|
Items in cart [4] |
Questions? Call 888-624-8373 |
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 93
Knowledge Gaps and
Research Needs
The statement of task for this study directed the committee to identify
significant knowledge gaps relevant to subsurface contamination prob-
lems at DOE sites and to provide recommendations for a long-term
basic research program to fill those gaps (see Sidebar 1.1~. In this
chapter, the committee identifies what it judges to be the significant
knowledge gaps that emerged from its review of DOE's subsurface
contamination problems in Chapter 2 and, for each identified gap, the
committee provides a short discussion of basic research needs. This
information will be used to formulate recommendations for a long-term
research program in Chapter 6.
For purposes of this discussion, the committee defines "knowledge
gap" as a deficiency in scientific or engineering understanding that is
now, or likely will be in the future, a significant impediment to DOE's
efforts to complete its mission to clean up, stabilize, or contain subsur-
face contamination. Perhaps the most direct manifestation of a DOE
knowledge gap is a technology gap, that is, a deficiency in technical
capabilities to identify and deal with contamination problems. The
committee has not focused on technology gaps in this report; that is the
topic of another recent N RC report (N RC, 1 999~. Rather, the committee
has focused on the identification of the knowledge gaps that underpin
those technology gaps.
The committee has been selective in the identification of subsurface
contamination knowledge gaps and research needs for the EM Science
Program. The identification of knowledge gaps involves an appreciable
element of judgment on the part of the committee, especially in inter-
preting the significance of the subsurface contamination problems (see
Chapter 2) and the scope and objective of other federal research pro-
grams. The committee believes that the gaps it has identified are highly
C h a p t e r 5
93
OCR for page 94
significant and that they must be addressed through basic research if
the DOE cleanup program is to succeed. Further, the committee
believes that a focus on these knowledge gaps is likely to yield the
greatest payoffs for DOE in terms of enhanced cleanup capabilities at
reduced costs and risks at major DOE sites. This is especially true
given the small size of the EM Science Program relative to the scope of
the DOE cleanup mission. The annual budget for the EM Science
Program budget is on the order of $30 mi l l ion to $50 mi l l ion and is
used to support basic research related to all aspects of the cleanup mis-
sion. This is less than 0.1 percent of the total EM annual budget of $5.8
hi l l ion. Without a sign if icant i ncrease i n its budget, the EM Science pro-
gram is unlikely to have a significant impact on DOE cleanup effective-
ness and costs.
Organizing Scheme Used in This
Analysis
The committee identified significant knowledge gaps and research
needs through discussions and analyses of the "snapshot" of DOE's sub-
surface contamination problems presented in Chapter 2. To organize
this analysis and ensure its completeness, the committee developed the
organizing scheme shown in Figure 5.1. This organizing scheme is
O O O O O
. . , . , . . , , ,, ~ , , ~ , .
based partly on the approach used by the Subsurface contaminants
Focus Area to organize its technology development programs (see
Figure 3.2), but it also includes the data collection and analysis steps
that provide the supporting information needed to make appropriate
corrective action decisions.2 The committee's organizing scheme, here-
after referred to as the framework for site remediation, is described
briefly in the following paragraphs.
Was discussed in Chapter 2, the major sites represent DOE's largest future mort-
gages and longest-term commitments.
2The committee uses the term "corrective action" in the following discussion to
refer to actions taken by DOE to address its subsurface contamination problems. A
corrective action can range from no action in cases where the subsurface contami-
nation is thought to pose minimal hazards to humans or the environment, or
where remediation is infeasible, to aggressive actions to treat, remove, or contain
contamination that poses significant hazards. As noted in Chapter 1, the term is
sometimes used interchangeably with terms like "cleanup" and "remedial action,"
but it really encompasses a broader range of possible options for dealing with con-
tamination, because it includes the no action (i.e., no cleanup or no remedial
action) option.
S U B S U R F A C E S C ~ E N C E
OCR for page 95
6.
Corrective
Action
A. Take no L
| remedial action |
characterize
model behavior
1. :
~-1
2.
Assess
risk
Validate performance of remedial action
The boxes in Figure 5.1 represent each of the major steps in the
remediation process, and the arrows represent decision and assessment
points. Boxes 1 through 5 represent the process that could be followed
to develop information to make an appropriate corrective action deci-
sion. The initial step is focused on locating and characterizing the conta-
minants of concern (Box 1~. This step involves determining the spatial
distributions, types, amounts, and physical and chemical states of subsur-
face contaminants, as well as the subsurface properties that affect con-
taminant fate and transport behavior. Locating and characterizing con-
tamination in the subsurface may be done using direct (e.g., drilling and
sampling) and indirect (e.g., surface and borehole geophysical) techniques.
The location and characterization data obtained in the first step are
then used to develop a conceptual model of the site (Box 2), that is, a
description of the subsurface as estimated from knowledge of the
known site geology and hydrology and the physical, chemical, and bio-
logical processes that govern contaminant behavior. The conceptual
model provides a descriptive framework for assessing how the subsur-
face system will behave with passing time and in response to potential
corrective actions. As noted later in this discussion, the conceptual
model is improved over time as more information on subsurface condi-
tions and processes becomes available.
The conceptual model provides a basis for constructing more quan-
titative dynamic models that can be used to predict behavior (Box 3) of
C h a p t e r 5
it{
I B. Contain & |
| stabilize r
C. In situ treat
or transform
| D. Remove
| hotspots
E. Remove
contamination r
Monitor
FIGURE 5. 1 Framework for site
remediation.
95
OCR for page 96
the subsurface system over a specified time period. These predictive
models are developed from mathematical representations of the con-
ceptual model. In current practice, most predictive models are discrete
representations of the physically continuous subsurface system and are
typically solved with such numerical techniques as finite elements and
finite difference.3 The parameters in these models represent the physi-
cal, chemical, and biological characteristics at each point in the subsur-
face. Parameters of interest in predictive models quantify the relation-
ship between the driving forces (e.g., hydraulic gradient and chemical
concentration gradient) and the resulting behavior (e.g., flow and trans-
port). In the case of discrete models, the parameter values are meant to
represent volume-averaged properties around each modeled point.
The predicted system behavior is then compared to the observed
behavior as measured in the field through monitoring activities (Box 4~.
A feedback loop (Arrow 1 ) updates the conceptual and predictive mod-
els when the behaviors do not match according to some specified mea-
surets) of comparison. The process of testing the predictive model to
determine whether it appropriately represents the system behavior of
interest is usually referred to as model validation.4 Because of uncer-
tainties in the model and data any match between predicted and
observed behaviors is only possible in a statistical sense. Consequently,
validation is best thought of as a process of confidence building
through increased understanding of fundamental mechanisms in the
underlying system rather than as a process to confirm or prove the cor-
rectness of a model.
The predictive model can be used to understand the present behav-
ior of the subsurface system and to estimate future contaminant migra-
tion to assess risk to human and environmental health (Box 5~. A cor-
rective action decision (Box 6A-E) that will reduce risk to acceptable
levels is then made using the information developed in the risk assess-
ment. The corrective action can range from no action (Box 6A) to
remove contamination (Box BE).
During and following the corrective action, monitoring activities
(Box 7) are again employed to assess the efficacy of that action. Long-
term monitoring is usually required to confirm the effectiveness of, or to
gain regulatory approval for, approaches that involve no action or con-
tainment and stabilization. Inconsistencies between the measured and
3Continuous representations are sometimes used in analytical models for
screening-level assessments.
41n model development protocol, a step referred to as model verification
involves evaluating whether the numerical model solves the mathematical equa-
tions of the conceptual model with acceptable accuracy. In this discussion, model
verification is included as a step in the validation process.
S U B S U R F A C E S C ~ E N C E
96
OCR for page 97
predicted performance of the corrective action may indicate that the
conceptual model of the system is deficient or that the parameters of
the model are not sufficiently resolved, and it may be extremely diffi-
cult to know which is the case. In these cases, there is a feedback loop
(Arrow 2) to the conceptual model (Box 2) through the predictive
model, which must be updated so that the corrective action decision
process can be revisited.
Although the framework for site remed iation shown i n Figu re 5.1 is
presented as a linear process, it is in reality an observational procedure
that follows both parallel and iterative paths. The framework may be tra-
versed many times as new information is acquired and incorporated
into the conceptual and predictive models and as corrective action per-
formance is assessed.
There is some correspondence between the organizing scheme out-
lined in Figure 5.1 and the technology development organizing scheme
shown in Figure 3.2. For example, the identify function in Figure 3.2 is
roughly equivalent to the locate and characterize function in Figure 5.1.
Similarly, the validate function in Figure 3.2 is roughly equivalent to the
validate performance of remedial action function (Arrow 2) in Figure
5.1. The remaining functions in Figure 3.2 have no directly equivalent
functions in Figure 5.1, and there are many functions in Figure 5.1 that
are not represented at all in Figure 3.2 (e.g., the develop conceptual
model and predict system behavior functions). The organizing scheme
shown in Figure 5.1 is more complete than that given in Figure 3.2.
Knowledge Gaps
The committee identified significant knowledge gaps in the follow-
ing process steps in the framework for site remediation shown in Figure
5.1:
location and characterization of subsurface contaminants and
characterization of the subsurface (Box 1~;
conceptual modeling (Box 2~;
containment and stabilization (Box 6B); and
· monitoring and validation (Boxes 4 and 7 and Arrows 1 and 2~.
These knowledge gaps do not include those associated with active
remediation of subsurface contamination (Boxes 6C-6E), with the
exception of remediation monitoring. This may come as a surprise to
some readers, given that the current EM Science Program portfolio is
heavily focused on this area (see the closing section of Chapter 3~. The
C h a p t e r 5
OCR for page 98
committee did not highlight knowledge gaps on these process steps,
because subsurface contamination is highly distributed at many DOE
sites, making cost-effective remediation infeasible, and because EM
Science Program resources are limited and there is much work on these
topics in other federal research programs (see Chapter 4~.
Location and Characterization of Subsurface
Contaminants and Characterization of the Subsurface
An important conclusion that emerges from the committee's analysis
of subsurface contamination problems in Chapter 2 is that the capabili-
ties to locate and characterize subsurface contaminants at many DOE
sites are incomplete. This conclusion is perhaps best supported by the
following three examples from Chapter 2:
subsurface radionuclide contamination in the 200 Area at
Hanford;
mixed contam
and
contaminant plumes and hot spots in waste burial grounds at the
Savannah River Site.
inant plumes at Test Area North at the Idaho Site;
Locating contamination in the subsurface at DOE sites has focused
on three interrelated approaches: (1 ~ information derived from h istorical
operations and records; (2) direct observations of contamination on the
surface, in surface water, and in boreholes; and (3) indirect geochemi-
cal and geophysical measurements from the surface and in boreholes.
These three sources of information have been used at some sites to
develop predictive models of contaminant movement in the subsurface
and the predictive models have been tested by further direct observa-
tions and measurements. Frequently, these models have not captured
the essential behavior of the contaminant, either in direction or speed
of movement.
The challenges of locating subsurface contamination are magnified
by the wide range of contaminant types (e.g., mixtures of organic sol-
vents, metals, and radionuclides) in the subsurface at many DOE sites
(see Chapter 2~; the wide variety of geological and hydrological condi-
tions across the DOE complex (see Table 2.2~; and the wide range of
spatial resolutions at which this contamination must be located and
characterized, ranging from widely dispersed contamination in ground-
water plumes to small isolated hot spots in waste burial grounds.
Moreover, because contaminant migration involves dynamic transport
processes, continuous temporal information on contaminant locations
is required. In effect, location, characterization, and continuing moni-
S U B S U R F A C E S C ~ E N C E
98
OCR for page 99
taring efforts must be integrated to assure an adequate database for
planning and implementing appropriate corrective actions.
Fundamental advances in capabilities to locate and characterize
subsurface contamination and important subsurface properties will help
DOE better assess the potential hazards of its contamination problems
and to design and implement appropriate corrective action strategies
(e.g., see Sidebar 5.1~. Moreover, research on subsurface heterogeneity
in geology, geochemistry, hydrology, and microbiology will provide a
framework for assessing the fate and transport of contaminants.
Examples of significant knowledge gaps include the following:
· Locating contaminants in the subsurface. At many sites, the
points of entry of contaminants into the subsurface (e.g., through
a leaking waste burial ground or injection well) are at least
approximately known. However, the determination of the spatial
distributions of contaminants (that may or may not change with
time) once they enter the subsurface remains a major knowledge
gap. Currently available indirect measurement methods (e.g.,
geophysical methods) are inadequate for locating most types of
contaminants in the subsurface, and direct methods such as
drilling are both expensive and limited in effectiveness, because
they only provide samples from specific points in the subsurface
along the borehole. Moreover, boreholes provide potential con-
taminant transport pathways through the subsurface.
· Characterizing contaminants in the subsurface. Once contami-
nants enter the subsurface, they can act as long-term sources of
pollution to ground or surface water. Understanding how to
characterize the concentrations, speciations, and release rates of
contaminants in the subsurface is a significant knowledge gap
across the DOE complex. In general, there are poor records of
contaminant discharges to the subsurface, so contaminant quan-
tities are highly uncertain. Moreover, once contaminants enter
the subsurface they can move long distances, either diffusing
through the fluid medium or migrating as a distinct plume, lead-
ing to contaminant distributions that are variable in size, shape,
and concentration. Currently available direct and indirect
observing technologies5 have limited effectiveness for character-
5Direct observing technologies allow in situ measurements or samples to be
obtained (e.g., by drilling). Indirect observing technologies allow measurements
to be made remotely (e.g., through geophysical measurements of the subsurface).
The terms "invasive" and "noninvasive" are sometimes used synonymously, but this
usage is not strictly correct. Indirect measurements can be obtained through inva-
C h a p t e r 5
99
OCR for page 100
.
.
izing site conditions and defining the extent and concentrations
of contaminant bodies.
Characterizing physical, chemical, and biological properties of
the subsurface, including improved approaches to understanding
the properties of the geologic system and relating them to conta-
minant fate and transport. The subsurface characteristics at a site
place fundamental controls on contaminant fate and transport
behavior. Subsurface characteristics also govern the selection of
conceptual and predictive models as well as the application and
effectiveness of appropriate corrective actions. The knowledge
gaps include understanding which characteristics control fate
and transport behavior in the subsurface and also understanding
how those characteristics can be measured at the appropriate
scales over large subsurface volumes, using both indirect and
direct techniques. The integration of direct measurements of sub-
surface geologic properties with indirect measurements (e.g.,
from geophysical methods) has been used very successfully in
the petroleum industry to develop conceptual and quantitative
models of subsurface transport. Such methods are potentially
applicable to DOE sites.
Characterizing highly heterogeneous systems. This knowledge
gap is a special case of the previous knowledge gap and is a sig-
nificant problem at many DOE sites, which are very large in spa-
tial extent and exhibit intra- and inter-site variations in geologic
and hydrologic conditions (see Chapter 2~. Heterogeneity arises
from the spatial variability in geological, chemical, and biologi-
cal properties of the subsurface. A fundamental understanding of
these properties, and especially the geological framework, is a
necessary prerequisite to understanding the fate and transport of
contaminants. Heterogeneity may occur at several spatial scales
in complex subsurface systems, but they may control contami-
nant fate and transport processes only at one or a few scales. The
primary knowledge gaps are in understanding the heterogeneity
scales that govern these processes, how to characterize this het-
erogeneity without having to perform an exhaustive characteriza-
tion of the subsurface, and how to represent this heterogeneity in
mathematical formulations.
Research Needs
In the committee's judgment, basic research can support the devel-
opment of new and improved capabilities to locate and characterize
sive means, as when borehole geophysical methods are employed to obtain sub-
surface measurements.
S U B S U R F A C E S C ~ E N C E
100
OCR for page 101
contamination in the subsurface, and also to characterize subsurface
properties at the scales that control contaminant fate and transport
behavior. Development of the following capabilities is especially needed:
Improved capabilities for characterizing the physical, chemical,
and biological properties of the subsurface. These approaches
should provide information on the following system properties
and behaviors at the spatial and temporal scales that control
contaminant fate and transport behavior:
· contaminant locations and characteristics:
· transport pathways;
· subsurface properties and boundary conditions that control
contaminant fate and transport behavior; and
· physical, chemical, and biological interactions between con-
taminants and earth materials.
SIDEBAR 5.1 NEW APPROACHES FOR DIRECT OBSERVING
The major limitations on direct observations by conventional drilling and sampling have been high
costs and concerns that direct approaches may unwittingly exacerbate the spread of contaminants in
the subsurface.The use of reduced diameter driliholes (using 4- to 6-inch diameter drills) as a cost-sav-
ing method has been explored widely in the petroleum industry, but cost reductions have not been
encouraging. However, recent developments in miniaturized drilling and sampling technologies (e.g.,
Albright and Dreesen, 2000) hold promise for significantly reducing drilling costs and reducing the
potential for contaminant spread when these technologies are used at DOE sites.
A new technology, microdrilling, represents the kinds of advanced capabilities made possible by basic
scientific and engineering research.This technology uses coiled tubing, steerable miniature-diameter
(1 3/8 inches to 2 inches [3.5 centimeters to 5.1 centimeters]) down-hole motors, and down-hole micro-
instrumentation to obtain in situ measurements and samples of contaminated subsurface environ-
ments. Additionally, smaller diameter holes reduce contaminant migration potential and promote
more effective sealing.
The depth capabilities thus far demonstrated are adequate for almost all of the major DOE sites (down
to about 300 meters, or about 1,000 feet). Many aspects of this microborehole technology still require
extensive research and development, including work on sampling techniques, down-hole instrumenta-
tion for diverse measurements, and effective plugging; however, enough feasibility demonstrations
have been completed to indicate great promise for use at DOE and other contaminated sites.
Albright and Dreesen (2000) suggest that this technology may cut drilling costs by at least 70 percent
compared to conventional technologies.They also suggest that much greater cost savings are possible
as these techniques are refined.
C h a p t e r 5
101
OCR for page 102
Research on indirect observations could involve the develop-
ment of new approaches for measuring contaminant and subsur-
face properties (e.g., approaches utilizing "unconventional" geo-
physical wave attributes such as polarized and nonlinear wave
responses) or new ways of interpreting "conventional" observa-
tional data to obtain information on the system properties of
interest. For direct observations, the research must also address
how the observing process changes the system being measured.
Approaches for making direct and indirect observations in the
unsaturated zone are especially needed.
2. Improved capabilities for characterizing physical, chemical, and
biological heterogeneity, especially at the scales that control
contaminant fate and transport behavior. Approaches that allow
measurements or estimates of heterogeneity features to be
obtained directly (i.e., without having to perform a detailed char-
acterization of the subsurface) are especially needed.
3. Improved capabilities for measuring contaminant migration and
the system properties that control contaminant movement.
4. Methods to integrate data collected at different spatial and tem-
poral scales to better estimate contaminant and subsurface prop-
erties and processes, and also methods to integrate such data
into conceptual models.
Conceptual Modeling
As shown by several examples in Chapter 2, DOE is finding subsur-
face contamination in unexpected places:
.
Technetium was discovered in groundwater beneath the SX Tank
Farm in the 200 Area at the Hanford Site.
· Plutonium was discovered in colloids in groundwater near the
Benham Test at the Nevada Test Site.
Plutonium was discovered in groundwater beneath the
Radioactive Waste Management Complex at the Idaho Site.
.
These discoveries were "unexpected" because models of the subsur-
tace at these sites did not predict them (e.g., see Sidebar 2.6~. Concep-
tual and predictive models have been developed for subsurface conta-
minant fate and transport for many DOE sites, but in many cases these
models have proven ineffective for understanding and predicting conta-
minant movement, especially at sites that have thick unsaturated zones
or complex subsurface characteristics.
The conceptual model "problem" has many possible causes. The
models themselves may be deficient because they were developed
S U B S U R F A C E S C ~ E N C E
102
OCR for page 103
using insufficient data on subsurface characteristics, contaminant distri-
butions, or transport processes, or the models may simply have an inap-
propriate theoretical basis. Good conceptual models must be grounded
in sound theory and underpinned with sound and sufficient data. In the
committee's judgment, at least part of the problem is that conceptual
model development is not viewed as an explicit part of remediation
practice. Consequently, there are few standardized tools or accepted
methodologies for developing such models, which has led to ad hoc
and inconsistent approaches across DOE sites.
Accurate conceptualizations are essential for understanding the
long-term fate of contaminants in the subsurface and the selection and
application of appropriate corrective actions. The significant knowledge
gaps include the following:
.
· Contaminant fate and transport. Understanding the factors con-
trolling the long-term fate of contaminants in the subsurface is
important for assessing the potential for human and ecological
exposure and for selecting appropriate corrective actions.
Understanding the dominant contaminant transport processes
and pathways through the subsurface remains a significant
knowledge gap for building accurate and useful conceptual and
predictive models. The simplest formu ration of contaminant
transport uses porous media flow of a dissolved phase, but such
transport may be the exception at many DOE sites, where trans-
port can occur in several distinct manners (e.g., colloidal trans-
port) through both porous media and fractures and may involve
a variety of chemical and biological reactions. The myriad chem-
ical, biological, and physical processes operating in the subsur-
face operate at different time scales and are poorly understood,
especial Iy for metals and radionucl ides.
Coupling physical, chemical, and biological processes. The physi
Cal, chemical, and biological properties and processes that gov-
ern contaminant fate and transport do not act independently.
Rather, they interact (i.e., they are coupled) in complex and often
poorly understood ways. Many coupled processes operate over
very small spatial scales that are defined by a distribution of prop-
erties, making it difficult to incorporate representations of these
processes into conceptual and mathematical models. For exam-
ple, redox potential and pH (chemical properties related to bulk
mineralogy, biological activity, and fluid composition) can affect
either or both physisorption and chemisorption of contaminants
onto solid phases. The heterogeneous distribution of permeability
(a physical property related to the geological characteristics of the
C h a p t e r 5
103
OCR for page 104
subsurface) can result in highly variable rates of fluid flow (a
physical process). These processes combine to effect transport (a
coupled process) of certain metals and radionuclides over small
spatial scales. Similarly, the coupling of biomass availability (a
property with biological, physical, and chemical components)
and substrate availability (controlled by processes such as sorp-
tion, dissolution, and transport) with the distribution of electron
acceptors (also possessing biological, physical, and chemical
. . ~
.
controls) can result in spatially variable rates of in situ contami-
nant biodegradation (a coupled process). The coupling of process-
es and their control by subsurface properties are only beginning
to be understood. Moreover, little progress has been made on
how to represent coupled processes in predictive models.
Mode/ parameter development. Model parameters are well
understood and definable for very simple homogeneous subsur-
face systems. However, in highly complex subsurface systems,
parameter definition may require unobtainable amounts of
detailed characterization data. In these cases, it is important to
understand which processes are actually dominating the behav-
ior of the system and to define parameters appropriate to those
processes. Determ i n i ng how to make the appropriate si mpl if ica-
tions and approximations is the main thrust of conceptual mod-
eling research that leads to the identification of appropriate
model parameters.
The definition and estimation of model parameters requires a good
understanding of the subsurface system and transport processes being
modeled, which is not often the case at DOE sites. For example, the tra-
ditional approach for modeling porous media is to choose permeability
as a model parameter. If the porous medium is highly heterogeneous
(e.g., if it contains a few large and interconnected fractures) then the
generalized concept of permeability is not well defined, and permeabil-
ity may not be an appropriate characterization of the physical system.
Flow and transport may be dominated by the fracture system, and the
model parameters should represent the properties of these permeable
and connected pathways. Similarly, for fate and transport models, the
traditional approach is to assume equilibrium sorption and use the
equilibrium partition coefficient as a model parameter. If the sorption
reactions are not at equilibrium, however, then the equilibrium partition
coefficient by itself is not an appropriate parameter, and additional
parameters describing mass transfer kinetics must also be included. The
challenge is to define the right conceptualization of the physical, chem-
ical, or biological processes that dominate system behavior, which in
S U B S U R F A C E S C ~ E N C E
104
OCR for page 105
turn defines the appropriate model parameters to be used.
The field observations used to develop parameter estimates are made
at many different scales and times and provide information about differ-
ent properties of the subsurface system. Samples from drillhole core, for
example, can provide detailed information on the physical, chemical,
and biological properties of the subsurface at small (centimeter) spatial
scales. Borehole testing data (e.g., hydraulic pumping tests and tracer
tests) and indirect observations (e.g., seismic surveys) provide indirect
measurements of subsurface properties averaged over much larger
(meters to tens of meters) spatial scales. Observations of a given sub-
surface region using different measurement techniques can yield very
different results, and measurements from a single technique can show
significant variations over small spatial scales. One of the primary knowl-
edge gaps for model conceptualizations is understanding how to inte-
grate these field observations into the models and parameter estimates.
The knowledge gaps include understanding the scale effects and devel-
oping methods for data integration that take these effects into account.
Research Needs
Conceptual model development has not been an explicit topic for
basic research in its own right. Indeed, conceptual model development
is viewed as an inherently empirical and site-specific process using
observational approaches that are not easily generalized or tested. The
committee believes, however, that basic research that addresses the fun-
damental approaches and assumptions underlying conceptual model
development could produce a tool box of methodologies that are
applicable to contaminated sites both inside and outside the DOE com-
plex. This research should focus on the following topics:
New observational and experimental approaches and tools for
developing conceptual models that apply to complex subsurface
environments, including such phenomena as colloidal transport
and biologic activity.
2. New approaches for incorporating geological, hydrological,
chemical, and biological subsurface heterogeneity into concep-
tual model formulations at scales that dominate flow and trans-
port behavior.
3. Development of coupled-process models through experimental
studies at variable scales and complexities that account for the
interacting physical, chemical, and biological processes that gov-
ern contaminant fate and transport behavior.
4. Methods to integrate process knowledge from small-scale tests
and observations into model formulations, including methods for
C h a p t e r 5
105
OCR for page 106
incorporating qualitative geological information from surface and
near-surface observations into conceptual model formulations.
5. Methods to measure and predict the scale dependency of para-
meter values.
6. Approaches for establishing bounds on the accuracy of parame-
ters and conceptual model estimates from field and experimental
data.
The research needs outlined above call for more hypothesis-driven
experimental approaches that address the fundamental methods and
assumptions underlying the development of conceptual models. This
research will require expertise from a wide range of disciplines and
must be conducted at scales ranging from the laboratory bench top to
contaminated field sites.
Moreover, to have long-term relevance to the DOE cleanup mission,
this research must be focused on the kinds of subsurface environments
and contamination problems commonly encountered at major DOE
sites. One way to ensure this focus is to give researchers the opportuni-
ty to conduct research at contaminated DOE sites. The committee pro-
vides additional comments on this issue in the next chapter.6
Containment antl Stabilization
As noted by DOE in Paths to Closure (DOE, 1 998a) and as shown in
Chapter 2 of this report, a great deal of subsurface contamination is
likely to remain at DOE sites even after DOE's cleanup program is com-
pleted. It will include contaminant plumes in groundwater, contaminat-
ed soil, and waste burial grounds both the historical burial grounds
discussed in Chapter 2 and new burial grounds developed by DOE to
dispose of waste from its current and future cleanup operations. DOE is
responsible for the long-term management of this contamination and
must develop methods to contain and stabilize it until it no longer
poses a hazard to humans or the environment or until new methods
to remed late th is contami nation are developed. DOE's management
commitment potentially extends for many thousands of years. DOE's
containment and stabilization systems are likely to include surface
caps, subsurface barriers, and other in situ stabilization systems. Once
installed, these systems will have to be monitored to assure that they
perform as expected, and if these systems fail, additional corrective
actions may have to be taken to repair the barriers and remediate resid-
ual contamination. There has been an increasing emphasis and accep-
6See the section titled "Field Sites" in Chapter 6.
S U B S U R F A C E S C ~ E N C E
106
OCR for page 107
tance of waste contai n ment and stabi I ization i n recent years, both i n
DOE and by regulatory agencies. Decreasing cleanup budgets, evalua-
tions that show that containment is a low-risk choice for some prob-
lems, and recognition that some contamination cannot be remediated
either with current technologies or conceivable new technologies are
responsible for this change in phi losophy. This shift in emphasis is per-
haps first fully acknowledged by DOE in Paths to Closure (DOE,
1 998a), which lays out DOE's cleanup objectives, and appears to be a
developing trend across the DOE complex.7 A more recent DOE report
(DOE, 1 999) discusses the long-term stewardship chal lenges.
At some sites, containment and stabilization may be an interim
measu re and has its own set of associated techn ical problems. These
include particularly the availability of appropriate technologies to both
contain and stabilize the residual contamination and to monitor and
validate the long-term performance of containment and stabilization
systems themselves. There is little understanding of the long-term per-
formance of containment and stabilization systems, and there is a gen-
eral absence of effective methods to validate that such systems are
properly installed or that they can provide effective long-term perfor-
mance. To address this knowledge gap, advances in basic knowledge to
support the development of new and improved waste containment and
stabilization systems will be needed, as noted below.
The development of improved and novel containment and stabiliza-
tion approaches will likely have the highest potential for cost savings
and lowered risk of the four knowledge gaps identified by the commit-
tee. The committee believes that the significant knowledge gaps include
the following:
· Development of robust physical, chemical, and biological con-
tainment and stabilization systems. Trad itional contai n ment sys-
tems comprised of surface caps, in situ walls, and bottom barri-
ers employ low-permeability materials to reduce water infiltra-
tion and provide a barrier to contaminant migration. When
designed properly, these systems may provide effective contain-
7Another recent example of the shift in emphasis to containment strategies can
be found in a recent report on disposal of DOE low-level waste (DOE, 1 998e).
This report shows that DOE's estimates of the volume of its low-level waste requir-
ing disposal between 1998 and 2070 has decreased from about 32 million cubic
meters to about 8 million cubic meters, largely because DOE has decided to con-
tain much of this low-level waste in place at its sites, rather than removing it for
treatment or disposal elsewhere in the complex. Most of this waste exists in waste
burial grounds at the major DOE sites (see Chapter 2).
C h a p t e r 5
107
OCR for page 108
ment for periods of up to a few decades,8 but current designs do
not meet DOE's needs for containment of its long-lived radioac-
tive and hazardous waste both for wastes contained in place
and new waste sites developed from current and future cleanup
operations. Natural low-permeability materials for minimizing
infiltration (e.g., clays) work well in humid environments, but
they may not be effective in arid regions, where dessication can
lead to the development of preferred pathways.
To the committee's knowledge, there has been little or no
research or development work on longer-term systems for con-
tainment of subsurface contamination of the sort encountered at
DOE sites, either by DOE or by other organizations.9 The
knowledge gaps include understanding how to design more
effective and permanent barrier systems for long-term contain-
ment, especially in arid environments characteristic of the west-
ern DOE sites, including the development and application of
more durable materials for barrier systems materials that are
compatible with the surrounding environment and with the
waste that is being contained.
New containment approaches. Conventional barrier systems
seek to m i n i m ize water i nfi Itration i nto the contai ned waste and
to minimize the spread of waste from containment zones into
the environment. Surface barrier systems (caps) have proven very
.
effective for retarding water infi Itration into containment zones,
but they require ongoing maintenance to ensure their continued
integrity, and they have short lives relative to the hazard of the
contained waste. Moreover, subsurface infi Itration barriers (e.g.,
impermeable walls installed around or beneath waste burial
grounds) are extremely difficult to install and maintain, especial-
ly barriers emplaced beneath waste containment zones, and
their performance is also extremely difficult to monitor.
New approaches are needed to address DOE's needs for
long-term in situ containment and treatment of subsurface conta-
mination. The recent development of reactive barriers (i.e., barri-
ers that degrade or immobilize contaminants through geochemi-
cally and biochemically mediated reactions, such as ion
To the committee's knowledge, this supposition has never been tested at a
DOE site, so the actual longevity of such barrier systems is uncertain at best.
9There has been a great deal of research and development work in the United
States and other countries on long-term containment systems for spent fuel and
vitrified high-level waste, but this work does not appear to be directly applicable
to the contamination problems at DOE sites.
S U B S U R F A C E S C ~ E N C E
108
OCR for page 109
exchange or redox processes) is an example of the kind of new
approach that holds promise. The continued development of
reactive barriers and the development of other hybrid systems
(e.g., barrier systems that incorporate biological materials to
reduce maintenance requirements and enhance long-term per-
formance, or systems that use controlled water infiltration to
en hance waste decomposition or transformation) cou Id i mp rove
the technology for containment and in-situ stabilization of sub-
surface contaminants across the DOE complex.
Research Needs
The construction of stabilization and containment systems is proper-
ly within the province of applied technology development and will be
the responsibility of other DOE programs (e.g., the Subsurface Contami-
nants Focus Area). However, basic research focused on the following
tonics will be needed to support this technology development effort:
The mechanisms and kinetics of chemically and biologically
mediated reactions that can be applied to new stabilization and
containment approaches (e.g., reactions that can extend the use
of reactive barriers to a greater range of contaminant types found
at DOE sites) or that can be used to understand the long-term
reversibi I ity of chemical and biological stabi I ization methods.
2. The physical, chemical, and biological reactions that occur among
contaminants (metals, radionuclides, and organics), soils, and barri
er components so that more compatible and durable materials for
containment and stabilization systems can be developed.
3. The fluid transport behavior in conventional barrier systems, for
example, understanding water infiltration into layered systems,
including infiltration under partially saturated conditions and
under the influences of capillary, chemical, electrical, and ther-
mal gradients that can be used to support the design of more
effective infiltration barrier systems.
4. The development of methods for assessing the long-term durabil-
ity of containment and stabilization systems.
Monitoring and Validation
The abi I ity to man itor and val idate is essential to the successfu I
application of any corrective action to a subsurface contamination
problem, as is regulatory acceptance of that action. However, the
knowledge and technology bases to support these activities are not fully
developed and are receiving little attention in EM's science and tech-
nology programs. The monitor process step does not appear on the
C h a p t e r 5
109
OCR for page 110
Subsurface Contaminants Focus Area's remedial action flow chart (see
Figure 3.2), and its validate process step applies only to the confirma-
tion of the performance of a remedial action. As noted in Chapter 3,
very little research relevant to these activities is being supported cur-
rently by the EM Science Program.
As illustrated by Figure 5.1, monitoring and validation are important
at both the front and back ends of the site remediation process. At the
front end, monitoring and validation are used to support the develop-
ment of conceptual and predictive models of subsurface and contami-
nant behavior (Box 4 and Arrow 1~. At the back end, monitoring and
validation are used to gain regulatory acceptance for corrective actions
and to demonstrate the effectiveness of efforts to remove, treat, or espe-
cially to contain contamination (Box 7 and Arrow 2~. Such monitoring
and validation efforts can also improve the understanding of the conta-
minant fate and transport processes and can be used to recalibrate and
revise conceptual and predictive models important elements of the
model bu i Id i ng process.
Improvements in capabilities to monitor and validate could greatly
improve the technical success of DOE's efforts to contain and stabilize
contamination at its sites. The development of new containment and
stabi I Cation approaches cou Id lower the cost, accelerate regu latory
approvals for, and increase public confidence in efforts to address DOE
contamination problems. In the committee's judgment, the significant
knowledge gaps include the following:
.
.
Design of efficient and effective monitoring systems. There is I it-
tle experience with monitoring over the long (decadal to centen-
nial) time scales that are required at DOE sites. Consequently, a
great deal of basic knowledge is required to design efficient and
effective monitoring systems. The knowledge gaps include under-
standing what parameters need to be measured to assess system
performance (e.g., the performance of a subsurface barrier);
where, when, and how to obtain these measurements; and how
to relate these measurements to system behavior.
Unsaturated zone monitoring. Mon itori ng of the u nsatu rated
zone is a special case of the previous knowledge gap and is a
special need for DOE, because most of its containment and sta-
bilization systems are being constructed above the water table,
especially at the western U.S. sites. Unsaturated zone monitoring
is an especially difficult problem; the physics and chemistry of
unsaturated zone processes are more complicated than for the
saturated zone, and these processes have received far less atten-
tion from researchers. Contaminants may be present in both
~ . . . . .
S U B S U R F A C E S C ~ E N C E
OCR for page 111
liquid and gas phases in unsaturated zone environments and
under both aerobic and anaerobic conditions. The exchange,
degree of equilibration of these phases, and the transport of
these phases may occur by different processes with very different
rates. There is a great disparity between what can currently be
measured and what needs to be measured to predict the behav-
ior of contam i n ants i n many u nsatu rated setti ngs.
· Mode/ validation. A conceptual model is an estimate of the real-
world behavior and must be tested to ensure that it appropriately
represents the behaviors of interest. This testing is usually carried
out by comparing predictions made with the model against field
and experimental observations. This testing also allows the
model to be improved as new information on the subsurface sys-
tem is collected. The science of model testing, or validation, has
received relatively little attention until recently and is an area
where significant work is needed. The knowledge gaps include
understanding what measurements need to be collected to vali-
date a model (it is frequently the case that what can be calculat-
ed in a model cannot be measured in the field, and vice versa);
how to evaluate the relationships between measured and pre-
dicted behaviors; and understanding what diagnostic information
these differences provide for assessing and improving the accura-
cy of the models (e.g., see Sidebar 5.2~.
· Performance validation. Performance validation is a necessary
step to document the success, or lack thereof, with every step
shown in Figure 5.1 . The issues here are simi lar to those for
model validation, that is, how to assess whether the process is
performing as designed. The knowledge gaps include under-
standing what to measure, how to measure it, how to assess dis-
crepancies between designed and measured behavior, and deter-
mining what diagnostic information these differences provide for
assessing and improving performance.
For example, with regard to locating and characterizing con-
taminants, one must determine when enough information for risk
characterization and remedy selection has been gathered. Th is
requires tools to validate the assessments of contaminant
amounts, distributions, and mass release rates. Similar considera-
tions arise for validation of predictive models in the face of vari-
ability and uncertainty. The difficulty increases when probability
models are introduced to try to deal with uncertainty. With
regard to corrective action performance, validation is an essen-
tial step that is lacking for many innovative technologies, and has
prevented their selection for site remediation because of regula-
C h a p t e r 5
111
OCR for page 112
tory and stakeholder concerns. Knowledge gaps in performance
validation include understanding how to develop monitoring sys-
tems and sampling strategies, understanding the critical system
variables that need to be used, strategies for data collection in
highly heterogeneous systems, and the development of statistical
methods to be used in performance evaluation.
Research Needs
Many of the research needs for monitoring and validation have been
covered in previous sections; for example, research on locating conta-
minants and characterizing contaminant and subsurface properties and
research on data integration will provide new knowledge and capabili
ties for monitoring and validation. In addition, the committee believes
that basic research is needed on the following topics:
.
.
.
Development of methods for designing monitoring systems to
detect both the current conditions and changes in system behav-
iors. These methods may involve the application of conceptual,
mathematical, and statistical models to determine the types and
locations of observation systems and also will involve predicting
the spatial and temporal resolutions at which observations need
to be made. For example, such methods may help to determine
what types of measurements (e.g., core samples from a borehole
versus seismic images of the subsurface) can be used to validate
the model and also suggest where such measurements should be
made in both time and space.
Development of validation processes. The research questions
include (1 ) understanding what a representation of system
behavior means and how to judge when a model provides an
accurate representation of a system behavior the model may
give the right answers for the wrong reasons and thus may not be
a good predictive tool; and (2) how to validate the future perfor-
mance of the model or system behavior based on present-day
measurements. These questions might be addressed through
research projects that focus on the development of validation
methodologies using real-world examples at DOE sites.
Data for model validation. Determining the key measurements that
are required to validate models and system behaviors, the spatial
and temporal resolutions at which such measurements must be
obtained, and the extent to which surrogate data (e.g., data from
lab-scale testing facilities) can be used in validation efforts.
· Research to support the development of methods to monitor fluid
and gaseous fluxes through the unsaturated zone, and for differ-
S U B S U R F A C E S C ~ E N C E
112
OCR for page 113
entiating diurnal and seasonal changes from longer-term secular
changes. These methods may involve both direct (e.g., in situ sen-
sors) and indirect (e.g., using plants and animals) measurements
over long time periods, particularly for harsh chemical environ-
ments characteristic of some DOE sites. This research should sup-
port the development of both the physical instrumentation and
measurement techniques. The latter includes measurement strate-
gies and data analysis (including statistical) approaches.
·eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee
SIDEBAR5.2 MANAGING UNCERTAINTY
Management of uncertainty in model and performance validation is a theme that cuts across many of
the knowledge gaps identified in this chapter. Uncertainties arise in multiple ways. In field data they
can emerge in quality features: random measurement error, systematic errors from imperfectly cali-
brated instruments, and recording and other transmission errors. In mathematical models, uncertain-
ties stem from incorrect specifications and through propagation of errors in the data that are input to
the computational models. In the integration (or combination) of models and data, uncertainties are
affected by the need to link data and models that are on mismatched scales; some data may have to be
aggregated, other data may need to be disaggregated.
Quantifying the uncertainties in, for example, a site characterization problem can involve all the paths
described above.There may be several data sets of varying quality; missing data (measurements on
some contaminants may be found at some monitoring wells but missing at others); auxiliary data sets
(e.g., river flow data) on time scales very different from the frequency of sampled monitoring data; his-
torical records of differing content and quality; and transport models requiring uncertain input para-
meters. How best to combine the variety of information and assess the accuracy of results and predic-
tions is a great challenge.
Similar issues are found in validation and performance assessment.These may be compounded by the
need to perform detailed computer experiments to determine the impacts of uncertainties in data
quality and input specifications. For perfomance assessment and validation, attention has to be given
to design of future data collection: Where and when should collection be done to assure a desired level
of accuracy?
The daunting technical problem is how to respond to complex, though simple sounding, queries (e.g.,
Where is the contaminant plume now? Where will it be next year?) that demand intricate combinations
of computer and statistical models fed by several data sources. Powerful methods such as Bayesian
hierarchical modeling are emerging to break such complicated problems into components and,
through intensive computation, capture the uncertainties; but implementation is limited by the com-
plexity and scale of the problems typically encountered in subsurface contamination.
C h a p t e r 5
113
OCR for page 114
C ~scuss~on
.
As noted in the introduction to this chapter, the committee has been
selective in the identification of subsurface contamination knowledge
gaps and research needs for the EM Science Program. Indeed, the list of
knowledge gaps presented in this chapter is not exhaustive and is per-
haps notable for what it does not include, namely, the knowledge gaps
associated with assessment of risk (Box 5 in Figure 5.1~° and many of
the corrective actions associated with EM's cleanup program (Boxes 6C
through BE in Figure 5.1~. The committee has been selective because
(1 ) it believes that much of the subsurface at DOE sites cannot be reme-
diated cost effectively; (2) the contamination is highly distributed in
very large volumes of the subsurface; and (3) the EM Science Program
does not have the management or financial capital to support a com-
prehensive research program to address all of EM's cleanup problems.
Further, the committee recognizes that there is much good research on
these excluded topics being supported by other programs (see Chapter 4~.
The committee has selected the four research areas highlighted in
this chapter because, as illustrated by the examples in Chapter 2, these
themes cut across all DOE cleanup efforts, and the committee believes
that they are key to the long-term success of the DOE's cleanup pro-
gram. Further, the committee believes that a focused, sustained, and
adeauatelv funded research Program directed at the knowledge Baas
~~ ~ - ~ - r -o ~ - ----- lo-- - - -- --a- o~r-
could result both in significant improvements to L9OE cleanup capabili-
ties and the effectiveness of its cleanup actions.
The committee discussed whether it should prioritize these four
research areas, but decided against doing so. The selection of these four
research foci from among a much broader range of potential research
areas is in itself a significant prioritization. Further, the committee
believes that all four research foci are equally important for DOE's
need to be pursued aggressively if DOE is to
improve its capabilities to address its subsurface contamination prob-
lems. The new location, characterization, modeling, and monitoring
capabi I ities that can resu It from th is research, when appl fed appropri-
ately, will enable DOE to stay on a track that leads to success in its
efforts to clean up or contain its widespread subsurface contamination.
cleanup mission and wil I
4°Of course, the committee recognizes that the basic research needs outlined
in this chapter will produce new knowledge on contaminant locations and behavior
and thereby affect critical steps in the risk assessment.
S U B S U R F A C E S C I E N C E
Representative terms from entire chapter:
doe sites
