by Robert A. Garrott, P.J. White, Claire Gower, Matthew S. Becker, Shana Dunkley, Ken L. Hamlin, & Fred G.R. Watson
The effects of wolves on elk in the Greater Yellowstone Ecosystem have been contested among laypersons, politicians, and scientists—with some claiming devastation, others suggesting healing restoration, and most seeing something in between. In 1991, Montana State University initiated a study of about 400 to 600 elk inhabiting the Madison headwaters area in the west-central portion of Yellowstone National Park. The elk herd was nonmigratory and remained within the park year-round; therefore, the animals were not subject to harvest by human hunters. This high-elevation area has complex terrain, accumulates deep snow, and supports a mosaic of habitats including large tracts of burned and unburned forests interspersed with geothermal areas, meadows, rivers, and small lakes. The area is also an important winter range for bison that seasonally migrate west from their summer range in Hayden Valley. Prior to wolf restoration, coyotes were the only abundant mammalian predator, with some grizzly bears during spring and a few mountain lions. The study was initiated seven years before reintroduced wolves recolonized this portion of the park and continued thereafter, providing a rare opportunity to compare the responses of individual elk and the population as a whole to the restoration of a top predator that had been absent for approximately 70 years.
The protocol for the study was based on maintaining a representative sample of radio-collared female elk with biologists conducting extensive field work from November to May each year to monitor their behavior, nutrition, movements, pregnancy, survival, and population trends in response to forage, snow, predators, and other conditions. From 1991 to 2009, scientists amassed more than 12,000 person-days of field work and evaluated 15,000 observation periods of elk groups; 6,500 snow urine samples for assessing elk nutrition; 2,000 serum and fecal samples for assessing elk pregnancy; 1,000 plant samples for assessing biomass and nutrition; 17,000 measurements of vegetation; 4,175 kilometers (2,594 miles) of snow tracking along wolf trails; and 750 carcasses of ungulates killed by wolves. Also, 4,300 snow cores and more than 24,000 hours of wind data were collected to model spatial and temporal dynamics of the snow pack.
Prior to wolf restoration, the probability of an elk dying was related to its age, body condition, and snow pack. The primary cause of death was starvation, with younger and older elk more likely to die than elk in the prime of their life (3-9 years old) that have uniformly high survival rates. Elk rely on their teeth to obtain and break up plant materials, which are further broken down by microbes in their four-chambered stomachs to obtain energy and protein. Teeth wear with age, so older elk become less efficient at obtaining nutrients and accumulating the fat and protein reserves needed to survive winter when the availability of nutritious foods is low. This is especially true in the Madison headwaters region where high concentrations of silica in the soils and fluoride in the waters accelerate tooth wear—thereby leading to a shortened life span compared to elk in other areas. In addition, calves are smaller in body size, and as a result have smaller stores of fat and protein to metabolize during winter when forage was scarce. Deep, prolonged, or hard snow conditions also increased the risk of starvation of young and old elk by limiting access to forage under the snow and requiring more energy for them to forage and move about the landscape. As a result, the proportion of elk in the population dying from starvation each winter varied among years depending on winter severity. However, elk that frequently used geothermal areas (where heat from the interior of the earth reduced or eliminated snow pack) were less vulnerable to starvation.
Wolves recolonizing the Madison headwaters area strongly preferred elk as prey and killed comparatively fewer bison, even though bison were more abundant than elk from midwinter through spring. Bison kills were more frequent during late winter when animals were in poorer condition. The wolves' preference for elk probably reflects the formidable challenge of killing bison, which form groups to aggressively and cooperatively defend themselves and their young. In contrast, elk do not use group defenses and generally flee when attacked. Wolves strongly selected calves and older elk, which are the age classes most vulnerable to starvation mortality during winters of average to severe snow pack. However, the survival of elk calves was lower and less variable among years after wolf numbers increased, suggesting predation limited the recruitment of animals into the breeding population. The survival of adult female elk was 5-15% lower following wolf recolonization, primarily in the middle to older age classes. The diets and nutrition of elk remained similar to those prior to the arrival of wolves. Elk pregnancy rates remained high, but elk abundance decreased rapidly as breeding females were killed and wolf predation on calves consistently reduced recruitment to low levels. As elk numbers decreased due to wolf predation, wolf kill rates remained high and wolf numbers continued to grow. As a result, predation removed a higher portion of the elk population each year until elk became scarce. Thereafter, wolf kill rates decreased, strife among packs increased, wolf numbers declined, and packs began to hunt elsewhere for most of the year.
After wolves established in the Madison headwaters, the probability of an elk dying was strongly influenced by factors other than its physical condition, including characteristics of the landscape and weather that increased its susceptibility to predation by wolves. Elk at higher elevations with deeper snows were more likely to be killed by wolves, as were elk in thermal areas or meadows where they could be chased into habitat boundaries of deeper snow or burned timber with down-fall that impeded their escape. Conversely, elk on steep slopes with shallow snow and good visibility, or in areas where they could quickly escape to deep, swift, and wide rivers after encountering wolves, were less vulnerable to predation. As a result, in less than two decades, elk went from being numerous (~400-600 individuals) and broadly distributed throughout the Gibbon, Firehole, and Madison drainages during winter to scarce (less than 25 individuals) and constrained to relatively small refuges in the Madison drainage where they were more likely to observe approaching wolves and escape if detected and attacked. Wolves killed nearly all of the elk in the Firehole and Gibbon drainages where susceptibility to predation was high. Many of these elk were strong and in good condition, but were caught in "terrain traps" where they were unable to flee effectively. Wolves also substantially lowered adult survival and limited recruitment in the Madison drainage;but less than two dozen elk persisted in areas with shallower snow bordered by the swift, deep, and wide Madison River. Encounters with wolves remained high in these areas, but adult elk were sometimes able to flee to nearby refuge habitat.
Ultimately, this study demonstrated how behavioral, physical, and environmental factors interact to influence the vulnerability of elk to predation by wolves and, in the end, revealed wolves can have a dramatic effect on the abundance and distribution of elk across the landscape. While the Madison headwaters study may represent what could be considered a "worst-case" scenario with respect to the impacts of wolf restoration on elk, the processes documented in this study are similar to those documented in other wolf-elk systems throughout the Greater Yellowstone Ecosystem by other research teams. Integrating the results from this impressive body of scientific work, we conclude the impacts of wolf restoration can be substantial for elk herds spending winter in forested, mountainous environments where elk are quite vulnerable to predation due to a heterogeneous landscape with deeper snow pack. Predators tend to be more diverse and numerous in these areas due to lower susceptibility to human harvest and less conflict with livestock production. Conversely, the impacts of wolf restoration can be modest for elk herds spending winter in open, low-elevation valleys where elk are less vulnerable to predation due to a more homogeneous landscape with shallower snow. Also, predators tend to be less numerous in these areas due to high susceptibility to harvest and culls after livestock depredations. Over time, higher survival and recruitment in lower elevation valleys should lead to an increased proportion of elk spending winter in these areas. Indeed, a review of migratory elk populations throughout the Greater Yellowstone Ecosystem indicates broad-scale distribution shifts are occurring, with a higher portion of elk spending winter on lower-elevation ranges.
Certainly, many factors other than wolves, including human harvests, drought, and predation by bears and mountain lions, have had substantial effects on elk populations living in the Greater Yellowstone Ecosystem. However, the restoration of an additional top predator was a transformational event that eventually facilitated and maintained a substantive decrease in elk numbers and many other indirect effects to decomposers, other herbivores, predators, producers, and scavengers throughout the ecosystem. As a result, this bold restoration effort also led to a substantially improved understanding of the role of apex predators in terrestrial communities.
Pelican Valley & Mollie's Pack
by Douglas W. Smith, Travis Wyman, Daniel R. Stahler, & Daniel R. MacNulty
Unlike the northern range, wolf work in the interior can be tough. So tough, it was originally envisioned as aerial monitoring only, which is how most wolf studies accomplish the task of remote study. Bob Garrott and his team had successfully mastered ground data collection in the Madison-Firehole river drainages, but work elsewhere seemed infeasible. Most of these other locations were far from roads.
Then in 1998, the idea of working in Pelican Valley came up—a long famous place and what some would call the "heart" of Yellowstone. Situated in the middle of the park and vital to much wildlife, it certainly fits. A pack of wolves lived there, named Mollie's pack (after the late Director of the U.S. Fish and Wildlife Service who held her ground on wolf reintroduction despite criticism), and they seemed unique. Initially there were many hurdles to overcome; one was the uncertainty of success and, perhaps more importantly, some significant safety issues. The plan would entail camping out for two weeks in winter without a shelter and observing from a high point above the valley or an observation point (OP). Before that though, the first task was to see if the wolves were even in the valley enough to make observation worthwhile. A quick plotting of radio locations revealed wolves were in the valley a significant portion of the time, especially in late winter. We decided we might just have a project!
Clearing this new research with rangers and administrators took time. The administration's desire was for us to use Pelican Springs cabin, but if we did we would not be able to see the valley. A daily ski across the valley would disturb any wildlife we wished to observe. The decision was made to camp at the OP above the valley and stay put–no wildlife disturbance. In 1999, with scant winter equipment, we did just that. It proved to be a wise decision.
Once the hardship and struggle of hauling two weeks of gear across Pelican Valley was accomplished, with subpar equipment (especially sleds) and up a large hill, major scientific insights followed. At first we just watched and gathered behavioral data.
Quickly, we realized there were few elk, and later no elk due to the harsh environment, so Mollie's pack wolves had adapted to eating only bison in the winter. Quickly the story became about wolves and bison. Formidable prey compared to elk, killing bison presented a different challenge to the wolves. Several bison kills were witnessed, and a few were filmed, wetting the appetite to learn more and how their strategy differed from killing elk. Bison commonly stand their ground, whereas elk commonly flee–a major difference we noticed right away. Wolves facing a 1,000-2,000 pound animal presented a unique set of problems; taking the bison head on was out of the question. Wolves would have to work the environment to their advantage. Watching and waiting for the right moment to attack was critical. Wolves seem to have all the time in the world, so they were never in a hurry and waited. When they decided to attack, they chipped away: attack, wound, and wait; attack, wound, and wait...Using this strategy, some kills took up to nine hours. The wolves also had to use terrain to their advantage. Wind-blown hills had no snow and the bison favored such terrain for better footing; between the hills were troughs that collected snow, so the wolves favored these areas for attack as the snow hampered bison defense.
Confrontation between bison and wolves was stunning to watch;rarely observed nature in action. Pressuring bison for hours, wolves gradually drove them into deep snow and then jumped on them, many wolves at times, hanging from muscle and hide by their teeth. Once on firm ground, the bison shook the wolves off like water droplets, finally swinging their horns at them. Seemingly undeterred, the wolves waited for their next chance, or inexplicably left the bison, sensing an unseen cue or sign that made them abandon the effort.
At times, persistence paid off and a kill was made. But then another problem cropped up: who gets the spoils? This time of year a large bison carcass is a food bonanza. Every critter far and wide came in to grab what they could: weasels, foxes, coyotes, ravens, eagles, magpies, and grizzly bears. Once bears arrived it was over for the wolves. The carcass now belonged to them. Virtually every documented carcass in Pelican Valley from March through October attracted grizzly bears. It was not a matter of if but when, and the wolves had to grab as much meat as they could before the bears moved in. Up to 24 bears have been observed on one wolf kill at the same time. In March during our study, these carcasses became small "eco-centers" and most of the action in the valley occurred here.
Through time, our science became more sophisticated with fixed locations to observe from at regular intervals throughout the day, in addition to opportunistic observation of behavioral interactions. These observations indicated bison organized themselves differently when wolves were present in the valley versus when they were gone. Bison stayed closer to areas of good footing when wolves were around, and straying into riskier areas to forage when wolves were absent. Eventually the bison cow/calf groups left, probably because of wolf pressure, leaving about 40-80 hardy bulls for the wolves to deal with. So the valley changed, but in a vigorous way, and in fact gained some with the addition of wolves as they provided the carcasses that life hinged on in late winter.
Of course, we changed too. We purchased better equipment, especially sleds and light teepees that made living there for two weeks tolerable. We also dug into the snow and made caves to sleep in, and other years cut snow blocks with a saw to make an igloo. Crawling in either shelter, you could escape the near-constant roar of the wind or at night be oblivious to a foot of overnight snow that collapsed tents. For 16 straight years we managed the storms and wind that made Pelican Valley famous; and like with all things, we told stories, building memories that grew into a fondness for the place. After these years of study, it was felt our objectives had been achieved, so we turned things back to the valley, to the animals and plants that endure this harshness in the heart of Yellowstone.
Wolf Management: Den Closures, Habituation, & Hunting
by Douglas W. Smith, Kira A. Cassidy, Daniel R. Stahler, Erin E. Stahler, & Rick McIntyre
Although wolf reintroduction to Yellowstone National Park (YNP) was a very deliberate management action, and initially almost all of our work was management related, most of the wolf program today is monitoring and research. One reason for this is that there is almost no human safety threat posed by wolves. Why this is so is not entirely clear, but wolves seem to be naturally wary of people, or perhaps centuries of persecution have made them this way. Wolves are also less interested in human foods than other carnivores because they do not eat daily and are accustomed to the feeling of hunger. Therefore, it does not drive their behavior. Wolves commonly go days and sometimes a couple weeks without eating, so they do not become desperate for a meal. Wolves will feed on garbage, but when doing so are usually still wary of people (until conditioned). Overall, wolves are probably the least dangerous large carnivore. This does not mean we are not alert to the occasional wolf that may have received human food and is gradually losing its fear of people. Rather our management is not dominated by human-wolf interactions. Mostly we are focused on the flip-side, managing wolves so they are adequately protected from people—the other side of the National Park Service mandate. With Yellowstone being probably the best place in the world to view free-ranging wolves, much of our wolf management is geared toward people.
Protection of Wolf Dens & Rendezvous Sites
First and foremost are dens. Research has shown that wolves can be sensitive to human disturbance in the first six weeks after pups are born (Frame et al. 2007). Studies that have experimentally disturbed wolves during this time period found that sometimes the den will be relocated (Frame et al. 2007). Any time young pups are moved there is a risk of mortality, so this is the time period we try to protect wolves the most. The original Federal Special Regulations recommended protecting areas around dens until June 30. After this date, pups are mature enough to withstand disturbance and den relocation. In Yellowstone, we have only used this date as an approximate guideline because some circumstances are unique to a park. For example, we keep a den in Lamar Valley and a rendezvous site (the above ground site that wolves use after a den) in Hayden Valley, both popular viewing areas, closed for longer not only to protect wolves but also to allow for visitor enjoyment—a key national park policy objective. If we opened these areas, many people, with no ill intent at all, would approach the wolves hoping to see or photograph one, especially a pup, which would displace the wolves and make them less visible afterward. Despite our protection, many people mistakenly walk into the Hayden Valley rendezvous site. Packs that use this area have low pup production. The correlation between few pups and high disturbance is a concern. This possible relationship has caused us to keep the area closed after the recommended June 30 deadline, wanting to err on the side of resource protection. Finally, remote dens are left unmanaged mostly because it is unlikely they will be disturbed.
Although the risk of human injury from a wolf is almost zero, it is not actually zero, so another management activity is to keep wolves and humans apart. Our best tool for this is enforcing the park regulation that people must stay 100 yards from a wolf (or bear), and if the wolf moves closer, then the person must maintain this distance. This will keep wolves and people safe and prevent habituation.
When wolves and people do interact, Mark McNay, formerly of the Alaska Department of Fish and Game, summarized the outcomes for North America during the 20th century (McNay 2002). He found only 19 cases of aggression of non-rabid wolves toward people from 1900- 2001; these encounters excluded 20 incidents involving dogs or defensive behavior (protection by wolves of other wolves). There were no fatalities. Since 1969, there were 18 aggressive encounters and 11 of them involved habituated wolves (McNay 2002). Clearly, habituation needs to be prevented. How is this done? Keeping wolves and humans apart is one way, but keeping human food from them (similar to bears) is another.
Since McNay's study there have been three fatalities in North America, but the circumstances were similar. Wolves lost their natural fear of humans through exposure. It appears for wolves to attack humans they must first become familiar with them, lose their natural fear, and then attack, although this is very rare. This is not the case with other carnivores which may attack a person on their first encounter. In YNP, we have removed two wolves proactively because they had probably obtained human food and were exhibiting inappropriate behaviors (e.g., closely approaching humans). One wolf chased a person on a bicycle and a motorcycle. Another wolf walked up to several people and closely inspected anything they had in their hands (i.e., thinking it was food). In another situation this wolf tore apart a back-pack looking for food. Unfortunately, aversive conditioning did not work, so we removed them.
Before removal, park staff try everything they can to discourage habituated wolves (YNP 2003). Typically this means aversive conditioning. Confused with hazing, which is opportunistic negative reinforcement, aversive conditioning targets individuals. We start gently and escalate if there is no response--yelling, horns, and sirens first, graduating into cracker shells and ending with nonlethal bean bags or rubber bullets fired from a shotgun (YNP 2003). Initially, this was recommended against. We subjected these methods to professional review before we formulated our policy, and some comments were "don't bother with aversive conditioning, it doesn't work, just kill the wolf." This has been done in other places because wolves are so common and removing a tame one will have little impact on the population. This is true, but in a park setting we chose to respond differently. And to some people's surprise, we have found aversive conditioning to be effective.
Since wolf reintroduction, 55 wolves in 127 incidents have exhibited habituated behavior (this is different from McNay's "aggressive" category). Thirty-eight of these wolves were aversively conditioned 76 times; 49% of these actions immediately changed the wolves' behavior. Another 42% were probably successful, but not clearly so; because in eight cases we did nothing and the wolf never approached a person again. Finally, in 13 incidents the wolf either died or disappeared within six months of the incident. This is strong evidence that aversive conditioning does change habituated wolf behavior.
Where was habituation most common? Of the 127 events, 102 (80%) were on park roads, 14 (11%) in developments, and 11 (9%) were in the backcountry. Clearly the roads are a hot spot, and this is where the park has focused outreach and staff to avoid human-wolf contact. Most roadside encounters were in the spring/summer (May, June, and July; figure 1). This time period is also when pups become yearlings and many habituated wolves are young. Young wolves (yearlings in particular) have a lot of free time since older adults typically hunt and care for pups. Although yearlings do care for pups occasionally, they have less investment in pups so they take care of them less. They also explore and range widely and have a strong curiosity—which can lead them to humans. Some have compared this to human teenager behavior. The lesson is to be alert to young wolves in spring. Further, 54% of our habituated events have been confined to four packs, all of them road-adjacent packs: Canyon (23 incidents, 18%; pack lives in Hayden Valley area), Lamar Canyon (22 incidents, 17%), Druid Peak (13, 10%; Lamar Valley area combined with Lamar Canyon is 27% of incidents), and Hayden Valley (11 incidents, 9%; with Canyon, this is 27% of the incidents in the Hayden Valley area). Lamar and Hayden valleys are both open valleys with roads where visitors commonly encounter wolves. Arguably, these two valleys have more wolf-human proximity than any other location in North America. Certainly there are other places where wolf-human contact is more acute, but there might not be any other place where year-in and year-out wolves and humans coexist as much. Importantly, no one has been hurt, some aversive conditioning has occurred, and it has mostly been successful. Only one wolf (restricting the area to only these two valleys) has been removed.
Wolf Hunting Outside of Yellowstone
The last wolf management issue of concern is packs that primarily live in YNP, but wander outside of the park during the hunting season. Some of these wolves are legally harvested (figure 2). This is a difficult issue because the wolves are not aware of the boundary or the differing management objectives. These objectives are not mutually exclusive, but they are not the same. A gradual transition of regulations from inside the park to outside the park was necessary. Also, park wildlife that spend most of their time inside YNP are not conditioned to human hunting and are less wary and possibly more vulnerable to human take—wolves being one example. Many compare wolves to elk that are also cross-boundary, but migratory elk spend about half the year outside the park so they learn to be wary. To accomplish this goal the states have created small hunting units with quotas next to the park. Montana has created two special hunting districts north of Yellowstone that limit the number of wolves that can be taken. Wyoming has also created relatively small hunting units next to YNP which allows for precise control of harvest. Both of these actions have limited the harvest of wolves that primarily live within YNP (figure 2). This regulated hunting will ensure that human-take outside of YNP will not impact the wolf numbers inside the park. These actions do not control what wolves will get harvested; but it does reduce the chances that a commonly observed wolf, cherished by the public, will be removed. It also preserves the social fabric within wolf packs by not removing too many wolves of high social rank, thereby preserving the natural functioning of the pack and population dynamics. Overall, having wolves protected within YNP and harvested in a sustainable fashion outside the park is good for wolves in the long run. Such a mosaic of management practices protects wolves in some areas and limits them in areas of human conflict, which may reduce human dislike of wolves. Wolves are a polarizing issue for the public. Controlling problem wolves and hunting some of the others enhances acceptance of having them on human dominated landscapes. This is a foundational premise for all state agencies, and although questioned by some social scientists (Treves and Bruskotter 2014), has quelled some of the controversy over wolf restoration to the West. In the park, our mission is balancing wildlife protection with human enjoyment. It has taken some time; but we have achieved the proper balance for wolves to function as they should, and for people to observe and enjoy seeing wolves without harming them in a natural and wild setting.