Eagle Conservation Plan Guidance - April 2013

U.S. Fish and Wildlife Service

Eagle Conservation Plan Guidance
Module 1 – Land-based Wind Energy
Version 2

Credit: Brian Millsap/USFWS

U.S. Fish and Wildlife Service
Division of Migratory Bird Management
April 2013

i

Disclaimer

This Eagle Conservation Plan Guidance is not intended to, nor
shall it be construed to, limit or preclude the Service from
exercising its authority under any law, statute, or regulation,
or from taking enforcement action against any individual,
company, or agency. This Guidance is not meant to relieve
any individual, company, or agency of its obligations to
comply with any applicable Federal, state, tribal, or local
laws, statutes, or regulation. This Guidance by itself does not
prevent the Service from referring cases for prosecution,
whether a company has followed it or not.

ii

EXECUTIVE SUMMARY
1. Overview
Of all America’s wildlife, eagles hold perhaps the most revered place in our national history and
culture. The United States has long imposed special protections for its bald and golden eagle
populations. Now, as the nation seeks to increase its production of domestic energy, wind energy
developers and wildlife agencies have recognized a need for specific guidance to help make wind
energy facilities compatible with eagle conservation and the laws and regulations that protect
eagles.

To meet this need, the U.S. Fish and Wildlife Service (Service) has developed the Eagle Conservation
Plan Guidance (ECPG). This document provides specific in‐depth guidance for conserving bald and
golden eagles in the course of siting, constructing, and operating wind energy facilities. The ECPG
guidance supplements the Service’s Land‐Based Wind Energy Guidelines (WEG). WEG provides a
broad overview of wildlife considerations for siting and operating wind energy facilities, but does
not address the in‐depth guidance needed for the specific legal protections afforded to bald and
golden eagles. The ECPG fills this gap.

Like the WEG, the ECPG calls for wind project developers to take a staged approach to siting new
projects. Both call for preliminary landscape‐level assessments to assess potential wildlife
interactions and proceed to site‐specific surveys and risk assessments prior to construction. They
also call for monitoring project operations and reporting eagle fatalities to the Service and state and
tribal wildlife agencies.

Compliance with the ECPG is voluntary, but the Service believes that following the guidance will
help project operators in complying with regulatory requirements and avoiding the unintentional
“take” of eagles at wind energy facilities, and will also assist the wind energy industry in providing
the biological data needed to support permit applications for facilities that may pose a risk to
eagles.

2. The Bald and Golden Eagle Protection Act
The Bald and Golden Eagle Protection Act (BGEPA) is the primary law protecting eagles. BGEPA
prohibits “take” of eagles without a permit (16 USC 668‐668c). BGEPA defines “take” to include
“pursue, shoot at, poison, wound, kill, capture, trap, collect, molest or disturb,” and prohibits take of
individuals and their parts, nests, or eggs. The Service expanded this definition by regulation to
include the term “destroy” to ensure that “take” includes destruction of eagle nests. The term
“disturb” is further defined by regulation as “to agitate or bother a bald or golden eagle to a degree
that causes, or is likely to cause,….injury to an eagle, a decrease in productivity, or nest
abandonment” (50 CFR 22.3).

3. Risks to Eagles from Wind Energy Facilities
Wind energy development can affect eagles in a variety of ways. First, eagles can be killed by
colliding with structures such as wind turbines. This is the primary threat to eagles from wind
facilities, and the ECPG guidance is primarily aimed at this threat. Second, disturbance from pre‐
construction, construction, or operation and maintenance activities might disturb eagles at
concentration sites or and result in loss of productivity at nearby nests. Third, serious disturbance
or mortality effects could result in the permanent or long term loss of a nesting territory.
Additionally, disturbances near important eagle use areas or migration concentration sites might
stress eagles so much that they suffer reproductive failure or mortality elsewhere, to a degree that

iii
could amount to prohibited take. All of these impacts, unless properly permitted, are violations of
BGEPA.

4. Eagle Take Permits
The Service recognizes that wind energy facilities, even those developed and operated with the
utmost effort to conserve wildlife, may under some circumstances result in the “take” of eagles
under BGEPA. However, in 2009, the Service promulgated new permit rules for eagles that address
this issue (50 CFR 22.26 and 22.27).

Under these new rules the Service can issue permits that authorize individual instances of take of
bald and golden eagles when the take is associated with, but not the purpose of, an otherwise lawful
activity, and cannot practicably be avoided. The regulations also authorize permits for
“programmatic” take, which means that instances of “take” may not be isolated, but may recur. The
programmatic take permits are the most germane permits for wind energy facilities. However,
under these regulations, any ongoing or programmatic take must be unavoidable even after the
implementation of advanced conservation practices (ACPs).

The ECPG is written to guide wind‐facility projects starting from the earliest conceptual planning
phase. For projects already in the development or operational phase, implementation of all stages
of the recommended approach in the ECPG may not be applicable or possible. Project developers or
operators with operating or soon‐to‐be operating facilities and who are interested in obtaining a
programmatic eagle take permit should contact the Service. The Service will work with project
developers or operators to determine if the project might be able to meet the permit requirements
in 50 CFR 22.26. The Service may recommend that the developer monitor eagle fatalities and
disturbance, adopt reasonable measures to reduce eagle fatalities from historic levels, and
implement compensatory mitigation. Sections of the ECPG that address these topics are relevant to
both planned and operating wind facilities (Appendices E and F in particular). Operators of wind
projects (and other activities) that were in operation prior to 2009 that pose a risk to golden eagles
may qualify for programmatic eagle take permits that do not automatically require compensatory
mitigation. This is because the requirements for obtaining programmatic take authorization are
designed to reduce take from historic, baseline levels, and the preamble to the Eagle Permit Rule
specified that unavoidable take remaining after implementation of avoidance and minimization
measures at such projects would not be subtracted from regional eagle take thresholds.

5. Voluntary Nature of the ECPG
Wind project operators are not legally required to seek or obtain an eagle take permit. However,
the take of an eagle without a permit is a violation of BGEPA, and could result in prosecution. The
methods and approaches suggested in the ECPG are not mandatory to obtain an eagle take permit.
The Service will accept other approaches that provide the information and data required by the
regulations. The ECP can be a stand‐alone document, or part of a larger bird and bat strategy as
described in the WEG, so long as it adequately meets the regulatory requirements at 50 CFR 22.26
to support a permit decision. However, Service employees who process eagle take permit
applications are trained in the methods and approaches covered in the ECPG. Using other
methodologies may result in longer application processing times.

6. Eagle Take Thresholds
Eagle take permits may be issued only in compliance with the conservation standards of BGEPA.
This means that the take must be compatible with the preservation of each species, defined (in
USFWS 2009a) as “consistent with the goal of stable or increasing breeding populations.”

iv

To ensure that any authorized “take” of eagles does not exceed this standard, the Service has set
regional take thresholds for each species, using methodology contained in the National
Environmental Policy Act (NEPA) Final Environmental Assessment (FEA) developed for the new
eagle permit rules (USFWS 2009b). The Service looked at regional populations of eagles and set
take thresholds for each species (upper limits on the number of eagle mortalities that can be
allowed under permit each year in these regional management areas).

The analysis identified take thresholds greater than zero for bald eagles in most regional
management areas. However, the Service determined that golden eagle populations might not be
able to sustain any additional unmitigated mortality at that time, and set the thresholds for this
species at zero for all regional populations. This means that any new authorized “take” of golden
eagles must be at least equally offset by compensatory mitigation (specific conservation actions to
replace or offset project‐induced losses).

The Service also put in place measures to ensure that local eagle populations are not depleted by
take that would be otherwise regionally acceptable. The Service specified that take rates must be
carefully assessed, both for individual projects and for the cumulative effects of other activities
causing take, at the scale of the local‐area eagle population (a population within a distance of 43
miles for bald eagles and 140 miles for golden eagles). This distance is based on the median
distance to which eagles disperse from the nest where they are hatched to where they settle to
breed.

The Service identified take rates of between 1 and 5 percent of the total estimated local‐area eagle
population as significant, with 5 percent being at the upper end of what might be appropriate
under the BGEPA preservation standard, whether offset by compensatory mitigation or not.
Appendix F provides a full description of take thresholds and benchmarks, and provides suggested
tools for evaluating how these apply to individual projects.

7. An Approach for Developing and Evaluating Eagle ACPs
Permits for eagle take at wind‐energy facilities are programmatic in nature as they will authorize
recurring take rather than isolated incidences of take. For programmatic take permits, the
regulations require that any authorized take must be unavoidable after the implementation of
advanced conservation practices (ACPs). ACPs are defined as “scientifically supportable measures
that are approved by the Service and represent the best available techniques to reduce eagle
disturbance and ongoing mortalities to a level where remaining take is unavoidable” (50 CFR 22.3).

Because the best information currently available indicates there are no conservation measures that
have been scientifically shown to reduce eagle disturbance and blade‐strike mortality at wind
projects, the Service has not currently approved any ACPs for wind energy projects.

The process of developing ACPs for wind energy facilities has been hampered by the lack of
standardized scientific study of potential ACPs. The Service has determined that the best way to
obtain the needed scientific information is to work with industry to develop ACPs for wind projects
as part of an adaptive‐management regime and comprehensive research program tied to the
programmatic‐take‐permit process. In this scenario, ACPs will be implemented at operating wind
facilities with an eagle take permit on an “experimental” basis (the ACPs are considered
experimental because they would not currently meet the definition of an ACP in the eagle permit
regulation). The experimental ACPs would be scientifically evaluated for their effectiveness, as
described in detail in this document, and based on the results of these studies, could be modified in

v
an adaptive management regime. This approach will provide the needed scientific information for
the future establishment of formal ACPs, while enabling wind energy facilities to move forward in
the interim.

Despite the current lack of formally approved ACPs, there may be other conservation measures
based on the best available scientific information that should be applied as a condition on
programmatic eagle take permits for wind‐energy facilities. A project developer or operator will be
expected to implement any reasonable avoidance and minimization measures that may reduce take
of eagles at a project. In addition, the Service and the project developer or operator will identify
other site‐specific and possibly turbine‐specific factors that may pose risks to eagles, and agree on
the experimental ACPs to avoid and minimize those risks. Unless the Service determines that there
is a reasonable scientific basis to implement the experimental ACPs up front (or it is otherwise
advantageous to the developer to do so), we recommend that such measures be deferred until such
time as there is eagle take at the facility or the Service determines that the circumstances and
evidence surrounding the take or risk of take suggest the experimental ACPs might be warranted.
The programmatic eagle take permit would specify the experimental ACPs, if circumstances
warrant, and the permit would be conditioned on the project operator’s agreement to implement
and monitor the experimental ACPs.

Because the ACPs would be experimental, the Service recommends that they be subject to a cost cap
that the Service and the project developer or operator would establish as part of the initial
agreement before issuance of an eagle permit. This would provide financial certainty as to what
maximum costs of such measures might be. The amount of the cap should be proportional to
overall risk.

As the results from monitoring experimental ACPs across a number of facilities accumulate and are
analyzed, scientific information in support of certain experimental ACPs may accrue, whereas other
ACPs may show little value in reducing take. If the Service determines that the available science
demonstrates an experimental ACP is effective in reducing eagle take, the Service will formally
approve that ACP and require its implementation up front on new projects when and where
warranted.

As the ECPG evolves, the Service will not expect project developers or operators to retroactively
redo analyses or surveys using the new approaches. The adaptive approach to the ECPG should not
deter project developers or operators from using the ECPG immediately.

8. Mitigation Actions to Reduce Effects on Eagle Populations
Where wind energy facilities cannot avoid taking eagles and eagle populations are not healthy
enough to sustain additional mortality, applicants must reduce the unavoidable mortality to a no‐
net‐loss standard for the duration of the permitted activity. No‐net‐loss means that these actions
either reduce another ongoing form of mortality to a level equal to or greater than the unavoidable
mortality, or lead to an increase in carrying capacity that allows the eagle population to grow by an
equal or greater amount. Actions to reduce eagle mortality or increase carrying capacity to this no‐
net‐loss standard are known as “compensatory mitigation” in the ECPG. Examples of compensatory
mitigation activities might include retrofitting power lines to reduce eagle electrocutions, removing
road‐killed animals along roads where vehicles hit and kill scavenging eagles, or increasing prey
availability.

The Service and the project developer or operator seeking a programmatic eagle take permit
should agree on the number of eagle fatalities to mitigate and what actions will be taken if actual

vi
eagle fatalities differ from the predicted number. The compensatory mitigation requirement and
trigger for adjustment should be specified in the permit. If the procedures recommended in the
ECPG are followed, there should not be a need for additional compensatory mitigation. However, if
other, less risk‐averse models are used to estimate fatalities, underestimates might be expected and
the permit should specify the threshold(s) of take that would trigger additional actions and the
specific mitigation activities that might be implemented.

Additional types of mitigation such as preserving habitat – actions that would not by themselves
lead to increased numbers of eagles but would assist eagle conservation – may also be advised to
offset other detrimental effects of permits on eagles. Compensatory mitigation is further discussed
below (Stage 4 – Avoidance and Minimization of Risk and Compensatory Mitigation).

9. Relationship of Eagle Guidelines (ECPG) to the Wind Energy Guidelines (WEG)
The ECPG is intended to be implemented in conjunction with other actions recommended in the
WEG that assess impacts to wildlife species and their habitats. The WEG recommends a five‐tier
process for such assessments, and the ECPG fits within that framework. The ECPG focuses on just
eagles to facilitate collection of information that could support an eagle take permit decision. The
ECPG uses a five‐stage approach like the WEG; the relationship between the ECPG stages and the
WEG tiers is shown in Fig. 1.

Tiers 1 and 2 of the WEG (Stage 1 of the ECPG) could provide sufficient evidence to demonstrate
that a project poses very low risk to eagles. Provided this assessment is robust, eagles may not
warrant further consideration in subsequent WEG tiers, and Stages 2 through 5 of the ECPG and
pursuit of an eagle take permit might be unnecessary. A similar conclusion could be reached at the
end of Stage 2, 3, or 4. In such cases, if unpermitted eagle take subsequently occurs, the wind
project proponent should consult with the U.S. Fish and Wildlife Service to determine how to
proceed, possibly by obtaining an eagle take permit.

The following sections describe the general approach envisioned for assessing wind project impacts
to eagles (also see the Stage Overview Table at the end of the Executive Summary).

Tiers 1 and 2 of the WEG, Stage 1 of the ECPG
Tier 1 of the WEG is the preliminary site evaluation (landscape‐scale screening of possible
project sites). Tier 2 is site characterization (broad characterization of one or more
potential project sites). These correspond with Stage 1 of the ECPG, the site‐assessment
stage. As part of the Tiers 1 and 2 process, project developers should carry out Stage 1 of
the ECPG and evaluate broad geographic areas to assess the relative importance of various
areas to resident breeding and non‐breeding eagles, and to migrant and wintering eagles.
During Stage 1, the project developer or operator should gather existing information from
publicly available literature, databases, and other sources, and use those data to judge the
appropriateness of various potential project sites, balancing suitability for development
with potential risk to eagles.

To increase the probability of meeting the regulatory requirements for a programmatic take
permit, biological advice from the Service and other jurisdictional wildlife agencies should
be requested as early as possible in the developer's planning process and should be as
inclusive as possible to ensure all issues are being addressed at the same time and in a
coordinated manner. Ideally, consultation with the Service, and state and tribal wildlife

vii
agencies is done before wind developers make any substantial financial commitment or
finalize lease agreements.

Tier 3 of the WEG, Stages 2, 3, and 4 of the ECPG
During Tier 3 of the WEG, a developer conducts field studies to document wildlife use and
habitat at the project site and predict project impacts. These site‐specific studies are critical
to evaluating potential impacts to all wildlife including eagles. The developer and the
Service would use the information collected to support an eagle take permit application,
should the developer seek a permit. As part of Tier 3, the ECPG recommends project
developers or operators implement three stages of assessment:
 Stage 2 ‐ site‐specific surveys and assessments;
 Stage 3 ‐ predicting eagle fatalities; and
 Stage 4 ‐ avoidance and minimization of risk and compensatory mitigation.

Stage 2 – Site Specific Surveys and Assessments
During Stage 2 the Service recommends the project developer collect quantitative
data through scientifically rigorous surveys designed to assess the potential risk of
the proposed project to eagles. The Service recommends collecting information that
will allow estimation of the eagle exposure rate (eagle‐minutes flying within the
project footprint per hour per kilometer2), as well as surveys sufficient to determine
if important eagle use areas or migration concentration sites are within or in close
proximity to the project footprint (see Appendix C). In the case of small wind
projects (one utility‐scale turbine or a few small turbines), the project developer
should consider the proximity of eagle nesting and roosting sites to a proposed
project and discuss the results of the Stage 1 assessment with the Service to
determine if Stage 2 surveys are necessary. In many cases the hazardous area
associated with such projects will be small enough that Stage 2 surveys will not be
necessary.

Stage 3 – Predicting Eagle Fatalities
In Stage 3, the Service and project developers or operators use data from Stage 2 in
models to predict eagle risk expressed as the average number of fatalities per year
extrapolated to the tenure of the permit. These models can compare alternative
siting, construction, and operational scenarios, a useful feature in constructing
hypotheses regarding predicted effects of conservation measures and experimental
ACPs. The Service encourages project developers or operators to use the
recommended pre‐construction survey protocol in this ECPG in Stage 2 to help
inform our predictive models in Stage 3. If Service‐recommended survey protocols
are used, this risk assessment can be greatly facilitated using model tools available
from the Service. If project developers or operators use other forms of information
for the Stage 2 assessment, they will need to fully describe those methods and the
analysis used for the eagle risk assessment. The Service will require more time to
evaluate and review the data because, for example, the Service will need to compare
the results of the project developer or operator’s eagle risk assessment with
predictions from our models. If the results differ, we will work with the project
developers or operators to determine which model results are most appropriate for
the Service’s eventual permitting decisions.

viii
The Service and project developers or operators also evaluate Stage 2 data to
determine whether disturbance take is likely, and if so, at what level. Any loss of
production that may stem from disturbance should be added to the fatality rate
prediction for the project. The risk assessments at Stage 2 and Stage 3 are
consistent with developing the information necessary to assess the efficacy of
conservation measures, and to develop the monitoring required by the permit
regulations at 50 CFR 22.26(c)(2).

Stage 4 - Avoidance and Minimization of Risk and Compensatory Mitigation
In Stage 4 the information gathered should be used by the project developer or
operator and the Service to determine potential conservation measures and ACPs (if
available) to avoid or minimize predicted risks at a given site (see Appendix E). The
Service will compare the initial predictions of eagle mortality and disturbance for
the project with predictions that take into account proposed and potential
conservation measures and ACPs, once developed and approved, to determine if the
project developer or operator has avoided and minimized risks to the maximum
degree achievable, thereby meeting the requirements for programmatic permits
that remaining take is unavoidable. Additionally, the Service will use the
information provided along with other data to conduct a cumulative effects analysis
to determine if the project’s impacts, in combination with other permitted take and
other known factors, are at a level that exceed the established thresholds or
benchmarks for eagle take at the regional and local‐area scales. This final eagle risk
assessment is completed at the end of Stage 4 after application of conservation
measures and ACPs (if available) along with a plan for compensatory mitigation if
required.

The eagle permit process requires compensatory mitigation if conservation
measures do not remove the potential for take, and the projected take exceeds
calculated thresholds for the eagle management unit in which the project is located.
However, there may also be other situations in which compensatory mitigation is
necessary. The following guidance applies to those situations as well.

Compensatory mitigation can address pre‐existing causes of eagle mortality (such as
eagle electrocutions from power poles) or it can address increasing the carrying
capacity of the eagle population in the affected eagle management unit. However,
there needs to be a credible analysis that supports the conclusion that implementing
the compensatory mitigation action will achieve the desired beneficial offset in
mortality or carrying capacity.

For new wind development projects, if compensatory mitigation is necessary, the
compensatory mitigation action (or a verifiable, legal commitment to such
mitigation) will be required up front before project operations begin because
projects must meet the statutory eagle preservation standard before the Service
may issue a permit. For operating projects, compensatory mitigation should be
applied from the start of the permit period, not retroactively from the time the
project began. The initial compensatory mitigation effort should be sufficient to
offset the predicted number of eagle fatalities per year for five years. No later than
at the end of the five year period, the Service and the project operator will compare
the predicted annual take estimate to the realized take based on post‐construction
monitoring. If the triggers identified in the permit for adjustment of compensatory

ix
mitigation are met, those adjustments should be implemented. In the case where the
observed take was less than estimated, the permittee will receive a credit for the
excess compensation (the difference between the actual mean and the number
compensated for) that can be applied to other take (either by the permittee or other
permitted individuals at his/her discretion) within the same eagle management
unit. The Service, in consultation with the permittee, will determine compensatory
mitigation for future years for the project at this point, taking into account the
observed levels of mortality and any reduction in that mortality that is expected
based on implementation of additional experimental conservation measures and
ACPs. Monitoring using the best scientific and practicable methods available should
be included to determine the effectiveness of the resulting compensatory mitigation
efforts. The Service will modify the compensatory mitigation process to adapt to
any improvements in our knowledge base as new data become available.

At the end of Stage 4, all the materials necessary to satisfy the regulatory
requirements to support a permit application should be available. While a project
operator can submit a permit application at any time, the Service can only begin the
formal process to determine whether a programmatic eagle take permit can be
issued after completion of Stage 4. Ideally, National Environmental Policy Act
(NEPA) and National Historic Preservation Act (NHPA ) analyses and assessments
will already be underway, but if not, Stage 4 should include necessary NEPA
analysis, NHPA compliance, coordination with other jurisdictional agencies, and
tribal consultation.

Tier 4 and 5 of the WEG, Stage 5 of the ECPG
If the Service issues an eagle take permit and the project goes forward, project operators
will conduct post‐construction surveys to collect data that can be compared with the pre‐
construction risk‐assessment predictions for eagle fatalities and disturbance. The
monitoring protocol should include validated techniques for assessing both mortality and
disturbance effects, and they must meet the permit‐condition requirements at 50 CFR
22.26(c)(2). In most cases, intensive monitoring will be conducted for at least the first two
years after permit issuance, followed by less intense monitoring for up to three years after
the expiration date of the permit. Project developers or operators should use the post‐
construction survey protocols included or referenced in this ECPG, but we will consider
other monitoring protocols provided by permit applicants though the process will likely
take longer than if familiar approaches were used. The Service will use the information
from post‐construction monitoring in a meta‐analysis framework to weight and improve
pre‐construction predictive models.

Additionally in Stage 5, the Service and project developers or operators should use the post‐
construction monitoring data to (1) assess whether compensatory mitigation is adequate,
excessive, or deficient to offset observed mortality, and make adjustments accordingly; and
(2) explore operational changes that might be warranted at a project after permitting to
reduce observed mortality and meet permit requirements.

10. Site Categorization Based on Mortality Risk to Eagles
Beginning at the end of Stage 1, and continuing at the end of Stages 2, 3, and 4, we recommend the
approach outlined below be used to assess the likelihood that a wind project will take eagles, and if

x
so, that the project will meet standards in 50 CFR 22.26 for issuance of a programmatic eagle take
permit.

Category 1 – High risk to eagles, potential to avoid or mitigate impacts is low
A project is in this category if it:
(1) has an important eagle‐use area or migration concentration site within the project
footprint; or
(2) has an annual eagle fatality estimate (average number of eagles predicted to be
taken annually) > 5% of the estimated local‐area population size; or
(3) causes the cumulative annual take for the local‐area population to exceed 5% of the
estimated local‐area population size.

In addition, projects that have eagle nests within ½ the mean project‐area inter‐nest
distance of the project footprint should be carefully evaluated. If it is likely eagles
occupying these territories use or pass through the project footprint, category 1 designation
may be appropriate.

Projects or alternatives in category 1 should be substantially redesigned to at least meet the
category 2 criteria. The Service recommends that project developers not build projects at
sites in category 1 because the project would likely not meet the regulatory requirements.
The recommended approach for assessing the percentage of the local‐area population
predicted to be taken is described in Appendix F.

Category 2 – High or moderate risk to eagles, opportunity to mitigate impacts
A project is in this category if it:
(1) has an important eagle‐use area or migration concentration site within the project
area but not in the project footprint; or
(2) has an annual eagle fatality estimate between 0.03 eagles per year and 5% of the
estimated local‐area population size; or
(3) causes cumulative annual take of the local‐area population of less than 5% of the
estimated local‐area population size.

Projects in this category will potentially take eagles at a rate greater than is consistent with
maintaining stable or increasing populations, but the risk might be reduced to an acceptable
level through a combination of conservation measures and reasonable compensatory
mitigation. These projects have a risk of ongoing take of eagles, but this risk can be
minimized. For projects in this category the project developer or operator should prepare
an Eagle Conservation Plan (ECP) or similar plan to document meeting the regulatory
requirements for a programmatic permit. The ECP or similar document can be a stand‐
alone document, or part of a larger bird and bat strategy as described in the WEG, so long as
it adequately meets the regulatory requirements at 50 CFR 22.26 to support a permit
decision. For eagle management populations where take thresholds are set at zero, the
conservation measures in the ECP should include compensatory mitigation and must result
in no‐net‐loss to the breeding population to be compatible with the permit regulations. This
does not apply to golden eagles east of the 100th meridian, for which no non‐emergency
take can presently be authorized (USFWS 2009b).

Category 3 – Minimal risk to eagles
A project is in this category if it:

xi
(1) has no important eagle use areas or migration concentration sites within the project
area; and
(2) has an annual eagle fatality rate estimate of less than 0.03; and
(3) causes cumulative annual take of the local‐area population of less than 5% of the
estimated local‐area population size.

Projects in category 3 pose little risk to eagles and may not require or warrant eagle take
permits, but that decision should be made in coordination with the Service. Still, a project
developer or operator may wish to create an ECP or similar document or strategy that
documents the project’s low risk to eagles, and outlines mortality monitoring for eagles and
a plan of action if eagles are taken during project construction or operation. This would
enable the Service to provide a permit to allow a de minimis amount of take if the project
developer or operator wished to obtain such a permit.

The risk category of a project can potentially change as a result of additional site‐specific analyses
and application of measures to reduce the risk. For example, a project may appear to be in category
2 as a result of Stage 1 analyses, but after collection of site‐specific information in Stage 2 it might
become clear it is a category 1 project. If a project cannot practically be placed in one of these
categories, the project developer or operator and the Service should work together to determine if
the project can meet programmatic eagle take permitting requirements in 50 CFR 22.26 and 22.27.
Projects should be placed in the highest category (with category 1 being the highest) in which one
or more of the criteria are met.

11. Addressing Uncertainty

There is substantial uncertainty surrounding the risk of wind projects to eagles, and of ways to
minimize that risk. For this reason, the Service stresses that it is very important not to
underestimate eagle fatality rates at wind facilities. Overestimates, once confirmed, can be adjusted
downward based on post‐construction monitoring information with no consequence to eagle
populations. Project developers or operators can trade or be credited for excess compensatory
mitigation, and debits to regional and local‐area eagle‐take thresholds and benchmarks can be
adjusted downwards to reflect actual fatality rates. However, the options for addressing
underestimated fatality rates are extremely limited, and pose either potential hardships for wind
developers or significant risks to eagle populations.

Our long‐term approach for moving forward in the face of this uncertainty is to implement eagle
take permitting in a formal adaptive management framework. The Service anticipates four specific
sets of adaptive management decisions: (1) adaptive management of wind project siting and design
recommendations; (2) adaptive management of wind project operations; (3) adaptive management
of compensatory mitigation; and (4) adaptive management of population‐level take thresholds.
These are discussed in more detail in Appendix A. The adaptive management process will depend
heavily on pre‐ and post‐construction data from individual projects, but analyses, assessment, and
model evaluation will rely on data pooled over many individual wind projects. Learning
accomplished through adaptive management will be rapidly incorporated into the permitting
process so that the regulatory process adjusts in proportion to actual risk.

12. Interaction with the Service
The Service encourages early, frequent and thorough coordination between project developers or
operators and Service and other jurisdictional‐agency employees as they implement the tiers of the
WEG, and the related Stages of the ECPG. Close coordination will aid the refinement of the

xii
modeling process used to predict fatalities, as well as the post‐construction monitoring to evaluate
those models. We anticipate the ECPG and the recommended methods and metrics will evolve as
the Service and project developers or operators learn together. The Service has created a cross‐
program, cross‐regional team of biologists who will work jointly on eagle‐programmatic‐take
permit applications to help ensure consistency in administration and application of the Eagle
Permit Rule. This close coordination and interaction is especially important as the Service
processes the first few programmatic eagle take permit applications.

The Service will continue to refine this ECPG with input from all stakeholders with the objective of
maintaining stable or increasing breeding populations of both bald and golden eagles while
simultaneously developing science‐based eagle‐take regulations and procedures that are
appropriate to the risk associated with each wind energy project.

Stage Overview Table - Overview of staged approach to developing an Eagle Conservation Plan as
described in the ECPG. Stages are in chronological order. Stage 5 would only be applicable in cases where a
permit was issued at the end of Stage 4.

Stage

Objective

Actions

Data Sources

1

At the landscape level, identify
Broad, landscape‐scale
potential wind facility locations
evaluation.
with manageable risk to eagles.

Technical literature, agency files,
on‐line biological databases, data
from nearby projects, industry
reports, geodatabases, experts.

2

Obtain site‐specific data to
predict eagle fatality rates and
disturbance take at wind‐facility
sites that pass Stage 1
assessment. Investigate other
aspects of eagle use to consider
assessing distribution of
occupied nests in the project
area, migration, areas of
seasonal concentration, and
intensity of use across the
project footprint.

Site‐specific surveys and
intensive observation to
determine eagle exposure rate
and distribution of use in the
project footprint, plus locations
of occupied eagle nests,
migration corridors and
stopover sites, foraging
concentration areas, and
communal roosts in the project
area.

Project footprint: 800‐m radius
point count surveys and
utilization distribution studies.
Project area: nest surveys,
migration counts at likely
topographic features,
investigation of use of potential
roost sites and of areas of high
prey availability. Ideally
conducted for no less than 2
years pre‐construction.

3

As part of pre‐construction
monitoring and assessment,
estimate the fatality rate of
eagles for the facility evaluated
in Stage 2, excluding possible
additions of conservation
measures and advanced
conservation practices (ACPs).
Consider possible disturbance
effects.

Use the exposure rate derived
from Stage 2 data in Service‐
provided models to predict the Point count, nest, and eagle
annual eagle fatality rate for the concentration area data from
Stage 2.
project. Determine if
disturbance effects are likely and
what they might be.

xiii
Stage

4

Objective

Data Sources

Re‐run fatality prediction models
with risk adjusted to reflect
As part of the pre‐construction
application of conservation
assessment, identify and
measures and ACPs to determine
Fatality estimates before and
evaluate conservation measures
fatality estimate (80% upper
after application of conservation
and ACPs that might avoid or
confidence limit or equivalent).
measures and ACPs, using point
minimize fatalities and
Calculate required
count data from Stage 2.
disturbance effects identified in
compensatory mitigation
Estimates of disturbance effects
Stage 3. When necessary,
amount where necessary,
from Stage 3.
identify compensatory
considering disturbance effects,
mitigation to reduce predicted
if any. Identify actions needed to
take to a no‐net‐loss standard.
accomplish compensatory
mitigation.

Determine if regulatory
Permit
requirements for issuance of a
Decision
permit have been met.

5

Actions

Data from Stages 1, 2, 3 and 4;
results of NEPA analysis; and
The Service will issue or deny
considering information
the permit request based on an
obtained during tribal
evaluation of the ECP or other
consultation and through
form of application.
coordination with the states and
other jurisdictional agencies.

During post‐construction
monitoring, document mean
Conduct fatality monitoring in
annual eagle fatality rate and
project footprint. Monitor
effects of disturbance.
Determine if initial conservation activity of eagles that may be
disturbed at nest sites,
measures are working and
communal roosts, and/or major
should be continued, and if
foraging sites. Ideally, monitor
additional conservation
measures might reduce observed eagle use of project footprint via
fatalities. Monitor effectiveness point counts, migration counts,
and/or intensive observation of
of compensatory mitigation.
use distribution.
Ideally, assess use of area by
eagles for comparison to pre‐
construction levels.

Post‐construction survey
database for fatality monitoring,
Comparable pre‐ and post‐
construction data for selected
aspect of eagle use of the project
footprint and adjoining areas.

All post‐construction surveys
should be conducted for at least
2 years, and targeted thereafter
to assess effectiveness of any
experimental conservation
measures or ACPs.

1

Table of Contents

Disclaimer .............................................................................................................................................................. i
EXECUTIVE SUMMARY ........................................................................................................................................ ii
1. Overview ..................................................................................................................................................... ii
2. The Bald and Golden Eagle Protection Act ................................................................................................ ii
3. Risks to Eagles from Wind Energy Facilities .............................................................................................. ii
4. Eagle Take Permits .................................................................................................................................... iii
5. Voluntary Nature of the ECPG ................................................................................................................... iii
6. Eagle Take Thresholds ............................................................................................................................... iii
7. An Approach for Developing and Evaluating Eagle ACPs......................................................................... iv
8. Mitigation Actions to Reduce Effects on Eagle Populations ..................................................................... v
9. Relationship of Eagle Guidelines (ECPG) to the Wind Energy Guidelines (WEG) .................................... vi
Tiers 1 and 2 of the WEG, Stage 1 of the ECPG ........................................................................................ vi
Tier 3 of the WEG, Stages 2, 3, and 4 of the ECPG .................................................................................. vii
Tier 4 and 5 of the WEG, Stage 5 of the ECPG .......................................................................................... ix
10. Site Categorization Based on Mortality Risk to Eagles .......................................................................... ix
Category 1 – High risk to eagles, potential to avoid or mitigate impacts is low ...................................... x
Category 2 – High or moderate risk to eagles, opportunity to mitigate impacts ...................................... x
Category 3 – Minimal risk to eagles........................................................................................................... x
11. Addressing Uncertainty ........................................................................................................................... xi
12. Interaction with the Service .................................................................................................................... xi
INTRODUCTION AND PURPOSE .......................................................................................................................... 4
1. Purpose........................................................................................................................................................ 4
2. Legal Authorities and Relationship to Other Statutes and Guidelines ..................................................... 6
3. Background and Overview of Process ........................................................................................................ 8
a. Risks to Eagles ........................................................................................................................................ 9
b. General Approach to Address Risk ......................................................................................................... 9
ASSESSING RISK AND EFFECTS ....................................................................................................................... 12
1. Considerations When Assessing Eagle Use Risk .................................................................................... 12
a. General Background and Rationale for Assessing Project Effects on Eagles ..................................... 12
b. Additional Considerations for Assessing Project Effects: Migration Corridors and Stopover Sites... 14
2. Eagle Risk Factors ..................................................................................................................................... 15
3. Overview of Process to Assess Risk......................................................................................................... 16
4. Site Categorization Based on Mortality Risk to Eagles ........................................................................... 25
a. Category 1 – High risk to eagles, potential to avoid or mitigate impacts is low ................................ 25
b. Category 2 – High or moderate risk to eagles, opportunity to mitigate impacts ................................ 25
c. Category 3 – Minimal risk to eagles ..................................................................................................... 26
5. Cumulative Effects Considerations .......................................................................................................... 26
a. Early Planning ........................................................................................................................................ 26
b. Analysis Associated with Permits ........................................................................................................ 27

2

ADAPTIVE MANAGEMENT ................................................................................................................................ 28
EAGLE CONSERVATION PLAN DEVELOPMENT PROCESS ................................................................................ 29
1. Contents of the Eagle Conservation Plan ................................................................................................. 30
a. Stage 1 .................................................................................................................................................. 31
b. Stage 2 .................................................................................................................................................. 31
c. Stage 3 ................................................................................................................................................... 31
d. Stage 4 .................................................................................................................................................. 31
e. Stage 5 – Post-construction Monitoring .............................................................................................. 31
INTERACTION WITH THE SERVICE .................................................................................................................... 32
INFORMATION COLLECTION.............................................................................................................................. 33
GLOSSARY .......................................................................................................................................................... 34
LITERATURE CITED ............................................................................................................................................. 40
APPENDIX A: ADAPTIVE MANAGEMENT ......................................................................................................... 44
1. Adaptive Management as a Tool ............................................................................................................. 45
2. Applying Adaptive Management to Eagle Take Permitting .................................................................... 46
a. Adaptive Management of Wind Project Operations ............................................................................ 46
b. Adaptive Management of Wind Project Siting and Design Recommendations .................................. 47
c. Adaptive Management of Compensatory Mitigation ........................................................................... 47
d. Adaptive Management of Population-Level Take Thresholds ............................................................. 47
Literature Cited .............................................................................................................................................. 48
APPENDIX B: STAGE 1 – SITE ASSESSMENT ................................................................................................... 50
Literature Cited .............................................................................................................................................. 52
APPENDIX C: STAGE 2 – SITE-SPECIFIC SURVEYS AND ASSESSMENT ......................................................... 53
1. Surveys of Eagle Use ................................................................................................................................ 53
a. Point Count Surveys .............................................................................................................................. 53
b. Migration Counts and Concentration Surveys...................................................................................... 60
c. Utilization Distribution (UD) Assessment ............................................................................................. 62
d. Summary ................................................................................................................................................ 63
2. Survey of the Project-area Nesting Population: Number and Locations of Occupied Nests of Eagles.. 64
Literature Cited .............................................................................................................................................. 66
APPENDIX D: STAGE 3 – PREDICTING EAGLE FATALITIES............................................................................... 68
1. Exposure .................................................................................................................................................... 69
2. Collision Probability .................................................................................................................................. 71
3. Expansion .................................................................................................................................................. 72

3

4. Fatalities.................................................................................................................................................... 72
5. Putting it all together: an example ........................................................................................................... 72
a. Patuxent Power Company Example ...................................................................................................... 73
b. Exposure ................................................................................................................................................ 74
b. Collision Probability............................................................................................................................... 75
c. Expansion ............................................................................................................................................... 75
d. Fatalities ................................................................................................................................................ 75
6. Additional Considerations ........................................................................................................................ 76
a. Small-scale projects .............................................................................................................................. 76
Literature Cited .............................................................................................................................................. 77
APPENDIX E: STAGE 4 – AVOIDANCE AND MINIMIZATION OF RISK USING ACPS AND OTHER
CONSERVATION MEASURES, AND COMPENSATORY MITIGATION............................................................... 78
Literature Cited .............................................................................................................................................. 79
APPENDIX F: ASSESSING PROJECT-LEVEL TAKE AND CUMULATIVE EFFECTS ANALYSES .......................... 80
Literature Cited .............................................................................................................................................. 85
APPENDIX G: EXAMPLES USING RESOURCE EQUIVALENCY ANALYSIS TO ESTIMATE THE
COMPENSATORY MITIGATION FOR THE TAKE OF GOLDEN AND BALD EAGLES FROM WIND ENERGY
DEVELOPMENT ................................................................................................................................................... 86
1. Introduction ............................................................................................................................................... 86
2. REA Inputs ................................................................................................................................................. 86
3. REA Example – WindCoA ......................................................................................................................... 88
a. REA Language and Methods ................................................................................................................. 89
b. REA Results for WindCoA ..................................................................................................................... 91
c. Summary of Bald Eagle REA Results .................................................................................................... 92
d. Discussion on Using REA ...................................................................................................................... 93
e. Additional Compensatory Mitigation Example..................................................................................... 93
f. Take from Disturbance ........................................................................................................................... 93
Literature Cited .............................................................................................................................................. 94
APPENDIX H: STAGE 5 – CALIBRATING AND UPDATING OF THE FATALITY PREDICTION AND CONTINUED
RISK-ASSESSMENT ........................................................................................................................................... 96
1. Fatality Monitoring ................................................................................................................................... 96
2. Disturbance Monitoring ............................................................................................................................ 98
3. Comparison of Post-Construction Eagle Use with Pre-Construction Use................................................ 99
Literature Cited .............................................................................................................................................. 99

4

INTRODUCTION AND PURPOSE

The mission of the Service is working with others to conserve, protect and enhance fish, wildlife,
plants and their habitats for the continuing benefit of the American people. As part of this, we are
charged with implementing statutes including the BGEPA, MBTA (Migratory Bird Treaty Act), and
ESA (Endangered Species Act). BGEPA prohibits all take of eagles unless otherwise authorized by
the Service. A goal of BGEPA is to ensure that any authorized take of bald and golden eagles is
compatible with their preservation, which the Service has interpreted to mean allowing take that is
consistent with the goal of stable or increasing breeding populations. In 2009, the Service
promulgated regulations authorizing issuance of permits for non‐purposeful take of eagles; the
ECPG is intended to promote compliance with BGEPA with respect to such permits by providing
recommended procedures for:

(1) conducting early pre‐construction assessments to identify important eagle use areas;
(2) analyzing pre‐construction information to estimate potential impacts on eagles;
(3) avoiding, minimizing, and/or compensating for potential adverse effects to eagles; and
(4) monitoring for impacts to eagles during construction and operation.

The ECPG calls for scientifically rigorous surveys, monitoring, risk assessment, and research
designs proportionate to the risk to both bald and golden eagles. The ECPG describes a process by
which wind energy developers, operators, and their consultants can collect and analyze information
that could lead to a programmatic permit to authorize unintentional take of eagles at wind energy
facilities. The processes described here is not required, but project developers or operators should
coordinate closely with the Service if they plan to use an alternative approach to meet the
regulatory requirements for a permit.

1. Purpose
The Service published a final rule (Eagle Permit Rule) on September 11, 2009 under BGEPA (50
CFR 22.26) authorizing limited issuance of permits to take bald eagles (Haliaeetus leucocephalus)
and golden eagles (Aquila chrysaetos) ‘‘for the protection of ... other interests in any particular
locality’’ where the take is compatible with the preservation of the bald eagle and the golden eagle,
is associated with and not the purpose of an otherwise lawful activity, and cannot practicably be
avoided (USFWS 2009a). The ECPG explains the Service’s approach to issuing programmatic eagle
take permits for wind energy projects under this authority, and provides guidance to permit
applicants (project developers or operators), Service biologists, and biologists with other
jurisdictional agencies (state and tribal fish and wildlife agencies, in particular) on the development
of Eagle Conservation Plans (ECPs) to support permit issuance.

Since finalization of the Eagle Permit Rule, the development and planned development of wind
facilities (developments for the generation of electricity from wind turbines) have increased in the
range of the golden eagle in the western United States. Golden eagles are vulnerable to collisions
with wind turbines (Hunt 2002), and in some areas such collisions could be a major source of
mortality (Hunt et al. 1999, 2002; USFWS unpublished data). Although significant numbers of bald
eagle mortalities have not yet been reported at North American wind facilities, deaths have
occurred at more than one location (USFWS, unpublished data), and the closely related and
behaviorally similar white‐tailed eagle (Haliaeetus albicilla) has been killed regularly at wind
facilities in Europe (Krone 2003, Cole 2009, Nygård et al. 2010). Because of this risk to eagles,
many of the current and planned wind facilities require permits under the Eagle Permit Rule to be
in compliance with the law if and when an eagle is taken at that facility. In addition to being legally

5
necessary to comply with BGEPA and 50 CFR 22.26, the conservation practices necessary to meet
standards required for issuance of these permits should offset the short‐ and long‐term negative
effects of wind energy facilities on eagle populations. Because of the urgent need for guidance on
permitting eagle take at wind facilities, this initial module focuses on this issue. Many of the
concepts and approaches outlined in this module can be readily exported to other situations (e.g.,
solar facilities, electric power lines), and the Service expects to release other modules in the future
specifically addressing other sources of eagle take.

The ECPG is intended to provide interpretive guidance to Service biologists and others in applying
the regulatory permit standards as specified in the rule. They do not in‐and‐of themselves impose
additional regulatory or generally‐binding requirements. An ECP per se is not required, even to
obtain a programmatic eagle take permit. As long as the permit application is complete and
includes the information necessary to evaluate a permit application under 50 CFR 22.26 or 22.27,
the Service will review the application and make a determination if a permit will be issued.
However, Service personnel will be trained in the application of the procedures and approaches
outlined in the ECPG, and developers who choose to use other approaches should expect the review
time on the part of the Service to be longer. The Service recommends that the basic format for the
ECP be followed to allow for expeditious consideration of the application materials.

Preparation of an ECP and consultation with the Service are voluntary actions on the part of the
developer. There is no legal requirement that wind developers apply for or obtain an eagle take
permit, so long as the project does not result in take of eagles. However, take of an eagle without an
eagle take permit is a violation of BGEPA, so the developer or operator must weigh the risks in
his/her decision. The Service is available to consult with the developer or operator as he/she
makes that decision.

The ECPG is written to guide wind‐facility projects starting from the earliest conceptual planning
phase. For projects already in the development or operational phase, implementation of all stages
of the recommended approach in the ECPG may not be applicable or possible. Project developers or
operators with operating or soon‐to‐be operating facilities and who are interested in obtaining a
programmatic eagle take permit should contact the Service. The Service will work with project
developers or operators to determine if the project might be able to meet the permit requirements
in 50 CFR 22.26. The Service may recommend that the developer monitor eagle fatalities and
disturbance, adopt reasonable measures to reduce eagle fatalities from historic levels, and
implement compensatory mitigation. Sections of the ECPG that address these topics are relevant to
both planned and operating wind facilities (Appendices E and F in particular). Operators of wind
projects (and other activities) that were in operation prior to 2009 that pose a risk to golden eagles
may qualify for programmatic eagle take permits that do not automatically require compensatory
mitigation. This is because the requirements for obtaining programmatic take authorization are
designed to reduce take from historic, baseline levels, and the preamble to the Eagle Permit Rule
specified that unavoidable take remaining after implementation of avoidance and minimization
measures at such projects would not be subtracted from regional eagle take thresholds (U. S. Fish
and Wildlife Service 2009a).

The ECPG is designed to be compatible with the more general guidelines provided in the U.S. Fish
and Wildlife Service Land‐based Wind Energy Guidelines (WEG) http://www.fws.gov/
habitatconservation/windpower/wind_turbine_advisory_committee.html. However, because the
ECPG describes actions which help to comply with the regulatory requirements in BGEPA for an
eagle take permit as described in 50 CFR 22.26 and 22.27, they are more specific. The Service will
make every effort to ensure the work and timelines for both processes are as congruent as possible.

6

2. Legal Authorities and Relationship to Other Statutes and Guidelines
There are several laws that must be considered for compliance during eagle take permit application
review under the 50 CFR 22.26 and 22.27 regulations: BGEPA, MBTA, ESA, the National
Environmental Policy Act (NEPA) (42 U.S.C. 4321 et. seq.), and the National Historic Preservation
Act (NHPA) (16 U.S.C. 470 et seq.). BGEPA is the primary law protecting eagles. BGEPA defines
“take” to include “pursue, shoot, shoot at, poison, wound, kill, capture, trap, collect, molest or
disturb” and prohibits take of individuals, and their parts, nests, or eggs (16 USC 668 & 668c). The
Service expanded this definition by regulation to include the term “destroy” to ensure that “take”
includes destruction of eagle nests (50 CFR 22.3). The term “disturb” is defined by regulation at 50
CFR 22.3 as “to agitate or bother a bald or golden eagle to a degree that causes, or is likely to cause,
… injury to an eagle, a decrease in productivity, or nest abandonment…” (USFWS 2007). A goal of
BGEPA is to ensure that any authorized take is compatible with eagle preservation, which the
Service has interpreted to mean it can authorize take that is consistent with the goal of stable or
increasing breeding populations of bald and golden eagles (USFWS 2009b).

In 2009, two new permit rules were created for eagles. Under 50 CFR 22.26, the Service can issue
permits that authorize individual instances of take of bald and golden eagles when the take is
associated with, but not the purpose of an otherwise lawful activity, and cannot practicably be
avoided. The regulation also authorizes ongoing or programmatic take, but requires that any
authorized programmatic take be unavoidable after implementation of advanced conservation
practices. Under 50 CFR 22.27, the Service can issue permits that allow the intentional take of eagle
nests where necessary to alleviate a safety emergency to people or eagles, to ensure public health
and safety, where a nest prevents use of a human‐engineered structure, and to protect an interest
in a particular locality where the activity or mitigation for the activity will provide a net benefit to
eagles. Only inactive nests are allowed to be taken except in cases of safety emergencies.

The new Eagle Permit Rule provides a mechanism where the Service may legally authorize the non‐
purposeful take of eagles. However, BGEPA provides the Secretary of the Interior with the authority
to issue eagle take permits only when the take is compatible with the preservation of each species,
defined in USFWS (2009a) as “…consistent with the goal of stable or increasing breeding
populations.” The Service ensures that any take it authorizes under 50 CFR 22.26 does not exceed
this preservation standard by setting regional take thresholds for each species determined using
the methodology contained in the NEPA Final Environmental Assessment (FEA) developed for the
new permit rules (USFWS 2009b). The details and background of the process used to calculate
these take thresholds are presented in the FEA (USFWS 2009b). It is important to note that the
take thresholds for regional eagle management populations (eagle management units) and the
process by which they are determined are derived independent from this or any other ECPG
module.

Many states and tribes have regulations that protect eagles, and may require permits for purposeful
and non‐purposeful take. Project developers or operators should contact all pertinent state and
tribal fish and wildlife agencies at the earliest possible stage of project development to ensure
proper coordination and permitting. The Service will coordinate our programmatic take permits
with all such jurisdictional agencies.

Wind projects that are expected to cause take of endangered or threatened wildlife species should
still receive incidental take authorizations under sections 7 or 10 of ESA in order to ensure
compliance with Federal law. A project developer or operator seeking an Incidental Take Permit

7
(ITP) through the ESA section 10 Habitat Conservation Plan (HCP) process may be issued an ITP
only if the permitted activity is otherwise lawful (section 10(a)(1)(B)). If the project and covered
activities in the HCP are likely to take bald or golden eagles, the project proponent should obtain a
BGEPA permit or include the bald or golden eagle as a covered species in the HCP in order for the
activity to be lawful in the event that eagles are taken. When bald or golden eagles are covered in
an HCP and ITP, the take is authorized under BGEPA even if the eagle species is not listed under the
ESA (see 50 CFR 22.11(a)).

If bald or golden eagles are included as covered species in an HCP, the avoidance, minimization,
and other mitigation measures in the HCP must meet the BGEPA permit issuance criteria of 50 CFR
22.26, and include flexibility for adaptive management. If take of bald or golden eagles is likely but
the project developer or operator does not qualify for eagle take authorization (or chooses not to
request such authorization), an ITP may be issued in association with the proposed HCP. The
project proponent must be advised, in writing, that bald or golden eagles would not be included as
covered species and take of bald eagles or golden eagles would not, therefore, be authorized under
the incidental take permit. The project developer or operator must also be advised that the
incidental take permit would be subject to suspension or revocation if take of bald eagles or golden
eagles should occur.

In addition to ESA, wind project developers or operators need to address take under MBTA. MBTA
prohibits the taking, hunting, killing, pursuit, capture, possession, sale, barter, purchase, transport,
and export of migratory birds, their eggs, parts, and nests, except when authorized by the
Department of the Interior. For eagles, the BGEPA take authorization serves as authorization under
MBTA per 50 CFR 22.11(b). For other MBTA‐protected birds, because neither the MBTA nor its
permit regulations at 50 CFR Part 21 currently provide a specific mechanism to permit
“unintentional” take, it is important for project developers or operators to work proactively with
the Service to avoid and minimize take of migratory birds. The Service, with assistance from a
Federal Advisory Committee, developed the WEG to provide a structured system to evaluate and
address potential negative impacts of wind energy projects on species of concern. Because the
Service has the authority to issue a permit for non‐purposeful take of eagles, our legal and
procedural obligations are significantly greater, and therefore the ECPG is more focused and
detailed than the WEG. We have modeled as much of the ECPG as possible after the WEG, but there
are important and necessary differences.

NEPA applies to issuance of eagle take permits because issuing a permit is a federal action. While
providing technical assistance to agencies conducting NEPA analyses, the Service will participate in
the other agencies' NEPA to the extent feasible in order to streamline subsequent NEPA analyses
related to a project. For actions that may result in applications for development of programmatic
permits, the Service may participate as a cooperating agency to streamline the permitting process.

If no federal nexus exists, other than an eagle permit, or if the existing NEPA of another agency is
not adequate, the Service must complete a NEPA analysis before it can issue a permit. The Service
will work with the project developer or operator to conduct a complete NEPA analysis, including
assisting with data needs and determining the scope of analysis. Project developers or operators
may provide assistance that can expedite the NEPA process in accordance with 40 CFR §1506.5.
Additionally, there are opportunities to “batch” NEPA analyses for proposed projects in the same
geographic area. In these cases, project developers or operators and the Service could pool
resources and data, likely increasing the quality of the product and the efficiency of the process.
Developers should coordinate closely with the Service for projects with no federal nexus other than

8
the eagle permit. Close coordination between project developers or operators and the Service
regarding the data needs and scope of the analysis required for a permit will reduce delays.

Through 50 CFR 22.26 and the associated FEA, the Service defined “mitigation” as per the Service
Mitigation Policy (46 FR 7644, Jan. 23, 1981), and the President’s Council on Environmental Quality
(40 CFR 1508.20 (a‐e)), to sequentially include the following:

(1) Avoiding the impact on eagles altogether by not taking a certain action or parts of an action;
(2) Minimizing impacts by limiting the degree or magnitude of the action and its
implementation;
(3) Rectifying the impact by repairing, rehabilitating, or restoring the affected environment;
(4) Reducing or eliminating the impact over time by implementing preservation and
maintenance operation during the lifetime of the action; and
(5) Compensating for the impact by replacing or providing substitute resources or
environments.

Throughout this document we differentiate between mitigation, which covers all of the components
listed above, and compensatory mitigation, which is a subset of (5) above and directly targets
offsetting permitted disturbance and mortality to accomplish a no‐net‐loss objective at the scale of
the eagle management unit. The Service requires compensatory mitigation (potentially in addition
to other mitigation) where it has not been determined that eagle populations can sustain additional
mortality. The NEPA analysis on our permits and the discussion of mitigation in this document
follow this system, and in this ECPG we refer to (1) – (4) as conservation measures to avoid and
minimize take, of which ACPs are a subset, and to (5) as compensatory mitigation.

Eagles are significant species in Native American culture and religion (Palmer 1988) and may be
considered contributing elements to a “traditional cultural property” under Section 106 of the
NHPA. Some locations where eagles would be taken have traditional religious and cultural
importance to Native American tribes and thus have the potential of being regarded as traditional
cultural properties under NHPA. Permitted take of one or more eagles from these areas, for any
purpose, could be considered an adverse effect to the traditional cultural property. These
considerations will be incorporated into any NEPA analysis associated with an eagle take permit.

Federally‐recognized Indian tribes enjoy a unique government‐to‐government relationship with the
United States. The Service recognizes Indian tribal governments as the authoritative voice
regarding the management of tribal lands and resources within the framework of applicable laws. It
is important to recall that many tribal traditional lands and tribal rights extend beyond reservation
lands. The Service consults with Indian tribal governments under the authorities of Executive
Order 13175 “Consultation and Coordination with Indian Tribal Governments” and supporting DOI
and Service policies. To this end, when it is determined that federal actions and activities may
affect a tribe’s resources (including cultural resources), lands, rights, or ability to provide services
to its members, the Service must, to the extent practicable, seek to engage the affected tribe(s) in
consultation and coordination.

3. Background and Overview of Process
Increased energy demands and the nationwide goal to increase energy production from renewable
sources have intensified the development of energy facilities, including wind energy. The Service
supports renewable energy development that is compatible with fish and wildlife conservation.
The Service closely coordinates with state, tribal, and other federal agencies in the review and

9
permitting of wind energy projects to address potential resource effects, including effects to bald
and golden eagles. However, our knowledge of these effects and how to address them at this time is
limited. Given this and the Service’s regulatory mandate to only authorize actions that are
“compatible with the goal of stable or increasing breeding populations” of eagles has led us to adopt
an adaptive management framework predicated, in part, on the precautionary approach for
consideration and issuance of programmatic eagle take permits. This framework consists of case‐
specific considerations applied within a national framework, and with the outcomes carefully
monitored so that we maximize learning from each case. The knowledge gained through
monitoring can then be used to update and refine the process for making future permitting
decisions such that our ultimate conservation objectives are attained, as well as to consider
operational adjustments at individual projects at regular intervals where deemed necessary and
appropriate. The ECPG provides the background and information necessary for wind project
developers or operators to prepare an ECP that assesses the risk of a prospective or operating
project to eagles, and how siting, design, and operational modifications can mitigate that risk.
Implementation of the final ECP must reduce predicted eagle take, and the population level effect of
that take, to a degree compatible with regulatory standards to justify issuance of a programmatic
take permit by the Service.

a. Risks to Eagles
Energy development can affect eagles in a variety of ways. First, structures such as wind
turbines can cause direct mortality through collision (Hunt 2002, Nygård et al. 2010). This
is the primary threat to eagles from wind facilities, and the monitoring and avoidance and
minimization measures advocated in the ECPG primarily are aimed at this threat. Second,
activities associated with pre‐construction, construction, or operation and maintenance of a
project might cause disturbance and result in loss of productivity at nearby nests or
disturbance to nearby concentrations of eagles. Third, if disturbance or mortality effects
are permanent, they could result in the permanent or long term loss of a nesting territory.
All of these impacts, unless properly permitted, are violations of BGEPA (USFWS 2009a).
Additionally, disturbances near important eagle use areas or migration concentration sites
might stress eagles to a degree that leads to reproductive failure or mortality elsewhere;
these impacts are of concern as well, and they could amount to prohibited take, though such
effects are difficult to predict and quantify. Thus, the ECPG addresses both direct mortality
and disturbance. Many new wind projects are located in remote areas that have few, if any,
transmission lines. The Service considers new transmission lines and other infrastructure
associated with renewable energy projects to be part of a project. Accordingly, assessments
of project impacts should include transmission lines and other facilities, not merely wind
turbines.

b. General Approach to Address Risk
Applicants for permits under 50 CFR 22.26, non‐purposeful eagle take, are required to avoid
and minimize the potential for take of eagles to the extent practicable. Permits for wind‐
energy development are programmatic as they will authorize recurring take, rather than
isolated incidences of take. For programmatic take permits, the regulations at 50 CFR 22.26
require that any authorized take is unavoidable after implementation of ACPs. 50 CFR 22.3
defines “advanced conservation practices” as “scientifically supportable measures that are
approved by the Service and represent the best available techniques to reduce eagle
disturbance and ongoing mortalities to a level where remaining take is unavoidable.”

10
Because the best information indicates that there are currently no available scientifically
supportable measures that will reduce eagle disturbance and blade‐strike mortality at wind
projects, the Service has not currently approved any ACPs for wind‐energy projects.
The preamble to the Eagle Permit Rule envisioned the Service and industry working
together to identify and evaluate possible ACPs (USFWS 2009a). The process of ACP
development for wind‐energy facilities has been hampered because there has been little
standardized scientific study of potential ACPs, and such information can best be obtained
through experimental application of ACPs at operating facilities with eagle take permits.
Given this, and considering the pressing need to develop ACPs for wind‐energy facilities, the
Service believes that the best course of action is to work with industry to develop ACPs for
wind projects as part of the programmatic take permit process.

Under this scenario, ACPs would be implemented at operating wind facilities with an eagle
take permit on an “experimental” basis (the ACPs are considered experimental because they
would not yet meet the definition of an ACP in the eagle permit regulation). The
experimental ACPs would be scientifically evaluated for their effectiveness, and based on
the results of these studies, could be modified in an adaptive management regime.

Despite the current lack of available ACPs, the best available scientific information may
demonstrate that a particular avoidance, minimization, or other mitigation action should be
applied as a condition on an eagle programmatic take permit for wind‐energy facilities (see
50 C.F.R. 22.6(c)(1)). A project developer or operator will still be expected to implement
any reasonable avoidance and minimization measures that may reduce take of eagles at a
project. However, the Service and the project developer or operator will discuss and agree
on other site‐specific and possibly turbine‐specific factors that may pose risks to eagles and
experimental ACPs that might reduce or eliminate those risks if the risks are substantiated
by the best available science. Unless the Service determines that there is a reasonable
scientific basis to implement experimental ACPs up front, we recommend that such
measures be deferred until such time as there is eagle take at the facility or the Service
determines that the circumstances and evidence surrounding instances of take or risk of
take suggest the experimental ACPs might be warranted. This agreement would be
specified as a condition of the programmatic eagle take permit.

Because ACPs would be considered experimental in these situations, we recommend that
they be subject to a cost cap that the Service and the project developer or operator establish
as part of the initial agreement before issuance of a permit, thereby providing financial
certainty to the project operator or developer as to what maximum costs of such measures
might be. The amount of the cap should be relevant to the theorized risk factors identified
for the project, and proportional to overall risk.

If eagle take is confirmed through post‐construction monitoring, developers or operators
would be expected to implement the experimental ACP(s) and to monitor future eagle take
relative to the ACP(s) as part of the adaptive management process specified in Appendix A,
but all within the limits of the pre‐determined financial cap. As the results from monitoring
experimental ACPs across a number of facilities accumulates and is analyzed as part of the
adaptive management process, scientific information in support of certain ACPs may accrue,
whereas other ACPs may show little value in reducing take. If the Service determines that
the available science demonstrates an experimental ACP is effective in reducing eagle take,
the Service will approve that ACP and require its implementation up front on new projects
when and where warranted.

11

Where take is unavoidable and when eagle populations at the scale of the eagle
management unit (as defined in USFWS 2009b) are not estimated to be healthy enough to
sustain additional mortality over existing levels, applicants must reduce the effect of
permitted unavoidable mortality to a no‐net‐loss standard through compensatory
mitigation for the duration of the permitted activity. No‐net‐loss means that unavoidable
mortality caused by the permitted activities is offset by compensatory mitigation that
reduces another, ongoing form of mortality by an equal or greater amount, or which leads to
an increase in carrying capacity that allows the eagle population to grow by an equal or
greater amount. Compensatory mitigation may also be necessary to offset substantial
effects in other situations (USFWS 2009a), and mitigation designed to offset other
detrimental effects of permits on eagles may be advised in addition to compensatory
mitigation in some cases. The Service and the project developer or operator seeking a
programmatic eagle take permit should agree on the number of eagle fatalities to mitigate
and what actions will be taken if actual eagle fatalities differ from the predicted number.
The compensatory mitigation requirement and trigger for adjustment should be specified in
the permit. If the procedures recommended in the ECPG are followed, there should not be a
need for additional compensatory mitigation. However, if other, less risk‐averse models are
used to estimate fatalities, underestimates might be expected and the permit should specify
the threshold(s) of take that would trigger additional actions and the specific mitigation
activities that would be implemented if fatalities are underestimated. The approach
described in the ECPG is applicable for all land‐based wind energy projects within the range
of the bald and golden eagle where interactions with wind project infrastructure have been
documented or are reasonably expected to occur. The ECPG is intended to provide a
national framework for assessing and mitigating risk.

As part of the application process for a programmatic eagle take permit, the Service
recommends that project developers or operators prepare an ECP that outlines the project
development process and includes conservation and monitoring plans as recommended in
this ECPG. The ECPG provides examples of ways that applicants can meet the regulatory
standards in the rule, and while other approaches may be acceptable, the Service will
determine their adequacy on a case‐by‐case basis. As noted previously, an ECP is not
required, but if one is developed following the approach recommended here, it will expedite
Service review of the project.

There is substantial uncertainty surrounding the risk of wind projects to eagles, and of ways to
minimize that risk. For this reason, the Service strongly recommends that care be taken to protect
against the consequences of underestimating eagle fatality rates at wind facilities. Overestimates,
once confirmed, can be adjusted downward based on post‐construction monitoring information
with no consequence to eagle populations, and project developers or operators can trade or be
credited for excess compensatory mitigation. However, the options for addressing underestimated
fatality rates are extremely limited, and pose either potential hardships for wind developers or
significant risks to eagle populations.

12

ASSESSING RISK AND EFFECTS
1. Considerations When Assessing Eagle Use Risk
Bald eagles and golden eagles associate with distinct geographic areas and landscape features
throughout their respective ranges. The Service defines these “important eagle‐use areas” as “an
eagle nest, foraging area, or communal roost site that eagles rely on for breeding, sheltering, or
feeding, and the landscape features surrounding such nest, foraging area, or roost site that are
essential for the continued viability of the site for breeding, feeding, or sheltering eagles” (USFWS
2009a; 50 CFR 22.3). Migration corridors and migration stopover sites also provide important
foraging areas for eagles during migration (e.g., Restani et al. 2001, Mojica 2008) and result in
seasonal concentrations of eagles. As a result, the presence of a migration corridor or stopover site
on or near a proposed wind development project could increase the probability of encounters
between eagles and wind turbines. Although these sites are not specifically included within the
regulatory definition of an important eagle‐use area at 50 CFR 22.3, the presence of such a site on
or near a proposed wind project could increase the likelihood of collisions.

Wind energy projects that overlap, or are proximate to, important eagle use areas or migration
concentration sites may pose risks to the eagles for reasons described earlier. Project developers
or operators should identify the location and type of all important eagle use areas or migration
concentration sites that might be affected by a proposed wind project (e.g., within the project area).
If recent (within the previous 5 years) local data are available on the spacing of eagle nests for the
project‐area nesting population, those data can be used to determine an appropriate boundary for
such surveys (as described in Appendix H). Otherwise, for both species we suggest initial surveys
be conducted on and within 10 miles of a project’s footprint to establish the project‐area mean
inter‐nest distance. The project footprint is the minimum convex polygon (e.g., Mohr 1947) that
encompasses the wind project area inclusive of the hazardous area around all turbines and any
associated infrastructure, including utility lines, out‐buildings, roads, etc. We suggest a site‐specific
approach based on the spacing between nearest, simultaneously occupied nests for the species
present in the area. If data on nest‐spacing in the project area are lacking, project proponents or
operators may wish to survey up to 10 miles, as this is ½ the largest recorded spacing observed for
golden eagles in the Mojave/Sonoran deserts of western Arizona (Millsap 1981). . For subsequent
monitoring (e.g., post‐construction monitoring of occupancy and productivity of pairs potentially
disturbed by the project), the project‐area mean inter‐nest distance can be used to define a more
relevant project‐area boundary. The 10‐mile perimeter may be unnecessary for bald eagles in
some areas, and the Service acknowledges there needs to be flexibility in the application of this
approach to accommodate specific situations.

Evaluating the spatial area described above for each wind project is a key part of the programmatic
take permitting process. As described later, surveys should be conducted initially to obtain data to
predict effects of wind projects on eagles. After the project begins operating, studies should again
be conducted to determine the actual effects. The following sections include descriptions and
criteria for identifying important eagle use areas or migration concentration sites in these
assessments.

a. General Background and Rationale for Assessing Project Effects on Eagles
A synthesis of publicly available databases and technical literature are fundamental to the
pre‐construction assessment component of an ECP. In some instances, this work may
reveal information on use of a proposed project area by eagles that is strong enough to
support a decision on whether to proceed with the project. In most cases, if available

13
information warrants further consideration of a potential wind project site, on‐site surveys
should be implemented to further document use of the project area by eagles. The goal of
such surveys should be to quantify and describe use of the project area by breeding
(territorial) and non‐breeding eagles across seasons and years. A variety of survey
approaches may be needed to accomplish this goal.

Although potential for presence of all types of important eagle use areas or migration
concentration sites should be considered when beginning to assess a potential project site,
special attention is typically given to nests and nesting pairs. An eagle territory is defined in
50 CFR 22.3 as an area that contains, or historically contained, one or more nests within the
home range of a mated pair of eagles. We recognize that usage conflicts with the true
biological meaning of the term territory, but we use it herein in its regulatory context.
Newton (1979) considered the nesting territory of a raptor as the defended area around a
pair’s nest site and defined the home range as “...the area traveled by the individual in its
normal activities of food gathering, mating, and caring for the young.” For golden eagles at
least, the extent of the home range and territory during nesting season generally are
similar; the eagle defends its territory by undulating flight displays near the home range
boundaries and adjoining territories barely overlap (Harmata 1982, Collopy and Edwards
1989, Marzluff et al. 1997).

Avoidance zones, often distinguished by specific “buffer” distances, have been prescribed to
protect nests and other types of eagle use areas from disturbance. Recommendations for
the size of avoidance zones for nests of bald eagles and golden eagles have sometimes been
based on documented distances between nests and territory boundaries. For example,
McGrady et al. (2002) and Watson and Davies (2009) indicated nesting territories of golden
eagles extend to at least 4 miles from their nests. Garrett et al. (1993) found that bald eagle
territories extend at least 2 miles from nests, though studies in areas of densely packed
breeding territories of bald eagles suggest much smaller distances (Sherrod et al. 1976,
Hodges and Robards 1982, Anthony 2001). A recommendation for a spatial buffer to avoid
disturbance of eagle nests can hardly be applied throughout the entire range of either
species due to marked variation in the size and configuration of nesting territories. As such,
these avoidance prescriptions have been conservative because there are few site‐specific
data on spatial extent of territories in the published and unpublished literature. For bald
eagles, minimum‐distance buffers are prescribed by the Service to protect nests, foraging
areas, and communal roosts against disturbance from a variety of activities (USFWS 2007b).

The approach we recommend in the ECPG for evaluating siting options and assessing
potential mortality and disturbance effects of wind facilities on eagles is to conduct
standardized surveys (e.g., point counts) to estimate eagle exposure within the project
footprint. We further suggest augmenting these with surveys to determine locations of
important eagle use areas or migration concentration sites for the project‐area eagle
population. The project‐area eagle population is the population of breeding, resident non‐
breeding, migrating, and wintering eagles within the project area. As described previously
and in Appendix H, if recent data on the spacing of eagle nests in the project area are
available, it may be appropriate to use the mean species‐specific inter‐nest distance
(assuming there is no reason to suspect eagle territories in the project area are configured
such that the mean inter‐nest distance would be misleading) as the outer boundary of the
project area. Such a choice, however, also increases the importance of having adequate
eagle exposure information from the project footprint for all seasons. For example, a winter
communal night roost of eagles further than one mean inter‐nest distance from the project

14
boundary could produce a large influx of eagles into the footprint in winter. Inadequate
winter eagle exposure sampling (or sampling in only one year, if the night roost is not used
annually) in combination with selection of a project area based on nest spacing alone, could
result in a failure to detect this increased risk to eagles in winter. Unpredicted fatalities that
result from such an oversight will have to be addressed by the project developers or
operators eventually through increased compensatory mitigation, operational adjustments,
or both to continue operating under the authority of a valid eagle permit. Thus, it is
important that the combination of exposure and project‐area surveys adequately capture all
risks to eagles.

One‐half the mean inter‐nest distance has been used as a coarse approximation for the
territory boundary in a number of raptor studies (e.g., Thorstrom 2001, Wichmann et al.
2003, Soutullo et al. 2006). Eagle pairs at nests within ½ the mean project‐area inter‐nest
distance of the project footprint are potentially susceptible to disturbance take and blade‐
strike mortality, as these pairs and offspring may use the project footprint. We recommend
using this distance to delineate territories and associated breeding eagles at risk of
mortality or disturbance. Exposure surveys should adequately sample the parts of the
project footprint potentially used by these eagle pairs so they are captured in the fatality
estimates, and these nests should be included in post‐construction occupancy and
productivity monitoring (see Appendix H). This information is useful in decisions on
whether a wind project might meet permit requirements at 50 CFR 22.26 considering both
predicted take through fatalities and likely take from disturbance; for evaluating various
siting and project‐configuration alternatives; and in monitoring for disturbance effects
during the post‐construction period. In some situations, as where nests are concentrated
on linear features (such as cliffs for golden eagles or along rivers for bald eagles), ½ the
mean inter‐nest distance may not encompass all important parts of the territory. In these
situations inferences based on nest spacing should be used cautiously. The overall
effectiveness of this approach will be evaluated through post‐construction monitoring and
the adaptive management framework described later in this ECPG.

b. Additional Considerations for Assessing Project Effects: Migration Corridors and
Stopover Sites
Bald eagles and golden eagles tend to migrate along north‐south oriented cliff lines, ridges,
and escarpments, where they are buoyed by uplift from deflected winds (Kerlinger 1989,
Mojica et al. 2008). Bald eagles typically migrate during midday by soaring on thermal
uplift or on winds aloft, the onset of dally movements migration being influenced by rising
temperatures and favorable winds (Harmata 2002). Both species will forage during
migration flights, though for bald eagles foraging often is limited to lakes, rivers, streams,
and other wetland systems (Mojica et al. 2008). Both species use lift from heated air from
open landscapes to move efficiently during migration and seasonal movements, gliding
from one thermal to the next and sometimes moving in groups with other raptor species.

Passage rates and altitude of migrant eagles can be influenced by temperature, barometric
pressure, winds aloft, storm systems, weather patterns at the site of origin, and wind speed
(Yates et al. 2001). Both species avoid large water bodies during migration and funnel
along the shoreline, often becoming concentrated at the tips of peninsulas or in other
situations where movement requires water crossings (Newton 1979). Eagles annually use
stopover sites with predictably ample food supplies (e.g., Restani et al. 2000, Mojica et al.
2008), although some stopovers may be brief and infrequent, such as when optimal

15
migration conditions suddenly become unfavorable and eagles are forced to land and seek
roosts. Presence of a migration corridor or stopover site in the project area is best
documented and delineated by using a standard “hawk watch” migration count as
recommended in this ECPG as part of site‐specific surveys or, in some cases, by simply
expanding point count surveys to account for migration incidence during what normally
would be the peak migration period (Appendix C).

Much eagle mortality could occur if communal night roosts or communal foraging areas of
eagles are separated by strings of wind turbines from other areas used by eagles. Outside
the breeding season, both bald eagles and golden eagles can roost communally. Such roosts
can include individuals of all ages and residency status (Platt 1976, Craig and Craig 1984,
Mojica et al. 2008). During the breeding season, non‐breeding bald eagles also may roost
communally. Large roosts of eagles tend to be associated with nearby foraging areas.
Conversely, eagles also may congregate to forage at sites of unusually high prey or carcass
availability; such concentrations of bald eagles may number in the hundreds (Buehler
2000). Methods for documenting concentrations of eagles, and movements to and from
such areas in relation to the project footprint are provided in Appendix C.

2. Eagle Risk Factors
Factors that influence vulnerability of eagles to collisions with wind turbines are poorly known.
Theoretically, two major elements are likely involved: (1) eagle abundance, and (2) the presence of
features or circumstances that decrease an eagle’s ability to perceive and avoid collision. However,
the relative importance of these factors, and how they interrelate, remains poorly understood for
eagles and birds in general (Strickland et al. 2011). Table 1 lists some of the factors known or
postulated to be associated with turbineblade‐strike risk in raptors, but evidence for or against
these is equivocal, and may well vary between sites. While some of these factors are not known to
affect eagles, because of the similarity of flight behavior between eagles and some other soaring
raptors, we include them here because they may apply to eagles. Evidence across multiple studies
suggests that in addition to eagle abundance, two main factors contribute to increased risk of
collision by eagles: (1) the interaction of topographic features, season, and wind currents that
create conditions for high‐risk flight behavior near turbines; and (2) behavior that distracts eagles
and presumably makes them less vigilant (e.g., active foraging or inter‐ and intra‐specific
interactions).

Table 1. Factors potentially associated with wind turbine collision risk in raptors. Not all factors apply to
eagles, and the influence of these factors may vary in association with other covariates on a case-by-case basis.
Risk Factor
Bird Density

Bird Age

Status of Knowledge from Literature
Mixed findings; likely some
relationship but other factors have
overriding influence across a range of
species.
Mixed findings. Higher number of
fatalities among subadult and adult
golden eagles in one area. Higher
fatalities among adult white‐tailed
eagles in another.

Citations
Barrios and Rodriguez (2004), De
Lucas et al. (2008), Hunt (2002),
Smallwood et al. (2009), Ferrer et al.
(2011)

Hunt (2002), Nygård et al (2010)

16
Risk Factor

Interaction with
Other Birds

Status of Knowledge from Literature
White‐tailed eagle nesting areas close
to turbines have been observed to have
low nest success and be abandoned
over time.
Mixed findings. Higher risk to resident
adults in Egyptian vultures (Neophron
percnopterus). High number of
mortalities among subadults and
floating adults in golden eagles in one
other study.
Mixed findings. In some cases for some
species, risk appears higher in seasons
with greater propensity to use slope
soaring (fewer thermals) or kiting
flight (windy weather) while hunting.
Species most at risk perform more
frequent flights that can be described
as kiting, hovering, and diving for prey.
Higher risk when interactive behavior
is occurring.

Active Hunting/
Prey Availability

High risk when hunting close to
turbines, across a range of species.

Proximity to
Nests

Bird Residency
Status

Season

Flight Style

Turbine Height
Rotor Speed
Rotor‐swept
Area

Topography

Wind Speed

Mixed, contradictory findings across a
range of species.
Higher risk associated with higher
blade‐tip speed for golden eagles in one
study, but this finding may not be
generally applicable.
Meta‐analysis found no effect, but
variation among studies clouds
interpretation.
Several studies show higher risk of
collisions with turbines on ridge lines
and on slopes. Also a higher risk in
saddles that present low‐energy ridge
crossing points.
Mixed findings, probably locality
dependent.

Citations
Nygård et al (2010)

Barrios and Rodriguez (2004), Hunt
(2002)

Barrios and Rodriguez (2004), De
Lucas et al. (2008), Hoover and
Morrision (2005), Smallwood et al.
(2009)
Smallwood et al. (2009)
Smallwood et al. (2009)
Barrios and Rodriguez (2004), De
Lucas et al. (2008), Hoover and
Morrision (2005), Hunt (2002),
Smallwood et al. (2009)
Barclay et al. (2007), De Lucas et al.
(2008)
Chamberlain et al. (2006)

Barclay et al. (2007)
Barrios and Rodriguez (2004), De
Lucas et al. (2008), Hoover and
Morrission (2005), Smallwood and
Thelander (2004)
Barrios and Rodriguez (2004),
Hoover and Morrision (2005),
Smallwood et al. (2009)

3. Overview of Process to Assess Risk
This ECPG, and in particular the eagle fatality prediction model described in Appendix D, relies on
the assumption that there is predictable relationship between pre‐construction eagle occurrence
and abundance in the project footprint and subsequent fatalities. Assessing the veracity of this
operating hypothesis is a key element of the adaptive management component of the ECPG. The
ECPG outlines a decision‐making process that gathers information at each stage of project
development, with an increasing level of detail. This approach provides a framework for making

17
decisions sequentially at three critical phases in project development: (1) siting, (2) construction,
and (3) operations. The greatest potential to avoid and minimize impacts to eagles occurs if eagle
risk factors are taken into account at the earliest phase of project development. If siting and
construction have proceeded without consideration of risks to eagles, significant opportunities to
avoid and minimize risk may have been lost. This can potentially result in greater compensatory
mitigation requirements or, in the worst case, an unacceptable level of mortality for eagles.

The related, but more general, WEG advocates using a five‐tiered approach for iterative decision
making relative to assessing and addressing wildlife effects from wind facilities. Elements of all of
those tiers apply here, but the process for eagles is more specifically defined and falls into five
broadly overlapping, iterative stages that largely do not parallel the WEG’s five tiers (Figures 1 and
2).

Stage 1 for eagles (Appendix B) combines Tiers 1 and 2 from the WEG, and consists of an initial site
assessment. In this stage project developers or operators evaluate broad geographic areas to
assess the relative importance of various areas to resident breeding and non‐breeding eagles, and
to migrant and wintering eagles. The Service is available to assist project developers or operators
in beginning to identify important eagle use areas or migration concentration sites and potential
eagle habitat at this stage. To increase the probability of meeting the regulatory requirements for a
programmatic take permit, biological advice from the Service and other jurisdictional wildlife
agencies should be requested as early as possible in the developer's planning process and should be
as inclusive as possible to ensure all issues are being address at the same time and in a coordinated
manner. Ideally, consultation with the Service, and state and tribal wildlife agencies is done prior to
any substantial financial commitment or finalization of lease agreements. During Stage 1 the
project developer or operator should gather existing information from publicly available literature,
databases, and other sources, and use those data to judge the appropriateness of various potential
project sites, balancing suitability for development with potential risk to eagles.

Once a site has been selected, the next stage, Stage 2, is site‐specific surveys and assessments
(this is the first component of Tier 3 in the WEG; Appendix C). During Stage 2 the project developer
or operator should collect quantitative data through scientifically rigorous surveys designed to
assess the potential risk of the proposed project to eagles. In the case of small wind projects (one
or a few small turbines), the project developer or operator should apply the predictive model
described in Stage 3 (below) to determine if stage 2 surveys are necessary. In many cases, the
hazardous area associated with such projects will be small enough that Stage 2 surveys will not be
necessary to demonstrate that the project will likely not take eagles.

In Stage 3, the predicting eagle fatalities stage, the Service and project developers or operators
use data from Stage 2 in standardized models linked to the Service’s adaptive management process
to generate predictions of eagle risk in the form of average number of fatalities per year
extrapolated to the tenure of the permit (see Appendix D). These models can be used to
comparatively evaluate alternative siting, construction, and operational scenarios, a useful feature
in constructing hypotheses regarding predicted effects of conservation measures and ACPs. We
encourage project developers or operators to use the recommended pre‐construction survey
protocol in this ECPG in Stage 2 to help inform our predictive models in Stage 3. If Service‐
recommended survey protocols are used, this risk assessment can be greatly facilitated using model
tools available from the Service. If project developers or operators use other forms of information
for the Stage 2 assessment, they will need to fully describe those methods and the analysis used for
the eagle risk assessment, and more time will be required for Service biologists to evaluate and

18
review the data. For example, the Service will compare the results of the project developer or
operator’s eagle risk assessment with predictions from our models, and if the results differ, we will

Figure 1. Chart comparing Land-based Wind Energy Guideline tiers with Eagle Conservation Plan Guidance stages.

work with the project developers or operators to determine which model results are most
appropriate for the Service’s eventual permitting decisions. The Service and project developers or
operators also evaluate Stage 2 data to determine whether disturbance take is likely, and if so, at
what level. Any loss of production that may stem from disturbance should be added to the fatality
rate prediction for the project. The risk assessments at Stage 2 and Stage 3 are consistent with
developing the information necessary to assess the efficacy of conservation measures, and to
develop the monitoring required by the permit regulations at 50 CFR 22.26(c)(2).

Stage 4 is the avoidance and minimization of risk using conservation measures and ACPs and
compensatory mitigation (if required).

Conservation measures and ACPs. Regardless of which approach is employed in the Stage
3 assessment, in Stage 4 the information gathered should be used by the project developer
or operator and the Service to determine potential conservation measures and ACPs (if
available) that can be employed to avoid and/or minimize the predicted risks at a given site
(see Appendix E). The Service will compare the initial predictions of eagle mortality and
disturbance for the project with predictions that take into account proposed and potential
conservation measures and ACPs to determine if the project developer or operator has
avoided and minimized risks to the maximum degree achievable, thereby meeting the
requirements for programmatic permits in 50 CFR 22.26 that remaining take is
unavoidable. Additionally, the Service will use the information provided along with other

19
data to conduct a cumulative effects analysis to determine if the project’s impacts, in
combination with other permitted take and other known factors affecting the local‐area and

20

Figure 2. Figure 1 from WEG, adapted to show where and how eagles are considered in that process and which Stage and section of the ECPG
are applicable at each Tier of the WEG. Note that existing, operational wind energy projects enter the process between Tiers 3 and 4.

21
eagle management unit population(s), are at a level that exceed established thresholds or
benchmarks (see Appendix F). This final eagle risk assessment is completed at the end of
Stage 4 after application of conservation measures and ACPs along with a plan for
compensatory mitigation if required.

Compensatory Mitigation. Compensatory mitigation occurs in the eagle permitting
process if conservation measures and ACPs do not remove the potential for take, and the
projected take exceeds calculated thresholds for the species‐specific eagle management unit
in which the project is located. Compensatory mitigation may also be necessary in other
situations as described in the preamble to 50 CFR 22.26 (USFWS 2009a), and the following
guidance applies to those situations as well.

Compensatory mitigation can address any pre‐existing mortality source affecting the
species‐specific eagle management unit impacted by the project (e.g. environmental lead
abatement, addressing eagle electrocutions due to high risk power poles, etc.) that was in
effect at the time of the FEA in 2009 (USFWS 2009b), or it can address increasing the
carrying capacity of the eagle population in the affected eagle management unit. However,
there needs to be a credible analysis that supports the conclusion that implementing the
compensatory mitigation action will achieve the desired beneficial offset in mortality or
carrying capacity. All compensatory mitigation projects will be subjected to random
inspections by the Service or appointed subcontractors to examine efficacy, accuracy, and
reporting rigor.

For new wind development projects, if compensatory mitigation is necessary, the
compensatory mitigation action (or a verifiable, legal commitment to such mitigation) will
be required up front before project operations commence because projects must meet the
statutory and regulatory eagle preservation standard before the Service may issue a permit.
For operating projects that may meet permitting requirements, compensatory mitigation
should be applied from the start of the permit period, not retroactively from the initiation of
project operations. The initial compensatory mitigation contribution effort should be
sufficient to offset take at the upper 80% confidence limit (or equivalent) of the predicted
number of eagle fatalities per year for a five‐year period starting with the date the project
becomes operational (or, for operating projects, the date the permit is signed). No later
than at the end of the five year period, the predicted annual take estimate will be compared
to the realized take as estimated by post‐construction monitoring. If the triggers identified
in the permit for adjustment of compensatory mitigation are met, those adjustments should
be implemented. In the case where the realized take is less than predicted, the permittee
will receive a credit for the excess compensation (the difference between the actual mean
and the number compensated for) that can be applied to other take (either by the permittee
or other permitted individuals at his/her discretion) within the same eagle management
unit. Compensatory mitigation for future years for the project will be determined at this
point, taking into account the observed levels of mortality and any reduction in that
mortality that is expected based on implementation of additional experimental
conservation measures and ACPs that might reduce fatalities.

To illustrate an acceptable process for calculating compensatory mitigation, the Service has
prepared an example of a strategy using Resource Equivalency Analysis (REA) to quantify
the number of power pole retrofits needed to offset the take of golden eagles at a wind
project (see Appendix G). The Service used the example of eliminating electrocutions
because: (1) high‐risk power poles cause quantifiable adverse impacts to eagles; (2) the ‘per

22
eagle’ effects of high‐risk power pole retrofitting are quantifiable and verifiable through
accepted practices; (3) success of and subsequent maintenance of retrofitting can be
monitored; and (4) electrocution from high‐risk power poles is known to cause eagle
mortality and this can be corrected. The potential for take of eagles is estimated using
informed modeling, as described in Stage 3 of the ECPG (Appendix D). This fatality
prediction is one of several fundamental variables that are used to populate the REA (see
REA Inputs, Appendix G). The REA generates a project‐area eagle impact calculation
(debit), expressed in bird‐years, and an estimate of the quantity of compensatory mitigation
(credit) (e.g., power pole retrofits) necessary to offset this impact. Compensatory
mitigation would then be implemented either directly by the project developer or operator
or through a formal, binding agreement with a third party to implement the required
actions.

Effectiveness monitoring of the resulting compensatory mitigation projects should be
included within the above options using the best scientific and practicable method
available. The Service will modify the compensatory mitigation process to adapt to any
improvements in our knowledge base as new data become available.

At the end of Stage 4, all the materials necessary to satisfy the regulatory requirements to support a
permit application should be available. While the application can be submitted at any time, it is
only after completion of Stage 4 that the Service can begin the formal process to determine whether
a programmatic eagle take permit can be issued or not. Ideally, NEPA and NHPA analyses and
assessments will already be underway, but if not, Stage 4 should include necessary NEPA analysis,
NHPA compliance, coordination with other jurisdictional agencies, and tribal consultation.

If a permit is issued and the project goes forward, Stage 5 of the process is calibration and updating
of the fatality prediction and continued risk assessment, equivalent to Tier 4 and, in part, Tier 5 in
the WEG. During this stage, post‐construction surveys are conducted to generate empirical data for
comparison with the pre‐construction risk‐assessment fatality and disturbance predictions. The
monitoring protocol should include both validated techniques for assessing mortality, and for
estimating effects of disturbance to eagles, and they must meet the permit‐condition requirements
at 50 CFR 22.26(c)(2). We anticipate that in most cases, intensive monitoring to estimate the true
annual fatality rate and to assess possible disturbance effects will be conducted for at least the first
two years after permit issuance, followed by less intense monitoring for up to three years after the
expiration date of the permit, in accordance with monitoring requirements at 50 CFR 22.26(c)(2).
We recommend project developers or operators use the post‐construction survey protocols
included or referenced in this ECPG, but we will consider other monitoring protocols provided by
permit applicants. We will use the information from post‐construction monitoring in a meta‐
analysis framework to weight and improve pre‐construction predictive models. Additionally in
Stage 5 the Service and project developers or operators should use the post‐construction
monitoring data to (1) assess whether compensatory mitigation is adequate, excessive, or deficient
to offset observed mortality, and make adjustments accordingly; and (2) explore operational
changes that might be warranted at a project after permitting to reduce observed mortality and
ensure that permit condition requirements at 50 CFR 22.26(c)(7) are met.

Table 2 provides a summary of the roles of the project developer or operator and the Service,
responsibilities, and decision points at each stage.

23

Table 2. Roles, responsibilities of the project developers and operators and the Service, and decision
points at each stage of the ECP process.

Stage

Project developer/operator role


1






2






3








4


Conduct a desktop landscape‐level
assessment for known or likely
occurrence of eagles, including
reconnaissance visits to prospective
sites.
Consult with the Service on potential for
any obvious negative impacts on eagles
in at least general locale of prospective
sites.
Decision point: select site(s) for Stage 2
study, if appropriate.
Conduct detailed, site‐specific field
studies in the project area to inform
eagle fatality prediction model,
document important eagle use areas or
migration concentration sites, and
identify possible eagle disturbance
issues.
Coordinate in advance with the Service
and other jurisdictional agencies to
ensure studies will satisfy regulatory
requirements for permitting.
Decision point: choose whether to move
to Stage 3.

Service role










Optionally generate an estimated annual
eagle fatality prediction for the site(s)
and an assessment of eagle disturbance 
risk using data from Stage 2 and
model(s) of choice.
Report on all other germane aspects of

eagle use such as communal roosts and
nest or territory locations.

Decision point: choose whether to move
to Stage 4.

Identify conservation measures and ACP 
s that can be used to avoid and minimize
take identified in Stage 3.

Optionally generate revised fatality and
disturbance estimates, taking into

account conservation measures and
ACPs.

Identify and develop necessary
agreements for compensatory

mitigation to offset take, if required.

Recommend and help provide existing data and
input if requested.
Provide preliminary consultation on
appropriateness of application for eagle take
permits for sites considered and the likelihood
permits could be issued.
Review available Stage 1 data and advise what
Stage 2 data are recommended.
Decision point: none.

Consult on field study design and approach in
coordination with other jurisdictional agencies.
Decision point: None.

Generate an initial eagle fatality estimate for
site(s), using the Service model and survey data
from Stage 2.
Assess likelihood of disturbance to eagles;
quantify extent and impact of disturbance, if any
likely.
Make preliminary recommendation on risk
category.
Consult with developer/operator to interpret and
resolve discrepancies in conclusions and risk
category recommendation.
Decision point: None.
Re‐run Service fatality model to predict fatalities
with conservation measures and ACPs.
Re‐assess potential for disturbance take with
conservation measures and ACPs.
Coordinate with developer/operator to reach
agreement on predicted take and risk category.
Coordinate with developer/operator on
compensatory mitigation, if requested.
Provide revised preliminary assessment of
likelihood site(s) will be permittable if requested.

24
Stage

Project developer/operator role


Decision point: choose whether to
submit eagle take permit application.

Service role


Decision point: None.



Coordinate and consult on writing of ECP or
equivalent, including proposed plan for post‐
construction.
Convey adequacy of ECP or equivalent to
developer/operator.
Evaluate permit application for regulatory
sufficiency.
Draft permit conditions drawing on relevant
components of ECP or equivalent.
Conduct cumulative effects analysis.
Conduct NEPA review.
Conduct NHPA evaluation.
Coordinate with other jurisdictional agencies.
Consult with Tribes.
Establish limits on future operational adjustments
proportionate to risk, in coordination with
applicant.
Decision point: whether permit can be issued.
Monitor compliance with permit conditions.
Review post‐construction monitoring data,
including comparison of predicted and observed
annual fatality rate and disturbance.
At no more than 5‐year intervals, determine
whether revision of the estimated fatality rate,
adjustments to monitoring, implementation of
additional experimental conservation measures
and ACPs, and compensatory mitigation are
warranted.
Effect any necessary adjustments by crediting
back excess compensatory mitigation, or by
assessing additional compensatory mitigation for
fatalities in excess of predictions.
Combine monitoring data with that from other
projects for meta‐analysis within adaptive
management framework.
Decision point: determine what adjustments need
to be made to compensatory mitigation level, and
whether additional conservation measures and
ACPs are warranted or not.





Permit
Decision



Draft ECP or equivalent, including a plan
for post‐construction monitoring of
eagle fatality and disturbance.
Submit a permit application that meets
requirements at 50 CFR 22.26 or 22.27,
including ECP or equivalent information
as part of application package.
Choose whether to assist Service in
conducting NEPA.
Decision point: None.

















5




Implement post‐construction
monitoring in accordance with permit
conditions, including immediate
reporting of any eagle take.
Participate in scheduled reviews of
post‐construction monitoring results.
Effect additional compensatory
mitigation if necessary.
Implement and monitor additional
conservation measures and ACPS, if
warranted, within scope of permit
sideboards.
Decision point: choose whether to apply
for permit renewal near the end of
permit term.









25

4. Site Categorization Based on Mortality Risk to Eagles
We recommend the approach outlined below be used to categorize the likelihood that a site or
operational alternative will meet standards in 50 CFR 22.26 for issuance of a programmatic eagle
take permit.

a. Category 1 – High risk to eagles, potential to avoid or mitigate impacts is low
A project is in this category if it:
(1) has an important eagle‐use area or migration concentration site within the project
footprint; or
(2) has a species‐specific uncertainty‐adjusted annual fatality estimate (average number
of eagles predicted to be taken annually) > 5% of the estimated species‐specific
local‐area population size; or
(3) causes the cumulative annual take for the local‐area population to exceed 5% of the
estimated species‐specific local‐area population size.

In addition, projects that have eagle nests within ½ the mean project‐area inter‐nest
distance of the project footprint should be carefully evaluated (see Appendix H). If it is
likely eagles occupying these territories use or pass through the project footprint, category
1 designation may be appropriate.

Projects or alternatives in category 1 should be substantially redesigned if they are to at
least meet the category 2 criteria. Construction of projects at sites in category 1 is not
recommended because the project would likely not meet the regulatory requirements for
permit issuance and may place the project developer or operator at risk of violating the
BGEPA. The recommended approach for assessing the percentage of the local‐area
population predicted to be taken is described in Appendix F.

b. Category 2 – High or moderate risk to eagles, opportunity to mitigate impacts
A project is in this category if it:
(1) has an important eagle‐use area or migration concentration site within the project
area but not in the project footprint; or
(2) has a species‐specific uncertainty‐adjusted fatality estimate between 0.03 eagles per
year and 5% of the estimated species‐specific local‐area population size; or
(3) causes cumulative annual take of the species‐specific local‐area population of less
than 5% of the estimated local‐area population size.

Projects in this category will potentially take eagles at a rate greater than is consistent with
maintaining stable or increasing populations, but the risk might be reduced to an acceptable
level through a combination of conservation measures and reasonable compensatory
mitigation. These projects have a risk of ongoing take of eagles, but this risk can be
minimized. For projects in this category the project developer or operator should prepare
an ECP or similar plan to document meeting the regulatory requirements for a
programmatic permit. For eagle management populations where take thresholds are set at
zero, the conservation measures in the ECP should include compensatory mitigation and
must result in no‐net‐loss to the breeding population to be compatible with the permit
regulations. This does not apply to golden eagles east of the 100th meridian, for which no
non‐emergency take can presently be authorized (USFWS 2009b).

26

c. Category 3 – Minimal risk to eagles
A project is in this category if it:
(1) has no important eagle use areas or migration concentration sites within the project
area; and
(2) has a species‐specific uncertainty‐adjusted annual fatality rate estimate of less than
0.03 for both species of eagle; and
(3) causes cumulative annual take of the local‐area population of less than 5% of the
estimated species‐specific local‐area population size.

Projects in category 3 pose little risk to eagles and may not require or warrant eagle take
permits, but that decision should be made in coordination with the Service. Still, a project
developer or operator may wish to create an ECP that documents the project’s low risk to
eagles, and outlines mortality monitoring for eagles and a plan of action if eagles are taken
during project construction or operation. If take should occur, the developer or operator
should contact the Service to discuss ways to avoid take in the future. Such an ECP would
enable the Service to provide a permit to allow a de minimis amount of take if the project
developer or operator wished to obtain such a permit.

The risk category of a project has the potential to change from one of higher risk to one of lower
risk or one of lower risk to one of higher risk through additional site‐specific analyses and
application of measures to reduce the risk. For example, a project may appear to be in category 2 as
a result of Stage 1 analyses, but after collection of site‐specific information in Stage 2 it might
become clear it is a category 1 project. If a project cannot practically be placed in one of these
categories, the project developer or operator and the Service should work together to determine if
the project can meet programmatic eagle take permitting requirements in 50 CFR 22.26 and 22.27.
Projects should be placed in the highest category (with category 1 being the highest) in which one
or more of the criteria are met.

5. Cumulative Effects Considerations
a. Early Planning
Regulations at 50 CFR 22.26 require the Service to consider the cumulative effects of
programmatic eagle take permits. Cumulative effects are defined as: “the incremental
environmental impact or effect of the proposed action, together with impacts of past,
present, and reasonably foreseeable future actions” (50 CFR 22.3). Thorough cumulative
effects analysis will depend on effective analysis during the NEPA process associated with
an eagle permit. Scoping and other types of preliminary analyses can help identify
important cumulative‐effects factors and identify applicable past, present, and future
actions. Comprehensive evaluation during early planning may identify measures that would
avoid and minimize the effects to the degree that take of eagles is not likely to occur. In that
case, there may be no permit, and thus no need for NEPA associated with an eagle take
permit. When a wind project developer or operator seeks an eagle take permit, a
comprehensive cumulative effects analysis at the early planning stage will serve to
streamline subsequent steps, including the NEPA process.

The Service recommends that cumulative effects analyses be consistent with the principles
of cumulative effects outlined in the Council on Environmental Quality (CEQ) handbook,
"Considering Cumulative Effects under the National Environmental Policy Act (1997) (CEQ
handbook). The Service recommends consideration of the following examples from the CEQ

27
handbook that may apply to cumulative effects to eagles and the ecosystems they depend
upon:

(1) Time crowding ‐ frequent and repetitive effects on an environmental system;
(2) Time lags ‐ delayed effects;
(3) Space crowding ‐ High spatial density of effects on an environmental system;
(4) Cross‐ boundary ‐ Effects occur away from the source;
(5) Fragmentation ‐ change in landscape pattern;
(6) Compounding effects ‐ Effects arising from multiple sources or pathways;
(7) Indirect effects ‐ secondary effects; and
(8) Triggers and thresholds ‐ fundamental changes in system behavior or structure.

b. Analysis Associated with Permits
The cumulative effects analysis for a wind project and a permit authorization should include
whether the anticipated take of eagles is compatible with eagle preservation as required at
50 CFR 22.26, including indirect impacts associated with the take that may affect eagle
populations. It should also include consideration of the cumulative effects of other
permitted take and additional factors affecting eagle populations.

Whether or not a permit authorization is compatible with eagle preservation was analyzed
in the FEA that established the thresholds for take (USFWS 2009b). The scale of that
analysis was based upon eagle management units as defined in USFWS (2009b). However,
the scale for cumulative effects analysis of wind projects and associated permits should
include consideration of the effects at the local‐population scale as well.

The cumulative effects analyses for programmatic permits should cover the time period
over which the take will occur, not just the period the permit will cover, including the effect
of the proposed action, other actions affecting eagles, predicted climate change impacts, and
predicted changes in number and distribution of affected eagle populations. Effects
analyses should note whether the project is located in areas where eagle populations are
increasing or predicted to increase based on available data, over the lifetime of the project,
even if take is not anticipated in the immediate future. In addition, conditions where
populations are saturated should be considered in cumulative effects analyses. Numerous
relatively minor disruptions to eagle behavior from multiple activities, even if spatially or
temporally distributed, may lead to disturbance that would not have resulted from fewer or
more carefully sited activities (e.g., Whitfield et al. 2007). Additional detailed guidance for
cumulative impacts analyses can be found on the Council on Environmental Quality website
at http://ceq.hss.doe.gov/nepa/ ccenepa/ ccenepa.htm.

Specific recommendations for conducting cumulative effects analysis of the authorized take
under eagle programmatic take permits is provided in Appendix F.

28

ADAPTIVE MANAGEMENT

Management of wind facilities to minimize eagle take, through decisions about siting, design,
operation, and compensatory mitigation, is a set of recurrent decisions made in the face of
uncertainty. The Department of the Interior has a long history of approaching such decisions
through a process of adaptive management (Williams et al. 2007). The purpose of adaptive
management is to improve long‐term management outcomes, by recognizing where key
uncertainties impede decision making, seeking to reduce those uncertainties over time, and
applying that learning to subsequent decisions (Walters 1986).

In the case of managing eagle populations in the face of energy development there is considerable
uncertainty to be reduced. For example, evidence shows that in some areas or specific situations,
large soaring birds, specifically raptors, are vulnerable to colliding with wind turbines (Barrios and
Rodriguez 2004, Kuvlesky et al. 2007). However, we are uncertain about the relative importance of
factors that influence that risk. We are also uncertain about the best way to mitigate the effects of
wind turbine developments on raptors; we suspect some strategies might be effective, others are
worth trying. We also suspect that a few species, including golden eagles (USFWS 2009b), may be
susceptible enough to collisions with wind turbines that populations may be negatively affected.
Thus, there are uncertainties at several levels that challenge our attempts to manage eagle
populations: (1) at the level of understanding factors that affect collision risk, (2) at the level that
influences population trends, and (3) about the efficacy of various mitigation options. The Service,
our conservation partners, and industry will never have the luxury of perfect information before
needing to act to manage eagles. Our goal is to reduce that uncertainty through use of formal
adaptive management, thereby improving our predictive capability over time. Applying a
systematic, cohesive, nationally‐consistent strategy of management and monitoring is necessary to
accomplish this goal.

In the context of wind energy development and eagle management under the ECPG, there are four
specific sets of decisions that will be approached through adaptive management: (1) adaptive
management of wind project operations; (2) adaptive management of wind project siting and
design recommendations; (3) adaptive management of compensatory mitigation; and (4) adaptive
management of population‐level take thresholds. These are discussed in more detail in Appendix A.
The adaptive management process will depend heavily on pre‐ and post‐construction data from
individual projects, but analyses, assessment, and model evaluation will rely on data pooled over
many individual wind projects. Therefore, individual project developers or operators will have
limited direct responsibilities for conducting adaptive management analyses, other than to provide
data through post‐construction monitoring.

29

EAGLE CONSERVATION PLAN DEVELOPMENT PROCESS

The following sections of the ECPG, including attached appendices, provide a descriptive
instructional template for developing an ECP. Throughout this section, we use the term ECP to
include any other document or collection of documents that could be considered equivalent to an
ECP. The ECP is an integral part of the permit process, and the following chronological step‐by‐step
outline shows how the pieces fit together:

The ECPG provides guidance and serves as a reference for project developers or operators, the
Service, and other jurisdictional agency biologists when developing and evaluating ECPs. Using the
ECPG as a non‐binding reference, the Service will work with project developers or operators to
develop an ECP. The ECP documents how the project developer or operator intends to comply with
the regulatory requirements for programmatic permits and the associated NEPA process by
avoiding and minimizing the risk of taking eagles up‐front, and formally evaluating possible
alternatives in (ideally) siting, configuration, and operation of wind projects. The Service’s ability
to influence siting and configuration factors depends on the stage of development of the project at
the time the project developer or operator comes to us.

The Service recommends that project developers or operators develop an ECP following the five‐
staged approach described earlier. During Stages 1 through 4, projects or alternatives should be
placed in one of the three risk categories, with increasing certainty by Stage 4. The ECP should
provide detailed information on siting, configuration, and operational alternatives that avoid and
minimize eagle take to the point any remaining take is unavoidable and, if required, mitigates that
remaining take to meet the statutory preservation standard. The Service will use the ECP and other
application materials to either develop an eagle take permit for the project, or to determine that the
project cannot be permitted because risk to eagles is too high to meet the regulatory permit
requirements.

For permitted projects, the Service will use the 80% upper confidence limit or similar risk‐averse
estimate (e.g., the upper limit of the 80% credible interval is used in the Service’s predictive model
described in Appendix D) of the mean annual predicted unavoidable eagle take to determine likely
population‐level effects of the permit and compensatory mitigation levels, if required. For
predicted recurring eagle take that is in excess of calculated eagle management unit take
thresholds, the Service will either (a) approve a compensatory mitigation proposal from the project
developer or operator; or (b) accept, if sufficient, a commitment of funds to an appropriate
independent third party that is formally obligated (via contract or other agreement with the project
developer or operator) to perform the approved mitigation work. Under either (a) or (b), the
compensatory mitigation cost and actions will be calibrated so as to offset the predicted
unavoidable take, such that we bring the individual permit’s (and cumulatively over all such
permits’) predicted mortality effect to a no‐net‐loss standard. Compensatory mitigation will
initially be based on the upper 80% confidence limit of the predicted mean annual fatality rate (or
similar risk‐averse estimate) over a five year period, and it will be adjusted for future years based
on the observed fatality rate over the initial period of intensive post‐construction monitoring (no
less than 2 years). Compensatory mitigation, as well as other forms of mitigation aimed at reducing
other detrimental effects of permits on eagles, may also be necessary in other situations where
predicted effects to eagle populations are substantial and not consistent with stable or increasing
breeding populations of eagles.

Post‐construction monitoring may be required as a condition of an eagle programmatic take permit
and will be required for wind‐energy projects that may potentially take eagles. This monitoring

30
should be systematic and standardized to be suitable for use in a formal adaptive management
framework to evaluate and improve the predictive accuracy of our models. In addition, the
information will be used by the Service and the project developer or operator to determine if, after
no more than five years of post‐construction monitoring, the 80% upper confidence limit on the
predicted mean number of annual fatalities adequately captured the observed estimated mean
number of fatalities annually. If the observed and predicted estimates of annual fatalities are
different, either additional compensatory mitigation will be required retroactively to offset higher‐
than‐predicted levels of take (assuming the actual number of eagles taken was greater than the
number actually compensated for), or the permittee will receive a credit for the excess
compensation (the difference between the actual mean and the number compensated for) that can
be applied to other take (either by the permittee or other permitted individuals at his/her
discretion) within the same eagle management unit at any time in the future.

At no more than five‐years from the date a permit is issued, the permittee will compile and the
Service and the permittee will review fatality information for the project to determine if
experimental ACPs should be implemented to potentially reduce eagle mortalities based on the
observed, specific situation at each site. As discussed previously, at the time of permit issuance the
Service and the project developer or operator will agree to an upper limit on the cost of such future
experimental ACPs, which will only be implemented if warranted by eagle disturbance or mortality
data. If these experimental ACPs are likely to reduce mortalities at the project in the future, the
amount of future compensatory mitigation will be decreased accordingly (e.g. if ACPs are predicted
to reduce the fatality rate from three to two eagles annually, compensatory mitigation would only
be required to offset the future predicted take of two eagles per year). In such cases, additional
post‐implementation monitoring should be conducted to determine the effectiveness of the
experimental ACPs. In cases where observed fatalities exceed predicted to the degree category 1
fatality‐rate criteria are confirmed to have been met or exceeded by a permitted project, and for
whatever reason experimental ACPs or additional conservation measures cannot be implemented
to reduce fatalities to category 2 levels or below, the Service may have to rescind the permit for that
project to remain in compliance with regulatory criteria.

Programmatic eagle take permits will be conditioned to require access to the areas where take is
possible and where compensatory mitigation is being implemented by Service personnel, or other
qualified persons designated by the Service, within reasonable hours and with reasonable notice
from the Service, for purposes of monitoring the site(s). The regulations provide, and a condition of
any permit issued will require, that the Service may conduct such monitoring while the permit is
valid, and for up to three years after it expires (50 CFR 22.26(c)(4)). In general, verifying
compliance with permit conditions is a secondary purpose of site visits; the primary purpose is to
monitor the effects and effectiveness of the permitted action and mitigation measures. This may be
done if a project developer or operator is unable to observe or report to the Service the information
required by the annual report—or it may serve as a “quality control” measure the Service can use to
verify the accuracy of reported information and/or adjust monitoring and reporting requirements
to provide better information for purposes of adaptive management.

1. Contents of the Eagle Conservation Plan
This section provides a recommended outline for an ECP, with a short description of what should
be contained in each section. See previous sections and referenced appendices for details on the
stages and categories.

31

a. Stage 1
Data from Stage 1 should be presented and summarized in this section of the ECP. The
project developer or operator should work with the Service to place potential wind–facility
site in a category based on the Stage 1 information. For detailed recommendations on the
Stage 1 process, see Appendix B.

b. Stage 2
Data from Stage 2 should be presented and summarized in this section of the ECP. For
detailed recommendations on the Stage 2 methods and metrics, see Appendix C. The risk
categorization should be re‐assessed in this section, taking into account Stage 2 results.

c. Stage 3
In this section of the ECP, project developers or operators should work in coordination with
the Service to calculate a prediction of the annual eagle fatality rate and confidence interval
for the project using data generated from the Stage 2 assessment. The initial estimate of the
fatality rate should not take into account possible conservation measures and ACPs; these
will be factored in as part of Stage 4. For detailed recommendations on Stage 3 methods
and metrics, see Appendix D. The risk categorization should be re‐assessed in this section,
taking into account Stage 3 results.

d. Stage 4
This section of the ECP should describe how proposed conservation measures and ACPs
should reduce the fatality rate generated in stage 3, and what compensatory mitigation
measures will be employed to offset unavoidable take, if required. This section facilitates
demonstrating how conservation measures and ACPs have reduced the raw predicted
fatality rate to the unavoidable standard. For detailed recommendations on considerations
for the development of conservation measures and ACPs see Appendix E. The risk
categorization should be re‐assessed in this section, taking into account Stage 4 results. This
should be the final pre‐construction risk categorization for the proposed project. This
section should also fully describe the proposed compensatory mitigation approach (if
required). For detailed recommendations regarding compensatory mitigation, see
Appendix G.

e. Stage 5 – Post-construction Monitoring
In this section of the ECP, the project developer or operator should describe the proposed
post‐construction survey methodology for the project. Detailed recommendations for post‐
construction monitoring are in Appendix H. The Stage 5 post‐construction monitoring plan
is the final section of the ECP.

32

INTERACTION WITH THE SERVICE

As noted throughout this ECPG, frequent and thorough coordination between project developers or
operators and Service and other jurisdictional‐agency employees is crucial to the development of
an effective and successful ECP. Close coordination will also be necessary in the refinement of the
modeling process used to predict fatalities, as well as in post‐construction monitoring to evaluate
those models. We anticipate the ECPG and the recommended methods and metrics will evolve
rapidly as the Service and project developers or operators learn together. The Service has created a
cross‐program, cross‐regional team of biologists who will work jointly on eagle‐programmatic‐take
permit applications to help ensure consistency in administration and application of the Eagle
Permit Rule. This close coordination and interaction is especially important as the Service
processes the first few programmatic eagle take permit applications.

The Service will continue to refine this ECPG with input from all stakeholders with the objective of
maintaining stable or increasing breeding populations of both bald and golden eagles while
simultaneously developing science‐based eagle‐take regulations and procedures that are
appropriate to the risk associated with each wind energy project. As the ECPG evolves, the Service
will not expect project developers or operators to retroactively redo analyses or surveys using the
new approaches. The adaptive approach to the ECPG should not deter project developers or
operators from using it immediately.

33

INFORMATION COLLECTION

The Bald and Golden Eagle Protection Act authorizes us to collect information in order to issue
permits for eagle take. The Eagle Conservation Plan Guidance defines and clarifies the information
required for a permit application (FWS Form 3‐200‐71) and the associated annual report (FWS
Form 3‐202‐15). We use the collected information to evaluate whether the take is compatible with
the preservation of the eagle; to determine if take is likely and how it can be avoided and
minimized; to determine if the applicant will take reasonable measures to minimize the take; and to
assess how the activity actually affects eagles in order to adjust mitigation measures for that project
and for future permits.

We may not conduct or sponsor, nor are you required to respond, to a collection of information
unless it displays a currently valid Office of Management and Budget control number. The burden
for the information collection associated with eagle permits and reports is approved under OMB
Control No. 1018‐0022 (Federal Fish and Wildlife Permit Applications and Reports‐‐Migratory
Birds and Eagles) and OMB Control No. 1018‐0148 (Land‐Based Wind Energy Guidelines).

34

GLOSSARY

Active nest – see occupied nest.
Adaptive resource management – an iterative decision process that promotes flexible decision‐
making that can be adjusted in the face of uncertainties as outcomes from management actions
and other events become better understood.
Advanced conservation practices (ACP) – means scientifically supportable measures that are
approved by the Service and represent the best available techniques to reduce eagle
disturbance and ongoing mortalities to a level where remaining take is unavoidable. ACPs are a
special subset of conservation measures that must be implemented where they are applicable.
Adult – an eagle five or more years of age.
Alternate nests – additional sites within a nesting territory that are available to be used.
Avoidance and minimization measures – conservation actions targeted to remove or reduce
specific risk factors (e.g., avoiding important eagle use areas and migration concentration sites,
placing turbines away from ridgelines). A subset of conservation measures.
Benchmark – an eagle harvest rate at the local‐area population scale that should trigger
heightened scrutiny.
Breeding territory – equivalent to eagle territory.
Calculated take thresholds – annual allowable eagle take limits established in USFWS (2009b).
Collision probability (risk) – the probability that an eagle will collide with a turbine given
exposure.
Compensatory mitigation – replacement of project‐induced losses to fish and wildlife resources.
Substitution or offsetting of fish and wildlife resource losses with resources considered to be of
equivalent biological value. In the case of an the ECPG, an action in the eagle permitting process
that offsets the predicted take of eagles if ACPs and other conservation measures do not
completely remove the potential for take, and projected take exceeds calculated take thresholds
for the species or the eagle management unit affected (or in some cases, under other
circumstances as described in USFWS 2009a).
Conservation measures – actions that avoid (this is best achieved at the siting stage), minimize,
rectify, reduce, eliminate, or mitigate an effect over time. ACPs are conservation measures that
have scientific support and which must be implemented where they are applicable.
Discount rate – the interest rate used in calculating the present value of expected yearly benefits
and costs.
Disturb ‐ means to agitate or bother a bald or golden eagle to a degree that causes, or is likely to
cause, based on the best scientific information available, (1) injury to an eagle, (2) a decrease in
its productivity, by substantially interfering with normal breeding, feeding, or sheltering
behavior, or (3) nest abandonment, by substantially interfering with normal breeding, feeding,
or sheltering behavior.
Eagle Conservation Plans (ECP) – a document produced by the project developer or operator in
coordination with the Service that supports issuance of an eagle take permit under 50 CFR
22.26 and potentially 22.27 (or demonstrates that such a permit is unnecessary).
Eagle Management Unit – regional eagle populations defined in the FEA (USFWS 2009b). For
golden eagles, eagle management units follow Bird Conservation Regions (Figure 2), whereas
bald eagle management units largely follow Service regional boundaries (Figure 3).
Eagle exposure rate – Eagle‐minutes flying within the project footprint (in proximity to turbine
hazards) per hour (hr) per kilometer2 (km2).
Eagle nest (or nest) – any readily identifiable structure built, maintained or used by bald eagles or
golden eagles for the purposes of reproduction (as defined in 50 CFR 22.3).

35
Eagle territory – an area that contains, or historically contained, one or more nests within the
home range of a mated pair of eagles (from the regulatory definition of “territory” at 50 CFR
22.3). “Historical” is defined here as at least the previous 5 years.
Experimental ACPs – prospective conservation measures identified at the start of a programmatic
eagle take permit that are not implemented immediately, but are deferred pending the results
of post‐construction monitoring. If such monitoring indicates the measures might reduce
observed eagle fatalities, they should be implemented and monitored for a sufficient period of
time to determine their effectiveness.
Fatality monitoring – searching for eagle carcasses beneath turbines and other facilities to
estimate the number of fatalities.
Fatality rate – (1) in fatality prediction models, the fatality rate is the number of eagle fatalities per
hr per km 2 ; (2) elsewhere in the ECPG it is the number of eagles taken or predicted to be taken
per year.
Floater (floating adult) – an adult eagle that has not settled on a breeding territory.
Hazardous area – Rotor‐swept area around a turbine or proposed turbine (km2).
Home range – the area traveled by and eagle in its normal activities of food gathering, mating, and
caring for young. Breeding home range is the home range during the breeding season, and the
non‐breeding home range is the home range outside the breeding season.
Important eagle‐use area – an eagle nest, foraging area, or communal roost site that eagles rely on
for breeding, sheltering, or feeding, and the landscape features surrounding such nest, foraging
area, or roost site that are essential for the continued viability of the site for breeding, feeding,
or sheltering eagles (as defined at 50 CFR 22.26).
Inactive nest – a bald eagle or golden eagle nest that is not currently being used by eagles as
determined by the continuing absence of any adult, egg, or dependent young at the nest for at
least 10 consecutive days immediately prior to, and including, at present. An inactive nest may
become active again and remains protected under the Eagle Act.
Inventory – systematic observations of the numbers, locations, and distribution of eagles and eagle
resources such as suitable habitat and prey in an area.
Jurisdictional agency – a government agency with jurisdictional authority to regulate an activity
(e.g., a state or tribal fish and wildlife agency, a state or federal natural resource agency, etc.).
Juvenile – an eagle less than one year old.
Kiting – stationary or near‐stationary hovering by a raptor, usually while searching for prey.
Local‐area population – is as defined in USFWS (2009b), and refers to the eagle population within
a distance from the project footprint equal to the species median natal‐dispersal distance (43
miles for bald eagles and 140 miles for golden eagles).
Mean inter‐nest distance – the mean nearest‐neighbor distance between simultaneously occupied
eagle nests.
Meteorological towers (met towers) – towers erected to measure meteorological events such as
wind speed, direction, air temperature, etc.
Migration concentration sites – places where geographic features (e.g., north‐south oriented
ridgelines, peninsulas) funnel migrating eagles, resulting in concentrated use during migration
periods.
Migration corridors – the routes or areas where eagles may concentrate during migration (e.g.,
funneling areas along ridgetops, at tips of peninsulas) as a result of the interplay between
weather variables and topography.
Migration counts – standardized counts that can be used to determine relative numbers of diurnal
raptors passing over an established point during fall or spring migration.
Mitigation – avoidance, minimization, rectification, reduction over time, and compensation for
negative impacts to bald eagles and golden eagles from the permitted actions. In the ECPG, we

36
use the term compensatory mitigation to describe the subset of mitigation actions designed to
offset take to achieve the no‐net‐loss standard.
Monitoring – (1) a process of project oversight such as checking to see if activities were conducted
as agreed or required; (2) making measurements of uncontrolled events at one or more points
in space or time with space and time being the only experimental variable or treatment; (3)
making measurements and evaluations through time that are done for a specific purpose, such
as to check status and/or trends or the progress towards a management objective.
No‐net‐loss – no net change in the overall eagle population mortality or natality rate after issuance
of a permit that authorizes take, because compensatory mitigation reduces another form of
mortality, or increases natality, by a comparable amount.
Occupied nest – a nest used for breeding in the current year by a pair of eagles. Presence of an
adult, eggs, or young, freshly molted feathers or plucked down, or current year’s mutes
(whitewash) suggest site occupancy. In years when food resources are scarce, it is not
uncommon for a pair of eagles to occupy a nest yet never lay eggs; such nests are considered
occupied.
Occupied territory – an area that encompasses a nest or nests or potential nest sites and is
defended by a mated pair of eagles.
Operational adjustments – modifications made to an existing wind project that changes how that
project operates (e.g., increasing turbine cut in speeds, implementing curtailment of turbines
during periods of high eagle use).
Posterior distribution (Bayesian) – a distribution that quantifies the uncertainty in the model
parameters after incorporating the observed data. The distributions are usually summarized by
intervals around the median.
Present value – within the context of a Resource Equivalency Analysis (REA), refers to the value of
debits and credits based on an assumed annual discount rate (3%). This term is commonly
used in economics and implies that resources lost or gained in the future are of less value to us
today.
Prior distribution (Bayesian) – a distribution that quantifies the uncertainty in the model
parameters from previous data or past knowledge. A non‐informative prior can be used to
imply that little or nothing is known about the parameters.
Programmatic take – take that is recurring, is not caused solely by indirect effects, and that occurs
over the long term or in a location or locations that cannot be specifically identified (as defined
in 50 CFR 22.3).
Project area – the area that includes the project footprint as well as contiguous land that shares
relevant characteristics. For eagle‐take considerations, the Service recommends the project
area be either project footprint and a surrounding perimeter equal to the mean species‐specific
inter‐nest distance for eagles locally, or the project footprint and a 10‐mile perimeter.
Project‐area inter‐nest distance – the mean nearest‐neighbor distance between simultaneously
occupied eagle nests of a species (including occupied nests in years where no eggs are laid). We
recommend calculating this metric from the nesting territory survey in Stage 2, using all nesting
territories within the project area, ideally over multiple years.
Project‐area nesting population – number of pairs of eagles nesting within the project area.
Project‐area eagle population – the population of eagles, considering breeding, migrating, and
wintering eagles, within the project area.
Project footprint – the minimum‐convex polygon that encompasses the wind‐project area
inclusive of the hazardous area around all turbines and any associated utility infrastructure,
roads, etc.
Project developer or operator – any developer or operator that proposes to construct a wind
project.

37
Productivity ─ the number of juveniles ledged from an occupied nest, often reported as a mean
over a sample of nests.
Renewable energy – energy produced by solar, wind, geothermal or any other methods that do not
require fossil fuels.
Resource Equivalency Analysis (REA) – in the context of the ECPG, a methodology used to
compare the injury to or loss of eagles caused by wind facilities (debit) to the benefits from
projects designed to improve eagle survival or increase productivity (credits). Compensation is
evaluated in terms of eagles and their associated services instead of by monetary valuation
methods.
Retrofit – any activity that results in the modification of an existing power line structure to make it
bird safe.
Risk‐averse – a conservative estimate in the face of considerable uncertainty. For example, the
Service typically will use the upper 80% credible interval of the median estimated number of
annual eagle fatalities for permit decisions in an effort to avoid underestimating fatality rates at
wind projects.
Risk validation – as part of Stage 5 assessment, where post‐construction surveys are conducted to
generate empirical data for comparison with the pre‐construction risk assessment predictions
to validate if the initial assumptions were correct.
Roosting – activity where eagles seek cover, usually during night or periods of severe weather (e.g.,
cold, wind, snow). Roosts are usually found in protected areas, typically tree rows or trees
along a river corridor.
Seasonal concentration areas – areas used by concentrations of eagles seasonally, usually
proximate to a rich prey source.
Site categorization – a standardized approach to categorize the likelihood that a site or
operational alternative will meet standards in 50 CFR 22.26 for issuance of a programmatic
eagle take permit.
Stopover sites – areas temporarily used by eagles to rest, seek forage, or cover on their migration
routes.
Subadult – an eagle between 1 and 4 years old, typically not of reproductive age.
Survey –combined inventory and monitoring.
Take threshold – an upper limit on the annual eagle harvest rate for each species‐specific eagle
management unit. Thresholds were set in the Final Environmental Assessment on the Eagle
Permit Rule (USFWS 2009b).
Territory – area that contains, or historically contained, one or more nests within the home range
of a mated pair of eagles (from 50 CFR 22.3).
Unoccupied nest – those nests not selected by raptors for use in the current nesting season. See
also inactive nest.
U.S. Fish and Wildlife Service Draft Land‐based Wind Energy Guidelines (WEG) – a document
that describes a multi‐tiered process to site, construct, operate and monitor wind facilities in
ways that avoid, minimize, and mitigate impacts to wildlife.
Wind facilities – developments for the generation of electricity from wind turbines.
Wind project – developments for the generation of electricity from wind turbines.
Wind turbine – a machine for converting the kinetic energy in wind into mechanical energy, which
is then converted to electricity.

38

Figure 2. Map of golden eagle management units, from USFWS (2009b).

39

Figure 3. Map of bald eagle management units, from USFWS (2009b).

40

LITERATURE CITED

Anthony, R. G. 2001. Low productivity of bald eagles on Prince of Wales Island, Southeast Alaska.
Raptor Research 35:1‐8.
Arnett, E. B. 2006. A preliminary evaluation on the use of dogs to recover bat fatalities at wind
energy facilities. Wildlife Society Bulletin 34(5):1440–1445.
Arnett, E. B., D. B. Inkley, D. H. Johnson, R. P. Larkin, S. Manes, A. M. Manville, R. Mason, M. Morrison,
M. D. Strickland, and R. Thresher. 2007. Impacts of wind energy facilities on wildlife and
wildlife habitat. Technical Review 07‐2, The Wildlife Society, Bethesda, Maryland, USA.
Barclay, R. M. R., E. F. Baerwald, and J. C. Gruver. 2007. Variation in bat and bird fatalities at wind
energy facilities: assessing the effects of rotor size and tower height. Canadian Journal of
Zoology 85: 381–387.
Barrios, L. and A. Rodriguez. 2004. Behavioural and environmental correlates of soaring‐bird
mortality at on‐shore wind turbines. Journal of Applied Ecology 41:72‐81.
Buckland, S. T., D. R. Anderson, K. P. Burnham, J. L. Laake, D. L. Borchers, and L. Thomas. 2001.
Introduction to distance sampling. Oxford University Press, New York, New York, USA.
Buehler, D. A. 2000. Bald eagle (Haliaeetus leucocephalus). The Birds of North America no. 506 (A.
Poole, ed.). The Birds of North America Online, Cornell Lab of Ornithology, Ithaca, New York,
USA. http://bna.birds.cornell.edu/bna/species/506.
Chamberlain, D. E., M. R. Rehfisch, A. D. Fox, M. Desholm, and S. J. Anthony. 2006. The effect of
avoidance rates on bird mortality predictions made by wind turbine collision risk models. Ibis
148:198‐202.
Cole, S. 2009. How much is enough? Determining adequate levels of environmental compensation
for wind power impacts using equivalency analysis. In European Offshore Wind Conference
2009, 14‐16 September 2009, Stockholm, Sweden.
Collopy, M. W., and T. C. Edwards, Jr. 1989. Territory size, activity budget, and role of undulating
flight in nesting golden eagles. Journal of Field Ornithology 60:43‐51.
Craig, T. H., and E. H. Craig. 1984. A large concentration of roosting golden eagles in southeastern
Idaho. Auk 101:610‐613.
Craig, T. H., E. H. Craig, and L. R. Powers. 1984. Recent changes in eagle and buteo densities in
southeastern Idaho. Murrelet 65:91‐93.
De Lucas, M., G. F. E. Janss, D. P. Whitfield and M. Ferrer. 2008. Collision fatality of raptors in wind
farms does not depend on raptor abundance. Journal of Applied Ecology 45:1695–1703.
Ferrer, M. de Lucas, G. F. E. Janss, E. Casado, A. R. Munõz, M. J. Bechard, and C. P. Calabuig. 2011.
Weak relationship between risk assessment studies and recorded mortality in wind farms.
Journal of Applied Ecology doi: 10.1111/j.1365‐2664.2011.02054.x.
Fuller, M. R., J. J. Millspaugh, K. Church, and R. Kenward. 2005, Wildlife radiotelemetry. Pages 377‐
417 in C. E. Braun (ed.), Techniques for wildlife investigations and management, 6th edition. The
Wildlife Society, Bethesda, Maryland, USA.
Garrett, M. G., J. W. Watson, and R. G. Anthony. 1993. Bald eagle home range and habitat use in the
Columbia River estuary. Journal of Wildlife Management 57:19‐27.
Gregory, M. J. P, A. G. Gordon, and R. Moss. 2002. Impact of nest‐trapping and radio‐tagging on
breeding golden eagles Aquila chrysaetos in Argyll, Scotland. Ibis 145:113‐119.
Harmata, A. R. 1982. What is the function of undulating flight display in golden eagles? Raptor
Research 16:103‐109.
Harmata, A.R. 2002. Vernal migration of bald eagles from a southern Colorado wintering area.
Journal of Raptor Research 36:256‐264.

41
Hodges, J. I. and F. C. Robards. 1982. Observations of 3,850 bald eagle nests in southeast Alaska.
Pages 37‐46 in A symposium and workshop on raptor management and biology in Alaska and
western Canada. (W. N. Ladd and P. F. Schempf, eds.) U.S. Fish and Wildlife Service, Anchorage,
Alaska, USA.
Hoechlin, D. R. 1976. Development of golden eagles in southern California. Western Birds 7:137‐
152.
Hoover, S. L., and M. L. Morrison. 2005. Behavior of red‐tailed hawks in a wind turbine
development. Journal of Wildlife Management 69:150‐159.
Hunt, W. G. 1998. Raptor floaters at Moffat’s equilibrium. Oikos 82:191‐197.
Hunt, W. G., R. E. Jackman, T. L. Brown, and L. Culp. 1999. A population study of golden eagles in
the Altamont Pass Wind Resource Area: population trend analysis 1994‐1997. Report to
National Renewable Energy Laboratory, Subcontracts XAT‐5‐15174‐01, XAT‐6‐16459‐01.
Predatory Bird Research Group, University of California, Santa Cruz, California, USA.
Hunt, G. 2002. Golden eagles in a perilous landscape: predicting the effects of mitigation for wind
turbine bladestrike mortality. California Energy Commission Report P500‐02‐043F.
Sacramento, California, USA.
Huso, M. M. P. 2010. An estimator of wildlife fatality from observed carcasses. Environmetrics DOI:
10.1002/env.1052.
Kenward, R. E. 2001. A manual for wildlife tagging. Academic Press, London, UK.
Kerlinger, P. 1989. Flight strategies of migrating hawks. University of Chicago Press, Chicago,
Illinois, USA.
Kochert, M. N. 1972. Population status and chemical contamination in golden eagles in
southwestern Idaho. M.S. Thesis. University of Idaho, Moscow, Idaho, USA.
Kochert, M. N., K. Steenhof, C. L. Mcintyre, and E. H. Craig. 2002. Golden eagle (Aquila chrysaetos).
The Birds of North America No. 684 (A. Poole, Ed.). The Birds of North America Online. Cornell
Lab of Ornithology, Ithaca, New York, USA. http://bna.birds.cornell.edu/bna/species/684.
Krone, O. 2003. Two white‐tailed sea eagles (Haliaeetus albicilla) collide with wind generators in
northern Germany. Journal of Raptor Research 37:174‐176.
Kunz, T. H., E. B. Arnett, B. M. Cooper, W. P. Erickson, R. P. Larkin, T. Mabee, M. L. Morrison, M. D.
Strickland, and J. M. Szewczak. 2007. Assessing impacts of wind‐energy development on
nocturnally active birds and bats: a guidance document. Journal of Wildlife Management 71:
2449‐2486.
Kuvlesky, W. P. Jr., L. A. Brennan, M. L. Morrison, K. K. Boydston, B. M. Ballard, and F. C. Bryant.
2007. Wind energy development and wildlife conservation: challenges and opportunities.
Journal of Wildlife Management 71:2487‐2498.
Marzluff, J. M., M. S. Vekasy, M. N. Kochert, and K. Steenhof. 1997. Productivity of golden eagles
wearing backpack radiotransmitters. Journal of Raptor Research 31:223‐227.
McGrady, M. J., J. R. Grant, I. P. Bainbridge, and D. R. A. McLeod. 2002. A model of golden eagle
(Aquila chrysaetos) ranging behavior. Journal of Raptor Research 36(1) supplement:62‐69.
McLeod, D. R. A, D. P. Whitfield, and M. J. McGrady. 2002. Improving prediction of golden eagle
(Aquila chrysaetos) ranging in western Scotland using GIS and terrain modeling. Journal of
Raptor Research 36(1) Supplement:70‐77.
Millsap, B. A. 1981. Distributional status and ecology of Falconiformes in west central Arizona,
with notes on ecology, reproductive success, and management. U. S. Bureau of Land
Management Technical Note 355.
Mohr, C. O. 1947. Table of equivalent populations of North American small mammals. American
Midland Naturalist 37:223–249.
Mojica, E. K., J. M. Meyers, B. A. Millsap, and K. T. Haley. 2008. Migration of sub‐adult Florida bald
eagles. Wilson Journal of Ornithology 120:304‐310.

42
Moorcroft, P. R., M. A. Lewis, and R. L. Crabtree. 1999. Home range analysis using a mechanistic
home range model. Ecology 80:1656‐1665.
Newton, I. 1979. Population ecology of raptors. Buteo Books, Vermillion, South Dakota, USA.
Nygård, T, K., E. L. Bevanger, Ø. Dahl, A. Flagstad, P. L. Follestad, Hoel, R. May, and O. Reitan. 2010.
A study of white‐tailed eagle Haliaeetus albicilla movements and mortality at a wind farm in
Norway. in D. Senapathi (ed.), Climate change and birds: adaptation, mitigation and impacts on
avian populations. British Ornithologists’ Union Proceedings, Peterborough, United Kingdom.
http://www.bou.org.uk/ bouproc‐net/ccb/nygard‐etal.pdf (last visited 9 October 2011).
Palmer, R. S. 1988. Golden eagle Aquila chrysaetos. Pages 180‐231 in R. S. Palmer (ed.), Handbook
of North American birds, volume 5, diurnal raptors (part 2). Yale University Press, New Haven,
Connecticut, USA.
Platt, J. B. 1976. Bald eagles wintering in a Utah desert. American Birds 30:783‐788.
Phillips, R. L. and A. E. Beske. 1984. Resolving conflicts between energy development and nesting
golden eagles. Pages 214‐219 in R. D. Comer, J. M. Merino, J. W. Monarch, C. Pustmueller, M.
Stalmaster, R. Stoecker, J. Todd, and W. Wright (eds.). Proceedings of asymposium: Issues and
technology in the management of impacted western wildlife. Steamboat Springs, Colorado,
November 15‐17, 1982. Thorne Ecological Institute, Boulder, Colorado, USA.
Restani, M., A. R. Harmata, and E. M. Madden. 2000. Numerical and functional responses of migrant
bald eagles exploiting a seasonally concentrated food source. Condor 102:561‐568.
Seaman, D. E., and R. A. Powell. 1996. An evaluation of the accuracy of kernel density estimators for
home range analysis. Ecology 77:2075–2085.
Sherrod, S. K., C. M. White, and F. S. L. Williamson. 1976. Biology of the bald eagle on Amchitka
Island, Alaska. Living Bird 15:145‐182.
Smallwood, K. S. and C. G. Thelander. 2004. Developing methods to reduce bird fatalities in the
Altamont Wind Resource Area. Final report prepared by BioResource Consultants to the
California Energy Commission, Public Interest Energy Research‐Environmental Area, Contract
No. 500‐01‐019.
Smallwood, K. S., L. Rugge, M. L. Morrison. 2009. Influence of behavior on bird mortality in wind
energy developments. Journal of Wildlife Management 73:1082–1098.
Soutullo, A., V. Urios, M. Ferrer, and S. G. Penarrubia. 2006. Post‐fledging behaviour in golden
eagles Aquila chrysaetos: onset of juvenile dispersal and progressive distancing from the nest.
Ibis 148:307‐312.
Strickland, M.D., E.B. Arnett, W.P. Erickson, D.H. Johnson, G.D. Johnson, M.L., Morrison, J.A. Shaffer,
and W. Warren‐Hicks. 2011. Comprehensive Guide to Studying Wind Energy/Wildlife
Interactions. Prepared for the National Wind Coordinating Collaborative, Washington, D.C., USA.
Thorstrom, R. 2001. Nest‐site characteristics and breeding density of two sympatric forest‐falcons
in Guatemala. Ornitologia Neotropical 12:337‐343.
USFWS. 1983. Northern states bald eagle recovery plan. U.S. Fish and Wildlife Service, Division of
Migratory Bird Management, Washington D.C., USA.
USFWS. 2007. Protection of eagles; definition of “Disturb.” Federal Register 72(107):31132‐31140.
USFWS. 2009a. Eagle permits; take necessary to protect interests in particular localities. Federal
Register 74(175):46836‐46879.
USFWS. 2009b. Final environmental assessment. Proposal to permit take provided under the Bald
and Golden Eagle Protection Act. U.S. Fish and Wildlife Service, Division of Migratory Bird
Management, Washington D.C., USA.
Walker, D., M. McGrady, A. McCluskie, M. Madders, and D. R. A. McLeod. 2005. Resident golden
eagle ranging behaviour before and after construction of a windfarm in Argyll. Scottish Birds
25:24‐40.
Watson, J. W., and R. W. Davies. 2009. Range use and contaminants of golden eagles in Washington.
Progress Report. Washington Department of Fish and Wildlife, Olympia, Washington.

43
Whitfield, D. P., A. H. Fielding, M. J. P. Gregory, A. G. Gordon, D. R. A.McLeod, and P. F. Haworth.
2007. Complex effects of habitat loss on golden eagles Aquila chrysaetos. Ibis 149:26–36.
Wichmann, M. C., F. Jeltsch, W. R. J. Dean, K. A. Moloney and C. Wissel. 2003. Implication of climate
change for the persistence of raptors in arid savanna. Oikos 102:186–202.
Williams, B. K., R. C. Szaro, and C. D. Shapiro. 2007. Adaptive Management: The U.S. Department of
the Interior Technical Guide. Adaptive Management Working Group, U.S. Department of the
Interior, Washington, DC, USA.
Worton, B. J. 1989. Kernel methods for estimating the utilization distribution in home‐range
studies. Ecology 70:164–168.
Yates, R. E., B. R. McClelland, P. T. McClelland, C. H. Key, and R. E. Bennetts. 2001. The influence of
weather on golden eagle migration in northwestern Montana. Journal of Raptor Research
35:81‐90.

44

APPENDIX A: ADAPTIVE MANAGEMENT

Management of wind facilities to minimize eagle take through decisions about siting, design,
operation, and compensatory mitigation, is a set of recurrent decisions made in the face of
uncertainty. The Department of the Interior has a long history of approaching such decisions
through a process of adaptive management (Williams et al. 2007). The purpose of adaptive
management is to improve long‐term management outcomes, by recognizing where key
uncertainties impede decision making, seeking to reduce those uncertainties over time, and
applying that learning to subsequent decisions (Walters 1986).

Adaptive management is a special case of decision analysis applied to recurrent decisions (Lyons et
al. 2008). Like all formal decision analysis, it begins with the identification of fundamental
objectives—the long‐term ends sought through the decision (step 2, Fig. A‐1). These objectives are
the primary concern, and all the other elements are designed around them. With these objectives in
mind, alternative actions are considered, and the consequences of these alternatives are evaluated
with regard to how well they might achieve the objectives. But in many decisions, there is critical
uncertainty that impedes the decision (step 6, Fig. A‐1), that is, the decision‐maker is missing
knowledge that affects which alternative might be best. In recurrent decisions, there exists the
opportunity to reduce that uncertainty, by monitoring the outcomes of early actions, and apply that
learning to later actions. It is valuable to note that learning is not pursued for its own sake, but only
insofar as it helps improve long‐term management by reducing these uncertainties.

There are two hallmarks of a formal interpretation of adaptive management, like that described
above. The first hallmark is the a priori identification of the critical uncertainty. In this way,
adaptive management is not a blind search for some unspecified new insights, but a focused effort
to reduce the uncertainty that stands in the way of better decision‐making. The second hallmark is
that the means of adaptation is clear, that is, the way in which new information will be applied to
subsequent decisions is articulated.

There is, however, recognition that unanticipated learning does occur in any real system, and this
learning can sometimes lead to valuable insights. In so‐called “double‐loop learning” (Argyris and
Shon 1978), the learning might even lead to a re‐framing of the decision, a re‐examination of the
objectives, or consideration of new alternatives (this could be represented by a loop from step 7 to
step 1 in Fig. A‐1). In the context of eagle management at wind facilities, the Service’s focus is on
the inner‐loop learning (represented by the feedback from step 7 to 8 to 4 in Fig. A‐1), but
unanticipated learning will not be ignored.

In the case of managing eagle populations in the face of energy development, there is considerable
uncertainty to be reduced. For example, we believe that in some areas or specific situations, large
soaring birds, specifically raptors, might be especially vulnerable to colliding with wind turbines
(Barrios and Rodriguez 2004, Kuvlesky et al. 2007), but we are uncertain about the relative
importance of factors that influence that risk. We are also uncertain about the best way to mitigate
the effects of wind turbine developments on raptors; we suspect some strategies might be effective,
others are worth trying. We also suspect that a few species, including golden eagles (USFWS 2009),
may be susceptible enough to collisions with wind turbines that populations may be negatively
affected. Thus, there are uncertainties at several levels that challenge our attempts to manage eagle
populations: (1) at the level of understanding factors that affect collision risk, (2) at the level that
influences population trends, and (3) about the efficacy of various mitigation options. The Service,
our conservation partners, and industry will never have the luxury of perfect information before
needing to act to manage eagles. We are therefore left to make management decisions based on the

45
best available information with some inherent degree of uncertainty about the outcomes of those
decisions. Our goal is to reduce that uncertainty through use of formal adaptive management,
thereby improving our predictive capability over time. Applying a systematic, cohesive, nationally‐
consistent strategy of management and monitoring is necessary to accomplish this goal.

1

2

Problem
Framing

Elicit
Objectives

4

Update
Predictive
Models (Learn)

Develop
Alternatives

5,9 Identify

Evaluate
Consequences

8

3

Preferred
Alternative

6 Evaluate
Critical
Uncertainty

Implement
Action

7
Monitor

Figure A-1: A framework for adaptive resource management (ARM). At the core of adaptive
management is critical uncertainty that impedes the identification of a preferred alternative. When
decisions are recurrent, implementation coupled with monitoring can resolve uncertainty, and allow
future decisions to reflect that learning. (Figure from Runge 2011).

1. Adaptive Management as a Tool
Using adaptive management as a tool to manage wildlife populations is not new to the Service. We
and other agencies are increasingly using the principles of adaptive management across a range of
programs, including waterfowl harvest management (Johnson et al. 1997), endangered species
(Runge 2011), and habitat management at local and landscape scales (Lyons et al. 2008). Applying
adaptive management to complex resource management issues is promoted throughout the
Department of the Interior (Williams et al. 2007).

46

Waterfowl harvest management is the classic example of adaptive resource management. Hunting
regulations are reset each year in the United States and Canada through the application of adaptive
management principles (Johnson et al. 1997). A key uncertainty in waterfowl management is the
extent to which harvest mortality is compensated by reductions in non‐harvest mortality or by
increases in productivity (Williams et al. 1996). Various population models have been built based
on competing hypotheses to answer this question; these competing models make different
predictions about how the population will respond to hunting. Every year the Service and the
Canadian Wildlife Service monitor waterfowl and environmental conditions to estimate population
size, survival rates, productivity, and hunting rates. These data feed into the various competing
models, and the models are evaluated annually based on how well they predict changes in
waterfowl populations. Models that perform best year‐after‐year accrue increasing weight (i.e.,
evidence in support of the underlying hypothesis). Weighted model outputs directly lead to
recommended sets of hunting regulations (e.g., bag limits and season lengths) for the subsequent
year. Over time, by monitoring the population effects of various harvest rates on survivorship, and
environmental conditions on productivity, our uncertainty about the degree to which harvest is
compensated by other factors has been reduced, allowing for the setting of harvest rates with
greater confidence every year. The application of adaptive management principles to waterfowl
harvest regulation has helped the Service and its partners achieve or exceed population goals for
most species of waterfowl (NAWMP 2004).

Adaptive management is a central component of the Service’s approach to collaborative
management at the landscape scale, through strategic habitat conservation (NEAT 2006). The
principles of adaptive management are also embedded in endangered species management (Ruhl
2004, Runge 2011), including in recovery planning (Smith 2011) and habitat conservation planning
(Wilhere 2002). Indeed, the Service recognizes that adaptive management is a normative concept
in modern ecological decision‐making (Callicott et al. 1999), and embraces it as a fundamental tool.

2. Applying Adaptive Management to Eagle Take Permitting
In the context of wind energy development and eagle management under the ECPG, there are four
specific sets of decisions that are suitable for an adaptive management approach.

a. Adaptive Management of Wind Project Operations
The most immediate and direct opportunity for adaptive management is at the site‐level for
wind facilities after construction. The relevant uncertainty is in the predictions of eagle
take at the project, and the operational factors that influence the level of take. The role of
adaptive management at this scale will be analyzed and evaluated in the NEPA associated
with each permit. Under the ECPG, a wind project would initially work with the Service to
generate predictions of take, given the siting, design, and operational parameters of the
project. These predictions are made under uncertainty, and the risk to eagles associated
with this uncertainty is factored into the compensatory mitigation terms of the permit
under BGEPA. After a site becomes operational, ongoing surveys of realized take can be
compared to the predictions of take. At the review points of the permit (typically, every five
years), the Service and the operator will review the observed take. If the observed take
exceeds the predicted and permitted take, the Service will work with the operator to
identify measures that could be taken to reduce the take below the permitted threshold
(within the limits jointly agreed to at the outset of the permit period). The monitoring data
may provide clues about how this could be done, for example, by identifying where and
when most of the take is occurring. On the other hand, if the observed take is significantly

47
less than the predicted take, the Service can work with the operator to update the
predictions of take for the next review period, adjust the conditions for compensatory
mitigation, and return credits to the operator for any excess compensatory mitigation.

In a related manner, for both new and existing facilities, ongoing monitoring can provide
information to reduce uncertainty about the effectiveness of conservation measures and
ACPs. In particular, experimental conservation measures and ACPs are actions taken by the
operator that are thought to reduce mortality risk, but there is uncertainty about how
effective some of these measures can be. In the end, the purpose of adaptive management
of operations is to reduce mortality of eagles while also reducing the impact of conservation
measures and ACPs on power generation at wind facilities.

b. Adaptive Management of Wind Project Siting and Design Recommendations
Through the ECPG and the permit review process, the Service makes recommendations to
operators about how to site and design wind facilities to reduce eagle disturbance and
mortality. These recommendations are based on the best available science, but
acknowledge that our understanding of the interaction between eagles and wind facilities is
incomplete. Adaptive management provides the opportunity to respond to increasing
understanding about this interaction.

The particular focus of this layer of adaptive management is the predictions of take that are
made by considering pre‐construction surveys and risk factors (see APPENDIX D). The
proposed models are initially quite coarse in their ability to make predictions, but the
Service, in partnership with the U.S. Geological Survey (USGS), plans to refine these models.
The key uncertainties concern the risk factors that are important in predicting eagle take.
For example, how important is the proximity to nesting sites, prey concentrations, or
ridgelines in determining the risk posed by any wind turbine? Multiple models will be
developed to express uncertainty in these risk factors, and the predictions from these
multiple models will be compared to the patterns of observed take at existing facilities.
Using multiple models to express uncertainty allows inclusion and evaluation of alternative
models from different sources. The learning that emerges will be used to improve the
predictions from the models, which in turn, will allow future recommendations about siting
and design to be enhanced. In this case, the benefit of the monitoring at individual sites
accrues to the wind industry as a whole.

c. Adaptive Management of Compensatory Mitigation
The determination of appropriate levels of compensatory mitigation, such as through a
resource equivalency analysis (REA, see APPENDIX F), is based on two predictions: the level
of take expected at a project; and the amount of mitigation required to offset that take. As
noted above, site‐level learning, through observation of realized take, can be used to update
predictions of take, and compensatory mitigation can be adjusted accordingly. In addition,
the accrued experience across sites, through monitoring of the effectiveness of
compensatory mitigation projects and eagle population responses, can be used to update
the methods and parameters in the REA methods used to determine the appropriate level of
compensatory mitigation.

d. Adaptive Management of Population-Level Take Thresholds
Healthy, robust populations of animals can sustain some degree of incidental take, without
long‐term adverse impacts to the population or the ecosystem. The amount of take that is

48
sustainable and that can be authorized is a function of both scientific factors (e.g., the
intrinsic growth rate and carrying capacity of the population) and policy interpretation (e.g.,
the amount of potential growth that can be allocated to take, and the risk tolerance for
excessive take) (Runge et al. 2009). The capacity to sustain incidental take arises from the
resilience in populations due to the ability to compensate for that take by increasing
survival or reproductive rates.

At the scale of regional populations (e.g., bird conservation regions for golden eagles), the
central question for eagles is not altogether different than it is for waterfowl: to what extent
is mortality from energy development, or any other anthropogenic source, compensated by
reductions in mortality from other sources, or by increases in productivity? These
questions are best answered by building population models founded on competing
hypotheses that incorporate estimates of mortality, productivity, and the variation around
those vital rates. What is needed is a systematic effort to collect information on mortality,
breeding, and population status to feed those models. Similar to waterfowl management,
reducing uncertainty in population‐level models for eagle management will require rolling
up the results of local monitoring and research across the distribution of eagles. The results
will allow the Service to make more informed management recommendations to reach the
Service’s population goal of stable or increasing breeding populations for both eagle
species.

At present, the Service’s regulations call for no increase in net take of golden eagles, under a
protective concern that the current level of take exceeds a sustainable threshold. As our
understanding of golden eagle population size and status increases, and our knowledge of
vital rates and potential resilience improves, the Service and USGS will reanalyze the
potential for instituting take thresholds for golden eagles. Take thresholds for bald eagles
will also be re‐assed no less frequently than every five years (USFWS 2009). If thresholds
for either species are increased and additional take is authorized, continued population
monitoring will be critical in providing feedback on population response (i.e., step 4 to 8 in
Fig. A‐1).

Literature Cited
Argyris, C., and D. Shon. 1978. Organizational Learning: a Theory of Action Learning. Addison‐
Wesley, Reading, Massachusetts.
Barrios, L., and A. Rodriguez. 2004. Behavioural and environmental correlates of soaring‐bird
mortality at on‐shore wind turbines. Journal of Applied Ecology 41:72‐81.
Callicott, J. B., L. B. Crowder, and K. Mumford. 1999. Current normative concepts in conservation.
Conservation Biology 13:22‐35.
Johnson, F. A., C. T. Moore, W. L. Kendall, J. A. Dubovsky, D. F. Caithamer, J. R. Kelley, Jr., and B. K.
Williams. 1997. Uncertainty and the management of mallard harvests. Journal of Wildlife
Management 61:202‐216.
Kuvlesky, W. P., Jr, L. A. Brennan, M. L. Morrison, K. K. Boydston, B. M. Ballard, and F. C. Bryant.
2007. Wind energy development and wildlife conservation: challenges and opportunities. The
Journal of wildlife management 71:2487‐2498.
Lyons, J. E., M. C. Runge, H. P. Laskowski, and W. L. Kendall. 2008. Monitoring in the context of
structured decision‐making and adaptive management. Journal of Wildlife Management
72:1683‐1692.

49
National Environmental Assessment Team [NEAT]. 2006. Strategic Habitat Conservation. U.S. Fish
and Wildlife Service, Arlington, Virginia, USA.
North American Waterfowl Management Plan, Plan Committee [NAWMP]. 2004. North American
Waterfowl Management Plan 2004. Strategic Guidance: Strengthening the Biological
Foundation. Canadian Wildlife Service, U.S. Fish and Wildlife Service, Secretaria de Medio
Ambiente y Recursos Naturales.
Ruhl, J. 2004. Taking adaptive management seriously: A case study of the Endangered Species Act.
University of Kansas Law Review 52:1249‐1284.
Runge, M. C. 2011. Adaptive management for threatened and endangered species. Journal of Fish
and Wildlife Management 2.
Runge, M. C., J. R. Sauer, M. L. Avery, B. F. Blackwell, and M. D. Koneff. 2009. Assessing allowable take
of migratory birds. Journal of Wildlife Management 73:556‐565.
Smith, C. B. 2011. Adaptive management on the central Platte River ‐ Science, engineering, and
decision analysis to assist in the recovery of four species. Journal of Environmental
Management 92:1414‐1419.
USFWS. 2009. Final environmental assessment. Proposal to permit take provided under the Bald
and Golden Eagle Protection Act. U.S. Fish and Wildlife Service, Division of Migratory Bird
Management, Washington D.C., USA.
Walters, C. J. 1986. Adaptive management of renewable resources. Macmillan, New York, New York,
USA.
Wilhere, G. F. 2002. Adaptive management in habitat conservation plans. Conservation Biology
16:20‐29.
Williams, B. K., F. A. Johnson, and K. Wilkins. 1996. Uncertainty and the adaptive management of
waterfowl harvests. The Journal of wildlife management 60:223‐232.
Williams, B. K., R. C. Szaro, and C. D. Shapiro. 2007. Adaptive Management: The U.S. Department of
the Interior Technical Guide. Adaptive Management Working Group, U.S. Department of the
Interior, Washington, DC, USA.

50

APPENDIX B: STAGE 1 – SITE ASSESSMENT

Occurrence of eagles and their use of landscapes vary across broad spatial scales. The first step in
project development is to conduct a landscape‐scale assessment, based mainly on publicly available
information, to identify sites within a large geographic area that have both high potential for wind
energy and low potential for negative impacts on eagles if a project is developed. Stage 1
corresponds to Tiers 1 and 2 of the WEG and, along with Stage 2 herein and Tier 3 in the WEG,
comprise the pre‐construction evaluation of wind energy projects. Depending on the outcome of
Stage 1, developers decide whether to proceed to the next stage, “... requiring a greater investment
in data collection to answer certain questions” (referring to Tier 3, in the WEG; see also Table B‐1).
The WEG should be examined for general considerations relevant to Stage 1; this appendix and the
following APPENDIX C focus on considerations specific to eagles.

The Stage 1 assessment should evaluate wind energy potential within the ecological context of
eagles, including considerations for the eagle’s annual life‐cycle, i.e., breeding, dispersal, migration,
and wintering. The goal at this stage is to determine whether prospective wind project sites are
within areas known or likely to be used by eagles and, if so, begin to determine the relative
spatiotemporal extent and type of eagle use of the sites. Areas used heavily by eagles are likely to
fall into category 1; development in these areas should be avoided because the Service probably
could not issue project developers or operators a programmatic permit for take that complies with
all regulatory requirements. Stage 1 assessment is a relatively straightforward “desktop” process
that probably should conduct before significant financial resources have been committed to
developing a particular project.

Multiple data sources can be consulted when evaluating a prospective site’s value to eagles.
Wildlife biologists and other natural resource professionals from federal agencies including the
Service, and tribal, state, and county agencies should be consulted early in the Stage 1 process to
help ensure all relevant information is being considered. Information mainly encompasses
physiographic and biological factors that could affect eagle risk associated with wind energy
development. Questions generally focus on: (1) recent or historical nesting and seasonal
occurrence data for eagles at the prospective area; (2) migration or other regular movement by
eagles through the area or surrounding landscape; (3) seasonal concentration areas such as a
communal roost site in a mature riparian woodland or a prairie dog (Cynomys spp.) town serving as
a major forage base; and (4) physical features of the landscape, especially topography, that may
attract or concentrate eagles. “Historical” is defined here as 5 or more years; a search for historical
data should encompass at least the previous 5 years. Data from far longer time periods may be
available but should be cautiously scrutinized for confounding factors such as land use change that
diminish the data’s relevance.

Preliminary site evaluation could begin with a review of publically available information, including
resource databases such as NatureServe (http://www.natureserve.org/) and the American Wind
Wildlife Institute’s Landscape Assessment Tool (LAT; http://www.awwi.org/initiatives/
landscape.aspx); information from relevant tribal, state, and federal agencies, including the Service;
state natural heritage databases; state Wildlife Action Plans; raptor migration databases such as
those available through Hawk Migration Association of North America (http://www.hmana.org) or
HawkWatch International (http://www.hawkwatch.org); peer‐reviewed literature and published
technical reports; and geodatabases of land cover, land use, and topography (e.g., the LAT
integrates several key geodatabases). Additional information on a site’s known or potential value
to eagles can be garnered by directly contacting persons with eagle expertise from universities,
conservation organizations, and professional or state ornithological or natural history societies.

51
Some of this wide assortment of desktop information and certain knowledge gaps identified
probably will necessitate validation through site‐level reconnaissance, as suggested in the WEG.

Using these and other data sources, a series of questions should be considered to help place the
prospective project site or alternate sites into an appropriate risk category. Relevant questions
include (modified from the WEG):

1. Does existing or historical information indicate that eagles or eagle habitat (including
breeding, migration, dispersal, and wintering habitats) may be present within the
geographic region under development consideration?
2. Within a prospective project site, are there areas of habitat known to be or potentially
valuable to eagles that would be destroyed or degraded due to the project?
3. Are there important eagle use areas or migration concentration sites documented or
thought to occur in the project area?
4. Does existing or historical information indicate that habitat supporting abundant prey for
eagles may be present within the geographic region under development consideration
(acknowledging, wherever appropriate, that population levels of some prey species such as
black‐tailed jackrabbits (Lepus californicus) cycle dramatically [Gross et al. 1974] such that
they are abundant and attract eagles only in certain years [e.g., Craig et al. 1984])?
5. For a given prospective site, is there potential for significant adverse impacts to eagles
based on answers to above questions and considering the design of the proposed project?

We recommend development of a map that, based on answers to the above questions, indicates
areas that fall under site category 1, i.e., areas where wind energy development would pose
obvious, substantially high risks to eagle populations. Remaining areas could be tentatively
categorized as either moderate to high but mitigable risk or minimal risk to eagle populations
(category 2 or category 3). Prospective sites that fall into category 1 at this point are unlikely
candidates for a programmatic permit for take of eagles, although classification of a site at Stage 1
might be regarded as tentative (see “Assessing Risk and Effects; 4. Site Categorization Based on
Mortality Risk to Eagles” in the ECPG. If a site appears to be a category 1 site based on the outcome
of Stage 1, the developer can decide whether information at that stage adequately supports a
category decision or whether to invest in Stage 2 assessment to clarify preliminary indications of
Stage 1 (Table B‐1). Sites that tentatively fall into categories 2 or 3 at Stage 1 can move on to Stage
2 assessment, but could ultimately be excluded as permit candidates after more site‐specific data
are collected in Stage 2.

Again, the goal of Stage 1 site assessment in this ECPG is to determine whether prospective wind
project sites are within areas known or likely to be used by eagles and, if so, begin to assess the
spatiotemporal extent and type of eagle use the sites receive or are likely to receive. Thus, the
ultimate goal of Stage 1 is to determine whether sites exhibit any obvious substantial risk for eagles.
For those that do not, the Stage 1 site assessment will provide fundamental support for the design
of detailed surveys in Stage 2, decisions which influence optimal allocation of the financial
investment in surveys and quality of data collected. In some situations, the Stage 1 site assessment
may provide enough information to adequately estimate impacts and support decisions on site
categorization (and, where relevant, potential conservation measures and appropriate levels of
compensatory mitigation), rendering Stage 2 assessment unnecessary (Table B‐1).

52

Literature Cited
Craig, T. H., E. H. Craig, and L. R. Powers. 1984. Recent changes in eagle and buteo abundance in
southeastern Idaho. Murrelet 65:91‐93.
Gross, J. E., L. C. Stoddart, and F. H. Wagner. 1974. Demographic analysis of a northern Utah
jackrabbit population. Wildlife Monograph 40.

Table B-1. Framework for decisions on investment at Stage 2 level to address chief information needs.
A bidirectional arrow represents a continuum of conditions.

Area of
Information
Need

Seasonal
abundance

Nesting records

Migration
corridors

Communal
roosts

Prey availability
or foraging
hotspots

Strength of Stage 1 Information Base for Assessing Risk to
Eagles
Robust:
well investigated and supported, at
least semi‐quantitative
documentation from most recent 2‐5
years, encompassing potential site(s)
or adjoining areas from which reliable
inferences can be made

↔

Weak:
characterized by little
supportive information and
marginal certainty overall, at
best only general descriptions,
conjecture, or limited
inferences from other areas or
regions

↔

↔
↔

↔

↔

↔

Uncertain risk level – strong
survey effort at Stage 2 level
advised

Relevant areas of information need
are well‐addressed and risk level is
clearly low – Stage 2 may not be
Outcome and
implications for warranted or else modest or limited‐
focus survey effort at Stage 2 level
additional
recommended
assessment

needs at Stage 2 Relevant areas of information need
are well‐addressed and risk level is
level:
moderate or high – strong effort at
Stage 2 level advised

53

APPENDIX C: STAGE 2 – SITE-SPECIFIC SURVEYS AND ASSESSMENT
1. Surveys of Eagle Use
Information collected in Stage 2 is used mainly to generate predictions of the mean annual number
of eagle fatalities for a prospective wind energy project and to identify important eagle use areas or
migration concentration sites that could be affected by the project. Information from Stage 2 is also
used to assess the likelihood of disturbance take of eagles. An array of survey types could be used
to quantify use by eagles of a proposed project area. This section focuses on four types of surveys
recommended for assessing risk to eagles at proposed wind projects. The first three are surveys of
eagle use within the proposed project footprint. These include: (1) point count surveys, which
mainly generate occurrence data that form underpinnings of the risk assessment model
recommended herein; (2) migration (“hawk watch”) counts, documenting hourly passage rates of
eagles; and (3) utilization distribution (UD) assessment, an accounting of the intensity of use of
various parts of the home range within the project footprint; and (4) surveys of nesting territory
occupancy in the project area. Where uncertainties exist regarding survey methods, our
recommendations tend to be conservative such that biases in survey data, if any, are more likely to
favor greater rather than lower estimates of use and ultimately more rather than less protection for
eagles. This approach is consistent with the Service’s policy of taking a risk‐averse stance in the
face of existing uncertainty with respect to eagle programmatic take permits.

In addition to fatality estimation and informing a site categorization decision, Stage 2 studies of
eagles should help answer the following questions (modified from the WEG):
1. What is the distribution, relative abundance, behavior, and site use of eagles and to what
extent do these factors expose eagles to risk from the proposed wind energy project?
2. What are the potential risks of adverse impacts of the proposed wind energy project to
individual and local populations of eagles and their habitats?
3. How can developers avoid, minimize, and mitigate identified adverse impacts?
4. Are there studies that should be initiated at this stage that would be continued in post‐
construction?

a. Point Count Surveys
Point counts (i.e., circular‐plot surveys) often are used to assess relative abundance,
population trends, and habitat preferences of birds (Johnson 1995). The Service advocates
use of point count surveys as the means of providing primary input for models predicting
fatality rate of eagles associated with wind turbines. However, we acknowledge the term
point count survey does not accurately describe the approach we advocate for collecting
data to support fatality rate estimation at wind energy projects. The Service’s approach in
this regard is point‐based recording of activity duration (minutes of flight) within a three‐
dimensional plot. In contrast, point count surveys, as typically conducted, yield indices of
relative abundance or frequency of occurrence (in addition to trend, density estimation, and
habitat association, depending on how data are collected; Ralph et al. 1993). With that said,
most records of eagle flight duration are likely to be classified as 1 minute, per the approach
recommended in this section, and as such resemble records of occurrence for data from
point count surveys. Although a bit of a misnomer in this regard, “point count survey” is
applied broadly herein to include both point‐based records of flight time and traditional
point count surveys because sampling frameworks for each so closely overlap and both data
types can be gathered simultaneously, along with other information described in this
appendix. There may be other means of generating count data to support the fatality model

54
described in this document. Consideration of alternative approaches for predicting fatality
at such projects may require greater time and additional reviews.

The general approach for conducting a fixed‐radius point count survey is to travel to a pre‐
determined point on the landscape and record individual birds detected – whether
observed, only heard, or both observed and heard – within a circular plot, the boundary of
which is at a fixed distance from the point and is marked in the field in several places (Hutto
et al. 1986, Ralph et al. 1993). In addition to plot radius, the survey is standardized by count
duration. Sometimes a variable‐radius plot method (Reynolds et al. 1980) is used, yielding
species‐by‐species detectability coefficients to appropriately bound the plot radius (i.e.,
sampling area) for each species. A variety of point count survey methods have been used
specifically for raptors (reviewed in Anderson [2007]; the North American Breeding Bird
Survey [Sauer et al. 2009] is a random‐systematic, continent‐wide point count survey of
bird population trends, including those of many raptor species). However, a fixed‐radius
approach with circular plots of 800‐m radius typically is used for surveying eagles and
other large (greater than crow [Corvus spp.]‐size) diurnal species of raptors at proposed
wind energy projects in the United States (Strickland et al. 2011).

The optimal duration of point count survey for eagles is a focus of current research. For
now, for point count surveys of eagles at proposed wind energy projects, the Service
recommends counts of 1, 2, or more hours duration instead of 20‐ to 40‐minute counts
typically used (Strickland et al. 2011). Longer counts also facilitate integration of other
survey types (e.g., development of utilization distribution profiles). Many raptor biologists
have suggested that the likelihood of detecting an eagle during a 20‐ to 40‐minute point
count survey is extremely low in all but locales of greatest eagle activity and datasets
generated by pre‐construction point count surveys of this duration typically are replete
with counts of zero eagles, resulting in unwieldy confidence intervals and much uncertainty.
Moreover, time spent traveling to and accessing points for 20‐minute surveys may exceed
time spent conducting the observations. For example, 250 1‐hour surveys conducted
annually at a project of average size (e.g., 15 sampling points, 1 to 3 km apart) and travel
conditions require roughly the same total field time as needed for 500 20‐minute surveys,
yet yield 50% more observation hours (250 versus 167), with correspondingly greater
probability of detecting eagles. Another advantage of longer counts is that they reduce
biases created if some eagles avoid conspicuous observers as they approach their points
and begin surveys, although some observers may become fatigued and overlook eagles
during longer counts. A potential trade off of fewer visits, of course, is diminished
accounting of temporal variation (e.g., variable weather conditions or an abrupt migration
event). While counting at fewer points for longer periods might also reduce the ability to
sample more area, we advocate maintain the minimum spatial coverage of at least 30% of
the project footprint. Until there is more evidence that shorter count intervals are adequate
to estimate eagle exposure, we believe that a sampling strategy including counts of longer
duration, albeit fewer total counts, may in the end improve sampling efficiency and data
quality.

A key assumption of fatality prediction models based on data from point count surveys is
that occurrence of eagles at a proposed project footprint before construction bears a
positive relationship with turbine‐collision mortality after the project becomes operational
(Strickland et al. 2011). Support for this assumption from published literature is limited for
eagles and other diurnal raptors at this time, however. In a recent study of raptors at 20
projects in Europe, no overall relationship was evident between either of two pre‐

55
construction risk indices and post‐construction mortality (Ferrer et al. 2011). However, the
authors based risk indices only in part on data from pre‐construction point counts; factors
incorporated into risk indices included a somewhat subjective decision on species‐specific
sensitivity to collision and conservation status. Despite this, a weak relationship between
pre‐construction flight activity and post‐construction mortality was suggested for the most
common species, griffon vulture (Gyps fulvus) and kestrels (Falco spp.). Neither Aquila nor
Haliaeetus eagles occurred in the study. On coastal Norway, however, a high density, local
population of the white‐tailed eagle, a species closely related and ecologically similar to the
bald eagle, experienced substantial turbine‐collision fatality and loss of nesting territories
after development of a wind energy project (Nygård et al. 2010). The relationship between
pre‐construction occurrence and post‐construction mortality might be less clear if eagles
and other raptor species avoided areas after wind energy projects were constructed (e.g.,
Garvin et al. 2011), but in general such displacement seems negligible (Madders and
Whitfield 2006).

Precision, consistency, and utility of data derived from point count surveys depend greatly
on the sampling framework and field approach for conducting the counts, which in turn
depend somewhat on study objectives and the array of species under consideration.
Precision and reliability of data from point count surveys for eagles can be much improved
upon – and need for a risk‐averse approach lessened – by incorporating some basic,
common‐sense sideboards into the survey design. One of these, longer count duration, is
discussed above. Below are examples of ideal design features for point count surveys of
eagle use of proposed wind energy projects, particularly when fatality rate prediction is a
primary objective. Some of these extend from Strickland et al. (2011) and references
therein, although the first is not in accord with corresponding guidance in that document.

 Surveys of eagles and other large birds are exclusive of those for small birds, to
avoid overlooking large birds while searching at a much smaller scale for a much
different suite of birds. The relatively brief (e.g., 10‐minute) point counts for small
birds could be conducted during the same visit, but before or after the count of large
birds.
 In open areas where observers may be conspicuous, counts are conducted from a
portable blind or from a blind incorporated into a vehicle to reduce the possibility
that some individual eagles avoid observers, ,thus reducing likelihood of detection.
Blinds are designed to mask conspicuous observer movement while not impeding
views of surroundings.
 Point locations may be shifted slightly to capitalize on whatever vantage points may
be available to enhance the observer’s view of surroundings.
 Elevated platforms (e.g., blinds on scaffolding or high in trees, truck‐mounted lifts)
are used to facilitate observation in vistas obstructed by tall vegetation, topographic
features, or anthropogenic structures.
 The observer’s visual field at a point count plot, if less than 800 m (e.g., due to
obstruction by forest cover), is mapped. The percentage of the plot area that is
visible is factored into the calculation of area surveyed.
 Observers use the most efficient, logical route to move among points, changing the
starting point with the beginning of each survey cycle such that each point is
surveyed during a range of daylight hours.
 Systematic scans of the point count plot using binoculars alternating with scans via
the unaided eye to detect close and distant eagles, and with overhead checks for

56











eagles that may have been overlooked during peripheral scanning (Bildstein et al.
2007).
Observers are trained and their skills are tested, including accurate identification
and distance estimation (both horizontal and vertical; e.g., eagles greater than 600
m horizontal distance may not be detected by some observers and correction for
differences among individual observers may be warranted).
The boundary of each point count plot is identified via distinct natural or
anthropogenic features or marked conspicuously (e.g., flagging on poles) at several
points for distance reference. Distance intervals within the plot also are marked if
observations are to be categorized accordingly; rangefinder instruments are useful
in this regard.
Surveys are distributed across daylight hours (e.g., morning – sunrise to 1100 hours;
midday – 1101‐1600; evening 1601 to sunset). In areas or during seasons where
eagle flight is more likely during midday than in early morning or evening (e.g.,
migration [Heintzelman 1986]), sampling efficiency could be increased by
temporally stratifying surveys to more intensively cover the midday period.
A map (e.g., 1:24,000 scale topographic quadrangle) or aerial photographs
indicating topographic and other reference features plus locations of point count
plots is used as the primary recording instrument in the field. A GPS with GIS
interface may serve in this regard.
Time and position of each individual eagle is recorded on the map, e.g., at the
beginning of each minute of observation, if not more frequently.

The following examples of suggested sideboards pertain especially to point count surveys
supplying data for the fatality prediction method recommended in this document:

 Following a point count survey, the duration of observation of each eagle flying
within the plot is summarized in number of minutes, rounded to the next highest
integer (e.g., an eagle observed flying within the plot for about 15 seconds is 1 eagle‐
minute, another observed within for about 1 minute 10 seconds is 2 eagle‐minutes,
and so on; most observations likely will equal 1 eagle‐minute).
 Eagles are mapped when perched or when otherwise not flying, but the summary of
eagle‐minutes for a count excludes these observations and includes only eagles in
flight.
 Horizontal distance of each eagle‐minute is estimated and recorded as ≤ 800 m or >
800 m. Vertical distance of each eagle‐minute is estimated and recorded as ≤ 200 m
(at or below conservative approximation of maximum height of blade tip of tallest
turbine) or > 200 m. Thus, the point count “plot” is a 200‐m high cylinder with a
radius of 800 m.
 Surveys are done under all weather conditions except that surveys are not
conducted when visibility is less than 800 m horizontally and 200 m vertically.
 Data from point count surveys are archived in their rawest form to be available
when fatality is estimated as detailed in this document (APPENDIX D).

Other information recorded during point counts may prove useful in project assessment
and planning, or in additional data analyses (some requiring data pooled from many
projects), e.g.:

57








Flight paths of eagles, including those outside the plot, are recorded on reference
maps, using topographic features or markers placed in the field as location
references. Eagle flight paths are recorded also before and after point count surveys
and incidental to other field work. Flight paths are summarized on a final map, with
those recorded during point count surveys distinguished from others to roughly
account for spatial coverage bias. Documentation of flight paths can aid planning to
avoid areas of high use (Strickland et al. 2011).
Behavior and activity prevalent during each 1‐minute interval is recorded as (e.g.)
soaring flight (circling broadly with wings outstretched); unidirectional flapping‐
gliding; kiting‐hovering; stooping or diving at prey; stooping or diving in an
agonistic context with other eagles or other bird species; undulating/territorial
flight; perched; or other (specified).
Age class of individual eagles is recorded, e.g., juvenile (first year), immature or
subadult (second to fourth year), adult (fifth year or greater), or unknown.
Weather data are recorded, including wind direction and speed, extent of cloud
cover, precipitation (if any), and temperature (Strickland et al. 2011).
Distance measures are used to estimate detectability for improving estimates from
counts (Buckland et al. 2001) and could be used to assess whether eagles avoid
observers. Horizontal distance of each eagle‐minute is estimated and categorized,
e.g., in 100‐m intervals to > 800 m.

The key consideration for planning point count surveys at proposed wind energy projects is
sampling effort. We advise that project developers or operators coordinate closely with the
Service regarding the appropriate seasonal sampling effort, as sampling considerations are
complex and depend in part on case‐specific objectives. We also reiterate that these (and
most other) surveys should be conducted for at least 2 years before project construction
and, in most cases, across all seasons. In general, sampling effort should be commensurate
with the relative level of risk at a proposed project footprint if this can be surmised reliably
from the Stage 1 assessment. If Stage 1 information cannot support reasonably certain risk
categorization, Stage 2 surveys should be conducted as described here to clearly ascertain
whether eagles are known or likely to use the area. If a project is determined to be category
2, products of point count surveys should include data for the fatality model detailed in this
document (APPENDIX D). If there is compelling Stage 1 evidence indicating no use in a
given season, zero use could be assumed and point count surveys in that season might be
unnecessary.

In general, goals for the Stage 2 surveys are either to: (1) confirm category‐3 status for a
project, or (2) to generate a fatality rate estimate. Regardless of which of these survey goals
apply to a particular project, we recommend first identifying potential sites for wind
turbines, including alternate sites, then calculating the total area (km 2 ) encompassing a 1‐
km buffer around all the sites. We suggest 1 km because this approximates optimal spacing
of a generic 2.5‐MW turbine (Denholm et al. 2009), and the area outside this may not be
representative of topographic features and vegetation types that characterize turbine
strings within the project footprint. This approach assures close association between
sampling sites and likely turbine locations, as recommended by Strickland et
al. (2011). Next, we recommend that at least 30% of the area within 1 km of turbines be
considered as the total km 2 area to be covered by 800‐m radius point count plots (with a
sample area for each plot of 2 km2). Our recommended 30% minimum is based on the
actual minimum coverage at eight wind facilities under review by the Service at the time
version 2 of the ECPG was being developed.

58

The first case (i.e., (1) above) is the use of point count data to validate whether a proposed
project meets category 3 criteria when Stage 1 information is inadequate. Based on
experience with current parameters of the “prior term” in our predictive model (see
APPENDIX D), we calculate an average of 20 hours per turbine as an optimal level of annual
sampling via point count survey (e.g., equivalent of ten 4‐hour point count surveys at each
of 20 sample points for a 40‐turbine project; our 20‐hour recommendation considers the
hazardous area created by a generic 2.5‐MW turbine with a rotor diameter of about100 m;
sample effort for turbines with smaller rotor diameters would be less). As sampling effort
falls from this level, uncertainty regarding fatality risk rises sharply, calling for an
increasingly risk averse basis for risk categorization. Although 20 sample hours per turbine
may be necessary initially for validating category 3 determination where little Stage 1
information exists, we expect this will decrease as more projects are incorporated into the
adaptive management meta‐analyses that will refine the prior term.

The second case (i.e., (2) above) is where Stage 1 evidence is strong enough to support the
decision that a project is category 2 (or category 3 with potential for re‐evaluation as
category 2). Fatality rate estimation becomes the main objective of point count surveys and
demands for sampling effort can be reduced. We recommend a minimum of 1 hour of
observation per point count plot per month but at least 2 hours of observation per point
count is warranted for a season for which Stage 1 evidence is ambiguous or suggests high
use.

These ideas on minimum observation hours stem from the Service’s initial experience in
fatality estimation (see APPENDIX D: Stage 3 – Predicting Eagle Fatalities). However, as
noted above, with more field applications of our fatality prediction model we should be able
to refine our ability to characterize uncertainty based in part on site‐specific characteristics,
something the Service’s current model does not do. Again, to develop a reasonable,
informed sampling approach, we urge project developers to engage early with the Service in
discussions about sampling design and strategies.

The example below includes determination of the number of point count plots for a project.
Example
The site for a 100‐MW, 40‐turbine project proposed in open foothills of central New
Mexico encompasses 40 km2 (16 mi2). During the Stage 1 assessment, data from a
hawk watch organization indicates the area is 25 miles east of a north‐south
mountain ridge that sustains a moderate level of migration by golden eagles each
fall but receives little use in spring. According to the state ornithological society, the
region also is thought to attract golden eagles during winter, but this is based on
sparse anecdotal accounts. Aerial nesting surveys by the Service 5 years ago yielded
no evidence of eagle nests within 10 miles of the proposed project, although use of
the area by non‐breeding resident eagles during spring and summer cannot be ruled
out. Reconnaissance visits and review of land cover and other habitat layers in
geodatabases support the general indication that the area is important to golden
eagles during at least part of the year.

Stage 1 Summary: Of primary concern at the prospective project site is potential for
risk to golden eagles during fall migration. Evidence of this at the Stage 1 level is
somewhat equivocal, however, because the known migration pathway is outside the

59
project area. Further examination of use in spring, summer, and especially winter
also seems warranted. Questions include temporal (seasonal) and spatial
(distribution within project) use. The overarching goal is to quantify risk to eagles
posed by the proposed project, mainly by estimating fatality rate. If fatality is
anticipated, a secondary goal is to determine whether the predicted level is
acceptable and, if not, whether fatality can be avoided and minimized through
specified project design and operation features.

The primary tool for predicting fatality is the point count survey. However, if the
pre‐construction assessment is robust and optimally designed, point count surveys
will provide insight on distribution of use within the project footprint especially
near proposed turbine sites, and on migration timing and movement pathways.

Sampling Effort
A. Number of points, i.e., point count plots, and spatial allocation:
1. 40 turbines are proposed for project
2. potential sites for turbines have been selected
3. area within 1 km of turbines covers total of 100 km2
4. 30% of total area = 30 km2
5. number of 800‐m radius (area of each, 2‐km2) point count plots
recommended = 30/2 = 15 plots
6. survey points are distributed among turbine strings via random‐systematic
allocation, with each point no more than 1 km from a prospective turbine
site

B. Number of counts per point per season and duration of each point count survey:
1. Based on some Stage 1 evidence of low use in this example, 1 hour of
observation per point count plot per month seems appropriate during each
of winter (e.g., mid‐December through mid‐March), spring (mid‐March
through mid‐June), and summer (mid‐June through mid‐September)
seasons. A count duration of 1 hour is selected to maximize efficiency in the
field
2. Survey effort is doubled during the mid‐September through mid‐December
fall migration season for golden eagles, based on Stage 1 evidence of fall
migration nearby and need for more definitive data on eagle occurrence,
timing, and distribution within the footprint. This could be done by using
either two 1‐hour counts or a 2‐hour count per point per month; the latter is
chosen to maximize field efficiency and better emulate migration count
methods. The 1‐hour counts may lend better insight on temporal variation,
but in this example each monthly session of 15 2‐hour counts requires an
observer 3‐4 days to complete, affording some accounting of day‐to‐day
variation.
3. The total yearly effort in this example is nine 1‐hour counts and three 2‐
hour counts at each of 15 points, yielding 225 total observation hours.

The raw data, in number of eagle‐minutes, appear as follows (e.g., for the first fall
season sampled, with one 2‐hour count per point per month):

60

Point no.

Point count visit number – Fall Season, Year 1
1 (early fall)
2 (mid‐fall)
3 (late fall)

1

0

0

0

2

0

0

0

3

0

0

0

4

0

0

0

5

0

0

0

6

0

0

0

7

1

1

0

8

0

0

0

9

0

0

0

10

0

2

1

11

0

0

0

12

0

2

0

13

0

0

0

14

0

1

0

15

0

0

0

The first year’s fall point count survey totals 90 observation hours, the equivalent of
nine 10‐hour migration counts. Thus, the fall point count surveys could yield much
insight on eagle migration – perhaps even substituting for focused migration counts
– especially if the sample is stratified so point count surveys mainly cover the
midday period when eagles are most likely to be moving. (see b. Migration Counts
and Concentration Surveys, below). Observations made during point count surveys
in all seasons also could support a map of flight paths to roughly indicate the
distribution of use of the area by eagles relative to turbine sites (see c. Utilization
Distribution (UD) Assessment, below).

Fatality estimation should be adequately supported by the data, although multiple survey
years are likely needed to account for annual variation. Data for fatality estimation should
be made available to the Service in the rawest form, as in the above example.

b. Migration Counts and Concentration Surveys
Wherever potential for eagle migration exists, migration counts should be conducted unless
the Stage 1 assessment presents compelling evidence that the project area does not include
or is not part of a migration corridor or a migration stopover site. Migration counts convey
relative numbers of diurnal raptors passing over an established point per unit time
(Bildstein et al. 2007, Dunn et al. 2008), usually a migration concentration site. Examples of
sites include north‐south oriented ridges, cliff lines, or deeply incised river valleys; terminal
points or coast lines of large water bodies; or peninsulas extending into large water bodies
(Kerlinger 1989, Bildstein 2006, Mojica et al. 2008). Migration counts could be considered a
specialized type of point count, one for which the plot radius is unlimited (Reynolds et al.
1980) and the count period is quite long, from 6 hours to a full day.

61
In contrast to the allocation of sample points for point count surveys at proposed wind
energy projects, migration counts typically are conducted from one to a few points within or
adjacent to a proposed project footprint. Points are widely spaced, located primarily at
places that collectively provide greatest visual coverage especially of topographic features
likely to attract or funnel migrating raptors. At many proposed projects, however, survey
points for migration counts could be the same as or a subset of those used for point count
surveys, e.g., per the above example (under 1a. Point Count Surveys), such that migration
counts at a given point simultaneously contribute point count data. Consideration should
be given to restructuring point count surveys to this end, including temporal stratification
to more effectively account for potential eagle migration and improve precision of exposure
estimates. As another example, during an anticipated 6‐week peak of eagle migration in fall,
point count duration could be extended to 6 hours. If the surveys were to cover either the
first 6 hours or the last 6 hours of the day, the two survey periods would overlap by several
hours in midday, better covering the time of day when eagles are most likely moving
(Heintzelman 1986). The data may have to be adjusted slightly when used for fatality
estimation, however.

Strickland et al. (2011) summarize some important details for conducting raptor migration
counts at proposed wind energy sites. Counts should be conducted using standard
techniques (Bildstein et al. 2007, Dunn et al. 2008) during at least peak periods of passage
(see the Hawk Migration Association of North America’s [HMANA] website for information
on seasonal passage periods for eagles at various migration survey sites:
http://www.hmana.org). Migration counts may involve staffing survey points up to 75% of
days during peak passage (Dunn et al. 2008). If at least a modest eagle migration is
evidenced (i.e., multiple individuals observed passing unidirectionally during each of
multiple days), surveys should be continued for at least 2 years and into the operational
phase to validate initial observations and help assess evidence of collision and influence of
turbines on migration behavior. Migration count data should be provided to the Service as
an appendix to the ECP, using a reporting format similar to that used by HMANA. As with
point count surveys, training of migration survey staff should include assessment of raptor
identification skills and of ability of individuals to detect eagles in flight under a broad range
of distances and weather conditions.

Potential for non‐breeding (either winter or summer) season concentrations of eagles in or
near the project footprint should begin to be evaluated in Stage 1, including close scrutiny of
potential habitat via geospatial imagery and follow up reconnaissance visits (see APPENDIX
B). Non‐breeding bald eagles often use communal roosts and forage communally (Platt
1976, Mojica et al. 2008). Golden eagles may do so on occasion, with other golden eagles
and/or with bald eagles (Craig and Craig 1984). Both species can become concentrated on
spring and fall migration under particular combinations of weather and topographic
conditions, or may annually use traditional stopover sites during migration. The Stage 1
assessment may suggests that seasonal concentrations of eagles regularly occur within the
project area, either because of favorable conditions (e.g., clusters of large trees along rivers
offering potential roost sites, stopover concentrations of migrating waterfowl) or because of
indications from prior anecdotal or systematically collected records. The Stage 2
assessment should include surveys designed to further explore evidence of any such
occurrences. If, based on the outcome of Stage 1, there is no compelling reason to believe
concentration areas are lacking, an efficient way to begin to probe for concentration areas is
simply to extend the duration of point count surveys and perhaps conduct them more
frequently. Expanded point count surveys, distributed evenly across the day during the first

62
year of Stage 2, should provide at least a preliminary indication of regular movements to
and from what may be roosts or prey hotspots within or outside the project footprint.
Moreover, expanded point count surveys conducted near potential turbine sites (see design
recommendations in a. Point Count Surveys, above) can better inform turbine siting
decisions in relation to eagle use of concentration areas, if such areas exist. The increased
survey effort also could contribute towards a more precise indication of eagle exposure in a
fatality estimate for the proposed project (APPENDIX D).

Early in Stage 2, evidence from Stage 1 of concentration areas in the project area may be
corroborated or new evidence of concentrations may surface. In either case, focused
surveys (e.g., via direct observation or by aircraft) can be implemented to document their
locations and daily timing and spatial patterns of their use by eagles in relation to the
proposed project footprint throughout the season(s). For example, surveys for wintering
concentrations of bald eagles could be conducted, following USFWS (1983) guidance.
Direct, systematic observation from vantage points in early morning and evening is the
most practical means of documenting roost locations and movements of eagles to and from
roosts on a local scale (Steenhof et al. 1980, Crenshaw and McClelland 1989). Aerial
surveys may be needed for repeated surveys of eagles at extensive roosts (Chandler et al.
1995). Direct observation can be used to compare occurrence and activity of eagles before
and after construction and operation of a project (Becker 2002) and may be a valid means
to identify disturbance effects on roosting concentrations.

c. Utilization Distribution (UD) Assessment
UD can be thought of as animal’s spatial distribution or intensity of use of various parts of a
given area, such as its home range. A basic though perhaps labor‐intensive approach for
documenting spatial distribution of use across all or part of a proposed project footprint by
eagles is to systematically observe and record eagle movements and activities (e.g.,
territorial display, prey delivery flight) on maps in the field then convert the data into GIS
formats for standard analyses (e.g., Walker et al. 2005). For example, a grid of square cells,
each 0.5 x 0.5 km, can be framed by the Universal Transverse Mercator (UTM) system
across a map of the area of interest to record eagle observations in each 0.25 km2 cell. The
area of interest is divided into non‐overlapping observation sectors, each with a vantage
point that affords unobstructed viewing of grid cells to more than 1 km in all directions.
Observation periods last at least 4 hours and include all daylight hours and account for
roost sites. If necessary, two (or more) observers working from separate vantage points
can pinpoint locations of eagles through triangulation.

The data can be analyzed by simply counting the number of flights intersecting each cell. An
eagle’s distribution of use can then be estimated by using standard kernel analyses (Worton
1989, 1995, Seaman and Powell 1996, Kenward 2001) or other probabilistic approaches,
comparable to Moorcroft et al. (1999), McGrady et al. (2002), and McLeod et al. (2002).
Having concern over potential autocorrelation, Walker et al. (2005) randomly selected
independent locations of golden eagles along flight paths to establish a point database for
standard UD analyses. They determined that locations would be independent if separated
by at least 45 minutes. McGrady et al. (2002) conservatively used a 1‐hour minimum to
separate points, even though their data indicated a 20‐minute interval would suffice.
Concerns with autocorrelation in UD analyses have recently diminished, however (Feiberg
et al. 2010). Most study of eagle UD has focused on resident birds especially breeding
adults on their nesting territories. Size and shape of use areas can vary seasonally (Newton

63
1979), so documentation of spatial use by resident eagles should encompass all seasons in
addition to accounting for annual variation.

A substantial advantage of a direct observation approach compared to telemetry
techniques, which typically target only one or two resident eagles at a proposed project, is
that it disregards age and breeding and residency status. Included are overwintering
individuals; dispersing juveniles; post‐fledging young from nearby territories and juveniles
dispersing from other areas or regions; and adults from adjoining territories plus non‐
breeding adults (i.e., “floaters,” Hunt 1998) and subadults that may occur along boundaries
of breeding territories. In many instances, identification of individual eagles may not be
important and final results of a generalized UD analysis may be based on data pooled from
multiple birds, some of which were indistinguishable from each other in the field. A
disadvantage of this approach is that position accuracy based on direct observation across
expansive landscapes is coarse compared to using telemetry with GPS capability, and
generally declines with distance, increasing topographic and forest cover, and during early
morning and late evening hours. This can be resolved to some extent by limiting the size
and increasing the number of observation sectors (in addition to using multiple observers),
but for most pre‐construction information needs, a high degree of accuracy is unessential
for UD data. Last, it is unlikely that UD needs to be assessed across entire project footprints.
Instead, it is more likely used to target specific areas of concern, such as areas where eagles
nest or frequently forage, and to refine knowledge of use of particular areas to better inform
turbine siting decisions. The method obviously has little utility in areas of low eagle
occurrence.

Although we acknowledge telemetry offers some distinct benefits for assessing risks and
impacts of wind projects, use of the method for eagles has other drawbacks. Specific
individual eagles must be targeted for capture and not all eagles using a given project
footprint are equally likely to be captured or provide useful data (e.g., migrants may be
readily captured but leave the area before providing much data). More importantly,
capturing and radio‐marking eagles can have negative effects on behavior, productivity, and
re‐use of nest sites (e.g., Marzluff et al. 1997, Gregory et al. 2002), and recent information
suggests a negative effect in some cases on survival, especially of golden eagles captured as
adults and released with large (70‐ to 100‐g), solar‐charged transmitters (USFWS,
unpublished information). These effects must be better understood before routine use of
telemetry techniques can be recommended as components of wind‐facility assessments.
Until then, the Service discourages the use of telemetry in assessments of eagle use
associated with wind energy projects; survey approaches suggested herein do not require
telemetry.

d. Summary
The Service encourages development of cost‐effective sampling designs that simultaneously
address multiple aspects of use of proposed wind energy projects by eagles, though
emphasizes that high‐quality point count data to support fatality rate estimation should be
considered the highest priority. In many cases, the sampling framework for point count
surveys likely can be extended to reasonably assess migration incidence, UD, and other
objectives. Although field‐based data that directly support fatality estimation are most
important, development of methods for addressing other objectives is encouraged, such as
the use of digital trail cameras to document eagle occurrence at carcass stations.
Regardless, we recommend that pre‐construction surveys at proposed wind energy sites

64
encompass a minimum of 2 years, including at least 1 year characterized by robust
sampling that integrates multiple survey types.

2. Survey of the Project-area Nesting Population: Number and Locations of Occupied Nests of
Eagles
To evaluate project siting options and help assess potential effects of wind energy projects on
breeding eagles, we recommend determining locations of occupied nests of eagles within the
project area for no less than two breeding seasons prior to construction. The primary objective of a
survey of the project‐area nesting population is to determine the number and locations of occupied
nests and the approximate centers of occupied nesting territories of eagles within the project area.
If recent (i.e., within the past 5 years) data are available on spacing of occupied eagle nests for the
project‐area nesting population, the data can be used to delineate an appropriate boundary for the
project area as described in APPENDIX H. Otherwise, we suggest that project area be defined as the
project footprint and all area within 10 miles.

In this ECPG document we use raptor breeding terminology originally proposed by Postupalsky
(1974) and largely followed today (Steenhof and Newton 2007). An occupied nest is a nest
structure at which any of the following is observed: (1) an adult eagle in an incubating position, (2)
eggs, (3) nestlings or fledglings, (4) occurrence of a pair of adult eagles (or, sometimes subadults,
e.g., Steenhof et al. [1983]) at or near a nest through at least the time incubation normally occurs,
(5) a newly constructed or refurbished stick nest in the area where territorial behavior of a raptor
had been observed early in the breeding season, or (6) “A recently repaired nest with fresh sticks
(clean breaks) or fresh boughs on top, and/or droppings and/or molted feathers on its rim or
underneath” (Postupalsky 1974).

A nest that is not occupied is termed unoccupied. An occupied nesting territory includes one
occupied nest and may include alternate nests, i.e., any of several other nest structures within the
nesting territory. Sometimes “active nest” is used to encompass occupied nests in which eggs were
laid plus those at which no eggs were laid. Here, as elsewhere in the ECPG and in Postupalsky
(1974), an active nest is considered one in which an egg or eggs have been laid. A nest that is active
is also, by default, occupied. A nest that is not active is inactive, and there is a regulatory definition
for the term inactive nest (50 CFR 22.3. Not all pairs of bald eagles and golden eagles attempt to
nest or nest successfully every year (Buehler 2000, Kochert et al. 2002), and nesting territories
where pairs are present but do not attempt to nest could in some cases be misclassified as
unoccupied. Accurate comprehension of territory distribution and determination of occupancy
status is the crux of determining the project‐area nesting population.

The project‐area nesting population survey should include all potential eagle nesting habitat within
the project area. At least two checks via aircraft or two ground‐based observations are
recommended to designate a nest or territory as unoccupied, as long as all potential nest sites and
alternate nests are visible and monitored (i.e., alternate nests may be widely separated such that a
full‐length, ground‐based observation should be devoted to each). Ground‐based observations
should be conducted for at least 4 hours each (occupancy may be verified in less time), aided by
spotting scopes, from at least 0.8 km from the nest(s), during weather conducive to eagle activity
and good visibility. Surveys of occupancy should be conducted at least 30 days apart, ideally during
the normal courtship and mid‐incubation periods, respectively. Surveys later in the breeding
season are likely to overlook some territorial pairs that that did not lay eggs or failed early in the
nesting season. Timing of surveys should be based on local nesting chronologies; Service staff can
provide recommendations. If an occupied nest or a pair of eagles is located, the territory should

65
continue to be searched for alternate nest sites. This information can help determine the relative
value of individual nests to a territory if ever there are applications for permits to take inactive
nests, and when determining whether abandonment of a particular nest may result in loss of a
territory.

Use of aerial surveys followed by ground‐based surveys at targeted sites can be an ideal approach
to determine nest and territory occupancy. Helicopters are an accepted and efficient means for
inventory of extensive areas of potential nesting habitat for eagles, although fixed‐wing aircraft can
be used where potential nest sites are widely scattered and conspicuous. Aerial surveys for eagle
nests in woodland habitat may require two to three times as much time as aerial surveys for nests
on cliffs. When surveying rugged terrain by helicopter, cliffs should be approached from the front,
rather than flying over from behind or suddenly appearing from around corners or buttresses.
Inventories by helicopter should be flown at slow speeds, about 30 to 40 knots. All potentially
suitable nest sites should be scrutinized; multiple passes at several elevation bands may be
necessary to provide complete coverage of nest site habitat on large cliff complexes. Hovering for
up to 15 seconds no closer than 50 m from a nest may be necessary to verify the nesting species,
photograph the nest site, and, if late in the nesting season, allow the observer to count and estimate
age of young in the nest. Aerial surveys may not be appropriate in some areas such as bighorn
sheep lambing areas; to avoid such sensitive areas, state resource agencies should be consulted
when planning surveys. Additional guidelines for aerial surveys for eagles and other raptors are
reviewed in Anderson (2007).

Surveys should be conducted only by biologists with extensive experience in surveys of raptors and
appropriate training in aerial surveys (see review in Anderson 2007). Whether inventories are
conducted on the ground or aerially, metrics of primary interest to the Service for the project‐area
nesting population include:

1. number and locations of nest structures that are verified or likely to be eagle nests
2. number and locations of eagle nests currently or recently occupied based on criteria
outlined herein
3. estimated number and approximate boundaries and centers of eagle breeding territories,
based on records of nest site occupancy and clustering of nests.

Additionally, productivity (i.e., reproductive success, defined here as the mean number of nestlings
surviving to > 56 and ≥ 67 days of age per occupied nest for golden eagles and bald eagles,
respectively) may be of interest for assessing disturbance effects, although utility of productivity
data at a given project likely will be limited due to small sample size and factors confounding the
interpretation of results. A meta‐analysis approach based on productivity data from many projects
is contemplated as part of the adaptive management process accompanying the ECPG, and may
contribute to understanding of disturbance effects on this aspect of eagle breeding biology.
Moreover, abandonment of territories – the gravest manifestation and clearest evidence of
disturbance effects – could be documented through the occupancy surveys recommended herein, if
these surveys are repeated after project construction. We reiterate that accurate comprehension of
territory distribution and determination of occupancy status should be the primary goal of nesting
surveys.

66

Literature Cited
Anderson, D. E. 2007. Survey techniques. Pages 89‐100 in D. M. Bird and K. L. Bildstein (eds.),
Raptor research and management techniques. Hancock House, Blaine, Washington.
Becker, J. M. 2002. Response of wintering bald eagles to industrial construction in southeastern
Washington. Wildlife Society Bulletin 30:875‐878.
Bildstein, K. L. 2006. Migrating raptors of the world, their ecology and conservation. Cornell
University Press, Ithaca, New York.
Bildstein, K. L., J. P. Smith, and R. Yosef. 2007. Migration counts and monitoring. Pages 102‐115 in D.
M. Bird and K. L. Bildstein (eds.), Raptor research and management techniques. Hancock House,
Blaine, Washington.
Buckland, S. T., D. R. Anderson, K. P. Burnham, J. L. Laake, D. L. Borchers, and L. J. Thomas. 2001. An
introduction to distance sampling: estimating abundance of biological populations. Oxford
University Press, Oxford, United Kingdom.
Buehler, D. A. 2000. Bald Eagle Haliaeetus leucocephalus. No. 506 in A. Poole and F. Gill (eds.), The
Birds of North America. The Birds of North America, Inc., Philadelphia, Pennsylvania.
Chandler, S. K., J. D. Fraser, D. A. Buehler, and J. K. D. Seegar. 1995. Perch trees and shoreline
development as predictors of bald eagle distribution on Chesapeake Bay. Journal of Wildlife
Management 59:325‐332.
Craig, T. H., and E. H. Craig. 1984. A large concentration of roosting golden eagles in southeastern
Idaho. Auk 101:610‐613.
Crenshaw, J. G., and B. R. McClelland. 1989. Bald eagle use of a communal roost. Wilson Bulletin
101:626‐633.
Denholm, P., M. Hand, M. Jackson, and S. Ong. 2009. Land‐use requirements of modern wind power
plants in the United States. National Renewable Energy Laboratory Technical Report NREL/TP‐
6A2‐45834. www.nrel.gov/docs/fy09osti/45834.pdf (last visited October 22, 2011)
Dunn, E. H., D. J. T. Hussell, and E. R. Inzunza. 2008. Recommended methods for population
monitoring at raptor‐migration watch sites. Pages 447‐459 in K. L. Bildstein, J. P. Smith, E. R.
Inzunza and R. R. Veit (eds.). State of North America’s birds of prey. Series in Ornithology No. 3.
Nuttall Ornithological Club. Cambridge, Massachusetts and American Ornithologists' Union,
Washington, D.C.
Fieberg, J., J. Matthiopoulos, M. Hebblewhite, M. S. Boyce, and J. L. Frair. 2010. Correlation and
studies of habitat selection: problem, red herring, or opportunity? Philosophical Transactions
of the Royal Society Biological Sciences 365: 2233‐2244. doi: 10.1098/rstb.2010.0079.
Garvin, J. C., C. S. Jennelle, D. Drake, and S. M. Grodsky. 2011. Response of raptors to a windfarm.
Journal of Applied Ecology 48:199‐209.
Heintzelman, D. S. 1986. The migrations of hawks. Indiana University Press, Bloomington and
Indianapolis, Indiana.
Hutto, R. L., S. M. Pletschet, and P. Henricks. 1986. A fixed‐radius point count for nonbreeding and
breeding season use. Auk 103:593‐602.
Johnson, D. H. 1995. Point counts of birds: what are we estimating? Pages 117‐123 in C. J. Ralph and
J. R. Sauer (eds.), Monitoring bird populations by point counts. General Technical Report PSW‐
GTR‐149, U.S. Forest Service, Pacific Southwest Research Station, Albany, California.
Kerlinger, P. 1989. Flight strategies of migrating hawks. University of Chicago Press, Chicago,
Illinois.
Kochert, M. N., K. Steenhof, C. L. McIntyre, and E. H. Craig. 2002. Golden Eagle Aquila chrysaetos. No.
684 in A. Poole and F. Gill (eds.), The Birds of North America. The Birds of North America, Inc.,
Philadelphia, Pennsylvania.
Madders, M., and D. P. Whitfield. 2006. Upland raptors and the assessment of wind farm impacts.
Ibis 148:43‐56.

67
McGrady, M. J., J. R. Grant, I. P. Bainbridge, and D. R. A. McLeod. 2002. A model of golden eagle
(Aquila chrysaetos) ranging behavior. Journal of Raptor Research 36(1) supplement:62‐69.
Mojica, E. K., J. M. Meyers, B. A. Millsap, and K. T. Haley. 2008. Migration of sub‐adult Florida bald
eagles. Wilson Journal of Ornithology 120:304‐310.
Nygård, T, K., E. L. Bevanger, Ø. Dahl, A. Flagstad, P. L. Follestad, Hoel, R. May, and O. Reitan. 2010. A
study of white‐tailed eagle Haliaeetus albicilla movements and mortality at a wind farm in
Norway. in D. Senapathi (ed.), Climate change and birds: adaptation, mitigation and impacts on
avian populations. British Ornithologists’ Union Proceedings, Peterborough, United Kingdom.
http://www.bou.org.uk/bouproc‐net/ccb/nygard‐etal.pdf (last visited October 9, 2011).
Platt, J. B. 1976. Bald eagles wintering in a Utah desert. American Birds 30:783‐788.
Potupalsky, S. 1974. Raptor reproductive success: some problems with methods, criteria, and
terminology. Raptor Research Report 2:21‐31.
Ralph, C. J., G. R. Geupel, P. Pyle, T. E. Martin, and D. F. DeSante. 1993. Handbook of field methods for
monitoring landbirds. General Technical Report PSW‐GTR‐144, U.S. Forest Service, Pacific
Southwest Research Station, Albany, California.
Reynolds, R. T., J. M. Scott, and R. A. Nussbaum. 1980. A variable circular‐plot method for estimating
bird numbers. Condor 82:309‐313.
Sauer, J. R., J. E. Hines, J. E. Fallon, K. L. Pardieck, D. J. Ziolkowski, Jr., and W. A. Link. 2011. The North
American Breeding Bird Survey, results and analysis 1966 ‐ 2009. Version 3.23.2011. U.S.
Geological Survey, Patuxent Wildlife Research Center, Laurel, Maryland, USA.
Steenhof, K., S. S. Berlinger, and L. H. Fredrickson. 1980. Habitat use by wintering bald eagles in
South Dakota. Journal of Wildlife Management 44:798‐805.
Steenhof, K., M. N. Kochert, and J. H. Doremus. 1983. Nesting of subadult golden eagles in
southwestern Idaho. Auk 100:743‐747.
Steenhof, K., and I. Newton. 2007. Assessing nesting success and productivity. Pages 181‐191 in D.
M. Bird and K. Bildstein (eds.), Raptor Research and Management Techniques. Hancock House,
Blaine, Washington, USA.
Strickland, M. D., E. B. Arnett, W. P. Erickson, D. H. Johnson, G. D. Johnson, M. L. Morrison, J. A.
Shaffer, and W. Warren‐Hicks. 2011. Comprehensive guide to studying wind energy/wildlife
interactions. Prepared for the National Wind Coordinating Collaborative, Washington, D.C.
USFWS. 1983. Northern states bald eagle recovery plan. U.S. Fish and Wildlife Service, Division of
Migratory Bird Management, Washington D.C. http://www.fws.gov/midwest/eagle/recovery/
be_n_recplan.pdf (last visited October 9, 2011).
Walker, D., M. McGrady, A. McCluskie, M. Madders, and D. R. A. McLeod. 2005. Resident golden
eagle ranging behaviour before and after construction of a windfarm in Argyll Scottish Birds
25:24‐40.

68

APPENDIX D: STAGE 3 – PREDICTING EAGLE FATALITIES

The Service uses a Bayesian method (see Gelman et al. 2003) to predict the annual fatality rate for a
wind‐energy facility, using explicit models to define the relationship between eagle exposure
(resulting from the Stage 2 assessment, APPENDIX C), collision probability, and fatalities (verified
during post‐construction monitoring in Stage 5, APPENDIX H), and to account for uncertainty. The
relationships between eagle abundance, fatalities, and their interactions with factors influencing
collision probability are still poorly understood and appear to vary widely depending on multiple
site‐specific factors (see Assessing Risk and Effects; 2. Eagle Risk Factors in the ECPG). The baseline
model presented below is a foundation for modeling fatality predictions from eagle exposure to
wind turbine hazards. In addition to generating the fatality estimate that will be a component of the
Service's analysis of the permit application, the model also serves as a basis for learning and the
exploration of other candidate models that attempt to better incorporate specific factors and
complexity. The Service encourages project developers or operators to develop additional
candidate models (both a priori and post hoc) for direct comparison with, and evaluation of, the
baseline model and modeling approach. Our ability to learn over time and reduce uncertainty by
incorporating new information into our modeling approach through an adaptive management
framework (see APPENDIX A) enables us to improve site‐specific estimation of eagle fatalities,
reduce uncertainty in predictions, and, ultimately, improve management decisions relating to
eagles and wind energy in a responsible and informed way. Rigorous post‐construction monitoring
is a critical component of evaluating model performance over time (see APPENDIX H).

Variables used in the formulas below are summarized in Table D‐1 for ease of reference. The total
annual eagle fatalities (F) as the result of collisions with wind turbines can be represented as the
product of the rate of eagle exposure (λ) to turbine hazards, the probability that eagle exposure will
result in a collision with a turbine (C), and an expansion factor (ε) that scales the resulting fatality
rate to the parameter of interest, the annual predicted fatalities for the project:

.

Using the Bayesian estimation framework, we define prior distributions for exposure rate and
collision probability; the expansion factor is a constant and therefore does not require a prior
distribution. Next, we calculate the exposure posterior distribution from its prior distribution and
observed data. The expanded product of the posterior exposure distribution and collision
probability prior yields the predicted annual fatalities.

69

Table D-1. Abbreviations and descriptions of variables used in the Service method for predicting annual eagle
fatalities.
Abbreviation

Variable

Description

F

Annual fatalities

Annual eagle fatalities from turbine collisions

λ

Exposure rate

Eagle‐minutes flying below 200 m in height within the project
footprint (in proximity to turbine hazards) per hr per km2

C

Collision
probability

The probability of an eagle colliding with a turbine given exposure

ε

Expansion factor

Product of daylight hours and total hazardous area (hr∙km2)

k

Eagle‐minutes

Number of minutes that eagles were observed flying below 200 m
during survey counts

δ

Turbine
hazardous area

Rotor‐swept area around a turbine or proposed turbine from 0 to 200
m (km2)

n

Trials

Number of trials for which events could have been observed (the
number of hr∙km2 observed)

τ

Daylight hours

Total daylight hours (e.g. 4383 hr per year)

nt

Number of
turbines

Number of turbines (or proposed turbines) for the project

1. Exposure
The exposure rate λ is the expected number of exposure events (eagle‐minutes) per daylight hour
per square kilometer (hr∙ km2). We defined the prior distribution for exposure rate based on
information from a range of projects under Service review and others described with sufficient
detail in Whitfield (2009). The exposure prior predicts an exposure rate from a mixture distribution
of project‐specific Gamma distributions (Figure D‐1). We used the Gamma distribution because all
values are positive and real (see Gelman et al., 1995, p. 474–475). The mixture distribution is
summarized by a new Gamma distribution (our prior distribution for exposure) with a mean
(0.352) and standard deviation (0.357) derived from the conditional distributions (Gelman et al,
1995, equation 1.7 p. 20). The resulting prior distribution for exposure rate is:

~

∝,

, with shape and rate parameters of α = 0.97 and β = 2.76.

Simulation trials produced consistent results. The prior distribution is meant to include the range of
possible exposure rates for any project considered.

70

Figure D-1. The prior probability distribution Gamma (0.97, 2.76), for exposure rate, λ, with a mean of 0.352
(indicated by the reference line) and standard deviation of 0.357. The distribution is positively skewed such
that exposure is generally at or near 0 with fewer higher values shown by the black curve. The project-specific
distributions (gray curves) were used to determine the mixture distribution (dashed curve) which determined the prior
distribution parameters.

Eagle exposure data collected during the pre‐construction phase surveys (see APPENDIX C) can be
used to update this prior and determine the posterior distribution that will be used to estimate the
predicted fatalities. The Service may also be able to work with a project developer or operator on a
case‐by‐case basis to use the prior λ distribution to generate a risk‐averse fatality prediction for
projects where no pre‐construction survey data are available. Assuming the observed exposure
minutes follow a Poisson distribution with rate λ, the resulting posterior λ distribution is:

,
.
~
∝ ∑

The new posterior λ parameters are the sum of α from the prior and the events observed (eagle
minutes, ki), and the sum of β from the prior and the number of trials, n, for which events could
have been observed (the number of “trials” is the number of hr∙km2 that were observed). Note that
by including realistic time and area data from the pre‐construction surveys, the relative influence of
the prior λ distribution on the resulting posterior λ distribution for exposure rate becomes
negligible. In other words, with adequate sampling, the data will determine the posterior
distribution, not the prior. The posterior λ distribution can then be used to estimate the annual
fatality distribution.

71
In addition, this posterior λ distribution can now serve as a prior distribution for the next iteration
of the predictive model in an adaptive framework (see APPENDIX A), at least for the project under
consideration and potentially in a more general way as the posteriors from multiple sites are
considered; in this way, we build ongoing information directly into the predictive process.

2. Collision Probability
Collision probability C is the probability, given exposure (1 minute of flight in the hazardous area,
), of an eagle colliding with a turbine; for the purposes of the model, all collisions are considered
fatal. We based the prior distribution on a Whitfield (2009) study of avoidance rates from four
independent sites. Averaging avoidance from those sites yielded a mean and standard deviation for
collision probability of 0.0058, 0.0038, respectively (note this is consistent with eagle avoidance
rates in other risk assessment approaches, e.g. 99%). This in turn defined the prior C distribution
as:

~
, ´ , with parameters ν and ν´ of 2.31 and 396.69 (Figure D‐2).

The Beta distribution is used to describe values between 0 and 1 (Gelman et al.,1995, p. 476–477).
The prior C distribution attempts to include the range of possible collision probabilities across the
set of potential sites to be considered.

60
0

20

40

Density

80

100

120

Collision Probability Prior

0.000

0.005

0.010

0.015

Pr(Collision|Exposure Minute)

0.020

Figure D-2. The probability distribution for the collision probability prior, a Beta(2.31, 396.69) distribution
with a mean of 0.0058 (indicated by the reference line) and a standard deviation of 0.0038. The distribution
is positively skewed such that most collision probabilities will be small.

At the time of pre‐construction permitting, the prior C distribution will be used to estimate the
annual predicted fatalities. After construction, post‐construction monitoring can be used to
determine the posterior C distribution by updating the prior C distribution.

72

Assuming the observations of fatalities follow a binomial distribution with rate C, the posterior
distribution of the rate C will be a beta distribution (the beta distribution and the binomial
distribution are a conjugate pair):

~
, ´
,

where f is the number of fatalities estimated from the Stage 5 post‐construction monitoring, and g is
the estimated number of exposure events that did not result in a fatality. The posterior distribution
for C cannot be calculated until a project has been built, has started operations, and at least one
season of post‐construction monitoring has been completed. Once determined, the posterior C
distribution can then be used to generate a prediction for annual fatalities and can serve as a prior C
for the next iteration of the predictive model (see APPENDIX A).

3. Expansion
The expansion factor (ε) scales the resulting per unit fatality rate (fatalities per hr per km2) to the
daylight hours, τ, in 1 year (or other time period if calculating and combining fatalities for seasons
or stratified areas) and total hazardous area (km2) within the project footprint:

∑
,

where nt is the number of turbines, and δ is the circular area centered at the base of a turbine with a
radius equal to the rotor‐swept radius of the turbine; we define this as the hazardous area
surrounding a turbine. In this model, to simplify data requirements and assumptions, we consider
both eagle use and hazardous area as 2‐dimensional areas, since the height of the sampled and
hazardous areas are the same (200 m) and will cancel out in the calculations. Alternative models
that consider 3‐dimensional space could also be considered, though the expansion factor should be
adjusted accordingly. The units for ε are hr∙ km2 per year (or time period of interest).

4. Fatalities
Now we can generate the distribution of predicted annual fatalities as the expanded product of the
posterior exposure rate and the prior collision probability (once post‐construction data is available,
the posterior collision probability would be used to update our fatality distribution):

∙
∙
.

We can then determine the mean, median, standard deviation, and 80% quantile (this will be the
upper credible limit) directly from the distribution of predicted fatalities.

5. Putting it all together: an example
The Patuxent Power Company example below illustrates the calculation of predicted fatalities from
exposure data from a hypothetical project site. This data will normally come from the field surveys
in Stage 2, but for the purposes of this example, we have generated fabricated observation data.
The advantage of simulating data in such an exercise is that we can manipulate model inputs to
critically evaluate the performance of the model. Additional examples are provided at the end of
this document to illustrate the general approach and clarify specific considerations that may apply
to certain projects.

73

a. Patuxent Power Company Example
Patuxent Power Company conducted surveys for eagles at a proposed location for a small‐
to medium‐sized wind facility (18 turbines, each with a 50 meter rotor diameter) following
the recommended methods in the ECPG (see Table D‐2). They conducted 168 counts at 7
points and 60 eagle‐min of exposure were observed. Each count was 2‐hr in duration, and
covered a circular area of radius 0.8 km. Thus, 675.6 km2∙hr were observed in total.

Table D-2. Exposure data for Patuxent Power Company example. In this hypothetical example, 168 counts
were performed. Each count was 2-hr in duration and covered a 0.8 km radius circle. Thus, the total time and area
2
sampled was 675.6 km ·hr. In that time, 60 exposure events (eagle-min) were observed.
Visit
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Total

P1
0
0
0
0
0
0
0
0
0
0
1
0
0
2
0
0
0
1
0
0
0
1
1
0
6

P2
0
0
1
1
1
0
1
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
5

P3
2
1
2
0
0
1
0
0
0
0
1
0
1
0
0
0
0
1
0
2
0
0
0
0
11

P4
0
0
0
0
1
1
0
0
0
0
1
0
0
0
2
1
2
1
1
0
0
0
3
0
13

P5
2
0
0
0
0
0
0
0
0
0
0
1
0
0
2
0
0
0
0
1
1
0
0
0
7

P6
0
0
0
1
1
0
1
1
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
6

P7
1
1
1
1
1
1
1
0
0
0
0
0
1
2
1
0
0
0
0
0
0
1
0
0
12

Total
5
2
4
3
4
3
3
1
0
0
3
2
2
4
5
1
2
3
3
3
1
2
4
0
60

74

b. Exposure
The posterior distribution for the exposure rate is:

~
∝, , remember,
~
0.97, 2.76 , Figure D1; where,

0.97

60

60.97

678.31

∙

2.76

168

2

0.8

Thus,

~

60.97, 678.31 ; the units for λ are per hr per km2.

The posterior distribution is shown in Figure D‐3. The mean and standard deviation of
exposure rate are 0.09 and 0.01, respectively. Note that there is little influence of the prior
on this posterior, because the sampling effort was substantial.

Figure D-3. The posterior distribution for exposure rate for the example project, “Patuxent Power
Company.” This gamma distribution has a mean (indicated by the reference line) of 0.09 and a standard deviation of
0.01.

75

b. Collision Probability
We do not have any additional information about collision probability, C, so we will use the
prior distribution, which has a mean of 0.0058 and a standard deviation of 0.0038:

~
2.31,396.69 ; see Figure D‐2.

c. Expansion
The expansion rate, ε, is the number of daylight hours in a year (τ) multiplied by the
hazardous area (δ) around the 18 turbines proposed for the project:

∙ 18 154.9 ∙
.
4,383 ∙ 0.025

d. Fatalities
To determine the distribution for the predicted annual fatalities, the exposure and collision
risk distributions need to be multiplied by each other and expanded. The resulting
distribution cannot be calculated in closed form; it is easiest to generate it through
simulations. In this example, after running 100,000 simulations, the predicted distribution
for annual fatalities (Figure D‐4) has a mean of 0.082 and a standard deviation of 0.055.
The 80% quantile is 0.12 eagle fatalities per year.

0

2

4

Density

6

8

10

Predicted Annual Fatalities

0.00

0.05

0.10

0.15
Fatality Rate

0.20

0.25

0.30

Figure D-4. The probability distribution for predicted annual fatalities. The histogram shows the simulation
results. The mean (0.082) and 80% quantile (0.12) are represented by the reference lines (black and gray,
respectively). The standard deviation is 0.055.

76

The Service’s baseline model for the proposed Patuxent wind facility predicts that 80% of
the time that annual fatalities would be 0.12 eagles or fewer, suggesting that an eagle
collision fatality would be predicted to occur at the project site every 8‐9 years on average.
The facility had a medium amount of eagle activity at the site, but the small size of the
project kept the predicted fatality numbers lower than they would have been for a larger
project in the same location. Ideally, we would consider other candidate models alongside
the baseline model presented here and compare their relative performance using data
collected in Stage 5.

6. Additional Considerations
This initial estimate of fatality rate should not take into account possible conservation measures
and ACPs (e.g. changes in turbine siting or seasonal curtailments); these will be factored in as part
of Stage 4 (APPENDIX E). Additionally, any loss of production that may stem from disturbance is
not considered in these calculations, but should be added to these estimates and later adjusted
based on post‐construction monitoring as described in Stage 5. This stage and Stage 5 of the ECP
will require close coordination between the project developer or operator and the Service.

a. Small-scale Projects
Small‐scale projects (generally these will be residential or small‐business projects) may
pose a low enough risk that Stage 2 surveys are unnecessary to demonstrate that the
project is not likely not take eagles. This presumes that Stage 1 surveys are conducted and
show no important eagle use areas or migration concentration sites in the project area. In
such cases, the fatalities predicted by the collision fatality model are the expanded product
of the exposure prior and the collision probability prior; the exposure prior is not updated
to create a posterior as it would be for projects with survey data (Figure D‐5). With the
prior distributions currently used for exposure rate and collision probability (note that the
parameters for the priors distributions are part of the adaptive management framework
and will change as new information becomes available), the 80 percent quantile of the
predicted fatality distribution for projects with less than approximately 2.4x10‐3 km2 of
hazardous area predicts fatalities at a rate less than 1 eagle in 30 years (not likely to take
eagles). This is equivalent to a single turbine with a rotor diameter of approximately 55 m,
or more than 45 turbines with 8 m rotor diameter (each of which has the capacity to exceed
typical home energy needs). The calculation of hazardous area is presented in this
Appendix under ‘Expansion’. If the collision model prediction based on the exposure prior
predicts that take of eagles will occur (e.g., if the hazardous area is greater than 2.4x10‐3
km2), Stage 2 preconstruction sampling for eagle use of the project area is recommended
(see APPENDIX C). The data from Stage 2 surveys will be used to update the exposure prior
distribution and produce a project‐specific fatality prediction. Projects are encouraged to
consult with the Service early in the planning process as components of the fatality
prediction model will continue to evolve and may change over time.

77

Figure D-5. Predicted fatalities for projects with small hazardous areas based on the prior-only collision fatality
-3
2
model; projects with less than 2.4x10 km hazardous area are predicted to take less than 1 eagle in 30 years.

The Service is working on the development of additional tools to assist project developers or
operators with estimating predicted fatalities given different inputs and allowing for the flexibility
to incorporate other factors into additional candidate models. We encourage project developers or
operators to begin coordinating with the Service early in the process (Stage 1 or Stage 2) so that we
can collaboratively develop a suite of candidate models to consider.

Literature Cited
Gelman, A., Carlin, J. B., Stern, H. S., and D. B. Rubin. 2003. Bayesian Data Analysis, 2nd ed. London,
Chapman & Hall.
Whitfield, D. P. 2009. Collision avoidance of golden eagles at wind farms under the ‘Band’ collision
risk model. Report from Natural Research to Scottish Natural Heritage, Banchory, UK.

78

APPENDIX E: STAGE 4 – AVOIDANCE AND MINIMIZATION OF RISK USING ACPS AND OTHER
CONSERVATION MEASURES, AND COMPENSATORY MITIGATION

The most important factor when considering potential effects to eagles is the siting of a wind
project. Based on information gathered in Stage 2 and analyzed in Stage 3, the project developer or
operator should revisit the site categorization from the Stage 1 assessment to determine if the
site(s) still falls into an acceptable category of risk (at this stage, acceptable categories are 2 and 3,
and very rarely 1). When information suggests that a proposed wind project has a high eagle
exposure rate and presents multiple risk factors (e.g., is proximate to an important eagle‐use area
or migration concentration site and Stage 2 data suggest eagles frequently use the proposed wind‐
project footprint), it should be considered a category 1 site; we recommend relocating the project
to another area because a location at that site would be unlikely to meet the regulatory
requirements for a programmatic permit. If the site falls into categories 2 or 3, or rarely some
category 1 sites where there is potential to adequately abate risk, the ECP should next address
conservation measures and ACPs that might be employed to minimize or, ideally, avoid eagle
mortality and disturbance. To meet regulatory requirements, ACPs, if available, must be employed
such that any remaining eagle take is unavoidable.

In this section of the ECP, we recommend project developers or operators re‐run models predicting
eagle fatality rates after implementing conservation measures and available ACPs for all the
plausible alternatives. This re‐analysis serves two purposes: (1) it demonstrates the degree to
which minimization and avoidance measures might reduce effects to eagle populations compared
to the baseline project configuration, and (2) it provides a prediction of unavoidable eagle
mortality. Conservation measures and ACPs should be tailored to specifically address the risk
factors identified in Stage 3 of the ECP. This section of the ECP should describe in detail the
measures proposed to be implemented and their expected results.

The Service does not advocate the use of any particular conservation measures and merely
provides the below list as examples. Moreover, at this time none of these measures have been
approved as ACPs for wind projects. Ultimately, project developers or operators will propose and
implement site specific conservation measures and ACPs (as they become available) in cooperation
with local Service representatives in order to meet the regulatory standard of reducing any
remaining take to a level that is unavoidable.

Examples of conservation measures that could be considered before and during project
construction, depending on the specific risk factors involved, include:
1. Minimize the area and intensity of disturbances during pre‐construction and construction
periods.
2. Prioritize locating development on lands that provide minimal eagle use potential including
highly developed and degraded sites.
3. Utilize existing transmission corridors and roads.
4. Set turbines back from ridge edges.
5. Site structures away from high eagle use areas and the flight zones between them.
6. Dismantle nonoperational meteorological towers.
7. Bury power lines to reduce avian collision and electrocution.
8. Follow the Avian Power Line Interaction Committee (APLIC) guidance on power line
construction and design (APLIC 2006).
9. Minimize the extent of the road network.

79
10. Avoid the use of structures, or remove existing structures, that are attractive to eagles for
perching.
11. Avoid construction designs (including structures such as meteorological towers) that
increase the risk of collision, such as guy wires. If guy wires are used, mark them with bird
flight diverters (according to the manufacturer’s recommendation).
12. Avoid siting turbines in areas where eagle prey are abundant.
13. Avoid areas with high concentrations of ponds, streams, or wetlands.

Examples of avoidance and minimization measures that could be considered during project
operation, depending on the specific risk factors involved, include:

1. Maintain facilities and grounds in a manner that minimizes any potential impacts to eagles
(e.g. minimize storage of equipment near turbines that may attract prey, avoid seeding forbs
below turbines that may attract prey, etc.).
2. Avoid practices that attract/enhance prey populations and opportunities for scavenging
within the project area.
3. Take actions to reduce vehicle collision risk to wildlife and remove carcasses from the
project area (e.g. deer, elk, livestock, etc.).
4. Instruct project personnel and visitors to drive at low speeds (< 25 mph) and be alert for
wildlife, especially in low visibility conditions.

When post‐construction fatality information becomes available, the project developer or operator
and the Service should consider implementing all or a subset of the additional conservation
measures and experimental ACPs that were considered at the time the permit was issued (see
ASSESSING RISK AND EFFECTS, 3b. General Approach to Address Risks in the ECPG).

Examples of experimental ACPs that could be identified initially or after evaluation of post‐
construction fatality monitoring data, depending on the specific risk factors involved, include:

1. Seasonal, daily, or mid‐day shut‐downs (particularly relevant in situations where eagle
strikes are seasonal in nature and limited to a few turbines, or occur at a particular time of
day).
2. Turbine removal or relocation.
3. Adjusting turbine cut‐in speeds.
4. Use of automated detection devices (e.g. radar, etc.) to control the operation of turbines.

Literature Cited
Avian Power Line Interaction Committee (APLIC). 2006. Suggested practices for avian protection
on power lines: the state of the art in 2006. Edison Electric Institute, APLIC, and the California
Energy Commission. Washington D.C. and Sacramento, CA, USA. http://www.aplic.org/
SuggestedPractices2006(LR‐2watermark).pdf.

80

APPENDIX F: ASSESSING PROJECT-LEVEL TAKE AND CUMULATIVE EFFECTS ANALYSES

The Service is required to evaluate and consider the effects of programmatic take permits on eagles
at the eagle management unit, local‐area, and project‐area population scales, including cumulative
effects, as part of its permit application review process (50 CFR 22.26 (f)(1) and USFWS 2009). The
Service will rely on information a developer provides from the Stage 1 and Stage 2 assessments, as
well as all other available information on mortality and other population‐limiting effects at the
various population scales, when preparing its cumulative impact assessment. The Service’s NEPA
on the Eagle Permit Rule evaluated and set sustainable take levels at the eagle management unit
scale (USFWS 2009). However, that NEPA analysis did not assess impacts at other population
scales. A significant part of the cumulative effects evaluation is assessing the effect of the proposed
take in combination with take caused by previously authorized actions and reasonably foreseeable
future actions on the local‐area eagle population(s), and it is this analysis that is the focus of this
appendix.

The purpose of this part of the cumulative effects evaluation is to identify situations where take,
either at the individual project level or in combination with other authorized or foreseeable future
actions and other limiting factors at the local‐area population scale, may be approaching levels that
are biologically problematic or which cannot reasonably be offset through compensatory
mitigation. In previous assessments of the effect of falconry take on raptor populations (Millsap
and Allen 2006), the Service identified annual take levels of 5% of annual production to be
sustainable for a range of healthy raptor populations, and annual take levels of 1% of annual
production as a relatively benign harvest rate over at least short intervals when population status
was uncertain. This approach was used to establish take thresholds at the eagle management unit
scale (USFWS 2009). The Service considered several alternatives for benchmark harvest rates at
the local‐area population scale, and after comparative evaluation identified take rates of between
1% and 5% of the estimated total eagle population size at this scale as significant, with 5% being at
the upper end of what might be appropriate under the BGEPA preservation standard, whether
offset by compensatory mitigation or not. These local‐area harvest rate benchmarks are overlain
by the more conservative take thresholds for the eagle management units, so the overall harvest
rate at the eagle management unit scale should not exceed levels established in the Final
Environmental Assessment (USFWS 2009).

The Service recommends a top‐down approach for this assessment: (1) identify numbers of eagles
that may be taken safely at the national level (i.e., a national‐level benchmarks); (2) allocate take
opportunities among regional eagle management units (USFWS 2009) as a function of the
proportion of eagles in each unit (i.e., regional‐level benchmarks); (3) further allocate take
opportunities to the local‐area population scale as a function of inferred eagle population size at
that scale (assuming, in the absence of better data on eagle distribution at the scale of the eagle
management unit, a uniform distribution of that population); and (4) incorporating benchmarks
that can be used to assess the likely sustainability of predicted levels of take at the local‐area scale.
Through a spatial accounting system, permitted take is managed to ensure that the benchmarks
also consider cumulative effects at the local‐area eagle population scale as a guard against
authorizing excessive take at this scale.

In Table F‐1, we work through this approach using the hypothetical example of eight individual yet
identical projects, one in each bald eagle management unit. Each of these projects has a 314 mi2
footprint, and affects a local‐area bald eagle population over 8824 square mile (mi2) area. For this
example, we use a take rate of 5% of the local‐area bald eagle population per year as the maximum
acceptable take rate. In this example, the 5% benchmark take rate over the eight projects is 150

81
individual bald eagles per year, and the range of allowable take rates at this scale varies across
management units from <1 bald eagle per year in the southwest to 67 per year in Alaska. Table F‐2
provides population and eagle management unit area statistics for golden eagles to aid in
performing these calculations for that species.

As noted above, in cases where the local‐area eagle populations of proximate projects overlap, the
overlap should be taken into account in a cumulative effects analysis so that the cumulative take on
the local‐area population scale can be considered against population benchmarks. Figure F‐1
illustrates one method to do this, and Table F‐3 provides the calculations for this example. These
examples use bald eagles, but the same concept and approach can be used for golden eagles, with
Bird Conservation Regions (BCRs) defining the eagle management units. The example in Figure F‐1
involves bald eagles in Region 3. Project 1 (in green) has a footprint of 41 miles2 (mi2), and affects a
local‐area bald eagle population over 6854 mi2 (light green buffer around the project footprint).
Following the approach in Table F‐1, project 1 was issued a programmatic take permit with a
maximum annual project‐level take of 21 bald eagles per year (see Table F‐3). Project 2 (in red, the
same size as project 1) applied for a programmatic eagle take permit 5 years later. The calculated
project‐level bald eagle take for project 2 is 20 bald eagles per year, but under the 5% benchmark,
maximum take for 1563 mi2 of project 2’s local‐area bald eagle population (totaling 5 bald eagles
per year) was already allocated to project 1 (the hatched‐marked area of overlap between the local
areas of project 1 and project 2). Therefore, the calculated local‐area bald eagle take for project 2
exceeds the 5% benchmark. Thus, the decision‐maker for the permit for project 2 should carefully
consider whether this project can be permitted as designed under the requirements of our
regulations at 50 CFR 22.26.

The examples assume acceptable compensatory mitigation opportunities, when they are required,
are limitless. They are not, and where compensatory mitigation is necessary to offset the permitted
take, the availability of compensatory mitigation can become the proximate factor limiting take
opportunities.

A critical assumption of this approach is that eagle density is uniform across eagle regions. The
potential consequence of this assumption is to over protect eagles in areas of high density and
under protect them in areas of low density. As the Service and others develop more reliable models
for predicting the distribution of eagles within regional management populations at finer scales,
these approaches should be used in place of an assumption of uniform distribution in the analyses
suggested here.

82

Table F-1. Example of the proposed method to calculate local-area annual eagle take benchmarks. The
example uses bald eagles (BAEA), and is based on a hypothetical scenario where a single project with a circular
footprint of 10-mile radius is proposed in each BAEA region. See Figure F-1 for an example of how to assess the
cumulative effects of such permitted take over the local-area population.

BAEA
Management
Unit

Estimated
Population
Sizea

Region
Size (mi2)

R1

7105

245336

Maximum
Take Rate
(% local‐
area
population
per year)b
5.0

R2

797

565600

R3

27617

R4

Management
Unit Eagle
Density
(BAEA/ mi2)c

Local
Area
(mi2)d

Local‐area
5%
Benchmark
(eagles per
year)e

0.029

8824

13

5.0

0.001

8824

>1

447929

5.0

0.062

8824

27

13111

464981

5.0

0.028

8824

12

R5

14021

237687

5.0

0.059

8824

26

R6

5385

732395

5.0

0.007

8824

3

R7

86550

570374

5.0

0.152

8824

67

R8

889

265779

5.0

0.003

8824

1

Sum

155474

150

a Taken directly from USFWS (2009).

b A take rate of 5% is the Service’s upper benchmark for take at the local‐area population scale.
c Management unit eagle density = population size / management unit size.

d The local‐area for this example is the project footprint (in this case, a circle with radius of 10 miles) plus a

buffer of 43 additional miles (43 miles is the average natal dispersal distance for the BAEA) = 3.142 * 532 .
e The local‐area 5% benchmark = (Local‐area*Regional Eagle Density)*0.05.

83

Table F-2. Background information necessary to estimate the local-area take benchmarks for golden
eagles (GOEA). Columns are as in Table F-1. The local-area for golden eagles, which is not used in this table, is
calculated using the median natal dispersal distance of 140 miles (USFWS 2009).

GOEA Management Unit
Alaska
Northern Pacific Rainforest
Prairie Potholes
Sierra Nevada
Shortgrass Prairie
Coastal California
Sonoran and Mojave Desert
Sierra Madre Occidental
Chihuahuan Desert
Great Basin
Northern Rockies
Southern Rockies and
Colorado Plateau
Badlands and Prairies

Estimated
Population
Sizea

BCR Size
(mi2)b

Management Unit
Eagle Density
(GOEA per mi2)

5
11
15
18
32
33
34
35
9
10

2400
108
1680
84
1080
960
600
360
720
6859
6172

557007
68777
160794
20414
148540
63919
95593
47905
72455
269281
199666

0.0043
0.0016
0.0104
0.0041
0.0073
0.0150
0.0063
0.0075
0.0099
0.0255
0.0309

16

3770

199522

0.0189

17

7800

141960

0.0549

BCR
Number

Sum

32593

a Taken directly from USFWS 2009.

b BCR area values are from the North American Bird Conservation Region website at: http://www.bsc‐
eoc.org/international/bcrmain.html (last visited 8 December 2011).

84

Project 1

Project 2

60

0

60

120 Miles

Figure F-1. Example of the proposed method for ensuring local-area take benchmarks are not exceeded
through the cumulative take authorized over multiple projects. Project 1 is in green, project 2 is in red, and
the overlap in their local-area eagle bald eagle populations is the hatched-marked area (see text). This same
approach could be used to assess the cumulative effects of other forms of take and anthropomorphic impacts for
which data on population effects are available.

85

Table F-3. Calculations used to determine local-area bald eagle take for the example in Fig. F-1, where
project 1 is first-in-time, and the local-area bald eagle (BAEA) populations for the two projects overlap.
Calculations are as described in the footnotes to table F-1.

Project

Project 1 (first‐
in‐time)
Project 2,
unadjusted
Overlap Area
Project 2,
adjusted

Region 3
Region
BAEA
Size
Population
(mi2)
Size

Maximum
Take Rate
(% local‐
area
population
per year)b

Regional
Eagle
Density
(BAEA
per mi2)

Local‐
area
(mi2)

Local‐area
5%
Benchmark
(eagles per
year)e

27617

447929

5.0

0.062

6854

21

27617

447929

5.0

0.062

6550

20

27617

447929

5.0

0.062

1562

5

27617

447929

5.0

0.062

13404

15

Literature Cited
USFWS. 2007. Final environmental assessment, take of raptors from the wild under the falconry
regulations and the raptor propagation regulations. U.S. Fish and Wildlife Service, Division of
Migratory Bird Management, Washington, D.C.
USFWS. 2008. Final environmental assessment and management plan, take of migrant peregrine
falcons from the wild for use in falconry, and reallocation of nestling/fledgling take. U.S. Fish
and Wildlife Service, Division of Migratory Bird Management, Washington, D.C.
USFWS. 2009. Final environmental assessment, proposal to permit take as provided under the
Bald and Golden Eagle Protection Act. U.S. Fish and Wildlife Service, Division of Migratory Bird
Management, Washington, D.C.
USFWS. 2011. Draft eagle conservation plan guidance. U.S. Fish and Wildlife Service, Division of
Migratory Bird Management, Washington, D.C.

86

APPENDIX G: EXAMPLES USING RESOURCE EQUIVALENCY ANALYSIS TO ESTIMATE THE
COMPENSATORY MITIGATION FOR THE TAKE OF GOLDEN AND BALD EAGLES FROM WIND
ENERGY DEVELOPMENT
1. Introduction
This appendix provides Resource Equivalency Analysis (REA) examples developed by the Service to
illustrate the calculation of compensatory mitigation for the annual loss of golden (GOEA) eagles
and bald (BAEA) eagles caused by wind power if conservation measures and ACPs do not remove
the potential for take, and the projected take exceeds calculated thresholds for the species or
management population affected. These examples result in estimates of the number of high‐risk
electric power poles that would need to be retrofitted per eagle taken based on the inputs provided
below. Detailed explanatory documentation, literature, and supporting REA spreadsheets are now
located at: www.fws.gov/windenergy/index.html

As a framework for compensatory mitigation, it needs to be clear that the results provided below
are an illustration of how REA works given the current understanding of GOEA and BAEA life
history inputs, effectiveness of retrofitting high‐risk electric power poles, the expected annual take,
and the timing of both the eagle take permit and implementation of compensatory mitigation. As
would be expected, the estimated number of eagle fatalities and the permit renewal period affect
the number of poles to be retrofitted. Delays in retrofitting would lead to more retrofitted poles
owed. New information on changes in the level of take, understanding of the eagle life history, or
effectiveness of retrofitting could be used to change the number of retrofitted poles needed for
compensation. Finally, while only electric pole retrofitting is presented here in detail, the REA
metric of bird‐years lends itself to consideration of other compensatory mitigation options to
achieve the no‐net‐loss standard in the future. With enough reliable information, any
compensatory mitigation that directly leads to an increased number of GOEA and BAEA (e.g.,
habitat restoration) or the avoided loss of these eagles (e.g., reducing vehicle/eagle collisions,
making livestock water tanks ‘eagle‐safe’, lead ammunition abatement, etc.) could be considered for
compensation within the context of the REA.

2. REA Inputs
The best available peer‐reviewed, published data are provided in Tables G‐1 and G‐2. It should be
noted that additional modeling work within the REA may be needed, particularly on issues related
to migration, adult female survivorship, natal dispersal, age at first breeding, and population sex
ratio.

87

Table G-1. EXAMPLE INPUTS. REA Inputs to Develop a Framework of Compensatory Mitigation for Potential Take
of GOEA from Wind Energy Development

Parameter

REA Input

Reference

Start year of permit
Length of permit renewal
period
Estimated take

2012

Example.

5 years

Example.

1 eagle/year

Example.
28 years, 3 months, USGS Bird Banding
Lab.
Consistent with Cole (2010) approach.
 20% juveniles (age class (0‐1))
 35% sub‐adults (11.67% for each age
class from age class (1‐2) through age
class (3‐4))
 45% adults (1.73% for each age class
from age class (4‐5) through age class
(29‐30))
Assume age class is distributed evenly
over time. Age distribution derived
from models presented in USFWS 2009.

Average maximum
lifespan

Age distribution of birds
killed at wind facilities
(based on age
distribution of GOEA
population)

30 years

(0‐1)
(1‐4)
(4‐30)

Age start reproducing
Expected years of
reproduction
% of adult females that
reproduce annually
Productivity (mean
number of individuals
fledged per occupied nest
annually)
year 0‐1 survival
year 1‐2 survival
year 2‐3 survival
year 3‐4 survival
year 4+ survival
Relative productivity of
mitigation option

Discount rate

20%
35%
45%

Age 5
[age class (5‐6)]

Steenhof et al. 1984; Kochert et al. 2002

25 years

= (Maximum Lifespan) – (Age Start
Reproducing) (Harmata 2002)

80%

Steenhof et al. 1997

0.61

USFWS 2009

61%
79%
79%
79%
90.9%

USFWS 2009

Example. Compensatory mitigation
involves retrofitting high‐risk electric
0.0036 eagle
power poles, thus avoiding the loss of
electrocutions/pole/year
GOEA from electrocution (Lehman et al.
2010).
A 3% discount rate is commonly used
for valuing lost natural resource
3%
services (Freeman 1993, Lind 1982,
NOAA 1999; and court decisions on
damage assessment cases)

88

Table G-2. EXAMPLE INPUTS. REA Inputs to Develop a Framework of Compensatory Mitigation for Potential Take
of BAEA from Wind Energy Development

Parameter

REA Input

Start year of permit
Length of permit
renewal period
Estimated take

Relative productivity of
mitigation option

Discount rate

Example.

5 years

Example.

30 years

Age distribution of
birds killed at wind
(0‐1)
facilities (based on age (1‐4)
distribution of BAEA
(4‐30)
population)

Expected years of
reproduction
% of adult females that
reproduce annually
Productivity
year 0‐1 survival
year 1‐2 survival
year 2‐3 survival
year 3‐4 survival
year 4+ survival

2011
1 eagle/year

Average maximum
lifespan

Age start reproducing

Reference

15.4%
30%
54.6%

Age 5
[age class (5‐6)]
25 years

Example.
32 years 10 months; Longevity record
from USGS Bird Banding Lab. Consistent
with Cole (2010) approach.
 15.4% juveniles (age class (0‐1))
 30% sub‐adults (10% for each age
class from age class (1‐2) through age
class (3‐4))
 54.6% adults (2.1% for each age class
from age class (4‐5) through age class
(29‐30))
Assume age class is distributed evenly
over time. Age distribution derived
from models presented in USFWS 2009.
Buehler 2000
= (Maximum Lifespan) – (Age Start
Reproducing)

42%

Hunt 1998, per. comm. Millsap

1.3
77%
88%
88%
88%
83%

Millsap et al. 2004
Millsap et al. 2004

Example. Mitigation involves
0.0036 eagle
retrofitting high‐risk electric power
electrocutions/pole/year poles, thus avoiding the loss of BAEA
from electrocution (Lehman et. al 2010).
A 3% discount rate is commonly used
for valuing lost natural resource
3%
services (Freeman 1993; Lind 1982;
NOAA 1999; and court decisions on
damage assessment cases).

89

3. REA Example – WindCoA
The Service developed the following hypothetical scenario for permitting and compensatory
mitigation to be applied to the take of GOEA1 from wind power operations:

WindCoA conducted three years of pre‐construction surveys to determine relative abundance of
GOEA at their proposed wind project in Texas. The survey data was then used to populate a risk
assessment model to generate an eagle fatality estimate. The initial fatality estimate of two eagles
per year was further reduced after WindCoA implemented a few mutually agreed upon ACPs. The
final fatality estimate generated from the risk assessment model, after consideration of the
advanced conservation practices, was an annual take of one GOEA per year over the life of the
permit starting in 2012.

WindCoA decided to conduct an REA to determine the number of high‐risk power poles that would
need to be retrofitted to get to no‐net‐loss. The company used the Service’s GOEA REA inputs and
assumed the power pole retrofit would occur in calendar year 2012, thus offsetting the potential
loss of eagles at the newly operating wind project with avoidance of electrocution of an equal
number of GOEA. Through proper operation and maintenance (O&M), the retrofitted poles are
assumed to be effective in avoiding the loss of eagles for 10 years. The results of the model are
expressed in the total number of electric power poles to be retrofitted to equate to no‐net‐loss of 5
eagles for the 5‐year permit renewal period (1 eagle annually over five years). These results are
extrapolated over the expected operating life of the wind project, which is assumed to be 30 years,
for a total take of 30 eagles.

The results of the REA indicated that WindCoA needed to retrofit approximately 149 power poles
for the first 5‐year permit period (see Table G‐3). Using an estimated cost of $7500/pole, the
Service estimated that WindCoA could contribute $1,117,500 to a third‐party mitigation account or
contract the retrofits directly. After determining that they could fund the retrofits directly at a
lower cost, WindCoA decided to partner with UtilityCoB to get the required number of poles
retrofitted. UtilityCoB had previously conducted a risk assessment of their equipment and had
identified high‐risk poles that were likely to take golden eagles. Through a written agreement,
WindCoA provided funding to UtilityCoB to retrofit the required number of power poles and
maintain the retrofits for 10 years. In addition, WindCoA contracted with ConsultCoC to perform
effectiveness monitoring of the retrofitted power poles for 2 years. The contract required that
ConsultCoC visit each retrofitted power pole every 4 months (quarterly) to perform fatality
searches and check for proper operation and maintenance of the equipment. The Service reviewed
the compensatory mitigation project proposed by WindCoA and found it to be consistent with
requirements at 50 CFR 22.26. After reviewing the signed contract between WindCoA, UtilityCoB,
and ConsultCoC, the Service issued a programmatic eagle take permit to WindCoA.

a. REA Language and Methods
As discussed in greater detail in documents on the supporting website, this REA includes:

 The direct loss of GOEA/BAEA eagles from the take (debit in bird‐years);
 The relative productivity of retrofitting high‐risk power poles, which is the
effectiveness in avoiding the loss of GOEA/BAEA by electrocution as a mitigation
offset (measured in total bird‐years per pole); and

1

Using the inputs provided in Table G-2, this scenario may also be applied to BAEA.

90


The mitigation owed, which is the total debit divided by the relative productivity
(scaling) to identify the number of high‐risk power poles that need retrofitting to
completely offset the take of GOEA/BAEA eagles (credit).

There are up to 16 steps when conducting a REA. Depending on whether foregone future
reproduction (part of the debit) is included, there are up to 13 total steps involved in
calculating the injury side (debit) of a REA, and three additional steps involved in estimating
compensatory mitigation owed (credit). Please refer to the technical note “Scaling Directly
Proportional Avoided Loss Mitigation/Restoration Projects” on the supporting website
(www.fws.gov/windenergy) for further information on the development of REA inputs and
the inclusion of lost reproduction. Notably, in the case of an avoided loss project where the
estimated prevented loss of bird‐years (e.g., through mitigation) is directly proportional to
the loss of bird‐years (e.g., from “take”), the life history inputs (e.g., longevity, age
distribution, survival rates, reproduction) do not affect the final results of the credit owed.
That is, the retrofitting of high‐risk power poles is a directly proportional avoided loss, so
only the level of take (number of eagles annually), the avoided loss of eagles per mitigated
electric pole, the number of years the mitigated pole is effective in avoiding the loss of
eagles, and the timing of the mitigation relative to the take affect the final credit owed. It
should also be noted that the annual take of one eagle is used in the example because the
lost bird‐years associated with one eagle can be easily multiplied by the actual take to
estimate the total debit in bird‐years.

The following is a brief discussion of REA variables used in the Service’s WindCoA example
that affect the outcome of the compensatory mitigation calculation:
 Relative Productivity of Mitigation (0.0036 electrocutions/pole/year) – This
rate is taken directly from published literature on eagle electrocution rates in
northeastern Utah and northwestern Colorado and is specific to eagles (Lehman et
al. 2010). Although the referenced study also lists a higher rate (0.0066) that
includes all known eagle mortalities, this rate included eagles that may have died
from causes unrelated to electrocution.
 Years of Avoided Loss Per Retrofitted Pole (10 Years) – The Service uses a
period of 10 years for crediting the project developer or operator for the avoided
loss of eagles from power pole retrofits. This is a reasonable amount of time to
assume that power pole retrofits will remain effective. However, project developers
or operators should consider entering into agreements with utility companies or
contractors for the long‐term maintenance of retrofits. Evidence of this type of
agreement could increase the amount of credit received by the project developer or
operator and, as a result, decrease the amount of compensatory mitigation required.
 Permit Renewal Period (5 Years) – This will be the review period that is used by
the Service for adaptive management purposes and re‐calculation of compensatory
mitigation. The Service believes that this length of time will enable the project
developer or operator to continue to meet the statutory and regulatory eagle
preservation standard. This permit review tenure will remain the same regardless
of the overall tenure of the permit.
 Retrofit Cost/Payment ($7,500/pole) – The Service received input directly from
the industry regarding the actual costs to retrofit power poles. Estimates ranged
from a low of approximately $400 to over $11,000 given that costs vary according to
many factors. The Service believes that $7,500 represents a reasonable estimate for
the current cost to retrofit power poles in the United States. Project developers or

91
operators are encouraged to contract directly for retrofits as this will likely not be as
costly as contributing $7,500/pole to an eagle compensatory mitigation account.

b. REA Results for WindCoA
Using the WindCoA example described above, along with the REA inputs provided in Table
G‐1, Table G‐3 provides a summary of the results:

Table G-3. WindCoA Example: Compensatory Mitigation Owed for a 5-Year Permitted Take of 5 GOEA
Extrapolated to the 30-Year Expected Operating Life of the Wind Project (30 GOEA in Total).
Total Debit for Take of 1 GOEA

28.485

÷Relative Productivity of High‐
Risk Electric Pole Retrofitting

÷0.191

= Mitigation Owed for 5‐Year
Permitted Take
x # Cycles of 5‐Year Permit
Reviews
=Total Mitigation Owed
*PV=Present Value

=149.136
x 6 = 894.818

PV* bird‐years for 5 years of GOEA take
Avoided loss of PV bird‐years per
retrofitted pole
(assumes 10 years of avoided loss per pole
based on the commitment from
UtilityCoB)
Poles to be retrofitted to achieve no‐net‐
loss
Poles to be retrofitted to achieve no‐net‐
loss for the 30‐year expected operating life
of the wind project

If all of the REA inputs remain the same after the initial five years, then the estimated
149.14 poles may be multiplied by the expected number of permit reviews to provide an
estimate of the total number of poles that would eventually be retrofitted. For example, for
the 30‐year life cycle of the WindCoA wind project, 149.14 poles would be multiplied by 6
permit renewals to equal approximately 895 high‐risk power poles in total to be retrofitted
as compensatory mitigation for the take of 30 GOEA over 30 years (1 eagle annually). While
this example shows the effectiveness of the mitigation method as lasting for 10 years, it may
be the case that the method selected is more or less effective at avoiding the loss of eagles
(e.g., 5 years, more than 10 years). The REA can be adjusted for the expected effectiveness
of mitigation, and more or fewer high‐risk power poles would need to be mitigated. All
estimates of compensatory mitigation are contingent on proper operation and maintenance
being conducted by UtilityCoB or a contractor to ensure that the expected effectiveness is
achieved.

For purposes of illustration, should WindCoA choose to use the GOEA inputs provided in
Table G‐1 and their fatality estimate is that 5 GOEA will be taken annually, the results may
be easily adjusted as shown in Table G‐4:

92

Table G-4. WindCoA Example: Compensatory Mitigation Owed for a 5-Year Permitted Take of 25 GOEA
Extrapolated to the 30-Year Expected Operating Life of the Wind Project (150 GOEA in Total).
Total Debit for Take of 1 GOEA

28.485

PV bird‐years for 5 years of GOEA take
from Table F‐3

x Actual Annual Take of GOEA

x 5 =142.425

PV bird‐years for 5 years of GOEA take

÷ Relative Productivity of High‐
Risk Electric Pole Retrofitting
= Mitigation Owed for 5‐Year
Permitted Take
x # Cycles of 5‐Year Permit
Reviews = Total Mitigation
Owed
PV=Present Value

÷0.191
=745.681
x 6 =4474.086

Avoided loss of PV bird‐years per
retrofitted pole (assumes 10 years of
avoided loss per pole based on the
commitment from UtilityCoB)
Poles to be retrofitted to achieve no‐net‐
loss
Poles to be retrofitted to achieve no‐net‐
loss for the 30‐year expected operating life
of the wind project

c. Summary of Bald Eagle REA Results
Following the same process described above for GOEA (i.e., using the WindCoA example and
the BAEA REA inputs provided in Table G‐2), Table G‐5 provides a summary of the results
for bald eagles:

Table G-5. Example of Compensatory Mitigation Owed for a 5-Year Permitted Take of 5 BAEA Extrapolated to the
30-Year Expected Operating Life of the Wind Project (30 BAEA in Total).
Total Debit for Take of 1 BAEA
÷ Relative Productivity of High‐
Risk Electric Pole Retrofitting
= Mitigation Owed for 5‐Year
Permitted Take
x # Cycles of 5‐Year Permit
Reviews = Total Mitigation
Owed
PV=Present Value

20.229
÷0.136
=149.136
x 6 =894.818

PV bird‐years for 5 years of BAEA take
Avoided loss of PV bird‐years per
retrofitted pole
Poles to be retrofitted to achieve no‐net‐
loss
Poles to be retrofitted to achieve no‐net‐
loss for the 30‐year expected operating life
of the wind project

Although there are differences between GOEA and BAEA life history inputs (e.g., longevity,
age distribution, survival rates, reproduction), the estimated avoided loss of bird‐years
through mitigation is directly proportional to the loss of bird‐years from the take, so the life
history inputs do not affect the final results of the credit owed. Because there was no
change in the level of take (number of eagles annually), the avoided loss of eagles per

93
mitigated electric pole, the number of years the mitigated pole is effective in avoiding the
loss of eagles, or the timing of the mitigation relative to the take, there is no change in the
credit owed. To help illustrate, when comparing the results of BAEA to GOEA, both the
debit (20.23÷28.49) and the relative productivity of electric pole retrofitting (0.14÷0.19) for
BAEA are approximately 70% of GOEA, so the amount of retrofitting owed is the same. That
is, both the numerator of the scaling equation (total debit) and the denominator (relative
productivity of mitigation) were changed proportionally (approximately 70%), so there is
no change in the mitigation owed.

d. Discussion on Using REA
The ECPG does not mandate the use of REA. Rather, the Service recognized the need for a
reliable, transparent, reproducible, and cost‐effective tool to expedite wind power permits,
while ensuring sufficient compensatory mitigation for the take of golden eagles and bald
eagles from operations to meet regulatory permitting requirements. Although there is a
learning curve, REA meets these basic needs. This appendix and materials on the
supporting website explain the methods, share the tools to run REAs, and discuss how
changes in the different inputs can affect the results. Should project developers or
operators/applicants choose to use the provided inputs, methods, and tools, the Service will
be able to appropriately focus on the expected take of eagles. Project developers or
operators/applicants have the discretion to offer alternative REA inputs or use different
compensatory mitigation modeling methods. However, they will need to provide sufficient
evidence and tools (if necessary) to ensure that the Service can provide appropriate review
of the results, and should expect that such an effort will likely take additional time.

e. Additional Compensatory Mitigation Example
In the United States, another known cause of mortality to eagles, both bald and golden, is
vehicle collisions. Eagles are susceptible to being struck by vehicles as they feed on
carcasses along roadsides, particularly in areas of the United States where large numbers of
ungulates concentrate seasonally (e.g. winter, breeding season, etc.). As a compensatory
mitigation strategy, a project developer or operator may decide to collect data (or use
existing data if it is available) on the annual number of eagle mortalities that result from
vehicle collisions in a specified geographic area or along a specific stretch of roadway. This
data could then be used to generate an estimate of the number of eagle mortalities that
could be prevented in the same area by removing carcasses from roadsides. If there was
sufficient evidence that this was a valid project (e.g. quantifiable and verifiable), the project
developer or operator could contract to have these roadsides ‘cleaned’ of carcasses during
the time of year that ungulates concentrate and eagles are known to be struck. The credible
estimate of eagle mortalities that would be avoided through carcass removal would be the
value of the compensatory mitigation achieved.

f. Take from Disturbance
Project developers or operators should work with the Service to determine if take from
disturbance is likely to occur. This should be predicted in advance based on Stage 3 data,
and verified through post‐construction monitoring in Stage 5. The following are
recommended take calculations based on information contained within the FEA (USFWS
2009):

For the standard bald eagle population:

94



Take resulting from disturbance at one nest on only one occasion = take of 1.3
individuals
One nest take resulting in the permanent abandonment of a territory = take of 1.3
individuals for the first year, then take of 8 individuals annually until data show the
number of breeding pairs has returned to or exceeded the original estimated
number for the eagle management unit.

For the standard golden eagle population:
 Take resulting from disturbance at one nest on only one occasion = take of 0.8
individuals
 One nest take resulting in the permanent abandonment of a territory = take of 0.8
individuals for the first year, then take of 4 individuals annually until data show the
number of breeding pairs has returned to or exceeded the original estimated
number for the eagle management unit.

Using the data presented in the above WindCoA example, the compensatory mitigation
required for disturbance resulting in the loss of productivity from one GOEA nest for one
year would result in the following:

1. Disturbance take of one GOEA nest on one occasion = 0.8 GOEA,
2. From the REA, the take of one GOEA for one year = 6 PV bird‐years,
3. Six PV bird‐years/GOEA * 0.8 GOEA = 4.8 PV bird‐years, and
4. From the REA, 4.8 PV bird‐years ÷ 0.191 PV bird‐years/pole retrofitted (for 10 year
maintenance of poles) = 25.1 poles retrofitted.

WindCoA would be required to retrofit a total of 174.24 poles (149.14 poles for the lethal
take of 5 GOEA (see Table G‐3) + 24.5 poles for the disturbance take of one GOEA nest) to
cover the initial five year permitted take.

Literature Cited
Buehler, D. A. 2000. Bald Eagle (Haliaeetus leucocephalus), The Birds of North America Online (A.
Poole, Ed.). Ithaca: Cornell Lab of Ornithology; Retrieved from the Birds of North America
Online: http://bna.birds.cornell.edu/bna/species/506.
Cole, S. 2010. How much is enough? Determining adequate levels of environmental compensation
for wind power impacts using resource equivalency analysis: An illustrative and hypothetical
case study of sea eagle impacts at the Smola Wind Farm, Norway. Epsilon Open Archive
Publishing, Swedish Agricultural University.
Freeman, A.M. III. 1993. The Measurement of Environmental and Resource Values: Theory and
Methods. (Resources for the Future, Washington, DC).
Harmata, A. R. 2002. Encounters of Golden Eagles banded in the Rocky Mountain West. J. Field
Ornithol. 73:23‐32.
Hunt, W.G. 1998. Raptor floaters at Moffat’s equilibrium. Oikos 81:1‐7.
Kochert, M. N., K. Steenhof, C. L. Mcintyre and E. H. Craig. 2002. Golden Eagle (Aquila chrysaetos),
The Birds of North America Online (A. Poole, Ed.). Ithaca: Cornell Lab of Ornithology; Retrieved
from the Bird of North America Online: http://bna.birds.cornell.edu/bna/species/684.
Lehman, R. N., Savidge, J. A., Kennedy, P. L. and Harness, R. E. (2010), Raptor Electrocution Rates for
a Utility in the Intermountain Western United States. Journal of Wildlife Management, 74: 459‐
470.

95
Lind, R. 1982. A Primer on the Major Issues Relating to the Discount Rate for Evaluating National
Energy Options in Discounting for Time and Risk in Energy Policy, edited by R. Lind.
Washington: Resources for the Future.
Millsap, B.A., T. Breen, E. McConnell, T. Steffer, L. Phillips, N. Douglass, S. Taylor. 2004. Comparative
fecundity and survival of bald eagles fledged from suburban and rural natal areas in Florida.
Journal of Wildlife Management 68:1018‐1031.
NOAA. 1999. Discounting and the Treatment of Uncertainty in Natural Resource Damage Assessment.
Technical Paper 99‐1 (Silver Spring, MD: NOAA).
Steenhof, K., M.N. Kochert, and M. Q. Moritsch. 1984. Dispersal and migration of southwestern Idaho
raptors. J. Field Ornithol. 55: 357‐368.
Steenhof, K., M. N. Kochert, and T. L. McDonald. 1997. Interactive effects of prey and weather on
Golden Eagle reproduction. J. Anim. Ecol. 66: 350‐362.
USFWS. 2009. Final environmental assessment. Proposal to permit take provided under the Bald
and Golden Eagle Protection Act. U.S. Fish and Wildlife Service, Division of Migratory Bird
Management, Washington D.C., USA.

96

APPENDIX H: STAGE 5 – CALIBRATING AND UPDATING OF THE FATALITY PREDICTION AND
CONTINUED RISK-ASSESSMENT

Given the degree of uncertainty that currently exists surrounding the risk of wind facilities to eagles
and the factors that contribute to that risk, post‐construction monitoring is one of the most
significant activities that will be undertaken by eagle programmatic take permit holders. Post‐
construction monitoring has two basic components when applied to eagle take: (1) estimating the
mean annual fatality rate, and (2) assessing possible disturbance effects on neighboring nests and
communal roosts. Provided that assessments conducted during Stages 1‐4 are consistent, robust,
and reliably performed as suggested in this ECPG, the pre‐construction data should provide a solid
platform for development of the Stage 5 monitoring and assessment studies.

1. Fatality Monitoring
All wind facilities that are permitted to take eagles will need to conduct fatality monitoring to
ensure compliance with regulatory requirements. Fatality monitoring must be conducted at all
wind facilities that are permitted to take eagles. We anticipate that in most cases, intensive
monitoring to estimate the true annual fatality rate and to assess possible disturbance effects will
be conducted for at least the first two years after permit issuance, followed by less intense
monitoring for up to three years after the expiration date of the permit, in accordance with
monitoring requirements at 50 CFR 22.26(c)(2). However, additional intensive, targeted
monitoring may be necessary to determine the effectiveness of additional conservation measures
and ACPs implemented to reduce observed fatalities. Such monitoring should be rigorous and
sufficient to yield a reasonable estimate of the mean annual eagle fatality rate for the project.
General considerations for designing fatality monitoring programs can be found in Strickland et al.
(2011) and the WEG, and these sources should be consulted in the development of a post‐
construction study design. Because the post‐construction monitoring protocol will be included as a
condition of the programmatic take permit, the design of such monitoring will be determined
jointly by the permittee and the Service. Additionally, the Service and USGS are investing significant
resources into research to test and assess post‐construction monitoring approaches for eagles, thus
we expect to be able to offer useful input in the design of such monitoring programs. Fatality
monitoring for eagles can be combined with monitoring mortality of other wildlife so long as
sampling intensity takes into account the relative infrequency of eagle mortality events.

Fatality‐monitoring efforts involve searching for eagle carcasses beneath turbines and other
facilities to estimate the number of fatalities. The primary objectives of these efforts are to: (1)
estimate eagle fatality rates for comparison with the model‐based predictions prior to construction,
and (2) to determine whether individual turbines or strings of turbines are responsible for the
majority of eagle fatalities, and if so, the factors associated with those turbines that might account
for the fatalities and which might be addressed via conservation measures and ACPs.

Fatality monitoring results should be of sufficient statistical validity to provide a reasonably precise
estimate of the eagle mortality rate at a project to allow meaningful comparisons with pre‐
construction predictions, and to provide a sound basis for determining if, and if so which,
conservation measures and ACPs might be appropriate. The basic method of measuring fatality
rates is the carcass search. All fatality monitoring should include estimates of carcass removal and
carcass detection bias (scavenger removal and searcher efficiency) likely to influence those rates,
using the currently accepted methods. Fatality and bias correction efforts should occur across all
seasons to assess potential temporal variation. Where seasonal eagle concentrations were

97
identified in the Stage 2 assessment, sampling protocols should take these periodic pulses in
abundance into account in the sample design.

Carcass searches underestimate actual mortalities at wind turbines, but with appropriate sampling,
carcass counts can be adjusted to account for biases in detection (Kunz et al. 2007, Arnett et al.
2007, NRC 2007, Huso 2010). Important sources of bias and error include: (1) low or highly
variable fatality rates; (2) carcass removal by scavengers; (3) differences in searcher efficiency; (4)
failure to account for the influence of site (e.g., vegetative) conditions in relation to carcass removal
and searcher efficiency; and (5) fatalities or injured birds that may land or move outside search
plots. Strickland et al (2011) provide a concise overview of fatality prediction models and
considerations in the selection of a model. In the case of eagles, a primary consideration in the
selection of a model and in the sampling design is the relative rarity of collisions, even at sites
where fatality rates are comparatively high.

Regardless of the approach selected, we recommend the following data be collected for each search:
1. Date.
2. Start time.
3. End time.
4. Interval since last search.
5. Observer.
6. Which turbine area was searched (including decimal‐degree latitude longitude or UTM
coordinates and datum).
7. Weather data for each search, including the weather for the interval since the last search.
8. GPS track of the search path.

When a dead eagle is found, the following information should be recorded on a fatality data sheet:
1. Date.
2. Species.
3. Age and sex (following criteria in Pyle 2008) when possible.
4. Band number and notation if wearing a radio‐transmitter or auxiliary marker.
5. Observer name.
6. Turbine or pole number or other identifying character.
7. Distance of the carcass from the turbine or pole.
8. Azimuth of the carcass from the turbine or pole.
9. Decimal‐degree latitude longitude or UTM coordinates of the turbine or pole and carcass.
10. Habitat surrounding the carcass.
11. Condition of the carcass (entire, partial, scavenged).
12. Description of the carcass (e.g., intact, wing sheared, in multiple pieces).
13. A rough estimate of the time since death (e.g., <1 day, > a week), and how estimated.
14. A digital photograph of the carcass.
15. Information on carcass disposition.

In some cases, eagle take permits may specify other biological materials or data that should be
collected from eagle carcasses (e.g., feathers, tissue samples). Rubber gloves should be used to
handle all carcasses to eliminate possible disease transmission. All eagle fatalities (not just those
found on post‐construction surveys) and associated information should be immediately reported to
the Service’s Office of Law Enforcement and to the Service’s migratory bird permit issuing office if
the facility is operating under an eagle take permit. Eagle carcasses should not be moved until such
notification occurs, after which carcass disposition should be in accordance with permit conditions
or Service direction.

98

2. Disturbance Monitoring
Project developers or operators may also be required to monitor many of the eagle nesting
territories and communal roost sites identified in the Stage 2 assessments as stated in the permit
regulations at 50 CFR 22.26(c)(2)for at least two years after project construction and for up to
three years after the cessation of the activity. The objective of such monitoring will be to determine
post‐construction (1) territory or roost occupancy rates, (2) nest success rates, and (3)
productivity. On a project‐by‐project basis, changes in any of these reproductive measures may not
be indicative of disturbance. However, patterns may become apparent when the Service and USGS
pool data appropriately and analyze findings from many projects in the context of a meta‐analysis
within the adaptive management framework.

Eagle nesting territories most likely to be affected by disturbance from a wind project are those that
have use areas within or adjacent to the project footprint. The Service will accept an assumption
that all eagle pairs at or within the mean project‐area inter‐nest distance (as determined from the
Stage 2 assessment) of the project boundary are territories that may be at risk of disturbance (e.g.,
if the mean nearest‐neighbor distance between simultaneously occupied eagle territories in the
Stage 2 assessment is 2 miles, we would expect disturbance to most likely affect eagles within 2
miles of the project boundary; Figures H‐1 though H‐4). Eagle pairs nesting within ½ the project‐
area mean intern‐nest distance are the highest candidates for disturbance effects, and should
receive special attention and consideration.

Where nesting habitat is patchy or eagle nesting density is low such that nearest‐neighbors are
outside a 10‐mile wide perimeter of the project footprint, we recommend either: (1) extending the
project‐area survey outward to include the nearest‐neighbors for the purposes of estimating the
mean inter‐nest distance value, or (2) undertaking detailed observational studies of the eagles
occupying territories within the typical project‐area to assess use patterns and ranging behavior
relative to the project footprint. We recognize that selecting option (1) for golden eagles would
extend the project area beyond the maximum of 10 miles advocated in the ECPG, but in some areas
it is possible golden eagles using nests further than 10 miles from the project footprint may occur
there. Regardless of which approach is used, territories that meet this distance criterion should be
re‐sampled annually for no less than two years after the project is operational following identical
survey and reporting procedures as were used in the Stage 2 assessment.

If such monitoring shows strong evidence of direct disturbance from a project, project developers
or operators and the Service will consider additional conservation measures and ACPs that might
be effective in reducing the effect. Such measures would be within the sideboards established at
the time of permit issuance. Alternatively, the project developer or operator may be required to
provide compensatory mitigation to offset the estimated decreases in productivity to the extent
necessary to meet the statutory requirement to preserve eagles.

The Service and the project developer or operator should agree on a site‐specific, post‐construction
survey protocol for eagle concentration areas identified in Stage 2 and make an a priori decision on
how to interpret and act on potential outcomes. Mortalities of eagles using proximate communal
roosts will be accounted for through the protocol for monitoring post‐construction fatalities.
However, if communal roosts are no longer used by eagles because of disturbance, that effect
should be determined, evaluated, and where population‐level effects are indicated, mitigated.

99

3. Comparison of Post-Construction Eagle Use with Pre-Construction Use
As noted elsewhere, Service fatality models assume eagle use of the project footprint does not
change as a result of project development. However, there is little information to support this
assumption, and the ability to accurately predict fatality rates could be greatly improved by
comparative information on post‐construction eagle use. The Service encourages project
developers or operators to consider conducting exposure surveys similar in design and intensity to
pre‐construction survey work to test this assumption where and when feasible.

Literature Cited
Arnett, E. B. 2006. A preliminary evaluation on the use of dogs to recover bat fatalities at wind
energy facilities. Wildlife Society Bulletin 34(5):1440–1445.
Huso, M. M. P. 2010. An estimator of wildlife fatality from observed carcasses. Environmetrics DOI:
10.1002/env.1052.
Kochert, M. N., K. Steenhof, C. L. Mcintyre, and E. H. Craig. 2002. Golden eagle (Aquila chrysaetos).
The Birds of North America No. 684 (A. Poole, Ed.). The Birds of North America Online. Cornell
Lab of Ornithology, Ithaca, New York, USA. http://bna.birds.cornell.edu/bna/species/684.
Kunz, T. H., E. B. Arnett, B. M. Cooper, W. P. Erickson, R. P. Larkin, T. Mabee, M. L. Morrison, M. D.
Strickland, and J. M. Szewczak. 2007. Assessing impacts of wind‐energy development on
nocturnally active birds and bats: a guidance document. Journal of Wildlife Management 71:
2449‐2486.
National Research Council (NRC). 2007. Environmental impacts of wind‐energy projects. National
Academies Press. Washington, D.C., USA. www.nap.edu.
Strickland, M.D., E.B. Arnett, W.P. Erickson, D.H. Johnson, G.D. Johnson, M.L., Morrison, J.A. Shaffer,
and W. Warren‐Hicks. 2011. Comprehensive Guide to Studying Wind Energy/Wildlife
Interactions. Prepared for the National Wind Coordinating Collaborative, Washington, D.C., USA.

Figures H-1 to H-4 (following pages). Suggested approach for determining project-area and identifying eagle
nesting territories to monitor for disturbance effects during Stage 5.

100

101

102

103

U.S. Fish and Wildlife Service
Division of Migratory Bird Management
April 2013