Wildl. Soc. Bull. 21:356-357, 1993

WILDLIFE SOFTWARE

PUMA: A POPULATION SIMULATOR FOR
COUGAR CONSERVATION

PAUL BEIER, School of Forestry, Northern Arizona University, Flagstaff, AZ 86011-4098

     The computer program PUMA simulates the population dynamics of cougar (Felis
concolor) populations and quantifies the risk of extinction for small populations under
various scenarios. It is intended as a teaching tool for classes in conservation
biology, population ecology, or wildlife management.
     The user controls both the land-use decisions and the biological conditions
under which the program operates. Land-use decisions are incorporated by allowing the
user to specify the areas of 2 different habitat types (within the area used by a
single cougar population) and to specify the level of immigration into the population.
The latter corresponds to the availability of a wildlife movement corridor to allow
such immigration. Because biological parameters (e.g., carrying capacity, survival
rates) are difficult to estimate and may vary geographically, the user also may specify
these parameters. By using the model, the user can examine the importance of both
biological factors and land-use decisions.
     Beier (1993) summarized the results of several thousand simulations under a
variety of assumptions. Persons wanting answers can more easily obtain them from that
paper than by using this program, which I designed as an educational tool. The ability
to explicitly model habitat loss and {fragmentation makes the model particularly
relevant to classes in conservation biology. I wrote the program out of concern for
a specific cougar population threatened by changes in land use in southern California
(Beier 1993). The cougar-specific features (below) and the applicability of the model
to a real-world situation give the model an immediacy that should stimulate student
interest in exercising the model.
     The program is menu-driven. Before being allowed to change any variable, the
user is given a brief definition of that variable and how it is used by the program.
During each run, a graph depicting numbers of adult males, adult females, and cubs as
a function of year is displayed on the screen. After all simulations under a given
scenario have run, the user can select any of 11 different graphs depicting the
results, or can inspect an ASCII file with detailed results in matrix form.
 Unlike most commercial programs, PUMA has many routines that specifically model the
dynamics of cougar populations, especially small populations, with density-dependence
in both mortality (especially of juveniles) and reproduction (less important). The
following are examples of features that give the user a sense that the model is
relevant to real populations.

1. The model incorporates an "Allee effect" (Begon and Mortimer 1981:29), whereby at
     low population sizes and skewed sex ratios, some females may not be bred. Some
     generic models track only females and ignore males, allowing reproduction when
     adult males are absent.
2. When a litter dies, the mother's interbirth interval is reduced to 12 months (vs.
     24 months when the cubs survive to independence).
3. Litters orphaned before 6 months of age do
     not survive.

     The program also incorporates stochastic variation in survival rates, litter
size, primary sex ratio, and other biological parameters. The stochastic routines were
designed to function realistically even when there are only 1-2 individuals in an
age-sex class. The user also can specify the duration, frequency, and intensity of
catastrophes (e.g., droughts, a series of hard winters) that periodically reduce
ungulate numbers and thus carrying capacity for cougars. The user can select a model
with survival rates independent of density, or any of 4 models in which survival rates
vary with density.
     The model has several limitations. Perhaps the most important is that the model
ignores inbreeding effects. To the extent that inbreeding in small populations
increases the risk of extinction, that risk is underestimated by the program. Also,
the program only simulates a single population; a logical extension would be to link
several such populations to form a meta-population model. Finally, once specified by
the user, the duration and frequency of catastrophes are fixed and do not fluctuate
stochastically.
     PUMA is written in Turbo Pascal 5.0, and runs on any IBM-compatible machine; a
color monitor is helpful. The program executable code and user documentation are
available on the Bird Monitor bulletin board (301-498-0402) or from the author.
     Acknowledgments. - While developing this program, I was supported by California
Agricultural Experiment Station Project 4326-MS, California Department of Fish and
Game, and the County of Orange, California. R. H. Barrett, T. L. George, F. C. Dean,
E. A. Rexstad, and an anonymous reviewer offered helpful comments both on the program
and this paper.

                              LITERATURE CITED


BEGON, M., AND M. MORTIMER. 1981. Population ecology. Sinauer Assoc., Sunderland, Mass.
     200pp.

BEIER, P. 1993. Determining minimum habitat areas and corridors for cougars. Conserv.
     Biol. 7:94-108.

Received 25 February 1992.
Accepted 17 December 1992.
Software Editor: Rexstad.