NATIONAL PARK SERVICE Research in the Parks NPS Symposium Series No. 1

The northern winter range of Yellowstone National Park contains about 204,000 acres, supports six species of native ungulates with their complement of predators and scavengers, and is, in short, an ecological gem unmatched for this assemblage of species. Current research is designed to test hypotheses which relate to ungulate habitat relationships, biotic succession, and ecological homeostasis. Special emphasis has been placed upon the elk (Cervus canadensis) as the most abundant ungulate and because of the history of concern and controversy surrounding this species. I present here a review of the research approach and some preliminary results.

The primary purpose of Yellowstone National Park, as a natural area, is to maintain a representative ecosystem in as near pristine conditions as possible. Management of the park ecosystem generally involves preventing or compensating for modern man's altering of natural ecological relations. Criteria for management of a national park must therefore differ from criteria for other uses of land. The mission-oriented research in parks often involves documenting pristine conditions and processes, determining the completeness of park ecosystems, and developing management procedures to maintain or restore the ecosystem (Houston 1971a).

This is a fair description of the studies being conducted on the northern winter range and research hypotheses have been developed to guide problem definition and analysis. The major concern in posing hypotheses was that they be capable of being tested and rejected, and I have not always used a formal null form (Ghent 1966). Stating hypotheses capable of unequivocal support or rejection is a problem in ecology, and I suspect that those being used to guide this research program could generate lively discussions on the basis for rejection, the separation of causes from effects, the possibility of circular arguments, etc. Some of the hypotheses lend themselves to experiment; others must be tested using the method of successive approximations described by Poore (1962). Many of the hypotheses being tested are based upon results from other studies of ungulates that have been conducted in parks of the Rocky Mountain area since 1962. These have recently been reviewed by Cole (1971a).

HYPOTHESES CONCERNING HERBIVORES

The "northern range" extends down the elevational gradients of the Yellowstone and Lamar rivers from about 8500 to 5000 ft. The range within the park is contiguous with additional acreage on the Gallatin National Forest, making an overall total that approaches a quarter-million acres. Winter distribution and densities of ungulates along this elevational gradient are fluid, and appear to be influenced by severity of environmental conditions, population size, and various behavioral interactions. Although some ungulates occur on the area as year-long residents, highest densities occur during November-April periods.

An analysis of census data obtained from 1929-61 led to the hypothesis that the northern elk herd might be divided into a "migratory segment" that wintered in the lower portions of the Yellowstone River drainage both inside and outside of the park, and a "resident segment" that stays within the upper portions of the Yellowstone and Lamar river drainages (Fig. 1) even during severe winters (Cole 1969). The division between segments is tentative, but present studies suggest that 3000-4000 elk may occur in the postulated resident segment (Table 1). Research designed to test the hypothesis that the resident segment occurs on "ecologically complete habitat" and will be "naturally regulated" was initiated in 1968 when the periodic artificial regulation (involving trapping and shooting) was terminated. Natural regulation of ungulate populations has been defined as regulation of numbers without human influence (Cole 1971a). The actual mechanisms involved in natural regulation are difficult to document, but appear to be primarily due to the effects of density-influenced intraspecific competition for winter forage and the partially density-independent effects of severe weather on mortality and natality. Ecologically complete habitat for elk appears to consist of complexes of physiographic sites with interspersions of vegetation types that provide contingencies to obtain forage and maintain relatively stable populations in variable and periodically harsh environments (Cole 1971a).

Fig. 1. Map of Yellowstone Park showing the northern winter range divided into postulated areas for resident elk segment (C) and migratory segments inside (B) and outside (A) the park. Area B, is the "boundary line area" described in text.

The hypothesis for the natural regulation of elk will be evaluated by analyses of accumulated data on rates of natality and mortality, sex and age structures, historical records, and by conducting regular distribution and census counts. The basis for rejecting the hypothesis would be departures from natural conditions in interspecies homeostasis which resulted from an elk population eruption (Caughley 1970), retrogressive plant succession, or trends toward competitive exclusion among populations of sympatric native herbivores.

TABLE 1. Research hypotheses being tested on herbivores, carnivores, and vegetation on the northern winter range of Yellowstone National Park.

As shown in Table 1, the use of historical records and photographs is an important means of evaluating departures from natural conditions. A major project has been to assemble and interpret this reference information. Yellowstone has a wealth of historical information, most of which has not been used in this manner (M. Meagher, pers. comm.). It has become apparent that, without this background framework, any assessment of problem situations can be little more than personal judgments which may vary with the training and experience of individuals.

As part of this project I have reviewed over 200 publications, reports, etc., in an attempt to provide historical perspective on numbers and distribution of the northern elk herd. Narrative accounts from 1877 support the concept of the postulated resident segment for the northern herd. This review has not supported the reports of a population eruption and crash in the early decades of this century which seem to be entrenched in the literature of the northern herd. The reported population eruption appears to have resulted in part from changing the definition of the northern herd, and from the methods employed in early censuses and in calculations of increase rates. The northern herd has variously included elk wintering on the Gallatin, Madison, Firehole, and Shoshone rivers, in addition to those on the Yellowstone and Lamar rivers (the current definition). This apparently has not been previously recognized. The population crash appears to have been suggested in an administrative report as an explanation for the discrepancy between counts and calculations, and has subsequently been accepted as factual (Houston 1971b).

Historically, some portion of the migratory segment of the northern herd has moved beyond park boundaries during severe winters, which suggests that the lower range within the park does not represent an ecologically complete habitat. These elk are periodically available to be managed by public hunting conducted outside park boundaries. Present management objectives will test the hypothesis that it is possible to maintain the migratory segment in a 3000-5000 animal range by using public hunting, and that hunting will not result in conditioned avoidance behavior among elk, which in turn would cause unnatural concentrations within the park. The proposed variable quota hunting system, where allowable removals can fluctuate annually from 0 to 2500 elk, depending upon fall herd sizes and the numbers migrating outside park boundaries, has been described in detail elsewhere (Cole 1969).

Elk occur sympatrically on the northern range with populations of mule deer (Odocoileus hemionus), bighorn sheep (Ovis canadensis), moose (Alces alces), pronghorn (Antilocapra americana), bison (Bison bison), and a variety of smaller herbivores. The hypothesis concerning sympatric herbivores will test the concepts that the natural regulation or effects from hunting the different segments of the elk population will not result in competitive exclusion because of interspecific competition, and that populations of herbivores associated with elk will also be naturally regulated.

The entire northern elk herd probably numbered less than 4500 at the termination of artificial reductions in 1967-68. Calculations suggest that 8500-9000 elk should occur in the 1971 fall herd. The elk population is approaching levels where these hypotheses will be tested, but it would be premature to reject, modify, or claim support for them from ongoing field studies. Tentative support for several hypotheses has been provided from historical records.

HYPOTHESES CONCERNING CARNIVORES

Herbivores of the northern range support a complement of predators and scavengers which include the grizzly bear (Ursus arctos), black bear (U. americanus), cougar (Felis concolor), gray wolf (Canis lupus), coyote (C. latrans), wolverine (Gulo luscus), raven (Corvus corax), golden eagle (Aquila chrysaetos), and bald eagle (Haliaeetus leucocephalus). Except for the historic presence of a low-density population of aboriginal man, this predator-scavenger fauna appears to be largely intact, although gray wolves, cougars, and wolverines occur at very low densities. The hypothesis that predator-scavenger populations will increase in response to increased food supplies as a result of having naturally regulated elk will be evaluated by monitoring sightings and signs (Cole 1971b). A continuing analysis of predator-prey relationships (including rates of predation and scavenging, prey selection, etc.) will be used to evaluate the hypothesis that in a variably harsh environment such as Yellowstone's, predation functions as an assisting but nonessential adjunct to the regulation of ungulate population sizes. The latter is an example of an hypothesis that will be difficult to clearly support or reject. Both hypotheses must be qualified to the extent that population responses among avian scavengers such as eagles and perhaps for the cougar and gray wolf will depend upon the protection these species receive when they move beyond park boundaries.

HYPOTHESES CONCERNING PLANT COMMUNITIES AND HABITAT RELATIONSHIPS

The vegetation on the winter range is primarily a steppe composed of a variety of grassland types, which is intermixed with scattered coniferous forests. A fundamental hypothesis is that the density, composition, and successional trends of vegetation on the range do not depart from natural conditions, or that departures have not resulted from grazing by native ungulates (Table 1). (I omit the area adjacent to the north boundary from consideration here for reasons discussed below.) The corollary under study is that native ungulates on ecologically complete habitat do not have a capacity to progressively deplete food supplies that ultimately determine their own densities. Stated more specifically, no unnatural effects upon vegetation will result from the natural regulation of the resident segment of the elk population. Support for these hypotheses has been provided by rephotographing over 100 historical photos of the range. This in turn led to the development of subsidiary hypotheses that are more specific and which deal with vegetative conditions or changes that have been of major concern in the maintenance of the park ecosystem. A consideration of written historical records indicated that photos taken from 1871 to the late 1880s provide our best approximation of the pristine conditions. Ungulate populations would have been at pristine levels in the early 1870s; some may have been reduced through hide hunting by the mid-1880s.

A consideration of historical photos (Fig. 2), vegetation measurements, and studies of ungulate habitat use led to the hypothesis that low densities of herbaceous vegetation on certain ridgetop sites represent naturally occurring "zootic climax" vegetation, and do not illustrate retrogressive plant succession. After some preliminary descriptive sampling, I mapped in detail the distribution of those plant communities showing 60% or more bare ground. About 5900 acres were in this condition, and about 1100 of these acres were considered to be overwhelmingly the result of peculiar topoedaphic conditions. The remaining 4800 acres are considered to represent some combination of natural "zootic climax" vegetation and perhaps retrogressed disclimax vegetation adjacent to the north boundary. Given that the area of zootic sites might be underestimated by 25% (because of failure to locate them all), there may be 6000 acres of this vegetation that has been previously considered "overgrazed" and in "poor condition" when the usual range criteria are applied. This represents 3% of the winter range and the historic photos show the presence of these areas 100 years ago. I suggest that the condition of this vegetation does not reflect range deterioration (boundary line area possibly excepted). This represents a fundamental difference from interpretations made in the past.

Fig. 2. Northern winter range along Yellowstone River. The west-facing ridgetop receives heavy use by bighorn sheep and elk, shows little change in patterns of herbaceous vegetation after about 86 years, and is considered to be largely a "zootic climax" vegetation. Upper photo by J. P. Iddings circa 1885 (USGS No. 142), lower, D. B. Houston 1971.

Space precludes a complete discussion concerning the hypothesis that replacement of willow (Salix spp.) and aspen (Populus tremuloides) communities on the winter range does not represent retrogressive succession, but instead indicates types of natural primary and secondary succession. However, I shall use the changes in willow distribution to present a portion of this argument. Photographic evidence shows that willow communities have declined on limited areas of the winter range during the past century. However, willow has also declined on areas outside the winter range (Fig. 3) and in areas well outside the park. Willow has also persisted on, and appears to have colonized, pioneer substrates on the northern range in the presence of wintering ungulates. These observations support the hypothesis that changes in the distribution of willow represent primary plant succession which reflect some fundamental changes in hydrology and alluviation on streams.

Fig. 3. Soda Butte Creek. Yellowstone National Park. Replacement of willow by conifers has occurred well outside the usual ungulate winter range in about 85 years. These changes are interpreted as representing primary succession and not retrogressive succession due to ungulate browsing. Upper photo by J. P. Iddings circa 1885 (USGS No. 347), lower, D. B. Houston 1970.

Major vegetative changes illustrated in historical photos (Fig. 4) led to the hypothesis that natural fires have been a major influence in development of vegetation on the winter range. Evidence for past fires is abundant, and a series of fire-scarred trees were cut to investigate fire frequency. Individual fires have been dated as early as 1525, and best estimates of frequency suggested mean intervals of about 20-25 years between fires (Houston 1973). This has provided strong support for the hypothesis, and I suggest that few interpretations of range trends or assessments of departures from natural conditions can be valid unless the influence of natural fires is considered. A natural or man-influenced change in fire frequency has probably contributed substantially to the deterioration of aspen on the winter range, and could relate to postulated changes in stream hydrology affecting willow distribution. These hypotheses in turn are being subjected to further testing by again permitting natural fires to burn.

Fig. 4. Tower Junction, Yellowstone National Park. The increase in big sagebrush on Festuca idahoensis and Stipa spp. grasslands, the increase in the area and density of coniferous forest, and the decline in aspen typify the vegetative change that has occurred on the winter range. Changes are considered to support the hypothesis of a change in frequency of natural fires. Upper photo by J. P. Iddings circa 1885 (USGS No. 152), lower, D. B. Houston 1970.

A preliminary review also suggests the possibility of climatic changes dating from the Gannett Peak stage of neoglaciation (Benedict 1968; Porter and Denton 1967) that must be considered in interpretations of vegetative changes.

The range along the present north boundary of the park is an arid steppe with an annual precipitation of about 11 inches. The 11,000 acres (5% of the range) within 1-3 miles of the boundary provide winter habitat for elk, mule deer, pronghorn, and bighorn sheep. The condition of the vegetation on this area also nearly defies interpretation because of the various human uses that have occurred in the past. Much of this area was not within the original boundaries of Yellowstone National Park but was added in 1932. A partial list of past human influences would include developments such as homesteads with attendent grazing by domestic livestock, a townsite of several hundred people, a military rifle range, a racetrack and golf course, a fence along the original park boundary, feedgrounds for native ungulates, and accidental or deliberate introductions of exotic plants. Additionally, soils on portions of the area were derived from an unusual bentonitic parent material (Waldrop and Hyden 1963). I have constructed three hypotheses concerning the ecology of big sagebrush (Artemesia tridentata) to illustrate the complexity of interpretations (Table 1). The observed decline of this species on the area has been a major concern. It is quite possible to find support for each hypothesis, but most photographic evidence favors the hypothesis that high densities of big sage at this low elevation represented a disclimax community which resulted primarily from grazing by domestic livestock and its reduction represents a return to more natural conditions (Fig. 5).

Fig. 5. The Reese Creek area was added to Yellowstone Park in 1932. Upper photo by J. E. Haynes, 1941; lower by D. B. Houston, 1971. Foreground shows a reduction in big sagebrush and its replacement by an Agropyron spicatum-Stipa comata grassland. Sage persists on the area beyond the fence, which is within the park, but where grazing by domestic livestock still occurs. Other photographs to 1871 suggest that the present foreground vegetation more closely approximates pristine conditions.

An analysis of historical records and studies of habitat use by ungulates are being used to evaluate the hypothesis that vegetation on spring, summer, and fall ranges does not depart from natural conditions. Support for this hypothesis has been provided from comparisons of historic with recent photos and from results of detailed studies of summer range conditions in southern Yellowstone National Park and on adjacent forest areas (G. F. Cole, in prep.; G. Gruell 1973).

The purpose of this paper has been to review the approach to research that is being conducted on a portion of Yellowstone's ecosystem. The review of historical information has provided background support for some research hypotheses and has resulted in different interpretations concerning the history of the northern Yellowstone elk and their habitat. It is too early to claim conclusive support for most hypotheses and there will no doubt be modifications and rejections as studies continue. There are those hypotheses that, by the very nature of the complex interactions among causes and effects in ecosystems, may not lend themselves to decisive support or rejection.

REFERENCES

BENEDICT, J. B. 1968. Recent glacial history of an alpine area in the Colorado Front Range, U.S.A. J. Glaciology 7:77-87.

CAUGHLEY, G. 1970. Eruption of ungulate populations, with emphasis on Himalayan thar in New Zealand. Ecology 51:53-72.

COLE, G. F. 1969. Elk and the Yellowstone ecosystem. Res. Note Yellowstone Natl. Park. 13 p.

______. 1971a. An ecological rationale for the natural or artificial regulation of native ungulates in parks. Trans. N. Am. Wildl. Nat. Resour. Conf. 36:417-425.

______. 1971b. Status of gray wolf. Res. Note No.4, Yellowstone Natl. Park. 6 p.

GHENT, A. W. 1966. The logic of experimental design in the biological sciences. BioScience 16:17-22.

GRUELL, G. E. 1973. An ecological evaluation of Big Game Ridge. U.S. Forest Service Publ. Intermountain Region. 61 p.

HOUSTON, D. B. 1971a. Ecosystems of national parks. Science 172:648-651.

______. 1971b. History and demography of the northern Yellowstone elk. Unpubl. ms.

______. 1973. Wildfires in northern Yellowstone National Park. Ecology 54:1111-1117.

POORE, M. E. D. 1962. The method of successive approximation in descriptive ecology. Pages 35-68 in J. B. Cragg, ed. Advances in ecological research, Vol. I. Academic Press, New York.

PORTER, S. C., and G. H. DENTON. 1967. Chronology of neoglaciation in the North American Cordillera. Am. J. Sci. 265:177-210.

WALDROP. H. A., and H. J. HYDEN. 1963. Landslides near Gardiner, Montana. Pages E11-E14 in Short papers in geology, hydrology, and topography, U.S. Geol. Survey, Prof. Paper 450E.

Acknowledgments

I thank G. Cole, M. Meagher, L. Loope, and D. Despain for reviews of the manuscript. Dr. Meagher located many of the photos and records used in the historical review.