A Framework for Net Environmental
Benefit Analysis for Remediation or
Restoration of Petroleum-Contaminated
R. A. Efroymson
J. P. Nicolette
G. W. Suter II
Environmental Sciences Division
Reports produced after January 1, 1996, are generally available free via the U.S. Department
of Energy (DOE) Information Bridge.
Web site http://www.osti.gov/bridge
Reports produced before January 1, 1996, may be purchased by members of the public from
the following source.
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone 703-605-6000 (1-800-553-6847)
Web site http://www.ntis.gov/support/ordernowabout.htm
Reports are available to DOE employees, DOE contractors, Energy Technology Data
Exchange (ETDE) representatives, and International Nuclear Information System (INIS)
representatives from the following source.
Office of Scientific and Technical Information
P.O. Box 62
Oak Ridge, TN 37831
Web site http://www.osti.gov/contact.html
This report was prepared as an account of work sponsored
by an agency of the United States Government. Neither the
United States Government nor any agency thereof, nor any of
their employees, makes any warranty, express or implied, or
assumes any legal liability or responsibility for the accuracy,
completeness, or usefulness of any information, apparatus,
product, or process disclosed, or represents that its use
would not infringe privately owned rights. Reference herein to
any specific commercial product, process, or service by trade
name, trademark, manufacturer, or otherwise, does not
necessarily constitute or imply its endorsement,
recommendation, or favoring by the United States
Government or any agency thereof. The views and opinions
of authors expressed herein do not necessarily state or
reflect those of the United States Government or any agency
Environmental Sciences Division
A FRAMEWORK FOR NET ENVIRONMENTAL BENEFIT ANALYSIS FOR
REMEDIATION OR RESTORATION OF PETROLEUM-CONTAMINATED SITES
Rebecca A. Efroymson
Environmental Sciences Division
Oak Ridge National Laboratory
Joseph P. Nicolette
Glenn W. Suter II
National Center for Environmental Assessment
U.S. Environmental Protection Agency
U.S. Department of Energy
National Petroleum Technology Office
Budget Activity Number AC 10 15 00 0
OAK RIDGE NATIONAL LABORATORY
Oak Ridge, Tennessee 37831
U.S. DEPARTMENT OF ENERGY
under contract DE-AC05-00OR22725
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
1. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2. PRECEDENTS FOR NEBA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
3. ALTERNATIVE ACTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.1 NATURAL ATTENUATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.2 TRADITIONAL REMEDIATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.3 ECOLOGICAL RESTORATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
4. STRUCTURE OF NEBA FRAMEWORK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.1 PLANNING PHASE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.1.1 Management Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.1.2 Stressors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.1.3 Ecological Services and Other Ecological Properties . . . . . . . . . . . . . . . . 14
4.1.4 Comparative Metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.1.5 Temporal Measures of Exposure and Effects . . . . . . . . . . . . . . . . . . . . . . 18
4.1.6 Spatial Measures of Exposure and Effects . . . . . . . . . . . . . . . . . . . . . . . . 18
4.1.7 Conceptual Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4.1.8 Analysis Plan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.2 CHARACTERIZATION OF REFERENCE STATE . . . . . . . . . . . . . . . . . . . . . . . . 19
4.3 NEBA OF SINGLE ALTERNATIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.3.1 Time-integrated Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.3.2 Characterization of Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
220.127.116.11 Biodegradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
18.104.22.168 Bioavailability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
22.214.171.124 Chemical removal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.3.3 Exposure-response Relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.3.4 Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.4 COMPARISON OF MULTIPLE ALTERNATIVES . . . . . . . . . . . . . . . . . . . . . . . . 28
4.4.1 Challenges of Comparative Assessments . . . . . . . . . . . . . . . . . . . . . . . . . 29
5. ADDITIONAL CONSIDERATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
6. CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
7. REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
LIST OF FIGURES
Fig. 1. Framework for Net Environmental Benefit Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Fig. 2. Characterization of the contaminated reference state or natural attenuation . . . . . . . . . . 10
Fig. 3. Net Environmental Benefit Analysis of remedial alternatives . . . . . . . . . . . . . . . . . . . . . 11
Fig. 4. Net Environmental Benefit Analysis of ecological restoration . . . . . . . . . . . . . . . . . . . . 12
Fig. 5. Trajectory of assessment endpoint entity (service or other ecological property) with
time, following a petroleum spill; conditions that would have been expected to
prevail in the absence of the spill; expected trajectory of the remediated state;
expected trajectory of the restored state. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Research sponsored by the Fossil Energy Office, U.S. Department of Energy, under
contract DE-AC05-00OR22725 with UT-Battelle, LLC. The views in the paper do not
necessarily reflect the policies of the U. S. Department of Energy or the U.S. Environmental
Protection Agency. We thank F. Dexter Sutterfield, Nancy Comstock, Kathy Stirling, and Daniel
Gurney, project managers at the DOE National Petroleum Technology Office in Tulsa,
Oklahoma, for their support of this project. We thank Jim Loar of Oak Ridge National
Laboratory, Randy Bruins of the U.S. Environmental Protection Agency, and three petroleum
industry reviewers for comments on an earlier draft of this manuscript.
Efroymson, R. A., J. P. Nicolette, and G. W. Suter II. 2003. A framework for net environmental
benefit analysis for remediation or restoration of petroleum-contaminated sites. ORNL/TM-
2003/17. Oak Ridge National Laboratory, Oak Ridge, Tennessee.
Net environmental benefits are the gains in environmental services or other ecological
properties attained by remediation or ecological restoration, minus the environmental injuries
caused by those actions. A net environmental benefit analysis (NEBA) is a methodology for
comparing and ranking the net environmental benefit associated with multiple management
alternatives. A NEBA for chemically contaminated sites typically involves the comparison of the
following management alternatives: (1) leaving contamination in place; (2) physically,
chemically, or biologically remediating the site through traditional means; (3) improving
ecological value through onsite and offsite restoration alternatives that do not directly focus on
removal of chemical contamination or (4) a combination of those alternatives. NEBA involves
activities that are common to remedial alternatives analysis for state regulations and the
Comprehensive Environmental Response and Liability Act, response actions under the Oil
Pollution Act; compensatory restoration actions under Natural Resource Damage Assessment,
and proactive land management actions that do not occur in response to regulations: i.e., valuing
ecological services or other ecological properties, assessing adverse impacts, and evaluating
restoration options. This paper provides a framework for NEBA, with special application to
petroleum spills in terrestrial and wetland environments. A high-level framework for NEBA is
presented, with subframeworks for natural attenuation (the contaminated reference state),
remediation, and ecological restoration alternatives. Primary information gaps related to NEBA
include: non-monetary valuation methods, exposure-response models for all stressors, the
temporal dynamics of ecological recovery, and optimal strategies for ecological restoration.
1Restoration, as defined here, refers to actions that directly improve ecological services or other
ecological properties, onsite or offsite (the term “mitigation” is sometimes used), in contrast to
remediation, which focuses on chemical removal. Ecological restoration encompasses
restoration, rehabilitation, replacement or acquisition of the equivalent, as defined by the Oil
Pollution Act of 1990 (NOAA 1997).
Net Environmental Benefit Analysis (NEBA) is a methodology for identifying and
comparing net environmental benefits of alternative management options, usually applied to
contaminated sites. Net environmental benefits are the gains in environmental services or other
ecological properties attained by remediation or ecological restoration1, minus the environmental
injuries caused by those actions. A NEBA for chemically contaminated sites typically involves
the comparison of the following management alternatives: (1) leaving contamination in place; (2)
removing the contaminants through traditional remediation; (3) improving ecological value
through onsite or offsite restoration that does not involve removing contaminants; or (4) a
combination of those alternatives. Examples of combinations include remediation of localized
soil contamination combined with natural attenuation and the planting of trees, and the dredging
of sediment hotspots combined with local wetland restoration. NEBA involves valuing
ecological services or other properties, assessing adverse impacts, and evaluating restoration
options. These activities are common to remedial alternatives analysis for state regulations and
the Comprehensive Environmental Response and Liability Act (CERCLA); response actions
under the Oil Pollution Act (OPA); compensatory restoration actions under Natural Resource
Damage Assessment (NRDA), and proactive land management actions that do not occur in
response to regulations. NEBA often incorporates the comparative methodology of habitat
equivalency analysis (HEA), as described below. This methodology is in common use for
ecological restoration alternatives related to petroleum spills. However, NEBA has not been
formalized in a manner analogous to the U. S. Environmental Protection Agency (EPA)
ecological risk assessment framework (EPA 1998), and land managers would benefit from such a
framework for NEBA.
NEBA may be thought of as an elaboration of ecological risk assessment. That is, it is
risk-benefit analysis applied to environmental management actions. Hence, the EPA ecological
risk assessment framework (EPA 1998) could be adapted to perform NEBAs. However, because
risk assessment does not normally consider benefits, and risk assessors are not familiar with the
requirements of an assessment that estimates benefits, a new framework is useful to accentuate
the specific features of such analyses. In addition, NEBAs are usually performed by resource
management agencies that are not familiar with the ecological risk assessment formalism.
NEBA has the potential to help land managers avoid the possibility that the selected
remedial or ecological restoration alternative will provide no net environmental benefit over
natural attenuation of contaminants and ecological recovery. An alternative may provide no net
environmental benefit because: (1) the remedial or ecological restoration action is ineffective (the
action does not substantially change the risk) or (2) the remediation alternative causes
environmental injuries greater than the damage associated with the contamination because (a) the
need for remediation has been driven by human health risk, not ecological risk; (b) the ecological
injury from contamination has been overestimated because of conservative assumptions; or (c)
injuries associated with remediation were not properly addressed. Pitfall 2c is emphasized in this
discussion. Similarly, NEBA has the potential to help land managers plan an ecological
restoration alternative that provides a positive net environmental benefit over the hypothetical
state that would prevail in the absence of contamination. NEBA is recommended if any of the
remedial or restoration alternatives potentially has significant negative ecological effects or
minimal ecological benefits. Finally, NEBA is needed when the multiple alternatives are
beneficial, but the one with the greatest net benefits is not apparent without formal analysis.
This paper provides a framework for NEBA; demonstrates how residual injuries and
benefits from natural attenuation, traditional remediation, and ecological restoration options may
be compared systematically; and identifies key research needs. Principal aspects of the
framework include: (1) a single planning phase for analysis of all alternatives, (2) the
identification of a comprehensive set of ecological services or other focal ecological properties,
(3) the modular layout of certain components of the framework (e.g., chemical exposure and
effects analysis), (4) the development of temporally variable estimates of exposure (e.g., due to
biodegradation), (5) the development of credible, non-conservative exposure-response
relationships beyond simple toxicity thresholds or habitat area thresholds, (6) the development
and integration of temporal estimates of effects (e.g., due to recovery), and (7) the consideration
of habitat equivalency and other potential valuation metrics for comparing ecological states. The
emphasis of this paper and examples herein is on petroleum contamination in terrestrial and
wetland ecosystems, although the framework is equally applicable to aquatic environments.
Environmental management alternatives must also be considered in an economic context,
but cost issues (such as relative costs of alternatives, monetary value of ecological resources,
costs of monitoring, and NRDA liability costs) are not currently included in the framework; this
framework addresses net environmental benefits rather than net economic benefits. Similarly,
human health risks are typically external to NEBA, but would contribute significantly to most
management decisions about chemical contamination. If substantial human health risks are
present, the relative net environmental benefit of alternatives would hold less weight in the
2. PRECEDENTS FOR NEBA
methodological guidance for the assessment of contaminated sites. These range from federal and
state government examples to industry examples. The term NEBA is not commonly used in
CERCLA remedial feasibility analysis or NRDA contexts. It was probably coined by agencies
and industries evaluating options for marine oil spills, as in the report published by the U.S
National Oceanic and Atmospheric Administration (NOAA) in 1990 entitled Excavation and rock
washing treatment technology: Net environmental benefit analysis. In that study, representatives
of Exxon, NOAA, and the State of Alaska evaluated a remedial option for the Alaskan shoreline
affected by the Exxon Valdez oil spill “to determine if there were net environmental benefits
from the excavation and washing of oiled sediments [below 15 cm depth, with heated seawater],
and return of treated sediments to the excavated site over natural cleansing and the use of
approved 1990 treatments,” i.e., manual removal, spot washing, and bioremediation (NOAA
1990). Although the study provided an analysis of the potential adverse impacts associated with
the proposed remediation technology, and estimated relative recovery periods, it did not provide a
framework or propose metrics for comparison of injuries and benefits from the alternative
methods. The term NEBA is commonly associated with assessments of oil spill dispersants in
marine environments (Baker 2001, Fiocco and Lewis 1999, Lunel et al. 1997).
Several precedents for NEBA exist, but they provide little specific, procedural or
In another example of NEBA, the net environmental benefit of dredging part of an
estuary was investigated (J. P. Nicolette, CH2M Hill, personal communication, October 11, 2001;
Rubin et al. 2001). The approach was to adopt an estuary-wide sediment services strategy. Many
ecological services from the contaminated sediments had been lost due to biochemical reduction
within the anaerobic environment, which caused toxic levels of ammonia. However,
sedimentation was shown to be occurring at rates that were expected to reduce the bioavailability
of contaminants. Although natural attenuation was viewed as an attractive option because of its
cost and efficacy, the regulatory agencies were concerned about potential injuries that could
occur during the natural attenuation process. Therefore, a restoration action was proposed to
deliver sediment services with certainty to offset the potentially lost sediment services.
Additional applications of NEBA are listed in Table 1, though most applications of HEA, a subset
of NEBA, are not publicly available because of their use in litigation proceedings (Milon and
Dodge 2001), and thus many more NEBAs (especially terrestrial NEBAs) have been performed
than those of which we are aware.
Table 1. Examples of NEBA
Net environmental benefit (NEB) of excavation and rock
washing treatment technology versus natural attenuation
and approved treatments, Exxon Valdez oil spill
Quantification of wetland mitigation from petroleum pipeline
Nicolette et al. 2001
NEB of natural attenuation versus pump and treat technology
versus air sparge/vapor extraction of volatile organic
compounds in groundwater
J. Nicolette, CH2M Hill,
NEB of dredging versus not dredging an estuary, with
quantification of restoration needed to offset uncertainty in
Rubin et al. 2001
NEB of seagrass and mangrove restoration, following
undisclosed disturbance at John’s Island, Palm Beach
S. Friant, Entrix, personal
communication, May 23,
NEB of the use of dispersant following the grounding of the
Sea Empress in Great Britain
Lunel et al. 1997
Quantification of compensatory restoration of salt marsh
vegetation on dredge material placed on a barrier island,
given impacts from oil spill in Lake Barre, Louisiana
Penn and Tomasi 2002
Quantification of replacement habitat to compensate for coral
reef injuries from vessel groundings
Milon and Dodge 2001
Although the NEBA terminology is not normally used in CERCLA remediation
assessments, the concept is included in the guidance from the EPA Office of Emergency and
Remedial Response (Luftig 1999). “Even though an ecological risk assessment may demonstrate
that adverse ecological effects have occurred or are expected to occur, it may not be in the best
interest of the overall environment to actively remediate the site. At some sites, particularly those
that have rare or very sensitive habitats, removal or in-situ treatment of the contamination may
cause more harm (often due to wide spread physical destruction of habitat) than leaving it in
place. . . . The likelihood of the response alternatives to achieve success and the time frame for a
biological community to fully recover should be considered in remedy selection. Although most
receptors and habitats can recover from physical disturbances, risk managers should carefully
weigh both the short- and long-term ecological effects of active remediation alternatives and
passive alternatives when selecting a final response.” Similarly, the Great Lakes Water Quality
Board recommends that “prior to embarking on sediment remediation, [one should] have
developed some quantifiable expectation of result (ecological benefit) and a program to follow
the predicted recovery” (Zarull et al.1999).
Individual scientists have espoused NEBA-like concepts and methods (Principe 1995;
Baker 1999). P. P. Principe of the EPA National Exposure Research Laboratory has used the
term “Ecological Benefits Assessment” to refer to a procedure that could be used to assess
changes in resource service flows that would result from different management or control
alternatives at large spatial scales (Principe 1995). Principe (1995) emphasizes the importance of
benefits assessment, and describes a general taxonomy of benefits, but does not provide
Baker (1999) advocates the use of NEBA for evaluating oil spill clean-up alternatives.
She describes elements of a process for NEBA, including (1) collection of environmental data,
characterization of environmental services, and description of the remediation method; 2) review
of spill case studies that are relevant to the proposed remedial method; (3) prediction of likely
environmental outcomes; (4) comparison of advantages and disadvantages of remediation and
natural attenuation; and (5) balancing of advantages and disadvantages to proposed alternatives to
make a decision. These elements of NEBA have not been formalized in a framework.
At least three states endorse NEBA-related concepts or methodologies. That is, these
states allow or advocate the comparison of environmental benefits in environmental management
legislation. In addition, New Jersey, Massachusetts, Louisiana, Arkansas, Connecticut, Alaska,
Indiana, California, Pennsylvania, and Delaware have supported NEBA-type strategies for
evaluating remedial alternatives.
(1) The Texas Commission on Environmental Quality (TCEQ) (formerly the Texas
Natural Resource Conservation Commission, TNRCC) recommends “Ecological Services
Analysis” as an option for contaminated sites where chemical concentrations exceed ecologically
protective concentration levels (PCLs) but not human health PCLs (TNRCC 2001). The
potentially responsible party may propose compensatory ecological restoration after quantifying
benefits and risks associated with alternative remedial actions or natural attenuation. HEA,
described below, is one recommended comparative methodology, and others may be proposed to
natural resource trustees. In addition, ecological services analysis and compensatory ecological
restoration require approvals of the natural resource trustees for the state of Texas, obtained
through the TCEQ.
(2) The State of Florida Department of Environmental Protection (DEP) may enter into a
voluntary “ecosystem management agreement”with regulated entities and other government
entities if the DEP determines that “implementation of such agreement meets all applicable
standards and criteria so that there is a net ecosystem benefit to the subject ecosystem more
favorable than operation under applicable rules” and “implementation of the agreement will result
in a reduction in overall risks to human health and the environment compared to activities
conducted in the absence of the agreements” (State of Florida 2001). This “team-permitting”
approach to environmental management was proposed by the business community and supported
by the Florida DEP (Barnett 1999).
(3) Recent revisions to Washington State’s Model Toxics Control Act include provisions
for a “Disproportionate Cost Analysis” for the consideration of incremental benefits and costs in
the selection of a remedial alternative. The comparison of benefits and costs may be quantitative
or qualitative and need not be monetary. The analysis includes an evaluation of residual risks that
are associated with each alternative, such as whether or not remedies that are protective of human
health are also protective of ecological receptors (Washington State Department of Ecology
3. ALTERNATIVE ACTIONS
As stated above, alternative actions are divided into three principal categories: natural
attenuation, traditional remediation, direct ecological restoration, and combinations of these.
Comparisons are based on not only the state of contamination, but also ecological recovery.
3.1 NATURAL ATTENUATION
Natural attenuation is remediation through natural dilution and degradation processes,
without addition of electron acceptors, nutrients, or electron donors. This alternative is
equivalent to the “baseline” scenario for which risks are rather rigorously assessed in CERCLA
remedial investigations in order to determine whether remediation is needed (EPA 1989,
Sprenger and Charters 1997, Suter et al. 2000). Typically in CERCLA, the emphasis is on
current risks and one or two future time points, rather than a continuous temporal analysis. If
estimated health and ecological risks are sufficiently low, no remedial action is required and the
contaminants are naturally attenuated. If these risks are unacceptable, natural attenuation is
considered along with remedial alternatives that involve removal of contaminated media or
interventions to increase the rate of attenuation. Natural attenuation may be chosen as the best
alternative or part of the best alternative for meeting remedial goals if active remediation is
ineffective, cost-prohibitive or damaging to the environment. Swindoll et al. (2000) provide a list
of six situations where natural attenuation may be appropriate, including “there is no evidence of
an imminent threat to ecological resources” and “sufficient time is available for [natural
remediation].” Because active remediation may introduce new risks, natural attenuation may be a
viable option even if Swindoll’s two criteria are not met. One of EPA’s criteria for judging
natural attenuation is “to achieve site-specific remediation objectives within a time frame that is
reasonable compared to that offered by other more active methods” (EPA 1999). If these
remediation objectives relate to ecological properties, then net environmental benefit
determinations should reflect estimates of ecological recovery. Natural attenuation is
nonintrusive and has no incremental remedial hazards, only those associated with the original
contamination and its metabolites. Few data are available to compare the risk reduction provided
by natural attenuation to reductions from various remediation alternatives (Stahl and Swindoll
1999). Performance monitoring is important for this alternative (Heath 1999).
3.2 TRADITIONAL REMEDIATION
Excavation, incineration, burning, chemical remediation, microbial bioremediation, and
phytoremediation reduce risks by removing contamination or actively reducing chemical
concentrations in environmental media. Excavation is the most common option for remediating
contaminated soils if the scale of contamination does not make the cost prohibitive. A physically
harsh remedial alternative, such as soil excavation, would usually have greater, immediate
adverse impacts to ecological receptors than concentrations of petroleum hydrocarbons at many
spill sites, especially given that many semi-volatile hydrocarbons and their metabolites are not
highly toxic to plants. Facilitated bioremediation can range from simple aeration (tilling) of soil
to the addition of electron donors or microorganisms. Phytoremediation of petroleum
hydrocarbons enhances rates of degradation in rhizosphere soil (Susarla et al. 2002). Some
remedial alternatives, such as burning of spills in marshes and fields, are used only in emergency
management situations (API 1999). Potential hazards posed by remedial interventions are listed
in Table 2.
Rigorous assessments are not typically required to evaluate risks associated with remedial
alternatives, and few guidance documents emphasize the importance of comparing risks from
various remedial alternatives and no-action alternatives (Suter et al. 2000, Reagan 2000).
Remediation is assumed to reduce risk. Remedial goals are defined based on health or ecological
risks from the contaminants, but the remedial technologies are chosen based primarily on two
engineering criteria: the ability to achieve those goals and cost-effectiveness. This focus on
engineering criteria rather than environmental goals tends to restrict the range of options
NEBA might have facilitated more rigorous ecological comparisons of alternatives to
support past remedial actions. For example, at a Department of Energy facility in Oak Ridge,
Tennessee, thousands of healthy, but PCB-contaminated, fish and other aquatic organisms were
rotenoned to prevent a probable reproductive decrement to a few individual herons, ospreys and
kingfishers feeding at the pond currently and in the future. Although substantial resources were
devoted to assessing risk to the piscivorous birds, little effort was devoted to assessing the risk
from the removal action. Similarly, recent research suggests that dredging of sediments in a canal
of the San Francisco Bay may not have provided net environmental benefits, as measured by
DDT and metabolite body burdens, and capping, more rigorous dredging, or an unevaluated
ecological restoration alternative might have provided an environmental benefit (Weston et al.
3.3 ECOLOGICAL RESTORATION
Ecological restoration is the direct restoration of certain ecological entities (services or
other properties of populations, communities, or ecosystems) or their habitats (specific wetland,
grassland, forest, or stream bed types). In NRDAs, ecological restoration may be proposed by
potentially responsible parties to replace time-integrated, lost services or other ecological
properties, in lieu of monetary compensation. The restoration may occur on the affected land or
on other land, usually in the same ecosystem. In either case, restoration is “compensatory,”
damages are “offset,” and the net environmental benefit compared to the uncontaminated
reference state is zero or positive. Ecological restoration of chemically contaminated land is
Table 2. Examples of ecological hazards posed by terrestrial remedial actions
Possibly increased bioavailability or toxicity of hydrocarbons or products
Devegetation due to tilling
Decreased plant diversity and aqueous contamination due to fertilization
Isolation (capping) of
Destruction of vegetation
Destruction of habitat and outmigration by vertebrates in excavated area
Removal of nutrient-rich surface soil and associated microorganisms and
Failure of soil ecosystem and vegetation to recover if nonindigenous fill soil
Destruction of ecosystem at borrow pit where fill is obtained and at landfill
where excavated soil is deposited.
Alarm and escape behavior of wildlife due to construction activity and noise
Burning of Spills,
Soil Incineration or
Decrease in air quality and associated risk to wildlife or plants
Destruction of above-ground vegetation, below-ground seeds and root
material from severe heat
Destruction of soil organic matter and potential loss of productivity
Change in chemistry of oil residue which may prevent emergence of new
Secondary fires, extending area of habitat destruction
Outmigration by vertebrates in burned area
Destruction of vegetation and outmigration by vertebrates in areas where
roads, parking areas or laydown areas are developed, or foot traffic is
Reduction in biodiversity and wildlife forage from mowing of excavated
area, cap or landfarm to maintain lawn
Decrease in air quality associated with increased truck traffic
sometimes combined with a localized remedial action, such as hot spot removal, or with
monitored natural attenuation (TNRCC 2001). Offsite ecological restoration could provide a net
environmental benefit with lower costs than excavation or bioremediation of decades-old,
refinery-contaminated land. Ecological restoration with natural attenuation would usually be
expected to provide a net environmental benefit, compared to natural attenuation, because
restoration provides benefits beyond the reduction in chemical risk through time.
Replacement habitats have included seagrasses, coral reefs, tidal wetlands, salmon
streams, estuarine soft-bottom sediments, mangroves, mud flats, salt marshes, riparian forests,
dune and swale ecosystems, and grasslands (http://contaminants.fws.gov/Issues/Restoration.cfm).
For example, a 50-foot wide buffer zone of native trees and shrubs were planted on the eroded
banks of a tributary of the Potomac River near Reston, Virginia, where a pipeline released diesel
fuel overland and into the stream. The replacement habitat is sometimes located at a distance
from the degraded habitat, particularly if site selection criteria are narrow. For example, seagrass
restoration requires a specific substrate (Fonseca et al. 2000).
Phytoremediation of terrestrial oil spills may also restore services. Planting native plant
species would restore primary production and wildlife habitats as well as aid in the removal of
petroleum contamination, likely resulting in substantial net environmental benefit under NEBA.
An example of the use of phytoremediation for this dual purpose is the planting of the marsh
grass Spartina alterniflora, supplemented with fertilizer in a petroleum-contaminated wetland
(Lin and Mendelssohn 1998).
Research is needed to define optimal strategies for ecological restoration of particular
ecological services and other ecological properties. Habitat which appears to be successfully
restored may support few individuals of a critical species or may not support sufficient
reproduction to balance mortality, thereby becoming a sink habitat that drains individuals from
other areas. A restored tidal marsh failed to create habitat for the endangered light-footed clapper
rail (Hackney 2000). Restoration that involves physical construction may subsequently fail,
resulting in ecological injuries (e.g., stream channel restorations that wash away or artificial
wetlands that are dry). Indeed, ecological restoration technologies tend to be evaluated on the
basis of engineering criteria, such as the ability to establish soil cover or to stabilize stream banks,
rather than ecological criteria. Estimates of restoration endpoints may be uncertain due to
temporal and spatial variability in precipitation and other environmental factors, natural variation
in growth of vegetation and animals, errors in site preparation and in use of transplant material,
predation on transplanted organisms, and human land use changes (Thom 2000, NRC 1992). In
addition, droughts or floods may affect the success of restoration. NRDA consent decrees
typically specify performance criteria and monitoring schedules (Penn and Tomasi 2002).
4. STRUCTURE OF NEBA FRAMEWORK
The high-level framework for NEBA is depicted in Fig. 1, and includes a planning
phase, characterization of reference state, NEBA of alternatives (including characterizations of
exposure of effects, including recovery), comparison of NEBA results, and possible
characterization of additional alternatives. Only ecological aspects of alternatives are included in
this framework. The figure also depicts the incorporation of cost considerations, the decision,
and monitoring and efficacy assessment of the preferred alternative, although these processes are
external to NEBA. Three subframeworks are presented: (1) characterization of the contaminated
reference state (Fig. 2), (2) NEBA for a remediation alternative (Fig. 3), and (3) NEBA for an
Characterization of reference state(s) (services or other ecological properties)
ecological characterization of
ecological characterization of contaminated
state (natural attenuation alternative)
Net Environmental Benefit Analysis of Management Alternatives
Management and assessment goals
Spatial and temporal scope of assessment
Ecological services and other properties
Comparative metrics (e.g., habitat equivalency)
Measures of exposure and effects
NEBA of ecological
NEBA of combined remediation
or restoration or natural
Comparison (ranking) of NEBA results,
relative to each reference state
Integration of NEBA results to
produce improved management
Division of net
benefit by cost
Fig. 1. Framework for Net Environmental Benefit Analysis. Dashed lines indicate
optional processes; circles indicates processes outside of NEBA framework.
current or past
exposure to chemical
future exposure to
Contaminant Risk Module
ecosystem disturbance by
habitat removal and
(services or other
Fig. 2. Characterization of the contaminated reference state or natural attenuation.
The net environmental benefit of natural attenuation, where the reference state is the trajectory of
ecological entities (services and other properties) under contaminated conditions, is zero.
current or past
plan for chemical
plan for physical
future ecological effects
(services or properties)
due to biological agent
properties) due to
benefit of remediation,
compared to reference
habitat removal and
ecological effects from
chemical, physical and
Fig. 3. Net Environmental Benefit Analysis of remedial alternatives. Dashed-borders
on boxes indicate that the contaminant risk module in Fig. 2 should be inserted here.
current or past
direct restoration action
(e.g, addition of riparian
Characterization of ecological endpoint
entities (services or properties)
to reference state
Fig. 4. Net Environmental Benefit Analysis of ecological restoration.
ecological restoration alternative (Fig. 4). Figure 2 also constitutes the steps of analysis for
NEBA of natural attenuation, if natural attenuation is compared to the uncontaminated reference
state. (The net environmental benefit of natural attenuation, compared to the contaminated
reference state, is zero, by definition.) If an alternative involves multiple actions (e.g., addition of
plants and chelation agents for phytoremediation or removal of hot spot contamination and
grassland restoration), the assessor can draw on Suter (1999) for recommendations concerning
how to estimate combined effects.
This detailed framework for NEBA does not preclude the use of more informal, NEBA-
like approaches in regulatory negotiations. As in many ecological risk assessments, the funds
available for for a NEBA may not allow the level of data collection that we recommend for
estimating past, present and future ecological states with confidence.
4.1 PLANNING PHASE
The planning phase for a NEBA includes: setting the goals of assessment; selecting a
limited and feasible suite of alternative actions (Sect. 3); defining the temporal and spatial scope
of assessment; identifying contaminant and remediation stressors; selecting environmental
services and other ecological properties of interest; selecting metrics and methodologies for the
comparison of alternatives; selecting a reference state (Sect. 4.2); establishing a link between
stressors and services (conceptual model); and developing an analysis plan (Fig. 1). The planning
phase is comparable to the planning and problem formulation phases in a risk assessment (EPA
1998). A comparative assessment such as a NEBA should have a plan that encompasses all
relevant, alternative actions. If NEBA is performed after the CERCLA or other baseline
ecological risk assessment has been completed, the risk assessment and related data collection
may need to be modified to suit the comparative purpose.
4.1.1 Management Goals
A common management goal for a NEBA may be to quantify net environmental benefits
of remediation and restoration alternatives to support a cost-benefit analysis of those alternatives.
Or, if a particular type of restoration is preferred by land managers and/or natural resource
trustees, the management goal may be to restore land to the extent that there is a positive
environmental benefit, compared to the uncontaminated reference state. For example, natural
resources trustees selected marsh construction on dredge spoil as the preferred type of restoration
at a pipeline spill site, and formal analysis focused on determining the amount of restoration
needed to achieve the desired net benefit (Penn and Tomasi 2002). Rules and regulations,
scoping assessments, or ad hoc decisions by regulatory agencies may define: (1) the ecological
services or other properties of concern; (2) the relative importance of past, present and future
injuries; (3) the reference state for the analysis (contaminated or uncontaminated); (4) acceptable
or recommended analytical methodologies; (5) preferred comparative methods (HEA is
recommended by TCEQ); or (6) preferred actions. For example, past damage is important in the
NRDA context, because NRDA aims to compensate for ecological services lost in the past,
present and future; but CERCLA remedial responses and related state regulations only draw on
present and future conditions. One management goal may be to streamline the relationship
between (1) risk assessment and response guidance of state regulations, CERCLA, and the Oil
Pollution Act and (2) resource liability estimates associated with NRDA provisions of the federal
acts. Emergency response may necessitate a decision before a formal NEBA can be undertaken.
The stressor that is common to the determination of net environmental benefits of all
alternatives is chemical contamination. Traditional remediation may impose the widest range of
potential stressors, including the physical stressor of excavation or tilling, the biological stressor
of introduced microorganisms or plants, the residual chemical stressor or added chelation agents,
nitrate, or peroxide (Table 2). Nonchemical stressors are seldom considered under restoration
scenarios, but it is possible that (1) vehicle movement, grading, tilling, or trampling could
constitute stressors in the process of restoring an ecosystem, (2) the restoration may fail and result
in physical damage, or (3) the restoration of habitat for one population could decrease habitat for
another. For example, because killdeer prefer gravelly surfaces for laying their eggs, restoring
soil and vegetation to these areas could reduce local populations. Similarly, the marsh restored
on dredge spoil in Penn and Tomasi (2002) would have been of greater value to shorebirds if it
had remained unvegetated. Moreover, the ecological service of protection from predation is in
direct conflict with the service of provision of food to predators.
NEBA also considers benefits that result from a decrease in a chemical stressor or a direct
ecological restoration effort. Although restoration is comprised of physical and biological
components that could be termed “beneficial agents,” we choose not to use that term because
restoration feeds into the NEBA at the characterization of effects stage of analysis (Sect 4.3.3)
and does not need to be separated into its component actions to determine exposure.
4.1.3 Ecological Services and Other Ecological Properties
NEBAs usually evaluate ecological services that are provided by an area of land or
wetland. Services have been emphasized because (1) they are more easily valued than other
ecological properties, prior to a cost-benefit analysis of alternatives; (2) services are the subject of
the TCEQ’s Ecological Services Analysis option (TNRCC 2001), and much petroleum activity is
located in Texas; (3) ecological services are often the subject of NRDAs; and (4) HEA (discussed
below) is a convenient methodology for comparing multiple services or multiple alternatives on a
The selection of services rather than population or community properties as focal entities
of NEBA might appear to be inconsistent with CERCLA ecological risk assessments. Risk
assessments associated with remedial investigations tend to emphasize multiple endpoint
properties of organisms (e.g., mortality or fecundity) or populations (e.g., abundance or
production) representing different trophic groups while NRDAs and NEBAs typically emphasize
ecological services and ecosystem value. However, services estimated or measured in NEBAs
are sometimes estimated or measured quantities in an ecological risk assessment (e.g., production
of a plant community, abundance of a food item or area suitable for mating, nesting). In addition,
the NEBA practitioner can choose other ecological properties as endpoints if they are consistent
with the management goals of the assessment (e.g., regulations list injuries to survival, growth,
reproduction, behavior, community composition, and community processes and functions as key
components to NRDA (Department of Commerce 1996)). Barnthouse et al. (1995) note that
resources and CERCLA assessment endpoints are “functionally equivalent,” but the entities or
properties may be different because trustees and CERCLA participants (DOE, EPA, state)
emphasize different goals.
Environmental services or other ecological properties should be selected with all
alternatives in mind. That is, if clapper rail habitat is injured or benefits from one alternative
2The term “habitat” in HEA refers to an ecosystem, rather than to species-specific habitat. In this
paper, we attempt to use the term “ecosystem” to refer to land areas with ecological value based
on their structure and functions and “habitat” to refer to species-specific habitat.
action, the state of clapper rail habitat under other alternatives, during the time and within the
spatial extent of the analysis, should be obvious or evaluated.
NEBA should measure many ecological services or other ecological properties or
demonstrate that the analysis of one is sufficient to represent others. The representation of all
populations of a particular trophic level by a single species is often acceptable in CERCLA
chemical risk assessments if species are not known to have differential sensitivity. Animals of
similar taxa and feeding habits (e.g., insectivorous birds) are often assumed to have similar
exposures and sensitivity to chemicals. However, if ecosystem area is lost during excavation or
gained during restoration, the use of representative species would require that species home
ranges and habitat requirements would be similar. More commonly, NEBAs use a single
restoration metric to represent all services.
4.1.4 Comparative Metrics
Few comparative methodologies and metrics exist. The most common methodology for
comparing ecological restoration alternatives at petroleum and other contaminated terrestrial and
marsh sites is HEA. HEA “is a habitat2-based approach that determines compensation in terms of
the amount of comparable habitat required to replace lost ecological services; [therefore], natural
resource injuries must be determined at the habitat level” (DOI et al. 1999). A typical metric for
comparing injured and replacement ecological services and other properties under HEA is the
total service integrated over area and time, or service-hectare-year. The metric is often converted
to present-day value.
In the simplest form of HEA, the analyst could assume that on a per area basis for a given
plot of land, all ecological services are proportional to each other. That is, if grassland primary
production is restored, then litter decomposition and the provision of nesting or lekking sites for
all bird species will be restored, and the single metric of primary production is sufficient for the
NEBA. In a more complex implementation of HEA, injuries that are not associated with
ecosystem-level disturbance (e.g., direct mortality of birds from contact with oil spill, either
measured or estimated) may be converted to habitat service metrics. Penn and Tomasi (2002)
converted individual bird losses to the habitat area that would have produced the biomass, based
on salt marsh production and inefficient energy exchange among trophic levels (but without
explicitly considering potential nesting sites or habitat connectivity). Although habitat metrics
simplify the comparative analyses, they are recommended for NEBA only if the correlations with
all ecological services or other ecological properties are obvious or established in the NEBA.
That is, a link should be made between the injured or restored ecosystem and the parameter used
to represent the service flows from that ecosystem.
The major advantage of habitat equivalency metrics is that few analyses are performed.
All ecological services are presumably represented by one or a few metrics, such as primary
However, at some sites habitat equivalency metrics could be improved. The result of a
HEA or NEBA can be sensitive to the metric used to estimate the net environmental gains of
ecological services or other ecological entities that are associated with a restoration alternative.
For example, Strange et al. (2002) found that the marsh service metrics of (1) primary
productivity, (2) provision of habitat for the endangered light-footed clapper rail (Rallus
3 or values equal the cost of the restoration plan, but this NRDA option is beyond the scope of
longirostris levipes), (3) provision of soil nitrogen, (4) provision of benthic invertebrate prey for
fish and shellfish and (5) secondary productivity all resulted in different compensatory restoration
quantities. In general, the “marginal contribution” of a land area to the abundance of an
endangered species is not well understood (Unsworth and Bishop 1994). Thus, compensatory
restoration should usually be determined based on multiple services, and the range of these results
will provide an estimate of one source of uncertainty in the NEBA.
Similarly, HEA can be made more rigorous if the species for which habitat is assessed are
specified. For example, rather than generic ecosystem equivalencies, species-specific habitat
metrics, such as provision of suitable substrate for pitcher plant and sundew or provision of food
for bog turtle could be evaluated. In that way, the quality of the habitat may be quantified in
terms of the number of individuals supported or even the viability of the specified populations.
Moreover, Milon and Dodge (2001) note that habitat equivalency is most applicable to uniform
landscapes with little difference in biological functions across the injured area; thus they had to
adjust basic habitat equivalency equations to account for different coral reef populations with
different area uses and different recovery times.
In addition, it is recommended that practitioners of HEA consider modifications of simple
habitat metrics that reflect more detailed habitat requirements of populations of concern, such as
length of edge, connectivity of habitat, and minimum patch size required by the species. Total
area of habitat is not a surrogate for the distribution of habitat. For example, if a hectare of land
is damaged in the middle of a habitat corridor, the affected population is much larger than that
which resides in the damaged area. Similarly, if replacement ecosystem is created at a distance
from the injured ecosystem, the connectivity may be lost. The DOI considered wildlife forage
range injury, which went beyond the damaged ecosystem, in its assessment of damages from the
Colonial Pipeline Spill in Virginia (DOI et al. 1999). Moreover, the area of habitat lost is not
correlated with population survival (and not a suitable metric for population injuries) when
toxicity (not ecosystem area loss) reduces forage vegetation or prey, or bioaccumulation leads to
toxicity in the ecological receptor.
Although some NEBA practitioners treat habitat equivalency service-area-year metrics as
the principal, or even sole, metrics for NEBA, we believe that NEBA is a broader concept and
that other comparative metrics and methodologies are worthy of discussion. For example, TCEQ
(TNRCC 2001) will consider other comparative metrics for ecological services analysis. They
state that “out-of-kind services can often be normalized such that they can be compared,” though
guidance on acceptable normalization methods is not provided. Habitat equivalency is an
example of a service-to-service (or resource-to-resource) approach to scaling restoration actions
(Chapman et al. 1998, NOAA 1995). Under the Oil Pollution Act, the preferred restoration
actions are those that restore resources of the same type and quality and of comparable value as
those injured. In contrast, in valuation approaches to scaling, lost and restored resources need not
be of same type and quality. Values of the original and replacement resources are comparable3
according to a chosen metric (Chapman et al. 1998, NOAA 1997).
As opposed to service-to-service approaches to comparing net benefits, valuation
approaches to comparison require that equivalencies between different types of services or
ecological properties be established (Chapman et al. 1998, NOAA 1997). In many cases,
equivalencies are derived through regulatory negotiation; i.e., natural resource trustees may use
economic valuation methods to establish adequate levels of compensatory services (NOAA
1995). Wetland compensation ratios are determined based on a combination of scientific criteria,
4Although this paper emphasizes ecological entities, environmental service valuation may
encompass a wide range of human use values, such as the value of drinking and irrigation water.
5 The precision of natural resource values is often low, as “willingness to accept” measures of
value, which are appropriate for resource damage determinations (where something is lost), are
often two or more times the analogous estimates of “willingness to pay” (Brown and Gregory
negotiations among stakeholders, and the permit applicant’s ability to pay (King and Adler 1991).
The planning phase of a NEBA would have to reflect whether equivalent services or other
ecological properties are those that are equally valued in economic terms by society, or whether
ecological value is more important.
Numerous valuation methods are available to estimate and to compare apparent dollar
values of ecological services4. The willingness to pay for some services can be inferred (termed a
“revealed preference”) from market prices or other estimates of present use of the resources. The
cost of replacing ecosystem services may be an estimate of their value. To determine non-use
values (also termed “intrinsic,” “existence,” or “passive” values) for ecological properties that are
not traded in markets, the willingness to pay for an entity or the willingness to accept the loss of
an entity can be expressed directly (contingent valuation, CV)5 or derived from values of groups
of attributes (conjoint analysis), both through surveys. The D.C. Circuit Court established
through three cases that natural resource trustees are not limited to specific valuation methods,
including CV (Jones 1997).
In a non-dollar approach to the comparison of ecological services or other properties,
injuries of different types (human health and ecological risks) are classified as insignificant (de
minimus), highly significant (de manifestis), or intermediate, and therefore requiring
consideration of non-risk factors prior to a remediation decision (Suter et al. 1995). Implicit in
this categorization is the assumption that an increased number of species, amount of area, or value
of species (according to regulation or local preference) affected constitutes a greater injury. De
minimus ecological risk, for example, is defined, based on regulatory precedents, as (1) “less than
20% reduction in the abundance or production of an endpoint population within suitable habitat
within a unit area,” (2) “loss of less than 20% of the species in an endpoint community in a unit
area,” or (3) “loss of less than 20% of the area of an endpoint community in a unit area.”
However, Suter et al. (1995) acknowledge that “the loss of all individuals from 20% of the range
of a population can be considered equivalent to loss of 20% of individuals from the entire range
of a population,” except for the fact that these entities would be expected to recover at different
rates. This type of comparative analysis is qualitative rather than quantitative, and would not
apply to instances where restoration must offset damage exactly.
Another non-economic type of comparative valuation metric is the past available solar
energy (“emergy”) required to produce goods and services (Odum and Odum 2000). Thus, a
deer population would have a higher value per joule than their food. This metric may be
correlated with the recovery time for these entities following ecosystem removal (e.g., via
excavation). In one example, the cost of constructing and operating Mississippi River diversions
to marshes were compared with benefits using an emergy analysis, whereby “natural and human
contributions required to construct and operate two diversions were expressed in common units of
solar energy” (Martin 2002).
4.1.5 Temporal Measures of Exposure and Effects
NEBA analyzes ecological gains and losses associated with alternative management
options through time. The time-scale of comparison should include the duration of injury
combined with the longest construction and recovery period of the alternative actions, past,
present and future. In NRDA, the lifetime of the replacement project is significant, if it is not
expected to persist indefinitely (NOAA 1995). If future benefits are discounted, as in most
NRDA analyses (Sect. 4.4.1), benefits after a few decades become negligible. Various stressors
or restoration actions act continuously (e.g., persistent chemicals in environmental media), and
others act almost instantaneously (e.g.,excavation). To estimate future effects, the rates of natural
attenuation (biodegradation and aging through sorption to soil), the rates of contaminant removal
or changes in bioavailability through remediation, and the rates of ecological recovery should be
estimated. These dynamics may be incorporated in the characterization of exposure or the
characterization of ecological effects (Fig. 2). Temporal analysis may be de-emphasized in a
NEBA if (1) estimates of current ecological states are much more certain than estimates of future
and past states and (2) temporal analysis is not required by the relevant statute.
4.1.6 Spatial Measures of Exposure and Effects
Because NEBA is a comparative analysis, it must include the largest spatial extent of
analysis of any single alternative. That is, if ecological restoration is proposed one kilometer
from the area of contamination, the state of the ecological services or other properties at that
location must be ascertained under the competing, alternative scenarios. TCEQ (TNRCC 2001)
requires that ecological restoration occur in the “same ecosystem.” In addition to offsite
restoration, offsite contamination is possible under natural attenuation and active remediation
4.1.7 Conceptual Model
A conceptual model, a concept that is borrowed from risk assessment, is a graphical
representation of the relationships between the chemical or nonchemical stressor and the
responses of ecological services or other properties (Suter et al. 2000). If there is no connection
between a stressor and a service, then the service does not need to be represented in the NEBA.
Contaminant exposure pathways should be considered for all alternatives in the NEBA. For this
reason, the “contaminant risk” module that is depicted in detail in Fig. 2 is also included in Fig. 3
and Fig. 4, the NEBAs for remediation and restoration, respectively. The conceptual model for
NEBA of remediation alternatives should include stressor-service pathways for remedial
technologies, such as the link between nutrients added in bioremediation and plant growth or
diversity, or the link between excavation and vegetation cover (Fig. 3). If multiple alternatives
include a particular remedial technology, such as dredging of hot spot contamination, this portion
of the conceptual model should be depicted in all alternatives. If injuries of contaminants are
indirect, i.e., if wildlife habitats or forage vegetation are directly affected by chemicals, but not
the animals themselves, the connections between habitat or food and vertebrate population
properties should be considered in the NEBA. Similarly, if the restoration plan calls for
restoration of an ecosystem, the conceptual model should show how ecological services and other
ecological properties are expected to be affected.
4.1.8 Analysis Plan
The analysis plan includes data collection, modeling, and logical analyses that are
described or implicit in the NEBA framework. The plan should explain how exposure will be
modeled, the exposure-response models that will be used, how recovery will be modeled, how net
environmental benefits of different alternatives will be compared (e.g, habitat equivalency), and
how uncertainty will be treated. Because a NEBA is a time-integrated analysis, the analysis plan
should explain how predictions forward and backward in time will be made. Examples of
assumptions include: first-order chemical degradation, instantaneous removal of plants during
excavation, or linear recovery of an ecological property. The analysis plan should describe how
sampling design decisions may influence the power to detect injuries relative to the reference
state (Peterson et al. 2001). The analysis plan may describe NEBA results that would cause
assessors to develop improved alternatives and to repeat the NEBA (Fig. 1).
4.2 CHARACTERIZATION OF REFERENCE STATE
NEBA involves the comparison of each alternative to a common reference state to
determine net environmental benefit. Two potential reference states are (1) the contaminated
reference state, equivalent to natural attenuation and ecological recovery, and (2) the
uncontaminated reference state. Particular reference states may be mandated by regulations. For
example, in NRDA, generally, a reference state consists of past, present and future conditions that
would have prevailed in the absence of disturbance, i.e., the uncontaminated reference state (Fig.
1). In a CERCLA remedial investigation, the current state of the environment is typically
characterized in the baseline assessment, and assessment endpoint properties associated with
proposed remedies are compared to the contaminated reference state. The term “baseline” is
avoided here because it has very specific but differing meanings in the CERCLA remedial
investigation and OPA NRDA cases. If net environmental benefits of all alternatives are ranked,
relative to the contaminated reference state, the ranking relative to the uncontaminated reference
state should not differ, because the net environmental benefits in the two comparisons should
differ from each other by the same constants. However, the absolute changes associated with
each alternative depend on which reference state is chosen, and thus the choice of reference state
for the NEBA could influence the decision.
Washington State’s Model Toxics Control Act requires that all remedial alternatives be
compared to the most “practicable” permanent remedial option that is evaluated in the feasibility
study (Washington State Department of Ecology 2001). If a NEBA ranking has been generated
relative to the contaminated or uncontaminated reference state, the same environmental ranking
should exist relative to the most “practicable” remedial option.
Characterizing the reference state is a challenge, and models may be required for analysis
of a temporally changing reference. Ecological services and other properties that are associated
with the uncontaminated state may sometimes be approximated by conditions at a neighboring,
uncontaminated site or conditions prior to the disturbance. The U. S. Department of the Interior
requires that injury quantification in NRDA be based on statistical comparisons between
biological properties in assessment and uncontaminated reference areas (DOI 1995, Barnthouse
and Stahl 2002). However, numerous environmental factors may have acted in concert with the
contamination to alter ecological services following the chemical disturbance. Aquatic
assessments typically utilize regional reference conditions that are bounded by analyses of several
streams, but the adequacy of reference streams is always in question. To the extent possible,
reference states are characterized by seasonal variability, meteorological variability, predator-
prey cycles, and stressors that are not associated with the contamination or remediation or
ecological restoration alternatives. However, in NEBAs, like in NRDAs, uncontaminated
reference states are most often depicted as constant through time, because the variability is
unknown (NOAA 1997).
The characterization of the contaminated reference state may precede or be a parallel
component of a NEBA (Fig. 2). Current and future exposures are estimated, and exposure-
response models are used to estimate injuries, as in CERCLA remedial investigations (Sect. 4.3.2
and Sect. 4.3.3). In Fig. 2, an indirect pathway whereby an ecosystem is disturbed and the areal
disturbance results in injuries, is explicit. Oil and brine spills that occur at exploration and
production sites often have little prolonged, direct toxicity, but large-scale ecosystem removal
may result in injuries to populations. A recovery model is also explicitly included in the
framework for characterization of the contaminated reference state (Fig. 2, Sect. 4.3.4).
4.3 NEBA OF SINGLE ALTERNATIVES
The net environmental benefit of each alternative is the benefit minus the injury of the
alternative. If the net environmental benefit is positive (compared to the reference state), it is
sometimes termed a credit; if it is negative, it is a debit. If the benefits are to different ecological
services than the injuries, either both need to be expressed to the decision maker, or both should
be normalized by a single metric. The comparative metrics discussed above apply.
The subframeworks for NEBA of remediation and ecological restoration are presented in
Fig. 3 and Fig. 4, respectively. Remediation can introduce physical, chemical, or biological
stressors that may partly or wholly balance the ecological benefits from reduced chemical
concentrations. An ecological restoration alternative is designed to have beneficial effects on
ecosystem-level ecological entities, but a rigorous NEBA is recommended. Both subframeworks
include the estimation of exposure, the use of exposure-response models, and the estimation of
recovery. In the ecological restoration subframework, the exposure analysis may be omitted, if
the beneficial effects are well described in the restoration plan. However, the net environmental
benefit of restoration should subtract the value of the services provided by the unrestored land.
The framework for NEBA of natural attenuation if the uncontaminated reference state is
used is presented in Fig. 2. If the contaminated reference state is used in NEBA, the net
environmental benefit of natural attenuation is zero, as these analyses are equivalent.
4.3.1 Time-integrated Analysis
The type of result that may be expected from a NEBA for each service or other ecological
property is presented in Fig. 5. In this hypothetical example, following an oil spill the state of
ecological services or other properties is rapidly degraded. However, the ecological property is
expected to improve with time during natural attenuation of the contamination, followed by
recovery (Fig. 5). The level of the ecological property associated with the uncontaminated
reference state is assumed to continue at approximately the pre-spill level. Although natural
variability is expected, the ecological property in the uncontaminated reference state is generally
considered to be a constant (Fonseca et al. 2000). The proposed remedial alternative is expected
to reduce the ecological property initially (as excavation would reduce vegetation production),
but recovery is expected to be completed more rapidly than in the natural attenuation case (Fig.
5). In the proposed ecological restoration alternative, restoration is achieved more rapidly than
reference statereference state
(natural attenuation)(natural attenuation)
Restored stateRestored state
Fig. 5. Trajectory of assessment endpoint entity (service or other ecological
property) with time, following a petroleum spill; conditions that would have been expected
to prevail in the absence of the spill; expected trajectory of the remediated state; expected
trajectory of the restored state.
ecological recovery in the natural attenuation alternative, and the final level of the ecological
property is greater than the pre-spill level.
The net environmental benefit of remediation, compared to the contaminated reference
state, is the area under the ecological property curve for the remediated state, minus the area
under the ecological property curve for the contaminated reference state (Fig. 5). Note that in this
instance, the net environmental benefit is close to zero and may be less than zero. The net
environmental benefit of restoration, compared to the contaminated reference state (natural
attenuation plus recovery), is above zero (Fig. 5). In one example of NEBA, the net benefit of
planting to restore marsh services was calculated (Penn and Tomasi 2002). The spatial scope of
the contaminated reference state included the unplanted marsh platform. In a non-monetary
restoration under NRDA, restoration is intended to offset the loss in the past and prior to
complete implementation of the alternative. Therefore, the net environmental benefit of
ecological restoration, compared to the uncontaminated reference state, should be zero or greater
to meet the NRDA management goal. The service is restored to a level above the pre-spill level
in Fig. 5 to compensate for past loss. In practice, the target net environmental benefits of
restoration might not be determined until after the injuries from chemical contamination are
estimated, because restoration may be intended to offset exactly the injuries. Net environmental
benefit is often expressed in integrated service-hectare-years. The type of analysis shown in Fig.
5 should be performed for each ecological service or other ecological property, or for a service
(e.g., primary productivity) that represents many other services.
4.3.2 Characterization of Exposure
If benefits and injuries associated with alternatives are not obvious, they may be
determined by exposure-response relationships if exposure is characterized. The characterization
of exposure is the estimate of the magnitude of contact or co-occurrence of a stressor with an
ecological service or other property. An analysis of a proposed ecological restoration alternative
could omit the characterization of exposure (Fig. 4), unless excavation, construction, tilling,
trampling or vehicle movement constitute significant stressors; benefits should be quantified in
the restoration plan.
Present estimates of exposure to chemicals may be determined by the measurement of
chemical concentrations in soil or water, with an assumption about the statistical distribution of
unmeasured contamination. Past and future contamination can be estimated with simulation
models or, if data permit, statistical forecasting or hindcasting. Although extensive and frequent
biological surveys (with reference locations) may obviate the need for exposure analysis to
estimate current effects of chemicals, these surveys would have to be accompanied by exposure
measurements and non-conservative modeling to estimate future injury. Chemical exposures
change through time through the processes of transport (leaching, volatilization), sorption,
degradation, and transformation (Fig. 2). Changes in bioavailability should be estimated, as they
are predictors of effects.
Exposures may be continuous or instantaneous. Chemical exposures are typically
considered as continuous functions. However, if the change in exposure occurs over a time
period that is short compared to the scale of the analysis (e.g, rapid degradation), such changes
may be treated in a step-wise fashion. That is, the curves depicted in Fig. 5 would have vertical
lines at certain time points if exposure was assumed to change instantaneously and ecological
services were assumed to change instantaneously with exposure. The removal of chemicals by
excavation may be assumed to cause instantaneous effects.
Because NEBA is a comparative analysis, exposure estimates for any particular stressor
or alternative should not be expressed conservatively, though uncertainties may be noted.
Conservative injury estimates for an alternative will lead to inappropriate ranking of alternatives
in the comparative part of the NEBA. In typical chemical risk assessments, several conservative
assumptions are made. An organism is often assumed to be exposed to the maximum, measured
concentration of a chemical across space and time, or at the very least, an upper confidence limit
on the mean of that concentration. Similarly, chemicals at non-detected concentrations are often
assumed to be present at the chemical detection limit for the medium. Biodegradation is
occasionally assumed to be zero. Disturbed ecosystems are sometimes assumed to be entirely
unavailable to biota, even when they are partially utilized. None of these are valid assumptions
for a comparative NEBA.
Determining the rate and extent of biodegradation is relevant to NEBAs for natural
attenuation and enhanced bioremediation. More models exist to aid in the estimation of
biodegradation rates in groundwater than in surface soil to which ecological receptors are
exposed. Rates depend on the concentrations of chemicals in soil, status of the microbial
populations, and soil types, among other factors. During a field test of phytoremediation, a first-
order model explained local chemical disappearance at some locations but not others (Nedunuri et
al. 2000). The assumption of rapid, first-order kinetics in soil is often erroneous, because of the
nutrient and oxygen limitations, insolubility of the bulk of a hydrocarbon mixture, sequestration
of hydrophobic constituents in soil pores, potential toxicity of chemicals and byproducts,
differential degradation of different hydrocarbon constituents, seasonal changes in rates, and time
required for microbial acclimation (Odermatt 1997, Duncan et al. 1999, Samson et al. 1994).
Dibble and Bartha (1979) have shown a good correlation between the rate of disappearance of
hydrocarbons and monthly average temperatures in the field.
One researcher has developed a predictive method for estimating the average extent of
petroleum hydrocarbon degradation in land farms, based on initial hydrocarbon composition
(Huesemann 1995). Because total petroleum hydrocarbon concentrations appear to level off by
20 weeks of treatment in this study, it may be advisable in a NEBA to treat the change in
hydrocarbon concentration as instantaneous, if the scale of a NEBA is two decades or more. If
degradation effectively ceased, only the proportion of chemical degraded would be needed to
Although changes in the bioavailability of hydrocarbons and other chemicals are known
to occur (e.g., aging), the rates of the sorptive and diffusive processes that contribute to these
changes are difficult to estimate. Ongoing research may provide rate constants for these
processes, as well as bioavailability factors for ingestion of hydrocarbons by mammals and birds.
126.96.36.199 Chemical removal
The dynamics of source removal and associated changes in exposure may be adequately
estimated in a remedial work plan. As stated above, excavation of soil to remove chemicals may
be treated as an instantaneous process, with respect to chemical contamination.
4.3.3 Exposure-response Relationships
NEBA practitioners may determine the trajectory of ecological services or other
properties through time, either directly, or based on one or more exposure-response models.
However, continuous chemical exposure-response relationships are rarely available for species in
soil. Exposure-response thresholds (Lowest Observed Adverse Effects Concentrations) or
estimated EC50s (median effective concentrations) are commonly available for soil
contamination, but the roles of soil type, receptor taxa, multiple chemicals, aging of chemicals,
and acclimation to toxicity are not well understood. Moreover, the magnitude of the exceedence
of a screening value does not reveal much about the magnitude of risk above the threshold (e.g.,
the percentage of the community that is injured) or the probability of injury, unless the exposure-
response relationship is known. In some aquatic NRDAs, agencies have used exceedences of
water and sediment-quality criteria as evidence of injury; however, elevated chemical
concentrations are not, by themselves, reliable indicators of adverse natural resource effects
(Barnthouse and Stahl, 2002).
In addition, toxicity threshold estimates tend to be conservative, and, as stated above,
conservatism should play no role in a comparative NEBA. Ecotoxicological benchmarks or
screening values, which usually represent estimates of thresholds, are conservative: they tend to
represent low values in the distribution of toxic thresholds (e.g., 10th percentile of values for
phytotoxicity, Efroymson et al. 1997), and they are often based on tests in soils to which
chemicals have been freshly added. Thus, toxicity of aged chemicals is sometimes not observed
in the field at these concentrations (Suter et al. 2000). Tests of field soils or measurements of
effects in the field should be relied on to the extent possible.
Future injuries from chemical contamination may be estimated for NEBA by field
measurement or modeling to determine current ecological states, combined with (1) modeling of
changes in exposure, followed by the use of toxicity and ecological relationships or (2) modeling
of recovery, under the assumption that contamination is below toxic concentrations (Fig. 2).
Toxicity tests performed at multiple times can indicate the approximate rate of reduction of
toxicity with time. For example, toxicity of diesel fuel to Tradescantia in artificial soil was
significantly removed by four weeks after planting (Green et al. 1996). Similarly, bioremediation
treatment (tilling, fertilization and liming) of a fuel-spill-contaminated soil led to the removal of
phytotoxicity after 20 weeks (Wang and Bartha 1990). Marwood et al. (1998) recommend a
battery of toxicity tests to monitor the progress of bioremediation. If tests are carried out in the
field, estimates of exposure may not be needed.
In addition, chemical contaminants such as petroleum can act by disturbing ecosystems
(Fig. 2). Therefore, models that relate area and distribution of disturbed ecosystems to population
sustainability would also be useful for estimating injuries from brine scars, some petroleum spills
(Fig. 2), roads, and trampled areas, as well as injuries from physical remedial alternatives (Fig. 3).
In contrast, assessing the direct impacts of excavation may only require the spatial and temporal
dimensions within which all vegetation is gone; thus Fig. 3 shows no exposure-response model
for physical remediation. Models that may be required to estimate ecological services and other
properties under the restoration alternative may include processes of: primary production,
colonization, succession, population demographics, bioenergetics, and predation (Fig. 4).
The confidence with which injuries should be demonstrated in a NEBA may depend on
the regulatory context. Under NRDA, an injury must be demonstrated with a high degree of
confidence, rather than with the “potential” qualifier that is sometimes used in risk assessment
Ecological recovery is a key determinant of net environmental benefit but is difficult to
quantify. Recovery typically refers to the colonization, growth or succession of an ecological
entity, following the effective removal of the direct pressure of a stressor. Recovery defines the
end of the NEBA analysis. As depicted in Fig. 2 and Fig. 3, recovery modeling estimates the
reduction over time of the effects of contamination or of remedial actions. Recovery models may
also apply to ecological restoration alternatives where services or other ecological properties are
restored as a consequence of the direct restoration goal. Although the term “recovery” is not used
in Fig. 4, the restoration-response models may be equivalent to recovery models, but the rate of
recovery is increased by the restoration action. Guidance from the TCEQ (TNRCC 2001) notes
that “estimates of recovery time may come from literature, site-specific information, or other
affected property investigations,” but this information is often difficult to obtain. In addition,
certain services may not be measurable at the spatial scale of the action. For example, a restored
riparian wetland was too small for investigators to measure recovery of small mammal
populations (Wike et al. 2000).
Estimates of recovered ecological services or other properties require either (1) a specific
function or (2) an approximate time to recovery and an assumption about the curve shape needed
to get there. Example times to recovery of ecological properties from petroleum mixtures in
various ecosystems (mostly in northern climates) are summarized in Table 3. Examples of times
Table 3. Select studies of ecological recovery from petroleum spills
fuel spills in lysimeters, followed by
tilling, fertilization and liming
0.4Wang and Bartha
crop productivitykerosene contamination, followed by
tilling, fertilization and liming
1Dibble and Bartha
JP-4 fuel-contaminated soil in 40 Air
Force bioassays (reflecting field
contamination incidents) with sorghum
or pinto bean
2 to 4 (0.5 with
Lillie and Bartine
crude oil applied to saturation at depth of
4 ft, with tillage, in Stillwater, Oklahoma
2 McKay and Singleton
experimental crude oil spill of 10L/m2 at
Masters Vig, Northeast Greenland
experimental crude oil spill of 10L/m2 at
Masters Vig, Northeast Greenland, moist
5 to 8, estimateHolt 1987
experimental crude oil spill of 10L/m2 at
Masters Vig, Northeast Greenland, dry
experimental crude oil spill of 10L/m2 at
Masters Vig, Northeast Greenland, dry
>8, estimate Holt 1987
crude oil spills in tundra communities of
Mackenzie Mountains, Northwest
20 Kershaw and
crude oil spills in tundra communities of
Mackenzie Mountains, Northwest
crude oil applied to saturation at depth of
4 ft, with tillage, in Stillwater, Oklahoma
5 McKay and Singleton
crude oil applied to saturation at depth of
4 ft, with tillage, in Stillwater, Oklahoma
>5McKay and Singleton
experimental crude oil spill of 3273 L on
simulated pipeline trench near Tulita,
Northwest Territories, Canada
1 Seburn et al. 1996
Table 3. Select studies of ecological recovery from petroleum spills
cover of majority
of plant species
experimental crude oil spill of 3273 L on
simulated pipeline right-of-way near
Tulita, Northwest Territories, Canada
>3 Seburn et al. 1996
experimental crude oil spill sprayed at
9.1 L/m2 on Low Arctic tundra near
Tuktoyaktuk, Northwest Territories,
10 to 15,
shrub, moss and
lichen cover )
crude oil sprayed on soil surface at
18L/m2 in black spruce taiga forest,
20 Racine 1994
vegetation covercrude oil spill from pipe at 100-250L/m2
in black spruce taiga forest, interior
Alaska; locations with oil remaining in
20 Racine 1994
vegetation covercrude oil spill from pipe at 100-250L/m2
in black spruce taiga forest, interior
Alaska; locations with asphalt-like
surface oil remaining
crude oil applied at 11L/m2 to plots in
Port Harcourt, Nigeria
>0.8 Kinako 1981
crude oil applied at 11L/m2 to plots in
Port Harcourt, Nigeria
>0.5 Kinako 1981
oil spills10 to 20,
Booth et al. 1991
to recovery of ecological properties from physical stressors and bioremediation are summarized
in Table 4. One type of estimate of the minimum time to recovery could be provided by the
average age of the lost vegetation (Vasek et al. 1975). Recovery of total vegetation cover from
petroleum spills may often occur more rapidly than recovery from physical disturbance, although
comparative tests of both types of disturbance in the same ecosystem have not been undertaken.
Recovery from summer spills may be slower than recovery from winter spills (Freedman and
Hutchinson 1976). To determine whether or not studies of ecological recovery from various
stressors are relevant to recovery from oil spills or oil spill remediation alternatives, additional
studies of oil spill recovery are needed. Time periods preceding recovery may sometimes be
extrapolated from measurements taken at another site; however, most data on recovery relate to
aquatic rather than terrestrial ecosystems (Niemi et al. 1990, Booth et al. 1991, NOAA 1990), and
few published studies relate to recovery from remedial actions (Tamis and Udo de Haes 1995).
The recovery trajectory is often assumed to be linear (NOAA 1995). In reality, the
dynamics of recovery may be complex, for example, encompassing recovery from multiple
Table 4. Select studies of ecological recovery from types of disturbance that may be associated
with remediation alternatives
soil structureWheel-rutted and other compacted
soils from tree harvesting by
skidder in Northern Mississippi
8 to 12 Dickerson 1976
soil structurecompaction of soil in Wahmonie, a
Mojave desert ghost town
>100, estimate Webb and Wilshire
soil bulk density compaction of soil from timber
harvesting by bulldozer or skidder
in New South Wales, Australia
>5 Croke et al. 2001
crop root weightsclay loam mechanically compacted
in plow furrow in southwestern
>9Blake et al. 1976
burn of oil spill in high marsh in
7-8, estimate Tunnell et al. 1997,
desert road, subsequently
>87Bolling and Walker
pipeline construction in the
southern Mojave desert (trenching,
piling and refilling)
Vasek et al. 1975
open-cast coal mining and
reclamation (ploughing and
3 to 15 Rushton 1986
climax community of
thermal cleaning of soil 100, estimateTamis and Udo de
climax community of
bioremediation <10, estimateTamis and Udo de
processes, such as soil compaction, colonization and succession of vegetation. Vasek et al.
(1975) suggest that the recovery of properties of scrub vegetation, such as composition and
percentage ground cover, would occur with sigmoidal temporal dynamics. The temporal
dynamics of recovery can be estimated by monitoring during the assessment period, but two data
points are never sufficient to establish the shape of the curve.
Factors that influence the recovery of ecosystems from disturbance include: current state,
disturbance severity and frequency, successional history, history of disturbance, preferred state,
management of the disturbance, and random factors such as weather (Fisher and Woodmansee
1994). Recolonization time is dependent on the size of the site and the proximity to a
recolonization source. Species that are characteristic of early successional communities recover
relatively rapidly from disturbance to colonize disturbed areas, due to their high reproductive
rates and rapid dispersal mechanisms (Booth et al. 1991). These factors are incorporated into
Restoration measures may reduce the time to recovery in remediation or natural
attenuation scenarios. It should be noted that the recovery of one ecological service or other
property can be impeded by restoration of another; for example, the maintenance of caps requires
that deeply rooted vegetation and burrowing mammals be kept off a site, inhibiting recovery of
some potential endpoints (Suter et al. 1993).
Recovery following excavation and landfarming is likely to be of longer duration than
toxic effects of low concentrations of petroleum hydrocarbons. So, while biodegradation may be
estimated as an instantaneous process, recovery should not be. The error associated with the use
of linear estimates of recovery will depend on the duration of the recovery, relative to the time
scale of the NEBA.
4.4 COMPARISON OF MULTIPLE ALTERNATIVES
Following the net environmental benefit calculation for individual alternatives, the net
environmental benefits of each alternative are compared (Fig. 1). As stated in the previous
section, benefits and injuries of different types can only sometimes be normalized by a single
metric in the NEBA for single alternatives. Similarly, the net environmental benefits of multiple
alternatives may be ranked in the NEBA only if normalizing metrics are available (Sect. 4.1.4). If
net environmental benefits of different alternatives are expressed in different units, the land
managers or trustees may rank the alternatives subjectively. However, the ranking is likely to be
more acceptable to stakeholders if relative values were established during the planning phase.
If alternatives are compared to the contaminated reference state, both remediation and
restoration alternatives may have positive net environmental benefit. However, if alternatives are
compared to the uncontaminated reference state, and the analysis includes the period of past
damage as in NRDA, a contaminant removal alternative may ultimately provide the level of
ecological services that were lost, but will never compensate for the past lost services and will not
have a net environmental benefit. In NRDA compensatory actions, restoration is required to
provide a net environmental benefit.
In a special case of NEBA, the net environmental benefits of each restoration alternative
(relative to the uncontaminated reference state) are compared to the net injuries from
contamination using HEA. As stated above, the management goal in NRDA and TCEQ’s
ecological services analysis is to compensate for the lost services. HEA is commonly used to
identify ecosystem replacement projects that provide resources and ecological services of the
types that have been lost and will be lost prior to the complete restoration. That is, the net
environmental benefit, compared to the uncontaminated reference state, must be at least as great
as the debit of services associated with contamination, compared to the uncontaminated reference
state. HEA can include the monetary value of services, but only the ecological aspect is
discussed here. HEA considers the recovery time-path of injured resources and services, the
relative productivity of restored ecosystems, community succession in restored ecosystems, and
the project life span (DOI et al. 1999). In one example of HEA related to petroleum damage, 7.5
ha of marsh plant strips were calculated to be needed for compensatory restoration, and 15.9 ha
was the area more rapidly colonized because of the strips (Penn and Tomasi 2002).
If a single comparative metric is used in HEA (e.g., primary production service-hectare-
years), a single graph of ecological services through time for each alternative (analogous to Fig.
5) can depict the dynamics from which relative net environmental benefits of alternatives can be
calculated. If multiple services or multiple, species-specific habitats are explicitly considered,
then multiple analyses of ecological services through time are performed, and compensatory
restoration may be determined through a weight of evidence. The mathematics of HEA are
illustrated in Penn and Tomasi (2002) and Milon and Dodge (2001). In addition to the temporal
estimate of the service-hectare-years lost or gained, the discount rate is incorporated into the
4.4.1 Challenges of Comparative Assessments
The challenges of conducting comparative assessments within the NEBA framework
include the few metrics available for relative valuation of alternatives or subtraction of risks from
benefits within a single alternative (discussed above), as well as (1) incomparable conservatism
among assessment results for individual alternative actions, (2) inadequate exposure-response
models to quantify the absolute magnitude of risk, (3) limited utility of the weight-of-evidence
approach to risk assessment, (4) relative valuation of differing magnitudes of uncertainty when
comparing net benefits of competing alternatives, and (5) relative valuation of past, present, and
future conditions within an alternative.
First, comparisons of ecological properties under remediation, natural attenuation, and
restoration alternatives require assumptions of comparable conservatism. Ecological risk
assessments commonly generate conservative estimates of exposure and effects, so estimates of
injury generated independently of the NEBA analysis may be high (therefore a comprehensive
planning phase for NEBA is suggested). (See discussions of conservatism in Sect. 4.3.2 and Sect.
Second, assessors may be limited by the uncertainty of existing, empirical models and
measurements. We are much more confident in predicting the decrement in biomass of plants
where the surface soil has been excavated (i.e., no vegetation) than we are in predicting the
percentage biomass decrement where a particular concentration of petroleum hydrocarbons or
metals is found.
Third, the weight-of-evidence approach that is common in ecological risk assessment
seldom results in a single value for magnitude of injury, because it is usually intended to aid an
assessor in determining whether or not injury is above reference levels. It is difficult to compare
net benefits across alternatives (e.g., how much to increase productivity in a restoration scenario
for that alternative to be recommended) when our most quantitative risk assessments are not very
quantitative. A weight-of-evidence approach could support NEBA if multiple models or
measurements resulted in different estimates of magnitude of injury or benefit. For example, a
weighting of different metrics could be used to estimate compensatory restoration, as discussed
Fourth, uncertainty has a role in comparative ecological valuation. For example, Arrow
and Fisher (1974) assert that uncertainty and irreversibility of ecological states can “lead to a
reduction in net benefits from an activity with environmental costs.” Irreversibility reduces the
“option value,” which is an important component of an environmental benefit calculation (Chavas
2000). However, uncertainty may make a NEBA comparison indeterminate if the uncertainty is
much greater than the differences in net benefits among the alternatives.
Fifth, to compare present effects in one scenario to future effects in another, future effects
may be discounted to present value, and past service flows should be compounded, through
methods analogous to economic discounting. Discounting is typically practiced in HEA, as well
as in non-HEA NEBAs. NOAA usually recommends applying a three percent rate for
discounting interim losses and gains (NOAA 1997), but the concept of discounting needs to be
more thoroughly examined by practitioners and managers of NEBA before a recommendation is
made. Some investigators recommend a zero discounting rate for ecological functions with no
economically quantifiable human use (Milon and Dodge 2001). Moreover, NEBA results are
highly sensitive to the choice of discounting rate (Milon and Dodge 2001). The discount rate
should be adjusted if the value of services is not constant over time, for example, if the marginal
value of wetland increases because its land area is decreasing, or if the marginal value of wetland
decreases because the cost of creating new wetlands or substituting for its services is reduced
(Unsworth and Bishop 1994).
5. ADDITIONAL CONSIDERATIONS
NEBA provides important information about environmental benefits and injuries of
alternative actions to decision-makers. As shown in Fig. 1, cost-effectiveness is also an important
criterion for the decision, although it is outside of the NEBA framework. Essentially, the net
environmental benefit, divided by cost, results in an estimate of cost-effectiveness. Human health
risk will probably also inform the decision, and sometimes a human health risk assessment or
screening analysis may be required for a NEBA to be considered by regulatory agencies
(TNRCC 2001). The OPA regulations specify criteria for selecting a preferred restoration
alternative: (1) cost of the alternative; (2) “extent to which each alternative is expected to meet
the trustees’ goals and objectives in returning the injured natural resources and services to
baseline and/or compensating for interim losses” (NOAA 1997); (3) likelihood of success; (4)
extent to which the alternative will prevent future injury from the incident and avoid collateral
injury from implementing the alternative; (5) extent to which the alternative benefits more than
one natural resource or service; and (6) effect of the alternative on health and safety (NOAA
1997). All of these criteria are compatible with NEBA, and #2 and #4 recommend NEBA-type
A NEBA may be performed iteratively as alternative actions are optimized, preferably
before an action is implemented (Fig. 1). Monitoring and efficacy assessment may be considered
external to the NEBA framework, but these processes produce data that may result in a decision
to alter the preferred alternative and possibly to perform the NEBA again. The Great Lakes
Water Quality Board recommends “that much greater emphasis be placed on post-project
monitoring of effectiveness of sediment remediation” (Zarull et al. 1999). Restoration actions
sometimes fail to meet their goals, and adaptive management can lead to effective redesign of
restoration alternatives (Thom 2000), especially if ecological properties have been well-specified
in the NEBA planning phase. The NEBA framework is open in that it does not demand that net
environmental benefit be the only criterion for a decision-maker.
A framework for Net Environmental Benefit Analysis (NEBA) of contaminated sites has
been developed. NEBA is expected to be useful for decision-making related to petroleum spills
and other contaminated sites. NEBA takes a holistic approach to decision-making for chemical
contamination, considering both expected risks from remedial actions and direct benefits that are
possible through ecological restoration. The methodology is expected to be most useful for
decisions where certain types of injuries (e.g., from remedial actions) or benefits (e.g., from direct
ecological restoration or enhanced remediation) are not consistently, rigorously assessed. NEBA
provides a methodology whereby state and federal remedial investigations and NRDAs may be
undertaken under a single framework. Similarly, NEBA provides a framework whereby
proactive, nonregulatory-driven restoration may occur. A comparative assessment such as a
NEBA should have a single planning phase that encompasses all relevant, alternative actions:
natural attenuation, traditional remediation, and restoration alternative(s). Comparative metrics
should be agreed upon. NEBA utilizes a reference state (contaminated or uncontaminated) that is
consistent with management goals. A NEBA may result in additional alternatives that are
optimized for net environmental benefit or that are targeted to provide offsetting environmental
benefit, compared to the uncontaminated reference state. Supporting knowledge that requires
further development includes: non-monetary valuation metrics; non-conservative, quantitative
exposure-response models; models of recovery; and strategies for ecological restoration. The use
of NEBA should result in better decisions, resulting in greater improvements in environmental
quality at lower cost.
API (American Petroleum Institute). 1999. Compilation and review of data on the environmental
effects of in situ burning of inland and upland oil spills. Health and Environmental Sciences
Department. Publication Number 3684. Washington, D.C.
Arrow, K. J., and A. C. Fisher. 1974. Environmental preservation, uncertainty, and
irreversibility. Quarterly J. Econ. 88:312–319.
Baker, J. M. 1999. Ecological effectiveness of oil spill countermeasures: How clean is clean?
Pure Appl. Chem. 71:135–159.
Barnett, E. 1999. Florida’s incentive-based environmental protection: Alternative regulatory
approach. Presentation. Environmental policies for the new millennium: Incentive- based
approaches to environmental stewardship. Conference, November 2-3 1999. World Resources
Institute. Washington, D.C. http://www.igc.apc.org/wri/incentives/barnett.htm
Barnthouse, L. W., J. J. Bascietto, S. A. Deppen, R. W. Dunford, D. E. Gray, and F. E. Sharples.
1995. Natural Resource Damage Assessment Implementation Project: Savannah River Site. U.S.
Department of Energy, Office of Environmental Policy & Assistance, Washington, D.C.
Barnthouse, L. W., and R. G. Stahl. 2002. Quantifying natural resource injuries and ecological
service reductions: challenges and opportunities. Environ. Manage. 30:1–12.
Blake, G. R., W. W. Nelson, and R. R. Allmaras. 1976. Persistence of subsoil compaction in a
mollisoil. Soil Sci. Soc. Am. J. 40:943–948.
Bolling, J. D., and L. R. Walker. 2000. Plant and soil recovery along a series of abandoned
desert roads. J. Arid Environ. 46:1–24.
Booth, P. N., D. S. Becker, R. A. Pastorok, J. R. Sampson, and W. J. Graham. 1991. Evaluation
of restoration alternatives for natural resources injured by oil spills. Publication Number 304.
American Petroleum Institute. Washington, D.C.
Brown, T. C., and R. Gregory. 1999. Why the WTA-WTP disparity matters. Ecolog. Econ.
Chapman, D., N. Iadanza, and T. Penn. 1998. Calculating resource compensation: an
application of the service-to-service approach to the Blackbird Mine hazardous waste site.
National Oceanic and Atmospheric Administration Damage Assessment and Restoration program
Technical Paper 97-1.
Chavas, J.-P. 2000. Ecosystem valuation under uncertainty and irreversibility. Ecosystems
Croke, J., P. Hairsine, and P. Fogarty. 2001. Soil recovery from track construction and
harvesting changes in surface infiltration, erosion and delivery rates with time. Forest Ecol.
Department of Commerce. 1996. Oil Pollution Act Regulations. Part 990. Natural Resource
Damage Assessments. 15 CFR Part 990.
Dibble, J. T., and R. Bartha. 1979. Rehabilitation of oil-inundated agricultural land: a case
history. Soil Sci. 128:56–60.
Dickerson, B. P. 1976. Soil compaction after tree-length skidding in Northern Mississippi. Soil
Sci. Soc. Am. J. 40:965–966.
DOI (U.S. Department of the Interior). 1995. 43 CFR Subtitle A, Part 11. Natural Resource
DOI (U. S. Department of the Interior), Virginia Department of Environmental Quality, and
District of Columbia Department of Health. 1999. Colonial Pipeline Oil Spill, Reston, Virginia.
Final Restoration Plan and Environmental Assessment. National Park Service, U.S. Department
of the Interior, Washington, D.C.
Duncan, K. E., R. Kolhatkar, G. Subramaniam, R. Harasimhan, E. Jennings, S. Hettenbach, A.
Brown, C. McComas, W. Potter, and K. Sublette. 1999. Microbial dynamics in oil-impacted
prairie soil. Appl. Biochem. Biotechnol. 77–79:421–434.
Efroymson, R. A., M. E. Will, G. W. Suter II and A. C. Wooten. 1997. Toxicological
Benchmarks for Screening Contaminants of Potential Concern for Effects on Terrestrial Plants:
1997 Revision, ES/ER/TM-85/R3. Oak Ridge National Laboratory, Oak Ridge, Tennessee.
EPA 1989. Risk Assessment Guidance for Superfund. Vol. I. Human Health Evaluation
Manual. EPA 540/1-89/002. Office of Emergency and Remedial Response, U.S. Environmental
Protection Agency, Washington, D.C.
EPA 1998. Guidelines for Ecological Risk Assessment. EPA/630/R-95/002F. Risk Assessment
Forum, U.S. Environmental Protection Agency, Washington, D.C.
EPA 1999. Use of Monitored Natural Attenuation at Superfund, RCRA Corrective Action, and
Underground Storage Tank Sites. OSWER Directive 9200.4-17P. Washington, D.C.
Fisher, S. G., and R. Woodmansee. 1994. Issue paper on ecological recovery. pp. 7–1 to 7–54.
in Ecological Risk Assessment Issue Papers. EPA/630/R-94/009. U.S. Environmental Protection
Agency, Washington, D.C.
Fiocco, R. J., and A. Lewis. 1999. Oil spill dispersants. Pure Appl. Chem. 71:27–42.
Fonseca, M. S., B. E. Julius, and W. J. Kenworthy. 2000. Integrating biology and economics in
seagrass restoration: How much is enough and why? Ecol. Eng. 15:227–237.
Green B. T, C. T. Wiberg, J. L. Woodruff, E. W. Miller, V. L. Poage, D. M. Childress, J. A.
Feulner, S. A. Prosch, J. A. Runkel, R. L. Wanderscheid, M. D. Wierma, X. Yang, H. T. Choe,
and S. D. Mercurio. 1996. Phytotoxicity observed in Tradescantia correlates with diesel fuel
contamination in soil. Environ. Exper. Bot. 36:313–321.
Freedman, W., and T. C. Hutchinson. 1976. Physical and biological effects of experimental
crude oil spills on Low Arctic tundra in the vicinity of Tuktoyaktuk, N.W.T., Canada. Can. J.
Hackney, C. T. 2000. Restoration of coastal habitats: expectation and reality. Ecolog. Eng.
Heath, J. S. 1999. Introduction: guidance on natural attenuation in soils and groundwater. J.
Soil Contam. 8:3–7.
Holt, S. 1987. The effects of crude and diesel oil spills on plant communities at Mesters Vig,
Northeast Greenland. Arctic Alpine Res. 19:490–497.
Huesemann, M. H. 1995. Predictive model for estimating the extent of petroleum hydrocarbon
biodegradation in contaminated soils. Environ. Sci. Technol. 29:7–18.
Jones, C. A. 1997. Use of non-market valuation methods in the courtroom: recent affirmative
precedents in Natural Resource Damage Assessments. Water Resources Update, Issue 109,
Autumn 1997. Universities Council on Water Resources, Carbondale, Ill.
Kershaw, G. P., and L. J. Kershaw. 1986. Ecological characteristics of 35-year-old crude-oil
spills in tundra plant communities of the Mackenzie Mountains, N.W.T. Can. J. Bot.
Kinako, P. D. S. 1981. Short-term effects of oil pollution on species numbers and productivity
of a simple terrestrial ecosystem. Environ. Pollut. Ser. A 26:87–91.
King, D. M., and K. J. Adler. 1991. Scientifically defensible compensation ratios for wetland
mitigation. U. S. Environmental Protection Agency Office of Policy, Planning and Evaluation.
Lillie, T. H., and R. W. Bartine. 1990. Protocol for evaluating soil contaminated with fuel or
herbicide. pp. 198–203. In W. Wang W, J. W. Gorsuch, and W. R. Lower (eds.), Plants for
Toxicity Assessment. ASTM, Philadelphia, Pa.
Lin, Q., and I. Mendelssohn. 1998. The combined effects of phytoremediation and
biostimulation in enhancing habitat restoration and oil degradation of petroleum contaminated
wetlands. Ecol. Eng. 10:263–274.
Luftig, S.D. 1999. Issuance of Final Guidance: Ecological Risk Assessment and Risk
Management Principles for Superfund Sites. OSWER Directive 9285.7-28. Office of Emergency
and Remedial Response, U.S. Environmental Protection Agency; Washington, D.C.
Lunel, T., J. Rusin, N. Bailey, C. Halliwell, and L. Davies. 1997. The net environmental benefit
of a successful dispersant operation at the Sea Empress incident. Pp. 185–194. In Proceedings of
the 1997 International Oil Spill Conference. Fort Lauderdale, Fla.
Martin, J. F. 2002. Emergy valuation of diversions of river water to marshes in the Mississippi
River Delta. Ecolog. Eng. 18:265–286.
Marwood, T.M., K. Knoke, K. Yau, H. Lee, J. T. Trevors, A. Suchorski-Tremblay, C. A.
Flemming, V. Hodge, D. L. Liu, and A. G. Seech. 1998. Comparison of toxicity detected by five
bioassays during bioremediation of diesel fuel-spiked soils. Environ. Toxicol. Water. Qual.
McKay, D., and P. C. Singleton. 1974. Time required to reclaim land contaminated with crude
oil–an approximation. Agricultural Extension Service, University of Wyoming, Laramie, Wyo.
Milon, J. W., and R. E. Dodge. 2001. Applying habitat equivalency analysis for coral reef
damage asssessment and restoration. Bull. Mar. Sci. 69:975–988.
Nedunuri, K. V., R. S. Govindaraju, M. K. Banks, A. P. Schwab, and Z. Chen. 2000. Evaluation
of phytoremediation for field-scale degradation of total petroleum hydrocarbons. J. Environ.
Eng. 126: 483–490.
Nicolette, J. P., M. Rockel, and M. J. Kealy. September 2001. Quantifying ecological changes
helps determine best mitigation. Pipe Line Gas Ind. 52–57.
Niemi, G. J., P. DeVore, N. Detenbeck, D. Taylor, A. Lima, J. Pastor, J. D. Yount, and R. J.
Naiman. 1990. Overview of case studies on recovery of aquatic ecosystems from disturbance.
Environ. Manage. 14:571–587.
NOAA Damage Assessment and Restoration Program. 1995. Habitat equivalency analysis: an
overview. National Oceanic and Atmospheric Administration. Silver Spring, Md.
NOAA Damage Assessment and Restoration Program. 1997. Scaling compensatory restoration
actions, guidance document for Natural Resource Damage Assessment. National Oceanic and
Atmospheric Administration. Silver Spring, Md.
NOAA Hazardous Materials Response Branch. 1990. Excavation and rock washing treatment
technology: Net environmental benefit analysis. National Oceanic and Atmospheric
Administration. Seattle, Wash.
NRC (National Research Council). 1992. Restoration of aquatic ecosystems: science,
technology, and public policy. National Academy Press, Washington, D.C.
Odermatt, J. R.. 1997. Simulations of intrinsic biodegradation using a non-linear modification of
first-order reaction kinetics. J. Soil Contam. 6:495–508.
Odum, H. T., and E. P. Odum. 2000. The energetic basis for valuation of ecosystem services.
Penn, T., and T. Tomasi. 2002. Calculating resource restoration for an oil discharge in Lake
Barre, Louisiana. Environ. Manage. 29:691–702.
Peterson, C. H., L. L. McDonald, R. H. Green, and W. P. Erickson. 2001. Sampling design
begets conclusions: the statistical basis for detection of injury to and recovery of shoreline
communities after the ‘Exxon Valdez’ oil spill. Mar. Ecol. Prog. Ser. 210:255–283.
Principe, P. P. 1995. Ecological benefits assessment: A policy-oriented alternative to regional
ecological risk assessment. Hum. Ecol. Risk Assess. 1:423–435.
Racine, C. H. 1994. Long-term recovery of vegetation on two experimental crude oil spills in
interior Alaska black spruce taiga. Can. J. Bot. 72:1171–1177.
Reagan, D. P. 2000. Natural remediation in the risk management process: goals, options, and
monitoring, pp. 9–29. In M. Swindoll, R. G. Stahl, Jr., and S. J. Ells (eds.), Natural Remediation
of Environmental Contaminants: Its Role in Ecological Risk Assessment and Risk Management.
SETAC Press, Pensacola, Fla.
Rubin, R. R., M. J. Brennan, M. C. Bracken, R. L. Hirsch, C. Leonard, and F. M. Rundlett. 2001.
Options for enhancing the transfer of excess and surplus military property for conservation or
recreational use. Bahr, Inc. Middleton, Wis.
Rushton, S. P. 1986. Development of earthworm populations on pasture land reclaimed from
open–cast coal mining. Pedobiologia 29:27–32.
Samson, R., C. W. Greer, T. Jawkes, R. Desrochers, C. H. Nelson, and M. St-Cyr. 1994.
Monitoring an aboveground bioreactor at a petroleum refinery site using radiorespirometry and
gene probes: effects of winter conditions and clayey soil. pp. 329–333. In R. E. Hinchee, B. E.
Alleman, R. E. Hoeppel, and R. N. Miller (eds.), Hydrocarbon Bioremediation. Lewis
Publishers, Boca Raton, Fla.
Seburn, D. C., G. P. Kershaw, and L. J. Kershaw. 1996. Vegetation response to a subsurface
crude oil spill on a subarctic right-of-way, Tulita (Fort Norman), Northwest Territories, Canada.
Sprenger, M. D., and D. W. Charters. 1997. Ecological risk assessment guidance for Superfund:
Process for designing and conducting ecological risk assessment, Interim Final. Environmental
Response Team, U.S. Environmental Protection Agency, Edison, N.J.
Stahl, R. G., and C. M. Swindoll. 1999. The role of natural remediation in ecological risk
assessment. Hum. Ecol. Risk Assess. 5:219-223
State of Florida. 2001. Ecosystem Management Agreements. The 2001 Florida Statutes, Title
XXIX Public Health, Chapter 403, Sect. 403.0752. http://www.leg.state.fl.us/Statutes/
Strange, E., H. Galbraith, S. Bickel, D. Mills, D. Beltman, and J. Lipton. 2002. Determining
ecological equivalence in service-to-service scaling of salt marsh restoration. Environ. Manage.
Susarla, S., V. F. Medina, and S. C. McCutcheon. 2002. Phytoremediation: an ecological
solution to organic chemical contamination. Ecol. Eng. 18:647–658.
Suter, G. W., II. 1999. A framework for assessment of ecological risks from multiple activities.
Hum. Ecolog. Risk Assess. 5:397–413.
Suter, G. W., II, B. W. Cornaby, C. T. Hadden, R. N. Hull, M. Stack, and F. Zafran. 1995. An
approach for balancing health and ecological risks at hazardous waste sites. Risk Analysis
Suter, G. W., II, R. A. Efroymson, B. E. Sample, and D. S. Jones. 2000. Ecological risk
assessment for contaminated sites. Lewis Publishers, Boca Raton,FL.
Suter, G. W., II, R. J. Luxmoore, and E. D. Smith. 1993. Compacted soil barriers at abandoned
landfill sites are likely to fail in the long term. J. Environ. Qual. 22:217–226.
Swindoll, M., R. G. Stahl, Jr., and S. J. Ells. 2000. Natural remediation of environmental
contaminants: its role in ecological risk assessment and risk management, pp. 1–8 In Swindoll,
M., R. G. Stahl, Jr., and S. J. Ells (eds.) Natural Remediation of Environmental Contaminants:
Its Role in Ecological Risk Assessment and Risk Management. SETAC Press, Pensacola, FL.
Tamis, W. L. M., and H. A. Udo de Haes. 1995. Recovery of earthworm communities
(Lumbricidae) in some thermally and biologically cleaned soils. Pedobiologia 39:351–369.
Thom, R. M. 2000. Adaptive management of coastal ecosystem restoration projects. Ecol. Eng.
TNRCC 2001. Guidance for conducting ecological risk assessments at remediation sites in
Texas. Toxicology and Risk Assessment Section, Texas Natural Resource Conservation
Commission, Austin, Tex.
Tunnell, J. W., K. Withers, and B. Hardegree. 1997. Environmental impact and recovery of the
Exxon pipeline spill and burn site, Upper Copano Bay Texas: Final Report. TAMU-CC-9703-
CCS. Texas A&M University-Corpus Christi, Tex.
Unsworth, R. E., and R. C. Bishop. 1994. Assessing natural resource damages using
environmental annuities. Ecol. Econ. 11:35–41.
Vasek, F. C., H. B. Johnson, and D. H. Eslinger. 1975. Effects of pipeline construction on
creosote bush scrub vegetation of the Mojave Desert. Madro o 23:1–13.
Wang X, and R. Bartha. 1990. Effects of bioremediation on residues, activity and toxicity in soil
contaminated by fuel spills. Soil Biol. Biochem. 22:501–505.
Washington State Department of Ecology. Amended February 12, 2001. Model Toxics Control
Act. Chapter 173–340 WAC. Publication No. 94-06 Seattle, Wash.
Webb, R. H., and H. G. Wilshire. 1980. Recovery of soils and vegetation in a Mojave desert
ghost town, Nevada, U.S.A. J. Arid. Environ. 3:291–303.
Weston, D. P., W. M. Jarman, G. Cabana, C. E. Bacon, and L. A. Jacobson. 2002. An evaluation
of the success of dredging as remediation at a DDT-contaminated site in San Francisco Bay,
California, USA. Environ. Toxicol. Chem. 21:2216–2224.
Wike, L. D., F. D. Martin, H. G. Hanlin, and L. S. Paddock. 2000. Small mammal populations in
a restored stream corridor. Ecol. Eng. 15:S121–S129.
Zarull, M. A., J. H. Hartig, and L. Maynard. 1999. Ecological benefits of contaminated sediment
remediation in the Great Lakes Basin. 1999. Sediment Priority Action Committee, Great Lakes
Water Quality Board, International Joint Commission, Washington, D.C.
Page 50 Download full-text