Vortex Manual as a PDF

Version 9.21
VORTEX
A Stochastic Simulation
of the Extinction Process
User’s Manual
Manual Written by:
Philip S. Miller, Conservation Breeding Specialist Group (SSC/IUCN)
Robert C. Lacy, Chicago Zoological Society
Software Written by:
Robert C. Lacy, Max Borbat, and JP Pollak
A contribution of the IUCN/SSC Conservation Breeding Specialist Group in collaboration with the Chicago
Zoological Society.
VORTEX is provided at no cost, in order to further conservation and science. It is distributed without warranty of its
suitability for any particular use, and neither the program or this manual is guaranteed to be free of errors, bugs, or
potentially misleading information. It is the responsibility of the user to ensure that the software is appropriate for
the uses to which it is put.
VORTEX is owned and copyrighted by the Chicago Zoological Society. The software is not copy-protected. In
addition to making back-up copies, individuals, not-for-profit organizations, and governmental agencies are hereby
given licenses for making unlimited copies of VORTEX for the purpose of furthering conservation, teaching, and
research.
Distribution of VORTEX is restricted to:
• distribution by the Chicago Zoological Society;
• distribution by the IUCN/SSC Conservation Breeding Specialist Group;
• downloading of the program from the Internet (http://www.vortex9.org/vortex.html) by individuals, notfor-profit organizations, and governmental agencies for their own research and conservation applications;
• redistribution without charge of the unmodified executable program for the purposes described above.
Unauthorized redistribution of VORTEX, in whole or in part by any for-profit organization or for any profit-making
purposes is expressly forbidden.
Cover Artwork: Linda Escher, Escher Illustrations.
Citation of this manual:
Miller, P.S., and R.C. Lacy. 2003. VORTEX: A Stochastic Simulation of the Extinction Process. Version 9.21 User’s
Manual. Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Citation of the software program:
Lacy, R.C., M. Borbat, and J.P. Pollak. 2003. VORTEX: A Stochastic Simulation of the Extinction Process.Version 9.
Brookfield, IL: Chicago Zoological Society.
Additional printed copies of VORTEX User’s Manual and installation CDs can be ordered through the IUCN/SSC
Conservation Breeding Specialist Group, 12101 Johnny Cake Ridge Road, Apple Valley, MN, 55124, USA. Send
checks for US $75.00 (for printing and shipping costs) payable to CBSG; checks must be drawn on a US bank.
Funds may be wired to First Bank NA ABA No. 091000022, for credit to CBSG Account No. 1100 1210 1736.
Lower costs for bulk orders may be arranged.
The CBSG Conservation Council
These generous contributors make the work of CBSG possible
Benefactors ($20,000 and above)
Chicago Zoological Society
Minnesota Zoological Gardens
Omaha’s Henry Doorly Zoo
SeaWorld, Inc.
Toronto Zoo
Conservators ($15,000 -$19,999)
Columbus Zoological Gardens
Saint Louis Zoo
Walt Disney’s Animal Kingdom
Wildlife Conservation Society - NYZS
World Association of Zoos &
Aquariums -WAZA
Zoological Society of London
Guardians ($7,000-$14,999)
Cleveland Zoological Society
Nan Schaffer
Toledo Zoological Society
White Oak Conservation Center
Zoological Society of San Diego
Protectors ($1,000-$6,999)
African Safari Wildlife Park
Albuquerque Biological Park
Allwetter Zoo Munster
ARAZPA
Audubon Zoological Gardens
Bristol Zoo
Caldwell Zoo
Calgary Zoo
Chester Zoo
Cincinnati Zoo
Colchester Zoo
Copenhagen Zoo
Denver Zoological Gardens
Detroit Zoological Park
Durrell Wildlife Conservation Trust
Everland Zoo
Federation of Zoological Gardens of
Great Britain & Ireland
Fort Wayne Zoological Society
Fort Worth Zoo
Fossil Rim Wildlife Center
Gladys Porter Zoo
Great Plains Zoo
Greater Los Angeles Zoo Association
Japanese Association of Zoological
Parks & Aquariums –JAZGA
Robert Lacy
Leisure & Cultural Services Department
of Hong Kong
Lincoln Park Zoo
Living Desert
Loro Parque
Marwell Zoological Park
Memphis Zoo
Milwaukee County Zoo
North Carolina Zoological Park
Oklahoma City Zoo
Paignton Zool. & Botanical Gardens
Parco Natura Viva Garda Zool. Park
Philadelphia Zoological Garden
Phoenix Zoo
Pittsburgh Zoo
Rotterdam Zoo
Royal Zoological Society of Antwerp
Royal Zoological Society of Australia
Royal Zoological Society of Scotland
Saitama Children’s Zoo
San Antonio Zoo
San Francisco Zoo
Schonbrunner Tiergarten
Sedgwick County Zoo
Taipei Zoo
Thrigby Hall Wildlife Gardens
Twycross Zoo
Union of German Zoo Directors
Wassenaar Wildlife Breeding Centre
Wilhelma Zoological Garden
Woodland Park Zoo
Zoologischer Garten Koln
Zoologischer Garten Zurich
Stewards ($500-$999)
Aalborg Zoo
Alice D. Andrews
Alameda Park Zoo
Arizona-Sonora Desert Museum
Banham Zoo & Sanctuary
BioSolutions Division of SAIC
Cotswold Wildlife Park
Dickerson Park Zoo
Dutch Federation of Zoological Gardens
Fota Wildlife Park
Givskud Zoo
Granby Zoo
Knoxville Zoo
Knuthenborg Park
Little Rock Zoo
National Aviary in Pittsburgh
National Zoological Gardens of Pretoria
Odense Zoo
Oregon Zoo
Ouwehands Dierenpark
Perth Zoo
Potter Park Zoo
Riverbanks Zoological Park
Rolling Hills Refuge Conservation Center
Staten Island Zoo
Tierpark Rheine
Wellington Zoo
Welsh Mountain Zoo
Wildlife World Zoo, Inc.
John S. Williams
Zoologischer Garten Rostock
Curators ($250-$499)
Dr. Edward & Marie Plotka
Emporia Zoo
Lee Richardson Zoo
Racine Zoological Society
Roger Williams Park Zoo
Rosamond Gifford Zoo at Burnet Park
The Animal Park-Gulf Breeze
Tokyo Zoological Park Society
Topeka Zoo, Friends of
Zoo de la Casa de Campo
Sponsors ($50-$249)
African Safari
American Loriinae Conservancy
Apenheul Zoo
Arbeitskreis Natur-u Artenschutz in den
Bighorn Institute
Brandywine Zoo
Darmstadt Zoo
Elaine Douglas
Folsom Children's Zoo
Jardin aux Oiseaux
Jean P. LeDanff
Kew Royal Botanic Gardens
Lisbon Zoo
Miller Park Zoo
National Birds of Prey Centre
Nigel Hewston
Steven J. Olson
Palm Beach Zoo at Dreher Park
Parc Zoologique de Thoiry
Prudence P. Perry
Safari Parc de Peaugres
Teruko Shimizu
Steinhart Aquarium
Tautphaus Park Zoo
Touro Parc-France
Supporters ($15-$49)
Oglebay's Good Children's Zoo
Judy Steenberg
Thank You!
August 2003
VORTEX Version 9 User’s Manual
Contents
Chapter 1. Introduction ..................................................................1
What’s New in VORTEX Version 9?........................................................... 1
How to Use This Manual......................................................................... 2
A Note about Regional Windows Settings ............................................. 3
VORTEX Technical Support....................................................................... 3
Chapter 2. Getting Started with Vortex ..........................................5
System Requirements ........................................................................... 5
Installation
.................................................................................. 5
Running VORTEX
.................................................................................. 6
Size Limitations on VORTEX Analyses ...................................................... 6
Getting Around in VORTEX ....................................................................... 8
A Quick Tour of VORTEX ........................................................................... 9
Chapter 3. Creating a Project: Data Input ....................................21
Creating a Project ................................................................................ 21
Documenting Your Input with Notes ................................................... 25
Creating a Scenario.............................................................................. 26
Scenario Settings
....................................................................................... 26
Species Description
....................................................................................... 28
Contents
i
VORTEX Version 9 User’s Manual
Labels and State Variables ............................................................................... 38
Dispersal Rates
....................................................................................... 41
Reproductive System ....................................................................................... 43
Reproductive Rates
....................................................................................... 48
Mortality
....................................................................................... 52
Catastrophes
....................................................................................... 53
Mate Monopolization ....................................................................................... 56
Initial Population Size...................................................................................... 57
Carrying Capacity
....................................................................................... 58
Harvest
....................................................................................... 61
Supplementation
....................................................................................... 62
Saving your Input and Running the Simulation................................... 64
Adding and Deleting Scenarios ............................................................ 66
Adding Scenarios to Your Project .................................................................... 66
Deleting Scenarios from Your Project.............................................................. 68
Reordering scenarios ....................................................................................... 68
Chapter 4. Viewing Model Results: Text, Tabular, and Graphical
Output ..................................................................69
Text Output
................................................................................ 69
Input Summary
....................................................................................... 69
Deterministic Calculations ............................................................................... 70
ii
Contents
Output Summary
....................................................................................... 73
Other Output
....................................................................................... 74
VORTEX Version 9 User’s Manual
Graphs and Tables................................................................................ 76
Data Graphs
Project Report
....................................................................................... 78
................................................................................ 81
Access to Other Stored Output............................................................. 82
Chapter 5. Using Functions in Vortex ...........................................83
Introduction
................................................................................ 83
Specification of Demographic Rates as Functions ............................... 84
Using Random Numbers in Functions .................................................. 88
Notes Regarding Function Syntax and Use .......................................... 88
Using Functions to Examine Genetic Evolution.................................... 90
Examples of Rate Functions ................................................................. 90
Appendix I. An Overview of Population Viability Analysis Using
VORTEX .................................................................101
Introduction
.............................................................................. 101
The Dynamics of Small Populations ................................................... 101
What is Population (and Habitat) Viability Analysis? ........................ 103
Population Viability Analysis (PVA).................................................... 104
Population and Habitat Viability Analysis (PHVA) ............................. 106
Methods for Analyzing Population Viability ....................................... 109
Modeling and Population Viability Analysis ....................................... 110
Contents
iii
VORTEX Version 9 User’s Manual
Dealing with uncertainty ................................................................... 112
Questions that can be explored with PVA models ............................. 113
The VORTEX Population Viability Analysis Model................................. 114
Demographic stochasticity............................................................................. 115
Environmental variation................................................................................. 115
Catastrophes
..................................................................................... 115
Genetic processes
..................................................................................... 116
Deterministic processes ................................................................................. 117
Migration among populations ........................................................................ 117
Output
..................................................................................... 118
Sequence of program flow ............................................................................. 118
Appendix II. Literature Cited......................................................123
Appendix III. Vortex Bibliography..............................................129
Appendix IV. Reprints .................................................................139
Lacy, R.C. 2000. Structure of the VORTEX simulation model for
population viability analysis. Ecological Bulletins
48:191-203............................................................ 139
Lacy, R.C. 2000. Considering threats to the viability of small
populations using individual-based models.
Ecological Bulletins 48:39-51. ............................... 139
iv
Contents
VORTEX Version 9 User’s Manual
Chapter
1
Introduction
VORTEX is an individual-based simulation model for population
viability analysis (PVA). This program will help you understand
the effects of deterministic forces as well as demographic,
environmental, and genetic stochastic (or random) events on the
dynamics of wildlife populations. VORTEX models population
dynamics as discrete, sequential events (e.g., births, deaths, catastrophes, etc.) that occur according to
defined probabilities. The probabilities of events are modeled as constants or as random variables that
follow specified distributions. Since the growth or decline of a simulated population is strongly
influenced by these random events, separate model iterations or “runs” using the exact same input
parameters will produce different results. Consequently, the model is repeated many times to reveal the
distribution of fates that the population might experience under a given set of input conditions.
VORTEX simulates a population by stepping through a series of events that describe the typical life cycle
of sexually reproducing, diploid organisms. The program was written originally to model mammalian and
avian populations, but its capabilities have improved so that it can now be used for modeling some
reptiles and amphibians and perhaps could be used for fish, invertebrates, or even plants—if they have
relatively low fecundity or could be modeled as if they do.
The purpose of this manual is to provide you with complete instructions on how to install and use
VORTEX. It is not intended as a primer on population biology; you must be conversant with this discipline
to use the program appropriately and effectively. In addition, you must know something about the biology
of the species that you intend to model. You should gather as much information as possible in order for
VORTEX simulations to be meaningful. The old computer adage of “garbage in, garbage out” is aptly
applied to population viability analysis, and PVA using VORTEX is certainly no exception. Having said
this, it is important to recognize that many of the questions VORTEX asks as you construct your population
model cannot be answered simply because the data do not exist. The only recourse that you will have is to
enter your best guess. Oftentimes, your best guess is not yours alone; most (if not all) population viability
analyses have succeeded through the efforts of many. Two or more heads are usually better than one
when you find yourself faced with a VORTEX question with no known answer. Further information about
VORTEX and the structure of the model is provided in publications reprinted as appendices to this manual.
What’s New in VORTEX Version 9?
The biggest change from prior versions of VORTEX is that the program is now a Windows application.
Although the user interface is now totally new, experienced VORTEX users will quickly recognize that the
content of the program (the input variables, the information output, etc.) is still very much like that of the
old MS-DOS versions of VORTEX. In fact, unless you invoke one of the few new features of the overall
model, results generated by the Windows version should match (except for stochastic uncertainty) the
results produced by the earlier DOS versions.
Chapter 1
Introduction
1
VORTEX Version 9 User’s Manual
You cannot directly import input files from prior (DOS) versions of VORTEX into version 9. However, for
an experienced VORTEX user, it usually takes only a few minutes to re-enter the input values from a prior
analysis. Attempts will be made to make future updated versions of VORTEX capable of importing projects
from all previous Windows versions of the program (version 9.0 and higher).
The user interface for entering input values, running simulations, seeing tabular and graphical
representations of output, and obtaining help are clearly very different in the Windows VORTEX than in the
earlier DOS versions. The easiest way to get a feel for these differences is to open the program and
explore it. A “Quick Tour” of the new program is provided in Chapter 2, and new users are encouraged to
use it to become more familiar with VORTEX version 9.
In addition to the switch to a Windows interface, there have been a few upgrades to the underlying
population biology model available in VORTEX. The most significant one (and the one that would be most
noticeable to users) is the availability now of “Individual State” variables. These optional variables allow
you to create descriptors of states or characteristics of individuals in the populations. These states can be
anything you want them to be – for example, dominance status, body condition, location on the landscape,
or territory quality. If you specify Individual State variables, then you must define how they are initially
determined for each individual at the outset of the simulation and when individuals are born, and how an
individual’s state can change across years. Once defined, these variables can be used as modifiers of any
of the demographic rates – such as probability of breeding, litter size, mortality, susceptibility to
catastrophes, and dispersal. Using this feature of the VORTEX model, you can create very complex and
detailed models of population dynamics. (For example, breeding could be function of the dominance
status of individuals, which could in turn be determined by maternal dominance status and a random
component.) However, we will not hide the fact that appropriate and wise use of Individual State
variables can be very difficult, and we strongly caution new VORTEX users to stay away from creating
such complex models.
How to Use This Manual
By following the detailed instructions provided in the VORTEX User’s Manual, you should be able to
construct surprisingly complex models of stochastic growth dynamics of wildlife populations. In addition
to this instruction, the User’s Manual provides you with supplementary information designed to help you
get the most out of VORTEX and to see how it and related software packages are used in practical
applications of conservation biology.
¾ Chapter 2, GETTING STARTED WITH VORTEX, gives you information on the program’s modest
¾
¾
¾
¾
2
system requirements and shows you how to install, run, and close the program.
Chapter 3, CREATING A PROJECT: DATA INPUT, provides a wealth of information on the types of
biological data necessary for developing a VORTEX population model and the mechanics of entering
data into the program.
Chapter 4, VIEWING MODEL RESULTS, describes how to view your model results, in text, tabular
and graphical form and how to work with output data files to assist in effective data analysis.
Chapter 5, USING FUNCTIONS IN VORTEX, presents a detailed description of how to use this major,
but complex, feature.
Appendix I, AN OVERVIEW OF POPULATION VIABILITY ANALYSIS USING VORTEX, gives a brief
introduction to the principles of small population biology and describes in more general terms the
use of population viability analysis to assist with wildlife management and endangered species
recovery.
Chapter 1
Introduction
VORTEX Version 9 User’s Manual
¾ Appendix II, LITERATURE CITED, provides a complete listing of the scientific literature referenced
throughout this User’s Manual.
¾ Appendix III, VORTEX BIBLIOGRAPHY, includes what we hope is a reasonably complete list of
references to papers that discuss VORTEX as a tool for population viability analysis, and to those
specific examples of the use of VORTEX in PVAs across a diverse taxonomic range. Authors
wishing to have their publications listed in future editions of this manual should email the citations
to help@vortex9.org.
¾ Appendix IV, REPRINTS, provides copies of two papers that describe in detail the structure of the
VORTEX model and some of the concepts behind the model.
Throughout the text you will find various aids that will enhance your overall use of VORTEX:
Brief explanatory notes that will help you remember important points, clarify some
commonly-used terms, etc.
Text boxes that will provide additional information on general concepts in
population biology and genetics, statistics, and simulation modeling.
Case Studies that show you real examples of how data have been used to develop
VORTEX simulation models. These case studies are gleaned from the many Population
and Habitat Viability Assessment (PHVA) workshops conducted by CBSG over the past
decade.
A Note about Regional Windows Settings
Although we cannot guarantee that VORTEX will work correctly with all possible configurations of MS
Windows, we believe that it will adapt appropriately to most Regional Settings of date, time, and numeric
formats. Throughout this manual, screen displays are shown from a system configured with the American
English regional settings. For example, the ‘.’ is used as the decimal delimiter (so that the number three
would be shown as 3.0). If your operating system is configured to use the ‘,’ as the decimal delimiter,
then you would use that format throughout VORTEX for input and output (so that the number three would
be shown as 3,0). Do not use any delimiter between thousands (e.g., thirty thousand would be 30000,
not 30,000 nor 30 000, nor 30.000). VORTEX will try to automatically convert input and output files
to the data format specified by your Windows Regional Settings.
VORTEX Technical Support
If you are having trouble using VORTEX and want additional information, there are a number of resources
available to you. Be advised, however, that CBSG is unable to provide the kind of technical support you
have come to expect (but rarely receive!) from large software companies. In this context, the phrase “you
get what you pay for” is particularly appropriate. VORTEX is provided on the Internet free of charge
because of our commitment to promoting the use of science in the service of biodiversity conservation.
Significant resources have been provided over the years by the Chicago Zoological Society and the CBSG
to support the development and continual improvement of VORTEX. Neither organization recovers these
costs of development, nor receives any funding to provide ongoing support to VORTEX users.
Chapter 1
Introduction
3
VORTEX Version 9 User’s Manual
Nevertheless, we are committed to doing everything we can to help you get the most out of your VORTEX
modeling experience. Towards this end, we suggest the following support options:
¾ This User’s Manual. We hope that we have provided you with all of the information necessary to
¾
¾
¾
¾
¾
navigate your way through the program.
Tooltips and Input Prompts. Most icons, commands, menus, and input boxes have tooltips that pop
up with explanatory messages when the cursor is paused over them. In addition, during input,
prompts will appear at the bottom of the window when the user clicks on a data entry box.
On-screen Help. Chapters 1-5 of this manual are provided in the Help menu of the program.
The VORTEX Listserver. To help VORTEX users in their use of the program for PVA, a VORTEX Users
email discussion group (Listserve) has been established. The VORTEX Listserve facilitates the
exchange of ideas, questions, answers, and suggestions among the many users of VORTEX. The
listserve also provides a medium for announcing updates, bug fixes, and suggestions provided by
CBSG or by the program’s developer. To get information about the listserve, or to subscribe, go to
https://listhost.uchicago.edu/mailman/listinfo/vortex. Because there is no registration of VORTEX
users, the listserve is only way we can assure that users will hear about updates, bug fixes, and
other announcements.
VORTEX on the Web. Explore the VORTEX home page at http://www.vortex9.org/vortex.html
to download updated programs or documentation files, to report program bugs, or to obtain other
information pertinent to the effective use of VORTEX.
Contact the CBSG Office. As a last resort, if you are unable to solve your problem by the means
suggested above, you can reach the CBSG Office directly to get help. Our contact information is:
Telephone: 1-952-997-9800; Fax: 1-952-432-2757; E-mail: office@cbsg.org.
We urge you to read the entire User's Manual not only to better your understanding of VORTEX, but also
to enhance your appreciation of the perils facing small populations of threatened wildlife. For a more indepth treatment of population viability analysis and models for use in risk assessment, we recommend
Starfield and Bleloch (1986) and Burgman et al. (1993) as excellent introductions to these topics.
A Note about Cost
Vortex is provided free of charge because of the commitment of the Chicago Zoological Society to
making it widely available to further biodiversity conservation. Similarly, the manual, developed by the
CBSG, is provided for downloading because the CBSG cares about saving species and their habitats.
However, the initial development and continuing improvement of the software and manual do represent a
significant commitment by these conservation organizations. The rate at which improvements can be
made is determined by the resources available to support that work.
If your budget allows it, please consider making a donation to support the further development of Vortex.
If you find the software to be especially valuable to you, consider donating perhaps US$100 (a wild guess
about the investment of resources per user that have gone into Vortex), or more or less as you feel is
appropriate.
If you find the manual to be especially helpful, consider donating to the CBSG. As a side benefit to US
tax-payers, donations to either the Chicago Zoological Society or the CBSG are tax-deductible. Donations
to the Chicago Zoological Society should be as a check written to the Chicago Zoological Society, sent to
“Vortex donation, Department of Conservation Biology, Brookfield Zoo, Brookfield, IL 60513 USA".
Donations to the CBSG should be sent to "Vortex donation, CBSG, 12101 Johnny Cake Ridge Road,
Apple Valley, MN 55124, USA".
4
Chapter 1
Introduction
VORTEX Version 9 User’s Manual
Chapter
2
Getting Started
with VORTEX
System Requirements
VORTEX version 9 was developed as a C++ program (for the simulation code) presented within an
interface developed in MS Visual Basic. To our knowledge, it will install and run properly on computers
with Pentium (or newer) processors running Win95, Win98, WinXP, Win2000, or WinNT operating
systems. We believe that VORTEX will work properly with a diversity of Windows Regional settings – for
example, it can use common European data formats. However, we cannot guarantee that it will work with
all system configurations. At this time, the user interface of VORTEX is presented in American English. At
a future time, versions in Spanish, German, French, or other languages may be made available.
For many analyses, VORTEX will use much of your computer’s system resources. Faster results and better
performance will be obtained if you do not try to run other large applications (such as MS Word, Excel,
or Outlook) at the same time that VORTEX is running. The program may not run properly with less than 64
MB of system memory (RAM), and even more RAM will be required if you want to run other
applications concurrently. In addition, the size of the populations that can be analyzed will be determined
by the available RAM. For example, simulation of a population of 5000 living animals can require up to
about 50 MB of RAM for storage of inbreeding calculations. The program requires much more memory if
you include inbreeding depression in your analyses, so omitting inbreeding depression (see Chapter 3)
will allow analysis of larger populations and will run much faster.
Installation
To install VORTEX from the CD:
Place the VORTEX installation CD into the appropriate disk drive, and then run the SETUP
program. (Either double-click on SETUP from Windows Explorer, or go to START>RUN, and then
enter D:\SETUP.EXE, in which D: is your CD drive.)
To install VORTEX from the Internet:
Go to http://www.vortex9.org/vortex.html and download the current version of the installation
program. Save this downloaded file to any temporary directory of your hard disk or to your
desktop. Double-click on the downloaded file to unzip the installation package files. Unzip them
to the directory where you saved the downloaded file, or to any directory of your choosing, other
than the directory to where you wish to install VORTEX. Run the SETUP program. (Either doubleclick on SETUP from Windows Explorer, or go to START>RUN, and then enter
Chapter 2
Getting Started with VORTEX
5
VORTEX Version 9 User’s Manual
C:\TEMP\SETUP.EXE, in which C:\temp is the drive and directory where you placed the installation
files.)
You may want to put a short-cut to VORTEX on your Desktop.
VORTEX is copyrighted but not copy-protected. You can make as many copies as you wish, and you may
give copies of the program to others free of charge. You may not sell the program or any components of
it, or otherwise represent it as your personal property.
Running VORTEX
Start VORTEX by double-clicking on a short-cut icon, or by double-clicking on the VORTEX program itself.
To exit the VORTEX program, just click on the Close button (marked by an ×) in the top right corner of the
program window. before closing, the program will always prompt you to determine if you wish to save
any open projects.
Size Limitations on VORTEX Analyses
VORTEX allocates computer memory as it needs it, depending on the characteristics of the population or
metapopulation you are modeling. VORTEX will make optimal use of all available memory to carry out the
simulations, but the available RAM on your computer may limit the size of analysis you can complete.
However, there are also some absolute limits to how large or complex a simulation can be. These limits
are listed below.
¾
¾
¾
¾
¾
¾
¾
¾
Number of iterations
10000
Duration of simulation
2000 years
Number of populations
50
Types of catastrophes
25
Maximum age
250 years
Maximum litter size
50*
Initial population size
30000 individuals
Carrying capacity
60000 individuals
*Only if specifying an exact distribution; see page 44 for an alternative method to increase maximum litter size.
Some combinations of parameters can require large amounts of memory. For example, if you are
including inbreeding depression in your simulation, and have chosen to model it as only partially due to
the presence of lethal alleles, more than 50 megabytes of memory may be required to analyze a
population that reaches 5000 living animals. In these cases, it is possible that VORTEX will abort an
analysis if there is insufficient memory available. Even if the program does not abort, it may run
exceedingly slowly as each individual and its pedigree is tracked throughout its lifetime. If you have
frequent problems with aborted or slow analyses, consider taking one or more of the following steps:
Change the mechanism by which inbreeding depression is modeled. The program will run
noticeably faster if the population’s genetic load is due entirely to lethal as opposed to detrimental
alleles.
¾ Construct a simpler general model. Often a large population (or metapopulation) may exist, but
the real concern may be whether smaller fragments are at risk of local extirpation. VORTEX will
simulate small populations much more rapidly than large populations or constellations of patches
within a metapopulation. If local patches do not exchange migrants, analyze them separately
rather than as parts of a larger, more complex metapopulation.
¾
6
Chapter 2
Getting Started with VORTEX
VORTEX Version 9 User’s Manual
¾
Think about using a different PVA software package. If VORTEX is running so slowly as to cause
you much grief, the types of populations you are analyzing are probably so large that the kinds of
random forces modeled explicitly by VORTEX —demographic and environmental stochasticity,
inbreeding, and genetic drift—are likely to be irrelevant to the population growth dynamics. In
these cases, it may be more appropriate to use a population-based model such as one of the
packages in the RAMAS family of software (these are produced and distributed by Applied
Biomathematics, Setauket, NY), or to use analytical methods (e.g., life-table analysis) that
exclude most or all stochastic factors entirely.
Box A: Is VORTEX the Best PVA Model for Your Analysis?
Different PVA models have different strengths and weaknesses with respect to what kinds of life
histories they can model, what range of processes can be examined, what aspects of population
dynamics are modeled well, and how easy they are to use for different analyses. Below is a list of
some considerations for evaluating whether VORTEX is more or less appropriate for your analysis.
While they are certainly not hard and fast rules, they should help you make informed decisions about
how to best conduct your analysis.
VORTEX may be less appropriate or
may not be needed
VORTEX is more appropriate
and may be necessary
High fecundity
Low fecundity
Short lifespan
Long lifespan
Polyploid
Genetic effects of little interest
Local population (N) > 500
> 20 populations modeled
Demographic rates not estimable
Diploid
Changes in genetic variation of interest
Local population (N) < 500
< 20 populations modeled
Age-specific fecundity and survival rates estimable
(only population growth trajectories known)
Stage- or size-dependent demography
Demographic rate fluctuations not estimable
No catastrophic events of interest
Only polygamous breeding
Random breeding
Age-dependent fecundity and survival rates
Fluctuations in rates can be estimated
Catastrophic events modeled
Polygamous or monogamous breeding
Some adults excluded from breeding
Non-random distribution of fecundity
Population starts at stable age distribution
Constant sex ratio
No trends in habitat expected
No manipulation of animal numbers
Fish, amphibian, invertebrate, or plant
You have lots of money
(for buying software)
Starting population not at stable age distribution
Unequal sex ratio
Trends projected in habitat quality or area
Managed removal, supplementation, or translocation
Bird, mammal, or reptile
You have lots of time
(for running analyses and summarizing results)
Chapter 2
Getting Started with VORTEX
7
VORTEX Version 9 User’s Manual
Getting Around in VORTEX
Your work in VORTEX will be structured as Projects and Scenarios. A Project will contain all your input,
output, and notes about a case that you are exploring. Often, a Project will contain all the analyses about a
given species or population. You could split your analyses of a species among multiple Projects, but that
would preclude you from easily copying input, results, or settings among the separate Projects containing
your work. On the other hand, there is no advantage to combining work on different species or cases into
one Project, and it may be more useful and less confusing to keep distinct PVAs in separate VORTEX
Projects. It is probably a good idea to specify a new directory for storing each Project (and this is the
default in VORTEX), although you can store all your work in one directory if you wish.
Each Scenario within a Project contains a discrete set of input values and (if it has been run) output. Thus,
for a given species (Project) you may decide to test several or even many different Scenarios, each of
which would have an alternative set of input values, representing an alternative view of the population.
For example, different Scenarios may represent various plausible input values to be explored during
sensitivity testing, or may represent alternative management options that might be applied to a population.
The VORTEX interface has separate screens (windows) or tabs for Project Settings, Simulation Input, Text
Output, Graphs and Tables, and a Project Report. Each of these are specific to an open Project, and you
can toggle among the Scenarios of a Project within the input and output screens. You can open
concurrently multiple VORTEX Projects within a VORTEX session, although there may only rarely be cases
in which it is useful to have more than one Project open at the same time.
A caution: It is almost inevitable that VORTEX contains some bugs, and it may be that you will make some
mistakes while working with VORTEX. Thus, it is possible that after you spend hours working on a
VORTEX Project, the program will suddenly crash. It is also possible that you will accidentally change a
very useful analysis into something that is worthless. It is strongly recommended that you periodically
save your work, and even save it under a new name in a new directory (see below). Hard disk space is
cheap – use it!
8
Chapter 2
Getting Started with VORTEX
VORTEX Version 9 User’s Manual
A Quick Tour of VORTEX
The first screen you will see when you begin a new VORTEX session is shown in Figure 1.
Figure 1. The Vortex opening screen
After admiring this artistic representation of the extinction vortex, and appreciating the fact that the
Chicago Zoological Society devoted a lot of resources to develop VORTEX for your use, click on the
“Close graphic …” message to enter the program.
The next screen, shown in Figure 2, asks what Project you wish to open. (Note: in Figure 2, and many
subsequent figures, the image shown is just the sub-window that is relevant to the point being made.)
Select the Open Project tab, so that we can use an existing sample Project. A number of sample projects
are copied into the Projects subdirectory when you install VORTEX. More sample projects will be made
available at http://www.vortex9.org/vortex.html, and we encourage users to contribute their project files
to this site so that others may explore those data sets.
Chapter 2
Getting Started with VORTEX
9
VORTEX Version 9 User’s Manual
Figure 2. The dialog box for starting a Project.
Figure 3 shows the dialog box for opening an existing Project. This screen may look a little different on
your computer, depending on where you installed the VORTEX program. Navigate to the \ZPG directory
and then select the ZPG Project by double-clicking on the ZPG.vpj file, or single-clicking on it and then
selecting OK. Note that windows in VORTEX can usually be resized or moved with the cursor.
Figure 3. The dialog box for opening an existing Project.
The ZPG Project does not represent any particular species. It has a set of input values that define a
population that would have an expected long-term zero population growth (r = 0.0), based on the mean
birth and death rates as modified by occasional catastrophes. This projection of zero population growth is
dependent upon an assumption that stochastic processes – such as demographic stochasticity, temporary
10 Chapter 2
Getting Started with VORTEX
VORTEX Version 9 User’s Manual
mate limitation, inbreeding, and annual fluctuations (“environmental variation”) – do not reduce mean
population performance. The population size in this case, however, is low enough that these stochastic
processes are important impacts on the population, causing it to be unstable, often decline, and be highly
vulnerable to extinction (as we will see). The ZPG Project has the same input values as the default values
in the earlier DOS versions of VORTEX.
When you open a Project, VORTEX opens a screen that shows the Project Settings interface and the tabs
for the other screens – Simulation Input, Text Output, Graphs and Tables, and Project Report (Figure 4).
On the Project Settings screen, you can specify a different Project name, enter the names of any
collaborators, and add any Project Notes text that you wish to describe your Project. It is wise to take the
time to document your work by typing Project Notes and, on a screen you will see later, on Input Notes.
At the time you are working, it may be seem obvious what decisions you were making when you created
your Project. However, months later it may be very difficult for you (or others) to recall what led you to
design the Project as you did. If you want ever to go back to a Project, take the time now to document
your work within the VORTEX Project (see Chapter 3 for more information on entering Input Notes).
Figure 4. The Project window.
Chapter 2 11
Getting Started with VORTEX
VORTEX Version 9 User’s Manual
Get into the habit of adding Notes to your Projects to document what you are doing. You will be
glad that you did when you later need to tell others what you did in your analyses.
Also on the Project Settings screen is a button that will send the Project Settings information to your
Project Report. Your Report is a note pad utility (much like MS Notepad) that lets you build
documentation of your Project. We will take a look at the Project Report soon; for now, click on the
“Send all to Report” button to capture the settings information in your Report.
Send information to your Report whenever you think that you may want it documented. It is
easy to delete parts of the Report, but it is hard later to see something that you never sent !
The last item on the Project Settings screen is a Special Options button. These options are ones that most
users will never need to use, so we won’t look at them now. Click now on the Simulation Input tab, to
take you to the screen shown in Figure 5.
Figure 5. The Simulation Input window.
12 Chapter 2
Getting Started with VORTEX
VORTEX Version 9 User’s Manual
Simulation Input is arranged as 13 screens that each request values for a section of input parameters
(“Scenario Settings”, “Species Description”, etc.). Clicking on one of the section labels in the list on the
left side of the screen takes you to that section of input. Within a section, it does not matter what order
you enter values, and the input sections can be accessed in any order you wish. However, it makes sense
to enter values in the order they appear in the program, so that you don’t forget to specify some critical
value. In addition, some input sections will use values already entered from prior sections to compute
useful values (such as the stable age distribution) during input. Notice that one section label, “Dispersal
Rates”, is greyed out and disabled. That is because the current Scenario has only one population, so there
can be no dispersal among populations. Similarly, some other sections and individual input boxes will
become disabled if values you have specified would make that section meaningless.
Take a quick look at the data input boxes on the Scenario Settings screen. As you click on any box, a
message will be displayed at the bottom of your screen with hints about what you need to enter into that
box. Now click on the “Species Description” label on the left to take you to that input section (Figure 6).
Note again that some input boxes are disabled, because they pertain only to metapopulation models.
Figure 6. Species Description input section.
Chapter 2 13
Getting Started with VORTEX
VORTEX Version 9 User’s Manual
Step through each of the input sections, looking at the input values that are requested by VORTEX and the
values that were entered for this ZPG Scenario. In some sections, you may need to use vertical or
horizontal scroll bars to see all of the data entered. You can also make the input screens larger by clicking
and dragging the corner of the window. Go now to the Catastrophes section (Figure 7).
Figure 7. Catastrophes input section.
Because it was specified in the Species Description that the model should contain two types of
catastrophes, the Catastrophes section has buttons to toggle between these two types. Hit the button to go
to input for Catastrophe 2. In this particular Scenario, the two types of catastrophes have the same input
values, so it is not obvious that you are moving between the two types (but you are).
When you are done looking through all the input sections, click on the Run icon (the green triangle on the
icon bar) to open up the Run Simulation dialog box shown in Figure 8. Check the box to select Scenario
ZPG1, and then click Run!
14 Chapter 2
Getting Started with VORTEX
VORTEX Version 9 User’s Manual
Figure 8. Run Simulation window.
Now, sit back and watch the simulation work. The lines on the screen (Figure 9) show the changing
population size over 100 years for 100 different iterations for the ZPG1 Scenario. When the simulation is
complete (which should take only a few seconds with this small population), the VORTEX Simulation
display window will show a few summary statistics along the top. When you are done viewing this
graphical display of the simulations, click on its close icon (×) in the upper right corner.
Figure 9. VORTEX Simulation display window.
Chapter 2 15
Getting Started with VORTEX
VORTEX Version 9 User’s Manual
The VORTEX Simulation window cannot be resized, and toggling to another window during the
VORTEX simulations may leave you with a blank VORTEX Simulation window when you return to it.
It is best not to move off of this window while the simulations are running. When the simulations
are complete, close the window before doing anything else.
The results of the simulation you just completed are now stored with your Project on the computer. There
are two modes in which you can view the results – “Text Output” and “Graphs and Tables”. Go to Text
Output by clicking on its tab. Within the Text Output section are four tabbed subsections – Input
Summary, Deterministic Calculations, Output Summary, and Other Output (Figure 10).
Figure 10. The Output Summary section of Text Output
The Input Summary section shows a text listing of all the input values used in this Scenario. Deterministic
Calculations show a text summary of the deterministic population growth that would be projected from
the specified mean demographic rates, if stochastic processes were not acting on the population. This
section also shows a simple graph of the deterministic population trajectory. The Output Summary section
gives a text description of the status of the population at each year of the simulations as well as summary
16 Chapter 2
Getting Started with VORTEX
VORTEX Version 9 User’s Manual
statistics for the Scenario. Other Output provides some tables with basic summary statistics for the
Scenario and for each iteration. Note that the sections of Text Output provide dropdown lists to allow you
to move among Scenarios and Populations. Buttons are also provided to allow you to save these texts to
simple text files, to print the text summaries, or to send the summaries to the Project Report. Send the
Output Summary to the Report now, so that we can view and edit it later as part of our Project Report.
All of the information shown on Text Output screens is stored automatically in text files that are
placed into your project directory. While you can access and edit these files (using, for example,
MS Notepad or Word), it is better to first save the text to your own files, so that they are safely
stored under names that you specify and will not be overwritten if you run the simulation again.
Click now on the Graphs and Tables tab (Figure 11). The program may display a warning message stating
that it could not find the data from one or more Scenarios that have not yet been run. That is OK for now
(as long as you don’t need yet to see the results from these other Scenarios). The Graphs and Tables
section has two subsections – Data Specification and Data Graphs. Data Specification is where you will
identify which results you wish to put into your table and graph.
Figure 11. The Graphs and Tables window.
Chapter 2 17
Getting Started with VORTEX
VORTEX Version 9 User’s Manual
In the lower left list of Scenarios, make sure that ZPG1 is checked, and then double-click on the box
under Columns. This will bring up a window that lets you specify which years you want to show as the
columns of your table and the x-axis of your graph (Figure 12).
Figure 12. Specification of years as the columns for tabular output.
To specify years, you can Select All, select individual years by clicking on their boxes, or select rows (all
years in a decade) or columns (years in decadal intervals) from the table. For example, to select years 0,
10, 20, …, click on the number 0 at the top of the first column. (Do this now.) The method by which years
are selected may be a little confusing at first, but once you learn how it works it does allow very rapid and
flexible specification of sets of years. If you add years to your selection in a non-sequential order, you
will want then to sort them by clicking on one of the sort buttons on the right. After you have selected the
years you want, click on OK.
When changing your selection of reporting years, you may obtain better results if you Unselect All
before making your new selection rather than unchecking multiple boxes.
Next click on the box below Rows, in order to specify which populations you want to list as rows of your
data table and as separate lines on the graph that will be created. The way you select populations is the
same as selecting years, except that in this Scenario there is only one population so selecting it is fairly
trivial. Select Population 1 and hit OK. You will see now that a table has been created, displaying one of
the result statistics for the years and population(s) that you specified. With the dropdown lists in the left
side of the screen, you can change the table to display other output statistics.
Change the Variable to N(all), and then hit the Data Graphs tab to show a graph of the values in the table
(Figure 13). The labels, legend, and line thickness can all be changed. By right-clicking on the graph
itself, you can also access a broader set of graph properties. By clicking on labels at the lower left, the
graph can be sent to the Project Report (do this now), or printed, or saved as a bitmap (.bmp) file on your
disk.
18 Chapter 2
Getting Started with VORTEX
VORTEX Version 9 User’s Manual
Figure 13. A Data Graph.
You should note that you have the option of adding bars to your graph to show standard errors (SE) of the
means, or standard deviations (SD) across the iterations. Click on the “Add SD bars” command to see
these bars.
Finally, to wrap up our Quick Tour of VORTEX, click on the Project Report tab. This will take you to a
note pad that contains information we have sent to the Report from other screens (Figure 14). You will
need to use the scroll bar or your cursor to move up and down through your Project Report. Any
information in the Project Report can be edited, using standard Windows editing tools (delete, cut, paste,
font settings, etc.). The Project Report is saved in Rich Text Format (and .rtf) file, and it can be edited in
Word or other programs.
Chapter 2 19
Getting Started with VORTEX
VORTEX Version 9 User’s Manual
Figure 14. The Project Report.
You have now completed your Quick Tour of VORTEX, and we have looked at most of the main features
of the program. Spend some time exploring other aspects of the program – change some of the input
values, run additional scenarios, create some more tables and graphs. Whenever you exit VORTEX, the
program will ask if you want to save your Project. If you do, all input, output, and report information will
be saved so that it can be loaded again later.
20 Chapter 2
Getting Started with VORTEX
VORTEX Version 9 User’s Manual
Chapter
3
Creating a Project:
Data Input
Creating a Project
When you open VORTEX, you must first choose whether to create a new Project or open an existing
Project (Figure 15). To create a new Project, double-click on the Blank Project (or click on the “Blank
Project” and then hit OK). The Open Project tab will allow you to browse to find an existing project. The
Recent Projects tab will allow you to select from a list of the 10 most recent Projects that you have
worked on. You can get to these same options to create a new Project or open an existing Project from
either the menu or the tool bar at the top of the VORTEX screen.
Figure 15. The welcome window for starting a VORTEX Project.
If you choose to start with a Blank Project, the only input values that will be pre-filled are a few that are
necessary to define the basic Project and Scenario properties. It is often easier to start a new Project by
opening an existing Project, and then changing those input parameters that are different. However, be sure
to go through every input screen to confirm that you have set the input parameters to the new values, and
be sure to save the Project under a new name. When you chose to create a new Project, you next need to
specify a Project name, and you have the option of recording your name as the Project creator (Figure 16).
Chapter 3 21
The Data Input Process
VORTEX Version 9 User’s Manual
You can also specify the directory in which the Project files will be stored, but it usually is reasonable to
accept the default, which is a subdirectory with the same name as the Project. Click OK to continue.
Figure 16. Dialog box for entering a new project name.
Figure 17. The Project Settings window.
22 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
In the Project Settings windows (Figure 17), you have the option of listing the names of the user team that
is developing the project (this documentation may be especially helpful in workshop or classroom
settings), and add any notes that you wish.
We strongly encourage you to take the time to add notes to your Project at this screen, during
specification of input parameters, and in your Project Report. The extra few minutes you spend
documenting your work may save you and others many hours of work later, when you try to
remember what information and logic was used to create the project. Unfortunately, many PVAs
are irreproducible because the authors did not fully document their work.
The Project Settings screen also has a button to send all of the settings to your Project Report (this is
always a good idea, so that your settings are documented in any printed reports that you create), and
another button that takes you to a screen for specifying Special Options.The Special Options will not be
needed by most users. They include options to:
•
change the way population sizes are graphed during the simulations,
•
use the last population as a holding site for individuals that are harvested from one population and
then supplemented into others (if this option is chosen, then you specify also what percent of the
individuals die during this translocation among populations),
•
omit the last population from metapopulation tallies (this is useful if the last population is
considered an outside source for immigration into a metapopulation),
•
prevent individuals from dispersing into populations that are at their carrying capacity (where the
immigrant or some individual would therefore die because of the population exceeding capacity),
•
define extinction as any reduction in population size (this is useful when the management goal is
to prevent further population declines),
•
produce files with more detailed results,
•
include more neutral loci in the genetic model (useful for examining the impacts of population
dynamics on genetic structure),
•
invoke other options that may from time to time be made available, usually on a test basis or for
special circumstances. (To use this option, you would need to know the undocumented codes for
using these additional options.)
To begin entering the values for the parameters that will specify the Scenarios of your Project, click on
the Simulation Input tab.
Getting Help when Entering Input Data
VORTEX will accept most of the input that you provide, as long as the values are biologically possible and
within the rather wide limits set by the program (see above). When entering input data, brief hints about
the values to be entered will be displayed in a line at the bottom of the Project screen. These messages
appear when you click on a data entry box, and sometimes they will appear as pop-up tooltip messages
when your cursor passes over a data entry box for more than a few seconds (Figure 18).
If you try to enter a value for input that is of an incorrect type (e.g., a letter when a number is required) or
outside the acceptable bounds (e.g., a negative number for a mortality rate), then VORTEX will usually
display a message that the value is invalid, and it will force you to enter a valid value before you proceed
with data entry.
Chapter 3 23
The Data Input Process
VORTEX Version 9 User’s Manual
Figure 18. Scenario Settings within Input, showing a tooltip for the Extinction Definition.
It is important to remember that VORTEX will accept input values that are mathematically possible but
biologically implausible. While VORTEX provides help on many data input questions you may have, such
as when to enter the data as a proportion or as a percent, the ultimate responsibility for entering valid data
that will result in a meaningful model rests with you, the user.
Most of the material in Chapters 1-5 of the manual is available through the Help menu of the program.
Selecting “Contents” on the Help menu will take you to a Table of Contents, which provides links to each
section of the Help manual. (Click on a section heading in the Table of Contents to jump to that topic in
the manual.) Selecting “Context-sensitive” help from the menu or clicking on the ‘?’ icon on the toolbar
will open the Help file and jump right to the place in the file that describes the current program screen.
24 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
Documenting Your Input with Notes
Whether you are using VORTEX as a researcher, a wildlife manager, or a student, it is highly likely that
you will need to document for others (and even for yourself when you return to a Project that had been set
aside for a period of time) why you used the input values you did to create your Scenarios. When you are
eager to get a Project running, it is very tempting to skip over the task of documenting the sources and
reasons for your input values. However, you may save yourself later hassles if you take the time at the
outset to record why program options and parameter values were chosen.
VORTEX provides a utility to attach a note with each piece of data requested as input. You can access the
Input Notes by any of three methods: clicking on the Notes icon at the lower right corner of the
Simulation Input screen; clicking on the same icon on the program toolbar; or directly typing into the long
text box below the input window. After you type a note into the text box showing on the Simulation Input
screen, hit the ‘+’ button to add that Note to your Project, or just hit Enter. Your Note is then associated
with the input parameter or question that last had cursor control. (Be careful not to enter a Note and then
immediately click on another data entry box. If you didn’t hit ‘+’ or Enter, your Note will be discarded.)
When you open Input Notes by clicking on its icon, you then select the Input section input parameter for
which you wish to enter a note (Figure 19). You then enter the text of your note in the box at the right.
Figure 19. Input Notes pop-up utility.
As you move among input screens and boxes, the Note for the input box that is selected is displayed in
the text box below the input window.
The Input Notes screen provides commands for pasting the displayed Note or All Notes into your Project
Report, and for printing the displayed Note or All Notes. Thus it is easy to quickly insert your Input Notes
into a report of your work. Input Notes are always saved when you save a Project.
Chapter 3 25
The Data Input Process
VORTEX Version 9 User’s Manual
Creating a Scenario
Input of model parameters into VORTEX is accomplished in 13 sections, each containing questions
pertaining to a category of model parameters. You move among the input sections by clicking on their
labels in the list on the left hand side of the Simulation Input screen. You can move among the sections to
enter data in any order, although the list provides a logical sequence for data input. After you have visited
an input section within a VORTEX session, the label for that section will be in italics. This may help you to
quickly check whether you have completed data entry for every section.
If you jump around among input sections, you risk forgetting to visit a section, and then running
models that are missing some parameters or that have values from scenarios used as templates.
In addition, VORTEX uses answers on some early screens to complete intermediate calculations
(such as the stable age distribution) that are useful when you reach later input sections.
Scenario Settings
The first data input screen you encounter when creating a Scenario asks for some basic Scenario Settings
(Figure 18 above). Subsequently, you will need to step through 12 more input screens to complete the
process of specifying the values for all of the input parameters needed by VORTEX. Below are described
all the input parameters requested on these screens.
Scenario Name: Within each project you create scenarios that are defined by their sets of parameter
values. As you will see, after you have defined one scenario (often a “Baseline” or “Best Guess”
scenario), it is easy to create additional scenarios that change one or a few of the input values. The default
scenario name for a new project is just “Scenario 1”. On the Scenarios Settings screen, you should change
this to a more descriptive name.
Number of Iterations: The answer to this question instructs VORTEX on how many times you wish to
repeat the simulation, given the data that you provide in the subsequent steps. Each repetition is generally
defined as a “run” or “iteration”. Because VORTEX uses a random number generator to simulate random
events in the life cycle, no two iterations will be identical. Thus, to obtain a more complete “picture” of
your simulated population, you will want to generate multiple iterations of your model.
As a first step in the development of a sound population model, you may want to make sure that the
simulated population is behaving in a manner that is similar to your expectations. To check this, you can
limit the number of iterations to just 10 or 20. If you wish to obtain a relatively crude picture or your
results, use 100 iterations. Once you are comfortable with the model and wish to obtain a more rigorous
description of the simulated population’s behavior, it is not excessive to enter 500 or even 1000 iterations
in this field. Note that commas are not used when specifying larger numbers during the input process,
even if your computer is set to use American data formats .
Number of years: How far into the future do you wish to project your population? The usual answer to
this question is 100 to 200 years, although a shorter duration can be entered so that you can assess the
validity of your input parameters, or to examine the short-term viability of a population. If you simulate
your population for just a few decades, however, you should be aware that processes controlling
population dynamics might be leading the population toward extinction but, especially for long-lived
species, the final extinction may not occur until a later time. By the time that the factors influencing
extinction are apparent, the process may be so far along as to be almost irreversible. One of the major
advantages of PVA modeling is that it can reveal the instability of a population long before it would be
apparent through field observations.
26 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
An important point to keep in mind is that VORTEX does not necessarily require “years” to be defined as
calendar years. Rather, the program operates more broadly in terms of “time cycles”. If the species you
are modeling has a short generation time and life span, on the order of weeks or months—such as mice or
shrews, for example—true calendar years would be an inappropriate time scale to use for modeling
population dynamics. In this case, a “year” for this type of species may actually represent only one or a
few months. When calculating your demographic inputs, it is vitally important that you make this
adjustment consistent throughout your calculations (see Case Study I for more information).
Case Study I:
Calculating input parameters when the “time cycle” is less than one year
Consider a hypothetical rodent population where the average generation time is 180 days. In order
to model this population most effectively in VORTEX, the user must adjust the “time cycle” to account
for this shortened generation time. In this case, we will define a VORTEX “year” as 90 days.
Consequently, events whose occurrences are typically described on an annual or per-generation
basis must be redefined in terms of the new definition of “year”.
For example, consider a major catastrophic flood that is thought to occur on average once every 100
years. The annual probability of occurrence, then, is 0.01. Because of the altered definition of
“year”, the rodent model must define the probability that this flood will occur in any given 90-day
interval. The number of 90-day time cycles in a calendar year is T = 365 / 90 = 4.06. Therefore,
Pr(flood)90 =
Pr(flood)365 0.010
=
= 0.0025 .
T
4.06
The same considerations must be applied to all other demographic rates, such as mortality, age of
first and last breeding, etc. In addition, appropriate migration, harvesting and supplementation rates
must be established relative to the revised time cycle.
Extinction Definition:VORTEX gives you three methods to define “extinction” of your population. For
most sexually reproducing species, ultimate biological extinction is assured whenever the population has
declined to the point that it no longer has individuals of both sexes. In the first (and most common)
choice, extinction is simply defined as the absence of at least one sex.
You also have the option to assess the probability of a population dropping below a user-defined
threshold size – termed quasi-extinction. The use of quasi-extinction risk offers a useful alternative to the
standard extinction risk. If you chose to have the simulation tally quasi-extinctions, you need to specify
the threshold “critical size” below which a population is considered extinct. The simulation will, however,
continue to run, as the population may grow again to a size above this threshold. Such recovery from
quasi-extinction would be tallied as a recolonization event. A third option is available under Special
Options on the Project Settings screen, which defines extinction as any decline in population size.
Number of Populations: VORTEX can model a single, isolated population or a complex metapopulation
composed of up to 50 populations. A metapopulation is a group of populations which, because they often
occupy fragmented, discontinuous habitat, exchange individuals with varying frequency. Note that,
because of the added complexities associated with metapopulations, these models will often run
considerably slower than the corresponding single-population models.
If there is no exchange of individuals among populations (i.e., dispersal) in your model, it
maybe faster to run several individual simulations (with each one modeling an isolated
population) instead of a more complex metapopulation model.
Chapter 3 27
The Data Input Process
VORTEX Version 9 User’s Manual
Enter the number of populations that comprise your metapopulation model or enter 1 for a simulation
composed of a single population. If you intend to build a metapopulation model, you will later need to
specify dispersal rates and some other parameters.
Species Description
The next section of input includes a set of basic questions about the species being modeled (Figure 20).
Figure 20. Species Description section of Input.
Inbreeding depression: Check this box if you want to include inbreeding depression in your model, as a
reduction in first-year survival among inbred individuals. (See Box B for more information). Although
most diploid species that have been studied show depressed fitness when inbred, you may sometimes
want to leave inbreeding depression out of your model so that you can compare results with and without
inbreeding depression – thereby allowing you to document what impacts inbreeding depression could
have on population viability.
28 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
Box B: Quantification of Inbreeding Depression
Inbreeding depression is the reduction in fitness commonly observed when individuals are produced
by matings between genetic relatives. Inbreeding depression seems to affect most (perhaps even all)
species of sexually reproducing organisms, and can cause reduction in survival (of infants, juveniles,
and adults), mate acquisition, fertility, fecundity, number of progeny per litter or brood, and a
variety of physiological measures related to fitness such as growth rate, disease resistance, stress
resistance, metabolic efficiency, sensory acuity, and behavioral dominance (see Lacy 1997 and
references therein).
Although inbreeding depression can affect many components of fitness, often the overall effect can
be reasonably well summarized by or combined into an effect on infant survival. For example, if
inbreeding causes a 10% reduction in litter size, and then a 10% reduction in survival of those
individuals born, the cumulative effect would be the same as a 19% reduction in infant survival
(resulting in 81% of the yearlings which would have been produced if no inbreeding had occurred).
Also, most of the published literature on inbreeding depression in wild species of animals deals only
with effects on juvenile survival (Ralls et al. 1988; Lacy et al. 1993). Therefore, the primary way in
which inbreeding depression is incorporated into VORTEX is through a reduction in first-year survival
of inbred individuals. (If desired, inbreeding effects on later survival, reproduction, carrying capacity,
and even dispersal can be modeled using functions of inbreeding to specify demographic rates: see
Chapter 5 for more information.)
While inbreeding depression is widely known (and has been for centuries), understanding the various
possible underlying mechanisms, the ways of quantifying it, and the consequences for population
survival and viability is not at all simple. Inbreeding depression may result from recessive deleterious
alleles (which are exposed more frequently in homozygous inbred individuals), or from a general
disadvantage of homozygotes relative to heterozygotes, or from other genetic mechanisms (see
Charlesworth and Charlesworth 1987; Lacy 1993b). In studies of Drosophila flies, it has been
observed that about half of the effect of inbreeding depression on survival is due to recessive lethal
alleles (Simmons and Crow 1977). The relationship between survival and inbreeding caused by the
presence of recessive lethal alleles is described by an exponential decline:
S = S0 e −bF
in which S0 is the survival of non-inbred individuals, F is the inbreeding coefficient, b is the average
number of lethal alleles per haploid genome (half the number per diploid individual), and S is the
resultant survival rate (Morton et al. 1956). Figure B-1 gives the expected relationship between the
extent of inbreeding and juvenile survival for a series of hypothetical scenarios differing in the total
number of lethal equivalents.
Even if the overall inbreeding depression is due only partly to recessive lethal alleles, the relationship
between inbreeding and survival might be expected to be roughly an exponential decline of this
form. By observing the relationship between survival and inbreeding, the coefficient b in the above
equation can be measured. The value b is a measure of the severity of the effects of inbreeding (not
in terms of how inbred the population is—as that is measured by F—but rather in terms of how much
fitness is depressed for any given level of inbreeding), and it is the number of recessive lethal alleles
per haploid genome that would cause the observed rate of inbreeding depression. This concept is
called the number of “lethal equivalents” in the population. A population with 4.0 lethal equivalents
per diploid individual (b = 2.0) might have 4 lethal alleles per individual, or it might have 8 alleles
per individual which each cause 50% reduction in survival when homozygous, or it might have 2
lethal alleles and four 50% lethals, or any other combination of deleterious alleles which have the
same total effect.
VORTEX uses this concept of lethal equivalents to quantify the severity of depression of first-year
survival due to inbreeding. Thus, the user must specify how many lethal equivalents characterize the
population under study. For only a few species, however, has the number of lethal equivalents been
measured in careful breeding studies. Among those species that have been studied, the number of
lethal equivalents per diploid (2b) ranges from 0 to more than 30, but it is usually in the range of 1
to 5. (Isn’t it depressing to know that you probably carry from 1 to 5 alleles which would be fatal
Chapter 3 29
The Data Input Process
VORTEX Version 9 User’s Manual
Box B (Continued)…
1.0
Juvenile Survival
0.8
Figure B-1. Expected juvenile survival as a function of
inbreeding coefficient under alternative levels of
inbreeding depression severity, defined as the number of
lethal equivalents (LE) per haploid genome (see text for
details).
0.6
0.4
2.0 LE
0.2
4.0 LE
6.0 LE
0.0
0.0
0.1
0.2
0.3
0.4
0.5
Inbreeding Coefficient (F)
genetic defects if you had two copies of any one of those alleles? Aren’t you glad that you are
diploid?) To date, no clear patterns have emerged to suggest that certain taxonomic, ecological, or
other categories of species typically have high or low number of lethal equivalents – it seems to be
largely a matter of chance whether a population is severely affected by inbreeding or not.
How does VORTEX use “lethal equivalents”?
VORTEX simulates inbreeding depression in two ways, because different genetic mechanisms of
inbreeding depression can have different consequences for population viability. Recessive lethal
alleles are rather efficiently removed from a population by natural selection when inbreeding occurs.
As a result, many individuals may die in the early generations of inbreeding, but when they die they
take their lethal alleles with them to the grave, and subsequent generations of individuals have
fewer lethal alleles to cause inbreeding depression. (This process is often referred to as “purging the
genetic load” of lethal alleles. See Hedrick 1994; Ballou 1997; and Lacy and Ballou 1998.) On the
other hand, selection is ineffective at purging inbreeding depression when the inbreeding depression
results from a general advantage of heterozygotes over all homozygotes (or, to a lesser extent,
when it is caused by recessive sub-lethal alleles).
To model the effects of lethal alleles, which can be removed by selection during generations of
inbreeding, VORTEX assigns to each individual at the start of a simulation some unique lethal alleles.
If inbred descendants happen to receive two copies of the same lethal allele, they are killed. To
model the component of inbreeding depression that is not effectively reduced by selection, VORTEX
calculates the inbreeding coefficient of each individual and then applies an exponential equation like
the one above (but using just a part of the total lethal equivalents) to determine how much that
individual’s survival is reduced. To incorporate these two mechanisms of inbreeding depression,
VORTEX needs to know (i.e., you need to tell it) how much of the overall inbreeding depression (lethal
equivalents) to assign to lethal alleles vs. other genetic mechanisms. As mentioned above, for
Drosophila flies, it has been reported that about half of the lethal equivalents are due to actual lethal
alleles. Almost no other species have been studied in sufficient detail to quantify the contributions of
different types of alleles to inbreeding depression, but the scant data available are not inconsistent
with about half of the inbreeding effects being due to lethals in other species as well.
In summary: if you don’t know what to enter for inbreeding depression in VORTEX, use the default
values of 3.14 lethal equivalents (the median of 40 mammalian populations surveyed by Ralls et al.
1988) with 50% of that due to lethal alleles.
30 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
VORTEX runs much more slowly when inbreeding depression is included and is modeled with less
than 100% of the impact due to lethal recessives (see below). This is because the program will
need to calculate and store all pairwise kinships among individuals in the population (N2 kinships,
where N is the maximum population size attained). Therefore, if your population is expected to
remain moderately large (perhaps > 250), so that inbreeding will be a rare event, you may want to
obtain much greater speed by assuming in your model that inbreeding has no impact on fitness.
VORTEX includes a detailed simulation of genetic change in the populations. At the beginning of a
simulation, each founder individual is assigned two unique alleles at each of a number of loci. Each
offspring is then randomly assigned one of the two alleles from each parent at each locus. VORTEX
normally models allele transmission at 5 loci that may contain lethal alleles (allowing up to five unique
and independent lethal recessive alleles per founder) and also at one neutral locus (with no impact on
inbred progeny). In the Special Options on the Project Settings screen, you have the option of asking
VORTEX to model alleles at a greater number of neutral loci. Doing so will produce more precise results
for genetic trends, and the details of the emergent genetic patterns at the modeled loci can optionally be
output to a file for further examination.
In its simplest form, inbreeding depression is modeled in VORTEX as a reduction in the survival of
offspring during the first year of life. As a result, the program generally underestimates the impact of
inbreeding as it can also depress other components of fitness such as adult survival, fecundity, and/or
success in competition for mates. (More complex relationships between inbreeding and demographic rates
can also be modeled; see Chapter 5 for more on this subject.)
Lethal Equivalents: This box and the next ask you to specify the severity and nature of inbreeding
depression in your simulated population. Enter the average impact of inbreeding on first-year survival,
quantified as a number of “lethal equivalents” per diploid individual. As described more fully in Box B,
the default value of 3.14 is a summary statistic based on a survey of 40 captive mammalian populations
(Ralls et al. 1988). If you have specific data indicating a different genetic load, you can enter it here.
Percent Due to Recessive Lethals: Enter here the percent of the total genetic load (quantified by the lethal
equivalents you entered into the previous box) that is due to recessive lethal alleles. The number of lethal
alleles per founder is limited to 10; therefore, the product of the number of lethal equivalents and the
percent of the total genetic load attributable to lethals cannot exceed this number. The lethal alleles are
distributed randomly among 10 autosomal loci; thus, the number of lethals per founder will be distributed
approximately as a Poisson distribution. A plausible value – one that is consistent with data on
Drosophila and a few other species that have been studied well – would be 50%. However, cases have
been reported in which nearly all of the genetic load is due to lethals, while – in other populations –
virtually none of the effects of inbreeding appears consistent with the action of recessive lethal alleles
(Lacy et al. 1996). You may wish to test low and high values to see if it affects your simulations of
population dynamics. (It probably won’t, because it is difficult to maintain a population for long at the
very small population sizes at which effective purging of recessive lethal alleles would occur.)
EV Concordance of Reproduction & Survival: Environmental variation (EV) is the annual variation in
the probabilities of reproduction and survival that arise from random variation in environmental
conditions. (For a more detailed introduction to this topic, refer to Boxes C and D.) EV impacts all
individuals in the population simultaneously. The sources of this environmental variation are outside the
population; examples include weather, predator and prey population densities, and parasite loads. These
factors can affect reproduction and survival independently or simultaneously. Check this box if you think
that good years for reproduction are also good years for survival.
Chapter 3 31
The Data Input Process
VORTEX Version 9 User’s Manual
Box C: A Brief Statistics Primer
Many demographic characteristics among wildlife species (e.g., birth and death rates, litter size, etc.)
fluctuate randomly in magnitude from one year to the next. In order to be able to describe this
variability, and to use VORTEX most effectively, you must have at least some basic knowledge of a
few concepts in statistics.
Population Statistics vs. Sample Statistics
It is important to keep in mind the distinction between the value of a variable or statistic in a
population and the value of the variable across a smaller set of observations sampled from that
population. Usually, we do not know the true value for the entire population, and that is something
we wish to estimate by examining a sample from that population. For some statistics, such as the
mean, the best (and unbiased) estimate of the true population value is simply that statistic
calculated for the observed sample. For some other statistics, such as the variance and standard
deviation, the statistic calculated on the sample is a biased measure of the value for the whole
population, so that correction factors must be applied to get a better estimate of the population
statistic. Below we somewhat loosely follow a common convention of using Greek letters to
symbolize the true (but often unknown) population statistic, Roman letters for sample statistics, and
letters with “hats” (for example, ĥ ) to symbolize estimated values.
Measures of Central Tendency and Variability
When a biologist studies a particular demographic characteristic in a wildlife population over some
period of time, one generally notes an abundance of values clustered near the middle of a range of
annual observations. In the language of statistics, the description of this concentration near the
midpoint is a measure of central tendency. The most common measure of central tendency is the
arithmetic mean or, more simply, the mean. The mean of a sample of observations is calculated as
n
∑ Xi
X=
i =1
n
,
which says that the sample mean equals the sum of all measurements in the sample divided by the
number of measurements in that sample. Another common measure of central tendency is the
median, which is the value at which 50% of the observations fall below and the remainder fall above
that value. For a symmetrical distribution, the median will approximate the mean.
To complete our initial description of these data, we must define a measure of variability in the data.
The most commonly used measure of variability is the variance, usually denoted by s2:
s2 =
∑ ( X i − X )2 .
n −1
The “n-1” in the denominator is a necessary correction factor to ensure that the estimate from the
sample is an unbiased estimate of the variance in the population. (If we measured the variance on
the entire population, we would not need this correction and could simply use n in the denominator.)
From this equation, it is evident that s2 gets larger as the amount of variability about the mean
increases. The standard deviation (s), often abbreviated as SD, is the positive square root of the
variance and is another very common descriptor of variation in a sample of observations. As you
enter data into VORTEX for your species, you will be defining the variation in demographic rates in
terms of standard deviations. We can also describe variability through the use of the coefficient of
variation (here labeled CV), in which sample variability is expressed as a percentage of the mean:
CV =
s
⋅ 100% .
X
Finally, the simplest measure of variability is the range of values observed in the sample.
Unfortunately, the observed range is highly sensitive to the number of observations that are made.
32 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
Box C (Continued)…
The Binomial Distribution
When summarizing a dataset consisting of a number of individual observations, it is useful to present
that summary graphically in the form of a frequency distribution, usually in the form of a bar graph.
Often we are dealing with data on a dichotomous variable—such as alive or dead, breeding or not
breeding, male or female—and we are interested in tallying
the frequency of each possibility in a sample of n
0.4 (a)
p = 0.5
q = 0.5
observations (also known as cases or trials). The probability
0.3
of any given case belonging to one or the other category is
p(X)
denoted p and q, with p + q = 1. The frequency of samples
0.2
consisting of 0, 1, 2, …, or n observations of a specific
category in a sample of n cases is described by a binomial
0.1
distribution. Figure C-1 shows a pair of binomial distributions
0.0
for n = 5.
0.4
(b)
p = 0.4
q = 0.6
0.3
p(X)
0.2
0.1
0.0
0
1
2
3
4
5
Figure C-1. The binomial distribution for sample size
n = 5 and (a) p = q = 0.5; and (b) p = 0.4, q = 0.6. The
bars give the probabilities p(X) of obtaining a particular
number of observations of the first category. For example,
with p = 0.4 and q = 1 – p = 0.6, the probability of obtaining
a sample consisting of one observation (out of a possible 5)
in the first category is 0.259.
X
If n observations are sampled from a population with X belonging to one category and n – X in the
other, then the population parameter p, the proportion of the observations that is in the first
category, can be estimated by the observed proportion within that sample:
X
pˆ =
.
n
If this sampling procedure were to be repeated a number of times, each estimate of p̂ would likely
be different. The variance of all possible values of p̂ would be
σ 2pˆ =
pq
,
n
in which p is the true probability in the population, and q = 1 – p. If p is estimated from the sample,
however, then this variance is biased (underestimated); replacing n by n – 1 results in a better
estimate of the variance of p across samples. The standard deviation of p̂ is therefore estimated by
s pˆ =
pˆ qˆ
.
n −1
The Normal Distribution
In a manner similar to the categorical data
described by a binomial distribution, continuous
variables are generally observed to have an
abundance of values nested around the mean with
f(X)
progressively fewer observations near the
maximum or minimum values. When these
observations are viewed graphically, the resulting
frequency distribution takes on the look of the
familiar “bell-shaped” curve, particularly when the
number of observations (n) becomes large. This
curve is more formally described as a normal
distribution. Figure C-2 shows a typical normal distribution.
µ - 3σ
µ - 2σ
µ−σ
µ
µ+σ
µ + 2σ µ + 3σ
X
Figure C-2. A normal distribution.
Chapter 3 33
The Data Input Process
VORTEX Version 9 User’s Manual
Box C (Continued)…
Note in Figure C-2 the relationship between the population mean (µ) and standard deviation (σ) in a
normal distribution. Using a fairly simple method known as normalizing a distribution, we can calculate
that 68.3% of all observations in a normally-distributed population fall within the range of µ ± σ,
95.5% fall within µ ± 2σ, and 99.7% fall within µ ± 3σ. This kind of information is helpful when
estimating standard deviations in demographic rates caused by environmental variation when only a
range of observations are available (see Box E for additional details).
It is also noteworthy that the binomial distribution becomes quite close to a normal distribution when
the number of observation per sample (n) is large—say, when n > 20. Observe that even when n is as
small as 5, the distributions shown in Figure C-1 look like approximate bell curves. However, one
important distinction between the binomial and normal distributions is the binomial distributions are
always bounded at 0 and n, while normal distributions have “tails” that are infinitely long, but rapidly
diminishing.
For more information on the theory and applications of these and many other concepts relevant to an
understanding of population dynamics and risk projections, see Caughley (1977), Sokal and Rohlf
(1994), and Zar (1996).
Case Study II:
Correlating environmental variation for reproduction and survival
North America’s whooping crane (Grus americana) shows a classic migratory pattern typical of many
bird species. The last remaining substantial population breeds in Alberta’s Wood Buffalo National
Park and spends the winter at Aransas National Wildlife Refuge along the Gulf Coast of Texas.
Because of this movement pattern, the environmental conditions affecting chick production are quite
different from those impacting mortality during the majority of the year (Mirande et al. 1991).
Consequently, we would expect EV affecting these processes to be uncorrelated when constructing a
VORTEX model
EV Correlation Among Populations: You specify here the correlation of EV among populations
(applicable, of course, only when more than one population is modeled). If this value is set to 0.0, then
EV will be completely independent among populations. If this value is set to 1.0, then EV in reproduction
and in survival will be completely synchronized among populations. As a result, good years for
reproduction and / or survival in one population will lead to similarly good years in all other populations.
If this degree of correlation is set to an intermediate value, then EV will be partly correlated among
populations.
Environmental variation in the metapopulation context can be considered to exist at two levels: local
(population-specific) and global (acting across all populations). The total EV, when expressed as a
variance rather than a standard deviation as entered by the user, is simply the sum of the EV existing at
these two levels. The correlation of EV among populations that you enter, then, is simply the proportion
of the total EV (when expressed as a variance) that is global in scope (i.e., common to all populations).
34 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
Box D: The Statistics of Demographic Stochasticity
and Environmental Variability
Now that you have reviewed some of the general definitions of central tendency and variability (see
Box C), as well as some characteristics of the binomial and normal distributions, we can discuss the
statistical nature of demographic and environmental stochasticity.
Demographic stochasticity is the random fluctuation in observed birth rate, death rate, and sex ratio
of a population resulting from stochastic sampling processes, even if the probabilities of birth and
death remain constant over time. In such cases, the annual variation in numbers of individuals that
are born, that die, and that are of a given sex can be specified from statistical theory and would be
expected to follow binomial distributions. Environmental variability is the annual fluctuation in
probabilities of birth and death arising from random fluctuations in the environment (e.g., weather,
abundance of prey or predators, prevalence of nest sites, etc.). Annual fluctuations in the
probabilities of reproduction and mortality are modeled in VORTEX as binomial distributions, while
environmental variation in carrying capacity (see Box F for more on this topic) is modeled as a
normal distribution.
Note that the distinction between demographic stochasticity and environmental variability is a subtle
one (even some professional population biologists have been confused by this!). Demographic
stochasticity is the variation in an observed vital rate due to the sampling variation that is inherent
because each individual (an observation) is an independent and random sample from a population
with a given mean or probability. Hence, it is the variation in sample means ( X ) around a fixed
population mean (µ). Environmental variation, on the other hand, is variation (due to extrinsic
factors that vary over time) in the population mean itself (i.e., µ is different each year).
Putting this information together, we conclude that the variation across years in the frequencies of
births and deaths—both in real populations and our simulated VORTEX populations—will have two
components: the demographic variation resulting from binomial sampling around the mean for each
year, and additional fluctuations due to environmental variability. In actuality, catastrophic events
(to be discussed in more detail later in the User’s Manual) also contribute to the overall observed
variation across many years of data, but they are treated separately from standard annual
environmental variability.
Figure D-1. Left panel: Expected values for a given annual demographic rate, showing binomial sampling
variance arising from demographic stochasticity with a sample size (n) of 100 individuals. Right panel:
Frequency distribution of that same demographic rate based on observed mean and variance of annual
values. Solid and dashed curves show the normal distributions that most closely fit the observed data with
and without the catastrophic “outlier”, while the dotted line shows the normal approximation to the
binomial distribution expected solely from environmental variability (and excluding the outlier). The
difference between the solid and dotted lines gives the variation attributable to demographic stochasticity.
80
Excluding outlier
Including outlier
EV only
10
9
8
60
Frequency (%)
"Outlier"
Year
7
6
5
4
3
40
20
2
1
0
10
20
30
40
50
60
70
Annual Demographic Rate (%)
80
90
100
0
0
10
20
30
40
50
60
70
80
90
100
Annual Demographic Rate (%)
Chapter 3 35
The Data Input Process
VORTEX Version 9 User’s Manual
Box D (Continued)…
The left panel of Figure D-1 shows ten years of expected values of a given demographic rate—say,
juvenile mortality—in a simulated wildlife population. Each “bell-shaped” curve depicts the
probability distribution we would expect from demographic stochasticity acting on that rate in that
year (Actually, these little curves of demographic stochasticity would be binomial, but the normal
distributions are close enough for illustration purposes). For example, the expected rate (µ) in year 1
is 15.2%. However, when the fate of each juvenile in the population is considered, it is possible that
the actual rate may deviate from 15.2% solely from this sampling process. In addition, the expected
mortality changes from year to year due to environmental variation, with each annual curve again
reflecting the sampling variance (demographic stochasticity) expected for that year’s value. Note
that these curves become tighter (the standard deviation resulting from demographic stochasticity
decreases) as the means deviate from near 50%. In addition, notice that the mortality rate in year 7
is particularly high; perhaps a catastrophic event occurred in that year to produce such high
mortality.
With annual rate data in hand, we can actually calculate the relative contributions that demographic
stochasticity (DS) and environmental variability (EV) make to the total observed variance. Consider
the example presented in Figure D-1. The mean mortality rate calculated from these annual data is
0.387 with a standard deviation (combining effects of DS and EV) of 0.148. Note, however, that the
catastrophe shown as the outlier in the dataset was not included in this calculation; if it were, the
mean and standard deviation would change to 0.435 and 0.204, respectively. If we consider the data
with the outlier absent, we can calculate the standard deviation due to EV:
2
2
2
σ EV = σ EV
= σ TOT
− σ DS
2
2
where σ TOT
is the total variance across the data and σ DS
is the mean sampling (binomial) variance
across the individual rates (see Box B for how to calculate a binomial variance). In the example
above, the mean binomial variance turns out to 0.0022. Therefore,
2
= 0.0219 − 0.0022
σ EV = σ EV
= 0.140
which is the variation across years of the mean (peak) values for each curve in the left panel of the
figure. This calculation tells us that the contribution of demographic stochasticity to the total
variance observed in our nine years of mortality data (remember, we removed the outlier from the
analysis) is quite small—the variance attributable to environmental variability is almost 90% of the
total variance in mortality. This is shown graphically in the right panel of Figure D-1. In order for
demographic stochasticity to make a significant contribution to the total observed variance, the
number of individuals sampled for a given rate (n) would have to be quite small—on the order of a
few tens.
The right panel of Figure D-1 also shows the frequency distribution obtained by including the
catastrophe outlier in the calculation of overall mean and variance. The inclusion of this single
observation results in a significantly poorer fit to the data, as the overall distribution of values (the
mean of all values in the left panel) does not look much like a normal distribution. This helps in part
to illustrate why catastrophes—events that are infrequent in occurrence yet severe in population
impact—are treated separately from more typical annual or seasonal fluctuations.
Finally, it is instructive to note that each of the distributions in the right panel of Figure D-1 extend
beyond 0.0 and/or 1.0. As this is biologically implausible, we need to truncate these distributions in
order to allow their proper use in defining probabilities. Partly for this reason, VORTEX usually uses a
binomial distribution (which does not extend beyond 0% and 100%, but which otherwise looks much
like a normal distribution) to represent EV. For ease of calculation, VORTEX sometimes does use a
normal distribution when it is a very close approximation to the binomial, but it then truncates the
normal curve symmetrically about the mean to avoid creating any bias.
36 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
Box D (Continued)…
The above methods are a bit complex. Because DS is usually quite small when the sample sizes (n)
are at all large, a quick, somewhat generous, estimate of EV is simply the total variation in rates
observed across years (treating DS as an insignificant contributor to the observed variation).
Finally, keep in mind that the VORTEX simulation program generates DS automatically as it
determines whether each individual lives, whether it breeds, and what sex it is. Unlike some other
PVA programs, you do not specify that DS should be added into the model, and you cannot exclude
it (from the model or from real life). You do need to specify the magnitude of EV, however, as EV
results from external processes rather than being an intrinsic and inevitable part of all population
dynamics. The size of DS is a consequence of the population size; the size of EV depends on the
constancy of the environment.
Number of Types of Catastrophes: Catastrophes can be thought of as extremes of environmental variation
that strongly impact reproduction and/or survival. Types of catastrophes might include sudden habitat
destruction, floods, forest fire, epidemic disease outbreaks, etc. Catastrophes can be significant threats to
small, isolated populations. For example, disease decimated the last population of black-footed ferrets,
and a hurricane killed half of remaining wild Puerto Rican parrots. It is up to you to determine what types
of catastrophe, if any, may impact your population. Later in the data input process, you will be given the
opportunity to define how each type of catastrophe will impact reproduction and survival.
You may be able to identify historical catastrophes by examining birth and/or death rate data over
several past years for your species of interest. If you find a demographic rate that is significantly
different than that described by normal levels of variation—for example, at least 2 standard
deviations from the mean value—you may use that as evidence of a catastrophic event.
Dispersal Among Populations
If you are modeling a metapopulation, you now need to specify a few parameters that help to define the
system of dispersal of individuals among populations. In a later section you will define the rates of
dispersal of individuals between populations.
Age Range – Youngest and Oldest: In these boxes, enter the youngest and oldest ages of those individuals
that move between populations. If both sexes are capable of moving between populations, and the ages at
which males and females disperse are different, you must decide which age(s) you use for these fields.
This decision will be based largely upon how conservative you want to be about your estimation of
potential risk. For example, if males begin moving among populations at 3 years of age and females at 5
years of age, entering 3 as the youngest age to disperse may underestimate the risk of population decline
and/or extinction since females are allowed to move at an earlier age in the model.
Dispersing Sex(es): Check the appropriate box(es) to specify whether males, females, or both can
disperse from the natal population.
Percent Survival of Dispersers: Often, dispersal among populations occupying discrete areas of suitable
habitat is dangerous. Traversing the matrix of unsuitable habitat between populations may expose an
individual to additional risks of predation or lack of food, and entry into a new population may require
competition with the established residents. Enter here the survival rate (as a percent) of individuals that
are dispersing between populations. The dispersal mortality is imposed separately from other mortality
detailed elsewhere in the program. (More specifically, this dispersal mortality is imposed after annual
mortality. See Appendix IV for a detailed description of VORTEX program flow.) A dispersal survival rate
Chapter 3 37
The Data Input Process
VORTEX Version 9 User’s Manual
of 80% means that there is a 20% chance that an individual will die during the process of moving from
population A to population B.
Dispersal Modifier Function: Dispersal patterns can be very complex and determined by many factors.
VORTEX does not provide a full model of dispersal across complex landscapes, but instead models
movements among discrete populations, with the user specifying the rate of movement between each pair
of populations (as you will do in a later Input section). However, this box provides you with the
opportunity to customize dispersal, in perhaps very complex ways. Any function entered here will be used
as a modifier of the rates to be entered later. For example, you could cause dispersal of males to be twice
as high as the specified rate (and twice as high as for females), by entering “=D*(1+(S=’M’))”. The
parameter ‘D’ in the equation stands for the specified dispersal rate between any two populations. Such
dispersal modifier functions can be used to cause dispersal to be dependent on sex, age, inbreeding,
population density, and many other characteristics of the individuals and populations.
With respect to dispersal or other aspects of population dynamics, the standard VORTEX model
may not match precisely the behavior of your species. Often, the differences between the model
and reality will not cause substantial differences in long-term population dynamics and risk.
(Although definitive confirmation of this assumption may require testing a more complex and
complete model to see if the refinements do matter.) In many cases, VORTEX does provide the
capability to create models that are more complex – sometimes much more complex! – than the
standard VORTEX model. These more complex population models are built by using functions
rather than constants for input values. Using such functions provides considerable flexibility, but
you should use them cautiously if you are not yet fully familiar with the VORTEX model.
When you are finished with entering Species Description parameters, click on the heading for Labels and
State Variables on the left-hand list.
Labels and State Variables
In this Input section, you can enter optional labels for your populations, as well as define parameters that
describe characteristics or “states” of your populations and individuals.
Population Labels and State Variables: VORTEX allows the user to enter a label for each population being
modeled (Figure 21). The labels can be any text, up to 20 characters long. These population labels will be
used as headers for entry of population-specific demographic rates (on subsequent data entry screens) and
as labels in the output files. One population label is entered per line.
When entering a population label, the user may also specify up to two numbers for use as “state
variables” describing some characteristic of the population. These state variables must be numeric values
that are entered on the same line as the population label. State variables may describe characteristics such
as measures of habitat quality or habitat suitability for the population, elevation or some other descriptor
of the habitat, or perhaps an identifying code for the subspecies or local population. This option of
entering population state variables is provided so that demographic rates (such as fecundity, mortality,
and carrying capacity) can be specified to be functions of these state variables. (See Chapter 5 for a
description of the use of functions for demographic rates.) In such functions, the state variables are
symbolized (in order) as B and C. (Symbol A is used for age.) Functions characterizing the demographic
rates for each population could always be entered in earlier versions of VORTEX even without using state
variables, but the use of state variables can allow for a more consistent representation of demographic
rates across populations and easier testing of the effects of varying habitat or population characteristics.
These capabilities will be explained in much greater detail in Chapter 5.
38 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
Figure 21. Labels and State Variables Input section.
Meta-model linkages to other programs: VORTEX provides the capacity to run the population simulation
simultaneously (functionally in parallel) with one or more other models that might describe the dynamics
of parts of an overall system. For example, an epidemiological modeling program, OUTBREAK, can model
the dynamics of an infectious disease in the population. VORTEX and OUTBREAK can be run at the same
time on the same simulated population, with VORTEX simulating demographic and genetic changes and
constantly informing OUTBREAK of the current census of the population, while OUTBREAK models the
changes of disease state (susceptible, latent infected, infectious, recovered) and constantly informs
VORTEX which individuals are in each state. The disease states can then be used to modify reproduction,
survival, dispersal or other demographic rates for individuals in the meta-model. More information about
OUTBREAK will be made available at http://www.vortex9.org/outbreak.html.
At this time, OUTBREAK is the only other model that is provided for linking with Vortex. In the future, we
plan to provide also built-in links to GIS models of landscape change, models of animal movements on
the landscape, and perhaps other models. However, VORTEX already provides the capability for a user to
Chapter 3 39
The Data Input Process
VORTEX Version 9 User’s Manual
create (or otherwise obtain) his or her own model and link it to VORTEX as a dynamic, multi-component
meta-model. This capability of VORTEX to link with other models is still being developed and tested. Full
explanation of the proper use of meta-models is beyond the scope of this manual, and users are strongly
cautioned to not expect to be able to build and use meta-models without the assistance of an expert. Lacy
and Miller (2002) discuss the conceptual need to link PVA models with other kinds of knowledge.
Individual State (IS) Parameters: VORTEX provides the user with the option of creating up to nine
“Individual State” parameters that define characteristics of individuals. These state parameters may
represent any feature of the organism that can be specified or coded by a numeric value. For example,
dominance status might be encoded as Dominant = 1.0; Subdominant = 2.0; and Subordinate = 3.0. Or a
state variable might be used to represent some measure of body condition. Or two state variables might be
used to track the x and y coordinates of each individual’s location on a landscape.
To create one or more individual state variables, check the box, then indicate how many variables you
will be creating. For each variable, you then enter into the table a label, which can be any text that will
help you to remember what parameter you were representing. The VORTEX program, however, will track
the Individual State variables with the labels IS1, IS2, etc., as indicated in the first column of the table.
For each IS variable, you need to enter three functions (or constants) to define: (a) an initialization
function (Init fn) – the starting value for each individual at the beginning of the simulation; (b) a birth
function (Birth fn) – the value for each newborn individual; and (c) a transition function (Transition fn) –
the change in state (if any) each year of the simulation. These functions are entered in the same way as
other functions that can be used to specify demographic rates (see Chapter 5).
Case III:
Using an Individual State Variable to model transmission of mitochondrial DNA
haplotypes
Mitochondrial DNA haplotypes are inherited from the maternal parent. This matrilineal transmission
can be modeled by creating an Individual State Variable (IS1, labeled “mtDNA”). To assign randomly
one of 10 haplotypes (encoded 1 through 10) to the founder individuals, specify a Initialization
Function of “=CEIL(RAND*10)”. The maternal inheritance is defined simply with the birth function of
“=IS1”, because the individual state parameters for use in the birth function are set at birth to be
the values from the maternal parent (pending redefinition in the birth function). In the absence of
mutation, the transition function also would be “=IS1”, to preserve the value for each individual
across years. To model mutation to one of the original haplotypes, with mutation rate 0.0001, the
transition function could be set to “=(RAND< 0.00011)*(CEIL(RAND*10))”. The mutation rate used
in the function has to be elevated by 10% to account for the cases in which mutation would be to
the existing haplotype, leaving the value unchanged.
Once the mtDNA haplotype is defined as an individual state variable, demographic rates can be
specified to be functions of an individual’s haplotype. The final frequencies of haplotype will not be
tallied by VORTEX (because VORTEX doesn’t even know that the Individual State Variable you created
is a categorical variable). However, you can obtain a complete listing of all individuals at the end of
the simulation, including their state variables, by selecting the Special Option (from the Project
Settings screen) to “Produce a file of all living individuals at the end of each iteration.” You can then
analyze those data in whatever spreadsheet or other utility software you prefer.
40 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
Dispersal Rates
This next section of input (Figure 22) is accessible only if you specified (in the Scenario Settings) that
your scenario is to have more than one population. In the grid on this screen, you enter dispersal rates to
specify the probability that a given individual of the appropriate age-sex class will disperse from
population A to population B in a specific year. That is, a rate of 1.00 indicates a 1% probability that an
individual will migrate from population A to population B. Equivalently, if a population consists of 100
one-year old females with a dispersal rate between two populations of 1%, then one of these females
would, on average, be expected to disperse in that direction in any given simulation year.
Dispersal rates need not be symmetric among populations; enter whatever probability you deem
appropriate for each pair of populations. Enter 0 to indicate no exchange of individuals between a pair of
populations. The values on the diagonal of the grid – the percents of individuals that do not disperse each
year – is automatically calculated by the program so that the rows will sum to 100%.
Figure 22. Dipersal Rates input section.
Chapter 3 41
The Data Input Process
VORTEX Version 9 User’s Manual
It is important to remember that while most input data should be in the form of
percentages, a few others are input as proportions. Check the labels and
prompts for clarification of the required data format.
On the Dispersal Rates screen are five commands that can make it easier to enter dispersal rates. “Import
Rate Matrix” allows you import the grid values from a semi-colon delimited text file. This file can be
created in Excel or whatever software you choose. It must contain values for all cells of the grid,
including the labels (although the labels in the file will be ignored and will not over-write what shows on
the screen). The easiest way to see the format of the rate matrix file is to select “Export Rate Matrix”, and
then look at the file that was created. With these commands, you can create a large matrix in a
spreadsheet program, and then import it into VORTEX, and you can export rate matrices for modification or
for re-use in other VORTEX projects. When you have only a few populations, these commands are usually
not worth using. However, if you have, for example, 40 populations, you very well may want to use Excel
to generate your dispersal rate matrix rather than typing in all 1560 pairwise dispersal rates!
The commands to “Make Cells Square” and “Make Cells Original Shape” are simply there to help you
show the grid in a format that is easier for you to view. The fifth command available allows you to apply
a multiplier to each non-diagonal cell in the grid. By entering a value and hitting “Apply Multiplier of ”
you can shift all of the dispersal rates upwards or downwards. This makes it much easier to test a range of
dispersal rates across Scenarios of your Project. For example, you might enter an initial set of rates, and
then apply multipliers of 0, 2, and 4 in order to test no dispersal, and 2x, and 4x increases in dispersal.
VORTEX provides you with significant flexibility in defining dispersal rates for individuals within a
metapopulation. That is, rates may be inversely proportional to distance, directly proportional to habitat
area, or they may be defined through a more complex determining function. However, you have the task
of calculating these rates for each pair of populations—VORTEX does not calculate them for you based on
a set of internal rules. A considerable body of literature exists on the methods for estimating dispersal
rates between populations. Capture-recapture studies can provide some of the best data for this process of
estimation, although experimental difficulties do exist (see Ims and Yoccoz (1997) for more information).
An example of estimating dispersal rates from molecular data is described in Case Study IV.
Case Study IV:
Estimating migration rates from DNA sequence data
The Anacapa Island deer mouse (Peromyscus maniculatus anacapae) is endemic to the Anacapa
islands off the coast of California near Los Angeles. Pergams et al. (1999) conducted a population
viability analysis to develop a comprehensive management plan involving the potential for captive
breeding, reintroduction, or translocation of individuals following eradication of introduced rats.
Nucleotide sequences were obtained from the mitochondrial DNA (mtDNA) cytochrome oxidase
subunit II gene of mice sampled across each of the three Anacapa Islands. Based on the average
amount of nucleotide sequence divergence among mice from the different islands, the authors were
able to directly calculate an estimate of gene flow, Nm, where N is the average size of a pair of islands
and m is the average rate of migration between those islands per generation (see Nei 1982 for
details). For example, Nm between Middle and West Anacapa was calculated to be 7.27 individuals per
generation. Since a generation in Anacapa Island deer mice is only 84 days, and assuming a time
cycle of 21 days for the analysis, the number of individuals migrating between the two islands is
(7.27)(0.25) = 1.8175 per time cycle. The average population size across Middle and West Anacapa
Islands was estimated to be 19,044 mice, so the migration rate is then calculated to be
1.8175/19,044 = 0.0000954. Finally, the authors assumed a symmetrical pattern of migration, so that
the estimate of the rate of migratin from Middle to West Anacapa (equivalent to that from West to
Middle Anacapa) is (0.5)(0.0000954) = 0.0000477.
42 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
Reproductive System
Monogamous, Polygamous, Long-term Monogamy, or Long-term Polygamy: VORTEX models breeding
systems as monogamous vs. polygamous, and short-term vs. long-term. With monogamous breeding,
there must be a male for every breeding female; males may therefore become a limiting factor restricting
breeding. In polygamous models, there only needs to be at least one male for all females to have an
opportunity to breed. However, in a later section (Mate Monopolization) you can specify that only a
subset of males have opportunities to breed. For example, you can create a polygynous system in which
some males control harems of typically 5 females, while the remaining males are excluded from breeding.
If you do not choose a “Long-term” option, then VORTEX will assume that mates are randomly reshuffled
each year and that all available individuals have an equal probability of breeding. If you do specify one of
the “Long-term” models, then once pairs are formed, those pairs will remain together across years of the
simulation until either the male or the female dies or disperses to a different population. Demographically,
it will not matter whether you choose long-term pairings or re-arrangement of pairs each year.
Genetically, there may be a small effect on the rate of loss of genetic diversity from the population.
Figure 23. Reproductive System input section.
Chapter 3 43
The Data Input Process
VORTEX Version 9 User’s Manual
VORTEX does not fully customize the details of mating systems because of the complexities of considering
a wide variety of species and their particular characteristics. More complex breeding systems can
substantially impact genetic variation, but are less likely to seriously alter the demographic performance
of a population. In the future, VORTEX will also allow you to model a species with hermaphroditic
breeding—that is, a species in which each individual is both male and female and can therefore
potentially mate with any other individual, including itself. Presently, this option is not yet enabled.
Age of First Reproduction for Females (and Males): VORTEX defines breeding as the time when the first
offspring are born, not the age of onset of sexual maturity or the age of first conception. The program also
assumes that breeding (and, for that matter, all other events) occurs at discrete intervals—usually years,
but this can be described in terms of whatever you have defined as a suitable time cycle. Thus, breeding
age must be entered as an integer value; you cannot enter 2.5 years as the first age of breeding but must
enter either 2 or 3 years. In addition, you should enter the median age of first breeding, not the earliest
age ever observed since the earliest observed age may not be typical of the normal population behavior.
Case Study V:
Estimating age of first breeding in males and females
The babirusa (Babyrousa babyrussa) is one of the more interesting endemic mammals on the
Indonesian island of Sulawesi. Individuals in captivity can reach sexual maturity as early as 5
months of age. However, most captive animals approach the age of one year before reaching sexual
maturity. Even at this age, the animals are quite small and it is considered unlikely that they will
reproduce until they are older than one year of age. Taking into account the gestation length of
about 5 months (usually 155 – 158 days), it is likely that female babirusa in the wild will have their
first litter at the age of 2 years. On the other hand, males of this age will have to deal with strong
competetion for mates among older and stronger males. Consequently, Manansang et al. (1996)
developed VORTEX models for this species in which the age of first breeding among males was
estimated to be delayed until 4 years.
Maximum Age of Reproduction: VORTEX assumes that individuals can breed at a rate typical for that
species throughout their adult lifespan. If your species does not reproduce throughout its entire adult life,
do not enter the species’ maximum life expectancy. For example, humans typically reach reproductive
senescence before they die. In this case, if the typical life expectancy is 70 years but they cease
reproduction at 50 years, then enter 50 as the maximum breeding age.
Maximum Number of Progeny per Year: Enter the most individuals born to a given female during a year
(time cycle). If your species produces more than one set of offspring (in the form of litters, clutches, pods,
etc.) per year, but you are using a year as your time cycle, add each set together and then enter the total
number born during the year. You can enter the maximum number that has ever been recorded—even
though such an event may be quite rare—and later on during the data input process you can then assign a
low probability of occurrence to this maximum value.
You have the option of entering a mean and standard deviation for the distribution of offspring numbers,
rather than specifying the percentage of females producing each possible number of young (see below).
This removes the limitation on the size of offspring numbers that can be modeled, and therefore makes it
much easier to model species with high fecundities. To choose this option, enter 0 for the maximum
number. Subsequent screens will allow you to specify the nature of this normal distribution (see p. 51).
44 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
When annual offspring numbers per female are not large—for example, on the
order of five or fewer—it is recommended to specify the exact distribution, rather
than using the optional normal distribution.
Case Study VI:
Modeling species with high fecundity
The distribution of the winged mapleleaf mussel (Quadrula fragosa) has been reduced to only a few
sites in the St. Croix River between Minnesota and Wisconsion in the United States. As is typical of
nearly all freshwater bivalves, this species reproduces by broadcasting a large number of larval
offspring known as glochidia into the water column. Kjos et al. (1998) estimated that an adult
female mussel produced a mean number of 171,000 glochidia in a typical breeding cycle. Only those
glochidia that locate and encyst within the gills of its host fish, excyst following metamorphosis, and
then settle onto suitable substrate on the river bottom will survive to the subadult age class.
Because of the impossibility (at least with today’s computer systems) of tracking such a large
number of individual offspring, VORTEX cannot normally deal with such high levels of fecundity.
However, this situation can be made much more tractable by simply redefining what is meant by
“reproduction”; instead of defining it in terms of the production of glochidia, we can define it as the
number of individuals that successfully settle onto suitable river-bottom substrate. Kjos et al. (1998)
estimated that only about 0.2% of the glochidia find and successfully encyst onto a host, that about
15% of those successfully metamorphose and excyst, and about 20% of those excysted juveniles
settle onto suitable substrate. In other words,
Final brood size = [171,000 glochidia]×[0.002 encyst]×[0.15 excyst]×[0.2 settle] = 10
By condensing a series of mortality events from the very early stages of the mussel’s life cycle, the
authors were able to define a system of reproduction that was amenable to the VORTEX modeling
system. Other types of organisms that would benefit from this type of simplification include
amphibians, fish, and even insects.
Sex Ratio at Birth: Enter here a number between 0.0 and 100.0 to represent the average percentage of
newborn offspring that are male. This number is typically very near 50%, signifying a roughly equal
male:female sex ratio at birth. If relatively more or fewer males are born to a given female per year, enter
the appropriate percentage (e.g., 55 for 55% males).
Density Dependent Reproduction: Does the reproductive rate of your species change with changing
population size? That is, is reproduction low when the population is small due to a difficulty in finding
mates or, conversely, does reproduction drop off when the population is large (more specifically, at high
density) due to limited resources or territories, intraspecific competition, crowding, stress, etc.? If so,
check the box and then enter the subsequent parameters defining the nature of the density dependence.
If your population’s reproduction is density dependent, you will need to model this relationship. VORTEX
models density dependence with an equation that specifies the proportion of adult females that reproduce
as a function of the total population size. Normally, the proportion of females breeding would decrease as
the population size becomes large. In addition, it is possible to model an Allee effect—a decrease in the
proportion of females breeding at low densities due, for example, to difficulty in finding mates. The
equation that VORTEX uses to model density dependence is:
(K )
B
P ( N ) = ( P (0) − [( P (0) − P ( K )) N ])
N
N+A
Chapter 3 45
The Data Input Process
VORTEX Version 9 User’s Manual
in which P(N) is the percent of females the breed when the population size is N, P(K) is the percent that
breed when the population is at carrying capacity (K, to be entered later), and P(0) is the percent of
females breeding when the population is close to zero (in the absence of any Allee effect). The exponent
B can be any positive number and determines the shape of the curve relating the percent breeding to
population size, as the population becomes large. If B = 1, the percent breeding changes linearly with
population size. If B = 2, P(N) is a quadratic function of N. Figure 24A shows representative densitydependence curves for different values of B in the absence of an Allee effect.
The term A in the density dependence equation defines the magnitude of the Allee effect. One can think
of A as the population size at which the percent of females breeding falls to 50% of its value in the
absence of the effect (Akçakaya 1997). Figure 24B shows several density-dependence curves for different
values of A with a steep decrease in breeding at high densities (B = 8). Figure 24C gives the same curves
as in Figure 24A, but with the addition of an Allee effect (A = 1).
By inspecting the density-dependence equation, one can see that when the population is at carrying
capacity, P(N) = P(K). When the population is very small (N is near 0), then P(N) = P(0) if there is no
Allee effect. It is also apparent that the Allee effect term [N / (N + A)] will have a strong impact on the
value of P(N) when N is small. When N is much larger than A, the Allee term will have very little effect
on the value of P(N). Fowler (1981) provides a review of density-dependence functions and presents
some density curves for large mammals. We have chosen to model density dependence using the equation
given above because it provides the user with considerable flexibility—despite its relatively simple
% Adult Females Breeding per Year [P(N)]
100
B
A
90
B = 16
80
70
2
A=0
0.5
8
4
1
1
60
A=4
0.5
50
B = 0.25
40
30
20
10
P(0) = 80%; P(K) = 40%; K = 100; B = 8
P(0) = 80%; P(K) = 40%; K = 100; A = 0
100
% Adult Females Breeding per Year [P(N)]
2
10
C
90
B = 16
70
60
1
0.5
B = 0.25
40
30
20
10
P(0) = 80%; P(K) = 40%; K = 100; A = 1
0
40
50
60
70
80
90
Figure 24. Plots of the default density
dependence relationship as used by
VORTEX in the absence of an Allee
effect (A = 0; panel A), in the
presence of a steep decrease in
breeding success at high population
densities (B = 8; panel B), and both a
steep decrease in breeding success
at high population densities and an
Allee effect (A = 1 and B = 8; panel C).
8
4
2
0
30
Population Size (N)
80
50
20
10
20
30
40
50
60
Population Size (N)
46 Chapter 3
The Data Input Process
70
80
90
100
100
VORTEX Version 9 User’s Manual
formulation. Fowler (1981) suggests that density dependence in reproductive success can often be
modeled quite well with a quadratic function, that is, with B = 2.
It is best to derive the values of P(0), P(K), A, and B from a regression analysis of data on the breeding
rate of your population. If these data are unavailable, but you can estimate P(0) and P(K), then you may
want to explore several different combinations of A and B to come up with a curve that seems appropriate
for your population. You could use graphics or statistical software—or even graph paper and a
calculator—to construct a range of hypothetical curves, using different combinations of parameters, as
was done to produce Figure 24. In any event, once you have decided on a particular set of parameters to
use you should always graph the resulting curve to verify that it represents the mode of density-dependent
behavior you are looking for. After you have entered the appropriate parameters as shown on Figure 25
below, VORTEX can display a graph of the specified density dependence function for you so that you can
verify the intended nature of the relationship. Select the population from the drop-down list and then hit
the “View” command to see your graph. (Note: you will need to specify at the bottom of the graphing box
that you want to plot the function against the ‘N’ parameter.)
Depending on the shape of the density dependence curve you have specified, and the mortality
rates you will enter later, your population may never be able to reach the carrying capacity K (also to
be specified later). The combination of density-dependence in both reproduction and survival will
determine over what range of sizes the population is expected to experience average net growth
and over what range it would be expected to decline since deaths outnumber births.
Figure 25. Specifying and graphing density dependence.
Chapter 3 47
The Data Input Process
VORTEX Version 9 User’s Manual
Case Study VII:
Modeling density dependence in reproductive success
Figure 26. Density dependence functions as
modeled in VORTEX for Peary caribou (top panel:
Gunn et al. 1998) and the winged mapleleaf
mussel (bottom panel: Kjos et al. 1998).
90
80
70
60
50
40
30
20
10
Peary Caribou (Rangifer tarandus)
0
% Females Producing Brood
In contrast, the winged mapleleaf mussel (Quadrula fragosa)
inhabits isolated stretches of the St. Croix River in Minnesota
and Wisconsin (United States), with individuals separated
from one another by as much as 20-25 meters. While no
evidence points to a suppression of reproductive success at
higher densities, the mode of reproduction in these mussels
suggests that Allee effects may play a major role in
influencing reproduction as population size (and density)
declines (Kjos et al. 1998). In fact, reproductive success is
thought to drop off rapidly as populations are reduced to
below about 30% of the estimated carrying capacity. To
model this phenomenon , P(0) =, P(K) = 19.4%, and A = 4
(the exponential steepness B is set to zero when
reproductive success is unaffected at high densities). This
relationship is shown in the bottom panel of Figure VI-1.
% Females Producing Calves
Peary caribou (Rangifer tarandus) are distributed in fragmented populations across the Canadian
Arctic. These animals are continually stressed by food resource limitation in the harsh winter
climate, with this stress becoming magnified as population density increases. Consequently, a risk
analysis conducted on this taxon (Gunn et al. 1998) included density-dependent reproductive
success at high population densities (Allee effects were not considered to be a factor). Field data
show 80% of adult females are able to breed under optimal density conditions, while only 10% are
expected to be successful when the population reaches carrying capacity. In other words, P(0) =
80% and P(K) = 10%. To approximate the expected shape of the density dependence curve
modeled in VORTEX, the exponential steepness parameter B was set to 4. The functional form of this
relationship is shown in the top panel of Figure V-1.
16
12
8
4
Winged Mapleleaf Mussel (Quadrula fragosa)
0
0
10
20
30
40
50
60
70
80
90
100
Population Density, (N/K)*100
Reproductive Rates
The next section of input (Figure 27) asks for parameter values that specify reproductive rates. Note that
you decide when in the development of the next generation the “birth” is defined to occur. For mammals,
you would probably use parturition as the point at which offspring are tallied. For oviparous species,
however, you can start to tally offspring at egg-laying, or at hatching, or at fledging, or at any other
developmental stage that makes sense to you and for which you can specify the demographic rate
parameters. For amphibians, you may choose to start each animal’s life in the VORTEX model at
metamorphosis. Whenever you define an individual’s life to begin, you must make sure that the first year
mortality rates you specify in the next input section are appropriate for the choice you made about when
to start recording offspring. For example, if you tally offspring starting at hatching, then the clutch sizes
you specify on this screen will be in terms of the number of hatches, and your first year mortality will be
from hatching through the subsequent 12 months. If you choose to start offspring at fledging, then the
clutch size will be specified in terms of the number of fledglings, and survival will be from fledging
onwards.
48 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
Figure 27. Reproductive Rates input section.
% Adult Females Breeding: Here you specify the mean percentage of adult females that breed in a given
year (or, stated another way, the probability that a given adult female will successfully produce offspring
in a given year). Data on the interbirth interval, or the timespan between successive birth events for a
given female, can be useful for estimating the percentage of adult females breeding annually. A simple
example: if the average length of time between successive births for adult females is 2 years, then 50% of
all adult females are expected to breed in a given year (this assumes, of course, that animals can breed
throughout their normal lifespan). A more detailed example is presented below in Case Study VIII.
EV in % Breeding: Environmental variation (EV) in reproduction is modeled by the user entering a
standard deviation (SD) for the percent females producing litters of offspring (see Box C for a refresher
on some basic concepts in statistics and their calculation). VORTEX then determines the percent breeding
for a given year by sampling from a binomial distribution with the specified mean and standard deviation.
Chapter 3 49
The Data Input Process
VORTEX Version 9 User’s Manual
Case Study VIII:
Using interbirth intervals to estimate annual % adult females breeding
In a risk assessment for the Ugandan population of the eastern subspecies of chimpanzee Pan
troglodytes schweinfurthii, Edroma et al. (1997) used interbirth interval (IBI) data from a number of
well-studied populations in Uganda and neighboring countries. The mean IBI for these populations
was set at 5 years. However, the early death of an infant will cause the IBI for a given female to
shorten dramatically. Data suggest that if an infant dies at 6 months of age, the IBI will be reduced
to about 2 years. This rapid response means that the effect of infant mortality on population growth
may be quite small, so long as the mother is not killed or harmed by the same forces responsible for
the death of her infant. In addition, infant mortality was set at 16%, and the frequency of nonsterile females in the population was estimated at 91%. Given these data, a corrected interbirth
interval can be calculated as:
IBIcorr =
(0.84)(5) + (0.16)(2) = 4.97
0.91
Based on this estimate of the interbirth interval, the percentage of adult females producing an
offspring in a given year is (4.97)-1 = 20.1%.
Box E: A Quick and Easy Way to Estimate a Standard Deviation
from Scant Data
Ideally, to estimate the standard deviation of a demographic rate across years, we would want to
have many years (perhaps 10 or more) of field data. However, we often have information on just a
few years, and often only the best and worst years in recent times. Fortunately, the expected range
observed in a sample of n values from a normal distribution can be specified (see below, and Rohlf
and Sokal 1981). To estimate the standard deviation of a distribution, the observed range (best –
worst years) can be divided by the expected range. For example, across 15 years of observations on
Sonoran pronghorn antelope (Hosack 1998), the mortality rate of fawns was 85% in the worst year
and 55% in the best year. Dividing the observed range (30%) by the expected range for a normal
distribution (3.47 SD units), provides us with an estimate of the SD of 30% / 3.47 = 8.64%.
Table E-1. Mean ranges (in SD units) for
samples of a normal distribution (from Table 26
in Rohlf and Sokal 1981)
Number of observations
Range
2
3
4
5
6
7
8
9
10
15
20
25
30
40
50
1.13
1.69
2.06
2.33
2.53
2.70
2.85
2.97
3.08
3.47
3.74
3.93
4.09
4.32
4.50
Although environmental variation in birth and death rates can have a substantial impact on the viability of
a population, it is often difficult to obtain the data needed to estimate EV. Long-term field studies are
needed in order to determine the amount of fluctuation that occurs in the demographic rates of your
50 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
population. If these data are available, the standard deviation in mean birth rate can be simply calculated
using the methods in Box C. If your dataset is small but you are comfortable with making a rough
quantitative estimate of the variability, you can use the technique presented in Box E.
A common problem in estimating annual fluctuations in demographic rates is that the data might be so
sparse that it is difficult or impossible to estimate the parameters on an annual basis. If this is the case,
you might be forced to admit that the data are not sufficient to allow estimation of the variability around
the mean values. The only alternatives are to guess at the fluctuations in reproductive and mortality rates,
based on a general understanding of the natural history of the species, or to omit environmental variability
from the model altogether (by entering 0 when each standard deviation is requested). In this case, you
must recognize that a potentially important component of population variability and instability is being
ignored in your analysis.
Another difficulty with these approaches, which may add a significant bias if sample sizes are small, is
that some of the year-to-year variation observed in reproductive and mortality rates is actually due to the
expected demographic stochasticity resulting from random sampling of individuals, even if environmental
factors do not cause fluctuations in the annual probabilities of birth and death. Refer to Box D for
methods of removing this source of variation as a means of estimating EV alone.
In order of ease of use (easiest to most difficult) and precision (least precise to most precise),
your options for estimating environmental variation (EV) in population demographic rates are:
guess at the “typical” fluctuations in your species’ reproduction and mortality rates; calculate the
variation across years in these rates from long-term field data; adjust the observed variation by
subtracting the variance due to demographic stochasticity (random sampling), even if the
probabilities of birth and death remain constant through time.
Use Normal distribution approximation/Specify exact distribution: Previously, you defined the maximum
number of offspring produced annually per female; you are must now specify the percentage of
litters/clutches/broods produced by the breeding adult females that are of a given size. You have two
options for specifying the distribution of numbers of progeny. You can use a Normal distribution
approximation or you can fully specify the probabilities of each number of progeny. When you use the
Normal approximation, VORTEX will randomly select a number of progeny for each breeding female by
sampling from a normal distribution with the specified mean and standard deviation, and then choosing
the closest whole number of offspring to the value sampled. The distribution will be symmetrically
truncated, if necessary, in order to prevent the specification of negative litter sizes and to prevent bias in
the sampling of the distribution.
If you have data on the percents of females producing each possible number of progeny, and if the
maximum number produced is not very large, then it is more accurate to enter that exact distribution. You
accomplish this by entering the percentage (i.e., a number between 0.0 and 100.0) for each specified size
up to the maximum. For example, if the maximum litter size is 5 but the average litter is comprised of just
2 individuals, you would enter a much higher percentage of females producing smaller litters (say, 60%
produce a litter of 2 but only 5% produce a litter of 5). The total must add to 100%, and the final value
will be entered automatically by the program in order to sum to 100%.
Copy input values from: When modeling a metapopulation, you often will want to use the same
values for most input parameters across all of the populations. On the left side of the Simulation
Input screens is a tool that allows you to copy input values from any one population to all
subsequent populations. You can copy only those values in the current Section of input, or copy
values in all the input Sections.
Chapter 3 51
The Data Input Process
VORTEX Version 9 User’s Manual
Mortality
In this next section of input (Figure 28), you enter the age-sex specific mortalites. In the language of
matrix life-table analysis (e.g., Caughley 1977; Caswell 2001), VORTEX defines mortality as the mortality
rate qx, or the percentage of animals alive at age x that die before reaching age x + 1. Enter mortality rates
as a percent (between 0 and 100) for each age-sex class. Once reproductive age is reached, the annual
probability of mortality remains constant over the life of the individual and is entered only once (but see
Chapter 5 for further information on how to relax this assumption).
Mortality of Females (Males) as %: In these tables, enter the mean mortality rates for each age class, and
enter also a standard deviation (SD) for each mean to describe the environmental variation (EV) in each
rate. For information on the calculation and statistical treatment of variance in demographic parameters
used in VORTEX, see Boxes C and D.
Figure 28 Mortality Rates input section.
52 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
Be aware that if you enter a standard deviation for each mean mortality rate that is at least half of
the survival rate (100% - mortality rate) – in other words, the coefficient of variation in survival is at
least 50% – in occasional years the mortality rate will be set at 100% and the population will
immediately go extinct. For example, if all age-specific mortality rates are 50% and the standard
deviations are set at 25%, then in about 1 in 40 years the mortality rate after adjustment for EV will
be 100% (since the rate will exceed the mean by 2 standard deviations about 2.5% of the time).
A substantial literature exists on the many methods by which one can estimate age-sex specific mortality
rates in wild populations; Caughley (1977) is a good text from which to start an exploration of this body
of information.
Case Study IX:
Estimating mortality rates from sightings of banded Whooping Cranes
As of 1991, the last remaining population of the whooping crane (Grus americana) could be tracked
and censused from its breeding grounds in Wood Buffalo National Park (Alberta, Canada) to its
wintering grounds along the Texas Gulf Coast at Aransas National Wildlife Refuge. During a
Conservation Viability Assessment for this population, Mirande et al. (1991) estimated mortality
rates for the population based on recorded sightings of banded birds. From 1976 through 1989,
about 234.5 cranes were observed to hatch at Wood Buffalo NP (taking the midpoint of the possible
range in those few years in which counts were imprecise), of which 172 arrived in Aransas the
following winter. This yields an estimated juvenile survival rate of 73.3%. During the 14 years of
close monitoring of the Wood Buffalo population, the observed variance around the mean
survivorship of 0.733 was 0.047. The variance that would be expected from random binomial
sampling based on this mean is 0.013. The difference (V = 0.034, or SD = 0.184) can be attributed
to environmental variation.
Mortality after the first year can similarly be determined from either data on banded birds of known
age, or from winter census reports from Aransas filed since 1938 (young of the year are
distinguishable from older birds upon arrival at Aransas). Since 1938, a total of 2359 birds older
than 1 year of age returned to Aransas, out of a total of 2594 birds that departed Aransas the
previous spring. This yields an estimated annual mortality after the first year of 9.06%. Among the
banded birds, 89.9% annual survival was observed in 386 bird-years, but band loss after several
years could have accounted for some of the “mortality” recorded among banded individuals. No
variation was detectable statistically among mortality rates calculated separately for each age class
beyond the first year.
The observed annual variation in survival rates from 1938 to 1990 was V = 0.00255; the variation
expected due to binomial sampling from a constant probability is V = 0.00220. The difference can be
attributed to environmental variation in the probability of surviving, with V = 0.00035, or SD =
0.019. This value turned out to be very close to the intuitive estimate provided by workshop
participants – that annual fluctuations in mortality rates would be about ± 2%.
Catastrophes
In the next section of input (Figure 29), you enter data to characterize catastrophes that might strike your
populations. Note that you must do this for each type of catastrophe you identified (back in the Species
Description section) across each population comprising your metapopulation (if you are modeling more
than one population). You toggle among the types of catastrophes by clicking on the buttons arrayed
across the top of the Catastrophes window.
Don’t forget to enter data for Catastrophe 2 (if any) and all further catastrophes
after you have completed entering data for Catastrophe 1.
Chapter 3 53
The Data Input Process
VORTEX Version 9 User’s Manual
Figure 29. Catastrophes input section.
Global/Local: Each catastrophe is to be specified as global or local in scope (this is applicable only when
more than one population is modeled). You are given considerable flexibility in how the scope of each
catastrophe is specified, so it is important to read the following information carefully in order to correctly
model your metapopulation.
A global catastrophe will occur in the same years in all populations, but the severity of effects can be
entered as different or equal across populations. Local catastrophes occur independently among the
populations.
To cause a catastrophe to be regional in scope, affecting only a subset of the populations, you can specify
that it is global, but then set the severity factors to 1.0 (see below) for those populations which are not
affected by that regional catastrophe.
54 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
You may also specify that a catastrophe is global for some populations, but local for others. In that case,
the catastrophe happens concurrently across the populations for which it is global, but occurs
independently in those populations for which it is local.
Normally, the frequency of a global catastrophe would be set to be the same in each population affected
by that global catastrophe. However, you can specify different frequencies for a global catastrophe among
the populations. In that case, when the catastrophe hits a population, it will also hit all other populations
in which that catastrophe has at least as high a frequency of occurrence. The catastrophe will sometimes
occur in the populations that have higher frequencies while not occurring in populations with lower
frequencies.
Case Study X:
A “catastrophe sampler”
1. Attwater’s Prairie Chicken
The impact of hurricanes was assessed by Seal (1994) in a Population and Habitat Viability
Assessment for Attwater’s prairie chicken (Tympanuchus cupido attwateri), an endangered bird
that was reduced to just 3 disjunct subpopulations in coastal southeastern Texas. Based on data
from the National Oceanic and Atmospheric Administration, it was assumed that hurricanes strike
this area on average once every 70 years. Species biologists indicated that populations in Refugio
County dropped from 1,200 – 1,500 in the spring prior to Hurricane Beulah to approximately 250
in October following that storm. Therefore, it was assumed that catastrophic hurricanes would
result in 80% mortality of the adult (post-fledging) population. This would translate into a
severity factor with respect to survival of 0.20. Because hurricanes typically occur during late
summer and autumn, it was assumed that such an event would not affect reproductive success
as breeding occurs in the spring.
2. Winged Mapleleaf Mussel
Kjos et al. (1998) identified major upriver chemical spills as the primary catastrophe impacting
the last surviving populations of the winged mapleleaf mussel, Quadrula fragosa. A chemical spill
of this type could occur as a result of, for example, an accident involving a vehicle carrying
hazardous materials. A detailed analysis by the Minnesota Department of Transportation suggests
that the probability of such an event could be quite small (see Kjos et al. (1998) for a detailed
description of the calculation). A very conservative estimate of the probability was set at 0.20%;
i.e., it is thought to occur perhaps, on average, once every 500 years. However, if it were to
occur, it would have major effects on the mussel populations: in the year that such an event
occurs, both reproductive success (proportion of adult females breeding) and survival (spread out
across all age classes) would be reduced by 30%, equivalent to a pair of severity factors equal to
0.70.
3. Mountain Gorilla
The primary catastrophic event modeled by Werikhe et al. (1998) in their evaluation of mountain
gorilla (Gorilla gorilla beringei) viability was the spread of disease from humans to gorillas. As the
extent of human-gorilla interaction increases with rising human population pressures, the
likelihood of passing human diseases to gorillas is thought to be markedly higher. Data from
discussions with primate veterinarians at the PHVA Workshop led to the construction of the
following major disease events:
• Influenza-like disease – 10% annual probability of occurrence; 5% reduction in survivorship;
no effect on reproduction
• Severe, but not pandemic, viral disease – 10% annual probability of occurrence; 25%
reduction in survivorship; 20% reduction in the proportion of adult females breeding
• Hypothetical viral disease with chronic cyclicity, targeting the organ reproductive system – 4%
annual probability of occurrence; 25% reduction in survivorship; 100% reduction in the
proportion of adult females breeding (i.e., no reproduction during a catastrophe year).
Chapter 3 55
The Data Input Process
VORTEX Version 9 User’s Manual
Frequency %: Once the scope of the catastrophe is identified, you need to define the probability that a
given catastrophe will occur in a particular year. Enter this as a percent from 0.0 to 100.0. For example, a
value of 1.0 means that there is a 1% chance that this particular event will occur in any one year. Stated
another way, a catastrophe given a frequency of occurrence of 1.0% means that, in a simulation lasting
100 years, this event is expected to occur one time on average.
Severity (proportion of normal values): For each catastrophe, you need to define the severity with respect
to reproduction (percentage of adult females breeding) and survival. The fecundity and survival rates for
years in which a catastrophe occurs are obtained by multiplying those rates in a “normal”, noncatastrophe year by the specified factor. These severity factors range from 0.0 to 1.0. Entering 0.0
indicates a total loss of reproduction or survival for the population and 1.0 means that the catastrophe,
when it occurs, will have no effect. For example, entering 0.75 for the severity factor with respect to
reproduction means that, if 50% of adult females breed in a normal year, then only (50%)(0.75) = 37.5%
breed in a year with a catastrophe.
Catastrophe severity factors greater than 1.0 can be used in your model. This would result in
“catastrophes” having beneficial effects on reproduction and/or survival.
Mate Monopolization
You are now asked to specify the male breeding characteristics of your population. This information is
important for those species that may have an established social structure and, consequently, exclude some
adult males from the pool of available breeders. The look of this screen (Figure 30) will depend on
whether you specified Monogamous or Polygamous breeding back in the Reproductive System section.
If you previously specified that the breeding system in your population was polygynous, you must specify
how many of the males are breeders. To describe the degree of polygyny, you will need to provide an
value in one of the following three lines of the grid:
a) % Males in Breeding Pool
b) % Males Successfully Siring Offspring (producing at least one offspring in the average breeding
cycle or year)
c) Mean # of Mates/Successful Sire (the mean number of litters sired by successful males in each year)
Whichever one of the three parameters you specify, VORTEX will calculate the other two values based on
the assumption that some males are excluded from the breeding pool and breeding success among males
in the pool of available breeders is described by a Poisson distribution. Only the top number is actually
used in the simulation model; the other two are provided as alternative ways to enter the degree of
polygyny.
If you earlier specified that the species has a monogamous breeding system, you are asked specify only
the percentage of the total adult male population that makes up the pool of available breeders (and each
male can breed with only one female each year). Remember that not all males within this pool may breed
in a given year, depending on the number of adult females that are successful breeders.
56 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
Figure 30. Mate Monopolization input section.
Initial Population Size
In this section of input (Figure 31), you specify the number of individuals at the start of your simulation.
Stable Age Distribution/Specified Age Distribution: You have two options for entering initial population
sizes. If you do not know the precise age-class distribution in your population (as is usually the case), you
can initialize your population according to the stable age distribution (see Box G in Chapter 4 for a
definition of and methods for calculating this distribution). If you select the stable age distribution, you
then enter the total Initial Population Size for each population. VORTEX will assign individuals to each
age-sex class proportionate to the stable age distribution, and will show that distribution in the grids.
If instead you choose to enter a Specified Age Distribution, you will then enter the actual size of each age
class for both males and females. You may need to use scroll bars to view the older age classes in the
grids. As you enter these values, you will notice that the total initial population size changes accordingly.
Chapter 3 57
The Data Input Process
VORTEX Version 9 User’s Manual
Figure 31. Initial Population Size input section.
Carrying Capacity
(see Figure 32)
Carrying Capacity (K): The carrying capacity describes the upper limit for the size of your simulated
population within a given habitat and must be specified by the user in this next section of input (see Box
F for a more in-depth discussion of this parameter). If the population size N exceeds K at the end of a
particular time cycle, additional mortality is imposed across all age and sex classes in order to reduce the
population back to this upper limit. The probability of any animal dying during this truncation process is
set to (N – K / N), so that the expected population size after the additional mortality is K.
SD in K Due to EV: If you think that the habitat carrying capacity varies over time due to environmental
variation (EV), you can enter a standard deviation (SD) here to account for this variability. For example,
the habitat might support different population sizes in different years due to changing conditions such as
rainfall or food resources. The standard deviation should be entered as a number of animals, not as a
percentage of K; for example, if K = 2000 with a standard deviation of 10%, then enter 200.
58 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
Figure 32. Carrying Capacity input section.
Be careful! If you enter a standard deviation for the carrying capacity that is greater than K/3,
then the value for K could drop to zero during your simulation – resulting in an unwanted
extinction event. If you’re not convinced, see Box C for an explanation of why this is so.
Trend in K? VORTEX allows you to simulate changes in the carrying capacity. Such changes may be
positive or negative and result from human activities such as resource utilization or corrective
management strategies, or from intrinsic ecological processes such as forest succession. To include a
trend in carrying capacity, check the box. Then specify the time period during which the trend will occur,
and the annual rate of change in K. The trend will begin in year 1 and continue for the specified duration.
The program will model a liner trend over this time period. Note: more complex patterns of changing K
can be specified by entering a function in the box for Carrying Capacity (see Chapter 5).
Chapter 3 59
The Data Input Process
VORTEX Version 9 User’s Manual
Box F: What Exactly Is Carrying Capacity, Anyway?
“Carrying capacity—rarely in the field of resource management
has a term been so frequently used to the confusion of so many.”
(MacNab 1985)
The definition and use of the concept of carrying capacity is one of the more tricky issues in
population viability analysis (and, for that matter, in much of population ecology). Pick up any
number of textbooks on ecology or wildlife management and you are likely to find that each one
presents a slightly different formal definition of carrying capacity. In fact, some authors (e.g.,
Caughley 1977) choose not to use the term altogether in their presentation of the mathematics of
population growth. In the context of wildlife management, the habitat carrying capacity for a
particular population can be defined as the maximum number of individuals that environment can
sustain over time in the absence of unnatural disturbances, and without inducing harmful trends in
the abundance of the resources required by that population.
We can gain more insight into this concept by considering the familiar (and admittedly simplistic)
logistic equation for population growth:
dN
K − N 
= rN 

dt
 K 
where r is the intrinsic rate of population increase, N is population size, and K is carrying capacity.
Mathematically, K can be thought of as the population size limit at which the rate of growth dN/dt is
equal to zero. Some ecologists define K as a ratio of the total rate of food production in the habitat
(P) to the per capita rate of food consumption necessary for survival (M). Since a population of size
N must consume food resources at a rate of NM just to stay alive, there are P – NM resources
available for the production of new individuals. If NM exceeds P, then resources available for
reproduction become negative and the population must decrease in size.
When N is very small—for example, when a population is re-established in its native habitat—the
potential growth rate is very close to exponential. If the population exceeds its carrying capacity, the
number of individuals will be reduced until N ≤ K. The carrying capacity, then, can be thought of as
representing a stable equilibrium population size. Many population ecologists describe the gradual
approach towards this equilibrium in terms of damped oscillations in population growth.
Empirically, one could estimate the habitat carrying capacity for a given animal species by
calculating the total food supply appropriate for that species that is available in the habitat, and
dividing that value by the rate of that species’ consumption of its available food supply (for a
detailed discussion of this technique, see Hobbs and Swift (1985)). For example, Petit and Pors
(1996) estimated population sizes, flower availability and nectar output for each of three species of
columnar cacti on Curaçao. Carrying capacities for the two species of nectar-feeding bats dependent
on these cacti could then be estimated based on the daily availability of mature flowers and the field
energy requirements of the bats, with additional energy requirements associated with pregnancy and
lactation taken into account.
If detailed data such as these are unavailable, a rough estimate of habitat carrying capacity can be
generated using long-term data on population size. If the size of the population of interest appears
to be relatively constant over the period of observation (and in the absence of significant human
impact), one can fairly safely assume that the population is at or near its carrying capacity. If this
equilibrium is observed in the presence of major human influences, such as a strong hunting
pressure, then historical data could be consulted to determine if this stable size is indeed natural or
purely artificial. One could also calculate K for a given habitat using population density data from
undisturbed habitats elsewhere in the species’ range.
60 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
Harvest
In this section, VORTEX gives you the option of removing individuals during a simulation (Figure 33).
Figure 33. Harvest input section.
Harvest can mimic hunting, culling, research-related removals, removal of young individuals for
translocation programs, etc. Check the Population Harvested? box to request a regular harvest of
individuals.
The harvest can begin and end at any time during the stipulated length of the simulation. Enter the First
Year of Harvest and the Last Year of Harvest. For example, if you wish to begin harvesting in year 10 and
end in year 25, enter 10 and 25 for these two questions, respectively. No harvest will be allowed before
or after the time frame that you have specified.
If you wish to harvest every year within the specified time frame, enter 1 for the Interval Between
Harvests. If you wish to harvest animals every other year, enter 2. As another example, if the first year of
harvest is 10, the final year is 50 and the interval is 10 years, then harvesting will take place in years 10,
20, 30, 40 and 50.
Chapter 3 61
The Data Input Process
VORTEX Version 9 User’s Manual
Optional Threshold for Harvest: You can specify here some criteria that will restrict harvesting to occur
only if the population status meets certain conditions. You enter this as a function (see Chapter 5). For
example, if you enter “=(N/K)>0.8” then harvests will only occur if the ratio of the population size to
the varying capacity is at least 0.8 (and if it is a harvest year as defined above). Leave the grid cell entry
as 1 if you do not want to provide any harvest threshold criteria.
Female (Male) Ages being Harvested: Enter the number of females and/or males that you will harvest at
each time interval defined for each age class through adults. Enter 0 for no individuals to be harvested in
a given age class. VORTEX will conduct the harvest immediately prior to calculating the year’s breeding
pairs, so the youngest individual that can be harvested is one year old. If the program attempts to harvest
individuals from an age class and finds an insufficient number of individuals, the simulation will continue
without the harvest of those individuals determined not to exist. VORTEX will then report at the end of the
simulation that some of the attempted harvests could not be carried out.
Supplementation
You also have the option of adding any number of juvenile or adult, male or female individuals to each
population (Figure 34). This option can simulate supplementation through, for example, a translocation or
releases from a captive breeding program. As with the harvest option, supplemental individuals can be
added at any time and interval within the specified time frame for the simulation. Furthermore, you are
allowed to both harvest and supplement individuals in a fashion independent of one another.
VORTEX assumes that the individuals that are being added to the recipient population are
unrelated to both each other and to any other individual in the recipient population. Consequently,
supplementation is a means of increasing genetic diversity as well as total numbers of individuals
within a population.
If you want to supplement your population(s), check the Population Supplemented? box. You must then
provide the values on the subsequent lines to define the nature of the supplementation.
First (Last) Year of Supplementation: The supplementation can begin and end at any time during the
stipulated length of the simulation. Enter the years in which you wish to begin and end supplementation.
For example, if you wish to begin supplementing in year 20 and end in year 50, enter 20 and 50 for these
two questions, respectively. No supplementation will be allowed before or after the time frame that you
have specified.
Interval Between Supplementations: If you wish to supplement every year within the specified time
frame, enter 1. If you wish to supplement every other year, enter 2.
Optional Criterion for Supplementing: You can specify here some criteria that will restrict
supplementation to occur only if the population status meets certain conditions. You enter this as a
function (see Chapter 5). For example, if you enter “=(N/K)<0.25” then supplements will be added only
if the ratio of the population size to the varying capacity is at less then 0.25 (and if it is a supplementation
year as defined above). Leave the grid cell entry as 1 if you do not want to provide any criterion for
supplementing.
Female (Male) Ages being Supplemented: Enter the number of females and/or males that you will add at
each time interval defined for each age class through adults. Enter 0 for no individuals to be
supplemented in a given age class. These parameters differ slightly from the partameters defining
harvesting in that the last age class listed on the screen corresponds to the first year of adulthood instead
of the aggregate adult stage. This difference results from the fact that while harvesting selects any adult
individual regardless of age, VORTEX must assign a specific age class to each adult that is being added to
62 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
the recipient population. The age of adults added to the population is always equal to the age at which
breeding commences.
Figure 34. Supplementation input section.
Chapter 3 63
The Data Input Process
VORTEX Version 9 User’s Manual
Saving your Input and Running the Simulation
After you have completed entering values for all of the input parameters, it is probably wise to save your
project. (You would probably be disappointed if you spent a long time entering input values, did not save
them, and then the program crashed during the simulation because one number was wrong.) To save your
Project, click on the save icon (the disk) on the toolbar, or select Save or SaveAs from the File menu. If
you save as a file name other than what you originally specified for the Project, VORTEX will prompt you
to find out if you wish to change the Project name to match the file name. (It makes sense to keep the file
name and the project name the same, although you do not have to do so.)
To run your simulation, click on the Run icon (the green triangle) on the toolbar or select Run Simulation
from the VORTEX menu. The Run Simulation box (Figure 35) will then appear. Check the boxes to
indicate which scenario(s) you wish to run and then hit the Run! command.
Figure 35. Run Simulation selection window.
During the simulation, a graph of the changing population size will be displayed (Figure 36). If the
simulation runs slowly (as it will if you have a very large population size) you can hit buttons to Stop,
Pause, or Clear Lines. Stopping will terminate the simulation without completing all of the iterations.
Pausing will temporarily stop the simulation (as you might want to do while describing the model to
others), and then will Resume when you hit that button (which appears after you hit Pause). Clear Lines
just gives you a way to erase all prior lines if the screen is getting cluttered and unreadable.
VORTEX uses a very fast method to write all of the lines to the screen. (If they are slow, that is
because the simulation itself, rather than the graphing, is taking a long time.) Consequently, the
program may not refresh the display if you drag the graph window or toggle out to another Windows
application and back. Therefore, it is best to not try to do anything else on your computer while the
simulation is running. (This also helps by leaving all memory and system resources available to
VORTEX.) If you feel that you must do other work (e.g., using Word or Excel or check email) while the
VORTEX simulation is working, you should specify the Special Option (under Project Settings) of “Do
not show graphs during iterations.”
64 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
When the requested simulations are complete, VORTEX displays a few summary statistics at the top of the
graph. When you are done viewing the graph, close it by clicking on the ‘x’ icon in the upper right corner.
If you want to print or save this graph of the simulation, use the Windows PrntScrn button to copy the
image to your Windows clipboard. You can then paste it into Word, or PowerPoint, or any other program
that can display images. (However, data for tabulating and graphing summary results have already been
saved on your disk – as we will see in the next Chapter, so there is usually no need to save these rather
cluttered displays of all iterations.)
Figure 36. VORTEX Simulation population size graph.
Chapter 3 65
The Data Input Process
VORTEX Version 9 User’s Manual
Adding and Deleting Scenarios
Congratulations! You have created and run your first scenario. However, it is highly unlikely that you are
now done with your analysis of the particular species, population, and issue at hand. Almost certainly,
you were uncertain about some of the input parameters you entered. Possibly the estimate was based on
few (or no!) data. It may be that the data came from field studies conducted in a different part of the
species range. And it may be that the data are accurate descriptions of the past or even present status of
the population, but are not likely to be good descriptors of the future performance of the population, in
light of expected or planned changes to the habitat or populations. In any case, a major part of almost all
population viability analyses is “sensitivity testing” – the examination of the impacts of varied input
parameters on the projected population performance. You should refer to the background material in
Appendix I and in the references for a more thorough discussion of the topic of sensitivity testing.
To conduct sensitivity testing, or any investigation of alternative scenarios that may be used to describe
the population, you need to create, run, and analyze scenarios that vary from your initial case for one or a
few input parameters. One way to do this would be to start from the beginning and create a new analysis
in VORTEX. However, the program makes it very easy for you to copy all input from a prior Scenario into
one or more new Scenarios in your Project. In these newly copied Scenarios, you can then change the few
input parameters that you want to vary, and re-run the simulations for each case.
Adding Scenarios to Your Project
To add a scenario, from any Simulation Input window, click on the “Add Scenario” command in the
upper left. The pop-up window shown in Figure 37 will appear.
Figure 37. New Scenario window for adding Scenarios based on a prior Scenario.
Click on the existing Scenario that you want to use as a template for new Scenarios, then select below the
number of copies you wish to create of this Scenario, and then click on “OK”. You may notice that there
66 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
is a disabled option for “Sensitivity Analysis”. This option has not yet been implemented, but it will (in a
later version) provide you with a more automated way to create new Scenarios that systematically vary in
one or more parameters.
The new Scenarios you create will initially be named as “Scenario1 – Copy 1”, Scenario1 – Copy 2”, etc.,
in which “Scenario1” is the name of whatever Scenario you chose to use as the template. After you create
new Scenarios, you can toggle among the Scenarios in any of three ways: (1) you can use the drop-down
list to select the Scenario you wish to work on, (2) you can click on the small ‘<’ and ‘>’ buttons
alongside the dropdown list to move backwards and forwards through the list of Scenarios, and (3) you
can click on the buttons to the right with the Scenario names. (See Figure 38.)
Figure 38. Multiple means for moving among alternative Scenarios.
The first task you should complete after creating new Scenarios is to change their names to something
more descriptive. You do this in the top input box (Scenario Name) of the Scenario Settings screen. After
Chapter 3 67
The Data Input Process
VORTEX Version 9 User’s Manual
changing the Scenario name to something more descriptive, go to the Input sections where you want to
vary the parameters in the new Scenarios, and make the desired changes.
Once you have many Scenarios in your Project, it is easy to accidentally change the labels and input
values for the wrong Scenario, because the screens for the various Scenarios look so similar.
Before you start editing input values, make sure that you are working on the intended Scenario.
After you create a new set of Scenarios, it is always wise to save the Project so that you won’t lose your
Scenarios if something later goes wrong. (However, VORTEX always prompts you to save each open
Project when you exit the program, so you do have a chance later to save everything – if nothing does go
wrong.)
Deleting Scenarios from Your Project
If you have created Scenarios that you no longer want, you can delete them by hitting the “Delete
Scenario” command at the top of the Input window. This will delete the current Scenario. A prompt will
first warn you that you cannot recover from this action (other than by re-creating the Scenario again).
Reordering scenarios
After you add and delete Scenarios from your Project, you may find that the Scenarios are not listed in the
order you would like them to have. VORTEX provides a Scenario Manager (Figure 39), which you access
by clicking on the “re-order” command next to the dropdown list of Scenario names. The Scenario
Manager lets you change the order of the Scenarios in your Project (moving them up or down in the
indexed list). A feature not currently implemented will allow you also to specify that some Scenarios are
to be considered to be grouped as “sub-scenarios” of others (moving them to higher or lower levels of a
Scenario tree structure). The level at which a Scenario is placed, and the Scenario under which lower
level Scenarios are nested, has little or no meaning other than to the Sensitivity Analysis utility that has
not yet been implemented.
Figure 39. Scenario Manager window.
68 Chapter 3
The Data Input Process
VORTEX Version 9 User’s Manual
Chapter
4
Viewing Model Results:
Text, Tabular, and
Graphical Output
Once you have entered all of the necessary input parameters and given the
command to run your first VORTEX model, you are ready to look at your simulation results. A display of
changing population sizes was displayed (perhaps very quickly) on your screen as the simulation was
running (Figure 36). While you may want to capture that image onto your Windows clipboard with a
PrntScrn system command, the more useful presentations of model results are generally those that provide
mean results across iterations, perhaps with standard deviations to indicate the variation among iterations
or standard errors to indicate the precision with which the means were estimated from the finite number
of iterations run. In addition, there are many measures of population performance and viability other than
just the projected population size. This chapter describes how to use VORTEX to view text, tabular, and
graphical summaries of your results.
Text Output
Click on the Text Output tab to access simple text descriptions of your Project input or results (Figure
40). There are four subsections of Text Output.
Input Summary
The first section provides a summary of the input values you entered for each Scenario. It is wise to scroll
through this Input Summary in order to be sure that you entered the values as you intended. In addition,
you can cut and paste from this text summary into any reports that you need to create of your work. At the
top of the Input Summary, and similarly at the top of the other three sections of Text Output, are three
command buttons. These allow you to send the text to the Project Report, print the text, or save the text as
a file that you would specify. The Project Report will be described in detail later – it is a simple word
processor that allows you to start building a report of your analyses while you are working within the
VORTEX program. (You can then access that report from MS Word or other software to further edit and
refine it.) If you chose to print the text, VORTEX will open the standard Windows Print utility. If you save
the text, VORTEX saves it as a simple text file, which could later be viewed, edited, or printed from
NotePad, MS Word, or other word processing software.
Chapter 4 69
Viewing Model Results
VORTEX Version 9 User’s Manual
Figure 40. Input Summary section of Text Output.
At the top of the text display window are dropdown lists that let you move among the reports for the
various Scenarios and Populations (within Scenarios). In the Input Summary, all Populations in each
Scenario are contained within the same text file, so you could find them by scrolling down. However, the
file can be very large, so it is often faster to use the dropdown list to jump to the place where text on a
Population starts.
Deterministic Calculations
The second section of Text Output provides both text and a simple graph to display the deterministic
projections of population size (Figure 41). The text window shows the exponential rate of increase, r, the
annual rate of change, λ, and the per generation rate of change or “net replacement rate”, R0, as
determined from life table analysis of the mean rates of reproduction and survival in your model. The
mean generation time and a stable age distribution (calculated from age-specific birth and death rates) are
70 Chapter 4
Viewing Model Results
VORTEX Version 9 User’s Manual
Figure 41. Deterministic Calculations shown in Text Output.
also given. See Box G for a brief description of these deterministic calculations.
The graph given with the Deterministic Calculations is fairly simple and crude, but it shows the
exponential growth (or decline) projected from the life table calculations (up to the limit set by the
carrying capacity). The graph can be sent to your Project Report, printed, or exported to a bitmap (.bmp)
file for import into other programs.
The deterministic calculations are performed by VORTEX as soon as you enter the input values for your
model. Therefore, you can view these results (and also the Input Summary text) even before you run your
simulations. It is often helpful (and always recommended) to look at the deterministic projections before
proceeding with stochastic simulations, so that you will know whether the rates of reproduction and
survival are at least minimally adequate to allow for population growth in the absence of random
fluctations and other destabilizing processes (such as inbreeding and harvest).
Chapter 4 71
Viewing Model Results
VORTEX Version 9 User’s Manual
Box G: Deterministic Calculations in VORTEX
Before the actual stochastic simulation begins, VORTEX performs a standard life table analysis to
calculate the deterministic mean population growth rate (r, the exponential growth rate; or λ
(“lambda”), the annual multiplicative growth rate), the mean generation time for males and females,
and the stable age distribution used (optionally) to initialize the starting population. These
calculations will provide accurate long-term averages, if stochastic variation (due to demographic
stochasticity, environmental variation, catastrophes, and inbreeding effects) is minimal. Life table
analyses implicitly assume that age-specific birth and death rates are constant through time; they
yield over-estimates of long-term population growth if there is any variation in demographic rates.
The deterministic population growth rate r is calculated by solving the Euler equation:
∑ (l x mx e −rx ) = 1 ,
in which lx and mx are the age-specific mortality and fecundity rates, respectively for age class x to
x+1, and the summation is over all age classes. Lambda is related to r by:
λ = er .
The stable age distribution, or the proportion of the population at each age class cx, is given by:
cx =
l x e −rx
∑ (lxe−rx )
.
The determination of a stable age distribution for males in a given population becomes a bit more
complicated if the mortality schedules are different for the two sexes. In this case, r is calculated
based on female life history tables (since females control population growth), but the lx’s used in the
age distribution equation are those for males. Moreover, the exact form of the equation is dependent
on when the age classes were censused. In the above equation, c0 is the proportion of the
population aged 0 plus a small increment, just after the breeding season but before mortality is
imposed. In order to build the initial population, VORTEX omits age class 0 (because the simulations
begin just before the breeding season), and re-scales the age distribution in order to sum to 1.0.
The life table calculations assume that there is no limitation of mates (i.e., there are always enough
breeding males to mate with all breeding females). Other complications arise if there are
catastrophes in the simulation model. In those cases, VORTEX adjusts the fecundity and mortality
rates to account for the effect of catastrophes, averaged across years.
For more information on the details of life table analysis, refer to any number of general ecology or
population biology texts such as Pielou (1977), Krebs (1994), or Caughley (1977).
It is important to look at the deterministic projections of population growth for any analysis. If
r is negative, the population is in deterministic decline (the number of deaths outpace the
number of births), and will become extinct even in the absence of any stochastic
fluctuations. The difference between the deterministic population growth rate and the growth
rate resulting from the simulation can give an indication of the importance of stochastic
factors as threats to population persistence.
72 Chapter 4
Viewing Model Results
VORTEX Version 9 User’s Manual
Output Summary
The third section of Text Output lists the basic status of each population at each year of the simulations.
Figure 42. Output Summary section of Text Output.
The statistics reported in this file are:
The cumulative number of iterations (populations) that have become extinct or remain extant;
The probability of population extinction (PE) or survival (equivalent to the proportion of
iterations (populations) that have become extinct or remain extant);
¾ The mean population size, reported separately for all populations (N-all) and only for those
remaining extant (N-extant), with standard error (SE) and standard deviation (SD) across
iterations;
¾ The mean stochastic growth rate (r), both with and without harvest or supplementation (if
applicable) as well as across all years of the simulation;
¾
¾
Chapter 4 73
Viewing Model Results
VORTEX Version 9 User’s Manual
The mean “expected heterozygosity” (or “gene diversity”) remaining in the extant populations,
with standard error and standard deviation across iterations;
¾ The mean “observed heterozygosity” (equal to 1 – [mean inbreeding coefficient]) remaining in
the extant populations, with standard error and standard deviation across iterations;
¾ The mean number of alleles remaining (AllelN) within extant populations (from an original
number equal to twice the number of founder individuals), with standard error and standard
deviation;
and, if the inbreeding depression option is included in the simulation,
¾ The number of lethal alleles remaining per diploid individual, with standard error and standard
deviation, determined by the nature and extent of the genetic load identified in the input process,
and the intensity of inbreeding the population undergoes.
¾
Following these yearly reports, the output file presents a series of final summary information that
includes:
¾
¾
¾
¾
¾
¾
¾
¾
The final probability of population extinction and, the converse, the probability of population
persistence;
If at least 50% of the iterations went extinct, the median time to extinction;
Of those iterations that suffer extinctions, the mean time to first population extinction, with SE
and SD across iterations;
The mean times to re-colonization and re-extinction of those simulations that went extinct;
The mean final population size, with SE and SD across iterations, for all populations, including
those that went extinct (e.g., had a final size of 0);
The mean final population size for those iterations that do not become extinct, with SE and SD
across iterations;
The final age-sex composition of the extant populations; and
The mean population growth rate, with SE and SD across iterations. When harvesting or
supplementation are included in your model, VORTEX will report the mean population growth rate
for years without harvest or supplementation, for years with harvest or supplementation, and
averaged across all years.
Additional summary information will be provided when you have built a metapopulation model. For
example, the output file will also include the same set of summary data for the global metapopulation, and
will also present a set of within-population means. These means are unweighted averages across
populations, while the standard deviations are means of the individual standard deviations of the
populations. Population sizes and genetic metrics are averaged across only those populations that survived
some simulations. Similarly, times to extinction, recolonization, and re-extinction are averaged across
only those populations that had some extinctions. Finally, if any recolonization events occurred during the
simulation, VORTEX will report the frequency of recolonization, the mean time to recolonization, and the
frequency and mean time to population re-extinction if appropriate.
Also given for metapopulations will be some tables of genetic distances (and identities), with standard
errors, to show the amount of genetic differentiation that existed between populations at the end of the
simulations.
Other Output
The fourth section of Text Output provides two summary tables, in grid formats (Figure 43). One table
provides a line of basic summary statistics for each Population of each Scenario that has been run. The
summary statistics tabulated are the number of iterations (#Runs), the deterministic growth rate (det-r),
74 Chapter 4
Viewing Model Results
VORTEX Version 9 User’s Manual
the mean stochastic growth rate (stoc-r) experienced in the simulations, the SD of the stochastic
population growth [SD(r)], and final values (at the end of the simulation) for many of the descriptive
statistics listed above in the Output Summary. If a Scenario has been run several times, then the table will
show the results from these multiple sets of simulations. Comparing results across repeated sets of
simulations can indicate whether the number of iterations was large enough to give results that are
sufficiently stable (precise) for your purposes. (You can also get an indication of this by looking at the
reported standard errors.)
Figure 43. Scenario Summaries table in Other Output.
The other table in Other Output provides Iteration Summaries – a tabulation for each iteration of the year
of extinction (if extinction occurs), final population size, and, if extinction does not occur, the final gene
diversity, mean inbreeding coefficient, and number of founder alleles remaining. The data in this Iteration
Summary table can be used to analyze the full distribution (not just the mean and SD) of the times to
extinction, final population sizes, and genetic statistics. The “Export” button for the Other Output tables
Chapter 4 75
Viewing Model Results
VORTEX Version 9 User’s Manual
will export these tables as text files delimited with semi-colons. These files can be imported directly into
Excel or other spreadsheet software for further analysis (the column headers should transfer over).
Graphs and Tables
Click on the Graphs and Tables tab to access a section that lets you build tables and graphs displaying
results for a variety of summary statistics (Figure 44). In the Data Specification tab of Graphs and Tables,
you specify what Scenarios, Populations, Years, and Variables you want to put into the table that will be
displayed at the right. VORTEX then also creates a graph of the data in your table, and gives you access to
tools to customize your graph.
The Data Specification area can be confusing, but it gives you considerable flexibility in what you put
into your tables and graphs. To understand how the Data Specification works, you should first realize that
the results from your analyses can be considered to be a series of data points in a three-dimensional space.
Figure 44. Data Specification and Table screen of the Graphs and Tables for output.
The three dimensions of the data are: Years (the year of the simulation at which the data point was taken),
Populations/Scenarios (which Population from which Scenario is being observed), and Variables (which
summary variable is being tabulated). These three dimensions need to be assigned to the columns, rows,
76 Chapter 4
Viewing Model Results
VORTEX Version 9 User’s Manual
and cell values of the table you create. For example, a common table would be to display some Variable
(such as the mean population size) as the cell contents, in a grid with a series of Years (such as 0, 10, 20,
…) as the column headings and Populations (such as Population1, Population2, and Metapopulation) as
the row headings. (See Figure 44 for an example of a table specification similar to this.)
However, you could instead want the Years to be arrayed down the side of the table as the rows, with the
Populations across the top as the columns. To do this, you would select Populations from the dropdown
list for Columns, and Years from the dropdown list for Rows. You could instead tabulate multiple
Variables (such as PE, N, SD(N), and MedianTE) as columns with Populations as rows, or Variables as
rows with Populations as columns. In the specification of the remaining dimension (Years) you then give
from which Year of the simulation you want the value of these Variables displayed. (Usually, but not
always, you would select the final year.) Figure 45 shows an example of a table of this type.
Figure 45. An example of a Table with Variables as Columns.
VORTEX can display a multi-part table, with results from a different Scenario shown in each part. You
select which Scenario(s) you want to examine by checking the boxes in the grid on the lower left of the
screen. To then specify which Years, Populations, and Variable(s) you want in your table, you need to
specify the Columns and Rows in the grid in the lower left, and pick from the dropdown list which
Variable/Year/Population is to be tabulated. The specification of Columns and Rows can be done by
Chapter 4 77
Viewing Model Results
VORTEX Version 9 User’s Manual
typing the numbers of the desired Years or Populations into the grid under Columns or Rows, as
appropriate.
Alternatively, you can click on the “…” button for Rows or Columns. That will open up a window for
specification of the Years, Populations, or Variables, as appropriate (Figure 46).
Figure 46. Year Specification window for creating tables and graphs.
Specification of the Years in this window can be confusing, but it is also very fast once you learn how to
use it. You can quickly “Select All” years by clicking on the command, if you want every year shown in
your table and graph. (You can also “Deselect All” if you want to remove a prior selection.) To specify
that only some years should be selected, you can check on those years in the grid. The boxes in the grid
are arrayed across rows for each decade. That is, the check boxes in the first row of the grid are years 0, 1,
2, …, 9; the check boxes in the second row of the grid are years 10, 11, 12, …, 19; etc. If you want to
select all of the years in one column (such as years 0, 10, 20, …), you can click on the column heading.
Similarly, to select all of the years in one decade (e.g., 20, 21, 22, …, 29), you can click on the row label.
The order in which you select the years is shown in the list that accumulates on the right, and this would
be the order that the years show up in your data table. Presumably, you would want the years to be in
ascending order in your table. To re-order the years, click on the “Sort Ascending” command. After you
have selected (and possibly sorted) the years you want in your table, click “OK” to send that selection to
the Data Specification screen.
Selection of Populations or Variables is done in a similar way, except that they are more straightforward
to select because they are displayed in a single vertical list of check boxes.
After you have completed specification of your data table, you can click on commands to print the table,
send it to your Project Report, or export it to a text file delimited with semi-colons.
Data Graphs
If you click on the Data Graphs tab, VORTEX will display a graph of the data that are in your table (Figure
47). The dimension used for table columns is plotted as the x-axis; the dimension used for the rows of the
78 Chapter 4
Viewing Model Results
VORTEX Version 9 User’s Manual
table will create separate lines on the graph; the values displayed in the table will be the y-axis of the
graph. To remember how the graph axes relate to the table lay-out, think of each row of the table as being
plotted as a line on the graph, with the columns (place in the row) being the dependent (x-axis) variable
and the value given in the table being the dependent (y-axis) variable.
Figure 47. Data Graphs window.
The initial size of the Data Graphs window that opens up will often force your graph to be squeezed into a
fairly small window, and often the legend will not be displayed (because it doesn’t fit in the small
window). To see a better image of the graph, click on the corner of the Project window and drag it out to a
larger size.
On the Data Graphs window, you are provided with several Graph Options for changing the look of your
graphs. You can change the text or position of the legend, the text or font of the main title or the axis
titles, and the line thickness. You can also toggle between a line graph and a bar graph, although for many
kinds of data that would be plotted only one of these two graphs would be reasonable. Command buttons
are provided to allow you to Print the graph, send it to the Project Report, or Export the graph to a bitmap
(.bmp) file. You can also add error bars to show standard errors (SE) around the plotted means, or
standard deviations (SD) across iterations. The SE and SD options are available only for those plotted
variables for which they would be applicable. (E.g., there is no error around deterministic growth rates.)
Chapter 4 79
Viewing Model Results
VORTEX Version 9 User’s Manual
While these basic Graph Options will be adequate for many purposes, VORTEX also provides you with
access to a much more extensive graphing utility. To fine-tune your graphs, right-click on the graph and
select Properties. That will bring up an Advanced Graph Properties window that gives you considerable
control over almost all aspects of the graph (Figure 48).
Figure 48. Advanced Graph Properties.
It is always good to watch the VORTEX website for upgrades! The Tables and Graphs section of
VORTEX is still undergoing considerable improvements, as are many other sections!
80 Chapter 4
Viewing Model Results
VORTEX Version 9 User’s Manual
Project Report
The Project Report tab takes you to a simple word processor utility that you can (and should) use to
document your work and begin to develop a report for sharing with others (Figure 49).
Figure 49. The Project Report.
In the earlier sections for Project Settings, Text Output, and Graphs and Tables, you had the option to
send text, tables, and graphs to your Project Report. This is where you were sending them! It is always a
good idea to liberally send information to your Project Report whenever you think that it may be
information that you will want to capture for inclusion in project reports of any sort. It is always easy later
to edit or delete sections of your Project Report, but it may be difficult to later resurrect information that
you neglected to send to the report.
Chapter 4 81
Viewing Model Results
VORTEX Version 9 User’s Manual
The Project Report gives you tools for standard format changes, including fonts, colors, italics, bold,
alignment, and bullets. You also have the option to Save, Open, or Print reports. VORTEX automatically
saves your Report with the Project, and re-loads it when you open the Project again. You only need to
explicitly Save a Report if you wish to save it under a new filename. When you save a report, it is saved
in Rich Text Format (an .rtf file). Such files can be imported directly into MS Word and most other word
processors, where you can use their more powerful editing capabilities to further refine your report.
Access to Other Stored Output
The results made available to you in the Text Output and Tables and Graphs sections of VORTEX are all
stored in text files (with extensions .inp, .det. .out, .dat, .sum, .run, and .rtf) placed into your Project
directory. You can access these files directly if you wish to view the data within other word processing,
spreadsheet, database, or graphical software. However, if you wish to edit these files in any way, you
should first make a copy of the files and then edit only the copy. If you change the files that were created
by VORTEX, you may make the data inaccessible when you re-open your Project. Another file that you
should not edit is the file with extension .vpj. This “vortex project” file is the master file that stores the
data for your Project. Any changes made to this file outside of the VORTEX program can cause your
Project to be corrupted and possibly un-openable.
In addition to the results made available to you within VORTEX, the program stores other results in
additional files that are placed into your Project directory. Some of these additional files contain more
detailed results (often, far more detailed than most users would care to examine). These files typically are
text files with semi-colon delimiters, formatted so that they can be opened directly into Excel and many
other spreadsheet and database programs. Which additional files are created depends on the settings you
select in the Special Options of the Project Settings. Available files include (in which “project” is
replaced by the name of your Project, and “scenario” is replaced by the name of the Scenario):
project_scenario.run
A listing for each iteration of the year the population went extinct (if it
did go extinct), and (if it did not go extinct) the final population size,
gene diversity, mean inbreeding, and number of founder alleles.
project_scenario.ani
A listing for each iteration of the animals living at the end of the
iteration, including their sex, age, inbreeding coefficient, and genotypes
at the modeled loci.
project_scenario.gen
A listing of final allele frequencies and probabilities of allele persistence,
averaged across the iterations, for the first locus modeled.
project_scenario_Iter#.gp
A listing in GenePop format of the individuals alive at the end of the
iteration, including their sex, age, inbreeding coefficient, and genotypes
at the modeled loci.
82 Chapter 4
Viewing Model Results
VORTEX Version 9 User’s Manual
Chapter
5
Using Functions
in VORTEX
Introduction
VORTEX provides the option of modeling demographic rates as functions of population or individual
parameters. The population descriptors that can be used as variables in the functions include time (year in
the simulation), iteration, population, population size, carrying capacity, numbers of juveniles (animals in
the first age class), subadults (greater than 1 year, but not yet breeding age), adult females, adult males, all
females, or all males, and gene diversity (expected heterozygosity). Individual characteristics that can be
entered as variables in these functions include ID#, sex, age, number of mates (0 or 1 for females and
monogamous males; possibly more for polygamous males), inbreeding coefficient, and genotypes at
modeled loci. Almost all demographic rate parameters – such as the percent of females breeding each
year, environmental variation in breeding, litter/clutch size, sex ratio, mortality rates, environmental
variation in mortality, catastrophe frequency and severities, carrying capacity, dispersal, dispersal
mortality, occurrence of harvest and supplementation, and definition of extinction – can be specified to be
functions of the above population and individual variables.
The flexibility to specify population rates as functions rather than as fixed constants has been added to
VORTEX so that users can model specific population dynamics that might be known to be appropriate for
some species, or that are of interest in a theoretical analysis. With some creativity and perhaps
considerable effort, VORTEX can now model many of the kinds of population dynamics that can be
envisioned. As just a few examples:
¾ it might be known that carrying capacity will change at some determined date in the future;
¾ it might be believed that reproductive rates will change over time, perhaps due to some
management action;
¾ the density dependence observed in reproduction might not fit the shapes of the curves allowed in
previous versions of VORTEX;
¾ mortality rates might change over time, or respond in a complex way to population density;
¾ inbreeding might impact fecundity, adult survival, or might affect the two sexes differently;
¾ dispersal might be age and sex dependent;
¾ fecundity, mortality, or the effects of catastrophes might be age-dependent;
¾ environmental variation might occur with a periodicity that is longer than a year, or catastrophes
might have multi-year effects.
Note that VORTEX includes within the basic data entry screens the option to model reproduction as a
density dependent function, and an option to model carrying capacity as having a linear change over a
Chapter 5 83
Using Functions in VORTEX
VORTEX Version 9 User’s Manual
specified number of years. Easy access to these two particular functions are provided because they are
needed more frequently than are detailed functional dependencies of most other rates. Even for these two
rates, however, you can specify these functions to have almost any shape if you use the function editor to
specify the rates.
For most users and for most purposes there will be no need or desire to model demographic rates as
functions; it is usually fully adequate to specify fixed demographic rates rather than functions.
Specification of rates as functions can be difficult: the appropriate form of the function is rarely known,
the function parameters are usually very difficult to estimate, and it is not trivial to enter a function
correctly. If alternative functions need to be examined in sensitivity testing, the number of combinations
of input parameters to be explored can quickly become overwhelming. Consequently, we would not
recommend that novice users or students use the function option within VORTEX.
Specification of Demographic Rates as Functions
Dependencies of demographic rates on population and individual parameters are entered into VORTEX by
specifying the functional relationships. There are two ways that you can enter a function rather than a
constant for an input variable: you can type the function directly into the input box for specifying the rate,
or you can open a Function Editor to help you develop the function to describe the relationship. VORTEX
provides an option (and this option is the default for new Projects) to have the Function Editor open
automatically whenever you enter a “=” as the first character in an input box for a demographic rate. The
other way to open the Function Editor is to click on the Function Editor icon on the toolbar. If you type a
function directly into an input box, you must precede the function with an “=” sign (to distinguish the
specification of a rate as a function rather than as a constant). If you open the Function Editor by typing
an “=” in an input box, and then develop a function within the Function Editor, then the Function Editor
will insert the function back into the active input box when you accept the function. It is usually easier
(and safer) to build a function first within the Function Editor and then send it over to the input screen.
Figure 50. The Function Editor utility.
84 Chapter 5
Using Functions in VORTEX
VORTEX Version 9 User’s Manual
When you open the Function Editor (Figure 50), you can enter the desired function by typing it into the
box at the top of the window, or you can use the screen keypads to select numbers, operators, and
functions, which will then be pasted into the function box. It is probably fastest to just type the function
you want, but the keypads can help remind you what operators, trigonometric functions, Boolean
operators, and other functions are available for use. The population and individual variables available for
use in functions are listed in a box in the lower part of the Function Editor window (see Table 1). The
Function Editor also keeps a list of recently used functions, and they can be recalled (and further edited, if
desired) by clicking of the drop-down list in the Function box.
In specifying a function, you should make no assumptions about the order of precedence of operators (see
Table 2 for list of valid operators). For example, the function A+B*C is interpreted as (A+B)*C.
(Precedence is left to right, so VORTEX interprets A*B+C in the standard way, but it is risky to assume that
VORTEX will read a function the way you would read it.) Always use parentheses to specify the order in
which operations are to be performed. (Parentheses), [brackets], and {braces} may be used
interchangeably to indicate the order of operations.
Functions cannot contain any spaces or extraneous punctuation. Case of function names and variables is
ignored. All letters that are entered in a function within the Function Editor are converted to upper case by
VORTEX.
It can be difficult to correctly specify the function that describes the relationship you want for a
demographic rate. To help you confirm that you have specified the correct function, VORTEX can display a
simple x-y Function Preview plot of any function entered into the Function Editor. From the drop-down
list above the graph, you must select which dependent variable from your function should be plotted
along the x-axis. The plot will show the relationship to the selected x-variable, with each other dependent
variable fixed at some simple value (e.g., sex = female, K = 100, N = 100, numbers of males, females,
juveniles, and subadults = 50, iteration = 1, year = 1, population = 1, gene diversity = 100%, inbreeding =
0). You can change the range and increment of the x-axis if you wish. The Function Preview graph will
not display until you click the “Update Graph” command.
If you know how to read reverse Polish notation for functions, you can click the command to “Show
Polish Notation” in order to confirm that VORTEX is interpreting the parentheses to produce the order of
operations that you desired.
After you have confirmed that you have the function that you want, you can cut and paste it into an input
box, or (if you triggered the Function Editor by typing an ‘=’ into an input box) just hit ‘OK’ to transfer
your function over to the input box you had left.
Chapter 5 85
Using Functions in VORTEX
VORTEX Version 9 User’s Manual
Table 1. Valid Function Variables
Population descriptors
N = Population size
K = Carrying capacity
Y = Year
P = Population identifier
R = Run (simulation iteration)
M = Number of adult males in the population
F = Number adult females
J = Number of juveniles (age 0-1)
U = Number of subadults (age > 1, < breeding age)
X = Number of females (all ages)
W = Number of males (all ages)
G = Percentage of initial gene diversity (expected heterozygosity) remaining in the population
B and C = population state variables (entered with population labels)
D = Dispersal rate (from the matrix entered), used only in functions modifying dispersal.
BB(p), CC(p), etc. = the parameter above (B, C, etc.) for population p.
Individual descriptors
A = Age
O = Individual ID (an arbitrary integer assigned to each individual)
S = Sex (0 or ‘F’ for female, 1 or ‘M’ for male)
I = Inbreeding coefficient (must be expressed as a percentage)
Q = Number of mates (0 or 1 for females and monogamous males, possibly more for
polygamous males)
V = Paternal allele identifier
Z = Maternal allele identifier
VV(i) and ZZ(i) = the paternal and maternal alleles at the ith modeled locus.
IS1, IS2, IS3, … , IS9 = Individual State Variables previously defined.
86 Chapter 5
Using Functions in VORTEX
VORTEX Version 9 User’s Manual
Table 2. Valid Operators (Note that there are alternative names for several operators.)
Function
Description
Example
Unary Operators
ABS
NEG
CEIL
FLOOR
ROUND
SQRT, SQR
LN, LOG
LOG10
EXP
Absolute value
Negative
Ceiling
Truncate
Round
Square root
Natural logarithm
Base 10 logarithm
e raised to specified power
ABS(-10) = 10
NEG(-10) = 10
CEIL(3.12) = 4
FLOOR(3.12) = 3
ROUND(3.12) = 3 ; ROUND(5.5) = 6
SQR(1.44) = 1.2
LN(1.60) = 0.47
LOG10(1.60) = 0.20412
EXP(0.47) = 1.60
Binary Operators
+
*
/
POW, ^
MAX
MIN
MOD, %
Addition
Subtraction
Multiplication
Division
Exponentiation
Maximum
Minimum
Modulus (Division remainder)
1.0+2.0 = 3.0
2.0–1.0 = 1.0
2.0*3.0 = 6.0
6.0/2.0 = 3.0
POW(10,0.20412) = 10^0.20412 = 1.60
MAX(3.12,4.21) = 4.21
MIN(3.12,4.21) = 3.12
MOD(33,10) = 33%10 = 3; MOD(33.5,5) = 3.5
Logical (Boolean) Operators
==, =
Is equal to
NOT, !
Negation
!=, #
Not equal to
AND, &&
And
OR, ||
Or
>
Greater than
<
Less than
>=
Greater than or equal to
<=
Less than or equal to
(3=2) = 0 = FALSE
!(3=4) = 1 = TRUE
(3#4) = 1
((3=4)AND(3#4)) = 0
((3=4)OR(3#4)) = 1
(3>4) = 0
(3<4) = 1
(3>=3) = 1
(3<=3) = 1
Trigonometric Operators
SIN
Sine
COS
Cosine
TAN
Tangent
ASIN
Arcsine
ACOS
Arccosine
ATAN
Arctangent
SIN(PI/2)
COS(PI/2)
TAN(PI/4)
ASIN(1.0)
ACOS(0.0)
ATAN(1.0)
Defined Constants
PI
3.1415927
E
2.7182818
TRUE
1
FALSE
0
SIN(PI/4) = 0.7071067
LN(E) = 1.0
(10>5) = TRUE
!TRUE = FALSE
=
=
=
=
=
=
1.0
0.0
1.0
1.5707963
1.5707963
0.7853981
Random Number Generators
RAND
Uniform random (0 – 1)
RAND = 0.2341 or 0.8714 or 0.9151 or …
NRAND
Normal random deviate
NRAND = 0.512 or –0.716 or –2.376 or …
SRAND
A “seeded” random number generator; hence, SRAND(x) provides a random number between
0 and 1 with a given seed value x
SNRAND
A “seeded” random normal deviate; hence, SNRAND(x) returns a number from a (0,1) normal
distribution with the seed value x
Chapter 5 87
Using Functions in VORTEX
VORTEX Version 9 User’s Manual
Using Random Numbers in Functions
Random number generators can be used to create a wide variety of stochastic events (for example, a 5-year drought that occurs on average once every 30 years), but the proper use of these functions requires
careful consideration of how the “seed” values (implicit, as in RAND and NRAND, or explicit, as in SRAND
and SNRAND) determine when new random numbers are selected. Repeated calls to the random number
return the same value if the same seed is specified. Random numbers produced with different, even
sequential, seeds will not be correlated. The “unseeded” forms (RAND and NRAND) set their own unique
(or nearly so) seed each time they are called. The very first use of a random number generator in VORTEX
uses a seed based on the number of seconds elapsed since the turn of the century. Each call to an
unseeded random number generator also sets a new seed for the next call for an unseeded random
number. Thus, two identically configured computers starting the same simulation at exactly the same
second on their clocks would produce identical results for an analysis. This synchrony may require,
however, that all memory storage locations (including hard disk caches) and even the hard disk contents
are identical on the two systems (because they will affect the time required for each read or write to the
disk).
The specification of random number seeds allows synchronization of sequences of random numbers. This
can be used to create synchrony of events, such as catastrophes or environmental variation across
populations, or autocorrelations among years (time lags or cycles). If several different demographic rates
are specified by functions containing random number generators (perhaps to trigger separate catastrophes
impacting survival and fecundity), care must be taken to create the desired synchrony or lack of
synchrony. If two functions contain the same seed values, they will return the same random number. Seed
values must be distinct to create independence of random numbers. (See examples below.)
Proper use of random number seeds can be difficult. Think carefully about the effect of any seed that you
use in a function, to be certain that it will produce the same random numbers when you want them, and
independent random numbers otherwise. Any variable (e.g., A for age, Y for year, R for run, P for
population) included within the seed will cause the same “random” number to be chosen for each case
with the same value for those variables (A, Y, R, P). For example, if you specify SRAND(P) within a
function, then each population will get an independent random number, and that set of random numbers
will be the same over all calls to evaluate that function (such as for every year, every run, and every
individual within each population). If you specify SRAND((P*100)+Y), then each population will get a
new independent random number each year of the simulation, but the set of random numbers will be the
same across all runs of the simulation. You would normally want to include the variable R in the random
number seeds (e.g., SRAND((R*10000)+(P*100)+Y)), in order to cause the random numbers to be
independent among runs of the simulation. See the examples below for further information about random
number seeds.
The seeds used by VORTEX will be converted to integers between 0 and 65536. Non-integer seeds will be
truncated [hence, SRAND (35.23) = SRAND(35.89)] and values above 64K will be “wrapped” [the
modulus taken, so that SRAND(65636) = SRAND(100)].
Notes Regarding Function Syntax and Use
•
•
Variables of trigonometric functions are assumed to be in radians.
The operator NEG is the same as using a minus (-) sign before a number. By the context, VORTEX
will interpret whether a minus sign signifies subtraction (a binary operation) or the negative (a
unary operation).
88 Chapter 5
Using Functions in VORTEX
VORTEX Version 9 User’s Manual
•
CEIL, FLOOR, and ROUND convert real numbers to integer values, but all expressions are
evaluated as real numbers. For example,
FLOOR(3.7)/FLOOR(4.1) = CEIL(2.1)/CEIL(3.7) = ROUND(3.1)/ROUND(3.6) = 0.75.
•
•
•
•
•
•
Numbers may be written with or without leading and trailing zeroes. Decimal points for integral
values are optional. For example, all of the following are valid expressions: 3, 3.00, 0.03,
.03, -0.30, -5.
Functions containing invalid mathematical expressions are prohibited, such as:
SQR(-10)
Square root of a negative number
LN(-10)
Natural log of a negative number or zero
5/0
Division by zero
TAN(1.5707963) Tangent of PI / 2
ASIN(1.1)
Arcsine or arccosine of a value greater than 1 or less than –1
Some mathematically valid functions would be ambiguous or meaningless. For example, functions
of carrying capacity (K) should not contain K as an independent predictor (of itself). Functions of K
should also not include A (age) or S (sex) as parameters, because the condition of exceeding the
carrying capacity is a population-level phenomenon, and K is assessed once for each population
each year. If K is a function of inbreeding (I), the value of I applied in the function will be the
mean for the population.
The total length of a function cannot exceed 512 characters. Functions cannot contain more than 24
numerical constants, or more than a total of 128 constants plus variables plus operators. (Often you
can find an alternative form that is shorter.)
Many of the variables that can be used in rate functions will themselves change during each year of
the simulation. In order to avoid unresolvable interdependencies of parameters and rates, the
population size (N) and sizes of subsets (J, F, M, U, X, W) and gene diversity (G) used in function
evaluations are the numbers that were tallied at the last (pre-breeding season) census.
Even without specifying rates as functions, many of the rates used in VORTEX can be specified to be
different for different years, sexes, ages, or inbreeding levels. (e.g., age-specific mortality,
inbreeding depression in juvenile mortality, linear trends in K, etc.) Be aware that the effects of any
functions entered are imposed on top of such dependencies that might be given in the standard input
format. For example, EV in carrying capacity could be specified via standard input, or via a
function of the type
K = 100+(10*NRAND),
thereby giving an annual level of EV in K equal to a standard deviation of 10. The advantage of
creating EV by specification within functions (rather than more simply as a parameter given to
VORTEX) is that you have greater control over how EV is implemented in the model. For example, it
is possible to specify that EV is concordant between two populations (but not with others):
RATE = 50+[10*SNRAND(Y+(R*100)+[(P>2)*100*SRAND(P)])].
In this function, the overall mean demographic rate is 50 with annual fluctuations due to EV (SD)
equal to 10. Populations 3 and greater will experience independent annual fluctuations, while
populations 1 and 2 will fluctuate synchronously. The use of year (Y) in the seed for the random
number causes a new random number to be used each year. The use of R (iteration or run) in the
seed causes the sequence of seed values to be different in each simulation. The inclusion of
(P>2)*100*SRAND(P) within the seed causes a different sequence of random numbers to be
chosen for each population after the first two have been evaluated. The seeds must include
Y+(R*100) to ensure that every year-iteration is independent. (If you use Y+R as the seed, then
year 3 of iteration 1 will have the same value as year 2 of iteration 2, etc.) In the simpler example of
EV in K given above, no seed was needed or specified, so an independent random number will be
selected each iteration, each year, and each population. These examples show how elaborate and
Chapter 5 89
Using Functions in VORTEX
VORTEX Version 9 User’s Manual
non-intuitive the functions can become when you want to create even moderately complex models
of population dynamics.
Using Functions to Examine Genetic Evolution
The parameters available for use in functions defining demographic rates include an individual’s
paternally inherited allele (V) and the maternally inherited allele (Z) of the (normally) non-selected locus,
which is monitored for tracking genetic diversity. (The symbols for these variables V and Z have no
intuitive meaning, but are rather the result of few letters remaining available for denoting additional
parameters for functions.) By specifying that demographic rates are functions of the alleles carried by an
individual, it is possible to model a wide variety of genetic processes impacting population dynamics,
including: the effect and fate of alleles that confer alternative life history strategies (e.g., lower fecundity
but higher survival); balancing, disruptive, or directional selection for alleles impacting demography;
hybrid vigor or outbreeding depression caused by introgression of alleles from a distinct taxon or
geographic population; and genetically based individual variation in demographic rates.
When the initial population is created, and when the population is supplemented with any new
individuals, the founders are assigned unique alleles sequentially. Hence the first individual of the
Population 1 is assigned alleles 1 and 2, the second individual is assigned alleles 3 and 4, and so on. Final
frequencies of all founder alleles in each population, averaged across all iterations, can be outputed to a
file if that option is selected within Special Options of Project Settings. The allele frequencies are placed
into a file with extension .gen.
Examples of Rate Functions
The easiest way to demonstrate the formats in which functions can be entered into VORTEX is with a series
of examples. The examples shown below include a plot of the function where appropriate as well as the
actual expression.
1. Continuous linear decline over time
RATE = 50–(0.2*Y)
50.0
45.0
Demographic Rate
This function specifies a starting rate
(perhaps for adult female breeding success or
for carrying capacity) equal to 50 in year 0,
with a decline of 0.2 per year resulting in a
rate equal to 30 after 100 years.
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
60
70
Year of Simulation
90 Chapter 5
Using Functions in VORTEX
80
90
100
VORTEX Version 9 User’s Manual
2. Linear decline limited to a period of years
RATE = 50–(0.2*MIN((Y-1),50))
50.0
45.0
Demographic Rate
In this case, the decline occurs only through
the first 50 years of the simulation. Note that
the decline is specified to start in year 2, so
that year 1 still has a rate of 50. This is the
form of the function used by VORTEX if the
user specifies a linear trend in carrying
capacity.
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
60
70
80
90
100
80
90
100
80
90
100
Year of Simulation
3. Linear decrease during intervals of years
RATE = 50–(5*(MIN(5,Y)+((Y-25)*(Y>25))))
50.0
45.0
Demographic Rate
The rate starts at 45 in year 1, declines to 25
by year 5, and again resumes the decline at a
rate of 5 per year after year 25.
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
60
70
Year of Simulation
4. Exponential decline
RATE = 50*(0.98^Y)
The rate declines by 2% each year.
50.0
Demographic Rate
45.0
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
60
70
Year of Simulation
Chapter 5 91
Using Functions in VORTEX
VORTEX Version 9 User’s Manual
5. Exponential decline with inbreeding
RATE = 50*EXP(-2.0*(I/100))
50.0
45.0
Demographic Rate
The rate declines from 50 in non-inbred
animals down to 6.7 in fully inbred animals (I
= 100%). Note that the inbreeding coefficient
(I) is expressed as a percent. An equation like
this might be used to specify a decline in
fecundity due to inbreeding.
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
60
70
80
90
100
70
80
90
100
70
80
90
100
Inbreeding Coefficient (%)
6. Age-dependent fecundity, with linear decline after the onset of breeding
RATE = (A>=5)*(50-((A-5)*2))
Breeding begins with a rate of 50 at age 5, but
then declines by 2 each year thereafter.
50.0
Demographic Rate
45.0
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
60
Age (Years)
7. Age-dependent fecundity, with a symmetrical peak at age 15
RATE = (A>=5)*(50-(ABS(A-15)*2))
50.0
Demographic Rate
45.0
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
60
Age (Years)
92 Chapter 5
Using Functions in VORTEX
VORTEX Version 9 User’s Manual
8. Age-dependent fecundity, with an asymmetrical peak at age 10
RATE = (A>=5)*[{[A<10]*([A-5]*10)}+([A>9]*[60-A])]
50.0
45.0
Demographic Rate
Different trends are specified for age intervals
0-4, 5-9, and 10+. Note the use of
parentheses, brackets, and braces to improve
readability.
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
60
70
80
90
100
70
80
90
100
80
90
100
Age (Years)
9. Increase in mortality with inbreeding
RATE = 100-(50*EXP(-1.57*(I/100)))
90.0
80.0
Demographic Rate
The survival rate declines exponentially
(described by the portion within the
outermost parentheses), while the percent
mortality is set at 100-survival. This is the
equation used by VORTEX to model increased
juvenile mortality if there are 3.14 lethal
equivalents and no recessive lethal alleles
contributing to inbreeding depression.
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
0
10
20
30
40
50
60
Inbreeding Coefficient (%)
10. Stepwise increase
RATE = 10+((Y>10)*10)+((Y>20)*10)+((Y>30)*10)
RATE = 50+(10*(MIN(3,FLOOR(
(Y-1)/10))))
90.0
80.0
Demographic Rate
The rate increases from 50 to 80 at 10-year
intervals. An alternative way to express the
same function would be
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
0
10
20
30
40
50
60
70
Year of Simulation
Chapter 5 93
Using Functions in VORTEX
VORTEX Version 9 User’s Manual
11. Different rates at different intervals
RATE = (10*(A<3))+(25*(A=3))+(30*((A=4)OR(A=5)))+(35*((A>5)AND(A<10)))+
(20*((A>=10)AND(A<15)))
50.0
45.0
Demographic Rate
The rate increases stepwise with age, then
drops to a lower level for years 10 through
14, and then drops to zero for animals 15
years and older. Note that although it can be
tedious, any rate function can be modeled by
specifying the rate for each interval of the
dependent variable.
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
60
70
80
90
100
70
80
90
100
80
90
100
Age (Years)
12. Cyclical response
RATE = 50+(10*SIN((PI*Y)/5))
60.0
50.0
Demographic Rate
Here the rate fluctuates between 40 and 60
according to a sine wave with a 10-year
periodicity.
40.0
30.0
20.0
10.0
0.0
0
10
20
30
40
50
60
Year of Simulation
13. Regular pulses of a higher rate
RATE = Y%8=0*20+30
50.0
45.0
Demographic Rate
Background rate of 30 jumps to 50 every 8th
year. Note that the order of operators was left
to the VORTEX default (left to right). This is
not a safe practice, but it does work in this
case.
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
60
70
Year of Simulation
94 Chapter 5
Using Functions in VORTEX
VORTEX Version 9 User’s Manual
14. Pulses of longer duration
RATE = (((Y%8)<3)*20)+30
50.0
45.0
Demographic Rate
The rate jumps to 50 for a 3-year time span
every 8th year. In this case, parentheses were
used (wisely) to be sure that the intended
order of operators was followed.
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
60
70
80
90
100
80
90
100
Year of Simulation
15. Random pulses: catastrophes
RATE = 50-(20*(SRAND(Y+(R*100))<0.05))
50.0
45.0
Demographic Rate
The background rate of 50 drops to 30 on
average once every 20 years. A seeded
random number is needed; otherwise the
years in which the rate drops would be
independent among individuals (effectively,
the rate would continuously be 49). The seed
of Y+(R*100) causes a different seed to be
used for each year of each iteration (if there
are 100 or fewer years). The above function is
equivalent to specifying a catastrophe with
frequency = 5% and severity = 0.60.
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
60
70
Year of Simulation
16. Random pulses independent among populations
RATE = 50-(20*[(SRAND[Y+(R*100)+(100*SRAND(P))])<0.05])
The “catastrophes” are independent among populations, because each population (P) sets a new (and
random) seed for the random number generator which tests whether the catastrophe occurs. Careful
use of parentheses (or brackets) is critical in this function, in order to ensure that the random number
seeds work as intended.
17. Catastrophes affecting only selected age class(es)
RATE = 50-((A<3)*20*(SRAND(Y+(R*100))<0.05))
The catastrophe affects only individuals of ages 1 and 2. The 2-D graphs of this function do not
illustrate the age-dependent relationship: The graphs against Y and R set A = 1 (and show the
catastrophes affecting young individuals); while the graph against A happens to display a year and
iteration (Y = R =1) in which no catastrophe for any age occurs.
Chapter 5 95
Using Functions in VORTEX
VORTEX Version 9 User’s Manual
18. Multi-year catastrophes
RATE = 50-(20*(SRAND((Y/2)+(R*100))<0.10))
90.0
80.0
Demographic Rate
The catastrophes have a 2-year impact,
because the seed value is converted to an
integer, giving pairs of subsequent years the
same random number. The frequency per year
is 10%, so that the frequency of an onset of a
2-year catastrophe is 5%.
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
0
10
20
30
40
50
60
70
80
90
100
Year of Simulation
19. Multi-year catastrophes with a decreased impact in year 2
RATE = 50-([10*(SRAND(Y+(R*100))<0.05)]+[8*(SRAND((Y-1)+(R*100))<0.05)])
50.0
45.0
Demographic Rate
The second seed is the same as the first seed
from the previous year. Thus the catastrophe
has a lesser impact (severity = 0.84 rather
than 0.80) in the second year. This approach
can also be used to model catastrophes which
impact survival in one year (using a function
with an expression like that in the first
brackets above) and fecundity in the second
year (using a function containing the
expression in the second set of brackets).
Note that in example #18, catastrophes
always start in even-numbered years, while in
this example catastrophes can begin in any
year.
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
60
70
80
90
100
80
90
100
Year of Simulation
20. Random variation across years
RATE = 50+[10*(SNRAND(Y+(R*100)))]
This is the same as imposing a mean rate of
50 with environmental variation of SD = 10.
80.0
Demographic Rate
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
0
10
20
30
40
50
60
70
Year of Simulation
96 Chapter 5
Using Functions in VORTEX
VORTEX Version 9 User’s Manual
21. Linear density dependence
RATE = 50*((K-N)/K)
The rate declines from 50, at N = 0, to 0 when
N = K.
50.0
Demographic Rate
45.0
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
60
70
80
90
100
70
80
90
100
Population Size
22. Density dependence used as the default for breeding in VORTEX
RATE = [50-(20*[(N/K)^4])]*[N/(1+N)]
50.0
45.0
Demographic Rate
The rate peaks near 50 when N is small,
declines at higher densities, and is 30 when N
= K. At very small N, the rate is also
depressed. For example, it is reduced by 50%
when N = 1, and reduced by 25% when N = 3.
In terms of the coefficients that can be
entered into the optional density-dependence
for breeding in VORTEX, P(0) = 50, P(K) =
30, B = 4, and A = 1.
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
60
Population Size
23. Sex-specific dispersal rates
RATE = D*[(S='M')OR(RAND>0.35)]
If the above function is used for the Dispersal Modifier Function, then 35% of the females are
prevented from dispersing. Thus, dispersal rates for females are effectively reduced by 35% relative
to male dispersal. An unseeded random number is used so that dispersal will be determined
independently each female. Note that the dispersal rates entered subsequently (D) will be those
applied to males, with females having lower rates. A similar approach can be used to create age-specific dispersal rates (or dispersal mortality).
Chapter 5 97
Using Functions in VORTEX
VORTEX Version 9 User’s Manual
24. Alleles confer differential reproductive rates
RATE = 40+(10*(V%2))+(10*(Z%2))
50.0
45.0
Demographic Rate
In this case, half of the alleles (specifically,
those with even numbers) cause an increment
of 10 in the breeding rate of their carriers. An
individual that is homozygous for an evennumbered allele will have a breeding rate
equal to 60%, while those homozygous for an
odd-numbered allele will have a rate equal to
40%.
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
40
50
Allele Identifier
25. Overdominance for survival, all unique founder alleles
RATE = 20+(10*(V=Z))
An infinite alleles model in which homozygotes have a mortality of 30%, while the rate for
heterozygotes is 20%.
26. Overdominance for survival, two functionally distinct founder alleles
RATE = 20+(10*((V%2)=(Z%2)))
A two allele model in which homozygotes with two odd or two even alleles have a mortality of 30%,
while the rate for heterozygotes is 20%.
27. Outbreeding depression for breeding rate upon introgression from supplemented individuals
RATE = 50-(10*((V<20)=(Z>19)))
50.0
45.0
Demographic Rate
With ten initial founders, and with some
number of individuals from another source
used as supplements at a later stage, the
breeding rate is 50% for an individual which
carries both of its alleles from the initial
founders, or both from the source population
of the supplements, vs. 40 for individuals
which are heterozygous, carrying an allele
from each source.
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0
10
20
30
Allele Identifier
98 Chapter 5
Using Functions in VORTEX
40
50
VORTEX Version 9 User’s Manual
28. Genetically-based individual variation in breeding success
RATE = 50+(5*(SNRAND((R*100)+V)+SNRAND((R*100)+Z)))
70.0
60.0
Demographic Rate
In this case, breeding rates vary around a
mean of 50 with a standard deviation equal to
5*SQRT(2).
50.0
40.0
30.0
20.0
10.0
0.0
0
10
20
30
40
50
Allele Identifier
29. Density dependence, relative to the average density across populations 1 and 2
RATE = 50*([KK(1)+KK(2)]-[NN(1)+NN(2)])/[KK(1)+KK(2)])
Note the use of KK(p) and NN(p) to indicate K and N for population p.
Chapter 5 99
Using Functions in VORTEX
VORTEX Version 9 User’s Manual
100 Appendix I
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
Appendix
I
An Overview of
Population Viability
Analysis Using VORTEX
Introduction
This Appendix presents an overview of processes threatening the health and persistence of wildlife
populations, the methods of population viability analysis, the VORTEX simulation program for PVA, and
the application of such techniques to wildlife conservation. Much of the following material is adapted
from Lacy (1993a) and Lacy (1993/4).
The Dynamics of Small Populations
Many wildlife populations that were once widespread, numerous, and occupying contiguous habitat have
been reduced to one or more small, isolated populations. The primary causes of the decline of many
species are obvious and deterministic: Populations are over-harvested; natural habitat is converted and
lost to the species, often involving the replacement of diverse ecological communities with monocultures;
environments are polluted, with the dumping of toxins into the air, water, and soil; local and now even
global climates are modified by the actions of humans; and numerous exotic competitors, predators,
parasites and diseases are introduced into communities that have never evolved defenses to the new
invaders. The primary causes of species decline are usually easy to understand, and often easy to study
and model, but usually, though not always, difficult to reverse. Even if the original causes of decline are
removed, a small isolated population is vulnerable to additional forces, intrinsic to the dynamics of small
populations, which may drive the population to extinction (Shaffer 1981; Soulé 1987; Clark and Seebeck
1990).
Of particular impact on small populations are stochastic, or random or probabilistic, processes. Indeed,
the final extinction of most populations often occurs not so much because of a continuation of the
pressures that led to the initial decline, but because of bad luck. Chance, or stochastic, processes usually
have little impact on long-term population dynamics, as long as the population is abundant and spread
over a wide geographic range and a number of habitats. Deterministic processes, such as those listed
above, predominate in widespread, still common species, while local chance events impacting subsets of
the population will average out across the broader, diverse range. When a population becomes small,
isolated, and localized, however, chance events can become important, even dominating the long-term
dynamics and fate of a population.
Appendix I 101
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
Many stages in the life history of an organism, and the processes that define the history of a biological
population, are essentially stochastic sampling phenomena. Births, deaths, dispersal, disease, sex
determination, and transmission of genes between generations all are largely probabilistic phenomena.
Small samples intrinsically have greater variance around the probabilistic mean or expectation than do
large samples, and therefore small populations will experience greater fluctuations in births, deaths, sex
ratio, and genetic variation than will larger populations. The fundamental problem facing small
populations is that the fluctuations they experience due to the multiple stages of sampling each generation
make it increasingly likely that the populations will, unpredictably, decline to zero. Once populations are
small, the probability that they will become extinct can become more strongly determined by the amount
of fluctuations in population size than in the mean, deterministic population growth rate. Thus, extinction
can be viewed as a process in which once common and widespread populations become reduced to small,
isolated fragments due to extrinsic factors, the small remnant populations then become subjected to large
fluctuations due to intrinsic processes, the local populations occasionally and unpredictably go extinct,
and the cumulative result of local extinctions is the eventual extinction of the taxon over much or all of its
original range (Gilpin and Soulé 1986; Clark et al. 1990).
The stochastic processes impacting on populations have been usefully categorized into demographic
stochasticity, environmental variation, catastrophic events, and genetic drift (Shaffer 1981). Demographic
stochasticity is the random fluctuation in the observed birth rate, death rate, and sex ratio of a population
even if the probabilities of birth and death remain constant. Assuming that births and deaths and sex
determination are stochastic sampling processes, the annual variations in numbers that are born, die, and
are of each sex can be specified from statistical theory and would follow binomial distributions. Such
demographic stochasticity will be most important to population viability perhaps only in populations that
are smaller than a few tens of animals (Goodman 1987), in which cases the annual frequencies of birth
and death events and the sex ratios can deviate far from the means.
Environmental variation is the fluctuation in the probabilities of birth and death that results from
fluctuations in the environment. Weather, the prevalence of enzootic disease, the abundances of prey and
predators, and the availability of nest sites or other required microhabitats can all vary, randomly or
cyclically, over time. The fluctuations in demographic rates caused by environmental variation is additive
to the random fluctuations due to demographic stochasticity. Thus, the difference between the observed
variation in a demographic rate over time and the distribution describing demographic variation must be
accounted for by environmental variation.
Catastrophic variation is the extreme of environmental variation, but for both methodological and
conceptual reasons rare catastrophic events are analyzed separately from the more typical annual or
seasonal fluctuations. Catastrophes such as epidemic disease, hurricanes, large-scale fires, and floods are
outliers in the distributions of environmental variation. As a result, they have quantitatively and
sometimes qualitatively different impacts on wildlife populations. (A forest fire is not just a very hot day.)
Such events often precipitate the final decline to extinction (Simberloff 1986, 1988). For example, one of
two populations of whooping crane was decimated by a hurricane in 1940 and soon after went extinct
(Doughty 1989). The only remaining population of the black-footed ferret (Mustela nigripes) was being
eliminated by an outbreak of distemper when the last 18 ferrets were captured (Clark 1989).
Genetic drift is the cumulative and non-adaptive fluctuation in allele frequencies resulting from the
random sampling of genes in each generation. This can impede the recovery or accelerate the decline of
wildlife populations for several reasons (Lacy 1993b). Inbreeding, not strictly a component of genetic
drift but correlated with it in small populations, has been documented to cause loss of fitness in a wide
variety of species, including virtually all sexually reproducing animals in which the effects of inbreeding
have been carefully studied (Wright 1977; Falconer 1981; O'Brien and Evermann 1988; Ralls et al. 1988;
Lacy et al. 1993; Lacy 1997). Even if the immediate loss of fitness of inbred individuals is not large, the
102 Appendix I
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
loss of genetic variation that results from genetic drift may reduce the ability of a population to adapt to
future changes in the environment (Fisher 1958; Robertson 1960; Selander 1983).
Thus, the effects of genetic drift and consequent loss of genetic variation in individuals and populations
negatively impact on demographic rates and increase susceptibility to environmental perturbations and
catastrophes. Reduced population growth and greater fluctuations in numbers in turn accelerates genetic
drift (Crow and Kimura 1970). These synergistic destabilizing effects of stochastic process on small
populations of wildlife have been described as “extinction vortices” (Gilpin and Soulé 1986).
What is Population (and Habitat) Viability Analysis?
Analyses which have used the VORTEX simulation for guiding conservation decisions refer variously to
“Population Viability Analysis (PVA)”, “Population and Habitat Viability Analysis (PHVA),”
“Population Vulnerability Analysis”, “Population Viability (or Vulnerability) Assessment”, and other
variants on the name. This diversity of terminology has caused some confusion among practitioners of the
PVA (or PHVA) approach, and probably even more confusion among wildlife managers who have tried
to understand what analysis was being described, and whether it could be a useful tool in their efforts to
conserve biodiversity. The diversity of perceptions about the PVA approach is not limited to its name.
Different people mean different things by PVA, and the definitions and practice of PVA are constantly
evolving. We don’t think it is not the case, as has sometimes been suggested, that some people are doing
PVA correctly, and others incorrectly, but rather that people are using different (if related) kinds of
analyses and labeling them with the same (or similar) terms. What analysis is correct depends on the need
and the application. Below, we attempt to clarify what PVA is, by suggesting a more consistent
terminology and by describing the features that characterize the application of the PVA approach to
conservation. The perspective offered here is necessarily biased by personal experiences in conservation;
we will not attempt an exhaustive historical account of this field.
Population viability analysis originally described methods of quantitative analysis to determine the
probability of extinction of a population. Shaffer (1981) first defined a minimum viable population
(MVP) as the size at which a population has a 99% probability of persistence for 1000 years, but it might
be more meaningful biologically to consider it to be the size below which a population's fate becomes
determined largely by the stochastic factors that characterize extinction vortices. One concept of
population viability analysis is any methodology used to determine an MVP (Shaffer 1990). More
broadly, PVA is the estimation of extinction probabilities and other measures of population performance
by analyses that incorporate identifiable threats to population survival into models of the extinction
process (Brussard 1985; Gilpin and Soulé 1986; Burgman et al. 1993; Lacy 1993/1994).
Shaffer's (1981) original term “minimum viable population” (MVP) has fallen into disfavor (Soulé 1987),
even as the PVA approach has risen in popularity. Shaffer stressed that an MVP was an estimate of the
population size below which the probability of extinction was unacceptably high, that different
populations would have different MVPs, and that the MVP determined for a population would depend on
the threatening factors that were considered. However, the term implied to some people that there was a
well-defined number below which extinction was certain and above which persistence was assured. Reemphasizing the probabilistic nature of the extinction process, a number of conservation biologists have
focused on methods for estimating the probability of extinction over defined time periods for a designated
population exposed to a specific scenario of environmental conditions, threats to persistence, and future
management actions and other foreseeable events (Brussard 1985; Starfield and Bleloch 1986; Soulé
1987; Simberloff 1988; Gilpin 1989; Shaffer 1990; Boyce 1992; Burgman et al. 1993). Thus, “Population
Viability Analysis” (or the synonymous “Population Viability Assessment” and “Population Vulnerability
Analysis”) came to describe any of the array of methods for quantifying the probability of extinction of a
Appendix I 103
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
population. Although PVA has been extended by some to encompass a broader approach to conservation
(see below), the term “Population Viability Analysis”, or PVA, should perhaps be reserved for its
original, yet still rather broad, meaning.
Beginning in about 1989 (Lacy et al. 1989; Seal and Lacy 1989; Seal et al. 1990), it became increasingly
recognized that PVA can often be most usefully incorporated into a strategy for the conservation of a
taxon if it is part of, and often central to, a conservation workshop that mobilizes collaboration among the
array of people with strong interest in or responsibility for a conservation effort (e.g., governmental
wildlife agencies, conservation NGOs, and the local people who interact with the species or its habitat) or
with particular expert knowledge about the species, its habitats, or the threats it faces (e.g., academic
biologists, conservation professionals, other wildlife biologists, experts on human demographics and
resource use). Conservation problems are almost always multi-faceted, involving not only complex
dynamics of biological populations, but also interactions with human populations, the past, present, and
future impacts of humans on habitats, and human political, social, and economic systems (Alvarez 1993;
Bormann and Kellert 1991; Clark 1989, 1993). Many people need to contribute knowledge, expertise, and
ideas in order to achieve the recovery of threatened species. Population viability analyses can provide a
framework for incorporating the many needed kinds of knowledge into species conservation efforts,
because PVAs do allow the assessment of many kinds of factors that threaten the persistence of
populations (Lacy 1993a; Lindenmayer et al. 1993).
The Conservation Breeding Specialist Group (CBSG) of the IUCN’s Species Survival Commission
especially has advocated and used workshops centered on PVAs to provide guidance to conservation
assessment and planning (see references to CBSG workshops in Appendix III). Over the past few years,
the PVA workshop as an approach to species conservation has expanded considerably beyond the
quantitative analysis of extinction probabilities as advanced by Shaffer (1981, 1990), Soulé (1987), Gilpin
(1989), Clark et al. (1991), Boyce (1992), and others. PVA workshops have incorporated consideration of
resource use and needs by local human populations (Seal et al. 1991; Bonaccorso et al. 1999), education
programs for the local human populations (Odum et al. 1993), trade issues (Foose et al. 1993), and trends
in human demographics and land use patterns (Walker and Molur 1994; Herrero and Seal 2000).
Recognizing that the conservation assessment workshops increasingly incorporated more than just the
population biology modeling (which still formed a core organizing and analysis framework for the
workshop), the CBSG has termed their workshops Population and Habitat Viability Analyses (PHVA).
We would recommend that the term Population and Habitat Viability Analysis (PHVA) be used to
describe the collaborative workshop approach to species conservation that centers on, but encompasses
more than, a Population Viability Analysis (in the narrow sense). The concept of a PHVA continues to
expand and evolve, as it should considering the need for more holistic and flexible approaches to
conservation (e.g., Ruggiero et al. 1994). Thus, in the usage I recommend, PVA is a quantitative analysis
of the probability of population persistence under defined sets of assumptions and circumstances. PHVA
is a workshop process that brings to bear the knowledge of many people on species conservation, eliciting
and assessing multiple options for conservation action, principally by using the tool of PVA as a way
evaluate present threats to population persistence and likely fates under various possible scenarios.
Population Viability Analysis (PVA)
Two defining characteristics of a PVA are an explicit model of the extinction process and the
quantification of threats to extinction. These features set PVA apart from many other analyses of the
threats facing species, including, for example, the IUCN Red Books of Threatened Species. As a
methodology to estimate the probability of extinction of a taxon, PVA necessarily must start with an
understanding, or model, of the extinction process (Clark et al. 1990).
104 Appendix I
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
Generally, the model of extinction underlying a PVA considers two categories of factors: deterministic
and stochastic. Deterministic factors, those that can shift species from long-term average population
growth to population decline include the well-known threats of over-harvest, habitat destruction, pollution
or other degradation of environmental quality, and the introduction of exotic predators, competitors, and
diseases. Singly or combined, these forces have driven many wildlife populations to low numbers and, for
some, to extinction. Once a population becomes small, and isolated from conspecific populations that
might serve as sources for immigrants that could stabilize demographics and genetics, its dynamics and
fate can become dominated by a number of random or stochastic processes (as outlined above and by
Shaffer 1981). Thus, even if the original deterministic causes of decline are stopped or reversed, the
instability caused by the action of stochastic processes acting on small populations can cause the
extinction of a population.
In nature, most threatening processes have both deterministic and stochastic features. For example, a high
level of poaching might be seen as a deterministic factor driving a wildlife population toward extinction,
but whether an individual animal is killed might be largely a matter of chance. In a PVA, poaching might
be modeled as a deterministic process by killing a determined proportion of the animals, or it might be
modeled as a stochastic process by giving each animal that probability of being killed but allowing the
exact numbers killed to vary over time. If the population is large and the percent of animals killed is high,
then these two ways of modelling the effects of poaching will yield the same results: the deterministic
component of poaching dominates the population dynamics. If the population is small or the percent of
animals killed is very low, then the numbers killed in a stochastic model (and in nature) might vary
substantially from year to year: the stochastic nature of poaching further destabilizes the population.
Which of the various deterministic and stochastic factors are important to consider in a PVA will depend
on the species biology, the present population size and distribution, and the threats it faces. For example,
orang utans may be threatened by forest destruction and other largely deterministic processes, but
inbreeding and randomly skewed sex ratios resulting from highly stochastic processes are unlikely to be
problems, at least not on a species-wide basis. On the other hand, even if the remnant Atlantic coastal
rainforest of Brazil is secured for the future, the populations of golden lion tamarins (Leontopithecus
rosalia) which can persist in that remnant forest are not sufficiently large to be stable in the face of
stochastic threats (Seal et al. 1990; Rylands 1993/4; Ballou et al. 1997). The identification of the primary
threats facing a taxon via a comprehensive PVA is important for conservation planning. For example,
tamarin populations might be stabilized by the translocations and reintroductions that are underway and
planned, but an orang utan PHVA recognized that releases of confiscated “pet” orang utans are unlikely
to have a conservation benefit for those populations which are facing habitat destruction, not stochastic
fluctuations and inbreeding. For many species, such as the whooping crane (Grus americana), the
temporarily extinct-in-the-wild black-footed ferret (Mustela nigripes), and the Puerto Rican parrot
(Amazona vitatta), only a single population persisted in the wild. Although those populations may have
been maintained or even increased for a number of years, the principal threat was that a local catastrophe
(e.g., disease epidemic, severe storm) could decimate the population (Clark 1989; Lacy et al. 1989;
Mirande et al. 1991). The primary recovery actions therefore needed to include the establishment of
additional populations. Tragically, some taxa, such the eastern barred bandicoot (Perameles gunnii) in
Australia, may be critically threatened simultaneously by deterministic factors and stochastic processes
(Lacy and Clark 1990).
PVA is formally an assessment of the probability of extinction, but PVA methods often focus on other
indicators of population health. Mean and variance in population growth (Lindenmayer and Lacy 1995a,
1995b, 1995c), changes in range, distribution, and habitat occupancy (Hanski and Gilpin 1991, 1997), and
losses of genetic variability (Soulé et al. 1986; Lande and Barrowclough 1987; Seal 1992; Lacy and
Lindenmayer 1995) can be analyzed and monitored. Although not yet common, monitoring of population
health could also utilize measures of developmental stability (Clarke 1995), physiological parameters
Appendix I 105
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
such as body condition (Altmann et al. 1993) or levels of the hormones related to stress and reproduction
(Sapolsky 1982, 1986), or the stability of behavior and the social structure of the population (Samuels and
Altmann 1991).
The interactions and synergism among threatening processes will often cause numerical, distributional,
physiologic, behavioral, and genetic responses to concordantly reflect species decline and vulnerability. It
remains important, however, to understand and target the primary causal factors in species vulnerability.
The recent proposal to base IUCN categories of threat on quantified criteria of probability of extinction,
or changes in such indicators as species range, numbers, and trends (Mace and Lande 1991; Mace et al.
1992; Mace and Stuart 1994; IUCN Species Survival Commission 1994) reflects the increased
understanding of the extinction process that has accompanied the development of PVA, and
simultaneously demands that much more progress be made in developing predictive models, gathering
relevant data on status and threats, and applying the PVA techniques.
Population and Habitat Viability Analysis (PHVA)
Population and Habitat Viability Analysis is a multi-faceted process or framework for assisting
conservation planning, rather than a singular technique or tool. It is often interwoven with other
techniques for managing complex systems, such as decision analysis (Maguire 1986; Maguire et al.
1990). Even when viewed as the PHVA workshop, all such conservation workshops involved and
required substantial pre-workshop and post-workshop activities. Some PHVA workshops have been
extended into multiple workshops and less formal, smaller collaborative meetings, often focused on
subsets of the larger problems of species conservation.
Although PHVAs are diverse and not well defined, the PHVA process contains a number of critical
components. First, it is essential to gather an array of experts who have knowledge of the species or
problem. A PHVA is not required to bring together experts, but it often facilitates such sharing of
expertise because the collective knowledge of many is essential for a useful PVA (in the narrow sense) to
be completed. In addition to a diversity of people, a PHVA workshop also requires and therefore
facilitates the involvement of a number of agencies and other concerned organizations. For example, the
PVA on the two endemic primates of the Tana River Primate Reserve in Kenya (Seal et al. 1991) was
convened by the Kenya Wildlife Service, facilitated by the IUCN SSC Captive Breeding Specialist
Group, benefited from the expertise contributed by members of the IUCN SSC Primate Specialist Group,
and was sponsored by the World Bank. The involvement of many agencies and interested parties is
critical to endangered species recovery.
An early requirement, or prerequisite, of a PHVA workshop is to determine the conservation problem to
be addressed, and to state the goals of the management plan. Many endangered species programs have not
clearly identified their goals. For example, at a PHVA and Conservation Assessment and Management
Plan workshop on the forest birds of the Hawaiian islands (Ellis et al. 1992a, 1992b), it became apparent
that the agencies responsible for the conservation of Hawaii's bird fauna had not determined whether their
goal was to prevent species extinctions, prevent taxa (species or subspecies) from becoming extirpated on
any of the islands they presently inhabit, preserve species in sufficient numbers and distribution to allow
them to continue to fill ecological roles in the biological communities, or the restoration of taxa to most or
all parts of the original ranges. The management actions required to achieve these various levels of
conservation are quite different. In contrast, a PHVA on the Grizzly Bear in the Central Rockies of
Canada (Herrero and Seal 2000) clearly identified that provincial policy called for maintenance of stable
or growing populations of the species. Thus, the criterion against which alternative management scenarios
were judged was whether the PVA projections indicated that the populations would not decline.
106 Appendix I
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
PHVA workshops facilitate the assembly of all available data. Often, important information is found in
the field notes of researchers or managers, in the heads of those who have worked with and thought about
the problems of the species, and in unpublished agency reports, as well as in the published scientific
literature. A pending PHVA can be the impetus that encourages the collection of data in anticipation of
presentation, review, and analysis at the workshop. For example, a Sumatran Tiger PHVA helped
stimulate the systematic collection of data on sightings and signs of tigers in protected areas throughout
the island of Sumatra, and collation and integration with a Geographic Information System (GIS) map of
habitats and human pressures on those habitats. The PHVA on the Grizzly Bear in the Central Canadian
Rockies Ecosystem provided the opportunity for detailed habitat mapping data to be integrated with
population biology data on the bears, resulting in the development of models which would allow
projection of the impacts of habitat changes on the bear populations.
It is important to specify the assumptions that underlay a PHVA, and any consequent management
recommendation. For example, the Hawaiian bird conservation efforts are constrained by a belief that no
birds bred outside of the islands should ever be brought back to the islands for release. While this position
derives from a reasonable concern for disease transmission (much of the decline of Hawaii's native birds
is thought to be due to introduced avian diseases) as much as from any political or philosophical stand,
any justification for the restriction must be questioned in light of the fact that wildlife agencies import and
release, without quarantine, 1000s of exotic gamebirds onto the islands annually.
Once experts are assembled, problems stated and goals set, data gathered, and assumptions specified, then
the PHVA process can proceed with what I describe as PVA in the narrow sense: estimation of the
probability of population persistence. The available data are used to estimate the parameters that are
needed for the model of population dynamics to be applied. Often, data are not available from which to
estimate certain key parameters. In those cases, subjective and objective, but non-quantified, information
might be solicited from the assembled experts, values might be obtained from data on related species, or a
factor might simply be omitted from the model. While such a non-precise process might consist simply of
intuitive judgements made by experts, it is important to specify how values for the parameters in the
model were obtained. The resulting limitations of the analyses should be acknowledged, and a decision
made if, how, by whom, and when the missing data would be collected so that more refined analyses
could be conducted. With the PVA model, projections of the most likely fate, and distribution of possible
fates, of the population under the specified assumptions are made.
Because so much of a PVA – the data, the model, and even the interpretation of output – is uncertain, a
PVA that provides an estimate of the probability of extinction under a single scenario is of very limited
usefulness. An essential component of the PHVA process, therefore, is sensitivity testing. Ranges of
plausible values for uncertain parameters should be tested, to determine what effects those uncertainties
might have on the results. In addition, several different PVA models might be examined at a PHVA
workshop, or the same general model tested under different structural assumptions. Different participants
in the process should assess and interpret the results. Such sensitivity testing reveals which components of
the data, model, and interpretation have the largest impact on the population projections. This will
indicate which aspects of the biology of the population and its situation contribute most to its
vulnerability and, therefore, which aspects might be most effectively targeted for management. In
addition, uncertain parameters that have a strong impact on results are those which might be the focus of
future research efforts, to better specify the dynamics of the population. Close monitoring of such
parameters might also be important for testing the assumptions behind the selected management options
and for assessing the success of conservation efforts.
Closely parallel to the testing of uncertainties in the present situation is the testing of options for
management. PVA modeling allows one to test the expected results of any given management action,
under the assumptions of the model and within the limitations of present knowledge, on the computer
Appendix I 107
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
before implementation in the field. This process can guide selection of the management options most
likely, given current knowledge, to be effective, and will define target recovery goals that should be
obtained if our knowledge is adequate and the recommended actions are followed. A PHVA workshop on
the Black Rhinoceros in Kenya's 11 rhino sanctuaries (Foose et al. 1993) suggested that periodic
movement of rhinos between fenced sanctuaries to reduce inbreeding and demographic fluctuations
would be necessary to stabilize the populations in the smaller parks. Moreover, the modeling provided
estimates of the rate at which the larger populations would be able to provide surplus animals for
translocation.
It would be an error to assume that any PVA model incorporates everything of interest. A PVA
simulation program can only include those processes that are known to the programmer. This will likely
be a subset of what might be known to the field biologists, which in turn will definitely be a subset of
those processes that impact natural populations. A number of variables affecting population dynamics and
viability are not yet commonly examined in PVA models. These include: social and ecological
determinants of dispersal; complex social processes, such as the role of non-breeders in group stability
and the impacts of other aspects of the social environment on reproductive success and survival;
competitive, exploitative, or mutualistic interactions with other species experiencing their own population
dynamics; and the effects of changes in the global environment. To date, most PVA models treat
organisms as independent actors in spatially homogeneous physical, biotic, and social environments.
There is tremendous opportunity and need for elaboration of PVA models, and it is likely that
increasingly sophisticated models will also become more specific to the individual taxa and environments
under study.
PHVA workshops must incorporate consideration of the assumptions of the PVA model used and the
biases or limitations in interpretation that could result. PHVAs consider only those threatening processes
of which we have knowledge, for which we can develop algorithms for modeling or other methods for
analysis, and for which we have some data. As a result, it is likely that PVAs will underestimate the
vulnerability of most populations to extinction, and that PHVA workshops will be less comprehensive
than is desirable. We need always to be cognizant of the limits of our understanding of wildlife
populations, and to include appropriate margins for error in our conservation strategies.
PVA is, by definition, an assessment of the probability of persistence of a population over a defined time
frame. Yet, persistence of a population, while a necessary condition for effective conservation of natural
systems, is often not sufficient. Prevention of extinction is the last stand of conservationists, but the goals
should be higher: conservation of functional biological communities and ecosystems. PVA usually
ignores the functional role of a species in a community, but a PHVA workshop should consider much
more than the prevention of the final biological extinction of the taxon. A species, such as the American
Bison (Bison bison), can be functionally extinct in terms of no longer filling its original role in nature,
even as it is praised as a conservation success story and would be considered safe from extinction and
viable.
The use of the PHVA process to help guide conservation decisions is not a singular event, in which an
analysis can be completed, management actions recommended and implemented, and conservation
thereby assured. The many uncertainties in the process mandate that PVA be used as a tool in an adaptive
management framework, and a PHVA workshop is just one stage of an effective conservation strategy. In
adaptive management, the lack of knowledge adequate to predict with certainty the best course of action
is recognized, management actions are designed in such a way that monitoring will allow testing of the
adequacy of our model and understanding, and corrective adjustments to management plans are made
whenever the accumulating data suggest that the present course is inadequate to achieve the goals and that
a better strategy exists (Holling 1978). The urgency of the biodiversity crisis will not permit us ethically
to refrain from aggressive conservation action until we have scientifically sound understanding of all the
108 Appendix I
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
factors that drive population, community, and ecosystem dynamics. PHVA provides a forum for making
use of the information we do have, in a well-documented process that is open to challenge and
improvement. PHVA workshops can, therefore, assist wildlife managers in the very difficult and
important job of using science to safeguard the future of wildlife populations.
In summary, Population Viability Analysis (PVA) and Population and Habitat Viability Analysis (PHVA)
refer to an array of interrelated and evolving techniques for assessing the survival probability of a
population and possible conservation actions. It might be useful to restrict the term PVA to its original
meaning -- the use of quantitative techniques to estimate the probability of population persistence under a
chosen model of population dynamics, a specified set of biological and environmental parameters, and
enumerated assumptions about human activities and impacts on the system. PHVA refers to a workshop
approach to conservation planning, which elicits and encourages contributions from an array of experts
and stakeholders, uses PVA and other quantitative and non-quantitative techniques to assess possible
conservation actions, and strives to achieve consensus on the best course of action from competing
interests and perspectives, incomplete knowledge, and an uncertain future.
Many of the components of PVAs and PHVAs, even when used in isolation, can be effective educational
and research tools. To be a useful framework for advancing the conservation of biodiversity, however,
PHVA must incorporate all of: (1) collection of data on the biology of the taxon, status of its habitat, and
threats to its persistence, (2) quantitative analysis of available data, (3) input of population status and
identifiable threats to persistence into analytical or simulation models of the extinction process, (4)
assessment of the probability of survival over specified periods of time, given the assumptions and
limitations of the data and model used, (5) sensitivity testing of estimates of extinction probability across
the range of plausible values of uncertain parameters, (6) specification of conservation goals for the
population, (7) identification of options for management, (8) projection of the probability of population
survival under alternative scenarios for future conservation action, (9) implementation of optimal actions
for assuring accomplishment of conservation goals, (10) continued monitoring of the population, (11)
reassessment of assumptions, data, models, and options, and (12) adjustment of conservation strategies to
respond to the best information available at all times. There are many uncertain aspects of population
dynamics, especially of endangered taxa, including few data on species biology and habitats, uncertain
political and social climate for implementing conservation actions, and the unpredictability inherent in
small populations due to the many stochastic forces that drive population dynamics.
The rapid development of PVA as a research and management tool, and the concurrent but not always
parallel expansion of the scope of what conservation threats, options, and actions are considered in PHVA
workshops, has led to confusion. Different people can describe rather distinct kinds of analyses with the
same terminology, while others use different terms to describe nearly identical approaches. The everchanging concepts of PVA and PHVA are confusing, but the flexibility of the processes is also their
strength. Current tools are inadequate to address fully the challenges of stemming the losses of
biodiversity. The PVA/PHVA framework allows and encourages rapid application of new tools, data, and
interpretations into increasingly effective conservation programs.
Methods for Analyzing Population Viability
An understanding of the multiple, interacting forces that contribute to extinction vortices is a prerequisite
for the study of extinction-recolonization dynamics in natural populations inhabiting patchy environments
(Gilpin 1987), the management of small populations (Clark and Seebeck 1990), and the conservation of
threatened wildlife (Shaffer 1981, 1990; Soulé 1987; Mace and Lande 1991).
Appendix I 109
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
Shaffer (1981) suggested several ways to conduct PVAs. Perhaps the most rigorous method, and the one
that would produce the most defensible estimates, would be an empirical observation of the stability and
long term fates of a number of populations of various sizes. Berger (1990) presented a good example of
this approach, in which he observed that populations of bighorn sheep in the mountains of the western
USA persisted only when the populations consisted of more than 100 animals. A few other studies of
wildlife populations have provided empirical data on the relationship between population size and
probability of extinction (e.g., Belovsky 1987; Thomas 1990), but presently only order of magnitude
estimates can be provided for MVPs of vertebrates (Shaffer 1987). More empirical studies are needed, but
the time and numbers of populations required for such studies are precluded in the cases of most species
threatened with extinction -- exactly those for which estimates of population vulnerability are most
urgently needed.
A more elegant and general approach to PVA is to develop analytical models of the extinction process
that will allow calculation of the probability of extinction from a small number of measurable parameters.
Goodman's (1987) model of demographic fluctuations, and applications to conservation of the classic
population genetic models of loss of genetic diversity by genetic drift (Franklin 1980; Soulé et al. 1986;
Lande and Barrowclough 1987) are valuable efforts in this direction. Unfortunately, our understanding of
population biology is not yet sufficient to provide fully adequate analytical models of the extinction
process. For example, none of the existing analytical models incorporate all three of demographic,
environmental, and genetic fluctuations, and thus they do not begin to model the array of extinction
vortices described by Gilpin and Soulé (1986). Moreover, the analytical models make extremely
simplifying assumptions about a number of the intricacies of population structure. For example, social
groupings or preferences are often assumed to be invariant or lacking, resulting in random mating; and
dispersal is usually assumed to be random between all sites (the "island model") or only to occur between
adjacent sites (the "stepping stone model"). Much more work is needed either to develop more complex
and flexible models or to demonstrate that the simple models are sufficient to provide guidance for
conservation.
A third method of conducting a PVA is the use of computer simulation modeling to project the
probability distribution of possible fates of a population. Simulation models can incorporate a very large
number of threatening processes and their interactions, if the processes can be described in terms of
quantitative algorithms and parameterized. Although many processes affecting small populations are
intrinsically indeterminate, the average long-term fate of a population and the variance around the
expectation can be studied with computer simulation models. The focus is on detailed and explicit
modeling of the forces impinging on a given population, place, and time of interest, rather than on
delineation of rules (which may not exist) that apply generally to most wildlife populations.
Modeling and Population Viability Analysis
A model is any simplified representation of a real system. We use models in all aspects of our lives, in
order to: (1) extract the important trends from complex processes, (2) permit comparison among systems,
(3) facilitate analysis of causes of processes acting on the system, and (4) make predictions about the
future. A complete description of a natural system, if it were possible, would often decrease our
understanding relative to that provided by a good model, because there is "noise" in the system that is
extraneous to the processes we wish to understand. For example, the typical representation of the growth
of a wildlife population by an annual percent growth rate is a simplified mathematical model of the much
more complex changes in population size. Representing population growth as an annual percent change
assumes constant exponential growth, ignoring the irregular fluctuations as individuals are born or
immigrate, and die or emigrate. For many purposes, such a simplified model of population growth is very
useful, because it captures the essential information we might need regarding the average change in
110 Appendix I
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
population size, and it allows us to make predictions about the future size of the population. A detailed
description of the exact changes in numbers of individuals, while a true description of the population,
would often be of much less value because the essential pattern would be obscured, and it would be
difficult or impossible to make predictions about the future population size.
In considerations of the vulnerability of a population to extinction, as is so often required for conservation
planning and management, the simple model of population growth as a constant annual rate of change is
inadequate for our needs. The fluctuations in population size that are omitted from the standard ecological
models of population change can cause population extinction, and therefore are often the primary focus of
concern. In order to understand and predict the vulnerability of a wildlife population to extinction, we
need to use a model which incorporates the processes which cause fluctuations in the population, as well
as those which control the long-term trends in population size. Many processes can cause fluctuations in
population size: variation in the environment (such as weather, food supplies, and predation), genetic
changes in the population (such as genetic drift, inbreeding, and response to natural selection),
catastrophic effects (such as disease epidemics, floods, and droughts), decimation of the population or its
habitats by humans, the chance results of the probabilistic events in the lives of individuals (sex
determination, location of mates, breeding success, survival), and interactions among these factors (Gilpin
and Soulé 1986).
Models of population dynamics which incorporate causes of fluctuations in population size in order to
predict probabilities of extinction, and to help identify the processes which contribute to a population's
vulnerability, are used in Population Viability Analysis (PVA). For the purpose of predicting vulnerability
to extinction, any and all population processes that impact population dynamics can be important. Much
analysis of conservation issues is conducted by largely intuitive assessments by biologists with experience
with the system. Assessments by experts can be quite valuable, and are often contrasted with "models"
used to evaluate population vulnerability to extinction. Such a contrast is not valid, however, as any
synthesis of facts and understanding of processes constitutes a model, even if it is a mental model within
the mind of the expert and perhaps only vaguely specified to others (or even to the expert himself or
herself).
A number of properties of the problem of assessing vulnerability of a population to extinction make it
difficult to rely on mental or intuitive models. Numerous processes impact population dynamics, and
many of the factors interact in complex ways. For example, increased fragmentation of habitat can make
it more difficult to locate mates, can lead to greater mortality as individuals disperse greater distances
across unsuitable habitat, and can lead to increased inbreeding which in turn can further reduce ability to
attract mates and to survive. In addition, many of the processes impacting population dynamics are
intrinsically probabilistic, with a random component. Sex determination, disease, predation, mate
acquisition -- indeed, almost all events in the life of an individual -- are stochastic events, occurring with
certain probabilities rather than with absolute certainty at any given time. The consequences of factors
influencing population dynamics are often delayed for years or even generations. With a long-lived
species, a population might persist for 20 to 40 years beyond the emergence of factors that ultimately
cause extinction. Humans can synthesize mentally only a few factors at a time, most people have
difficulty assessing probabilities intuitively, and it is difficult to consider delayed effects. Moreover, the
data needed for models of population dynamics are often very uncertain. Optimal decision-making when
data are uncertain is difficult, as it involves correct assessment of probabilities that the true values fall
within certain ranges, adding yet another probabilistic or chance component to the evaluation of the
situation.
The difficulty of incorporating multiple, interacting, probabilistic processes into a model that can utilize
uncertain data has prevented (to date) development of analytical models (mathematical equations
developed from theory) which encompass more than a small subset of the processes known to affect
Appendix I 111
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
wildlife population dynamics. It is possible that the mental models of some biologists are sufficiently
complex to predict accurately population vulnerabilities to extinction under a range of conditions, but it is
not possible to assess objectively the precision of such intuitive assessments, and it is difficult to transfer
that knowledge to others who need also to evaluate the situation. Computer simulation models have
increasingly been used to assist in PVA. Although rarely as elegant as models framed in analytical
equations, computer simulation models can be well suited for the complex task of evaluating risks of
extinction. Simulation models can include as many factors that influence population dynamics as the
modeler and the user of the model want to assess. Interactions between processes can be modeled, if the
nature of those interactions can be specified. Probabilistic events can be easily simulated by computer
programs, providing output that gives both the mean expected result and the range or distribution of
possible outcomes. In theory, simulation programs can be used to build models of population dynamics
that include all the knowledge of the system which is available to experts. In practice, the models will be
simpler, because some factors are judged unlikely to be important, and because the persons who
developed the model did not have access to the full array of expert knowledge.
Although computer simulation models can be complex and confusing, they are precisely defined and all
the assumptions and algorithms can be examined. Therefore, the models are objective, testable, and open
to challenge and improvement. PVA models allow use of all available data on the biology of the taxon,
facilitate testing of the effects of unknown or uncertain data, and expedite the comparison of the likely
results of various possible management options.
PVA models also have weaknesses and limitations. A model of the population dynamics does not define
the goals for conservation planning. Goals, in terms of population growth, probability of persistence,
number of extant populations, genetic diversity, or other measures of population performance must be
defined by the management authorities before the results of population modeling can be used. Because the
models incorporate many factors, the number of possibilities to test can seem endless, and it can be
difficult to determine which of the factors that were analyzed are most important to the population
dynamics. PVA models are necessarily incomplete. We can model only those factors which we
understand and for which we can specify the parameters. Therefore, it is important to realize that the
models probably underestimate the threats facing the population. Finally, the models are used to predict
the long-term effects of the processes presently acting on the population. Many aspects of the situation
could change radically within the time span that is modeled. Therefore, it is important to reassess the data
and model results periodically, with changes made to the conservation programs as needed.
Dealing with uncertainty
It is important to recognize that uncertainty regarding the biological parameters of a population and its
consequent fate occurs at several levels and for independent reasons. Uncertainty can occur because the
parameters have never been measured on the population. Uncertainty can occur because limited field data
have yielded estimates with potentially large sampling error. Uncertainty can occur because independent
studies have generated discordant estimates. Uncertainty can occur because environmental conditions or
population status have been changing over time, and field surveys were conducted during periods which
may not be representative of long-term averages. Uncertainty can occur because the environment will
change in the future, so that measurements made in the past may not accurately predict future conditions.
Sensitivity testing is necessary to determine the extent to which uncertainty in input parameters results in
uncertainty regarding the future fate of the population. If alternative plausible parameter values result in
divergent predictions for the population, then it is important to try to resolve the uncertainty with better
data. Sensitivity of population dynamics to certain parameters also indicates that those parameters
describe factors that could be critical determinants of population viability. Such factors are therefore good
candidates for efficient management actions designed to ensure the persistence of the population.
112 Appendix I
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
The above kinds of uncertainty should be distinguished from several more sources of uncertainty about
the future of the population. Even if long-term average demographic rates are known with precision,
variation over time caused by fluctuating environmental conditions will cause uncertainty in the fate of
the population at any given time in the future. Such environmental variation should be incorporated into
the model used to assess population dynamics, and will generate a range of possible outcomes (perhaps
represented as a mean and standard deviation) from the model. In addition, most biological processes are
inherently stochastic, having a random component. The stochastic or probabilistic nature of survival, sex
determination, transmission of genes, acquisition of mates, reproduction, and other processes preclude
exact determination of the future state of a population. Such demographic stochasticity should also be
incorporated into a population model, because such variability both increases our uncertainty about the
future and can also change the expected or mean outcome relative to that which would result if there were
no such variation. Finally, there is “uncertainty” which represents the alternative actions or interventions
that might be pursued as a management strategy. The likely effectiveness of such management options
can be explored by testing alternative scenarios in the model of population dynamics, in much the same
way that sensitivity testing is used to explore the effects of uncertain biological parameters.
Often, the uncertainty regarding a number of aspects of the population biology, current status, and threats
to persistence is too large to allow scientifically accurate and reliable projections of population dynamics.
Therefore, the predictions made from PVA models should be considered to be projections about what
would most likely happen to the population if various hypotheses about the status of the populations and
the threats are true. Conservation and management decisions must be made based on the most plausible
hypotheses about the population status, before sufficient data could be collected to test those hypotheses
scientifically. An important advantage of PVA models is that they forced systematic consideration and
specification of the assumptions and hypotheses that must be made in the absence of adequate data. This
facilitates careful reassessment and improvement in the analyses, as better data become available.
Questions that can be explored with PVA models
Below are some of the conservation and management questions that can be explored by Population
Viability Analysis modeling. References describing uses of VORTEX give many examples of these and
other applications of PVA techniques to guide conservation.
Using the best current information on the biology of the taxon and its habitat, are the populations
projected to persist if conditions remain as they are now? Beyond just the persistence of the population,
what is the most likely average population size, range of population sizes across years, and rate of loss of
genetic variation? If the population is at risk of extinction, is the extinction expected to result primarily
from negative average population growth (mean deaths exceeding mean births), from large fluctuations in
numbers, from effects of accumulated inbreeding, or from a combination of these factors?
Given that there is considerable uncertainty about several aspects of the species biology and its habitat, is
the population likely to persist across the plausible ranges of parameters that might characterize the
population? In particular, how sensitive are the population dynamics to varying estimates of reproductive
success, juvenile survival, adult survival, effects of natural catastrophes, initial population size, carrying
capacity of the habitat, and dispersal among populations? Are there critical values for any of these
parameters which demarcate a transition from a population that would be considered viable to one that is
not?
Which factors have the greatest influence on the projected population performance? If important factors
are identified, management actions might be designed to improve these factors or ameliorate the negative
effects. How much change would be required in aspects of the population in order to ensure population
survival?
Appendix I 113
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
What would be the effect of removing some individuals from the population? Would there be a significant
benefit from supplementing the population with individuals translocated from other populations or
released from captive breeding stocks? Can the population sustain controlled harvest? Can it sustain
poaching?
Would a corridor connecting fragmented habitats improve long-term viability? Could the same effect be
achieved by translocating a few individuals? What will happen to population viability if mortality
increases for individuals dispersing between habitat patches?
What will happen to the wildlife population if trends in human populations and human impacts on the
environment continue unabated?
The VORTEX Population Viability Analysis Model
The VORTEX computer program is a simulation of the effects of deterministic forces as well as
demographic, environmental and genetic stochastic events on wildlife populations. It is an attempt to
model many of the extinction vortices that can threaten persistence of small populations (hence, its name).
VORTEX models population dynamics as discrete, sequential events that occur according to probabilities
that are random variables following user-specified distributions. VORTEX simulates a population by
stepping through a series of events that describe an annual cycle of a typical sexually reproducing, diploid
organism: mate selection, reproduction, mortality, increment of age by one year, migration among
populations, removals, supplementation, and then truncation (if necessary) to the carrying capacity.
Although VORTEX simulates life events on an annual cycle, a user could model "years" that are other than
12 months duration. The simulation of the population is iterated many times to generate the distribution of
fates that the population might experience.
VORTEX is an individual-based model. That is, it creates a representation of each animal in its memory and
follows the fate of the animal through each year of its lifetime. VORTEX keeps track of the sex, age, and
parentage of each animal. Demographic events (birth, sex determination, mating, dispersal, and death) are
modeled by determining for each animal in each year of the simulation whether any of the events occur.
(See figure below.)
VORTEX Simulation Model Timeline
Immigrate
Breed
N
Supplement
Age 1 Year
Death
Census
Emigrate
Harvest
Carrying
Capacity
Truncation
Events listed above the timeline increase N, while
events listed below the timeline decrease N.
VORTEX requires a lot of population-specific data. For example, the user must specify the amount of
annual variation in each demographic rate caused by fluctuations in the environment. In addition, the
frequency of each type of catastrophe (drought, flood, epidemic disease) and the effects of the
114 Appendix I
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
catastrophes on survival and reproduction must be specified. Rates of migration (dispersal) between each
pair of local populations must be specified. Because VORTEX requires specification of many biological
parameters, it is not necessarily a good model for the examination of population dynamics that would
result from some generalized life history. It is most usefully applied to the analysis of a specific
population in a specific environment.
In the program explanation that follows, demographic rates are described as constants specified by the
user. Although this is the way the program is most commonly and easily used, VORTEX does provide the
capability to specify most demographic rates as functions of time, density, and other parameters (see
Chapter 5).
Demographic stochasticity
VORTEX models demographic stochasticity by determining the occurrence of probabilistic events such as
reproduction, litter size, sex determination, and death with a pseudo-random number generator. For each
life event, if the random value sampled from a specified distribution falls above the user-specified
probability, the event is deemed to have occurred, thereby simulating a binomial process. Demographic
stochasticity is therefore a consequence of the uncertainty regarding whether each demographic event
occurs for any given animal.
The source code used to generate random numbers uniformly distributed between 0 and 1 was obtained
from Maier (1991), based on the algorithm of Kirkpatrick and Stoll (1981). Random deviates from
binomial distributions, with mean p and standard deviation s, are obtained by first determining the
integral number of binomial trials, N, that would produce the value of s closest to the specified value,
according to:
N=
p(1 − p)
s2
N binomial trials are then simulated by sampling from the uniform 0-1 distribution to obtain the desired
result, the frequency or proportion of successes. If the value of N determined for a desired binomial
distribution is larger than 25, a normal approximation is used in place of the binomial distribution. This
normal approximation must be truncated at 0 and at 1 to allow use in defining probabilities, although,
with such large values of N, s is small relative to p and the truncation would be invoked only rarely. To
avoid introducing bias with this truncation, the normal approximation to the binomial (when used) is
truncated symmetrically around the mean. The algorithm for generating random numbers from a unit
normal distribution follows Latour (1986).
Environmental variation
VORTEX can model annual fluctuations in birth and death rates and in carrying capacity as might result
from environmental variation. To model environmental variation, each demographic parameter is
assigned a distribution with a mean and standard deviation that is specified by the user. Annual
fluctuations in probabilities of reproduction and mortality are modeled as binomial distributions.
Environmental variation in carrying capacity is modeled as a normal distribution. Environmental variation
in demographic rates can be correlated among populations.
Catastrophes
Catastrophes are modeled in VORTEX as random events that occur with specified probabilities. A
catastrophe will occur if a randomly generated number between zero and one is less than the probability
Appendix I 115
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
of occurrence. Following a catastrophic event, the chances of survival and successful breeding for that
simulated year are multiplied by severity factors. For example, forest fires might occur once in 50 years,
on average, killing 25% of animals, and reducing breeding by survivors 50% for the year. Such a
catastrophe would be modeled as a random event with 0.02 probability of occurrence each year, and
severity factors of 0.75 for survival and 0.50 for reproduction. Catastrophes can be local (impacting
populations independently), or regional (affecting sets of populations simultaneously).
Genetic processes
VORTEX models loss of genetic variation in populations, by simulating the transmission of alleles from
parents to offspring at a hypothetical neutral (non-selected) genetic locus. Each animal at the start of the
simulation is assigned two unique alleles at the locus. Each offspring created during the simulation is
randomly assigned one of the alleles from each parent. VORTEX monitors how many of the original alleles
remain within the population, and the average heterozygosity and gene diversity (or “expected
heterozygosity”) relative to the starting levels. VORTEX also monitors the inbreeding coefficients of each
animal, and can reduce the juvenile survival of inbred animals to model the effects of inbreeding
depression.
Inbreeding depression is modeled as a loss of viability of inbred animals during their first year. The
severity of inbreeding depression is commonly measured by the number of “lethal equivalents” in a
population (Morton et al. 1956). The number of lethal equivalents per diploid genome estimates the
average number of lethal alleles per individual in the population if all deleterious effects of inbreeding
were due entirely to recessive lethal alleles. A population in which inbreeding depression is one lethal
equivalent per diploid genome may have one recessive lethal allele per individual, it may have two
recessive alleles per individual, each of which confer a 50% decrease in survival, or it may have some
other combination of recessive deleterious alleles which equate in effect with one lethal allele per
individual.
VORTEX partitions the total effect of inbreeding (the total lethal equivalents) into an effect due to recessive
lethal alleles and an effect due to loci at which there is heterozygote advantage (superior fitness of
heterozygotes relative to all homozygote genotypes). To model the effects of lethal alleles, each founder
starts with a unique recessive lethal allele (and a dominant non-lethal allele) at up to five modeled loci.
By virtue of the deaths of individuals that are homozygous for lethal alleles, such alleles can be removed
slowly by natural selection during the generations of a simulation. This diminishes the probability that
inbred individuals in subsequent generations will be homozygous for a lethal allele.
Heterozygote advantage is modeled by specifying that juvenile survival is related to inbreeding according
to the logarithmic model:
ln(S) = A − BF
in which S is survival, F is the inbreeding coefficient, A is the logarithm of survival in the absence of
inbreeding, and B is the portion of the lethal equivalents per haploid genome that is due to heterozygote
advantage rather than to recessive lethal alleles. Unlike the situation with fully recessive deleterious
alleles, natural selection does not remove deleterious alleles at loci in which the heterozygote has higher
fitness than both homozygotes, because all alleles are deleterious when homozygous and beneficial when
present in heterozygous combination with other alleles. Thus, under heterozygote advantage, the impact
of inbreeding on survival does not diminish during repeated generations of inbreeding.
Unfortunately, for relatively few species are data available to allow estimation of the effects of
inbreeding, and the magnitude of these effects apparently varies considerably among species (Falconer
1981; Ralls et al. 1988; Lacy et al. 1993) and even among populations of the same species (Lacy et al.
1996). Even without detailed pedigree data from which to estimate the number of lethal equivalents in a
116 Appendix I
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
population and the underlying nature of the genetic load (recessive alleles or heterozygote advantage),
PVAs must make assumptions about the effects of inbreeding on the population being studied. If genetic
effects are ignored, the PVA will overestimate the viability of small populations. In some cases, it might
be considered appropriate to assume that an inadequately studied species would respond to inbreeding in
accord with the median (3.14 lethal equivalents per diploid) reported in the survey by Ralls et al. (1988).
In other cases, there might be reason to make more optimistic assumptions (perhaps the lower quartile,
0.90 lethal equivalents), or more pessimistic assumptions (perhaps the upper quartile, 5.62 lethal
equivalents). In the few species in which inbreeding depression has been studied carefully, about half of
the effects of inbreeding are due recessive lethal alleles and about half of the effects are due to
heterozygote advantage or other genetic mechanisms that are not diminished by natural selection during
generations of inbreeding, although the proportion of the total inbreeding effect can vary substantially
among populations (Lacy and Ballou 1998).
A full explanation of the genetic mechanisms of inbreeding depression is beyond the scope of this
manual, and interested readers are encouraged to refer to the references cited above.
VORTEX can model monogamous or polygamous mating systems. In a monogamous system, a relative
scarcity of breeding males may limit reproduction by females. In polygamous or monogamous models,
the user can specify the proportion of the adult males in the breeding pool. Males are randomly reassigned
to the breeding pool each year of the simulation, and all males in the breeding pool have an equal chance
of siring offspring.
Deterministic processes
VORTEX can incorporate several deterministic processes, in addition to mean age-specific birth and death
rates. Density dependence in mortality is modeled by specifying a carrying capacity of the habitat. When
the population size exceeds the carrying capacity, additional morality is imposed across all age classes to
bring the population back down to the carrying capacity. Each animal in the population has an equal
probability of being removed by this truncation. The carrying capacity can be specified to change over
time, to model losses or gains in the amount or quality of habitat.
Density dependence in reproduction is modeled by specifying the proportion of adult females breeding
each year as a function of the population size. The default functional relationship between breeding and
density allows entry of Allee effects (reduction in breeding at low density) and/or reduced breeding at
high densities.
Populations can be supplemented or harvested for any number of years in each simulation. Harvest may
be culling or removal of animals for translocation to another (unmodeled) population. The numbers of
additions and removals are specified according to the age and sex of animals.
Migration among populations
VORTEX can model up to 50 populations, with possibly distinct population parameters. Each pairwise
migration rate is specified as the probability of an individual moving from one population to another.
Migration among populations can be restricted to one sex and/or a limited age cohort. Emigration from a
population can be restricted to occur only when the number of animals in the population exceeds a
specified proportion of the carrying capacity. Dispersal mortality can be specified as a probability of
death for any migrating animal, which is in addition to age-sex specific mortality. Because of betweenpopulation migration and managed supplementation, populations can be recolonized. VORTEX tracks the
dynamics of local extinctions and recolonizations through the simulation.
Appendix I 117
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
Output
VORTEX outputs: (1) probability of extinction at specified intervals (e.g., every 10 years during a 100 year
simulation), (2) median time to extinction, if the population went extinct in at least 50% of the
simulations, (3) mean time to extinction of those simulated populations that became extinct, and (4) mean
size of, and genetic variation within, extant populations.
Standard deviations across simulations and standard errors of the mean are reported for population size
and the measures of genetic variation. Under the assumption that extinction of independently replicated
populations is a binomial process, the standard error of the probability of extinction is reported by
VORTEX as:
SE(p) =
p(1 − p)
n
in which the frequency of extinction was p over n simulated populations. Demographic and genetic
statistics are calculated and reported for each subpopulation and for the metapopulation.
Sequence of program flow
(1) The seed for the random number generator is initialized with the number of seconds elapsed since
the beginning of the 20th century.
(2) The user is prompted for an output file name, duration of the simulation, number of iterations, the
size below which a population is considered extinct, and a large number of population parameters.
(3) The maximum allowable population size (necessary for preventing memory overflow) is calculated
as:
Kmax = (K + 3s )(1 + L )
in which K is the maximum carrying capacity (carrying capacity can be specified to change during a
simulation, so the maximum carrying capacity can be greater than the initial carrying capacity), s is
the annual environmental variation in the carrying capacity expressed as a standard deviation, and L
is the specified maximum litter size.
(4) Memory is allocated for data arrays. If insufficient memory is available for data arrays then Nmax is
adjusted downward to the size that can be accommodated within the available memory and a
warning message is given. In this case it is possible that the analysis may have to be terminated
because the simulated population exceeds Nmax. Because Nmax is often several-fold greater than the
likely maximum population size in a simulation, a warning that it has been adjusted downward
because of limiting memory often will not hamper the analyses.
(5) The deterministic growth rate of the population is calculated from mean birth and death rates that
have been entered. Algorithms follow cohort life-table analyses (Ricklefs 1979). Generation time
and the expected stable age distribution are also calculated. Life-table calculations assume constant
birth and death rates, no limitation by carrying capacity, no limitation of mates, no loss of fitness due
to inbreeding depression, and that the population is at the stable age distribution. The effects of
catastrophes are incorporated into the life table analysis by using birth and death rates that are
weighted averages of the values in years with and without catastrophes, weighted by the probability
of a catastrophe occurring or not occurring.
(6) Iterative simulation of the population proceeds via steps 7 through 26 below.
118 Appendix I
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
(7) The starting population is assigned an age and sex structure. The user can specify the exact age-sex
structure of the starting population, or can specify an initial population size and request that the
population be distributed according to the stable age distribution calculated from the life table.
Individuals in the starting population are assumed to be unrelated. Thus, inbreeding can occur only
in second and later generations.
(8) Two unique alleles at a hypothetical neutral genetic locus are assigned to each individual in the
starting population and to each individual supplemented to the population during the simulation.
VORTEX therefore uses an infinite alleles model of genetic variation. The subsequent fate of genetic
variation is tracked by reporting the number of extant neutral alleles each year, the expected
heterozygosity or gene diversity, and the observed heterozygosity. The expected heterozygosity,
derived from the Hardy-Weinberg equilibrium, is given by
He = 1 −
∑ (pi2 )
in which pi is the frequency of allele i in the population. The observed heterozygosity is simply the
proportion of the individuals in the simulated population that are heterozygous. Because of the
starting assumption of two unique alleles per founder, the initial population has an observed
heterozygosity of 1.0 at the hypothetical locus and only inbred animals can become homozygous.
Proportional loss of heterozygosity through random genetic drift is independent of the initial
heterozygosity and allele frequencies of a population (Crow and Kimura 1970), so the expected
heterozygosity remaining in a simulated population is a useful metric of genetic decay for
comparison across scenarios and populations. The mean observed heterozygosity reported by
VORTEX is the mean inbreeding coefficient of the population.
(9) For each of the10 alleles at five non-neutral loci that are used to model inbreeding depression, each
founder is assigned a unique lethal allele with probability equal to 0.1 x the mean number of lethal
alleles per individual.
(10) Years are iterated via steps 11 through 25 below.
(11) The probabilities of females producing each possible size litter are adjusted to account for density
dependence of reproduction (if any).
(12) Birth rate, survival rates, and carrying capacity for the year are adjusted to model environmental
variation. Environmental variation is assumed to follow binomial distributions for birth and death
rates and a normal distribution for carrying capacity, with mean rates and standard deviations
specified by the user. At the outset of each year a random number is drawn from the specified
binomial distribution to determine the percent of females producing litters. The distribution of litter
sizes among those females that do breed is maintained constant. Another random number is drawn
from a specified binomial distribution to model the environmental variation in mortality rates. If
environmental variations in reproduction and mortality are chosen to be correlated, the random
number used to specify mortality rates for the year is chosen to be the same percentile of its binomial
distribution as was the number used to specify reproductive rate. Otherwise, a new random number
is drawn to specify the deviation of age- and sex-specific mortality rates from their means.
Environmental variation across years in mortality rates is always forced to be correlated among age
and sex classes.
The carrying capacity (K) for the year is determined by first increasing or decreasing the carrying
capacity at year 1 by an amount specified by the user to account for changes over time.
Appendix I 119
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
Environmental variation in K is then imposed by drawing a random number from a normal
distribution with the specified values for mean and standard deviation.
(13) Birth rates and survival rates for the year are adjusted to model any catastrophes determined to have
occurred in that year.
(14) Breeding males are selected for the year. A male of breeding age is placed into the pool of potential
breeders for that year if a random number drawn for that male is less than the proportion of adult
males specified to be breeding. Breeding males are selected independently each year; there is no
long-term tenure of breeding males and no long-term pair bonds.
(15) For each female of breeding age, a mate is drawn at random from the pool of breeding males for that
year. If the user specifies that the breeding system is monogamous, then each male can only be
paired with a single female each year. Males are paired only with those females which have already
been selected for breeding that year. Thus, males will not be the limiting sex unless there are
insufficient males to pair with the successfully breeding females.
If the breeding system is polygynous, then a male may be selected as the mate for several females.
The degree of polygyny is determined by the proportion of males in the pool of potential breeders
each year.
The size of the litter produced by that pair is determined by comparing the probabilities of each
potential litter size (including litter size of 0, no breeding) to a randomly drawn number. The
offspring are produced and assigned a sex by comparison of a random number to the specified birth
sex ratio. Offspring are assigned, at random, one allele at the hypothetical genetic locus from each
parent.
(16) The genetic kinship of each new offspring to each other living animal in the population is
determined. The kinship between new animal A, and another existing animal, B, is
f AB = 0.5(fMB + fPB )
in which fij is the kinship between animals i and j, M is the mother of A, and P is the father of A. The
inbreeding coefficient of each animal is equal to the kinship between its parents, F = fMP, and the
kinship of an animal to itself is f A = 0.5(1 + F ) . (See Ballou 1983 for a detailed description of this
method for calculating inbreeding coefficients.)
(17) The survival of each animal is determined by comparing a random number to the survival
probability for that animal. In the absence of inbreeding depression, the survival probability is given
by the age and sex-specific survival rate for that year. If a newborn individual is homozygous for a
lethal allele, it is killed. Otherwise, the survival probability for individuals in their first year is
multiplied by
e − b(1− Pr [Lethals ]) F
in which b is the number of lethal equivalents per haploid genome, and Pr[Lethals] is the proportion
of this inbreeding effect due to lethal alleles.
(18) The age of each animal is incremented by 1.
(19) If more than one population is being modeled, migration among populations occurs stochastically
with specified probabilities.
120 Appendix I
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
(20) If population harvest is to occur that year, the number of harvested individuals of each age and sex
class are chosen at random from those available and removed. If the number to be removed do not
exist for an age-sex class, VORTEX continues but reports that harvest was incomplete.
(21) Dead animals are removed from the computer memory to make space for future generations.
(22) If population supplementation is to occur in a particular year, new individuals of the specified
age-class are created. Each immigrant is assumed to be genetically unrelated to all other individuals
in the population, and it carries the number of lethal alleles that was specified for the starting
population.
(23) The population growth rate is calculated as the ratio of the population size in the current year to the
previous year.
(24) If the population size (N) exceeds the carrying capacity (K) for that year, additional mortality is
imposed across all age and sex classes. The probability of each animal dying during this carrying
capacity truncation is set to (N - K)/N, so that the expected population size after the additional
mortality is K.
(25) Summary statistics on population size and genetic variation are tallied and reported.
(26) Final population size and genetic variation are determined for the simulation.
(27) Summary statistics on population size, genetic variation, probability of extinction, and mean
population growth rate are calculated across iterations and output.
Appendix I 121
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
122 Appendix I
An Overview of Population Viability Analysis Using VORTEX
VORTEX Version 9 User’s Manual
Appendix
II
Literature Cited
Akçakaya, H.R. 1997. RAMAS Metapop: Viability Analysis for Stage-Structured Metapopulations (Version 2.0).
Setauket, NY: Applied Biomathematics.
Altmann, J., D. Schoeller, S.A. Altmann, P. Muruthi and R.M. Sapolsky. 1993. Body size and fatness of free-living
baboons reflect food availability and activity levels. American Journal of Primatology 30:149-161.
Alvarez, K. 1993. Twilight of the Panther. Biology, Bureaucracy, and Failure in an Endangered Species Program.
Myakka River Publ., Sarasota, Florida.
Ballou, J.D. 1983. Calculating inbreeding coefficients from pedigrees. Pages 509-520 in: Schonewald-Cox, C.M,
S.M. Chambers, B. MacBryde, and W.L. Thomas (eds.). Genetics and Conservation: A Reference for Managing
Wild Animal and Plant Populations. Menlo Park, California: Benjamin/Cummings.
Ballou, J.D. 1997. Ancestral inbreeding only minimally affects inbreeding depression in mammalian populations.
Journal of Heredity 88:169-178.
Ballou, J.D., R.C. Lacy, D. Kleiman, A. Rylands, and S. Ellis (eds.). 1997. Leontopithecus II. The Second
Population and Habitat Viability Assessment for Lion Tamarins (Leontopithecus). Apple Valley, MN:
Conservation Breeding Specialist Group (SSC/IUCN).
Belovsky, G.E. 1987. Extinction models and mammalian persistence. Pages 35-57 in: Soulé, M.E. (ed.). Viable
Populations for Conservation. Cambridge: Cambridge University Press.
Berger, J. 1990. Persistence of different-sized populations: an empirical assessment of rapid extinctions in bighorn
sheep. Conservation Biology 4:91-98.
Bonaccorso, F., P. Clark, P.S. Miller, and O. Byers (eds.). 1999. Conservation Assessment and Management Plan
for the Tree Kangaroos of Papua New Guinea and Population and Habitat Viability Assessment for Matschie’s
Tree Kangaroo: Final Report. Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Bormann, F.H. and S.R. Kellert. 1991. Ecology, Economics, and Ethics. The Broken Circle. New Haven: Yale
University Press.
Boyce, M.S. 1992. Population viability analysis. Annual Review of Ecology and Systematics 23:481-506.
Brussard, P. 1985. Minimum viable populations: how many are too few? Restoration and Management Notes
3:21-25.
Burgman, M., S. Ferson and H.R. Akçakaya. 1993. Risk Assessment in Conservation Biology. New York: Chapman
and Hall.
Appendix II 123
Literature Cited
VORTEX Version 9 User’s Manual
Caswell, H. 2001. Matrix Population Models. 2nd ed. Sunderland, MA: Sinauer.
Caughley, G. 1977. Analysis of Vertebrate Populations. London: John Wiley and Sons.
Charlesworth, D., and B. Charlesworth. 1987. Inbreeding depression and its evolutionary consequences. Annual
Reviews of Ecology and Systematics 18:237-268.
Clark, T.W. 1989. Conservation Biology of the Black-Footed Ferret. Philadelphia: Wildlife Preservation Trust
International.
Clark, T.W. 1993. Creating and using knowledge for species and ecosystem conservation: Science, organizations,
and policy. Perspectives in Biology and Medicine 36:497-525.
Clark, T.W., R.M. Warneke and G.G. George. 1990. Management and conservation of small populations. Pages 118 in: Clark, T.W. and J.H. Seebeck (eds.). Management and Conservation of Small Populations. Brookfield,
Illinois: Chicago Zoological Society.
Clark, T.W., G.N. Backhouse and R.C. Lacy. 1991. The population viability assessment workshop: A tool for
threatened species management. Endangered Species Update 8:1-5.
Clark, T.W., and Seebeck, J.H. (eds.) 1990. Management and Conservation of Small Populations. Brookfield,
Illinois: Chicago Zoological Society.
Clarke, G.M. 1995. Relationships between developmental stability and fitness: Application for conservation biology.
Conservation Biology 9:18-24.
Crow, J.F. and Kimura, M. 1970. Introduction to Population Genetics Theory. New York: Harper and Row.
Doughty, R.W. 1989. Return of the Whooping Crane. Austin : University of Texas Press.
Edroma, E.L., N. Rosen, and P.S. Miller (eds.). 1997. Conserving the Chimpanzees of Uganda: Population and
Habitat Viability Assessment for Pan troglodytes schweinfurthii. Apple Valley, MN: Conservation Breeding
Specialist Group (SSC/IUCN).
Ellis, S., K. Hughes, C. Kuehler, R.C. Lacy and U.S. Seal (eds.). 1992a. Alala, Akohekohe, and Palila Population
and Habitat Viability Assessment Reports. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Ellis, S., C. Kuehler, R.C. Lacy, K. Hughes and U.S. Seal (eds.). 1992b. Hawai`ian Forest Birds Conservation
Assessment and Management Plan. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Falconer, D.S. 1981. Introduction to Quantitative Genetics. 2nd ed. New York: Longman.
Fisher, R.A. 1958. The Genetical Theory of Natural Selection. 2nd ed. New York; Dover.
Fowler, C.W. 1981. Density dependence as related to life history strategy. Ecology 62:602-610.
Franklin, I.R. 1980. Evolutionary change in small populations. Pages 135-149 in: Soulé, M.E. and B.A. Wilcox
(eds.). Conservation Biology: An Ecological/Evolutionary Perspective. Sunderland, MA: Sinauer Associates.
Foose, T.J., R.C. Lacy, R. Brett and U.S. Seal (eds.). 1993. Kenyan Black Rhino Metapopulation Workshop Report.
Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Gilpin, M.E. 1987. Spatial structure and population vulnerability. Pages 125-139 in: Soulé, M.E. (ed.). Viable
Populations for Conservation. Cambridge: Cambridge University Press.
Gilpin, M.E. 1989. Population viability analysis. Endangered Species Update 6:15-18.
Gilpin, M.E., and M.E. Soulé. 1986. Minimum viable populations: processes of extinction. Pages 19-34 in: Soulé,
M.E. (ed.). Conservation Biology: The Science of Scarcity and Diversity. Sunderland, MA: Sinauer Associates.
Goodman, D. 1987. The demography of chance extinction. Pages 11 – 34 in: Soulé, M.E. (ed.). Viable Populations
for Conservation. Cambridge: Cambridge University Press.
Gunn, A., U.S. Seal, and P.S. Miller (eds.). 1998. Population and Habitat Viability Assessment Workshop for Peary
Caribou and Arctic-Island Caribou (Rangifer tarandus). Apple Valley, MN: Conservation Breeding Specialist
Group (SSC/IUCN).
124 Appendix II
Literature Cited
VORTEX Version 9 User’s Manual
Hanski, I.A and M.E Gilpin. 1991. Metapopulation dynamics: A brief history and conceptual domain. Biological
Journal of the Linnean Society 42:3-16.
Hanski, I.A and M.E Gilpin (eds.). 1997. Metapopulation Biology: Ecology, Genetics, and Evolution. London:
Academic Press.
Hedrick, P.W. 1994. Purging inbreeding depression and the probability of extinction: full-sib mating. Heredity
73:363-372.
Herrero, S., P.S. Miller and U. S. Seal (eds.). 2000. Population and Habitat Viability Assessment Workshop for the
Grizzly Bear of the Central Rockies Ecosystem (Ursus arctos horribilis). Apple Valley, MN: Conservation
Breeding Specialist Group (SSC/IUCN).
Hobbs, N.T., and D.M. Swift. 1985. Estimates of habitat carrying capacity incorporating explicit nutritional
constraints. Journal of Wildlife Management 49:814-822.
Holling, C.S. (ed.). 1978. Adaptive Environmental Assessment and Management. International Series on Applied
Systems Analysis 3, International Institute for applied systems analysis. Toronto: John Wiley and Sons.
Hosack, D.A. 1998. Population Viability Analysis Workshop for the Endangered Sonoran Pronghorn (Antilocapra
americana sonoriensis) in the United States. Washington, DC: Defenders of Wildlife.
Ims, R.A., and N.G. Yoccoz. 1997. Studying transfer processes in metapopulations: Emigration, migration, and
colonization. Pages 247 – 265 in: Hanski, I.A. and M.E. Gilpin (eds.). Metapopulation Biology: Ecology,
Genetics, and Evolution. London: Academic Press.
IUCN Species Survival Commission. 1994. IUCN Red List Categories. IUCN, Gland, Switzerland.
Kirkpatrick, S, and Stoll, E. 1981. A very fast shift-register sequence random number generator. Journal of
Computational Physics 40:517.
Kjos, C., O. Byers, P.S. Miller, J. Borovansky, and U.S. Seal (eds.). 1998. Population and Habitat Viability
Assessment Workshop for the Winged Mapleleaf Mussel (Quadrula fragosa): Final Report. Apple Valley, MN:
Conservation Breeding Specialist Group (SSC/IUCN).
Krebs, C.J. 1994. Ecology: The Experimental Analysis of Distribution and Abundance. 4th ed. New York: Harper
Collins.
Lacy, R.C. 1993a. VORTEX: A computer simulation model for Population Viability Analysis. Wildlife Research
20:45-65.
Lacy, R.C. 1993b. Impacts of inbreeding in natural and captive populations of vertebrates: Implications for
conservation. Perspectives in Biology and Medicine. 36:480-496.
Lacy, R.C. 1993/1994. What is Population (and Habitat) Viability Analysis? Primate Conservation 14/15:27-33.
Lacy, R.C. 1997. Importance of genetic variation to the viability of mammalian populations. Journal of Mammalogy
78:320-335.
Lacy, R.C., G. Alaks, and A. Walsh. 1996. Hierarchical analysis of inbreeding depression in Peromyscus polionotus.
Evolution 50:2187-2200.
Lacy, R.C. and J.D. Ballou. 1998. Effectiveness of selection in reducing the genetic load in populations of
Peromyscus polionotus during generations of inbreeding. Evolution 52:900-909.
Lacy, R.C. and T.W. Clark. 1990. Population viability assessment of the eastern barred bandicoot in Victoria. Pages
131-146 in: Clark, T.W. and J.H. Seebeck (eds.). Management and Conservation of Small Populations.
Brookfield, IL: Chicago Zoological Society.
Lacy, R.C, N.R. Flesness and U.S. Seal (eds.). 1989. Puerto Rican Parrot Population Viability Analysis. Report to
the U.S. Fish and Wildlife Service. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Lacy, R.C. and D.B. Lindenmayer. 1995. A simulation study of the impacts of population sub-division on the
mountain brushtail possum, Trichosurus caninus Ogilby (Phalangeridae: Marsupialia), in south-eastern
Appendix II 125
Literature Cited
VORTEX Version 9 User’s Manual
Australia. II. Loss of genetic variation within and between sub-populations. Biological Conservation 73:131142.
Lacy, R.C., and P.S. Miller. 2002. Incorporating human populations and activities into population viability analysis.
Pages 490-510 in S.R. Beissinger and D.R. McCullough (eds.), Population Viability Analysis. Chicago:
University of Chicago Press.
Lacy, R.C., Petric, A.M., and Warneke, M. 1993. Inbreeding and outbreeding depression in captive populations of
wild species. Pages 352-374 in Thornhill, N.W. (ed.). The Natural History of Inbreeding and Outbreeding.
Chicago: University of Chicago Press.
Lande, R. and G.F. Barrowclough. 1987. Effective population size, genetic variation, and their use in population
management. Pages 87-123 in Soulé, M.E. (ed.). Viable Populations for Conservation. Cambridge: Cambridge
University Press.
Latour, A. (1986). Polar normal distribution. Byte (August 1986):131-2.
Lindenmayer, D.B. and R.C. Lacy. 1995a. Metapopulation viability of Leadbeater's Possum, Gymnobelideus
leadbeateri, in fragmented old-growth forests. Ecological Applications 5:164-182.
Lindenmayer, D.B. and R.C. Lacy. 1995b. Metapopulation viability of arboreal marsupials in fragmented oldgrowth forests: Comparison among species. Ecological Applications 5:183-199.
Lindenmayer, D.B. and R.C. Lacy. 1995c. A simulation study of the impacts of population sub-division on the
mountain brushtail possum, Trichosurus caninus Ogilby (Phalangeridae: Marsupialia), in south-eastern
Australia. I. Demographic stability and population persistence. Biological Conservation 73:119-129.
Lindenmayer, D.B., T.W. Clark, R.C. Lacy and V.C. Thomas. 1993. Population viability analysis as a tool in
wildlife conservation policy: A review with reference to Australia. Environmental Management 17:745-758.
Mace, G.M. and R. Lande. 1991. Assessing extinction threats: Toward a re-evaluation of IUCN threatened species
categories. Conservation Biology 5:148-157.
Mace, G. et al. 1992. The development of new criteria for listing species on the IUCN Red List. Species 19:16-22.
Mace, G. and S. Stuart. 1994. Draft IUCN Red List Categories, Version 2.2. Species 21-22:13-24.
MacNab, J. 1985. Carrying capacity and related slippery shibboleths. Wildlife Society Bulletin 13:403-410.
Maguire, L.A. 1986. Using decision analysis to manage endangered species populations. Journal of Environmental
Management 22:345-360.
Maguire, L.A., R.C. Lacy, R.J. Begg and T.W. Clark. 1990. An analysis of alternative strategies for recovering the
eastern barred bandicoot in Victoria. Pages 147-164 in: Clark, T.W. and J.H. Seebeck (eds.). Management and
Conservation of Small Populations. Brookfield, IL: Chicago Zoological Society.
Maier, W.L. (1991). A fast pseudo random number generator. Dr. Dobb's Journal (May 1991):152-7.
Manansang, J., A. MacDonald, D. Siswomartono, P.S. Miller, and U.S. Seal (eds.). 1996. Population and Habitat
Viability Assessment for the Babirusa (Babyrousa babyrussa). Apple Valley, MN: Conservation Breeding
Specialist Group (SSC/IUCN).
Mirande, C., R. Lacy, and U. Seal (eds.). 1991. Whooping Crane (Grus americana) Conservation Viability
Assessment Workshop Report. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Morton, N.E., Crow, J.F., and Muller, H.J. 1956. An estimate of the mutational damage in man from data on
consanguineous marriages. Proceedings of the National Academy of Sciences, USA 42:855-863.
O'Brien, S.J. and Evermann, J.F. 1988. Interactive influence of infectious diseases and genetic diversity in natural
populations. Trends in Ecology and Evolution 3:254-9.
Odum, A., et al (eds.). 1993. Aruba Island Rattlesnake Population and Habitat Viability Assessment (PHVA)
Workshop. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Pergams, O.R.W., R.C. Lacy, and M.V. Ashley. 1999. Conservation and management of Anacapa Island deer mice.
Conservation Biology, in press.
126 Appendix II
Literature Cited
VORTEX Version 9 User’s Manual
Petit, S., and L. Pors. 1996. Survey of columnar cacti and carrying capacity for nectar-feeding bats on Curaçao.
Conservation Biology 10:769-775.
Pielou, E.C. 1977. Mathematical Ecology. New York: John Wiley and Sons.
Ralls, K., Ballou, J.D., and Templeton. A.R. 1988. Estimates of lethal equivalents and the cost of inbreeding in
mammals. Conservation Biology 2:185-93.
Ricklefs, R.E. 1979. Ecology. 2nd ed. New York: Chiron.
Robertson, A. (1960). A theory of limits in artificial selection. Proceedings Royal Society of London 153B:234-49.
Rohlf, F.J., and R.R. Sokal. 1981. Statistical Tables. 2nd ed. New York: W.H. Freeman and Company.
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.
Rylands, A.B. 1993/1994. Population viability analyses and the conservation of the lion tamarins, Leontopithecus, of
south-east Brazil. Primate Conservation 14/15: 34-42.
Samuels, A. and J. Altmann. 1991. Baboons of the Amboseli basin: Demographic stability and change. International
Journal of Primatology 12:1-19.
Sapolsky, R.M. 1982. The endocrine stress-response and social status in the wild baboon. Hormones and Behavior
15:279-292.
Sapolsky, R.M. 1986. Endocrine and behavioral correlates of drought in the wild baboon. American Journal of
Primatology 11:217-227.
Seal, U.S. (ed.). 1992. Genetic Management Strategies and Population Viability of the Florida Panther (Felis
concolor coryi). Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Seal, U.S. (ed.). 1994. Attwater’s Prairie Chicken Population and Habitat Viability Assessment. Apple Valley, MN:
Captive Breeding Specialist Group (SSC/IUCN).
Seal, U.S. and R.C. Lacy (eds.). 1989. Florida Panther Population Viability Analysis. Report to the U.S. Fish and
Wildlife Service. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Seal, U.S., J.D. Ballou and C.V. Padua (eds.). 1990. Leontopithecus Population Viability Analysis Workshop Report.
Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Seal, U.S., R.C. Lacy, K. Medley, R. Seal and T.J. Foose (eds.). 1991. Tana River Primate Reserve Conservation
Assessment Workshop Report. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Selander, R.K. 1983. Evolutionary consequences of inbreeding. Pages 201-215 in: Schonewald-Cox, C.M, S.M.
Chambers, B. MacBryde, and W.L. Thomas (eds.). Genetics and Conservation: A Reference for Managing Wild
Animal and Plant Populations. Menlo Park, CA: Benjamin/Cummings.
Shaffer, M.L. 1981. Minimum population sizes for species conservation. Bioscience 1:131-134.
Shaffer, M.L. 1987. Minimum viable populations: Coping with uncertainty. Pages 69-86 in: Soulé, M.E. (ed.).
Viable Populations for Conservation. Cambridge: Cambridge University Press.
Shaffer, M.L. 1990. Population viability analysis. Conservation Biology 4:39-40.
Simberloff, D.A. 1986. The proximate causes of extinction. Pages 259-276 in: Raup, D.M., and D. Jablonski (eds.).
Patterns and Processes in the History of Life. Berlin: Springer-Verlag.
Simberloff, D.A. 1988. The contribution of population and community biology to conservation science. Annual
Review of Ecology and Systematics 19:473-511.
Simmons, M.J., and J.F. Crow. 1977. Mutations affecting fitness in Drosophila populations. Annual Review of
Genetics 11:49-78.
Sokal, R.R., and F.J. Rohlf. 1994. Biometry. 3rd ed. New York: W.H. Freeman and Company.
Soulé, M.E. (ed). 1987. Viable Populations for Conservation. Cambridge: Cambridge University Press.
Appendix II 127
Literature Cited
VORTEX Version 9 User’s Manual
Soulé, M., M. Gilpin, W. Conway and T. Foose. 1986. The millenium ark: How long a voyage, how many
staterooms, how many passengers? Zoo Biology 5:101-113.
Starfield, A.M. and A.L. Bleloch. 1986. Building Models for Conservation and Wildlife Management., New York:
Macmillan.
Thomas, C.D. 1990. What do real population dynamics tell us about minimum population sizes? Conservation
Biology 4:324-327.
Walker, S. and S. Molur (eds.). 1994. Population and Habitat Viability Analysis (PHVA) Workshop for
Indian/Nepali Rhinoceros. Zoo Outreach Organisation/CBSG India, Coimbatore.
Werikhe, S., L. Macfie, N. Rosen, and P.S. Miller (eds.). 1998. Can the Mountain Gorilla Survive? Population and
Habitat Viability Assessment Workshop for Gorilla gorilla beringei. Apple Valley, MN: Conservation Breeding
Specialist Group (SSC/IUCN).
Wright, S. 1977. Evolution and the Genetics of Populations. Vol. 3. Experimental Results and Evolutionary
Deductions. Chicago: University of Chicago Press.
Zar, J.H. 1996. Biostatistical Analysis. 3rd ed. Englewood Cliffs: Prentice Hall.
128 Appendix II
Literature Cited
VORTEX Version 9 User’s Manual
Appendix
III
VORTEX Bibliography
We have attempted to compile an exhaustive list of articles and reports
that use VORTEX as a tool in population viability analysis. If we have
missed any other examples, or if you have published a contribution that
you would like to add to the growing list, please contact the Conservation
Breeding Specialist Group (see page 3 for details) and we will update this Bibliography.
Ahlmann, V., K. Collins, and U.S. Seal (eds.). 2000. Riverine Rabbit (Bunolagus monticularis): A Population and
Habitat Viability Assessment Workshop. Apple Valley, MN: Conservation Breeding Specialist Group
(SSC/IUCN).
Allendorf, F. and N. Ryman. 2002. The role of genetics in population viability analysis. Pages 50-85 in: Beissinger,
S.R. and D.R. McCullough (eds.). Population Viability Analysis. Chicago, IL: University of Chicago Press.
Araya, B., D. Garland, G. Espinoza, A. Sanhuesa, A. Simeone, A. Teare, C. Zavalaga, R. Lacy, and S. Ellis (eds.).
2000. Population and Habitat Viability Assessment for the Humboldt Penguin (Spheniscus humboldti). Final
Report. Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Armbruster, P., P. Fernando, and R. Lande. 1999. Time frames for population viability analysis of species with long
generations: An example with Asian elephants. Animal Conservation 2(1):69-73.
Armstrong, D.P. and J.G. Ewen. 2002. Dynamics and viability of a New Zealand robin population reintroduced to
regenerating fragmented habitat. Conservation Biology 16(4):1074-1085.
Armstrong, D.P., Perrott, J.K. and Castro, I. 1997. The effect of food supply on the viability of hihi populations: An
experimental study on Mokoia. Report to WWF-NZ, Wellington, New Zealand, 130pp.
Arteaga, A., I. Canizales, G. Hernandez, M. Cruz Lamas, A. De Luca, M. Muñoz, A. Ochoa, A. Seijas, J.
Thorbjarnarson, A. Velasco, S. Ellis, and U. Seal (eds.). 1997. Taller de Análisis de la Viabilidad Poblacional y
del Hábitat del Caiman del Orinoco (Crocodylus intermedius). Apple Valley, MN: Conservation Breeding
Specialist Group (SSC/IUCN).
Ashraf, N.V.K., R. Chellam, S. Molur, U.S. Seal, D. Sharma, and S. Walker (eds.). 1995. Asiatic Lion Population
and Habitat Viability Assessment Report. Apple Valley, MN: Conservation Breeding Specialist Group
(SSC/IUCN).
Asquith, N. M. 2001. Misdirections in conservation biology. Conservation Biology 15(2):345-352.
Aurioles Gamboa, D., C. Godínez Reyes, M.E. Durán Lizarraga, F.J. García Rodrígurz, C.J. Hernández Camacho, S.
Luque, P.S. Miller and S. Ellis (eds.). 1999. Conservación, Análisis y Manejo Planificado sobre los Pinnípedos
de México y Análisis de la Viabilidad de la Población y del Hábitat para el Lobo Marino de California
(Zalophus californianus californianus). Apple Valley, MN: Conservation Breeding Specialist Group
(SSC/IUCN).
Appendix III 129
VORTEX Bibliography
VORTEX Version 9 User’s Manual
Ballou, J.D., R.C. Lacy, D. Kleiman, A. Rylands, and S. Ellis (eds.). 1997. Leontopithecus II. The Second
Population and Habitat Viability Assessment for Lion Tamarins (Leontopithecus). Apple Valley, MN:
Conservation Breeding Specialist Group (SSC/IUCN).
Barongi, R., J. Ventocilla, P.S. Miller, and U.S. Seal (eds.). 1996. Population and Habitat Viability Assessment for
Baird’s Tapir (Tapirus bairdi). Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Beissinger, S.R. 2002. Population viability analysis: Past, present, future. Pages 5-17 in: Beissinger, S.R. and D.R.
McCullough (eds.). Population Viability Analysis. Chicago, IL: University of Chicago Press.
Beissinger, S.R., J.R. Walters, D.G. Catanzano, K.G. Smith, J.B. Dunning, S.M. Haig, B.R. Noon, and B.M. Smith.
2002. The use of models in avian conservation. Current Ornithology 17 (in press).
Belovsky, G.E., C. Mellison, C. Larson, and P.A. Van Zandt. 2002. How good are PVA models? Testing their
predictions with experimental data on the brine shrimp. Pages 257-283 in: Beissinger, S.R. and D.R.
McCullough (eds.). Population Viability Analysis. Chicago, IL: University of Chicago Press.
Berry, H., M. Bush, B. Davidson, O. Forge, B. Fox, J. Grisham, M. Howe, S. Hurlbut, L. Marker-Kraus, J.
Martenson, L. Munson, K. Nowell, M. Schumann, T. Shille, K. Venzke, T. Wagener, D. Wildt, S. Ellis, and U.
Seal (eds.). 1997. Population and Habitat Viability Assessment for the Namibian Cheetah (Acinonyx jubatus)
and Lion (Panthera leo). Workshop Report. Apple Valley, MN: Conservation Breeding Specialist Group
(SSC/IUCN).
Bonaccorso, F., P. Clark, P.S. Miller, and O. Byers (eds.). 1999. Conservation Assessment and Management Plan
for the Tree Kangaroos of Papua New Guinea and Population and Habitat Viability Assessment for Matschie’s
Tree Kangaroo: Final Report. Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Bowland, A.E., K.S. Bishop, P.J. Taylor, J.Lamb, F.H. van der Bank, E. van Wyk, and D. York. 2001. Estimation
and management of genetic diversity in small populations of plains zebra (Equus quagga) in KwaZulu-Natal,
South Africa. Biochemical Systematics & Ecology 29(6): 563-583.
Boyce, M.S. 2002. Reconciling the small-population and declining-population paradigms. Pages 41-49 in:
Beissinger, S.R. and D.R. McCullough (eds.). Population Viability Analysis. Chicago, IL: University of
Chicago Press.
Brito, D., and F.A.S. Fernandez. 2000. Metapopulation viability of the marsupial Micoureus demerarae in small
Atlantic forest fragments in south-eastern Brazil. Animal Conservation 3(3):201-209.
Brook, B.W., M.A. Burgman, and R. Frankham. 2000. Differences and congruencies between PVA packages: The
importance of sex ratio for predictions of extinction risk. Conservation Ecology 4(1): 6. [online] URL:
http://www.consecol.org/vol4/iss1/art6.
Brook, B.W., Cannon, J.R., Lacy, R.C., Mirande, C. and Frankham, R. 1999. A comparison of the population
viability analysis packages GAPPS, INMAT, RAMAS and VORTEX for the Whooping Crane (Grus americana).
Animal Conservation 2:23-31.
Brook, B.W. & Kikkawa, J. 1998. Examining threats faced by island birds: A PVA on the Capricorn silvereye using
long-term data. Journal of Applied Ecology 35: 491-503.
Brook, B.W., Lim, L., Harden, R. and Frankham, R. 1997a. Does population viability analysis software predict the
behaviour of real populations? A retrospective study on the Lord Howe Island woodhen Tricholimnas sylvestris
(Sclater). Biological Conservation 82: 119-128.
Brook, B.W., Lim, L., Harden, R. and Frankham, R. 1997b. How secure is the Lord Howe Island Woodhen? A
population viability analysis using VORTEX. Pacific Conservation Biology 3:125-133.
Bustamante, J. 1996. Population viability analysis of captive and released bearded vulture populations. Conservation
Biology 10:822-831.
Cancino, J., P.S. Miller, J. Bernal Stoopen, and J. Lewis (eds.). 1995. Population and Habitat Viability Assessment
for the Peninsular Pronghorn (Antilocapra americana peninsularis). Apple Valley, MN: Conservation Breeding
Specialist Group (SSC/IUCN).
130 Appendix III
VORTEX Bibliography
VORTEX Version 9 User’s Manual
Chapman, A., Brook, B.W., Clutton-Brock, T.H., Grenfell, B.T. and Frankham, R. 2001. Population viability
analysis on a cycling population: a cautionary tale. Biological Conservation 97(1):61-69.
Clark, T.W., G.N. Backhouse, and R.C. Lacy. 1991a. The population viability assessment workshop: A tool for
threatened species management. Endangered Species Update 8:1-5.
Clark, T.W., G.N. Backhouse, and R.C. Lacy. 1991b. Report of a workshop on population viability assessment as a
tool for threatened species management and conservation. Australian Zoologist 27:28-35.
Clark, T.W., G.N. Backhouse, and R.C. Lacy. 2002. The population viability assessment workshop: A tool for
threatened species management. Endangered Species Update 19(4):136-141.
Combreau, O., F. Launay, and M. Lawrence. 2001. An assessment of annual mortality rates in adult-sized migrant
houbara bustards (Chlamydotis (undulata) macqueenii). Animal Conservation 4(2):133-141.
Conservation Breeding Specialist Group (SSC/IUCN). 1996. Population and Habitat Viability Assessment for the
Striped Legless Lizard (Delma impar). Mosman, NSW: Australasian Regional Association of Zoological Parks
and Aquaria.
Conservation Breeding Specialist Group (SSC/IUCN). 2000. Conservation Assessment and Management Plan for
Arabian Carnivores and Population and Habitat Viability Assessment for the Arabian Leopard and Tahr. Apple
Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Conservation Breeding Specialist Group (SSC/IUCN). 2002. Evaluation et Plans de Gestion pour la Conservation
(CAMP) de la Faune de Madagascar: Lemuriens, Autres Mammiferes, Reptiles et Amphibiens, Poissons d’eau
douce et Evaluation de la Viabilite des Populations et des Habitats de Hypogeomys antimena (Vositse). Apple
Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Edroma, E.L., N. Rosen, and P.S. Miller (eds.). 1997. Conserving the Chimpanzees of Uganda: Population and
Habitat Viability Assessment for Pan troglodytes schweinfurthii. Apple Valley, MN: Conservation Breeding
Specialist Group (SSC/IUCN).
Ellis, S., and U.S. Seal. 1995. Tools of the trade to aid decision making for species survival. Biodiversity and
Conservation 4:553-572.
Ellis, S., K. Hughes, C. Kuehler, R.C. Lacy and U.S. Seal (eds.). 1992. Alala, Akohekohe, and Palila Population and
Habitat Viability Assessment Reports. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Ewins, P., M. de Almeida, P. Miller and O. Byers (eds.). 2000. Population and Habitat Viability Assessment
Workshop for the Wolves of Algonquin Park: Final Report. Apple Valley, MN: Conservation Breeding
Specialist Group (SSC/IUCN).
Fisher, A., E. Rominger, P. Miller and O. Byers. 1999. Population and Habitat Viability Assessment Workshop for
the Desert Bighorn Sheep of New Mexico (Ovis canadensis): Final Report. Apple Valley, MN: Conservation
Breeding Specialist Group (SSC/IUCN).
Foose, T.J., R.C. Lacy, R. Brett and U.S. Seal (eds.). 1993. Kenyan Black Rhino Metapopulation Workshop Report.
Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Forys, E.A. and S.R. Humphrey. 1999. Use of population viability analysis to evaluate management options for the
endangered Lower Keys marsh rabbit. Journal of Wildlife Management 63(1):251-260.
Galimberti, F., S. Sanvito, L. Boitani, and A. Fabiani, A. 2001. Viability of the southern elephant seal population of
the Falkland Islands. Animal Conservation 4(1): 81-88.
Gonzalez, L.M., B. Heredia, A. Araujo, I. Robinson, J. Worms, P.S. Miller, and U.S. Seal (eds.). 2002. Population
and Habitat Viability Assessment for the Mediterranean Monk Seal (Monachus monachus) in the Eastern
Atlantic. Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
González, S., M. Merino, M. Gimenez-Dixon, S. Ellis, and U.S. Seal (eds.). 1993. Population and Habitat Viability
Assessment for the Pampas Deer (Ozotoceros bezoarticus). Apple Valley, MN: Captive Breeding Specialist
Group (SSC/IUCN).
Appendix III 131
VORTEX Bibliography
VORTEX Version 9 User’s Manual
Guichard, C., S. Ellis, Y. Matamoros, and U.S. Seal (eds.). 2001. Análisis de la Viabilidad de Poblaciónal y del
Hábitat del Manatí en México. Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Gunn, A., U.S. Seal, and P.S. Miller (eds.). 1998. Population and Habitat Viability Assessment Workshop for Peary
Caribou and Arctic-Island Caribou (Rangifer tarandus). Apple Valley, MN: Conservation Breeding Specialist
Group (SSC/IUCN).
Haig, S.M. and J.D. Ballou. 2002. Pedigree analyses in wild populations. Pages 388-405 in: Beissinger, S.R. and
D.R. McCullough (eds.). Population Viability Analysis. Chicago, IL: University of Chicago Press.
Haig, S.M., J.R. Belthoff, and D.H. Allen. 1993. Population viability analysis for a small population of red-cockaded
woodpeckers and an evaluation of enhancement strategies. Conservation Biology 7:289-301.
Hamilton, S. and H. Moller. 1995. Can PVA models using computer packages offer useful conservation advice?
Sooty shearwaters Puffinus griseus in New Zealand: As a case study. Biological Conservation 73(2):107-117.
Heredia, B., P. Gaona, A. Vargas, S. Ellis, and U.S. Seal (eds.). 1999. Taller Análisis de la Viabilidad de Población
y del Habitat para el Lince Ibérico (Lynx pardinus). Apple Valley, MN: Conservation Breeding Specialist
Group (SSC/IUCN).
Herrero, S., P.S. Miller and U. S. Seal (eds.). 2000. Population and Habitat Viability Assessment Workshop for the
Grizzly Bear of the Central Rockies Ecosystem (Ursus arctos horribilis). Apple Valley, MN: Conservation
Breeding Specialist Group (SSC/IUCN).
Hosack, D.A., P.S. Miller, J.J. Hervert, and R.C. Lacy. 2002. A population viability analysis for the endangered
Sonoran pronghorn, Antilocapra americana sonoriensis. Mammalia 66(2):207-229.
Howells, O., and G.E. Jones. 1997. A feasibility study of reintroducing wild boar (Sus scrofa) to Scotland: Are
existing woodlands large enough to support minimum viable populations? Biological Conservation 81:77-89.
Jackson, S.M. 1999. Preliminary predictions of the impacts of habitat area and catastrophes on the viability of
Mahogany Glider Petaurus gracilis populations. Pacific Conservation Biology 5(1):56-62.
Jennings, M., R. Beiswinger, S. Corn, M. Parker, A. Pessier, B. Spencer, and P.S. Miller (eds.). 2001. Population
and Habitat Viability Assessment for the Wyoming Toad (Bufo baxteri). Apple Valley, MN: Conservation
Breeding Specialist Group (SSC/IUCN).
de Jong, Y. A., P. R. van Olst, and R. C. C. M. de Jong. 1997. Feasibility of reintroduction of the Eurasian lynx
(Lynx lynx) on 'De Veluwe', the Netherlands, by using the stochastic simulation programme VORTEX.
Zeitschrift fuerSaugetierkunde 62:44-51.
Kaiya, Z., S. Ellis, S. Leatherwood, M. Bruford, and U.S. Seal (eds.). 1994. Baiji Population and Habitat Viability
Assessment Report. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Kelly, B.T., P.S. Miller, and U.S. Seal (eds.). 1999. Population and Habitat Viability Assessment Workshop for the
Red Wolf (Canis rufus). Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Kjos, C., O. Byers, P.S. Miller, J. Borovansky, and U.S. Seal (eds.). 1998. Population and Habitat Viability
Assessment Workshop for the Winged Mapleleaf Mussel (Quadrula fragosa): Final Report. Apple Valley, MN:
Conservation Breeding Specialist Group (SSC/IUCN).
Kovács, T., Z. Korsós, I. Rehák, K. Corbett, and P.S. Miller (eds.). 2002. Population and Habitat Viability
Assessment for the Hungarian Meadow Viper (Vipera ursinii rakosiensis). Apple Valley, MN: Conservation
Breeding Specialist Group (SSC/IUCN).
Kumar, A., S. Molur, and S. Walker (eds.). 1995. Lion Tailed Macaque Population and Habitat Viability
Assessment Report. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Lacy, R.C. 1993. VORTEX: A computer simulation model for Population Viability Analysis. Wildlife Research 20:4565.
Lacy, R.C. 1993/1994. What is Population (and Habitat) Viability Analysis? Primate Conservation 14/15:27-33.
132 Appendix III
VORTEX Bibliography
VORTEX Version 9 User’s Manual
Lacy, R.C. 1996. Further population modelling of northern white rhinoceros under various management scenarios.
Appendix 3 in: Foose, T.J. (ed.). Summary – Northern White Rhinoceros Conservation Strategy Workshop.
Cumberland, OH: International Rhino Foundation.
Lacy, R.C. 2000. Structure of the VORTEX simulation model for population viability analysis. Ecological Bulletins
48:191-203.
Lacy, R.C. and T.W. Clark. 1990. Population viability assessment of the eastern barred bandicoot in Victoria. Pages
131-146 in:Clark, T.W. and J.H. Seebeck (eds.). The Management and Conservation of Small Populations.
Brookfield, IL: Chicago Zoological Society.
Lacy, R.C. and T.W. Clark. 1993. Simulation modeling of American marten (Martes americana) populations:
Vulnerability to extinction. Great Basin Naturalist 53:282-292.
Lacy, R.C, N.R. Flesness and U.S. Seal (eds.). 1989. Puerto Rican Parrot Population Viability Analysis. Report to
the U.S. Fish and Wildlife Service. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Lacy, R.C. and D.B. Lindenmayer. 1995. A simulation study of the impacts of population subdivision on the
mountain brushtail possum, Trichosurus caninus Ogilby (Phalangeridae: Marsupialia), in south-eastern
Australia. II. Loss of genetic variation within and between subpopulations. Biological Conservation 73:131142.
Lane, S.J. and J.C. Alonso. 2001. Status and extinction probabilities of great bustard (Otis tarda) leks in Andalucia,
southern Spain. Biodiversity & Conservation 10(6):893-910.
Lee, H., D. Garshelis, U.S. Seal, and J. Shillcox (eds.). 2001. Asiatic Black Bears PHVA: Final Report. Apple
Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Leighton, M; U.S. Seal, K. Soemarna, Adjisasmito, M. Wijaya, T.Mitra Setia, G. Shapiro, L. Perkins, K. TraylorHolzer, and R. Tilson. 1995. Orangutan life history and vortex analysis. Pages 97-107 in: Nadler, R.D., B.F.M.
Galdikas, L.K. Sheeran, and N. Rosen (eds.). The Neglected Ape. New York: Plenum Press.
de Leon, J., N. Lawas, R. Escalada, P. Ong, R. Callo, S. Hedges, J. Ballou, D. Armstrong, and U.S. Seal (eds.).
1996. Tamaraw (Bubalus mindorensis) Population and Habitat Viability Assessment Report. Apple Valley,
MN: Conservation Breeding Specialist Group (SSC/IUCN).
Li, X.H., D.M.Li, Y.G. Yong, and J. Zhang. 1997. A preliminary analysis on population viability analysis for giant
panda in Foping. Acta Zoologica Sinica 43(3):285-293.
Li, X. and D. Li. 1998. Current state and the future of the crested ibis (Nipponia nippon): A case study by population
viability analysis. Ecological Research 13(3):323-333.
Lindenmayer, D.B., M.A. Burgman, H.R. Akçakaya, R.C. Lacy, and H.P. Possingham. 1995. A review of the
generic computer programs ALEX, RAMAS/Space and VORTEX for modelling the viability of wildlife populations.
Ecological Modelling 82:161-174.
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.
Lindenmayer, D.B. and R.C. Lacy. 1995a. Metapopulation viability of Leadbeater's possum, Gymnobelideus
leadbeateri, in fragmented old-growth forests. Ecological Applications 5:164-182.
Lindenmayer, D.B. and R.C. Lacy. 1995b. Metapopulation viability of arboreal marsupials in fragmented oldgrowth forests: comparison among species. Ecological Applications 5:183-199.
Lindenmayer, D.B. and R.C. Lacy. 1995c. A simulation study of the impacts of population subdivision on the
mountain brushtail possum, Trichosurus caninus Ogilby (Phalangeridae: Marsupialia), in south-eastern
Australia. I. Demographic stability and population persistence. Biological Conservation 73:119-129.
Lindenmayer, D.B. and R.C. Lacy. 2002. Small mammals, habitat patches and PVA models: A field test of model
predictive ability. Biological Conservation 103(3):247-265.
Lindenmayer, D.B., R.C. Lacy, and M.L. Pope. 2000. Testing a simulation model for population viability analysis.
Ecological Applications 10:580-597.
Appendix III 133
VORTEX Bibliography
VORTEX Version 9 User’s Manual
Lindenmayer, D.B., R.C. Lacy, V.C. Thomas, and T.W. Clark. 1993. Predictions of the impacts of changes in
population size and environmental variability on Leadbeater's Possum, Gymnobelideus leadbeateri McCoy
(Marsupialia: Petauridae) using Population Viability Analysis: An application of the computer program
VORTEX. Wildlife Research 20:67-86.
Lindenmayer, D.B., V.C. Thomas, R.C. Lacy, and T.W. Clark. 1991. Population Viability Analysis (PVA): The
concept and its applications, with a case study of Leadbeater's Possum, Gymnobelideus leadbeateri McCoy.
Report to the Forest and Timber Inquiry (Resource Assessment Commission), Consultancy Series No.
FTC91/18, Canberra, Australia. 170 pp.
Lombard, K.B. 1996. A population viability analysis of the Hawaiian Monk Seal (Monachus schauinslandi). M. Sc.
dissertation. University of Michigan, Ann Arbor, MI.
Lorek, H. and M. Sonnenschein. 1999. Modelling and simulation software to support individual-based ecological
modelling. Ecological Modelling 115(2-3):199-216.
Maehr, D.S., R.C. Lacy, E.D. Land, O.L. Bass Jr., and T.S. Hoctor. 2002. Evolution of population viability
assessments for the Florida panther: A multiperspective approach. Pages 284-311 in: Beissinger, S.R. and D.R.
McCullough (eds.). Population Viability Analysis. Chicago, IL: University of Chicago Press.
Maguire, L.A., R.C. Lacy, R.J. Begg, and T.W. Clark. 1990. An analysis of alternative strategies for recovering the
eastern barred bandicoot in Victoria. Pages 147-164 in: Clark, T.W. and J.H. Seebeck (eds.). The Management
and Conservation of Small Populations. Brookfield, IL: Chicago Zoological Society.
Majluf, P., E.A. Babcock, J.C. Riveros, M.A. Schreiber, and W. Alderete. 2002. Catch and bycatch of sea birds and
marine mammals in the small-scale fishery of Punta San Juan, Peru. Conservation Biology 16(5):1333-1343.
Manansang, J., S. Hedges, D. Siswomartono, P.S. Miller, and U.S. Seal (eds.). 1996. Population and Habitat
Viability Assessment Workshop for the Anoa (Bubalus depressicornis and Bubalus quarlesi). Apple Valley, MN:
Conservation Breeding Specialist Group (SSC/IUCN).
Manansang, J., A. MacDonald, D. Siswomartono, P.S. Miller, and U.S. Seal (eds.). 1996. Population and Habitat
Viability Assessment for the Babirusa (Babyrousa babyrussa). Apple Valley, MN: Conservation Breeding
Specialist Group (SSC/IUCN).
Manansang, J., P.S. Miller, J.W. Grier, and U.S. Seal (eds.). 1997. Population and Habitat Viability Assessment for
the Javan Hawk-Eagle (Spizaetus bartelsi). Apple Valley, MN: Conservation Breeding Specialist Group
(SSC/IUCN).
Manansang, J. D. Siswomartono, T. Soejarto, U.S. Seal, P.S. Miller, and S. Ellis (eds.). 1997. Marine Turtles of
Indonesia: Population Viability and Conservation Assessment and Management Workshop. Apple Valley, MN:
Conservation Breeding Specialist Group (SSC/IUCN).
Marmontel, M., S. R. Humphrey and T. J. O'Shea. 1997. Population viability analysis of the Florida manatee
(Trichechus manatus latirostris), 1976-1991. Conservation Biology 11(2):467-481.
Marshall, K. and G. Edwards-Jones. 1998. Reintroducing capercaillie (Tetrao urogallus) into southern Scotland:
Identification of minimum viable populations at potential release sites. Biodiversity & Conservation 7(3):275296.
Matamoros, Y., G. Wong, and U.S. Seal (eds.). 1996. Population and Habitat Viability Assessment Workshop for
Saimiri oerstedi citrinellus. Final Report. Apple Valley, MN: Conservation Breeding Specialist Group
(SSC/IUCN).
Mathews, F. and D.W. Macdonald. 2001. The sustainability of the common crane (Grus grus) flock breeding in
Norfolk: Insights from simulation modeling. Biological Conservation 100(3):323-333.
McCann, K., A. Burke, L. Rodwell, M.Steinacker and U. S. Seal (eds.). 2000. Population and Habitat Viability
Assessment for the Wattled Crane (Bugeranus carunculatus) in South Africa. Apple Valley, MN: Conservation
Breeding Specialist Group (SSC/IUCN).
McCann, K., K. Morrison, A. Byers, P. Miller and Y. Friedmann (eds.). 2001. Blue Crane (Anthropoides
paradiseus): A Population and Habitat Viability Assessment Workshop. Apple Valley, MN: Conservation
Breeding Specialist Group (SSC/IUCN).
134 Appendix III
VORTEX Bibliography
VORTEX Version 9 User’s Manual
Miller, P.S. 1996. Impacts of the Hawaiian longline fishery on the Pacific Ocean population of loggerhead turtles
(Caretta caretta): An analysis using the VORTEX simulation modelling package. Pages 105-132 in: Bolten, A.B.,
J.A. Wetherall, G.H. Balazs, and S.G. Pooley (eds.). Status of Marine Turtles in the Pacific Ocean Relevant to
Incidental Take in the Hawaii-Based Pelagic Longline Fishery. U.S. Department of Commerce, NOAA
Technical Memorandum, NOAA-TM-NMFS-SWFCS-230, 167 p.
Miller, P. S. and R. C. Lacy. 1999. VORTEX Version 8 users manual. A stochastic simulation of the simulation
process. Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Mills, L.S., S.G. Hayes, C. Baldwin, M.J. Wisdom, J. Citta, D.J. Mattson, and K. Murphy. 1996. Factors leading to
different viability predictions for a grizzly bear data set. Conservation Biology 10:863-873.
Mills, M.G.L, S. Ellis, R. Woodroffe, A. Maddock, P. Stander, A. Pole, G. Rasmussen, P. Fletcher, M. Bruford, D.
Wildt, D. Macdonald, and U.S. Seal (eds.). 1998. Population and Habitat Viability Assessment for the African
Wild Dog (Lycaon pictus) in Southern Africa. Final Workshop Report. Apple Valley, MN: Conservation
Breeding Specialist Group (SSC/IUCN).
Mirande, C., R. Lacy, and U. Seal (eds.). 1991. Whooping Crane (Grus americana) Conservation Viability
Assessment Workshop Report. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Molur, S., R. Sukumar, and S. Walker (eds.). 1995. Great Indian One-Horned Rhinoceros Population and Habitat
Viability Assessment Report. Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Novellie, P.A., P.S. Miller, and P.H. Lloyd. 1996. The use of VORTEX simulation models in a long-term programme
of re-introduction of an endangered large mammal, the Cape mountain zebra (Equus zebra zebra). Acta
Œcologica 17:657-671.
Odum, A., et al (eds.). 1993. Aruba Island Rattlesnake Population and Habitat Viability Assessment (PHVA)
Workshop. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Ounsted, M., K. Soemarna, W. Ramono, U.S. Seal, A. Green, Rudyanto, and R. Ounsted (eds.). 1994. The WhiteWinged Wood Duck in Sumatra: Population and Habitat Viability Analysis Report. Apple Valley, MN: Captive
Breeding Specialist Group (SSC/IUCN).
Paolo, C., and L. Boitani. 1991. Viability assessment of the Italian wolf and guidelines for the management of the
wild and a captive population. (In Italian). Ricerche de Biologia della Selvaggina 89:1-58.
Parysow, P. and D.J. Tazik. 2002. Assessing the effect of estimation error on population viability analysis: an
example using the black-capped vireo. Ecological Modelling 155(2-3):217-229.
Pergams, O.R.W. 1998. Genetic, morphological, and population viability analysis of California Channel Island deer
mice. M.Sc. thesis, University of Illinois at Chicago.
Phillips, M., N. Fascione, P.Miller and O. Byers (eds.). 2000. Wolves in the Southern Rockies: A Population and
Habitat Viability Assessment: Final Report. Apple Valley, MN: Conservation Breeding Specialist Group
(SSC/IUCN).
Pinder, L., and U.S. Seal (eds.). 1994. Population and Habitat Viability Assessment Report for Cervo-do-Pantanal
(Blastocerus dichotomus). Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Plissner, J.H. and S.M. Haig. 2000. Viability of piping plover Charadrius melodus metapopulations. Biological
Conservation 92(2):163-173.
Portales, G.L., P. Reyes, H. Rangel, A. Velazquez, P.S. Miller, S. Ellis, and A.T. Smith (eds.). 1997. Taller
Internacional para la Conservación de los Lagomorfos Mexicanos en Peligro de Extinción. Reporte del Taller.
Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Pucek, Z., I. Udina, U.S. Seal, and P.S. Miller (eds.). 1996. Population and Habitat Viability Assessment for the
European Bison (Bison bonasus). Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Rao, R.J., D. Basu, S.M. Hasan, S. Molur, and S. Walker (eds.). 1995. Indian Gharial Population and Habitat
Viability Assessment Report. Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Appendix III 135
VORTEX Bibliography
VORTEX Version 9 User’s Manual
Reed, D.H., J.J. O’Grady, B.W. Brook, J.D. Ballou, and R. Frankham. 2003. Estimates of minimum viable
population sizes for vertebrates and factors influencing those estimates. Biological Conservation 113:23-34.
Rodríguez, J.P., and F. Rojas-Suárez. 1994. Population viability analysis of three populations of insular psitacids of
Venezuela. Pages 97-113 in: Morales, G., I. Novo, D. Bigio, A. Luy, and F. Rojas-Suárez (eds.). Biologia y
Conservacion de los Psitacidos de Venezuela. Caracas: SCAV, EBAFY, Econatura, SCAPNHP and Provita.
Rodríguez-Luna, E., L. Cortés Ortiz, P.S. Miller, and S. Ellis (eds.). 1996. Population and Habitat Viability
Assessment for the Mantled Howler Monkey (Alouatta palliata mexicana). Apple Valley, MN: Conservation
Breeding Specialist Group (SSC/IUCN).
Rylands, A., K. Strier, R. Mittermeier, J. Borovansky, and U.S. Seal (eds.). 1998. Population and Habitat Viability
Assessment Workshop for the Muriqui (Brachyteles arachnoides). Apple Valley, MN: Conservation Breeding
Specialist Group (SSC/IUCN).
Seal, U.S. (ed.). 1992a. Bali Starling Population Viability Assessment Report. Apple Valley, MN: Captive Breeding
Specialist Group (SSC/IUCN).
Seal, U.S. (ed.). 1992b. Black-Footed Ferret Recovery Plan Review. Apple Valley, MN: Captive Breeding Specialist
Group (SSC/IUCN).
Seal, U.S. (ed.). 1992c. Genetic Management Strategies and Population Viability of the Florida Panther (Felis
concolor coryi). Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Seal, U.S. (ed.). 1994a. Attwater’s Prairie Chicken Population and Habitat Viability Assessment. Apple Valley,
MN: Captive Breeding Specialist Group (SSC/IUCN).
Seal, U.S. (ed.). 1994b. Houston Toad Population and Habitat Viability Assessment Report. Apple Valley, MN:
Conservation Breeding Specialist Group (SSC/IUCN).
Seal, U.S. (ed.). 1994c. Population and Habitat Viability Assessment for the Greek Population of the Mediterranean
Monk Seal (Monachus monachus). Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Seal, U.S. (ed.). 1996. Kirtland’s Warbler (Dendroica kirtlandii) Population and Habitat Viability Assessment
Workshop Report. Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Seal, U.S., J.D. Ballou and C.V. Padua (eds.). 1990. Leontopithecus Population Viability Analysis Workshop Report.
Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Seal, U.S., and M.W. Bruford (eds.). 1991. Pink Pigeon (Columba [Nesoenas] mayeri) Conservation Viability
Assessment Workshop. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Seal, U.S., and T. Foose (eds.). 1989. Javan Rhinoceros Population Viability Analysis. Apple Valley, MN: Captive
Breeding Specialist Group (SSC/IUCN).
Seal, U.S., P. Garland, D. Butler, A. Grant, and C. O’Donnell (eds.). 1993. Population Viability Analysis for the Kea
(Nestor notabilis) and Kaka (Nestor meridionalis). Apple Valley, MN: Captive Breeding Specialist Group
(SSC/IUCN).
Seal, U.S., and S. Hereford (eds.). 1992. Mississippi Sandhill Crane Population and Habitat Viability Assessment
Report. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Seal, U.S. and R.C. Lacy (eds.). 1989. Florida Panther Population Viability Analysis. Report to the U.S. Fish and
Wildlife Service. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Seal, U.S. and R.C. Lacy. 1990. Florida Key Deer (Odocoileus virginianus clavium) Population Viability
Assessment. Report to the U.S. Fish and Wildlife Service. Apple Valley, MN: Captive Breeding Specialist
Group (SSC/IUCN).
Seal, U.S., R.C. Lacy, K. Medley, R. Seal and T.J. Foose (eds.). 1991. Tana River Primate Reserve Conservation
Assessment Workshop Report. Apple Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Seal, U.S., J. Manansang, D. Siswomartono, T. Suhartono, J. Sugarjito (eds.). 1996. Komodo Monitor (Varanus
komodoensis) Population and Habitat Viability Assessment Workshop Report. Apple Valley, MN: Conservation
Breeding Specialist Group (SSC/IUCN).
136 Appendix III
VORTEX Bibliography
VORTEX Version 9 User’s Manual
Seal, U.S., S. Walker, and S. Molur (eds.). 1995. Barasingha Population and Habitat Viability Assessment Report.
Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Sillero-Zubiri, C., J. Malcolm, S. Williams, J. Marino, Z. Tefera, K. Laurenson, D. Gottelli, A. Hood, D.
Macdonald, D. Wildt, and S. Ellis. 2000. Ethiopian Wolf Conservation Strategy Workshop: Final Report. Apple
Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Soberon, R., P. Ross and U.S. Seal (eds.). 2000. Cocodrilo Cubano Análisis de la Viabilidad de la Población y del
Hábitat: Borrador del Informe. Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Soemarna, K., R. Tilson, W. Ramono, D.W. Sinaga, R. Sukumar, T.J. Foose, K. Traylor-Holzer, and U.S. Seal
(eds.). 1994. The Sumatran Rhino in Indonesia: Population and Habitat Viability Analysis Report. Apple
Valley, MN: Captive Breeding Specialist Group (SSC/IUCN).
Somers, M.J. 1997. The sustainability of harvesting a warthog population: Assessment of management opinions
using simulation modelling. South African Journal of Wildlife Research 27(2):37-43.
Sommer, S., A.T. Volahy, and U.S. Seal. 2002. A population and habitat viability assessment for the highly
endangered giant jumping rat (Hypogeomys antimena), the largest extant endemic rodent of Madagascar.
Animal Conservation 5(4):263-273.
Song, Y.L. 1996. Population viability analysis for two isolated populations of Haianan Eld’s deer. Conservation
Biology 10:1467-1472.
South, A.B., S.P. Rushton, and D.W. Macdonald. 2000. Simulating the proposed reintroduction of the European
beaver Castor fiber to Scotland. Biological Conservation 93:103-116..
Strier, K. B. 2000. Population viabilities and conservation implications for muriquis (Brachyteles arachnoides) in
Brazil's Atlantic Forest. Biotropica 32(4B):903-913.
Supriatna, J., R. Tilson, K. Gurmaya, J. Manansang, W. Wardojo, A. Sriyanto, A. Teare, K. Castle, and U.S. Seal
(eds.). 1994. Javan Gibbon and Javan Langur Population and Habitat Viability Analysis Report. Apple Valley,
MN: Conservation Breeding Specialist Group (SSC/IUCN).
Taylor, B.L., P.R. Wade, U. Ramakrishnan, M. Gilpin, and H.R. Akçakaya. 2002. Incorporating uncertainty in
population viability analyses for the purpose of classifying species by risk. Pages 239-252 in: Beissinger, S.R.
and D.R. McCullough (eds.). Population Viability Analysis. Chicago, IL: University of Chicago Press.
Tilson, R., U.S. Seal, K. Soemarna, W. Ramono, E. Sumardja, S. Poniran, C. van Schaik, M. Leighton, H. Rijksen,
and A. Eudey (eds.). 1993. Orang utan Population and Habitat Viability Analysis Report. Apple Valley, MN:
Captive Breeding Specialist Group (SSC/IUCN).
Tilson, R., K. Soemarna, W. Ramono, S. Lusli, K. Traylor-Holzer, and U.S. Seal (eds.). 1992. Sumatran Tiger
Population and Habitat Viability Analysis Report. Apple Valley, MN: Captive Breeding Specialist Group
(SSC/IUCN).
Tilson, R., K. Soemarna, W. Ramono, R. Sukumar, U.S. Seal, K. Traylor-Holzer, and C. Santiapillai (eds.). 1994.
The Asian Elephant in Sumatra: Population and Habitat Viability Analysis Report. Apple Valley, MN: Captive
Breeding Specialist Group (SSC/IUCN).
Tunhikorn, S., W. Brockelman, R. Tilson, U. Nimmanheminda, P. Ratankorn, R. Cook, A. Teare, K. Castle, and
U.S. Seal (eds.). 1994. Population and Habitat Viability Analysis Report for Thai Gibbons: Hylobates lar and
H. pileatus. Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Wade, P.R. 2002. Bayesian population viability analysis. Pages 213-238 in: Beissinger, S.R. and D.R. McCullough
(eds.). Population Viability Analysis. Chicago, IL: University of Chicago Press.
Walker, S. (ed.). 1994. Manipur Brow-Antlered Deer (Sangai) Population and Habitat Viability Assessment. Apple
Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Walker, S. and S. Molur (eds.). 1994. Population and Habitat Viability Analysis (PHVA) Workshop for
Indian/Nepali Rhinoceros. Zoo Outreach Organisation/CBSG India, Coimbatore.
Appendix III 137
VORTEX Bibliography
VORTEX Version 9 User’s Manual
Wang, Y., S.-W. Chu, and U.S. Seal (eds.). 1994. Population and Habitat Viability Assessment Workshop Report for
the Asiatic Black Bear (Ursus thibetanus formosanus). Apple Valley, MN: Conservation Breeding Specialist
Group (SSC/IUCN).
Wang, Y., S.-W. Chu, D. Wildt, and U.S. Seal (eds.). 1995. Clouded Leopard (Neofelis nebulosa brachyurus)
Population and Habitat Viability Assessment. Apple Valley, MN: Conservation Breeding Specialist Group
(SSC/IUCN).
Wei, F., Feng, Z. and Hu, J. 1997. Population viability analysis computer model of giant panda population in
Wuyipeng, Wolong Natural Reserve, China. International Conference on Bear Research and Management
9(2):19-23.
Wemmer, C., A. Than, S.T. Khaing, S. Monfort, T. Allendorf, J. Ballou, and S. Ellis. 2000. Thamin Population and
Habitat Viability Assessment: Final Report. Apple Valley, MN: Conservation Breeding Specialist Group
(SSC/IUCN).
Werikhe, S., L. Macfie, N. Rosen, and P.S. Miller (eds.). 1998. Can the Mountain Gorilla Survive? Population and
Habitat Viability Assessment Workshop for Gorilla gorilla beringei. Apple Valley, MN: Conservation Breeding
Specialist Group (SSC/IUCN).
Whittington, P., R.J.M. Crawford, O. Huyser, D. Oschadleus, R. Randall, P. Ryan, L. Shannon, A. Woolfardt, J.
Cooper, R. Lacy, and S. Ellis (eds.). 2000. African Penguin Population and Habitat Viability Assessment. Final
Report. Apple Valley, MN: Conservation Breeding Specialist Group (SSC/IUCN).
Wilson, M.H., C.B. Kepler, N.F.R. Synder, S.R. Derrickson, F.J. Dein, J.W. Wiley, J.M. Wunderle, A.E. Lugo, D.L.
Graham, and W.D. Toone. 1994. Puerto Rican parrots and potential limitations of the metapopulation approach
to species conservation. Conservation Biology 8:114-123.
Xu, H., and H. Lu. 1996. A preliminary analysis of population viability for Chinese water deer (Hydropotes inermis)
lived in Yancheng. (In Chinese). Acta Theriologica Sinica 16:81-88.
Yuan, S.D., G. Yazhen, L. Xiaoping, Q. Yang, J. Sale, C. Kirkpatrick, J. Ballou, and U.S. Seal (eds.). 1999. CBSG
Guizhou Snub-nosed Monkey Conservation and PHVA Workshop Report. Apple Valley, MN: Conservation
Breeding Specialist Group (SSC/IUCN).
Zhang, X., D. Wang, and K. Wang. 1994. The VORTEX model and its application on the management of the Chinese
River Dolphin (Lipotes vexillifer) population. Chinese Biodiversity 2:133-139.
138 Appendix III
VORTEX Bibliography
VORTEX Version 9 User’s Manual
Appendix
IV
Reprints
Lacy, R.C. 2000. Structure of the VORTEX simulation model for
population viability analysis. Ecological Bulletins 48:191-203.
(Reprinted with permission of the publisher.)
Lacy, R.C. 2000. Considering threats to the viability of small
populations using individual-based models. Ecological Bulletins
48:39-51.
(Reprinted with permission of the publisher.)
Appendix IV 139
Reprints
Download PDF