Ecological Dynamics on Yellowstone's Northern Range (2002)

EVALUATING THE CONSEQUENCES of a natural regulation policy requires that the long-term dynamics of the Yellowstone landscape be understood in terms of variation in climate, disturbance regimes, vegetation patterns, animal populations, and human occupancy or use of the landscape. Yellowstone is a dynamic landscape, and we cannot determine whether management actions have forced components of the system beyond their historical range of variability unless we place recent dynamics in a longer time frame. Knowledge of prehistoric and historical environments is essential for creating a context for this evaluation. This chapter points out how dynamic the landscape of the GYE has been over time.

On a geological time scale, earth’s history is one of continuous change driven primarily by plate tectonics, with periodic extraplanetary influences, such as asteroid impacts and solar cycles. On this scale, the earth is fractured; subducted; uplifted; built up by volcanism and sedimentation; and worn away by ice, water, wind, heat, and gravity. Yellowstone National Park (YNP) is

centered on a geological mantle plume, or hotspot, where the molten magma comes close to the surface (Anders et al. 1989, Pierce and Morgan 1992, Pritchard 1999). The North American plate is thought to have moved to the southwest, leaving a series of volcanic traces from southwestern Idaho along the Snake River to the currently active Yellowstone region. Geomorphological changes have not stopped, for uplifting is continuing (Reilinger 1985). YNP is characterized by boiling mud pots, thermal pools, geysers, a volcanic caldera, ancient lava flows, and eroding rivers and waterfalls, which were the original reason for its protection as a park—wildlife and other life forms were an afterthought. Geomorphological changes may proceed gradually or may occur abruptly. Yellowstone witnessed a major earthquake as recently as 1959, whereas the Grand Canyon of Yellowstone is a prime example of wearing away by water on a grand scale.

Over the last 70 million years, climates have changed virtually continuously (Miller et al. 1987). The Pleistocene saw at least four major changes in climate due to alternating glacial and interglacial periods, although there may have been many important events on a shorter time scale. Over the last one million years, the earth oscillated between 90,000-year-long cold periods with ice accumulation and 10,000-year-long warm periods of ice melting (Muller and MacDonald 1997, Petit et al. 1999).

Pleistocene environments shaped the current Yellowstone landscape, flora, and fauna. Changes in the distribution of plants and animals and extinction of species were prevalent. Horses, camels, mammoths, and many other large mammals became extinct, whereas caribou, lemmings, musk ox, and other mammals retreated to the north at the end of the Wisconsin glaciation about 10,000 years before present (YBP). In addition, the abiotic changes must have caused adaptive responses—through both genetic evolution and behavioral or other plastic responses—in Yellowstone’s biota. However, information to evaluate these changes is sparse or completely lacking.

A paleobiological perspective is useful in identifying prehistoric processes that have shaped the Greater Yellowstone ecosystem (GYE), determining whether they continue to operate, and describing reasonable boundaries that may be placed on future variation.

The combination of species that simultaneously inhabit a given geographic location may be termed a community. However, communities tend not to be conserved in response to environmental change. Instead, individual species adapt and remain where they have been, or disperse at rates, times, and in directions in response to their individual tolerances to changing climatic, geological, anthropogenic, or other environmental conditions (Gleason 1926; Curtis 1955; Whittaker 1956, 1970; Davis 1976; Graham 1986). Thus, within the GYE, community composition is not stable over time scales of hundreds to thousands of years.

Many of the climate changes that occurred in the GYE region throughout the late Holocene (the past 4,000 years) were relatively small. If future climates do not have Holocene analogs (Bartlein et al. 1997), then the ranges of individual species may shift to areas that can support them outside the GYE. Although the current paleobiological record cannot document all the biotic changes associated with climate fluctuations, it does show various rates of change in climate and vegetation, some of which were extremely abrupt. Changes characteristic of the Pleistocene could happen within a human lifetime or less in the future (Alley 2000). Because the GYE is an “ecosystem island” within a larger human-dominated landscape, it is too small to accommodate environmental changes of the magnitude and frequency that were characteristic of the middle Holocene and late Pleistocene without changes in community composition, including local extinctions, greater than those seen during the past 4,000 years.

Changes in the late Quaternary (from about 14,000 YBP) vegetation and climate have been summarized by Barnosky et al. (1987) and are paraphrased here. With the initiation of deglaciation, 13,000 to 14,000 YBP, sagebrush (Artemisia spp.) dominated the vegetation of the subalpine forest zone. By about 11,500 YBP spruce began to colonize the area, followed by lodgepole pine, Douglas fir, and whitebark/limber pine during a period of warmer, wetter, and more stable climate (Taylor et al. 1997). By 4,500 YBP lodgepole pine became dominant with trace amounts of Douglas fir, both of which are common at lower elevations today.

Yellowstone’s pre-Holocene history includes a series of diverse faunal communities, many of which have no modern-day analog. Records from the late Pleistocene are incomplete but sufficient to illustrate the degree to which animal populations can respond to changing environmental conditions; they illustrate the types of change likely to occur in the future. Because these changes are not directly relevant to conditions over the past few millennia, a more complete description is given in Appendix A.

Prehistoric Native Americans exploited ungulates (pronghorn, deer, moose, sheep, elk, and bison) from the time they first entered the new world in the Late Pleistocene (Frison and Stanford 1982, Kay 1990, Cannon 1992). Bison remained a stable resource for Native Americans throughout the Holocene (Frison and Bradley 1991). At some early Holocene sites, hundreds of bison have been found in association with Paleo-Indian projectile points and stone tools (Wheat 1972, Frison 1974). At sites within the GYE, however, most bone beds associated with human hunting contain fewer than five individual animals (Cannon 1992).

Unlike bison, no massive bone beds of elk have been found associated with cultural materials, and there is no evidence of artificial elk traps or communal procurement practices like those used for bison (Frison and Bradley 1991). Arguments for low population densities of elk in the past have been based on bone frequencies from archaeological and paleontological sites, journal accounts of early travelers, and historical photographs (Kay 1990, 1994, 1995; Schullery and Whittlesey 1992; Kay and Wagner 1994). Elk bones are not numerically as common at archaeological sites as pronghorn, deer, and sheep bones (Kay 1990).

The use of bone frequencies from archaeological sites as a proxy for population levels depends on the assumption that people kill and eat ungulates in proportion to their actual abundance. This assumption probably is not always valid. Anthropological studies of modern hunters show that humans, like other predators, have preferences in prey selection that are independent of abundance. Therefore, that none of the archaeological sites in the GYE yields abundant elk remains does not necessarily indicate that elk were not abundant. Problems with the relationship between bone frequencies from anthropological and paleontological sites and local ungulate abundance are further detailed in Appendix A.

To assess the extent and nature of changes resulting from the adoption of natural regulation by the park, it is necessary to distinguish changes due to biological interactions during the past several decades from changes due to

abiotic factors. We review possible abiotic factors first and then turn to biological interactions.

The climate of YNP has fluctuated for the entire period for which we have records, sometimes quite rapidly (Dansgaard et al. 1993, Alley et al. 1996). Climate histories can be developed for periods far into the past, but we focus on the period that has most directly led to the current environments in the GYE.

At the peak of the last major glacial period about 14,000 YBP, large areas of North America were covered by glacial ice that had built up to a thickness of about 1.6 km over the previous 100,000 years. The continental glaciers penetrated only to the Canada-Montana border, but glaciers formed in the mountains of the GYE and spread to lower elevations and coalesced into the Yellowstone Ice Cap. Glaciers redistributed soils and sediments, widened valley bottoms, and blocked streams to form large lakes. These actions created some of the dominant landforms still present in YNP and the influence of glacial processes continues to the present. Streams draining from glaciers deposited sheets of gravel and sand in valleys below the glaciers. Some of these were later cut down to form terraces.

Climate is much more important than geological events in shaping the biota of the northern range at the scale of thousands of years. Seasonal extremes in temperature or moisture, rather than mean annual values, are probably the more important limiting factors for organismal distribution. Areas with the same average temperature and precipitation but highly different variation and seasonal patterns support very different biotas. Several important climate shifts strongly influenced Yellowstone during the Holocene. The warmest climates occurred in the early Holocene (7,000 to 9,000 YBP). The Medieval Warm Period dates to 1100 to 1300 AD (900 to 700 YBP) and it was manifest

in droughts in the northern plains of the United States (Laird et al. 1996, Woodhouse and Overpeck 1998).

Of particular interest for Yellowstone are the effects of the Little Ice Age, which lasted from 1450 to 1890 AD in two main pulses and appears to have been global (Crowley and North 1991). Temperatures averaged 1 to 1.5° C cooler (Crowley and North 1991) and the climate tended to be moister than it is today. During the Little Ice Age, glaciers advanced worldwide. Within this period, there were periods of warm and dry and cool and wet (second pulse of the Little Ice Age, 1860 to 1910) conditions, which affected the growth and distribution of plant species (Whitlock et al. 1995). In the GYE, this record is best documented by tree rings.

The stability of the present climate appears to be abnormal compared with recent millennia, although the climates of the past 200 years have included the warmest and coldest periods of the past 4,000 years (Bradley 2000).

On a decadal time scale, Yellowstone’s climate is measurably influenced by El Niño-Southern Oscillation (ENSO) patterns, primarily through the creation of extremes in precipitation and temperature. These extremes themselves probably play a major role in shaping the GYE’s environment and biota. However, although climate variability is subject to ENSO periodicity, overall annual precipitation shows no statistically significant trend over the past 100 years (Balling et al. 1992a).

Recurrent wildfire profoundly influences fauna, flora, and ecological processes in the northern Rocky Mountains (Habeck and Mutch 1973; Houston 1973; Loope and Gruell 1973; Taylor 1973; Wright and Heinselman 1973; Wright 1974; Arno 1980; Romme and Knight 1981, 1982; Romme 1982; Knight 1987, 1996; Romme and Despain 1989; Despain 1990; Turner et al. 1997). Although small fires occur frequently, total area burned is dominated by a few, very extensive fires (Johnson and Fryer 1987, Romme and Despain 1989, Johnson 1992). Meyer et al. (1995) found weak evidence of fire as long ago as 7,500 and 5,500 YBP and strong evidence for substantial episodes 4,600, 4,000, 2,500, 2,100, 1,800, 1,200, and 850 YBP. Between 9 and 12 maxima can be identified in the record between 2,000 YBP and the present, including the 1988 fire. Major fire events appear to have occurred recently at 100- to 300-year intervals (Romme 1982, Romme and Despain 1989) and at

various frequencies throughout the Holocene (Millspaugh and Whitlock 1995). Within the ungulates’ summer range, large, infrequent fires create vegetation mosaics that dominate the landscape until the next extensive fire.

On the northern range, tree-ring evidence and fire scar data indicate that 8 to 10 extensive fires occurred in the area during the last 300 to 400 years, which suggests that fires burned the winter range at intervals of 20 to 30 years before European settlement (Houston 1973, Barrett 1994).

Climate plays an important role in fire frequency and extent. Most large North American twentieth century fires have been associated with persistent high-pressure ridges or dry La Nia phases of ENSO (Bessie and Johnson 1995, Swetnam and Betancourt 1998). Within Yellowstone, the area burned correlates with a trend toward increasing late-winter aridity since 1895 (Balling et al. 1992a, 1992b). However, the potential for climate-induced changes in fire frequency and extent in the Northern Rocky Mountains (Flannigan and Van Wagner 1991, Romme and Turner 1991, Bartlein et al. 1997) underscores the importance of understanding the effects of extreme events and of considering long-term disturbance dynamics when considering alternative management strategies.

Lower elevation communities, such as those found on the northern range where fires were formerly frequent, have been altered by long-term fire suppression. Fire suppression was discontinued in 1972 in Yellowstone, after which lightning-caused fires were again allowed to burn. As with many other crown fire-dominated ecosystems, YNP is usually considered a nonequilibrium landscape (Romme 1982, Sprugel 1991, Turner and Romme 1994).

No large fires occurred during the twentieth century until those of 1988, the largest since the park’s establishment, affected more than 321,000 ha in YNP and the surrounding area and burned approximately 36% of the park (Schullery 1997). These fires were primarily the result of unusually prolonged drought and high winds (Renkin and Despain 1992, Bessie and Johnson 1995) and were consistent with the earlier pattern of punctuated episodes of extreme fires followed by long periods of small fires (Millspaugh and Whitlock 1995). Reconstructions suggest that the last time a fire of this magnitude occurred in Yellowstone was in the early 1700s (Romme and Despain 1989), which makes the 1988 fires unusual in size (Christensen et al. 1989, Turner et al. 1994a,b,c). However, fire suppression probably had only minimal influence on the extent and pattern of the 1988 fires (Romme and Despain 1989, Barrett 1994).

The effects of fire on wintering ungulates change through time following the fire. Initially, fire consumes aboveground plant biomass and reduces winter

forage supply, which is not regenerated until the following summer. This reduction in forage can increase ungulate mortality during the first winter after a major fire (Singer et al. 1989, Wu et al. 1996), an effect that is magnified during more severe winters. In subsequent years, primary productivity may be stimulated, resulting in improved forage quantity and palatability (Harniss and Murray 1973, Gruell 1980, Hobbs and Spowart 1984, West and Hassan 1985, Coppock and Detling 1986). Some plant communities on the northern range showed substantial increases in forage abundance and nutrient content in response to burning (Wallace et al. 1995, Tracy and McNaughton 1997). Ungulates on the northern range preferentially used herbaceous plants in burned areas of the landscape (Pearson et al. 1995). Boyce and Merrill (1991) hypothesized that this fire-enhanced forage base in burned forests (i.e., increased herbaceous ground cover) might enhance ungulate recruitment and population size for several years following the 1988 fires. However, enhancement of forage production in grassland areas was no longer detectable approximately 5 years after those fires (Singer and Harter 1996).

ENSO is not directly correlated with YNP fires because the quantity and ignitability of herbaceous fuels are strongly influenced by local weather and time since the most recent catastrophic burns. Nevertheless, if the GYE’s climate is warming, and ENSO events are becoming more frequent, then fire can be expected to be a more dominant process in the future (Millspaugh and Whitlock 1995).

Fire also has geomorphological consequences, because vegetation is removed by burning, and storm runoff occurs as overland flow. Charcoal layers that commonly accompany or follow fires appear in expanded alluvial sediments (Meyer et al. 1992). Erosion continues in subsequent years, even after revegetation, because deep incision contributes to slumping of fire basins into channels. The clearest demonstration of the impact of fire on erosion is the aftermath of the 1988 conflagration (Meyer et al. 1992, 1995). Meyer et al. documented major erosional debris flows that extended over several hundred meters in the Slough Creek and Soda Butte Creek drainages of northeastern YNP. Meyer et al. (1992) proposed an idealized scenario to explain how wet periods during the Holocene created widened and sinuous stream beds, and dry periods with burns led to erosion of steeper slopes and aggradation of alluvial fans and incised stream courses. Whether the processes are captured correctly by this scenario, clearly events of this magnitude reshape the character of streams, both large and small, and constitute a greater influence than elk browsing on willows and other riparian woody vegetation.

Such major forces affect not only stream morphology and seasonal flow but also the likelihood of persistence of beavers with their modification of stream flow, water-table height, and extent of riparian vegetation.

This evidence suggests that relatively minor climatic changes in the late Holocene could have caused major shifts in fire regimes, alluvial processes, and resulting morphology and vegetation of valley floors.

Although many factors influence runoff, typically there is a strong positive correlation between precipitation and stream flow. No long-term trend in stream flow at the Corwin Springs gage on the Yellowstone River just north of the park is apparent through the 91-year record from 1908 to 1999 (p=0.259). Although other studies have shown a decline in precipitation over the same period, this is true only for late winter measurements, which accounts for a minority of annual precipitation (Balling et al. 1992a). The proportion of snowmelt that contributes to stream flow is high relative to that of rainfall, much of which is absorbed in the soil and transpired. Like stream flow, snowpack showed no significant downward trend (p=0.40) but was highly subject to ENSO events (p=0.001). During an ENSO, precipitation at YNP is much greater than normal, which causes major changes in stream courses.

It is usually presumed that fire increases surface runoff and stream flow by destroying vegetation, thereby increasing the likelihood of erosion and bank instability. After the 1988 fires, however, no significant increase in stream flow was observed in the Yellowstone River.

Stream-course changes during the flood years of 1996 and 1997 were the greatest observed since 1954, primarily the result of flood flows (Mowry 1998). From 1954 until 1987, a period of relatively constant stream flow, stream channels narrowed, especially in areas where elk wintered, a process described by Lyons et al. (2000). From 1988 to 1997 there was increased channel instability, and most of the changes occurred in the exceptional flow years of 1996 and 1997. Most stream-bank erosion during these extreme runoff years occurred higher in the drainage—where there were few wintering elk and the streamside willows were robust—suggesting that increased erosion was primarily due to hydrological variables acting on high-roughness riparian vegetation that stabilized the stream bank during high-flow, nonflood years and

not to ungulate activity. This conclusion does not explain the long-term development of low-roughness vegetation (i.e., vegetation with little resistance to stream flow power) in areas with heavy winter elk use. However, these differences do explain Mowry’s (1998) finding that there was a greater change in stream reaches with willow-covered banks than in those with grass-dominated banks. Shifts in stream morphology associated with increased stream flow are expected because the stream slope is near the transition of meandering and braided stream courses based on the established relationship between channel slope and bankfull discharge (Leopold and Wolman 1957).

Climate, vegetation, and stream-course relationships interact in multiple ways, but it is difficult to assess the role of extreme events because they happen so infrequently. Nevertheless, it is apparent that both physical forces and browsing and streambank trampling by ungulates contribute to changes in streams, but the data are not sufficient to sort out their relative importance in causality.

Stream-bank erosion is only one of the issues raised about management of the northern range. Critics of the natural regulation policy in the northern range claim that degradation of upland slopes by overgrazing results in accelerated soil erosion (YNP 1997). One indication of the level of erosion off the slopes is the accumulation of sediments in depressions or lakes. Engstrom et al. (1991) studied several lakes on the northern range, looking at pollen changes and abnormal sediment deposition patterns. They concluded that their “investigation of the sedimentary record does not support the hypothesis that ungulate grazing has had a strong direct or indirect effect on the vegetation and soil stability in the lake catchments or on the water quality of the lakes.” A review of their data, however, shows many, but not all, lakes with increasing sediment accumulation starting after the beginning of the twentieth century. Most of the study lakes near the confluence area of the Yellowstone River, Lamar River, and Slough Creek showed sediment increases. One study lake in this area did not. Interpretation of these data might vary by investigator, and the magnitude of the increases may be “normal,” although sediment accumulation in the nineteenth century seems to be comparatively stable when compared to the twentieth century.

The GYE includes not only YNP but also most of the adjacent lands at an elevation above 1,524 m (5,000 ft). These surrounding lands are used for ranching, agriculture, and recreation. The population of Park County, north of and adjacent to the northern range of YNP, is growing faster than the populations in most Montana counties.

In 1880, the population of Park County was 200, but after the arrival of the Northern Pacific Railroad in 1881, it grew to 6,900 in 1890. Population growth has continued and today the county (population about 16,000) is dotted with many new developments and small ranches with increased fencing, as well as many semi-urban areas (Park County 2001).

This area was reported to be the ancestral wintering range of the northern range elk herd before YNP was established (Graves and Nelson 1919). The influence of development and farming was noted by Wylie (1882,¹ 1926²) as early as the late 1800s and early 1900s. He said, “The buffalo, deer, and elk were accustomed to living on this plateau during the summer. In winter they migrated to lower and warmer regions outside the park area until settlements of farmers in the country surrounding the park made it impossible for them to use their long used winter homes.” All these factors fragment habitat and impede ungulate movements and access to foraging areas. This alteration of the landscape outside of YNP may have as great a potential to affect ungulate populations, their behavior, and the use of vegetation as do changing climatic conditions and reintroduction of wolves. Little is known, however, about the relative importance of each of these factors or their interrelationships.