When researching changing ecosystems, we found an excellent opportunity for computer modeling in Yellowstone Park. Management had recently re-introduced wolves to the park, and significant changes in animal populations and habitats occurred as a result. Since many of these changes were the result of predator-prey relationships, we thought that Stella would be a useful device in replicating the Yellowstone ecosystem through computer modeling. By experimenting with multiple variables, we created a computer model that resembled the ecosystem in Yellowstone Park.

BACKGROUND

        Yellowstone National Park is the United States’ oldest and largest national park.  Established in 1872, it is home to one of the country’s best preserved ecosystems, with animals co-existing as they did hundreds of years ago.  The park is well known for its rare and controversial management policies, which try to stay as true to nature as possible.  This includes allowing all forest fires that don’t threaten human structures burn, as they did in 1988 when they scorched more than half the park.  Those fires caused a great debate about humans’ role in managing ‘natural’ ecosystems.  In the next few years, the fantastic resurgence of the forest and its main tree, the lodge pole pine- which reproduces predominantly through instances of extreme heat that cause its seeds to burst from clumps of sap- led people to accept the notion of a system so complex, it naturally regulated itself.  This laid the groundwork for the re-introduction of Canadian wolves in 1995, a species that had been exterminated by park rangers some 70 years earlier.

        The re-introduction of the wolves was a controversial philosophical and legal debate.  Nearby ranchers worried about attacks on cattle, park management was concerned about the possible effects on tourism, and environmentalists questioned whether any species, regardless of its former position, could integrate and adapt into a running ecosystem.  Nonetheless, the consensus among most people related to the park was one of apprehensive support.  This was one of the first re-introduction campaigns of such magnitude, and no one could be sure what was going to take place.

        Ecologists used computer modeling to make predictions about what would happen in the park, with large studies taking place in 1990 and 1994 at Princeton University and the University of Idaho.  The models were instrumental in producing support for the initiative and in the planning process itself.  Dr. Robert Crabtree, of the Yellowstone Ecosystem Studies (YES) institute, was one of the head research ecologists.  From his research and computer modeling he predicted three basic results of the reintroduction:

1.  Elk populations would decrease from direct and indirect wolf effects.
2. Wolves would act as obligate killers and provide a plethora of carrion.
3. Wolf expansion and carrion would benefit many other species.

        Thirty-three Canadian wolves were reintroduced to the Yellowstone ecosystem in 1995.  The results came rapidly and clearly.  Within two years their numbers grew to 97, and their effects were felt in most spheres of the park.  Wolves killed half the coyote population (killing 44), increasing the populations of coyote prey like deer and ground squirrels.  The increased prey populations benefited middle level predators like foxes, hawks, owls and pine martens, increasing species diversity and balance.  Ironically, the coyotes that did survive thrived on the edges of wolf habitats, benefiting from carcasses wolves left behind.  Such carcasses benefited not only the wolves, but also many scavengers.  The number of ravens doubled, and eagles increased fivefold.  Lastly the wolves have begun to solve one of the growing problems in the park, the abundance of the elk population.  Few factors besides space, which tends to create huge carrying capacities, were keeping the elk in check.  As their numbers reached 20,000, park management feared they may have to restrict their growth with human means.  The wolves actually curved their rapid growth, killing 124 during 1996.

        Many scientists who have observed the recent developments agree with Dr. Crabtree’s assertion that this is the most successful reintroduction of a species ever.  The results are causing national park and forest management across the country to re-examine their policies, and are giving greater credibility to the sentiments that mother nature may really know best.  Even in the wake of such success, there are those who still challenge the presence of the wolves, citing monetary, legal or long-term concerns.  Federal Judge William Downes ruled in December of 1997 that the Federal recovery program for the wolves was illegal because it did not provide enough protection for wolves that returned to Canada alone.  And many ranchers complain the presence of the wolves costs them in added security, even though the terms of the current agreement provide for monetary restitution for any cattle killed by wolves.  Unfortunately the wolves future is uncertain, but even in their short time at Yellowstone they have taught valuable lessons about ecosystems and predator-prey relationships.

METHODS AND RESULTS

        Our objective in using Stella was to make a computer model of the ecosystem in Yellowstone. We tried to use numbers in our model that were close to the actual numbers in the Yellowstone ecosystem.  We had such figures for the overall populations of the wolves, coyotes and elk, but we had to estimate for the deer and ground squirrels, as well as for the birth and death rates of most animals. We operated under a timetable of 4 years, using months as our unit of measurement.

        When laying the groundwork for our model we established a few assumptions.  Though some are great simplifications of reality, they are a necesary part of computer modeling.  As we learned throughout the semester, a simplified model can still function, teach and even make acurate predictions about a real system.  Our assumptions were:

1. Predators eat all animals that die.
2. Weather and seasonal changes have no effect on the animals in the system.
3. The diet of deer and ground squirrels is a constant related to birth rate, hence unaccounted for.
4. The dietary and hunting habits of predators do not deviate from a strict routine.

To begin the modeling process, we created a self-sufficient coyote  population similar to that of the guppies in the lab from week one.  Its graph, as we suspected, was exponential because there were no factors keeping the population in check, and the birth rate alone  was high enough to sustain life.

  

The coyote was the chief predator in Yellowstone before the wolves entered the  ecosystem.  We introduced deer as an organism that would serve as the coyote's  prey.  Before we connected the deer population to the coyote population, we tested  them individually to see if they would produce an exponential curve. We applied this test to every organism before we entered it into the model. We  had little difficulty  obtaining a predator-prey graph where the predator and prey populations fluctuated  at similar times, the predator population always remaining a little behind the prey population.

        At this point, we were at the same level as our lab explorations, and any new organisms we wanted in our computer ecosystem would add new dimensions that we had, hitherto, not explored.
  

We decided to add another prey to the system - the ground squirrel.  After testing out different predation rates and birth rates for the three  organisms, we obtained a graph that displayed the prey populations  fluctuating at different points from each other and from the predator  population. Although the prey populations are not exactly in sync,  they both experience highs and lows at the same points; whereas,  the predator population reached highs when the prey hit lows.

We introduced the raven next to see how a scavenger would fit into our ecosystem. Because it does not kill any organisms, but instead eats the carcass of dead animals,  we connected them to the deer death convertor. Since squirrels do not leave much of a carcass, we did not connect them to the ravens.

Finally, we added the elk to our ecosystem; however, we did not connect them to the other organisms because in the actual Yellowstone ecosystem, the elk do not interact with any of the other organisms in predator-prey relationships. To simulate the park's natural space limitations, and to inhibit the elk from growing exponentially, we included a carrying capacity of 21,000. Our initial number for elk was 17,000.

              This was the state of the Yellowstone ecosystem before the wolves were reintroduced. Adding the wolves increased the complexity of
the ecosystem substantially - much more than adding any of the other animals did. We had three results we hoped to achieve, which were predicted by ecological computer modeling prior to 1995, and occurred in Yellowstone through the introduction of the wolves:

1. A reduction of the coyote population by half.
2. An increase in the populations of ground squirrels and deer, both of which are prey of the coyote.
3. An increase in the scavenger population, represented by the ravens.

We began, as we had with all the other animals, by establishing the wolf  population independently. (The graph was exponential). At first, we tried to connect the wolves to the elk population. We were able to establish a  functionning predator-prey relationship, however, when we tried to  introduce the coyotes as an alternate prey, all three organisms crashed.

        Initially, we did not completely accomplish our objectives, so we revamped the wolf introduction process. First, we connected the wolves to both the elk and coyote populations before we began to experiment with different physiological efficiencies, birth rates, and predation rates. After trying various combinations of increases and decreases of rates, we found the figures that worked the best and produced an ecosystem that closely resembled that of Yellowstone Park. The population of coyotes was reduced by roughly half and the populations of the coyotes' prey, ground squirrels and deer, increased as well. By connecting ravens to the death convertor for elk, we increased the raven population in our Stella model. Elk carcass, the result of wolf predation, provided a new source of food for the ravens.  Perhaps most impressively, we produced wolf, coyote and elk populations that mirrored the new changes almost exactly.

        Despite the inate simplicity of computer modeling, our project bettered our understanding of the complexity of natural ecosystems.  We succesfuly simulated predator-prey relationships and the effects of a keystone species in the Yellowstone Ecosystem through our Stella models.  Based on our models and research, we predict that:

1. The wolves will grow to hold elk populations in check
2. The Coyote population will continue to decrease, but avoid extinction, due to spatial heterogenity and their adaptation to a scavenging subsistence.
3. Biodiveristy rates continue to increase in response to the wolves, though at a slower pace.

LINKS

http://www.wolftracker.com/  A scientific site devoted to the study of the wolves of Yellowstone.
http://www.yellowstone.org/Yellowstone Ecosystem Studies (YES) homepage.
http://www.greateryellowstone.org/  The Greater Yellowstone Coalition's homepage, with news and political articles..
http://www.yellowstone.net/  One of the most comhprehensive sites on Yellowstone National Park
http://www.poky.srv.net/~jjmrm/  An incredibly comprehensive site devoted completely to the Yellowstone wolves.
http://www.cco.caltech.edu/~salmon/wh-yellowstone.html  A general directory to Yellowstone National Park.
http://www.hps-inc.com/stella.html  Stella's computer modeling homepage.
http:www.sprl.umich.edu/GCL/  The University of Michigan's Global Change Class Homepage.

BIBLIOGRAPHY

Crabtree, Bob. 'Total Impact,' The Yellowstone Wolf Tracker. 1997.
Robbins, Jim. 'In 2 Years, Wolves Reshape Yellowstone,' The New York Times. December 1, 1997
Stevens, William. 'Debating the Nature of Nature in Yellowstone,' The New York Times. June 2, 1998
Informal correspondance with the office of Dr. Bob Crabtree at the Yellowstone Ecosystem's Studies headquarters, November 5-
November 30.
Internet research- please see all links listed above.
 

By Ari Melber and Matt Abelson    Copywrite December, 1998