ArticlePDF Available

Development of a practical modeling framework for estimating the impact of wind technology on bird populations

Authors:

North Carolina State University

One of the most pressing environmental concerns related to wind project development is the potential for avian fatalities caused by the turbines. The goal of this project is to develop a useful, practical modeling framework for evaluating potential wind power plant impacts that can be generalized to most bird species. This modeling framework could be used to get a preliminary understanding of the likelihood of significant impacts to birds, in a cost-effective way. The authors accomplish this by (1) reviewing the major factors that can influence the persistence of a wild population; (2) briefly reviewing various models that can aid in estimating population status and trend, including methods of evaluating model structure and performance; (3) reviewing survivorship and population projections; and (4) developing a framework for using models to evaluate the potential impacts of wind development on birds.

…

Figures - uploaded by Kenneth H Pollock

Content may be subject to copyright.

Content uploaded by Kenneth H Pollock

Content may be subject to copyright.

November 1997 Ÿ NREL/SR-440-23088

Development of a

Practical Modeling Framework

for Estimating the Impact of

Wind Technology on

Bird Populations

Michael L. Morrison

California State University

Sacramento, California

Kenneth H. Pollock

North Carolina State University

Raleigh, North Carolina

National Renewable Energy Laboratory

1617 Cole Boulevard

Golden, Colorado 80401-3393

A national laboratory of the U.S. Department of Energy

Managed by Midwest Research Institute

for the U.S. Department of Energy

under contract No. DE-AC36-83CH10093

November 1997

NREL/SR-440-23088 Ÿ UC Category: 1210

Development of a

Practical Modeling Framework

for Estimating the Impact of

Wind Technology on

Bird Populations

Michael L. Morrison

California State University

Sacramento, California

Kenneth H. Pollock

North Carolina State University

Raleigh, North Carolina

NREL technical monitor: Karin Sinclair

National Renewable Energy Laboratory

1617 Cole Boulevard

Golden, Colorado 80401-3393

A national laboratory of the U.S. Department of Energy

Managed by Midwest Research Institute

for the U.S. Department of Energy

under contract No. DE-AC36-83CH10093

Work performed under Subcontract No. CCD-5-15367-01

November 1997

NOTICE

This report was prepared as an account of work sponsored by an agency of the United States government.

Neither the United States government nor any agency thereof, nor any of their employees, makes any

warranty, express or implied, or assumes any legal liability or responsibility for the accuracy,

completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents

that its use would not infringe privately owned rights. Reference herein to any specific commercialproduct,

process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute

or imply its endorsement, recommendation, or favoring by the United States government or any agency

thereof. The views and opinions of authord expressed herein do not necessarily state or reflect those of

the United States government or any agency thereof.

Available to DOE and DOE contractors from:

Office of Scientific and Technical Information (OSTI)

P.O. Box 62

Oak Ridge, TN 37831

Prices available by calling (423) 576-8401

Available to the public from:

National Technical Information Service (NTIS)

U.S. Department of Commerce

5285 Port Royal Road

Springfield, VA 22161

(703) 487-4650

Printed on paper containing at least 50% wastepaper, including 10% postconsumer waste

iii

Foreword

One of the most pressing environmental concerns related to wind project development is the

potential for avian fatalities caused by the turbines. In order to understand the potential for this

problem, pre-construction avian surveys are often required. However, these can be expensive and

require significant amounts of time in order to gather the empirical data necessary to fully understand

the potential impacts. Then, the results may suggest that moving forward with such a project in the

wind resource area just evaluated may be ill advised. This report describes a modeling framework

for evaluating potential windplant impacts that can be generalized to most bird species. This

modeling framework could be used to get a preliminary understanding of the likelihood of significant

impacts to birds in a cost-effective way.

Drs. Morrison and Pollock produced this report under a subcontract with the National Wind

Technology Center of the National Renewable Energy Laboratory, with funding from the Wind

Program at the U. S. Department of Energy.

Karin C. Sinclair

National Wind Technology Center

National Renewable Energy Laboratory

1617 Cole Boulevard

Golden, Colorado 80401

Internet Address: karin_sinclair@nrel.gov

Phone: 303 - 384 - 6946

Fax: 303 - 384 - 6901

EXECUTIVE SUMMARY

The goal of this project is to develop a useful, practical modeling framework for evaluating

potential wind power plant impacts that can be generalized to most bird species. We accomplish

this by (1) reviewing the major factors that can influence the persistence of a wild population; (2)

briefly reviewing various models that can aid in estimating population status and trend, including

methods of evaluating model structure and performance; (3) reviewing survivorship and

population projections; and (4) developing a framework for using models to evaluate the potential

impacts of wind development on birds.

We begin with a review of demography. Demography is the study of population statistics,

including births, deaths, immigration, and emigration. From demography, we know that

conditions leading to extinction are most likely to occur in small populations. Demographic rates

vary because individuals do not survive for the same length of time, individuals vary in the

number of offspring they bear, individuals often have low birth rates, and so forth. Adult

survivorship is usually very high, especially in long-lived species (such as raptors). Therefore,

estimating adult survivorship tells one a lot about population status. In addition, in most

monogamous species, it is female survivorship that is most important to population persistence.

At a minimum, then, quantifying adult survivorship provides a preliminary, basic indication of the

status of the population. Modeling genetics is not likely to be as important as modeling

demographic and ecological processes in evaluating population persistence. This is based, in

part, on the lack of our sufficient understanding of genetics to use it as a basis for management.

Thus, practical considerations were the overriding factor in this conclusion. Still, genetics may

be a priority in small, isolated populations.

Random environmental events such as catastrophic fires, hurricanes, and disease can also have

pronounced effects on small populations. Such factors can also have pronounced effects on

large populations that are spatially divided into subpopulations. Here, factors such as dispersal

will determine the fate of a subpopulation driven to very low numbers, or even to extinction, by a

catastrophic event. Thus, the relative importance of environmental stochasticity must be based

on an understanding of the spatial distribution of the population under study.

Next we review the parameters necessary to develop rigorous population-projection models.

Life-history parameters are essential components of population-projection models. The

characteristics that we collectively call life-history parameters of animals include quantifiable

longevity, lifetime reproductive output, the young produced per breeding attempt, the age of

dispersal, survivorship, sex ratio, and the time between breeding attempts. For example,

combining various ranges of parameters can yield substantially different rates of population

change. Such analyses provide guidance on whether the population can be sustained under

varying expressions of life history traits. Once such relationships are understood, researchers

have the opportunity to monitor selected life history traits as part of an assessment of the status

of a population.

A central part of impact assessment--such as in wind power plants--is developing a model that

estimates the survival rates required to maintain a constant population. The strategy is to

determine the survival rates required to sustain the populations that exhibit the various

combinations of the other parameters governing population size. To be useful in a wide range of

environmental situations and useable for people with varying expertise, the model should be

based on simple mathematics.

Leslie matrix and similar stage-structured models can give great insight into the processes of

population growth. For example, the sensitivity of the population growth rate, r, to perturbations

in vital rates for a Leslie-type model can be solved analytically. Understanding how growth rate

changes in response to perturbations at various stages in the life table may help direct

management strategies. For example, adult survival tends to be a parameter to which a model is

extremely sensitive in long-lived species, whereas fecundity can be more important in short-lived

species.

To aid in providing general guidelines concerning the potential impacts of wind developments on

bird populations, we developed Leslie matrix models and conducted sensitivity analyses to

determine the effects of survival of age classes on population growth rates. We gathered data

from the literature on passerines, ducks, geese, gulls, and eagles. These analyses provide a

first approximation of how populations of these types of birds respond to hypothetical changes in

fecundity and survivorship. They can be used to help focus attention on species most likely to be

adversely affected by changes in fecundity and survivorship.

The simplest models assume that the number of animals in a population goes up or down by a

constant ratio, usually designated as lambda, with each unit of time. The annual geometric

growth rate of a population is thus represented by lambda, which is also known as the finite rate

of population increase. The population is increasing if lambda >1, is constant if lambda = 1, and

is decreasing if lambda <1. For example, if lambda = 1.04, then the population was growing at

the rate of 4% during the time period sampled.

With the models in place, each survival rate parameter was allowed to vary from zero to one

while the remaining parameters were held constant. The new value of lambda was calculated at

each new value of the changing parameter. Once these data were obtained for each survival

parameter, the results were plotted on a graph so as to see the different effects each parameter

had on the population growth rate. This can be viewed as a way of expressing the sensitivity of

lambda to the different survival parameters.

The curves for the passerine show that lambda is much more sensitive to changes in the juvenile

survival rate than to changes in the adult survival rate. Also, the juvenile survival rate curve has

a very steep slope as the juvenile survival gets very small. The curves for the duck show that

lambda is roughly equally sensitive to changes in the juvenile and adult survival rates. For

geese, the nonadult age classes survival rates seem to have little impact on the value of lambda.

For the adult age class, lambda is extremely sensitive to changes in the adult survival rate. For

gulls, except for very small survival rates, the changes in the adult age class gives the largest

change in lambda. The other classes all have very similar curves. The situation for the eagle is

very similar to the situation for the gull, but even more extreme: there is great sensitivity of

lambda to changes in adult survival rate.

One of our objectives was to evaluate the use of surrogates, or indices, of survival and

population trends. We found a highly significant negative relationship between adult survival and

annual fecundity. This analysis indicates that fecundity might be a suitable surrogate for survival

in passerines and woodpeckers. This does not imply, however, that fecundity is a suitable

indicator of abundance (i.e., increasing fecundity does not necessarily compensate for lower

survival). Raptors will leave poor habitat (e.g., low food availability), often moving many

kilometers in search of a suitable nesting site. In addition, raptors tend to change territories more

often when nesting is unsuccessful. Thus, as a generality, constancy of territory occupancy

seems to be an indicator of good habitat quality in raptors. The number of nonbreeding, adult

"floaters" in an area is an indicator of the general health of the bird population. This holds if

territory availability is constant or increasing. Additionally, an increase in the age of first

breeding, as well as an increase in adult aggression, are possible indicators of a population at or

above carrying capacity. In long-lived species with delayed age at first breeding, such as in

many raptors and some waterbirds, changes in survival rates have a greater effect on the

population than changes of similar magnitude in reproductive rates. Thus, the use of

reproductive success in long-lived species as a population indicator should likely be

supplemented with other indicators, such as territory occupancy and floater individuals. The use

of surrogates that we recommend here is not designed to determine the cause of population

change. Rather, surrogates are intended to only identify that change has occurred; whether or

not such change is caused by wind development will usually require more rigorous research

(e.g., field work and experimentation). Surrogates serve primarily as a coarse filter to help

narrow the scope of subsequent research.

We also present a series of steps that can be used to develop a strategy for evaluating the

influence of a project on a bird population. Based on our review it seems that the appropriate

hierarchial framework for evaluating population responses to perturbations is: (1) empirical data,

(2) surrogates, and (3) model with available data (Leslie matrices). A large set of empirical data

is, of course, the optimal situation.

vii

Table of Contents

Problem Statement 1

Development of Conceptual Framework 1

Demography 2

Genetics 3

Environmental Stochasticity 3

Life History 3

Ecological Factors 4

Models 4

Introduction 4

Life Tables 5

Simple Lotka Models 5

Leslie Matrix Models 6

Effective Population Size 8

Population Viability Analysis 9

Model Evaluation 11

Objectives 11

Model Description 11

Analysis of Model Reliability 12

Model Structure 12

Parameter Values 12

Primary Prediction of the Model 12

Secondary Predictions of the Model 13

Synthesis 13

Survivorship and Population Projections 13

Model Development: Examples for Wind-Development Applications 17

Results 18

Passerines 18

Ducks 18

Geese 18

Gulls 18

Eagles 18

Surrogates 24

Conclusions: Development of An Analytical Framework 25

Hierarchical Framework 25

Prioritization 26

Acknowledgments 27

Literature Cited 28

Additional Readings 32

viii

List of Figures

Figure 1. Relationship Between Nest Productivity and the Population

Trend of Bald Eagle Populations 16

Figure 2. Sensitivities of Lambda to Different Survival Parameters in

Selected Groups of Birds - Passerines 19

Figure 3. Sensitivities of Lambda to Different Survival Parameters in

Selected Groups of Birds - Ducks 20

Figure 4. Sensitivities of Lambda to Different Survival Parameters in

Selected Groups of Birds - Geese 21

Figure 5. Sensitivities of Lambda to Different Survival Parameters in

Selected Groups of Birds - Gulls 22

Figure 6. Sensitivities of Lambda to Different Survival Parameters in

Selected Groups of Birds - Eagles 23

Problem Statement

Evaluating the possible impacts of existing or proposed wind developments on bird populations is

a continuing challenge. Although methods are available for making empirically-based

estimations of potential impacts, they usually require intensive sampling over a number of years

for each species of concern; the cost of such studies can be prohibitive. Given these difficulties,

it would be desirable to develop a protocol that could provide at least a preliminary indication of

the potential responses of birds and bird populations to wind developments.

One method with a high potential for estimating the impacts of a technology on a bird population

is developing a model of that population over time. Such a model would include parameters that

would represent the impact of that technology on the birds. Then, by changing parameter

values, a first-order (i.e., initial) approximation of the impact of the technology would be obtained.

This approach parallels what is done to gain insight on plagues or medical advances in human

populations.

The goal of this project is to develop a useful, practical modeling framework for evaluating

potential wind-farm impacts that can be generalized to most bird species. We accomplish this by

(1) reviewing the major factors that can influence the persistence of a wild population; (2) briefly

reviewing various models that can aid in estimating population status and trend, including

methods of evaluating model structure and performance; (3) reviewing survivorship and

population projections; and (4) developing a framework for using models to evaluate the potential

impacts of wind development on birds.

Development of Conceptual Framework

The problem confronting the wind industry and society at large regarding potential impacts on

birds falls under the general classification of environmental impact assessment. Determining the

effects of an existing development on birds, or predicting the impacts of a planned development,

is difficult. Thus, the problem confronting the wind industry is similar to that confronting virtually

any planned development (e.g., residential or commercial building, power plants, recreational

development). The goal in such studies is to determine the effects of specific management

actions on individuals and/or populations that occur on a given site.

Managers need a framework for examining the various environmental costs associated with the

different options they have available to them. In most cases, regulations require that an

environmental impact analysis be developed. The federal National Environmental Policy Act of

1969 (NEPA), which includes the environmental impact assessment (EIA)/environmental impact

statement (EIS) process, is designed to review the likely impacts a proposed development will

have on the biotic and abiotic environment in the project area. Unfortunately, no consensus

exists on the type, quality, and amount of data that are necessary to develop a sound

environmental analysis. Further, even if such a consensus existed, it is unlikely that the

biological information necessary to develop a sound analysis would be available in the scientific

literature.

Therefore, managers and researchers alike have been exploring the options available to them to

foster the development of rational decisions. One requirement of any modeling process is clear

elucidation of the information needed to make a decision, and, further, a detailed description of

the strengths and weaknesses of each category of data. This is necessary because certain

categories of data will have a greater influence on the validity of the conclusions than other

categories, and not all categories will have the same quality and amount of data available. Such

a process falls under the general heading of decision analysis, an approach that allows

examination and comparison of expected benefits and costs among various management

options (Lindenmayer et al. 1993).

Making determinations of likely environmental effects is difficult because of the complex

interactions present in any biological system, the difficulties in determining the proper spatial and

temporal scales for analysis, and the inherent lack of data for most parts of a biological system.

Federal management agencies have adopted various strategies to formalize NEPA requirements

and other federal laws (e.g., Endangered Species Act [ESA]). The U.S. Fish and Wildlife

Service, for example, develops recovery plans for threatened and endangered (T&E) species

under ESA. These plans include a detailed review of the available published and unpublished

data to arrive at management plans that are designed to "recover" the species from the T&E list.

The U.S. Forest Service has developed a formal "biological evaluation" procedure wherein

biologists provide written documentation of their judgement about whether or not a proposed

management action will increase the likelihood of the species of concern becoming threatened or

endangered (modeling is not required). These judgements are formalized in a "determination of

effect" document (Ruggiero et al. 1994).

A severe and continuing problem in ecology and management is determination of the relevant

population of study. A typical definition of population illustrates the vagueness of the concept: a

population is a group of organisms of the same species living in a particular space at a particular

time (Krebs 1985). A population is quantified in terms of birth rate, death rate, sex ratio,

emigration-immigration, and age structure. Unfortunately, the absolute values for each of these

parameters is determined in large part by the boundaries the user draws. Recently, much

interest has developed in understanding the role that the spatial structure of a population has on

genetics, and ultimately, survival. A metapopulation is composed of many subpopulations. This

concept relates directly to determinating the impacts on animals. First, even impacts occurring in

a small geographic area can disrupt immigration and emigration between local subpopulations,

and thus the impact can have a much wider effect on the population than immediately evident.

Second, small impacts can have serious consequences to the persistence of small populations if

population size is allowed to drop below a certain threshold level (e.g., the effective population

size).

Below we review the major factors that can influence the persistence of a population. These

factors must be considered when developing a study plan for evaluating the potential impacts of

developments.

Demography

Demography is the study of population statistics, including births, deaths, immigration, and

emigration. Conditions leading to extinction are most likely to occur in small populations because

individuals do not survive for the same length of time, individuals vary in the number of offspring

they bear, individuals often have low birth rates, and so forth. Such effects are sensitive to

population size, and their influences decline as population size increases. Larger total sample

sizes are needed, however, if the population is effectively divided into many local subpopulations

(which is likely in the case of birds with regard to wind development). At low densities, a

threshold or critical population size can exist below which extinction is probable. For example,

limitations to juvenile dispersal can create an extinction threshold in territorial species (Lande

1987).

Partitioning a model into spatial subunits can be difficult to model, although an increasing desire

for understanding metapopulation dynamics is forcing such efforts (e.g., Wootton and Bell 1992).

Spatial heterogeneity and dispersal can stabilize population fluctuations, whether the fluctuations

are caused by natural or human-induced factors. Unfortunately, no general statements can be

made regarding the influence of corridors; minimum distances between habitat patches; or the

ability of dispersers to locate suitable areas; the size and shape of suitable habitat patches; the

ability of animals to travel across, or survive within, marginal habitats. Quantifying emigration

and immigration is an important, but problematic, aspect of research needs.

Adult survivorship is usually very high, especially in long-lived species (such as raptors).

Therefore, estimating adult survivorship tells one a lot about population status (Lande 1988). In

addition, in most monogamous species, it is female survivorship that is most important to

population persistence (e.g., Wootton and Bell 1992). At a minimum, then, quantifying adult

survivorship provides a preliminary, basic indication of the status of the population.

Genetics

Models of genetic variation have a central role in the conservation of populations. This is

because of the interest in how small population size results in more inbreeding among related

individuals and a concomitant reduction in genetic variation, both of which can lead to extinction

(Boyce 1992). However, local subpopulations may contain the genetic diversity that is necessary

to ensure survival of the species if individuals from such subpopulations occasionally meet and

breed with each other.

Boyce (1992) concluded that modeling genetics is not likely to be as important as modeling

demographic and ecological processes in evaluating population persistence. He based this

conclusion, in part, on the lack of our sufficient understanding of genetics to use it as a basis for

management. Thus, practical considerations were the overriding factor in his conclusion. Still,

genetics may be a priority in small, isolated populations.

Environmental Stochasticity

Random environmental events such as catastrophic fires, hurricanes, and disease can have

pronounced effects on small populations. Such factors can also have pronounced effects on

large populations that are spatially divided into subpopulations. Here, factors such as dispersal

will determine the fate of a subpopulation driven to very low numbers, or even to extinction, by a

catastrophic event. It is also important to understand the variance structure of the population;

that is, how does environmental stochasticity affect the individuals differently. A major problem

here, however, is that the difficulties of sampling a variance adequately may overwhelm any

attempts to decompose a variance into individual and environmental components. Thus, the

relative importance of environmental stochasticity must be based on an understanding of the

spatial distribution of the population under study.

Life History

The characteristics that we collectively call life-history parameters of animals include quantifiable

longevity, lifetime reproductive output, the young produced per breeding attempt, the age of

dispersal, survivorship, sex ratio, and the time between breeding attempts. The absolute

expression of each of these characteristics is usually determined by the age of the individual;

thus they change during the lifetime of an animal. For example, young and old individuals tend to

produce fewer viable young than do animals in their prime. In addition, these factors can interact

in various ways that modify the expression of other factors.

Life-history parameters are used in the development of population-projection models. For

example, combining various ranges of parameters can yield substantially different rates of

population change. Such analyses provide guidance on whether the population can be sustained

under varying expressions of life history traits. Once such relationships are understood,

researchers have the opportunity to monitor selected life history traits as part of an assessment

of the status of a population. For example, if previous work shows that the timing of breeding is

correlated with reproductive output, and thus with the population size for the year, monitoring the

time of breeding can provide an early warning of potential population-level problems.

Naturally, this is a simplistic view of the numerous factors that determine population size and

trend. However, it illustrates the utility of developing even rudimentary models as part of the

evaluation of the population status. The data necessary for input into such models are often

available in the literature (this topic is discussed in the Models section below).

Ecological Factors

Temple (1985) found that for birds currently endangered by extinction, 82% are associated with

habitat loss, 44% with excessive take, 35% by introductions of exotics, and 12% by chemical

pollution or the consequences of natural events. It is relatively easier to quantify and model

habitat parameters and their influence on some index of population abundance and life-history

traits, than it is to adequately quantify and model demographic parameters. However, it is

difficult to evaluate the reliability or precision of these indices without some type of calibration

with absolute values.

The weakness in ignoring demographics is, of course, the likelihood that many of the abundance-

habitat correlations will provide little information on why the correlation occurred. This is because

habitat models usually reflect some indirect measure of the population's response to the actual

underlying factor. For example, a negative correlation between animal abundance and canopy

cover could be masking the fact that a predator is at high abundance under high cover.

Resolving the problem of variables interacting and thus confounding results is problematic; this

problem has a long history in modeling of human disease processes. A detailed demographics

study would likely discover that excessive mortality of the prey species is occurring. Competitors,

parasites, and disease also influence populations in substantial ways that are ignored by most

habitat-based analyses.

Models

Introduction

A central part of impact assessment is developing a model that estimates the survival rates

required to maintain a constant population. The strategy is to determine the survival rates

required to sustain the populations that exhibit the various combinations of the other parameters

governing population size. To be useful in a wide range of environmental situations and useable

for people with varying expertise, the model should be based on simple mathematics.

The use of models (of all types) has soared in the past 10 years. In fact, modeling is now a

focus of much interest, research, and management action in wildlife and conservation biology.

But as in all aspects of science, models have certain assumptions and limitations that must be

understood before results of the models can be properly used. Modeling per se is neither "good"

nor "bad"; it is the use of model outputs that determines the value of the modeling approach.

Modeling requires that all terms be defined precisely. The process of thinking rigorously about

the modeling framework is often the most useful product of a modeling exercise.

The use of population models to make management decisions is fairly common. For example,

models play a role in management plans for such threatened and endangered species as the

spotted owl (Strix occidentalis, all subspecies), desert tortoise (Gopherus agassizi), Kirtland's

warbler (Dendroica kirklandii), and various kangaroo rats (Dipodomys spp). Models are valuable

because they analyze the effects of management proposals in ways that are usually not possible

using short-term data or professional opinion.

Two general uses of models can be distinguished: using models to give insight into how an

ecological system behaves versus using models to predict specific variables. Both types of

models help guide decisions when used in combination with other reliable data; and both should

involve testing model assumptions and results in a quantitative manner (i.e., model validation).

Life Tables

Life tables are one of the oldest means of examining mortality in animals; simply, they summarize

survivorship by age classes in a cohort of animals. Grier (1980) and Beuhler et al. (1991) used a

deterministic life-table model to calculate survivorship and population growth in bald eagles

(Haliaeetus leucocephalus).

There are, however, many limitations to using life tables in analyzing wildlife populations. A basic

life table requires only that age, the number of individuals surviving to the beginning of each age

classification, and the number of deaths in each age class be known; mortality and survival rates

can be calculated from these data. There is only one independent column in a life table; all the

others can be calculated from entries in any one column. This dependency requires that great

care be taken in constructing the table, and that large sample sizes be gathered. Other

restrictive assumptions include: (1) Fecundity and mortality are assumed to be unaffected by

population density and to be time invariant (i.e., stationary with regard to time). (2) Reproductive

success of females is assumed to be unaffected by the number of breeding males. (3) All

individuals are assumed to be subject to the same fecundity and mortality schedule. These

assumptions are usually violated, which renders results suspect. It is also difficult to estimate

parameters from life tables (Caughley 1977, Getz and Haight 1989). These parameters are

discussed in the next section.

Simple Lotka Models

The simplest Lotka model assumes that the number of animals in a population goes up or down

by a constant ratio, usually designated as lambda, with each unit of time. The annual geometric

growth rate of a population is thus represented by lambda, which is also known as the finite rate

of population increase. At time t the population size is lambda times population size time t - 1, N

= lambda(N ). The population is increasing if lambda >1, is constant if lambda = 1, and is

t-1

decreasing if lambda <1. For example, if lambda = 1.04, then the population was growing at the

rate of 4% during the time period sampled. For purposes of calculation, this formula is usually

presented as N = N e, where e is the base of natural logarithms, and r is the instantaneous rate

t t-1

of population increase (Johnson 1994).

Lambda should be viewed like any other impact indicator. That is, it is difficult to attribute a

change in lambda to treatment (impact) effect without control areas. Lambda seems most useful

in evaluating the overall status of the population, rather than as a means of isolating the cause of

that status (i.e., increasing, decreasing, or stable). This is because there may be many

environmental factors that have a substantial influence on population status (e.g., weather,

pollution, food availability, disease, predators).

Eberhardt (1990) developed a modeling scheme based on approximations of the Lotka equations

using the grizzly bear (Ursus arctos) as an example. The parameters used to develop the model

included the litter size, proportion of female cubs, breeding interval in years, and reproductive

rate. The utility of this approach was that estimates of these parameters were available in the

literature. Eberhardt used variations among these estimates (e.g., litter size ranged from 1.65 to

2.36) to calculate ranges of female survival rates to provide information about the scope of such

rates needed to sustain populations. Each user of the Eberhardt scheme could select the

particular combination of demographic parameters thought to be most appropriate for a particular

situation. The Eberhardt procedure is, of course, only as good as the empirical estimates used

as model input.

Given any age-specific survival rate, and assuming a stable age distribution, we can estimate the

average productivity that must exist in order that the population remains at a constant size.

Henny et al. (1970), for example, categorized species that had similar breeding strategies and

presented different formulas applicable to each group. They provided the production needed per

breeding-age female for a stable population to be maintained as well as the age ratio that

indicated a stable population. Their "special cases" can be summarized as follows:

1. For an animal that reproduces young at the end of its first year and which has a

constant survival rate after the second year. Most species of birds fall into this category,

including the Passeriformes, many ducks and gamebirds, and shorebirds; included in the

calculations were the American coot (Fulica americana), green-winged teal (Anas crecca),

canvasback (Aythya valisineria), redhead (Aythya americana), ring-necked pheasant (Phasianus

colchicus), and house sparrow (Passer domesticus).

2. For an animal that produces young for the first time at the end of the second year.

Many diving ducks, geese, hawks, owls, and herons were in this category; included in the

calculations were the great-horned owl (Bubo virginianus), herons, and snow goose (Chen

caerulescens).

3. For an animal that produces young for the first time at the end of the third year. Some

raptors and waterbirds fall into this category; included in the calculations were the Caspian tern

(Sterna caspia) and osprey (Pandion haliaetus).

Henny et al. (1970) also provided directions on calculating the "allowable mortality for

maintenance of a stable population level". The difficulty in their approach, of course, is obtaining

precise survival estimates. Henney et al. did state, however, that obtaining a 5-year average or

precise survival estimates periodically (e.g., once every 5 years) should be adequate in most

situations (i.e., non-endangered populations). Additionally, they assume that the age distribution

is stable. Although this assumption likely holds in the short term, changes in the age distribution

is often symptomatic of either an increasing or decreasing population size. Further, they assume

a constant survival rate after breeding age is obtained. Fortunately, this appears to hold for most

animals (although survival decreases as senescence is approached). The Henny et al. (1970)

procedure offers a relatively simple means of at least estimating the status of a population; large

departures from stability would indicate that further study is warranted to verify the situation.

Leslie Matrix Models

Leslie matrix and similar stage-structured models (Caswell 1989) can give great insight into the

processes of population growth. For example, the sensitivity of the population growth rate, r, to

perturbations in vital rates for a Leslie-type model can be solved analytically (although potentially

severe limitations exist, which are discussed below). Understanding how growth rate changes in

response to perturbations at various stages in the life table may help direct management

strategies. For example, adult survival tends to be a parameter to which a model is extremely

sensitive in long-lived species, whereas fecundity can be more important in short-lived species.

Matrix models subsume classical life table analysis as a special case but have capabilities that

go far beyond that analysis. As summarized by McDonald and Caswell (1993), they (1) are not

limited to classifying individuals by age; (2) they lead easily to sensitivity analysis; (3) they can be

constructed using the life cycle graph, an intuitively appealing graphical description of the life

cycle; and (4) they can be extended to include stochastic variation and density-dependent

nonlinearities. Caswell (1989) presents the theory of these models and McDonald and Caswell

(1989) present a detailed description of the formulation and application of matrix models to avian

demographic studies (see also Lebreton and Clobert 1991).

The numbers in the body of the matrix are transition probabilities for survival and progression into

other stages, while the numbers on the top row of the matrix represent stage-specific fecundity

values (see Shenk et al. 1996). A Leslie matrix can be built from estimates of fecundity and

survival probabilities, and population growth may be projected for any number of time periods by

pre-multiplying the age distribution at each time period by the Leslie matrix to get the new age

distribution for the next time period. Thus, we term this matrix the population projection matrix,

or more popularly, the Leslie matrix after its developer (Leslie 1945).

Population projections using Leslie matrices are a useful approach to the analysis of demography

(Jenkins 1988). They provide a numerical tool for determining growth rate and age structure of

populations. They are also useful for illustrating and studying the transient properties of

populations as they converge to the stable state.

Stage-based matrices (e.g., Lefkovitch models), analogous to the age-based Leslie, can be used

to analyze population growth for species in which it is difficult to age individuals, or where it is

more appropriate to classify them into life stages or size classes rather than by age; these

models are generally referred to as Lefkovitch (1965) stage-based models. It is extremely

difficult to determine the specific age of most birds and mammals after they reach adulthood. In

the case of raptors--the focus of concern in many wind developments--young and subadults can

usually be aged up until adulthood (through differences in plumage and soft tissues, and

sometimes eye color). Further, adult raptors can often be placed into categories based on

breeding status.

Lefkovitch models assume a single, well-mixed population with no spatial structure and no

density dependence in the variables. Thus, they assume homogeneous probabilities of

survivorship and breeding success within each stage, independent of other factors such as age.

The models can be modified to incorporate spatial population structure and analyze this structure

in the context of different management options for a population (e.g., see Wootton and Bell 1992

for development and review). Building models that include dispersal, for example, is not

technically difficult; spatial cells can be constructed that have the same status mathematically as

age classes. In addition, these can be developed in both density-independent and density-

dependent cases (Lebreton and Clobert 1991). The problem with building spatially relevant

models, however, is obtaining reliable empirical data on the movements of individuals in a real

population. Although demographic stochasticity may be negligible in large, homogeneous

populations, it could be considerable in smaller, subdivided populations.

An example of a Lefkovitch stage-based model is being developed to assess the impact of the

Altamont Wind Resource Area on the resident golden eagle (Aquila chrysaetos) population (see

Shenk et al. 1996 for modeling details). The model is being developed by a team of scientists

assembled by NREL; field data are being collected by Dr. Grainger Hunt, who is also assisting in

model development. The overall goal of the project is to determine the finite rate of population

growth (lambda) based on birth and death rates for the defined golden eagle population around

the WRA. If the estimate of lambda is >1, then the population will be assumed to be stable or

increasing. If lambda is <1, then no definite conclusion regarding the impact of the WRA on the

population can be made. Because of the lack of a controlled experiment, there could be many

reasons for a declining growth rate; however, it would suggest very detailed field studies should

be done. The model may, however, provide a quantitative understanding of the current status of

the eagle population. Further, parameter estimates of survival and fecundity will assist in

evaluating the status of the population through comparisons with the same information for other

populations. Thus, the approach selected combines a model-based approach with field data

(design-based approach).

The group developing the Altamont population model were faced with time (2-3 years for field

study) and monetary constraints. Because of these constraints, time effects will have to be

ignored (i.e., they cannot determine interyear variability), and sampling must focus only on those

elements essential to model development.

The modeling group has also evaluated the sample sizes necessary to estimate the survival rate

with a minimum precision of 10%. The eagle population can be broken into four general

categories: adult breeders (territory holders >4 years old); adult floaters (non-territory holders >4

years old); subadults (1-4 years old); and juveniles (<1 year old). Eight total categories resulted

when sex was considered. Preliminary analyses indicated that at least 25 individuals would be

necessary for each of the eight classes. Based on time and funding constraints, it was infeasible

to sample that many individuals. Further discussion indicated that all adult floaters and subadults

could be combined and considered "nonterritory holders". Although telemetry is being used to

determine survival, there was not enough time or funding to quantify immigration and/or

emigration. Therefore, immigration and emigration were assumed to have no influence on

lambda. This assumption, if incorrect, could have substantial ramifications on interpretations of

model output. As such, studies of immigration-emigration should be a high priority addition in

follow-up studies. The assumptions and constraints necessary in the Altamont eagle study are,

however, typical of real-world modeling situations, and do not negate the value of a modeling

exercise.

Effective Population Size

As discussed above, small populations are susceptible to extinction because of the random loss

of genetic variation. In the "ideal" theoretical population, the rate of loss of variation is inversely

proportional to the population size. Of course, the reproductive behavior of natural populations is

far from ideal. To link natural and idealized populations, Wright (1931) defined the "effective

population size" (N ) as the size of an ideal population whose genetic composition is influenced

by random processes in the same way as the natural population.

When N is small, the population can rapidly loose genetic variation. However, N has no set

e e

relationship to actual population size, and its precise estimation is complex. Two approaches

have been used to estimate N : genetic and ecological. The genetic methods directly quantify

the effects on genetics of a particular effective population size, whereas the ecological methods

are indirect and depend on the measurement of ecological parameters that are thought to

influence a particular effective population size.

There are several problems associated with determinating the (direct) effective population size

by a genetic method. First, the method requires gathering large amounts of genetic information.

Although new technologies are reducing this problem, it is still beyond the capabilities of most

researchers. Second, the confounding factors of immigration and population subdivision and the

possibility that even relatively low levels of some types of selection have a large influence on the

estimated N .

The indirect ecological methods depend on the theory linking particular ecological parameters,

usually based on demography or behavior, to changes in Ne. Wright (1938) established the

relationship linking the effective population size to the population sex ratio and to the variance in

reproductive success among individuals. For example, a variation in family size inflates the

variance in reproductive success and thus reduces the effective population size.

Various formulas have been developed to estimate the effective population size. Harris and

Allendorf (1989) evaluated several of these methods. Hill's (1972) original equation and its

derivatives were consistently the most accurate. Nunney and Elam (1994) developed a related

approach that required a minimum amount of information while still providing a good estimate of

N . Termed the "minimal" method, it requires the estimation of six parameters: (1)the mean

maturation time to adulthood for both males and females; (2)the mean adult life span for each

sex; (3) the estimation of generation time; estimation of variation in the (4) male and (5) female

reproductive success per breeding season; and (6) an estimation of the adult sex ratio. This

method is designed to provide an estimate of the effective population size in long-lived

populations by using the minimum data possible derived from the literature and short-term study.

The method is most effective if survivorship is age-independent, which is common in many

natural, long-lived populations (not including juveniles).

Nunney and Elam (1994) argued that the minimal method (and ecological methods in general)

provide data that can be used to predict changes in effective population size as the conditions

confronting the population change. Thus, it functions well in monitoring populations over time.

Genetic methods determine what the effective population size has been over the last or several

generations, but they provide no insight into why this has been the prevailing value. Therefore,

they recommend ecological methods when it is practical, so that the effect of different

management options on the effective population size can be estimated. They note, however,

that the demographic information needed to provide a reliable estimate of N can often be difficult

to obtain.

There has been continuing debate over the minimum population size necessary to maintain and

ensure long-term persistence. During the 1980s and into the 1990s, geneticists estimated that

the minimum effective population size was 500 or so breeding individuals. New genetic evidence

suggests, however, that this former estimate is far too low, and could easily range between 1,000

and 10,000 individuals. This new estimate is based on consideration of the effect that mutations

have on the fitness of the organism at low population sizes (Lande 1995, Lynch et al. 1995). It is

difficult to make broad generalizations on the effective population size of organisms. For

example, small (<100 adults) populations have been shown to persist for extended periods of

time because of adaptations to local environmental conditions (e.g., Reed et al. 1986, Grant and

Grant 1992, Nunney 1992).

Population Viability Analysis

A population viability analysis (PVA) is a complex process that considers all factors that affect the

processes of a species' population dynamics (that can lead to extinction). Such factors can

include demographic, genetic, and environmental stochasticity. Life history and habitat-use

parameters, dispersal, competition, and predation may also need to be considered. By formally

trying to understand these processes and how they might influence the species, our general

knowledge of how an environmental change might impact the species is formed. The

determination of biological effect made by the Forest Service (discussed above) in their biological

evaluation procedure is usually based on some type of PVA.

Models are used within the PVA process, ranging from simple verbal to complex mathematical

versions (as reviewed above). A PVA should thus be considered a "process" rather than a

specific model in and of itself. It entails evaluation of available data and models for a population

to anticipate the likelihood of population persistence over some period of time. The "minimum

viable population" (MVP) modeling scheme, in which an estimate of the minimum number that

constitutes a viable population is estimated, is closely related. PVA embraces MVP, but without

seeking to arrive at an absolute population minimum. MVP can be considered in overall

determination of the PVA.

There are few published or peer-reviewed PVAs, and most of those available provide only a

vague outline of model structure or use general "rules of thumb" that are burdened with strong

assumptions. There are no specific guidelines for completing a valid PVA. This is

understandable, however, because each situation is unique because of differing environmental

conditions and differences in the proposed impact(s). The advantage of the PVA is in the fact

that sufficient data to derive reliable estimates for all parameters to develop a MVP is not

practical in most cases (for the reasons developed above).

Hindering the wider use of PVAs in the decision-making process is a general misunderstanding

of their strengths and limitations. Boyce (1992) and Lindenmayer et al. (1993) carefully reviewed

this topic, with Boyce outlining the following strengths of a PVA: (1) it produces an explicit

statement of the ecology of the species and identifies missing data; (2) it synthesizes interacting

factors and identifies trends in population behavior; (3) it identifies processes threatening to the

species; (4) it can be used in defining minimum critical areas and designing reserves; (5) it can

enhance on-ground management and decision making; and (6) it has applications in species

recovery, reintroduction, and captive breeding programs. It is clear that PVAs have applicability

in a wide range of management scenarios, which increases the need for managers from all

disciplines to understand their functioning.

Of particular interest in the application of PVAs to wind development are Lindenmayer et al.

(1993) points 1, 2, 3, and 5. By describing the general ecology of the species (point 1), users

from all educational backgrounds are able to better understand the problems confronting the

scientist in making predictions regarding likely environmental impacts of development. For

example, the needs of a raptor for certain types and sizes of prey when feeding young, or the

influence of a skewed sex ratio on territory occupancy during breeding, are complicated issues

that must be described. These factors might interact (point 2) because only a certain sex is able

to efficiently exploit the prey available in the project area; development might change this prey

availability because of ground disturbance and changes in plant-species composition (point 3).

Development of points 1-3 naturally leads to fulfillment of point 5, namely enhancement of sound

decision-making regarding both permitting of the project, and in the case where permitting is

allowed, modification of the project to avoid potential impacts to the species of concern (in the

above example, avoiding unwanted changes in prey availability through habitat management).

As noted by Lindenmayer et al. (1993), PVAs are only as strong as the data available for use in

their development: "The more data the better for PVAs". And because all models are

simplifications of ecological interactions, PVAs by their sheer complexity tend to inflate errors.

Further, because of substantial differences in life-history parameters among species, no generic

PVA model is available. This greatly complicates the use of a PVA, because one must be

familiar both with the ecology of the species as well as a complex set of mathematical

formulations (Lindenmayer et al. lists many of the models available). As such, most models for

PVA analysis must be modified to meet the particular requirements of a given project.

The use of PVAs is thus hindered not by something inherently wrong with the concept per se, but

rather by the inherent complexity of biological systems. PVAs are simply an attempt to formalize

the complexity of nature. As such, PVAs greatly enhance decision-making by formally identifying

the process under study; thus they provide a list of the data available and the data still needed to

make a rational decision regarding project impacts. Researchers and managers alike are

increasing their use of PVAs (Mace and Lande 1991). The use of PVAs are also limited by the

fact that they are, at best, "what if" analyses. Few PVAs can be tested experimentally, given we

would have to wait and see if, indeed, the population remained viable (i.e., it persists long term).

Boyce (1992) and Lindenmayer et al. (1993) reviewed many of the PVAs that have been

constructed. The summary presented by Lindenmayer et al. (1993:Table 1) is especially useful

in highlighting the fact that each of the PVAs they reviewed identified and estimated the primary

cause of risk to the population. As they noted, habitat loss was the primary risk factor in most of

the situations evaluated. In cases where habitat loss was not the primary factor, a species- and

site-specific factor was identified as the primary concern. For example, hurricanes--which fall

into the general category of environmental stochasticity--were the risk factor for several small,

geographically isolated populations (e.g. Puerto Rican parrot [Amazona vittata] and key deer

[Odocoileus virginianus clavium]). Other site-specific impacts were found to be the primary risk

agent in other isolated populations, including ski resort development, hunting, and logging.

These results from the PVA are mostly intuitive to the biologist: one would expect that a small,

isolated population would be negatively impacted by factors heavily impacting the location where

members of the population remain. The ability of a population to adjust to changes in its habitat

can be predicted through careful study of the behavior of individuals in the population; that is,

through determination of the classification of individuals as either "specialists" or "generalists".

By definition, habitat is a species- or population-specific phenomenon (Morrison et al. 1992). As

such, changes to the habitat must have some impact on individuals in the area, given the habitat

has been properly characterized by the researcher.

The question arises, then, if results of PVAs are biologically intuitive, what value does actual

creation of a PVA offer? The answer is that by constructing a PVA, the researcher is able to

check, in a systematic and analytical fashion, whether her or his intuition was indeed correct.

Perhaps a proper test of a PVA would involve evaluating a hypothesis based on researcher

knowledge and intuition. Further, a PVA allows knowledge to be gained on the interactions of

various life-history parameters and their impact on population numbers.

Model Evaluation

Bart (1995) provided an excellent review of the steps necessary in evaluating the appropriate

uses of a population model: the model objectives, a model description, and an analysis of model

reliability. The latter component is further divided into four important criteria.

Objectives

All studies should list the specific objectives for which model outputs will be used and the

reliability needed for those outputs. Will the output be used only as part of a much larger set of

information, or will management decisions be based on model results? The precision needed

should be specified.

Model Description

The general structure and organization of the model must be detailed. This description should

include details such as the basis for classifying the environment (e.g., vegetation types used for

analysis), the number of sex and age classes, and the behavior of the animals (e.g., breeding

times and dispersal). For example, if sexes or age classes are lumped (because of sample-size

considerations), then the behavior of the sexes and age classes is assumed to be equal.

Likewise, if data on any aspect of the model are lumped across years, then time is held constant

and assumed to have no overriding impact on the model. Careful consideration and justification

of any decisions must be included in model description because they will usually reduce the

complexity and reality of the model.

Analysis of Model Reliability

There are four major types of model reliability to evaluate: structure, parameter values, primary

predictions, and secondary predictions. Each type should receive attention, with emphasis on

the particular type that the management will focus on.

Model Structure

The realism of each assumption should be fully assessed using any information available.

Naturally, the first source of information is the scientific literature about animal behavior, habitat

relationships, population structure, and demographics. If little information is available on the

species of interest, then data on related species should be consulted. The impact that each

assumption has on model results should be clearly discussed. Some assumptions will likely

have minimal impact, while others may have potentially substantial influence on the model. In

some cases, the decision will have to be made that insufficient information is available on this or

closely related species for any meaningful evaluation of the model to be made. In such cases,

the model--if developed--is of the purely descriptive form and should only function in identifying

likely areas upon which field research (to fill the data gaps) should focus. However, usually

enough information is available for at least a preliminary model structure.

Parameter Values

The most reasonable estimate of mean values, variances, and ranges for each parameter should

be developed. The literature should first be consulted, but it is likely that field studies will have to

be conducted to provide reasonable estimates of certain parameter values. Unfortunately, the

wildlife literature provides little in the way of strong data on survivorship of animals, especially

where data on specific sex-age classes are needed. The reality of the situation usually demands

that a short-term (1-3 year) study be initiated to provide the missing data. Because we are

usually interested in either rare species, or isolated populations, we often have to ignore yearly

variations and lump across time to achieve an adequate sample size. The ramifications of this

type of simplification must be carefully evaluated. In most animals, age cannot be readily

determined after adulthood is reached, so it is also almost always the case that certain age

classes (e.g., nonbreeding adults in raptors) will have to be combined.

Primary Predictions of the Model

Primary predictions are the outputs of primary interest used to make management decisions.

Predicted model results should be compared to reality either by comparing them with empirical

data or by running simulations (sensitivity analyses) that can be compared with known (past)

population values. That is, if the model fits past (known) trends, then it is more likely to be

properly forecasting future values. Unfortunately, little data are usually available because few

species have been adequately studied. Most evaluations of models, however, are not truly

independent. This is because available empirical data will likely have been used to initially

develop the model; testing the model predictions with the same data results in a biased

validation. Models should be tested with a new, independent data set collected at a different

time and/or place.

Inferences drawn from the results of sensitivity analysis can be misleading and will not always

indicate the best management actions. These analyses are best regarded as a guide for the

allocation of sampling effort if the aim of measuring demographic parameters is to estimate the

population multiplication rate (lambda)(Green and Hirons [1991]). Sensitivity analysis only

establishes the effect of a fixed absolute change in a parameter value on the population

multiplication rate.

Secondary Predictions of the Model

Secondary predictions are intermediate outputs of the model that can be used to better

understand the population and help evaluate the reliability of the final model. Each of these

outputs is a function of two or more input variables. Comparing them to empirical data, and to

data for similar species, helps identify how reliable the model will be (and where weaknesses

exist). Examples of secondary outputs include data such as the distribution of age classes at

first breeding or territory occupation.

Synthesis

The goal should be to present a realistic and unbiased evaluation of the model. It is preferable to

present both a best and worst case scenario for model outputs, so that the range of values

attainable by the model can be evaluated. For example, with a basic Leslie Matrix Model of

population growth, knowing whether the confidence interval for the predicted (mean) value for

lambda (rate of population growth) includes a negative value provides insight into the reliability of

the predicted direction of population growth.

The process of model development and evaluation may show that the predictions of the model

are sufficiently insensitive (i.e. robust) to the existing uncertainties about the animal's behavior

and demography so that high confidence can be placed in the model's predictions. Even a poor

model does not mean that modeling is inappropriate for the situation under study. Rather, even

a poor model (i.e., a model that does not meet study objectives) can provide insight into how a

population reacts to certain environmental situations. It can thus provide guidelines as to how

empirical data should be collected so that the model can be improved. Modeling is usually an

iterative process.

Survivorship and Population Projections

We reviewed major wildlife and ornithological journals (e.g. Journal of Wildlife Management,

Condor, Auk, and Journal of Raptor Research) published during the past 20 years to determine if

any commonality existed among species with regards to annual survivorship. Most data in the

articles examined were based on either short-term (usually 1-3 years) telemetry studies, or long-

term analyses of band returns. Most of the band return data were obtained from waterfowl

harvested by hunters.

In summary, only very broad generalizations can be drawn regarding "normal" survival rates of

avian populations. Further, yearly variability in survivorship is large even in healthy populations,

which makes short-term (1-2 years) evaluations of a population suspect. Bellrose (1980)

summarized survival rates for waterfowl, concluding that immature ducks show 60-70% first year

mortality, but that subsequent (adult) yearly loss is only 35-40% (or survival of about 60-65%).

More recent studies confirm Bellrose's general values. For example, Smith and Reynolds (1992)

found that survivorship in mallards (Anas platyrhynchos) ranged from about 0.6 to 0.7, and the

population showed no decline in abundance during the study. Unfortunately, most studies that

present survivorship data have no information on population trends or projected population

persistence; most showed survivorship values similar to those summarized by Bellrose (1980;

e.g., see Conroy et al. 1989, Chu et al. 1995, Reynolds et al. 1995). Haramis et al. (1993) and

Hohman and Pritchert (1993) found what they called "high" survivorship rates of over 90% in

canvasbacks. Arnold and Clark (1996) reported survival estimates of female dabbling ducks

based on mark-resighting analyses: mallard (juv. = 0.55, adult = 0.56); galdwall (Anas strepera;

adult = 0.57); northern shoveler (Anas clypeata; adult = 0.51); American wigeon (Anas

americana; adult = 0.64); and blue-winged teal (Anas discors; juv. = 0.29, adult = 0.49). Their

findings were similar to data they summarized from the literature that were based on band

recoveries. Mark-resighting studies require far fewer samples than are needed in banding

studies, and can be used in studies of non-hunted species. Mark-resighting studies do require

intensive field sampling to conduct the re-sighting. However, such re-sighting is under the

control of the researcher, in contrast to band recoveries, over which the researcher has little

control.

In glaucous-winged gulls (Larus glaucescens), Reid (1988) found 85% annual survival in adults,

80% in the second year, and 61% in first-year birds. Using these survival values and other

population parameters to construct a Leslie matrix model, he calculated a lambda of 1.05. This

rate of population growth compared favorably to the observed rate of growth in the field.

Foster et al. (1992) examined the survival of 213 radio-fitted northern spotted owls in four study

areas for 2-4 years. The annual survivorship ranged between 0.67 to 1.00 with most between

0.80 to 0.94; no information on population persistence was provided. In Maryland, Bowman et al.

(1995) provided survivorship data for 1-6-year-old bald eagles (their Table 1), and by using a

deterministic life-table model predicted a finite population growth rate (lambda) of 5.8% per year.

They found, however, that a simulated 12% decrease in minimum adult survival (from 83% to

73%) eliminated population growth. Their review of the literature showed that their estimated

survival rate exceeded those previously published. This study is important because it indicates

that even a relatively minor change in survivorship can have substantial population impacts.

Bowman et al. (1995) also found that survivorship for bald eagles in Alaska was about 0.90 after

the first-year, and survivorship within the first year was 0.71. They used a deterministic life table

to calculate a lambda of 1.02 (2% annual population growth). They showed that their model was

robust to changes in reproductive rates and annual survival rates for first-year eagles, but

sensitive to changes in survival rates for older (after first-year) eagles (their fig. 2). Sensitivity

analysis showed that lambda dropped to <1 when after first-year survival dropped to only 88%;

lambda dropped to about 0.93 when survival was lowered to 82%.

Conway et al. (1995) conducted an experiment to evaluate the effects of removing nestling

prairie falcons (Falco mexicanus) on the breeding population in an attempt to simulate the

impacts of falconry. They removed 138 of 451 nestlings (31% of natality) from 20 territories

during 1982-89, along with a control area. They found no overall difference in nesting success

and productivity between treatments and controls, although treatments were lower than controls

in 2 years of study. Their results suggested that intensive harvest of nestling prairie falcons may

adversely affect some local population parameters, but harvests were sustainable and probably

did not affect local population size. Because only about 0.2% of all prairie falcon natality is

harvested annually in the United States, such a loss has no impact on population numbers or

persistence (relative to the much higher level of harvest they simulated). This study is an

excellent example of experimental evaluation of the impacts of loss of young, and indicates the

resilience of raptor populations to such loss.

In his review of population limitations in birds of prey, Newton (1991) concluded that raptors have

the most stable breeding densities among birds. He stated that breeding densities below those

which would normally be supported by the food supply are held constant by shortages of nest

sites. For raptors limited by available habitat, an increase in numbers could be achieved by

either (1) increasing habitat availability, or (2) increasing the carrying capacity of existing habitat

by increasing food supply and/or nest sites. Green and Hirons (1991) cautioned, however, that

rapid population changes have been recorded in birds from many taxa and a wide range of body

sizes. They noted that even populations of large species, which might be expected to have high

survival rates and low rates of population change, can decline rapidly.

Rowley and Russell (1991) summarized numerous studies of demography in passerines. They

showed that adult survivorship in passerines worldwide was seldom >70%. In North America,

survivorship was usually from 50% to 69%.

Martin (1995) summarized data from the literature on annual adult survival for North American

Passeriformes and Piciformes. For shrub, grassland, and forested areas, survival ranged from

0.435 to 0.668. Excavating cavity nesters had the highest survival (0.668), whereas non-

excavating cavity nesters had the lowest survival regardless of vegetation type (0.435-0.444).

Ground nesters (0.569 in forests, 0.569 in shrub/grasslands), shrub nesters (0.529; 0.532), and

canopy nesters (0.609, forest only) had survival rates that fell between the extremes of

excavators and non-excavators. Martin also presented data on annual fecundity (defined as the

product of numbers of broods and clutch size). Fecundity ranged from a low of 4.99 in

excavators to a high of about 9 in non-excavators. Fecundity in ground, shrub, and canopy

nesters ranged between 5.14 and 7.75. Martin concluded that variation in adult survival and

fecundity was organized by nest site and most closely correlated with nest predation.

Swenson et al. (1986) summarized nest productivity (defined as the number of young fledged per

occupied or active nest) and the reported population trend of 14 study populations of bald eagles.

Study periods ranged from 5-15 years and averaged 8 years. We correlated nest productivity

and the associated population trend reported by Swenson et al. (1986:Table 11) and found a

significant (P = 0.0012), positive correlation (r = 0.596). A graph of our analysis (Fig. 1)

indicates that productivity is a reliable measure of population trend in bald eagles. From Figure 1

it appears that populations have stable or increasing population trends when productivity is

above 0.7 young fledged/nest; at 0.6 young fledged, both a stable and declining trend was

observed. At productivity about 0.8-1.2, populations were either stable or increasing; there was

no clear level of productivity where populations were consistently increasing. Nevertheless, it

appears that productivity >0.8 represents a stable or increasing population abundance.

02426111m

1.40

1.20

1.00

0.80

0.60

0.40

0.20

0.00

Productivity

-1.00

Decreasing

0.00

Stable

1.00

Increasing

Population Trend

Relationship between nest productivity (no. fledged/active nest) and the population trend

(abundance) of bald eagle populations (raw data taken from Swenson et al. 1986:Table 11).

Qualitative descriptions of trends were coded for this analysis (decreasing = -1, stable = 0,

increasing = 1).

Figure 1.

A '

0 0

0 P

n(t)'

t,1

t,2

t,3

Model Development:

Examples for Wind-Development Applications

To aid in providing general guidelines concerning the potential impacts of wind developments on

bird populations, we conducted sensitivity analyses to determine the effects of survival of age

classes on population growth rates. As detailed below, we gathered data from the literature on

passerines, ducks, geese, gulls, and eagles. These analyses provide a first approximation of

how populations of these types of birds respond to hypothetical changes in fecundity and

survivorship. They can be used to help focus attention on species most likely to be adversely

affected by changes in fecundity and survivorship.

These analyses use a Leslie matrix model (Caswell 1989). This model breaks the life cycle into

stage or age classes and places the survival and fecundity rates for each class into a square

projection matrix. This matrix can be multiplied by a vector containing current population sizes

for each class to obtain the population sizes for the next year, i.e.

n(t + 1) = A * n(t),

and thus projects population size into the future. This process may be repeated in order to make

future predictions or to study the growth rate.

For example, a three-age class model would have a projection (or Leslie) matrix

where F = fecundity of the ith age class (typically fledglings have F = 0) and P = the survival

i 1 i

rate for the ith class, which may be multiplied by the vector of population sizes time t,

where n is the size of the ith age class at time t.

t,i

Under certain regularity conditions on the matrix A the population projections converge to a

stable age distribution (Caswell 1989). The largest eigenvalue of the matrix A(eigenvalues solve

an equation of the form |A - 8I| = 0), is the population growth rate, lambda (8). If lambda = 1, the

population size is remaining constant, if lambda >1 then the population is increasing, and lambda

<1 indicates a declining population size.

In the analyses, two age classes were used for passerines and ducks, three age classes for

geese, seven for gulls, and eight for eagles. For passerines, ducks, and geese, models having

`reasonable' survival and fecundity parameters yielding lambda = 1 were constructed. Values for

fecundity and survival for age classes of eagles and gulls were modified from results in Bowman

et al. (1995) and Reid (1988), respectively. These values resulted in lambda slightly larger than

one (1.07 for eagles and 1.046 for gulls).

Passerine - Two Age Class Model 8 = 1

F1 = 0, F2 = 3, P1 = 0.2 and P2 = 0.4.

Duck - Two Age Class Model 8 = 1

F1 = 0, F2 = 1.33, P1 = 0.3 and P2 = 0.6.Goose - Three Age Class Model 8 = 1

F1 = 0, F2 = 0, F3 = 0.476, P1 = 0.3, P2 = 0.7 and P3 = 0.9.

Gull - Seven Age Class Model 8 = 1.046

F1 = 0, F2 = 0, F3 = 0, F4 = 0.0174, F5 = 0.43878, F6 = 0.66172, F7 = 0.71, P1 =

0.607, P2 = 0.8, P3 = 0.853, P4 = 0.853, P5 = 0.853, P6 = 0.853 and P7 = 0.853.

Eagle - Eight Age Class Model 8 = 1.07

F1 = F2 = F3 = F4 = F5 = F6 = F7 = 0, F8 = 0.74, P1 = 0.71, P2 = 0.95, P3 = 0.95, P4

= 0.95, P5 = 0.88, P6 = 0.88, P7 = 0.88 and P8 = 0.88.

With the models in place, each survival rate parameter was allowed to vary from zero to one

while the remaining parameters were held constant. The new value of lambda was calculated at

each new value of the changing parameter. Once these data were obtained for each survival

parameter, the results were plotted on a graph so as to see the different effects each parameter

had on the population growth rate. This can be viewed as a way of expressing the sensitivity of

lambda to the different survival parameters. It is important to distinguish between the true effect

on the population (which is unknown), and the effect on lambda, which assumes the model is the

"correct" model.

Results

Passerines

The curves for the passerine show that lambda is much more sensitive to changes in the juvenile

survival rate than to changes in the adult survival rate (Fig. 2). Also, the juvenile survival rate

curve has a very steep slope as the juvenile survival gets very small.

Ducks

The curves for the duck show that lambda is roughly equally sensitive to changes in the juvenile

and adult survival rates (Fig. 3).

Geese

The nonadult age class survival rates seem to have little impact on the value of lambda (Fig. 4).

For the adult age class, lambda is extremely sensitive to changes in the adult survival rate.

Gulls

Except for very small survival rates, the changes in the adult age class gives the largest change

in lambda (Fig. 5). The other classes all have very similar curves.

Eagles

The situation for the eagle is very similar to the situation for the gull, but even more extreme.

There is great sensitivity of lambda to changes in adult survival rate (Fig. 6).

Figure 2

Sensitivites of Lambda to Different Survival

Parameters in Selected Groups of Birds

Figure 3

Sensitivities of Lambda to Different Survival

Parameters in Selected Groups of Birds

Figure 4

Sensitivities of Lambda to Different Survival

Parameters in Selected Groups of Birds

Figure 5

Figure 6

Sensitivities of Lambda to Different Survival

Parameters in Selected Groups of Birds

Surrogates

One of our objectives was to evaluate the use of surrogates, or indices, of survival and

population trends. Using Martin's data (Martin 1995:Table 4) on adult survival and fecundity in

Passeriformes (reviewed above), we found a highly significant (P < 0.0001) negative relationship

(r = -0.9731, r = 0.9468) between adult survival and annual fecundity. This analysis indicates

that fecundity might be a suitable surrogate for survival in passerines and woodpeckers. This

does not imply, however, that fecundity is a suitable indicator of abundance (i.e., increasing

fecundity does not necessarily compensate for lower survival).

Territory occupancy has also been suggested as an indicator of population health (here, "health"

refers to the stability of population size and adequate reproduction). For example, consistent

interyear occupation of breeding territories (by the same or replacement individuals) has been

accompanied by stable population levels in many species of raptors (e.g., Newton 1979,

Swenson et al. 1986, Ferrer and Donazar 1996).

Newton (1979) noted that adult raptors will leave poor habitat (e.g., low food availability), often

moving many kilometers in search of a suitable nesting site. In addition, raptors tend to change

territories more often when nesting is unsuccessful. Thus, as a generality, constancy of territory

occupancy seems to be an indicator of good habitat quality in raptors.

The number of nonbreeding, adult "floaters" in an area is an indicator of the general health of the

bird population. This holds if territory availability is constant or increasing. Additionally, an

increase in the age of first breeding, as well as an increase in adult aggression, are possible

indicators of a population at or above carrying capacity (Newton 1979, Bowman et al. 1995). In

long-lived species with delayed age at first breeding, such as in many raptors and some

waterbirds, changes in survival rates have a greater effect on the population than changes of

similar magnitude in reproductive rates. Thus, the use of reproductive success in long-lived

species should likely be supplemented with other indicators, such as territory occupancy and

floater individuals.

Conceptually, changes in the age ratios of a population are indicative of the current or likely near-

future trend in population abundance. These ratios can be used in two primary fashions. First,

the current ratio can be compared to literature values to approximate the likely status of the

population under study. Second, temporal changes in ratios track trends in population status.

As the adult:subadult ratio nears 1, for example, we should determine why the population is

becoming adult-dominated (e.g., low young production or survival). Swenson et al. (1986) found

that the subadult:adult ratio was positively related to reproductive success in their study of bald

eagles in the Greater Yellowstone Ecosystem.

The use of surrogates that we recommend here is not designed to determine the cause of

population change. Rather, surrogates are intended to only identify that change has occurred;

whether or not such change is caused by wind development will usually require more rigorous

research (e.g., field work and experimentation). Surrogates serve primarily as a coarse filter to

help narrow the scope of subsequent research (see Temple 1985).

Temple (1985) suggested that the causes of a population decline might be readily identified by

measuring productivity and comparing it with the values expected for the species of concern. If

productivity seemed sufficiently high to balance the expected level of adult mortality, then the

cause of the decline could be identified by elimination, as low adult survival. Survival rates were

considered by Temple to be too difficult to measure directly given the time and money that is

usually available. Green and Hirons (1991) criticized the Temple suggestion, stating that

"expected rates" are highly suspect, both because of poor empirical data for most species, and

because vital rates do not have fixed values. As evidenced by our sensitivity analyses (see

above) and literature review, adult survivorship may be more important in some species.

Our review indicates, however, that productivity may serve as at least a crude indicator of the

trend in population abundance. Recall that our analysis of the Swenson et al. (1986) bald eagle

data indicated a significant (P < 0.001) and moderately strong (r = 0.6) correlation between

productivity and population trend. Further, our analysis of Martin's (1996) data showed a highly

significant (P < 0.0001) and very strong negative (r = -0.97) correlation between adult survival

and annual fecundity. This result indicates that sampling fecundity is a reliable surrogate for

adult survival. Both fecundity and survival can then be compared to published values for the

same or related species as a first approximation of population status. Further, these estimates

can be used in population models.

Beauchamp et al. (1996) documented declines in nest success for many species of prairie-

nesting ducks between the 1930s and 1992. For the gadwall and northern shoveler, however,

there were not concurrent declines in population abundance. The authors speculated that brood

survival and survival throughout the annual cycle were more important indicators of population

status than nesting success. Brood survival is, however, a measure of reproductive success.

For birds in general, brood survival measures the number of young that were able to fledge from

the nest and survive for a specific period of time (usually until independent from the adult).

Thus, it appears that Temple's (1985) suggestion to use productivity as a surrogate of survival

may be appropriate. The points raised by Green and Hirons (1991) are valid, however, and we

reiterate that such surrogates should only be used as first approximations of population status

and as a means of prioritizing species for more intensive study.

Conclusions: Development of An Analytical Framework

From our review of factors known to influence population persistence, the following conclusions

can be drawn: (1) It appears that there is too much uncertainty involving the minimum effective

population size (N ) to use this index in evaluation of population responses to wind

developments. In addition, even the 'minimal method' of calculation requires the estimation of

numerous population parameters. (2) PVA analysis offers a useful framework for advancing

understanding of the processes driving population responses to perturbations. Actual calculation

of a viable population level, however, is far too complicated and subject to far too much variation

to be useful in evaluating the responses of birds to wind developments. (3) The spatial structure

of populations of concern should be considered. However, the practical difficulties of modeling

more than one population, and including consideration of immigration and emigration, will usually

prevent development of spatial-structured models. (4) Carefully evaluate how life-history

parameters could interact to influence population persistence as a first approximation of the

influence of wind developments on bird populations (e.g., sex ratio and productivity); this can be

accomplished through the use of the Leslie matrix and related models. It also appears that

surrogates of many of these factors can be developed as an additional first approximation of the

status of a population. We present below a series of steps that can be used to develop a

strategy for evaluating the influence of a project on a bird population.

Hierarchical Framework

Based on our review it seems that the appropriate hierarchial framework for evaluating

population responses to perturbations is:

1. empirical data

2. surrogates

3. model with available data (Leslie matrices).

A large set of empirical data is, of course, the optimal situation. Unfortunately, the time and

funding necessary to obtain an adequate set of field data are seldom available for all of the

factors of interest. Data that are available, however, feed into items #2 and #3, above.

Perturbation analysis of a model that captures the major dynamics of a population should be an

integral part of a modeling exercise. Also, remember that the three items listed above are not

mutually exclusive: as the quantity and quality of empirical data grows, our ability to develop

good models and good surrogates increases.

As a subset of 3, above, Lebreton and Clobert (1991) summarized a hierarchical framework for

modeling the dynamics of bird populations for conservation and management:

1. demographic models with constant parameters, with emphasis on sensitivity

analysis

2. environmental variability and its effects on population growth rate

3. density-dependent models

4. models incorporating density dependence, spatial aspects, and various kinds of

stochasticity.

From our review of the literature it is evident that we can seldom venture beyond Lebreton and

Colbert's stage 3. Regardless, any modeling that is based on a data set with known variability

will provide a useful evaluation of the potential response of a population to perturbations.

Prioritization

Most project locations are inhabited by many species of birds. As such, it will be necessary to

select a subset of species for intensive study and modeling. To assist with the selection of

species for study, we present the following simple criteria, in order of priority, for study:

1. Locally rare population of an overall rare species

2. Locally common population of an overall rare species

3. Locally rare population of an overall common species

4. Locally common population of an overall common species.

This hierarchy can be viewed in the following table, where “yes” means high priority or concern

and “no” means lower priority or concern.

Species

Local Population Rare Common

Rare Yes (1) No (3)

Common Yes (2) No (4)

We may worry about #1 and #2 the most, although locally there is reason to worry about #3.

However, when priorities must be made, we suggest that this scenario is a reasonable starting

point to determine allocation of resources. Alternatively, from a public and regulatory

perspective, a rare species that is locally common (#2) may be legally protected. From a

biological perspective, this population may be integral to the survival of the species. Common

species that are rare locally (#3) may suggest that the area is on the periphery of the range or

that some other parameter (e.g., habitat) is creating the problem, making the study of windpower

impacts relatively unimportant.

Because of limited time and budgets, the results of these evaluations should be considered

approximations of population status and longer-term persistence. It is imperative to continue to

study the situation, should the wind development be approved, in order to improve the initial

approximation. This would also improve the evaluation process in future impact assessments by

providing additional empirical data.

Acknowledgments

We thank Holly Davis, Robert Thresher, and Karin Sinclair, National Renewable Energy

Laboratory, for initiating and funding this study; William Kendall, Lawrence Mayer, Dale

Strickland, and Kenneth Wilson for reviewing the text; and Ann Oberg for assisting with analyses.

Literature Cited

Arnold, T.W., and R.G. Clark. 1996. Survival and philopatry of female dabbling ducks in

southcentral Saskatchewan. Journal of Wildlife Management 60:560-568.

Bart, J. 1995. Acceptance criteria for using individual-based models to make management

decisions. Ecological Applications 5:411-420.

Beauchamp, W.D., R.R. Koford, T.D. Nudds, R.G. Clark, and D.H. Johnson. 1996. Long-term

declines in nest success of prairie ducks. Journal of Wildlife Management 60:247-257.

Bellrose, F.C. 1980. Ducks, geese, and swans of North America. Stackpole Books, Harrisburg,

Penn.

Bowman, T.D., P.F. Schempf, and J.A. Bernatowicz. 1995. Bald eagle survival and population

dynamics in Alaska after the Exxon Valdez oil spill. Journal of Wildlife Management 59:317-324.

Boyce, M.S. 1992. Population viability analysis. Annual Review of Ecology and Systematics

23:481-506.

Buehler, D.A., J.D. Fraser, J.K.D. Seegar, G.D. Therres, and M.A. Byrd. 1991. Survival rates

and population dynamics of bald eagles on Chesapeake Bay. Journal of Wildlife Management

55:608-613.

Caswell, H. 1989. Matrix population models: construction, analysis, and interpretation. Sinauer

Associates, Sunderland, Mass.

Caughley, G. 1977. Analysis of vertebrate populations. John Wiley and Sons, New York, N.Y.

Chu, D.S., J.D. Nichols, J.B. Hestbeck, and J.E. Hines. 1995. Banding reference areas and

survival rates of green-winged teal, 1950-89. Journal of Wildlife Management 59:487-498.

Conroy, M.J., G.R. Costanzo, and D.B. Stotts. 1989. Winter survival of female American black

ducks on the Atlantic coast. Journal of Wildlife Management 53:99-109.

Conway, C.J., S.H. Anderson, D.E Runde, and D. Abbate. 1995. Effects of experimental

nestling harvest on prairie falcons. Journal of Wildlife Management 59:311-316.

Eberhardt, L.L. 1990. Survival rates required to sustain bear populations. Journal of Wildlife

Management 54:587-590.

Foster, C.C. et al. (7 authors). 1992. Survival and reproduction of radio-marked adult spotted

owls. Journal of Wildlife Management 56:91-95.

Ferrer, M., and J.A. Donazar. 1996. Density-dependent fecundity by habitat heterogeneity in an

increasing population of Spanish imperial eagles. Ecology 77:69-74.

Getz, W.M., and R.G. Haight. 1989. Population harvesting: demographic models of fish, forest,

and animal resources. Princeton University Press, Princeton, N.J.

Grant, P.R., and B.R. Grant. 1992. Demography and the genetically effective sizes of two

populations of Darwin's finches. Ecology 73:766-784.

Green, R.E., and G.J.M. Hirons. 1991. The relevance of population studies to the conservation

of threatened species. Pages 594-633 in Perrins, C.M., J.-D. Lebreton, and G.J.M. Hirons, eds.

Bird population studies. Oxford University Press, New York, N.Y.

Grier, J.W. 1980. Modeling approaches to bald eagle population dynamics. Wildlife Society

Bulletin 8:316-322.

Haramis, G.M., D.G. Jorde, and C.M. Bunck. 1993. Survival of hatching-year female

canvasbacks wintering on Chesapeake Bay. Journal of Wildlife Management 57:763-771.

Harris, R.B., and F.W. Allendorf. 1989. Genetically effective population size of large mammals:

an assessment of estimators. Conservation Biology 3:181-191.

Henny, C.J., W.S. Overton, and H.M. Wright. 1970. Determining parameters for populations by

using structural models. Journal of Wildlife Management 34:690-703.

Hill, W.G. 1972. Effective size of populations with overlapping generations. Theoretical

Population Biology 3:278-289.

Hohman, W.L., R.D. Pitchert, J.L. Moore, and D.O. Schaeffer. 1993. Survival of female

canvasbacks wintering in coastal Louisiana. Journal of Wildlife Management 57:758-762.

Jenkins, S.H. 1988. Use and abuse of demographic models of population growth. Bulletin of

the Ecological Society of America 69:201-207.

Johnson, D.H. 1994. Population analysis. Pages 419-444 in T. A. Bookhout, ed., Research and

management techniques for wildlife and habitats. The Wildlife Society, Bethesda, Maryland.

Krebs, C.J. 1985. Ecology: the experimental analysis of distribution and abundance. Third ed.

Harper and Row, New York, N.Y.

Lande, R. 1987. Extinction thresholds in demographic models of territorial populations.

American Naturalist 130:624-635.

Lande, R. 1988. Demographic models of the northern spotted owl (Strix occidentalis caurina).

Oecologia 75:601-607.

Lande, R. 1995. Mutations and conservation. Conservation Biology 9:782-791.

Lebreton, J.D., and J. Clobert. 1991. Bird population dynamics, management, and conservation:

the role of mathematical modelling. Pages 105-125 in Perrins, C.M., J.-D. Lebreton, and G.J.M.

Hirons, eds. Bird population studies. Oxford University Press, New York, N.Y.

Lefkovitch, L.P. 1965. The study of population growth in organisms grouped by stages.

Biometrics 21:1-18.

Leslie, P.H. 1945. On the use of matrices in certain population mathematics. Biometrika

33:183-212.

Lindenmayer, D.B., T.W. Clark, R.C. Lacy, and V.C. Thomas. 1993. Population viability analysis

as a tool in wildlife conservation policy: with reference to Australia. Environmental Management

17:745-758.

Lynch, M., J. Conery, and R. Burger. 1995. Mutation accumulation and the extinction of small

populations. American Naturalist 146:489-518.

Mace, G.M., and R. Lande. 1991. Assessing extinction threats: toward a reevaluation of IUCN

threatened species categories. Conservation Biology 5:148-157.

Martin, T.E. 1995. Avian life history evolution in relation to nest sites, nest predation, and food.

Ecological Monographs 65:101-127.

McDonald, D.B., and H. Caswell. 1993. Matrix methods for avian demography. Current

Ornithology 10:139-185.

Morrison, M.L., B.G. Marcot, and R.W. Mannan. 1992. Wildlife-habitat relationships: concepts

and applications. University of Wisconsin Press, Madison.

Newton, I. 1979. Population ecology of raptors. Buteo Books, Vermillion, South Dakota.

Newton, I. 1991. Population limitations in birds of prey: a comparative approach. Pages 3-21 in

Perrins, C.M., J.-D. Lebreton, and G.J.M. Hirons, eds. Bird population studies. Oxford University

Press, New York, N.Y.

Nunney, L. 1992. Estimating the effective population size and its importance in conservation

strategies. Transactions of the Western Section of The Wildlife Society 28:67-72.

Nunney, L., and D.R. Elam. 1994. Estimating the effective population size of conserved

populations. Conservation Biology 8:175-184.

Reed, J.M., P.D. Doerr, and J.R. Walters. 1986. Determining minimum population sizes for

birds and mammals. Wildlife Society Bulletin 14:255-261.

Reid, W.V. 1988. Population dynamics of the glaucous-winged gull. Journal of Wildlife

Management 52:763-770.

Reynolds, R.E., R.J. Blohm, J.D. Nichols, and J.E. Hines. 1995. Spring-summer survival rates

of yearling versus adult mallard females. Journal of Wildlife Management 59:691-696.

Rowley, I., and E. Russell. 1991. Demography of passerines in the temperate southern

hemisphere. Pages 22-44 in Perrins, C.M., J.-D. Lebreton, and G.J.M. Hirons, eds. Bird

population studies. Oxford University Press, New York, N.Y.

Ruggiero, L.F., G.D. Hayward, and J.R. Squires. 1994. Viability analysis in biological

evaluations: concepts of population viability analysis, biological population, and ecological scale.

Conservation Biology 8:364-372.

Shenk, T.M., A.B. Franklin, and K.R. Wilson. 1996. A model to estimate the annual rate of

golden eagle population change at the Altamont Pass Wind Resource Area. Pages 47-54 In

Proceedings of National Avian-Wind Power Planning Meeting II. National Wind Coordinating

Committee, Washington, DC.

Smith, G.W., and R.E. Reynolds. 1992. Hunting and mallard survival, 1979-88. Journal of

Wildlife Management 56:306-316.

Swenson, J.E., K.L. Alt, and R.L. Eng. 1986. Ecology of bald eagles in the Greater Yellowstone

Ecosystem. Wildlife Monographs no. 95.

Temple, S.A. 1985. The problem of avian extinctions. Current Ornithology 3:453-485.

Wootton, J.T., and D.A. Bell. 1992. A metapopulation model of the peregrine falcon in

California: viability and management strategies. Ecological Applications 2:307-321.

Wright, S. 1931. Evolution in Mendelian populations. Genetics 16:97-159.

Wright, S. 1938. Size of population and breeding structure in relation to evolution. Science

87:430-431.

Additional Readings

Caswell, H., and M.C. Trevisan. 1994. Sensitivity analysis of periodic matrix models.

Ecology 75:1299-1303.

Cochran, M.E., and S. Ellner. 1992. Simple methods for calculating age-based life

history parameters for stage-structured populations. Ecological Monograph 62:345-

364.

DeAngelis, D.L., and L.J. Gross, eds. 1992. Individual-based models and approaches

in ecology. Chapman & Hall, New York, N.Y.

Goodman, L.A. 1968. An elementary approach to the population projection matrix, to

the population reproductive value and to related topics in the mathematical theory of

population growth. Demography 5:382-409.

Manly, B.F.J. 1990. Stage-structured populations: sampling, analysis and simulation.

Chapman and Hall, London, England.

Nichols, J.D., J.R. Sauer, K.H. Pollock, and J.B. Hestbeck. 1992. Estimating transition

probabilities for stage-based population projection matrices using capture-recapture

data. Ecology 73:306-312.

Plant, R.E. 1986. A method for computing the elements of the Leslie matrix.

Biometrics 42:933-939.

Vandermeer, J.H. 1975. On the construction of the population projection matrix for a

population grouped in unequal stages. Biometrics 31:239-242.

REPORT DOCUMENTATION PAGE

Form Approved

OMB NO. 0704-0188

Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources,

gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this

collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson

Davis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704-0188), Washington, DC 20503.

1. 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED

November 1997 Subcontractor Report

4. TITLE AND SUBTITLE 5. FUNDING NUMBERS

Development of a Practical Modeling Framework for Estimating the Impact of Wind Technology on C: CCD-5-15367-01

Bird Populations

TA: WE801410

6. AUTHOR(S)

Michael L. Morrison, Kenneth H. Pollock

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION

Michael L. Morrison

California State University

P.O. Box 310

Fiddletown, CA 95629

REPORT NUMBER

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORING/MONITORING

National Renewable Energy Laboratory

1617 Cole Blvd.

Golden, CO 80401-3393

AGENCY REPORT NUMBER

SR-440-23088

11. SUPPLEMENTARY NOTES

NREL Technical Monitor: Karin Sinclair

12a. DISTRIBUTION/AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE

National Technical Information Service UC-1210

U.S. Department of Commerce

5285 Port Royal Road

Springfield, VA 22161

13. ABSTRACT (Maximum 200 words)

One of the most pressing environmental concerns related to wind project development is the potential for avian fatalities caused by the

turbines. The goal of this project is to develop a useful, practical modeling framework for evaluating potential wind power plant impacts that

can be generalized to most bird species. This modeling framework could be used to get a preliminary understanding of the likelihood of

significant impacts to birds, in a cost-effective way. We accomplish this by (1) reviewing the major factors that can influence the persistence

of a wild population; (2) briefly reviewing various models that can aid in estimating population status and trend, including methods of

evaluating model structure and performance; (3) reviewing survivorship and population projections; and (4) developing a framework for

using models to evaluate the potential impacts of wind development on birds.

14. SUBJECT TERMS 15. NUMBER OF PAGES

Key words: wind energy—environmental impacts, wind turbine, avian, birds, modeling, population impact

16. PRICE CODE

17. SECURITY CLASSIFICATION 18. SECURITY CLASSIFICATION 19. SECURITY CLASSIFICATION 20. LIMITATION OF ABSTRACT

OF REPORT OF THIS PAGE OF ABSTRACT

Unclassified Unclassified Unclassified

NSN 7540-01-280-5500 Standard Form 298 (Rev. 2-89)

Prescribed by ANSI Std. Z39-18

298-102

The biology and demographic parameters of Aedes albopictus in northern peninsular Malaysia

Article

Full-text available

Dec 2011

To generate life table characteristics for the dengue vector Aedes albopictus (A. albopictus) under uncontrolled conditions, incorporating both the aquatic and the adult stages. Ten females derived from wild pupae were allowed to fully blood-feed on restrained mice. 774 eggs were hatched in seasoned water. F1 larvae were followed for development until their F2 counterparts emerged as adults. Some population parameters were monitored (F1) or estimated (F2). A. albopictus exhibited increased fecundity and egg hatch success. Immature development was quick. Immature survival was high, with lowest rate in the pupal stage. Adult emergence was about 81% and sex ratio was close to 1:1. Generational mortality (K) was about 28%. A high proportion of females completed a reproductive cycle and the obtained parity rate was predicted to lead to higher fecundity in the next generation. It can be concluded that natural A. albopictus populations in Penang seem largely determined by quick development in combination with low immature loss and increased oviposition.

Assessing population-level consequences of anthropogenic stressors for terrestrial wildlife

Article

Full-text available

Mar 2020

Human activity influences wildlife. However, the ecological and conservation significances of these influences are difficult to predict and depend on their population-level consequences. This difficulty arises partly because of information gaps, and partly because the data on stressors are usually collected in a count-based manner (e.g., number of dead animals) that is difficult to translate into rate-based estimates important to infer population level consequences (e.g., changes in mortality or population growth rates). However, ongoing methodological developments can provide information to make this transition. Here, we synthesize tools from multiple fields of study to propose an overarching, spatially explicit framework to assess population-level consequences of anthro-pogenic stressors on terrestrial wildlife. A key component of this process is using ecological information from affected animals to upscale from count-based field data on individuals to rate-based demographic inference. The five steps to this framework are (1) framing the problem to identify species, populations, and assessment parameters ; (2) field-based measurement of the effect of the stressor on individuals; (3) characterizing the location and size of the populations of interest; (4) demographic modeling for those populations; and (5) assessing the significance of stressor-induced changes in demographic rates. The tools required for each of these steps are well developed, and some have been used in conjunction with each other, but the entire group has not previously been unified together as we do in this framework. We detail these steps and then illustrate their application for two species affected by different anthropogenic stressors. In our examples, we use stable hydrogen isotope data to infer a catchment area describing the geographic origins of affected individuals, as the basis to estimate population size for that area. These examples reveal unexpectedly greater potential risks from stressors for the more common and widely distributed species. This work illustrates key strengths of the framework but also important areas for subsequent theoretical and technical development to make it still more broadly applicable.

A Methodology to Assess the National and Regional Impacts of U.S. Wind Energy Development on Birds and Bats. U.S. Geological Survey Scientific Investigations Report 2018–5157, 45 p., https://doi.org/10.3133/sir20185157. [Supersedes USGS Scientific Investigations Report 2015–5066.

Technical Report

Full-text available

Jul 2019

A Review on the Environmental Impacts of Wind Energy

Article

May 2018

Human is well aware to habituate the wind energy since thousands of years, both for producing the power and soiling the boats at land. From all renewable strength resources, wind strength is best in term of industrial evolution due to its renewability and accessibility, this electricity source is appealing. Potency for evolution and improvement is significant, and the sector’s capacity is some distance bigger than the sector’s overall strength expenditure. With yearly production of about 100 TWh, the total capacity of the world is about 60,000 MW has been located. For further development the major challenges are directly related to land usage, economy, net capacity and environment. Wind power plant requires time less than a year for their installation and construction as compare to nuclear power plant which requires 10 years for their construction and installation. Due to reduce installation cost and no fuel cost wind power is most economic. On environment the effects of wind power are considered to be positive due to potential displacement of mining activities, greenhouse effect and air pollution which are related to non-renewable energy sources. In the result the positive and negative impacts of renewable or non-renewable energy sources totally depends on the whole understanding of their environmental effects and economic effects.

Impacts on biodiversity of exploitation of renewable energy sources: the example of birds and bats

Article

Full-text available

Jan 2006

Wind turbines and birds. A background review for environmental assessment. Document prepared by Bird Studies Canada, for Environment Canada / Canadian Wildlife Service

Article

Jan 2005

This document will be reviewed and updated as required. To ensure that you have the most up-todate version, please consult our web site at

Predicting the response of bird populations to wind energy-related deaths

Conference Paper

Jan 1998

A Roadmap for PIER Research on Avian Collisions with Wind Turbines in California

Article

Overview of the U.S. Department of Energy/National Renewable Energy Laboratory avian research program

Article

Jan 1997

As wind energy use continues to expand, concern over the possible impacts of wind farms on birds continues to be an issue. The concern includes two primary areas: the effect of avian mortality on bird populations, and possible litigation over the killing of even one bird if it is protected by the Migratory Bird Treaty Act or the Endangered Species Act or both. In order to address these concerns, the US Department of Energy (DOE) and the National Renewable Energy Laboratory (NREL), working collaboratively with all stakeholders including utilities, environmental groups, consumer advocates, utility regulators, government officials, and the wind industry, has an active avian-wind power research program. DOE/NREL is conducting and sponsoring research with the expectation of developing solutions to educe or avoid avian mortality due to wind energy development throughout the US. This paper outlines the DOE/NREL approach and summarizes completed, current, and planned projects.

Wind Energy Development and Wildlife Conservation: Challenges and Opportunities

Article

Nov 2007
J WILDLIFE MANAGE

Wind energy development represents significant challenges and opportunities in contemporary wildlife management. Such challenges include the large size and extensive placement of turbines that may represent potential hazards to birds and bats. However, the associated infrastructure required to support an array of turbines—such as roads and transmission lines—represents an even larger potential threat to wildlife than the turbines themselves because such infrastructure can result in extensive habitat fragmentation and can provide avenues for invasion by exotic species. There are numerous conceptual research opportunities that pertain to issues such as identifying the best and worst placement of sites for turbines that will minimize impacts on birds and bats. Unfortunately, to date very little research of this type has appeared in the peer-reviewed scientific literature; much of it exists in the form of unpublished reports and other forms of gray literature. In this paper, we summarize what is known about the potential impacts of wind farms on wildlife and identify a 3-part hierarchical approach to use the scientific method to assess these impacts. The Lower Gulf Coast (LGC) of Texas, USA, is a region currently identified as having a potentially negative impact on migratory birds and bats, with respect to wind farm development. This area is also a region of vast importance to wildlife from the standpoint of native diversity, nature tourism, and opportunities for recreational hunting. We thus use some of the emergent issues related to wind farm development in the LGC—such as siting turbines on cropland sites as opposed to on native rangelands—to illustrate the kinds of challenges and opportunities that wildlife managers must face as we balance our demand for sustainable energy with the need to conserve and sustain bird migration routes and corridors, native vertebrates, and the habitats that support them. (JOURNAL OF WILDLIFE MANAGEMENT 71(8):2487-2498; 2007)

Determining minimum population size for birds and mammals

Article

Full-text available

Jan 1986

Presents a simple model for calculating effective population sizes for species with overlapping generations. This equation can be used in management programmes for determining viable population levels of vertebrate species. Minimum effective population sizes for short-term and long-term survival are discussed.-from Authors

Population limitation in birds of prey: a comparative approach

Article

May 1991

Ian Newton

Bird of prey populations are normally regulated, rather than fluctuating at random. Regulation in many species comes through competition for breeding space, and is helped by the presence of surplus adults, which breed only when an existing nesting territory becomes vacant. In habitat where nest sites are freely available, breeding density is limited by food supply. This may be inferred from: 1) species that exploit fairly stable (often varied) food sources show fairly stable densities, which differ between regions where food abundance differs; and 2) species that exploit annually fluctuating (often restricted) food supplies show fluctuating densities. In other areas, however, breeding density may be restricted by shortage of nest sites to a lower level than would normally occur with the available food-supply. This may be inferred from: 1) the absence of breeding pairs in areas that lack nest sites but are otherwise apparently suitable; and 2) the provision of artificial nest sites is sometimes followed by a big increase in breeding density. Hence, in the habitat available in any one region, breeding density is naturally limited by food supply or nest sites, whichever is most restricted. Many modern populations are below the level that would be permitted by the habitat because of pesticide use or other human action. -from Author

Use and abuse of demographic models of population growth

Article

Jan 1988
Bull Ecol Soc Am

S.H. Jenkins

Bird Population Studies for Suburban Students

Article

Sep 1975

J.H. Coggins

I recognize that there are some who feel winter feeding of birds is unwise since it may result in larger populations than would exist naturally. There are also critics who view studies of populations at a feeder as worthless since the birds are not in a completely “natural” setting. I contend that winter feeding is harmful only if it is not continuous. As to whether a bird feeder study is unworthy of scientific examination, I would point out that with humans spreading into more and more wildlife habitats, there cannot be enough knowledge gained about those situations where man and other animals interact. For the students, these population studies have provided more than experience in statistical manipulation and some wildlife handling techniques (fig. 2 and 3). They seem to be genuinely excited by the discovery of research opportunities in their own communities and to gain a heightened awareness and appreciation of their environment. For the instructor that is ample reward, and ample justification for continuing the studies. © 1994, The Regents of the University of California. All rights reserved.

Demography of passerines in the temperate Southern Hemisphere

Article

May 1991

Southern temperate species for which demographic data are available are mainly Australian. Some 75% of Australian passerines are of endemic origin and evolved in response to past and present environmental conditions. In temperate non-arid regions, seasonal climatic extremes have never been severe. Breeding seasons of 4-5 months are not shortened by the need to migrate. Although the main food resources of insects and nectar fluctuate, with a spring peak and decline in late summer/winter, food is always available and resident species are common. Clutch size (mean 2.7) of the old endemic Australian families is lower than that of passerines from equivalent latitudes in North Africa (4.3). Many Australian passerines are regularly multibrooded and variation in reproductive effort occurs through the number of breeding attempts rather than in clutch size; even so, annual productivity is low, generally <4 fledglings per female per year. Adult survival is higher than in most northern hemisphere species; only two of 22 species had annual survival <60% and eight of these species had survival of ≥80%. For many species, a significant proportion of adults survive for a long time becoming the experienced, productive breeders; 14 of 22 species showed maximum longevity >10 yr. The longer life, lower reproductive rate, and delayed sexual maturity of tropical passerines appear to be also characteristic of temperate southern hemisphere passerines. Management strategies need to bear in mind the inability of such populations to recover rapidly following a disaster. -from Authors

Size of population and breeding structure in relation to evolution

Article

Jan 1938

S. Wright

The relevance of population studies to the conservation of threatened birds

Article

May 1991

Reviews current knowledge of demographic parameters of threatened birds and examines the contribution that studies of population processes could make to their conservation, focussing on assessment of the risk of extinction as a basis for assigning priorities for scarce conservation resources, and the identification of the causes of changes in demographic parameters and the consequences of such changes for population size and trend. -from Authors

The Problem of Avian Extinctions

Article

Jan 1986

Stanley A. Temple

Studying endangered birds and developing programs to prevent their extinction have become principal endeavors of bird conservationists worldwide. More than perhaps any other group of vertebrate biologists, ornithologists have responded to the contemporary threat of accelerated extinction rates with intensive research and management efforts. This responsiveness may be owing to the fact that many threatened or endangered birds are relatively well known so that their extinctions would represent the loss of organisms whose importance has already been demonstrated in terms of scientific advancement, contributions to ecological systems, commercial and recreational activities, or esthetic considerations. Furthermore, the generally advanced state of our knowledge of birds has facilitated the rapid development of management techniques. Ornithologists are often able to quickly identify the threats to a species and propose a diverse arsenal of proven management approaches that can prevent the species’ extinction. As a result, management of endangered birds has set standards to which conservationists concerned with other taxonomie groups refer.

Matrix Methods for Avian Demography

Chapter

Jan 1993

Demography is a tool for understanding population-level dynamics in terms of events (birth, death, maturation, etc.) at the level of the individual. Demographic models are a critical component of theory in population genetics, life history evolution, mating systems, and population biology. Demography is of fundamental concern to conservation biology; the demographic rather than genetic consequences of rarity may be the imminent threat to species facing rapid habitat destruction in many parts of the world (Lande, 1988b).

Sensitivity analysis of periodic matrix model

Article

Jul 1994

Periodic matrix models are used to describe the effects of cyclic environmental variation, both seasonal and interannual, on population dynamics. If the environmental cycle is of length m, with matrices B-(1), B-(2),..., B-(m) describing population growth during the m phases of the cycle, then population growth over the whole cycle is given by the product matrix A = B-(m)B-(m-1)...B-(1). The sensitivity analysis of such models is complicated because the entries in A are complicated combinations of the entries in the matrices B-(i), and thus do not correspond to easily interpreted life history parameters. In this paper we show how to calculate the sensitivity and elasticity of population growth rate to changes in the entries in the individual matrices B-(i) making up a periodic matrix product. These calculations reveal seasonal patterns in sensitivity that are impossible to detect with sensitivity analysis based on the matrix A. We also show that the vital rates interact in important ways: the sensitivity to changes in a rate at one point in the cycle may depend strongly on changes in other rates at other points in the cycle.

Development of a practical modeling framework for estimating the impact of wind technology on bird populations

Abstract and Figures

Recommended publications

Reproductive success and survivorship of the Galapagos Hawk Buteo galapagoensis: Potential costs and...

Parental Care among the Amphibia

The epidemiology of failure in total knee arthroplasty: avoiding your next revision

Greater impacts of wind farms on bird populations during construction than subsequent operation: Res...