].
A variety of types of abandoned wells may exist within the delineated AoR of a proposed GS
project, including wells constructed prior to federal or state regulation (i.e., in the late 1800s or
early 1900s) and any recently decommissioned wells. Wells constructed during early oil
exploration, including cable-tool drilled wells, pose the largest risk because these wells may be
relatively deep and often consist of an open (i.e., non-cased) well bore over much of their length.
These older wells may also not have been documented in state or local records.
Historically, wells no longer in use may not have been plugged and abandoned by today's
common standards. Prior to the early 1900s, there were no regulations concerning well
abandonment and it is unlikely that those wells were abandoned properly. Even in states
regulating well abandonment, it is likely that any wells abandoned before 1952 may have
inadequate plugs (Ide et al., 2006). In 1952, the American Petroleum Institute (API) published its
standards for cements for oil and gas wells. Prior to that, cement often lacked sufficient additives
to achieve the proper cement setting in the conditions experienced in oil and gas wells. As a
result, the plugs in many of these older wells failed to set properly and may have experienced
channeling and/or cement failure because of fluid intrusion into the improperly set cement.
The potential also exists for more recently constructed wells to have been decommissioned
improperly. For example, owners or operators may have gone bankrupt and failed to plug their
wells or used substandard materials.
Depending on site conditions and corrosion, "properly" plugged wells may also contain zones
(i.e., annular spaces) that could serve as a conduit for fluid movement. In other cases, the well
plugs may have degraded over time because of a poor cement job and/or corrosive conditions.
Even properly plugged wells may have been plugged with types of cement that could degrade
when in contact with a carbon dioxide plume. See the UIC Program Class VI Well Construction
Guidance for information on compatibility of different materials with a carbon dioxide stream.
Detecting abandoned wells can be very challenging in certain locations because of the variety of
wells that may exist. In addition, steel casings, the primary detectable portion of the well, were
often removed from abandoned wells for recycling and use during World War II (Gochioco and
Ruev, 2006). These challenges are compounded by the potentially large AoR delineation
determined for a proposed Class VI injection project, and therefore the greater surface area that
will have to be evaluated for the presence of artificial penetrations. However, as discussed
UIC Program Class VI Well Area of Review 51
Evaluation and Corrective Action Guidance
-------
below, several methods and sources of information are available to identify those artificial
penetrations in a relatively efficient manner. The primary stages of an abandoned well
investigation within the AoR include historical research, site reconnaissance, review of aerial and
satellite imagery, and one or more geophysical surveys. The reader is referred to additional
standard references regarding identification of artificial penetrations for further information
(Jordan and Hare, 2002; Frischknecht et al., 1985; ASTM, 2005).
4.2.1. Historical Research
Most deep wells that may penetrate the primary confining zone of a proposed GS project site are
related to oil and gas exploration and production. Deep well drilling for oil and gas exploration
dates back to the 1870s. State and local databases of well exploration may include locations of
abandoned wells, and EPA recommends conducting a records review as the first step in
abandoned well identification within the delineated AoR for a proposed Class VI injection well.
In addition, state and local records will provide information on the time period and types of
exploration that have been conducted in an area, and they may also provide information on
typical completion and abandonment methods in a given field. This records search will provide a
list of known abandoned wells, and it may inform additional stages of abandoned well
identification.
State well databases will, in most cases, provide valuable information for assistance with the
identification of abandoned wells. Prior to well construction, a government permitting authority
requires owners or operators to seek a permit to drill from a specific agency, such as a state
natural resources agency, environmental quality agency, or geological survey. Most states
maintain records of drilled wells, including location, construction, operating, and plugging
information. Although these records can take many forms, many states now have comprehensive
databases of these well records that have been digitized and made available online. However,
when conducting this historical records search, owners or operators of proposed Class VI
injection wells should be aware that older well records may not have been entered into databases.
In some cases, the records from different time periods may be filed in separate locations or on
separate types of media.
For example, the Wyoming Oil and Gas Conservation Commission maintains a digital database,
accessible online, of wells within the state. (See http://wogcc.state.wy.us for more information
on this database.) Basic information is available to the public regarding each well, including
geophysical survey results where available. The database can be searched by location, well
name, and well number, among other fields. The state also has a "well book" available online,
which contains records of older wells not entered in the database.
In addition, county records, including survey maps, ownership records, and chain-of-title and
property lease history, maintained by local tax assessors and county clerks, list abandoned wells
in many cases. Such records may also indicate land use and indicate areas and timeframes in
which drilling activities likely occurred. Private data compilation services often maintain detailed
databases for the purpose of oil and gas exploration, including information regarding well
locations, plugging, and abandonment. Often these services will maintain maps of known well
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
-------
locations. While these maps can be out of date, most private services have been known to update
their database for a fee.
4.2.2. Site Reconnaissance
Site reconnaissance includes interviewing local residents and property owners, as well as
conducting a physical search for features indicative of abandoned wells. Initial site
reconnaissance may be informed by the historical database research. For example, the records
search may indicate that, with a great deal of confidence, certain regions of the AoR have never
been subject to oil and gas exploration, deep well injection, or any other activity that may result
in deep well penetration. In this case, the owners or operators may choose to exclude those areas
from any additional well identification efforts.
Local residents that may be well informed regarding abandoned wells include oilfield workers
and service company employees, including consultants, and property and drilling-rights
ownership brokers. Such informed residents may be able to provide information on the areas and
timeframes where past drilling occurred. They may also be able to give additional details in
response to specific questions and provide information on locations, completion methods, and
plugging of wells.
Surface features that may be indicative of abandoned wells include abandoned well derricks,
access roads, brine pits, or vegetation stress associated with brine leakage. Detection of these
features at a site indicates the possible likelihood of one or more wells in the area. EPA
recommends that, because the AoR is likely to cover a large area, a surface review for such
features is most effectively supplemented by use of aerial surveys or photos.
4.2.3. Aerial and Satellite Imagery Review
EPA recommends that historical aerial photographs and satellite imagery be used in the
identification of abandoned wells. Aerial photographic surveys, taken from airplanes, were
conducted beginning in the 1930s and are available from a variety of governmental and private
information services. All historical aerial photos within the AoR are recommended to be
reviewed for evidence of past drilling activity. Surface features that provide a "signature" of
drilling activity include drill derricks, rig platforms, brine pits, power sources, and access roads.
Depending on the resolution of the image, satellite (i.e., remote sensing) images may be used to
detect wellheads, derricks, and surface features indicative of abandoned wells. These include
spatial patterns indicative of a well site, brine pits, modified topography, and vegetation stress
associated with brine leakage.
4.2.4. Geophysical Surveys
Geophysical surveys, including magnetic, ground penetrating radar (GPR), and electromagnetic
methods, can be used in the detection of abandoned wells. EPA recommends conducting
geophysical surveys throughout regions of the AoR that may have been subject to oil and gas
UIC Program Class VI Well Area of Review 53
Evaluation and Corrective Action Guidance
-------
exploration, deep well injection, or any other activity that may result in deep well penetration.
Geophysical methods will supplement other identification methods, discussed above.
Geophysical methods can help to pinpoint locations of known wells where surface evidence of
the well has been removed or can help to identify abandoned wells that are undocumented. The
type(s) of geophysical surveys conducted at a proposed Class VI injection well site are based on
known site subsurface and surface conditions. In general, at least two different types of
geophysical surveys are recommended in order to parse data background noise and to inform the
interpretation of survey results. As discussed below, ground or aerial (e.g., aeromagnetic)
surveys may be conducted, depending on the size of the area of interest.
4.2.4.1. Magnetic Methods
The magnetic method is one of the oldest and most well developed geophysical techniques, and
it is the standard method used for abandoned well detection. Magnetic surveys measure a
component of the magnetic field near the land surface. Any anomalies in the magnetic field are
caused by subsurface features, which could include abandoned well bores with iron or steel
casings. Anomalies associated with well casings are typically distinguishable from the
background magnetic field.
Magnetic surveys are applicable to abandoned wells with iron or steel casings or to wellheads in
areas with relatively low background magnetic signatures. Areas with significant cultural
development on the surface or in the shallow subsurface may have high interferences. Airborne
magnetic surveys can detect most wells constructed with approximately 200 feet or more of at
least 8-inch casing, and in some cases, very large cavities (Frischknecht et al., 1985). However,
open well bores, non-steel casings, or severely corroded casings cannot typically be detected
with a magnetic survey.
Ground or aerial (i.e., aeromagnetic) surveys may be conducted, depending on the size of the
area of interest. Aeromagnetic surveys will likely be more practical for most GS projects due to
the anticipated size of the delineated AoR, as they can collect large amounts of data in a
relatively short amount of time. Both ground and aerial surveys are conducted along straight-line
transects. EPA recommends that that ground survey transect spacing be no larger than 20-30
feet, and aerial survey transect spacing be no larger than 50-100 feet (Jordan and Hare, 2002).
Magnetic surveys may be conducted to measure the total magnetic field, or the vertical or
horizontal field gradients. For the purpose of locating abandoned well bores, the total magnetic
field measurement type is recommended. During these surveys, EPA recommends that the
operator periodically return to a common point to ensure instrument repeatability, continuously
measure diurnal variation in the magnetic field, and avoid high magnetic gradients. Data
processing of magnetic surveys includes incorporation of spatial positioning data, correction for
diurnal variation, and data filtering.
Figure 4-1 compares aeromagnetic survey results for the Coon Creek oil field in Oklahoma to
abandoned wells identified from aerial imagery (USGS, 1995). As shown in the figure, magnetic
anomalies associated with well casings are typically apparent. However, this figure also reveals
UIC Program Class VI Well Area of Review 54
Evaluation and Corrective Action Guidance
-------
some of the typical challenges that may be faced by owners or operators in abandoned well bore
identification. One challenge is that, due to the presence of other buried infrastructure (e.g.,
pipes), certain regions exhibit larger magnetic field values even if wells are not present.
Additionally, some wells may not be identified in the aeromagnetic survey, most likely because
of well casing removals. These challenges demonstrate the benefits of using multiple survey
techniques in order to properly identify abandoned wells.
ExpHin-ttian
Wnlte deiitrfioc n
Figure 4-1: Total Field Aeromagnetic Map, Cook Creek Oil Field, Arcadia, Oklahoma.
From: USGS (1995).
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
55
-------
4.2.4.3. Electromagnetic Methods
Electromagnetic methods used for abandoned well bore detection include frequency-domain and
time-domain electromagnetic surveys. These surveys consist of an electromagnetic transmitter
that establishes an electromagnetic field measured by a receiver. Similar to magnetic surveys,
electromagnetic surveys are non-invasive, as both the transmitter and receiver are positioned
above the ground surface. Both surface and aerial electromagnetic surveys are possible. The
depth at which these instruments are able to detect objects depends on the size and geometry of
the sensor, the size and conductivity of the target, and the potential interference from other
sources, such as fences and pipelines. Generally, object detection at depths ranging from a few
meters to several hundred meters is possible. Larger and more complex arrays are required at
greater depths. Aerial surveys are not likely able to detect small objects, such as well casings, but
may detect brine plumes, which may indicate the presence of abandoned wells (Jordan and Hare,
2002).
Abandoned well bore detection using electromagnetic methods is based on the larger
conductivity of steel casings and other well materials compared to surrounding soils and
geologic formations. These methods may detect anomalous fluids associated with leakage from
an abandoned well. Frequency-domain electromagnetic methods can measure current induced in
the subsurface by the electromagnetic field established by the transmitter. Induced current
establishes a secondary electromagnetic field detected by the receiver. The magnitude of the
induced current is a function of subsurface conditions, including conductivity. Time-domain
electromagnetic methods measure the decay of the secondary magnetic field created by the
induced current, and they can be especially useful for detection of brine leakage.
4.2.4.4. Ground Penetrating Radar
GPR may be used in abandoned well bore detection and in finding other artificial penetrations.
Unlike other geophysical methods, GPR does not rely on the presence of a steel or iron well
bore, so it may be able to detect open boreholes and non-metallic materials. GPR uses high
frequency radio waves to measure the transmission of electromagnetic energy. The investigation
depth possible depends on the frequency of the radio waves and the conductivity of the ground.
The greater the depth, the less resolution the instrument will have. For small objects, such as well
casings, depths are limited to a few meters (Jordan and Hare, 2002). This depth limitation may
lessen the value of GPR in areas with large topographical changes. GPR is also slower than
magnetic or electromagnetic methods. GPR is likely not as practical to use throughout the
delineated AoR as an initial larger scale survey method because the distance between transect
lines for sufficient resolution is too small. Instead, EPA recommends using GPR to determine the
exact location of abandoned well bores within a given area that have already been identified by
earlier, larger scale surveys.
4.3. Assessing Identified Abandoned Wells
After all artificial penetrations within the AoR that may penetrate the confining zone have been
identified, the owners or operators of a proposed Class VI injection well must evaluate the
UIC Program Class VI Well Area of Review 56
Evaluation and Corrective Action Guidance
-------
potential for each artificial penetration to serve as a conduit for fluid movement. In particular,
owners or operators must establish which abandoned wells in the AoR, if any, have not been
plugged in a manner that would prevent the movement of carbon dioxide or other fluids that may
endanger USDWs [40 CFR 146.84(c)(3)]. To prevent fluid movement, abandoned wells should
include a cement plug through the primary confining zone, and/or across the injection
zone/confining zone contact, with sufficient integrity to contain separate-phase carbon dioxide
and elevated pressures. The type of plugging that is sufficient to contain carbon dioxide and
formation fluids from the injection zone will be site specific and should be reviewed with the
UIC Program Director. In the absence of an adequate plug across the confining zone, cross-
migration may occur where fluids enter a permeable zone below the lowermost USDW and then
migrate upward from that zone. See Figure 4-2 for more information. EPA recommends cement
surface plugs (typically required by well abandonment regulations), and the UIC Program
Director may require additional plugs based on site-specific circumstances.
Evaluation of the wells in the AoR requires a two-step approach. The first step is to review
whatever records are available, as outlined in Section 4.3.1, for information relevant to proper
plugging. The second step is to perform physical tests on wells that are suspect or for which no
records are available.
4.3.1. Abandoned Well Plugging Records Review
A records review can aid in reducing the number of identified wells that may need to be
evaluated by future field testing. Records of wells that have been recently abandoned, have no
mentions of any difficulties experienced during the abandonment procedure, are cased holes, and
have plugs and cement situated to isolate the injection zone from other fluid containing zones
may be used to justify reduction in the number of follow-up field investigations. If records are
incomplete or indicate that the well has not been plugged or was inadequately plugged, follow-
up field investigations should be performed. Identified undocumented wells will have no records
and will require field investigation in order to determine the quality of plugging, as required in
the Class VI Rule [40 CFR 146.84(c)(3)]. The owner or operator may also choose to plug any
questionably abandoned wells rather than go through the expense of evaluating the plugs.
There are many elements in existing reports that can help in determining the adequacy of
abandonment procedures for identified wells located within the AoR. Some key elements to
review include, but are not limited to:
Well depth and completion;
Well abandonment date;
Open hole or cased hole;
Location of plugs;
Casing and cementing records;
Records of mechanical integrity tests (MITs) or logs performed; and
Well deviation.
UIC Program Class VI Well Area of Review 57
Evaluation and Corrective Action Guidance
-------
Examples of Improperly Abandoned Wells (Open Boreholes),
Plugs Not Set Through Confining Zone
Long string casing
Cement plug
Injection packer
Injection Zone
Note: Figure not to scale
Source: Daniel B. Stephens & Associates, Inc.
* Drilling fluid
Cement plug
Figure 4-2: Examples of Carbon Dioxide Leakage Through Improperly Abandoned Wells.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
58
-------
The well completion depth is important in determining if the identified abandoned well may
penetrate the proposed confining zone(s). If the well completion depth is above the confining
zone(s), no further action would likely need to be taken. The date of abandonment may also
provide information as to the adequacy of the abandonment procedure. Whether the well was
abandoned with casing or as an open hole is an important consideration in determining the
likelihood that the well might act as a conduit for fluid movement. Open holes are susceptible to
cross-migration between aquifers. If the hole is open and there is not a proper plug located at a
depth corresponding to the primary confining zone, fluids may migrate out of the injection zone
and into a USDW. For cased holes, EPA recommends that integrity of the casing be evaluated.
The location and type of plugs are also important factors, especially in open-hole wells. The plug
locations must be reviewed in order to determine the quality of plugging, as required in the Class
VI Rule [40 CFR 146.84(c)(3)]. For example, EPA recommends that the injection zone be
isolated from all other formations with plugs. This may be especially important if a well was
completed in a formation deeper than the proposed injection zone. EPA recommends that any
length of the well in the proposed injection zone be properly isolated by means of plugs and
casing. Mechanical plugs and cemented casing are not sufficient for the long-term isolation of
carbon dioxide, as eventually the metal is likely to corrode and the plug will fail (Randhol et al.,
2007). Therefore, cement plugs are considered superior to mechanical plugs for preventing the
movement of fluids into or between USDWs. EPA recommends that cement plugs be located
across the bottom of any casings, at the base of the lowermost USDW, and that plugging fluid
(i.e., composition, specific gravity) characteristics be considered, as drilling fluid of sufficient
weight may resist displacement by the injectate or mobilized fluids.
The integrity of any existing casing and cement must be determined in order to assess the quality
of well construction and plugging, as required in the Class VI Rule [40 CFR 146.84(c)(3)]. EPA
recommends reviewing the casing and cement quality through the proposed injection zone in
order to ensure that they are appropriate for contact with carbon dioxide, as well as reviewing
any additional well records that may indicate unusual conditions experienced during casing and
cementing. Events such as a loss of circulation, well bore stability problems, lack of the use of
centralizers, and/or improper removal of drilling mud before cementing can all lead to premature
cement or casing failure. Reviewing load calculations, if available, and comparing them to actual
events recorded in the drilling log may give the owners or operators an indication of an under-
designed casing that may be susceptible to failure. For example, if the casing had a low axial
loading stress and stuck pipe was experienced during casing placement, it is possible that the
casing may have experienced damage. The materials used for the well casing and cement must
also be assessed to see if they are compatible with carbon dioxide, in order to comply with Class
VI Rule requirements [40 CFR 146.84(c)(3)]. See the UIC Program Class VI Well Construction
Guidance for more information on compatibility of different materials with a carbon dioxide
stream.
Any tests performed on the well prior to its abandonment can also be useful information. An
MIT such as a pressure test, noise log, temperature log, or cement evaluation log can provide
information on any known or suspected leaks. If leaks were encountered, EPA recommends
determining if the source of the leak was found and repaired. If the leaks were not sealed,
UIC Program Class VI Well Area of Review 59
Evaluation and Corrective Action Guidance
-------
corrective action would be required to be taken to plug the leaks as discussed below [40 CFR
146.84(d)]. Drilling records can yield clues as to areas that might be susceptible to failure. Mud
logs and open-hole caliper logs can show areas of weak formations. Weak formations are
susceptible to well bore instability and subsequent cement failure. Cement evaluation logs and
temperature logs taken at the time of completion can also give an idea of the condition of the
cement, although degradation is always possible after well completion. Any corrosion logs will
help provide information on the condition of the casing. Results from mechanical caliper logs,
electromagnetic thickness logs, or down-hole video can show the casing condition when the well
was abandoned. For more information regarding these logs and tests, see the UIC Program Class
VI Well Testing and Monitoring Guidance.
Evaluation of well records for deviation during drilling may also identify wells more likely to be
in need of corrective action, as deviated wells are far more likely to fail than wells with no
deviation (Watson, 2009). Events such as well bore collapse during drilling or conditions that
placed unusual loads on the casing may also indicate a higher chance of failed well bore
integrity. EPA recommends that the design casing load also be checked to ensure adequacy for
the actual loads faced by the well.
4.3.2. Abandoned Well Field Testing
After all the available records have been reviewed, any wells located within the AoR that cannot
be proven to have plugs adequate to prevent migration of carbon dioxide or formation fluids out
of the injection zone must be evaluated by field tests in order to determine the quality of
plugging, as required in the Class VI Rule [40 CFR 146.84(c)(3)]. Evaluation and corrective
action for wells which the plume is not expected to reach in the near future may be phased. If the
owner or operator chooses and the UIC Program Director agrees, the evaluation may be omitted
and the wells re-plugged. If the integrity of the bottom plug or cement is in question, and records
cannot prove that the plugging is adequate, EPA recommends that the surface plug and possibly
additional plugs down-hole be drilled out and tests conducted to determine the adequacy of
abandonment. There are numerous field tests available to evaluate the integrity of abandoned
wells. Several of these tests are discussed in detail in the UIC Program Class VI Well Testing
and Monitoring Guidance. Additionally, the owner or operator must demonstrate guaranteed site
access to wells potentially needing corrective action in the future [40 CFR 146.84(b)(iv)]. The
owner or operator is encouraged to consult the UIC Program Director regarding any difficulties
in gaining site access in order to evaluate and perform corrective action on any identified
improperly plugged abandoned wells.
EPA recommends that both the casing and the cement plugs be evaluated. Casing failure is most
common at joints and in weak formations where instability around the well bore can lead to
failed cement and to casing buckling. Weak formations are also common areas for cement
failure, as are high pressure formations, due to fluid intrusion. Tools used to evaluate the cement
and casing include, but are not limited to:
Multi-finger caliper log;
Sonic scanner:
UIC Program Class VI Well Area of Review 60
Evaluation and Corrective Action Guidance
-------
Ultrasonic imaging tool;
Cement evaluation log;
Radioactive tracer;
Cased hole dynamic tester;
Modular sidewall coring tool; and
Cased hole fluid test.
Multi-finger caliper logs measure the radius of the borehole in a non-destructive way. They can
give a 360-degree picture of the inside of the casing and identify any defects caused by
corrosion, erosion, or other events (e.g., dropped tools).
A sonic scanner sends out sound waves and measures the returned waves in receivers. The log
provides information on the quality of the casing-cement bond and the cement-formation bond.
The sonic scanner averages the results for the entire radius and therefore cannot provide three-
dimensional pictures of the cement bond, or determine the reasons for a poor quality cement
bond. An ultrasonic imaging tool is another non-destructive tool that uses ultrasonic transmitters
and receivers to determine information about the casing and cement. The ultrasonic imaging tool
can return 360-degree information on casing thickness, cement thickness, and cement bond.
More information on these tools can be found in Duguid and Crow (2007) and Close et al.
(2009).
A cement evaluation log is another tool used and log results include information on both the
cement and the bond quality. This log provides results that are averaged over the circumference
of the well, and testing is typically conducted in combination with an ultrasonic imaging tool to
provide more complete information on the three-dimensional picture of the well. In some cases,
the cement hardens while the well casing is under pressure and, when pressure is released,
microannuli can form between the casing and cement. If unconnected to other cracks, these
microannuli cannot transmit fluid, but they will appear in logging results as a potential poor
bond. This artifact can be evaluated by performing the cement evaluation log under pressure
(Randhol et al., 2007). Radioactive tracers also can be used to detect leaks in casing and cement
and fluid leaking along channels in the well bore. Radioactive tracers are injected down the well,
and gamma detectors are used to detect any fluid flow.
Cased-hole dynamic testers measure mobility or porosity. They can be used to determine the
porosity of the cement. They are semi-destructive tests as they do create a small hole in the
casing and cement; however, the hole is patched after the test is run. The instrument works well
in highly permeable formations or in cement, while performance in lower porosity formations is
still under investigation.
Modular sidewall coring tools take small cores of the casing and cement for analysis in the
laboratory. Laboratory analyses can include scanning electron microscopy, X-ray diffraction, and
measurements of permeability and density. This is a more destructive test that leaves
approximately 1-inch diameter holes in the side of the well, which is then patched with a
remedial cement squeeze after testing is completed. Cased-hole fluid testers can be run with the
UIC Program Class VI Well Area of Review 61
Evaluation and Corrective Action Guidance
-------
cased-hole mobility tool, using optical instruments to determine what fluids are present in the
formation outside the well bore.
In general, EPA recommends that these tests be run sequentially, from the simplest and least
destructive tests to the more complicated and destructive tests. This way, if a flaw is found with a
simpler test that determines that the well should be plugged or otherwise remediated, the more
expensive and destructive tests may be avoided. The typical order of running the tests is caliper
log, sonic and ultrasonic tools, cased-hole mobility and fluid tests, and then sidewall cores
(Duguid and Crow, 2007). This set of tools can be used to determine the quality of the casing and
cement; if flaws such as degraded cement porosity, casing corrosion, microannuli in the cement,
channels between the cement and casing or cement and formation, or missing cement are found,
the Class VI Rule requires that corrective action be performed on the well [40 CFR 146.84(d)]. A
brief summary of the main methods for evaluating cement and casing condition along with major
benefits and disadvantages are included in Table 4-1 below.
Table 4-1: Tools for Assessment of the Integrity of Abandoned Wells.
Tool
Multifinger calipers
Sonic Logs
Ultrasonic Logs
Cement evaluation log
Tracers
Dynamic Cased Hole
Tester
Sidewall coring
Target
Casing
Cement
Casing,
Cement
Cement
Leak
detection
Cement
Cement
Advantages
Non-destructive, relatively
simple
Non-destructive, yields
information on cement bond
Non-destructive, can detect
flaws in casing and cement,
provides three-dimensional
images
Non-destructive, yields
information on quality of
cement bond
Can pinpoint routes of leaks,
channeling
Can determine porosity of
cement
Can give detailed analysis of
cement condition
Disadvantages
Only examines interior, only
detects casing damage
Results averaged over well
circumference, can't indicate
reasons for poor quality bond
Sensitive to well fluids
Results averaged over well
circumference
Radioactive tracers require special
handling and may have negative
public perception
Semi-destructive, untested in low
porosity conditions
Destructive
4.4. Performing Corrective Action on Wells Within the AoR
The Class VI Rule requires that owners or operators of Class VI injection wells perform
corrective action on all artificial penetrations in the AoR that may penetrate the confining zone
and are determined to have been plugged and abandoned in a manner such that they could serve
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
62
-------
as a conduit for fluid movement and endanger USDWs [40 CFR 146.84(d)]. In performing
corrective action, owners or operators must use methods designed to prevent the movement of
fluid into or between USDWs, including use of materials compatible with the carbon dioxide
stream, where appropriate [40 CFR 146.84(d)]. Figure 4-3 presents a decision tree that illustrates
how the various evaluation tools can be used together to evaluate abandoned wells in an efficient
and logical manner.
As described above, the Class VI Rule allows owners or operators to perform corrective action
on a phased basis, if approved by the UIC Program Director. If a phased approach is approved
for performing corrective action for a GS project, EPA recommends that all required corrective
action on wells identified as deficient during the permit application process (or AoR
reevaluations) receive corrective action prior to the end of the injection phase.
It is possible that some corrective action may be performed during the post-injection phase. For
example, if the plume and pressure front movement were to deviate from predictions, this may
necessitate corrective action for newly identified artificial penetration during the post-injection
phase.
Performing corrective action on improperly abandoned wells is intended to prevent the
movement of carbon dioxide or other mobilized fluids into or between USDWs. Acceptable
forms of corrective action include well plugging and/or remedial cementing of the improperly
abandoned well. In addition to corrective action, EPA recommends performing enhanced
monitoring in the vicinity of improperly abandoned wells, including ground water monitoring
and using indirect geophysical techniques for obtaining monitoring results. Appropriate
monitoring for Class VI injection wells is discussed in the UIC Program Class VI Well Testing
and Monitoring Guidance.
UIC Program Class VI Well Area of Review 63
Evaluation and Corrective Action Guidance
-------
Do state records or
geophysical surveys
indicate the presence
of abandoned wells?
No
->[ Evaluation complete.
Yes
Do records indicate the wells were plugged in
a manner that will prevent carbon dioxide
plume or formation fluid migration and that is
compatible with the carbon dioxide plume and
formation fluids?
V t
Yes
No
Or
Drill out surface plugs
and other plugs above
the injection zone.
Plug questionable
wells.
Perform caliper log on
well.
Does the caliper log indicate holes in the
casing?
Yes
No
Perform remedial
cementing or plug
faulty section and re-
plug well.
Run cement evaluation
logs.
Do the cement evaluation logs indicate
channels or missing cement that could allow
migration of fluids?
Yes
No
Run cased hole dynamic
tester or sidewall core.
Does cement testing indicate cement
degradation that will allow fluid movement?
Yes
No
T
Figure 4-3: Well Evaluation Decision Tree.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
64
-------
4.4.1. Plugging of Wells within the AoR
Plugging of Class VI injection wells at the cessation of the injection phase of the project is
discussed in detail in the UIC Program Guidance on Class VI Well Plugging, Post-Injection Site
Care, and Site Closure. This section focuses on the plugging of improperly abandoned wells
within the AoR prior to the commencement of injection. However, because similarities exist in
plugging techniques for abandoned wells and former injection wells, the reader should refer to
the UIC Program Guidance on Class VI Well Plugging, Post-Injection Site Care, and Site
Closure for further detail regarding well plugging techniques for Class VI injection wells.
A well requires plugging if records indicate that an abandoned well was not plugged, was
plugged and abandoned improperly, or has not been plugged in a manner that prevents
movement of carbon dioxide or other fluids that may endanger USDWs [40 CFR 146.84(c) and
(d)]. In addition, where records indicate that a well plug does not exist at a depth corresponding
to the primary confining layer of the GS project, EPA recommends that the well have an
additional plug set at this depth to meet the requirements of the Class VI Rule. Where records
indicate that there are no well plugs below USDWs or other permeable formations that may
exhibit cross flow of mobilized fluids, additional plugs may be required by the UIC Program
Director for proper corrective action in these zones. Also, in wells that were plugged but the
evaluation techniques discussed in Section 4.3 of this guidance document reveal cracks,
channels, or annuli in the plug that would allow fluid migration, EPA recommends drilling out
and replacing the plug. In addition, if the plug material may corrode in a carbon-dioxide rich
environment, EPA recommends replacing it. For wells where casing exists at depths
corresponding to the injection and/or confining zone and the annular space may serve as a
conduit for fluid movement if not properly cemented, remedial cementing may be necessary or
the casing may need to be removed and replaced with a cement plug. See Section 4.4.2 of this
guidance document for more information on remedial cementing.
For the plugging of improperly abandoned wells within the AoR, EPA recommends that a plug
be set at a depth interval corresponding to the primary confining zone overlying the injection
zone of the Class VI injection well. In the absence of an adequate plug across the confining zone,
cross-migration may occur wherein fluids enter a permeable zone below the lowermost USDW
and then migrate upward from that zone. See Figure 4-2 of this guidance document for more
information. However, in order to supplement the confming-zone plug, ideal additional plugging
zones include the bottom of any casings and across any USDWs. A surface plug would also
typically be required by local well abandonment regulations to ensure that there is no risk of
anyone physically falling into the well bore.
To provide the best possible barrier to carbon dioxide migration out of the injection zone, EPA
recommends that corrective action be conducted in a manner to provide multiple barriers to
carbon dioxide migration and avoid underground cross-flow. Materials that are compatible with
the carbon dioxide must be used where appropriate [40 CFR 146.84(d)]. Material compatibility
with carbon dioxide is discussed further in the UIC Program Class VI Well Construction
Guidance.
UIC Program Class VI Well Area of Review 65
Evaluation and Corrective Action Guidance
-------
4.4.2. Remedial Cementing
Properly cementing improperly abandoned wells located within the delineated AoR between any
existing well casing and the geologic formation, especially through the injection zone, provides
an important fluid migration barrier. EPA recommends performing remedial cementing in order
to meet the corrective action requirements of the Class VI Rule if a well has been properly
plugged but the records, or any testing such as that described in Section 4.3 of this guidance,
indicate that the cement surrounding the well bore has failed or has cracks, channels, or annuli
that could allow migration of carbon dioxide. Key areas on which to focus remedial cementing
include depths corresponding to the injection zone and through any other permeable zones.
Remedial cementing is performed through squeeze cementing, where the cement is emplaced
into the affected area. For more information on cement squeezes, refer to Reynolds and Kiker
(2003). Increased pressure on the cement forces water out of the cement slurry leaving behind
the partially dehydrated cement. Cement squeezes can either be low pressure or high pressure.
Low pressure squeezes are used to set a small amount of cement in a given area and operate at a
pressure lower than the fracture pressure of the formation. Higher pressure squeezes are used
when channels or disconnected microannuli are to be cemented. The higher pressure squeezes
may fracture the formation and then allow the cement to flow into disconnected channels.
Cement squeezes can be performed using either a packer or a bradenhead squeeze. The methods
differ in how the treated section is isolated from the rest of the well. In the packer squeeze,
packers isolate the area to be treated, and a bridge plug isolates the area below the area to be
cemented, while a modified packer with a bypass valve isolates the area above the treated area.
Cement retainers are used if significant back pressure is expected. A bradenhead squeeze only
isolates the area below the area to be cemented. It is typically used only if the casing above the
treated area is strong enough to withstand the squeeze pressure. In cement squeezes, either
drillable packers or retrievable packers can be used. Drillable packers allow less freedom in
placement but better control of the cement. They are preferred if high pressures are maintained
on the cement after the squeeze.
Cements used in squeeze cementing can vary depending on the nature of the defect. The Class
VI Rule requires that all materials used for cementing of abandoned wells be compatible with the
carbon dioxide stream, where appropriate [40 CFR 146.84(d)]. Traditional cements may be
supplemented with or replaced by materials such as polymer gels and acrylic grouts. Acrylic
grouts can be used for small casing leaks or cases where pressure leak off is detected. High
concentration low molecular weight polymers can be used for small to moderate leaks. High
molecular weight polymers are typically used for channeling and lost circulation applications.
Cement or cement/polymer blends are typically used for severe leaks (Randhol et al., 2007).
4.5. Reporting Well Identification, Assessment, and Corrective Action to the UIC
Program Director
As discussed in the UIC Program Class VI Well Project Plan Development Guidance, the AoR
and Corrective Action Plan, submitted with the initial stage of the permit application, must
UIC Program Class VI Well Area of Review 66
Evaluation and Corrective Action Guidance
-------
indicate what well identification and assessments will be used and how corrective action will be
conducted [40 CFR 146.84(b)(2)(iv)]. The plan is a condition of the permit and is subject to UIC
Program Director approval [40 CFR 146.84(b)].
Owners or operators seeking a Class VI injection well permit are required to report the following
information regarding abandoned wells within the AoR that may penetrate the primary confining
zone: the well's type, construction, date drilled, location, depth, record of plugging and/or
completion, and any additional information required by the UIC Program Director [40 CFR
146.82(a)(4)]. This information may be found in acceptable public and private databases, where
available. See Section 4.2.1 for more information. In cases where available records do not
provide the necessary information or indicate that the well was plugged improperly, in a
questionable manner, or with materials inappropriate for contact with carbon dioxide, then site
investigations must be performed to establish the condition of the well, as discussed previously
[40 CFR 146.84(c)(3)].
The UIC Program Director will review the submitted well information to ensure completeness
and may consult with officials at oil and gas or water agencies to ensure that the well search was
thorough. The UIC Program Director will also review well completion records to determine
those wells that may penetrate the primary confining zone and will likely compare this list to
wells scheduled for corrective action and submitted with the Class VI injection well permit
application. For those identified abandoned wells that have been determined by the owner or
operator to not require corrective action, the UIC Program Director will likely review the records
of plugging and any field testing conducted, to verify that the well does not require corrective
action. If information on the depth or condition of the plug(s) is missing, the UIC Program
Director may request additional tests or require the well to be re-plugged.
Reports of any tests done on abandoned wells must be submitted to the UIC Program Director
with the permit application along with a list of wells for which corrective action will be
conducted, as part of the AoR and Corrective Action Plan required at 40 CFR 146.84(b). Before
receiving authorization to inject, the owner or operator must submit a report on the status of
corrective action [40 CFR 146.82(c)(6)], indicating the number, type, and location of the plugs.
EPA recommends that owners or operators also submit any records of remedial cementing with
the Class VI injection well permit application, along with cement logs showing the methods used
and the results of the remedial cementing. Testing and remedial cementing records for wells
which are part of a planned later stage of corrective action may be submitted after that phase is
completed.
UIC Program Class VI Well Area of Review 67
Evaluation and Corrective Action Guidance
-------
5. AoR Reevaluation
The Class VI Rule requires owners or operators of permitted Class VI injection wells to
reevaluate the AoR delineation on a regular basis, at a frequency of at least once every five (5)
years [40 CFR 146.84(e)]. The purpose of AoR reevaluation is to ensure that the initial model
predictions are adequate for predicting the extent of the separate-phase carbon dioxide plume and
pressure front. To this end, AoR reevaluation consists of a comparison of modeling predictions
and the required site monitoring data [40 CFR 146.90] and a revision of the model used to
delineate the AoR when necessary. Because Class VI injection well permits are granted for the
lifetime of the project, AoR reevaluation is the primary opportunity for the owner or operator and
the UIC Program Director to assess the project's operation and take additional appropriate
actions, if necessary, to protect USDWs. If a revision of the AoR delineation is necessary, a
revision of the AoR and Corrective Action Plan is also required [40 CFR 146.84(e)(4)], along
with other related project plans that may be dependent on the extent of the delineated AoR,
including the Testing and Monitoring Plan [40 CFR 146.90(j)]. It is important to note that a
change in the AoR and/or the AoR and Corrective Action Plan after the permit is issued may
constitute a modification of the Class VI permit, and would be subject to public notice [40 CFR
144.39(a)(5)(i)].
Reevaluations of the AoR must continue throughout the life of the GS project, including the
post-injection phase [40 CFR 146.84(e)]. It is likely that, following cessation of injection, the
area of increased pressure will reduce in size as pressures dissipate; therefore, EPA expects that
the reviews will entail an examination of monitoring data and confirmation and communication
to the UIC Program Director that no modifications to the AoR or amendments to any plans are
needed. However, this step is necessary to ensure that USDWs are not endangered and that all of
the plans in force (including the PISC and Site Closure Plan and the Emergency and Remedial
Response Plan) remain protective of USDWs.
5.1. Class VI Rule Requirements Related to AoR Reevaluation
The following Class VI Rule requirements pertain to reevaluation of the AoR:
40 CFR 146.84(e): At the minimum fixed frequency, not to exceed five years, as
specified in the AoR and Corrective Action Plan, or when monitoring and operational
conditions warrant, owners or operators must:
(1) Reevaluate the AoR in the same manner specified in 40 CFR 146.84(c)(l);
(2) Identify all wells in the reevaluated AoR that require corrective action in the
same manner specified in 40 CFR 146.84(c);
(3) Perform corrective action on wells requiring corrective action in the
reevaluated AoR in the same manner specified in 40 CFR 146.84(d); and
(4) Submit an amended AoR and Corrective Action Plan or demonstrate to the
UIC Program Director through monitoring data and modeling results that no
amendment to the AoR and Corrective Action Plan is needed. Any amendments
UIC Program Class VI Well Area of Review 68
Evaluation and Corrective Action Guidance
-------
to the AoR and Corrective Action Plan must be approved by the UIC Program
Director, must be incorporated into the permit, and are subject to the permit
modification requirements at 40 CFR 144.39 or 144.41, as appropriate.
5.2. Conditions Warranting an AoR Reevaluation
AoR reevaluation is required at a minimum fixed frequency of at least once every five years, or
when monitoring and operational conditions warrant [40 CFR 146.84(e)]. EPA recommends that
monitoring and operational conditions that may warrant a reevaluation of the AoR include:
Significant changes in site operations that may alter model predictions and the AoR
delineation;
Monitoring results for the injected carbon dioxide plume and/or the associated pressure
front that differ significantly from model predictions; or
New site characterization data obtained that may significantly change model predictions
and the delineated AoR.
Any site-specific criteria that will trigger an AoR reevaluation for a particular project must be
included in the AoR and Corrective Action Plan [40 CFR 146.84(b)(2)(ii)].
5.2.1. Minimum Fixed Frequency
As stated above, the owners or operators of permitted Class VI injection wells must reevaluate
the AoR delineation at least once every five years [40 CFR 146.84(e)]. The planned fixed
frequency for reevaluation must be included in the AoR and Corrective Action Plan [40 CFR
146.84(b)(2)(i)]. The AoR may need to be reevaluated more frequently than the previously
scheduled timeframe based on other factors. In these cases, the schedule for AoR reevaluation
may be updated appropriately. At no time may AoR reevaluation occur less than once every five
years [40 CFR 146.84(e)].
5.2.2. Significant Changes in Operations
Significant changes in operation of the GS project and/or individual Class VI injection wells
mandate an AoR reevaluation [40 CFR 146.84(e)]. The UIC Program Director may require an
AoR reevaluation prior to approving any operational changes. If allowed by the UIC Program
Director, operational changes may occur prior to reevaluation of the AoR. In these cases, EPA
recommends that the AoR reevaluation be submitted to the UIC Program Director within an
agreed-upon timeframe of instituting such changes, as described in the AoR and Corrective
Action Plan.
EPA recommends that proposed operational changes warranting an AoR reevaluation may
include, but are not limited to, a change in the location or number of Class VI injection wells
injecting into the same injection zone and/or a change in carbon dioxide injection rates, volumes,
or pressures outside of the limits of the original permit and AoR delineation. Additional
UIC Program Class VI Well Area of Review 69
Evaluation and Corrective Action Guidance
-------
operational changes that may warrant an AoR reevaluation, if required by the UIC Program
Director, include a change in the composition of the injectate or changes in fluid production rates
from the injection or overlying zones. Based on the discretion of the UIC Program Director,
short-term routine operational changes (e.g., temporary well shut-ins) may not warrant
reevaluation of the AoR.
In addition, the owner or operator may choose to perform an AoR reevaluation based on other
operational changes, with the approval of the UIC Program Director. Specific operational
triggers for an AoR reevaluation for a particular Class VI injection well must be included in the
AoR and Corrective Action Plan submitted with the permit application for that particular
injection well [40 CFR 146.84(b)(2)(ii)]. Operational changes that trigger a reevaluation may be
associated with the GS project under which the permitted Class VI injection well operates or
with separate projects that inject carbon dioxide into the same injection formation.
5.2.3. Results from Site Monitoring that Differ From Model Predictions
EPA recommends that collection of any monitoring data (required under 40 CFR 146.90) that
indicate carbon dioxide and/or pressure front migration significantly different than that predicted
by the current AoR delineation model warrant an AoR reevaluation. Specific criteria for
differences in monitoring data and model predictions that may trigger an AoR reevaluation for a
particular project must be included in the AoR and Corrective Action Plan [40 CFR
146.84(b)(2)(ii)]. In such cases, when monitoring data and modeling predictions differ, the
owner or operator is encouraged to notify the UIC Program Director and submit an AoR
reevaluation within timeframes that have been established in the AoR and Corrective Action
Plan. Methods for monitoring the evolution of the carbon dioxide plume and associated pressure
front are discussed in more detail in the UIC Program Class VI Well Testing and Monitor ing
Guidance. An example of evaluation of monitoring results during AoR reevaluation is provided
in Box 5-1.
The owner or operator is required to perform monitoring to track the extent of the carbon dioxide
plume and the presence or absence of elevated pressure [40 CFR 146.90(g)]. Pressure monitoring
is required using direct methods (e.g., pressure transducers) within the injection zone, and
indirect methods for plume tracking are also required unless the UIC Program Director
determines, based on site-specific geology, that such methods are not appropriate [40 CFR
146.90(g)(l) and (2)]. Additionally, the owner or operator is required to perform periodic
monitoring of ground water quality and geochemical changes above the primary confining zone
that may be a result of carbon dioxide movement through the confining zone [40 CFR
146.90(d)]. On a site-specific basis, the UIC Program Director may require additional
geochemical monitoring within the injection zone as a component of carbon dioxide plume
tracking [40 CFR 146.90(i)].
EPA recommends that pressure measurements indicative of pressure-front migration further than
that predicted by the current computational model warrant an AoR reevaluation. In practice, this
would be indicated by an observed increase in pressure at monitoring wells greater than
predicted by the computational model. In some cases, pressure measurements may fluctuate, and
UIC Program Class VI Well Area of Review 70
Evaluation and Corrective Action Guidance
-------
short-term temporary pressure increases (e.g., spikes) may not warrant an AoR reevaluation.
EPA recommends that the specific pressure monitoring results that would trigger an AoR
reevaluation be included in the AoR and Corrective Action Plan. For example, the owner or
operator may specify the magnitude and duration of increased pressure that would trigger an
AoR reevaluation for each monitoring well.
Results of carbon dioxide plume and pressure-front tracking using indirect methods, such as
periodic geophysical surveys, may also be used for comparison to model predictions.
Geophysical survey results provide information over relatively large areas, as opposed to "point"
measurements provided by monitoring wells. Geophysical survey results are intended to provide
an estimate of the extent of the separate-phase carbon dioxide plume and, in some cases, pressure
changes. EPA anticipates that results of indirect monitoring that indicate carbon dioxide
migration (1) outside of the boundaries of the current AoR delineation, or (2) at rates
significantly greater than current model estimates would also warrant an AoR reevaluation.
EPA also recommends performing an AoR reevaluation if the results of the ground water
geochemical sampling indicate separate-phase (i.e., supercritical, liquid, or gaseous) carbon
dioxide migration outside of the boundaries of the current AoR delineation, or at rates
significantly greater than predicted by the computational model. The presence of separate-phase
carbon dioxide in the sampled fluids above the confining zone is evidence of carbon
dioxide/fluid migration out of the injection zone and is cause to notify the UIC Program Director
pursuant to 40 CFR 146.91(c)(l). In addition, elevated carbon dioxide aqueous concentrations
may indicate the presence of separate-phase carbon dioxide in the immediate vicinity of the
monitoring well.
5.2.4. Ongoing Site Characterization
Site characterization is not a one-time exercise at GS project sites. As additional site
characterization data are collected via geophysical surveys, the drilling of new injection or
monitoring wells, or from other sources, the data must be subsequently incorporated into the
existing computational model used for AoR delineation [40 CFR 146.84(c)(l) and (e)(l)]. Types
of data that may be incorporated into a reevaluation include newly identified potential conduits
for fluid movement, updated information regarding injection or confining zone extent and
thickness, or further characterization of formation heterogeneity. The UIC Program Director may
also require an AoR reevaluation based on any newly available site characterization data that
may impact current modeling predictions.
5.3. Performing an AoR Reevaluation
The first step in performing an AoR reevaluation for a Class VI injection well is a comparison of
the available monitoring data and the model predictions. If Class VI owners or operators believe
that monitoring and modeling data are consistent and that revision of the model is not necessary,
they must demonstrate this to the UIC Program Director in lieu of revising the computational
model [40 CFR 146.84(e)(4)]. However, if monitoring data and modeling predictions differ
significantly, then the Class VI owner or operator must submit an amended AoR and Corrective
UIC Program Class VI Well Area of Review 71
Evaluation and Corrective Action Guidance
-------
Action Plan and revise both the computational model and the AoR delineation results [40 CFR
146.84(e)(l)and(4)].
Box 5-1. Hypothetical Example of an AoR Reevaluation
An AoR reevaluation consists of comparing monitoring results of plume and pressure-front
movement to model predictions. In this hypothetical example, a continuation of the scenario
presented earlier in Boxes 3-1 and 3-2, the AoR reevaluation required after 20 years of
injection is illustrated below. In this example, the previously required AoR reevaluations at 5,
10, and 15 years did not result in any AoR delineation modifications.
Comparison of Plume Monitoring Data
In this hypothetical scenario, monitoring data are available from eight monitoring wells
screened within the injection zone and from an indirect geophysical monitoring technique.
Monitoring well data are used to assess the potential presence of separate-phase carbon
dioxide at each location. The data indicate that separate-phase carbon dioxide is present at one
of the monitoring wells. These data are compared to initial model predictions of plume
evolution for 20 years after the commencement of injection (Figure 5-1). Carbon dioxide is
detected at MW-6, outside of the areas predicted by the model to exhibit carbon dioxide.
,
^**'*» - ^
\
MW-2
Injection
- ^^
'7
Explanation
Injection
well «M
O Injection well
MW-1
«.« "
0 Monitor well with detection of supercritical carbon dioxide
Geophysical monitoring results, supercritical plume extent
Model-predicted supercritical plume extent (20 years)
i area of review
well #3
SOUK* D*n«B.
<**.
Hypothetical Geologic Sequestration Site:
Comparison of Model Predictions and
Plume Monitoring Results at 20 Years of Injection
Figure 5-1: Hypothetical Geologic Sequestration Site: Comparison of Model Predictions and Plume
Monitoring Results at 20 Years of Injection.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
72
-------
Box 5-1. Hypothetical Example of an AoR Reevaluation, continued
Compared to monitoring well data, geophysical data provide a larger-area estimate of the
extent of separate-phase carbon dioxide. The geophysical and monitoring-well data are
consistent in their general evaluation of where separate-phase carbon dioxide is present.
Geophysical data and model results are generally consistent for the plume emanating from
Injection Well #3, and inconsistent for Injection Wells #1 and #2. The carbon dioxide plume
may have migrated differently than originally predicted for several reasons, as discussed
below.
Comparison of Pressure Monitoring Data
Bottom-hole pressure data are collected at all of the eight monitoring wells. This example
focuses on data collected at three of the wells, MW-1, MW-2, and MW-6. For actual projects,
EPA recommends that data from all monitoring wells be considered. Graphs of pressure
monitoring data over the first 20 years of the project, compared to modeling results, are
presented in Figure 5-2.
MW-1 /
/
A
15.0
£14.5
| 14.0 -
|«5
13.0
(
Well MW-1
n. -tVin rj ?
^^°
) 5 10 15 20
Time (years)
Model Resuilss Monitoring Dala
Injection
well #3
5 10 15 20
Time (years!
Model Results " Monitoring Data
Injection
well#1
O Injector, well
MW-7
Monitor well
Initially delineated area or review
Source: .O.M.I ;.' fl Stephens « Associates. Inc.
MW-8
*
Hypothetical Geologic Sequestration Site:
Comparison of Model Predictions and
Pressure Monitoring Results at 20 Years of Injection
Figure 5-2: Hypothetical Geologic Sequestration Site: Comparison of Model Predictions and Pressure
Monitoring Results at 20 Years of Injection.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
73
-------
Box 5-1. Hypothetical Example of an AoR Reevaluation, continued
Pressure monitoring data are consistent with modeling predictions on the western edge of the
project (MW-1). The general scatter in the monitoring data are expected, and there is no
significant bias (i.e., less than, greater than) in comparing the monitoring data and modeling
results. Data from the northern portion of the project (MW-2) indicate that actual pressure
increases in the injection zone are lower than model predictions. This area has exhibited less
of a pressure perturbation caused by injection than originally predicted. In contrast, data from
the eastern portion of the site (MW-6) indicate that there has been a larger pressure increase
than originally predicted. These data are generally consistent with the plume migration data,
presented above, which showed that the plume has migrated further east than originally
predicted.
Outcome of Monitoring Data and Model Comparison
This comparison indicates that, after 20 years of injection, modeling results and monitoring
data compare favorably in some regions of the site. However, the plume and pressure front
appear to have migrated further to the east than initially predicted. This disparity may be due
to several factors. Examples include the presence of a high-permeability pathway within the
injection zone that had not been fully characterized during initial site characterization, or the
dip angle at the injection zone/confining zone interface being larger than originally assumed.
Based on this comparison, the operator of the project site, in consultation with the UIC
Program Director, decided to calibrate the AoR model and re-delineate the AoR. See Box 5-2
for more information.
UIC Program Class VI Well Area of Review 74
Evaluation and Corrective Action Guidance
-------
5.3.1. Demonstrating Adequate Existing AoR Delineation
An AoR reevaluation does not necessarily need to result in revisions or updates to the site
computational model. If the owner or operator determines that no changes are necessary, the
required reevaluation may consist of demonstrating this to the UIC Program Director [40 CFR
146.84(e)(4)]. EPA recommends that demonstrating the adequacy of the current AoR delineation
includes verification that existing operational and site characterization data have been
incorporated into the model and that existing monitoring data agree with the modeled
predictions.
EPA recommends that the Class VI injection well owners or operators submit any new
operational, monitoring, or site characterization data that have been received since the last AoR
reevaluation to the UIC Program Director. EPA also recommends that details regarding how this
information has been incorporated into the site computational model be presented, as newly
received operational or site characterization data may impact model input parameter values.
Integral to demonstrating that the current AoR delineation is adequate is the comparison of
monitoring data and model predictions. EPA recommends that this comparison take the form of
graphics and informative maps showing the general agreement between monitoring results and
model predictions, and that all available monitoring data be considered, including fluid
geochemistry monitoring, pressure monitoring, and geophysical surveys.
5.3.2. Modifying the Existing AoR Delineation
Any significant differences between operational monitoring results and the existing model
predictions that are the basis for the AoR delineation, for example as discussed in Section 5.2.3
of this guidance document, warrant a modification to the existing AoR delineation [40 CFR
146.84(e)]. The steps in revision of the AoR delineation include adjusting the site conceptual
model, model calibration (i.e., adjusting model parameters), and presentation of adjusted model
results and the newly delineated AoR to the UIC Program Director.
EPA recommends that the site conceptual model be revised based on new site characterization,
operational, and, in some cases, monitoring data. The new conceptual site model schematic may
be provided to the UIC Program Director along with the AoR reevaluation information, with any
changes highlighted. Examples of changes to the conceptual model include new injection wells,
newly elucidated geologic features (e.g., stratigraphic layers), or a revised permeability field.
Following revision of the site conceptual model, revision of the existing AoR delineation may
require model calibration in order to minimize the differences between monitoring data and
model simulations (see Section 2.5 of this guidance). EPA recommends that the relative error
difference between monitoring data and model predictions be quantified via the use of
calibration statistics (e.g., ME, MAE, RMSE). To the extent possible, the value of the calibration
statistics should be minimized during model calibration. The value of the model calibration
statistics also informs the expected uncertainty and error in model predictions of future
UIC Program Class VI Well Area of Review 75
Evaluation and Corrective Action Guidance
-------
conditions. Following model calibration, the AoR delineation may be revised using methods
described in Section 3.4 of this guidance.
In reporting an AoR computational model and delineation revision, EPA recommends that all
model attributes, as given in Section 3.5 of this guidance document, be re-submitted to the UIC
Program Director. In addition, EPA recommends that the model calibration process and final
AoR delineation results be presented in detail as part of the submission with:
Adjusted input parameter values listed;
Graphs comparing observed and modeled values of carbon dioxide migration and fluid
pressure;
Model results showing carbon dioxide and pressure front migration over time included;
and
The value of the model calibration statistics.
The newly delineated AoR may be presented on maps which would highlight similarities and
differences in comparison with previous AoR delineations. See Box 5-2, below, for an example
of a hypothetical AoR reevaluation.
If a revision of the AoR delineation is necessary, an amendment to the AoR and Corrective
Action Plan is also required, along with possible amendments to other related project plans [40
CFR 146.84(e)(4) and (f)]. EPA recommends that the amended AoR and Corrective Action Plan
explain any differences in corrective action activities that result from AoR revision, including a
demonstration of adequate surface access rights in order to perform the required corrective action
activities. See Section 4 for more information on performing corrective action. Furthermore, in
some cases, GS project attributes that are outside the scope of the Class VI Rule and the UIC
Program, such as pore-space ownership rights, may be related to the size of the AoR. In these
cases, the owners or operators are encouraged to consult with the UIC Program Director, or
another applicable regulatory agency, following a revision of the AoR in order to proceed.
UIC Program Class VI Well Area of Review 76
Evaluation and Corrective Action Guidance
-------
Box 5-2. Hypothetical Example of a Presentation of the Revised AoR
After the site computational model has been revised through model calibration to monitoring
data, and/or updating with new operational or site characterization parameters, the AoR must
be re-delineated [40 CFR 146.84(e)]. See Box 3-2 for more information on AoR delineation;
the same general methods should be used during the reevaluation. Once the AoR has been
revised, it may be presented on a site base map in comparison to the former AoR delineation
(Figure 5-3).
In this hypothetical example, the AoR reevaluation has resulted in an AoR delineation that
extends generally farther toward the east than previously. This is consistent with the
monitoring data (Box 5-1) indicating further plume and pressure front migration toward the
east. The model was revised to match monitoring data by adjusting intrinsic permeability
values within the injection zone and the dip angle at the injection/confining zone interface.
The region newly identified as located within the delineated AoR (between the purple and
blue contour lines) must be subjected to the artificial penetration identification, assessment,
and corrective action procedures as discussed in Section 4 of this guidance document [40 CFR
146.84(e)(2) and (3)]. Furthermore, the revision of the AoR requires revisions to the AoR and
Corrective Action Plan and other project plans, as discussed in the UIC Program Class VI
Well Project Plan Development Guidance. Changes to the AoR and Corrective Action Plan
may demonstrate a need to secure new surface access rights for the newly included area. The
owner or operator may also contact the applicable regulatory agency for other project
attributes (e.g., new pore space ownership rights) that are outside the scope of the Class VI
Rule and the UIC Program.
Source: ftmte* B. Stephens A Assocuuvs. Inc.
Hypothetical Geologic Sequestration Site:
Initial Area of Review Delineation and After Reevaluation
Figure 5-3: Hypothetical Geologic Sequestration Site: Initial AoR Delineation and Delineation after
Reevaluation.
UIC Program Class VI Well Area of Review
Evaluation and Corrective Action Guidance
77
-------
References
Ajo-Franklin, J., Doughty, C., & Daley, T.M. 2008. Combining Analysis of Continuous Active-
Source Seismic Monitoring (CASSM) and Multiphase Flow Modeling for CO2
Sequestration: The Frio II Brine Pilot. Presented at the 7th Annual Conference on Carbon
Capture and Sequestration. May 5-8, Pittsburgh, Pennsylvania.
ASTM International. 2005. Standard guide for locating abandoned wells. Designation D 6285-99
(Reapproved 2005).
Bachu, S., & Bennion, B. 2008. Effects of in-situ conditions on relative permeability
characteristics of CO2-brine systems. Environmental Geology, 54:1707-1722.
Bandilla, K.W., Kraemer, S.R., & Birkholzer, J.T. 2012. Using semi-analytic solutions to
approximate the area of potential impact for carbon dioxide injection, International
Journal of Greenhouse Gas Control, 8:196-204, ISSN 1750-5836,
10.1016/j.ijggc.2012.02.009.
Birkholzer, J.T., Nicot, J.P., Oldenburg, C.M., Zhou, Q., Kraemer, S., & Bandilla, K. 2011.
Brine flow up a well caused by pressure perturbation from geologic carbon sequestration:
Static and dynamic evaluations. InternationalJournal of Greenhouse Gas Control,
5:850-861.
Carle, S.F. 1999. T-PROGS: Transition Probability Geostatistical Software, Version 2.1.
Celia, M. A., and Nordbotten, J.M. 2009. Practical modeling approaches for geologic storage of
carbon dioxide. Ground Water, 47:627-638.
Close, D., Cho, D., Horn, F., & Edmundson, H. 2009. The Sound of Sonic: A Historical
Perspective and Introduction to Acoustic Logging. CSEG Recorder pp. 34-43
Doughty, C., & Pruess, K. 2004. Modeling supercritical carbon dioxide injection in
heterogeneous porous media. Vadose Zone Journal, 3(3):837-847.
Doughty, C. 2007. Modeling geologic storage of carbon dioxide: Comparison of non-hysteretic
and hysteretic characteristic curves. Energy Conversion and Management 48: 1768-
1781.
Doughty, C., Freifeld, B.M., & Trautz, R.C. 2007. Site characterization for CO2 geologic storage
and vice versa - the Frio brine pilot, Texas, USA as a case study. Environmental
Geology, DOI 10.10007/s00254-007-0942-0.
Duguid, A., & Crow, W. 2007. CO2 Well Integrity and Wellbore Monitoring. Presented at the
Carbon Sequestration Leadership Forum (CSLF) 2nd Capacity Building in Emerging
Economies Workshop. October 18-19, Porto Alegre, Brazil.
UIC Program Class VI Well Area of Review 78
Evaluation and Corrective Action Guidance
-------
Fetter, C.W. 2001. AppliedHydrogeology, 4th Edition. Prentice Hall, Upper Saddle River, NJ.
598p.
Finsterle, S. 2004. Mulitphase inverse modeling: Review and iTOUGH2 applications. Vadose
Zone Journal, 3:747-762.
Fischer, D.W., LeFever, J.A., LeFever, R.D., Helms, L.D., Sorensen, J.A., Smith, S.A.,
Steadman, E.N., & Harju, J.A. 2005. Mission Canyon Formation Outline. PCOR
Partnership Fact Sheet, May 2005.
Flett M., Gurton, R., & Weir, G. 2007. Heterogeneous saline formations for carbon dioxide
disposal: Impact of varying heterogeneity on containment and trapping. Journal of
Petroleum Science and Engineering, 57:106-118.
Friskchknecht, F.C., Raab, P.V., & van Ee, JJ. 1985. Project Summary: Location of abandoned
wells with geophysical methods. EPA-600/SR-84-085.
Gaus, I, Azaroual, M., & Czernichowski-Lauriol, I. 2005. Reactive transport modeling of the
impact of carbon dioxide injection on the clayey cap rock at Sleipner (North Sea).
Chemical Geology, 217:319-337.
Gerritsen, M.G., & Durlofsky, LJ. 2005. Modeling fluid flow in oil reservoirs. Annual Reviews
in Fluid Mechanics, 37:211-238.
Green, D.W., & Perry, R.H. 2008. Perry's Chemical Engineers' Handbook, 8th Edition.
McGraw-Hill Book Co., New York, NY.
Gochioco, L.M., & Ruev Jr, F. 2006. Detecting and imaging hard-to-fmd abandoned wells and
pipelines. The Leading Edge, March 2006, 358-361.
Han, W.S., & McPherson, B. 2008. Comparison of two equations of state for application of
carbon dioxide sequestration. Advances in Water Resources, 31:877-890.
Ide, T., Friedmann, S.J., & Herzog, H. 2006. Carbon Dioxide Leakage through Existing Wells:
Current Technology and Regulations, presented at the 8th International Conference on
Greenhouse Gas Control Technologies. June 2006, Trondheim, Norway.
Illinois State Geological Survey (ISGS). 2009. Enhanced Oil Recovery I: Loudon Single-Well
Huff 'n' Puff. Midwest Geologic Sequestration Consortium. Report Issued December 31.
U.S. DOE Contract: DE-FC26-03NT41994.
Izgec, O., Demiral, B., Bertlin, H., & Arkin, S. 2008. CO2 injection into saline carbonate aquifer
formations II: Comparison of numerical simulations to experiments. Transport in Porous
Media, 73:51'-74.
UIC Program Class VI Well Area of Review 79
Evaluation and Corrective Action Guidance
-------
Jordan, P.W., & Hare, J.L. 2002. Locating Abandoned Wells: A Comprehensive Manual of
Methods and Resources. Prepared for Solution Mining Research Institute, Research
Project Report No. 2002-SMRI.
Juanes, R., Spiteri, E.J., Orr Jr., P.M., & Blunt, M.J. 2006. Impact of relative permeability
hysteresis on geological CC>2 storage. Water Resources Research., 42. W12418.
Knauss K.G., Johnson, J.W., & Steefel, C.I. 2005. Evaluation of the impact of CC>2, co-
contaminant gas, aqueous fluid and reservoir rock interactions on the geologic
sequestration of CC>2. Chemical Geology, 217:339-350.
Kumar, A., Noh, M., Pope, G.A., Sephehrnoori, K., Bryant, S., & Lake, L.W. 2004. Reservoir
Simulation of Carbon Dioxide Storage in Deep Saline Aquifers. SPE/DOW Fourteenth
Symposium on Improved Oil Recovery. April 2004, Tulsa Oklahoma.
Law, D.H.-S., & Bachu, S. 1996. Hydrogeological and numerical analysis of CO2 disposal in
deep aquifers in the Alberta sedimentary basin. Energy Conversion and Management,
37:1167-1174.
Lindeberg, E. 1997. Escape of CO2 from aquifers. Energy Conversion and Management,
38:8235-8240.
Midwest Geologic Sequestration Consortium (MGSC). Undated. Screening Criteria for Geologic
Carbon Dioxide Sequestration, http://www.sequestration.org/sc model.htm
Miller, C.T., Christakos, G., Imhoff, P.T., McBride, J.F., Pedit, J.A., & Trangenstein, J.A. 1998.
Mulitphase flow and transport modeling in heterogeneous porous media: challenges and
approaches. Advances in Water Resources, 21(2): 77-120.
National Energy Technology Laboratory (NETL). 2011. Risk Analysis and Simulation for
Geologic Storage of CO2. DOE/NETL-2011-1459.
http://www.netl.doe.gov/technologies/carbon seq/refshelf/BPM RiskAnalysisSimulation
.pdf
National Research Council (NRC). 2007. Models in Environmental Regulatory Decision
Making. Prepared by Committee on Models in the Regulatory Decision Process.
Nicot, J.P., Oldenburg, C.M., Bryant, S.L., & Hovorka, S.D. 2008. Pressure perturbation from
geologic carbon sequestration: Area-of-review boundaries and borehole leakage driving
forces. Energy Procedia.
Obi, E.-O. I, & Blunt, M.J. 2006. Streamline-based simulation of carbon dioxide storage in a
North Sea aquifer. Water Resources Research, 42: W03414.
UIC Program Class VI Well Area of Review 80
Evaluation and Corrective Action Guidance
-------
Peng, D.Y., & Robinson, D.B. 1976. A new two-constant equation of state. Industrial and
Engineering Chemistry Fundamentals, 15:59-64.
Perrin, J.-C., Krause, M., & Benson, S.M. 2008. Relative permeability properties of the
CO2/Brine system in saline aquifers: An experimental study. Presented at the 7th Annual
Conference on Carbon Capture and Sequestration. May 5-8, Pittsburgh, Pennsylvania.
Perry, R.H., & Green, D. 1984. Perry's Chemical Engineers' Handbook, Sixth Edition.
McGraw-Hill Book Co., New York, NY.
Finder, G.F., & Gray, W.G. 2008. Essentials of Multiphase Flow in Porous Media. John Wiley
& Sons, Inc., Hoboken, NJ.
Plug, W. J., & Bruining, J. 2007. Capillary pressure for the sand- CO2-water system under
various pressure conditions. Application to CO2 sequestration. Advances in Water
Resources. 30:2339-2353.
Pruess, K., and Spycher, N. 2007. ECO2N - A fluid property module for the TOUGH2 code for
studies of CO2 storage in saline aquifers. Energy Conversion and Management. 48:1761-
1767.
Pruess, K., Garcia, J., Kovscek, T., Oldenburg, T.C., Rutqvist, J., Steefel, C., & Xu, T. 2004.
Code intercomparison builds confidence in numerical simulation models for geologic
disposal of CO2. Energy, 29(9-10): 1431-1444.
Pruess, K. 2008. Leakage of CO2 from geologic storage: Role of secondary accumulation at
shallow depth. InternationalJournal of Greenhouse Gas Control, 2:37-46.
Randhol, P., Valencia, K., Taghipour, A., Akervoll, I, & Carlsen, I.M. 2007. Ensuring Well
Integrity in Connection with CO2 Injection. SINTEF report 31.6920.00/02/07.
Reynolds, R.R., & Kiker, R.D. 2003. Produced Water and Associated Issues A Manual for the
Independent Operator. Oklahoma Geological Survey Open File Report 6-2003. Prepared
for the South Midcontinent Region of the Petroleum Technology Transfer Council.
Rutqvst, J., Birkholzer, J.T., & Tsang, C.F., 2008. Coupled Reservoir-Geomechanical Analysis
of the Potential for Tensile and Shear Failure Associated with CO2 Injection in
Multilayered Reservoir Caprock Systems. InternationalJournal of Rock Mechanics and
Mining Sciences, 45(2): 132-143.
Saripalli, P., & McGrail, P. 2002. Semi-analytical approaches to modeling deep well injection of
CO2 for geologic sequestration. Energy Conversion and Management, 43:185-198.
Schnaar, G., & Digiulio, D.C. 2009. Computational modeling of the geologic sequestration of
carbon dioxide. Vadose Zone Journal, 8:389-403.
UIC Program Class VI Well Area of Review 81
Evaluation and Corrective Action Guidance
-------
Soeder, DJ. 1988. Porosity and Permeability of Eastern Devonian Gas Shale. SPE Formation
Evaluation. March 1988, 116-124.
Sorensen, J.A., Smith, S.A., Fischer, D.W., Steadman, E.N., & Harju, J.A. 2005. Potential
Carbon Dioxide Storage Capacity of the Saline Portions of the Lower Cretaceous Aquifer
System in the PCOR Partnership Region. PCOR Partnership Fact Sheet, October 2005.
Spycher, N., Pruess, K., & Ennis-King, J. 2003. Carbon dioxide-water mixtures in geologic
sequestration of carbon dioxide. I. Assessment and calculation of mutual solubilities from
12 to 100 degrees Celsius and up to 600 bar. Geochemical et Cosmochimica Acta,
67:3015-3031.
Spycher, N., & Pruess, K. 2005. Carbon dioxide-water mixtures in geologic sequestration of
carbon dioxide. II. Partitioning in chloride brines at 12-100 degrees Celsius and up to 600
bar. Geochemical et Cosmochimica Acta, 69:3309-3320.
Stone, H.L. 1970. Probability model for estimating three-phase relative permeability. Trans.
AIME 249:214-218.
Tamimi, A.E., Rinker, B., & Sandall, O.C. 1994. Diffusion coefficients for hydrogen sulfide,
carbon dioxide, and nitrous oxide in water over the temperature range 293-368 K.
Journal of Chemical & Engineering Data., 39:330-332.
Thornhill, J.T., Short, I.E., & Silka L. 1982. Application of the Area of Review Concept.
Ground Water, 20:32-38.
United States Environmental Protection Agency (USEPA). 2002. Drinking Water Academy,
Underground Injection Control Modules, Lesson 8: Area of Review.
http://www.epa.gov/safewater/dwa/pdfs/2%20uic%20modules/dwaUIC-uicpermit2.pdf
USEPA. 2003. Draft Guidance on the Development, Evaluation, and Application of Regulatory
Environmental Models. Prepared by The Council for Regulatory Environmental
Modeling.
United States Geological Survey (USGS). 1995. Fact Sheet 163-95: Magnetic Surveys for
Locating Abandoned Wells.
Watson, T.L. 2009. CC>2 Storage: Wellbore Integrity Evaluation and Integrity Across the
Caprock, presented at the 2009 SPE Conference on CO2 Capture. November 2-4, San
Diego, CA.
Xu, T., Sonnenthal, E., Spycher, N., & Pruess K. 2006. TOUGHREACTA simulation program
for non-isothermal multiphase reactive geochemical transport in variably saturation
geologic media: Applications to geothermal injectivity and CC>2 geological sequestration.
Computers & Geosciences, 32:145-165.
UIC Program Class VI Well Area of Review 82
Evaluation and Corrective Action Guidance
-------
Xu, T., Apps, J.A., Pruess, K., & Yamamoto, H. 2007. Numerical modeling of injection and
mineral trapping of CC>2 with H2S and 862 in a sandstone formation. Chemical Geology,
242:319-346.
Yamamoto, H., & Doughty, C. 2009. Investigation of gridding effects for numerical simulation
of carbon dioxide geologic sequestration," Proceedings of TOUGH Symposium, LBNL.,
September 2009, Berkeley, CA.
Zhang, K., Doughty, C., Wu, Y.-S., & Pruess, K. 2007. Efficient parallel simulation of CO2
geologic sequestration in saline aquifers, SPE 106026, presented at 2007 SPE Reservoir
Simulation Symposium. February 26-28, Houston, TX,.
Zhou, Q., Birkholzer, J.T., Tsang, C.F., & Rutqvist, J. 2008. A method for quick assessment of
CC>2 storage capacity in closed and semi-closed saline formations. InternationalJournal
of Greenhouse Gas Control, 2(4): 626-63 9.
Zhou, W., Stenhouse, M.J., Arthur, R., Whittaker, S., Law, D.H.-S., Chalaturnyk, R., & Jazrawi,
W. 2004. The IEA Weyburn CC>2 monitoring and storage project - modeling of the long-
term migration of CC>2 from Weyburn. In: Wilson M., Gale, J., Rubin, E.S., Keith, D.W.,
Gilboy, C.F., Morris, T., & Thambimuthu K. (Eds). Greenhouse gas control technologies:
proceedings of the 7th International Conference on Greenhouse Gas Control
Technologies. September 5-9, Vancouver, BC.
UIC Program Class VI Well Area of Review 83
Evaluation and Corrective Action Guidance
------- |