Influences of Adjacent Forest Management Activities on Migratory Elk of Mount Rainier National Park
Kurt J. Jenkins, Edward E. Starkey

CH. 2: ELK-WINTER RANGE RELATIONSHIPS

Elk are important economic and ecologic components of montane and subalpine ecosystems of the Cascade Mountains. Nearby urban centers enhance the recreational value of large elk herds in the Mount Rainier ecosystem, both for viewing opportunities and for traditional consumptive uses. But high densities of elk resulting from forest management activities could cause unacceptable impacts to subalpine meadows on summer ranges within Mount Rainier National Park (MORA). As demands for elk, timber, and park recreational resources increase, better understanding of relationships between logging and nutritional limiting factors of elk populations will be needed to accommodate diverse resource management objectives in the region.

Factors limiting elk populations are poorly understood in the western Cascades and Pacific coastal forests. Most studies have concluded that availability of digestible energy, particularly during winter, is the most limiting nutritional factor (Schoen 1977, Janz 1983, Leslie et al. 1984). Dietary nitrogen levels are generally thought to meet minimum ecological requirements provided that concentrations of protein-binding phenolic compounds do not interfere significantly with protein digestion (Janz 1983, Leslie et al. 1984). Recent studies, however, suggest that insoluble tannin-protein complexes may result in protein deficiencies for elk consuming browse diets in the Pacific Northwest (Hanley et al. 1987, Happe et al., in prep).

Evaluating effects of forest management activities on winter nutrition of elk requires information on availability of nutrients, and composition and quality of elk diets following logging. Previous studies have documented vegetation trends following logging in the Cascade Mountains (Dyrness 1973, Long and Turner 1975). Those studies, however, emphasized peak standing crop biomass during summer, and failed to distinguish between phytomass available and unavailable to browsing herbivores. Previous researchers also have documented diets of elk in the Cascade Mountains (Schoen 1977, Merrill et al. 1987) and nearby coastal forests (Janz 1983, Leslie et al. 1984). None, however, have compared diets and nutritional planes of elk inhabiting distinct stages of forest succession.

The goal of this study was to determine influences of recent logging practices on winter nutrition of elk in the White River drainage, Washington. Specific objectives were (1) to compare available biomass of elk forages among various plant communities and successional stages on managed forests adjacent to MORA, (2) to compare composition and nutrient levels of elk diets from unmanaged old-growth forests within MORA and adjacent managed second-growth forests, and (3) to compare forage values among plant communities on managed elk range.

Methods

Forage Biomass Trends

Standing crop biomass of available elk forage was estimated in 14 plant communities outside MORA along the White River from 7 October 1986-15 April 1987. Available forage was defined as principal forage species available above snow and below 225 cm in height. Principal forage species were those that comprised at least 1% of the reported autumn-spring diets of elk from throughout the Pacific Northwest (Swanson 1970, Schoen 1977, Janz 1983, Leslie et al. 1984, Harper 1985, Merrill et al. 1987). Plant communities chosen for sampling included a chronosequence of stand ages present an the managed forest, both on mesic bottomlands and on comparatively xeric uplands (Table 1.2).

Available biomass was measured in 3 - 8 randomly selected replicate stands of each plant community. Autumn standing crops were measured using 10 1-m2 frames systematically placed along a 100 m transect in each sample stand. Canopy coverages and heights were measured for principal forbs, grasses, sedges, evergreen shrubs, and abscised leaves of deciduous trees. Additionally, current years shoots (>4 cm) of deciduous shrubs, conifers and ferns were counted and tabulated according to their heights above ground: 0-25, 26-50, 51-75, 76-100, and 101-225 cm.

Autumn biomass of forbs, grasses, sedges, evergreen shrubs, and abscised leaves were determined from cover and/or height measurements using simple linear regression models. Regression models relating biomass to cover were developed for individual species of principal forbs, evergreen shrubs and abscised leaves. General models based an cover and height were developed for grasses and sedges (Appendix I). Regression models were developed from clipping and weighing vegetation within 12 .25-m2 plots after estimating species coverage and height. These plots were distributed randomly throughout the study area within strata containing low, moderate, and high coverage of each species. Vegetation in each plot was clipped to a height of 1 cm, and was oven-dried prior to developing regression models.

Autumn biomass of deciduous shrubs, conifers and ferns was determined as a product of twig density (twigs/m2) and average dry weight of twigs (g/twig). Average even-dried weights of twigs were determined from random samples of 50-284 twigs of each species obtained from throughout the study area (Appendix II).

Seasonal changes in available biomass were described throughout the winter by monitoring canopy coverage of principal forages within 100 1-m2 permanently staked plots. Plots were located systematically along 10 100-m transects distributed randomly among 5 plant communities (2 transects per community) on managed forests adjacent to MORA. Species cover was estimated on 1 November, 15 January, 15 February, and 15 April 1986-87. Seasonal correction factors were developed for adjusting autumn estimates of available biomass based an ratios of winter cover: November cover in the permanent reference plots. If correction factors were not available for a principal forage species or specific plant community, correction factors for similar species and plant communities were substituted.

Forage Quality

In vitro digestible dry matter (IVDDM) and crude protein (CP) levels were determined throughout a winter for 34 principal forages of elk. Composite samples of each principal forage were collected on or about 1 November, 15 January, 15 February and 15 April 1986-87. Composite samples were collected from a minimum of 20 plants throughout the study area and consisted of plant parts believed to be selected by elk. Plant parts selected by elk, which in all cases consisted of current year's growth, were determined by examining adjacent browsed plants. Samples of deciduous shrubs collected during November contained both leaves and stems, which were separated prior to conducting nutritional analyses. Shrubs collected later in winter consisted of stem material only. All samples were oven-dried at 40 C within 6 hours of clipping, were ground through a 1 mm sieve, and were then stored in airtight plastic bags at room temperature until they were analyzed. IVDDM was determined using the two-stage procedures of Goering and Van Soest (1970) using inoculum from a fistulated heifer maintained an ryegrass hay and supplemental grain. Crude protein was measured using the micro-Kjeldahl technique (Horowitz 1980:858).

Diet Composition and Quality

Composition and quality of winter diets of elk were compared between an unmanaged forest ecosystem within MORA and a managed forest ecosystem adjacent to the park. Four composited fecal samples were obtained on each site on or about 1 November, 15 January, 15 February, and 15 April 1986-87 for dietary analyses. Composited samples consisted of 10 fecal pellets from each of 5-8 individual elk within a herd. Composited samples were obtained from each of 4 replicate elk herds an each study site (4 sample dates x 2 study sites x 4 replicates = 32 samples). All samples were stored frozen until they were oven-dried and ground through a 1-mm sieve in preparation for analysis.

Diet composition was determined from microhistological examination of fecal fragments following procedures of Sparks and Malechek (1968). Frequencies of occurrence of each plant species were determined from 20 microscope fields-of-view on 20 microscope slides prepared from each composited fecal sample. Frequencies were converted to percentage relative density (Fracker and Brischle 1944), which was assumed to be proportional to oven-dried weight.

Relative preferences of major forage classes in elk diets were assessed using Ivlev's Electivity Index (Ivlev 1961). Relative preference indices (RPI) (i.e., Ivlev's index) compared proportions of forage classes in diets to proportions of each forage class available in the environment. Available forage was determined from availability in each plant community weighted by the area of each community within the composited home range of elk (Cooper 1987). RPI of forage classes may range from -1, indicating complete absence in the diet, to +1, indicating maximum preference.

Nutritional quality of elk diets was estimated from botanical composition of diets and nutrient content of specific forages. Dietary levels of IVDDM and CP were computed following Westoby (1974) as the nutrient value of each forage weighted by its proportion in the seasonal diet. If nutrient data were absent for a specific forage, mean levels for similar species were substituted.

Two-factor analysis of variance was used to determine significant seasonal and site differences in composition and quality of elk diets. If diets varied among seasons, Fisher's protected LSD test was used to determine significance of all pairwise seasonal comparisons.

Relative Forage Values of Plant Communities

Traditionally, forage values have been con pared among plant communities on the basis of available forage biomass. Such comparisons may be misguided if forages vary widely in digestible energy or in palatability. Therefore, we derived an index for comparing relative forage values of plant communities based on estimates of forage biomass, digestibility, and dietary preferences of elk. Seasonal forage value indices (FVI) were computed for each plant community as the sum of digestible dry matters of forage classes weighted by forage preference indices of elk. Specifically, FVI was computed for each plant community as follows:

It was necessary to rescale Ivlev's RPI, as described in the previous section, so that it ranged from 0, indicating complete avoidance, to 1, which indicated maximum forage preference. Although influences of forage availability and nutrient content on diet selection remain poorly described in the literature, we believe that the above formulation, although simple, describes the nutritional interactions of nutrient quality and preference in a biologically reasonable way.

Results

Forage Biomass Trends

Seasonal trends of forage availability were influenced strongly by variation in snowpacks. Snowfall occurred sporadically throughout winter, but snow accumulated for only a two-week period prior to and during the January vegetation sampling period. Accumulation of snow averaged 25 cm under open canopies, 22 cm in alder communities, 20 cm in 30-yr-old Douglas-fir communities and 14 cm under old-growth forests. Excluding the January sampling period, our data reflect snowfree conditions during the remainder of winter.

Biomass of available forage was greatest during November and was least during January following a snowfall (Table 2.1). Biomass increased from February to April following disappearance of snow and initiation of green-growth. Mid-winter snowfall covered nearly all forbs, grasses, ferns and abscised leaves during January. Many grasses and sedges were matted down following snow melt; thus, availability of grasses and forbs remained low in February. Availability of grasses and forbs increased from February to April, reflecting an interplay between rapid plant growth and intensive cropping by elk. Evergreen shrubs, primarily low- lying (< 50 cm) salal and Oregongrape were largely covered by snow in January, although elk were observed pawing through snow to feed on low evergreens during this period, so our estimates of availability may underestimate the true amounts accessible to browsers. Availability of deciduous shrubs and conifers were less affected by snow than were low evergreen shrubs, and availability was relatively constant throughout winter.

Post-logging patterns of forage succession differed between mesic bottomlands and xeric uplands in the White River drainage (Fig. 2.1). Grasses, forbs and deciduous shrubs were more abundant in mesic than in xeric plant communities. Grasses and forbs were abundant for up to 40 years in grass-sedge communities that developed on wet sites after logging. Standing water, dense herbaceous vegetation, and intensive herbivory all appeared to inhibit overstory establishment and to perpetuate a long-lived herbaceous community. Grasses and forbs were also abundant in red alder communities that occurred extensively in bottomlands of the White River following logging. Grasses and forbs decreased in red alder communities approximately 20-40 years following logging, after a dense overstory of red alder had developed. Although the development of a red alder overstory reduced availability of grasses and forbs, abscised alder leaves were an abundant and important forage for elk during leaf-fall in October and November.

Figure 2.1. Successional patterns of available current annual growth (g DM/m2) following logging on mesic and xeric sites along the White River, Washington. Successional patterns shown for grass-forbs in mesic seres correspond to grass-sedge (a) and red alder (b) successional pathways (see text). Successional patterns for xeric seres correspond to unthinned 20-40 year old Douglas-fir stands.

The xeric sere was strongly dominated by evergreen shrubs, primarily salal and Oregongrape (Fig. 2.1, Table 2.1). Grasses, forbs and deciduous shrubs reached peak biomass during the first 15 years following logging. After approximately 15 years, regenerating Douglas-fir shaded out herbs and shrubs. Evergreen shrubs increased in the early stages following logging and remained abundant in mid- and late-seral stages of succession.

Successional patterns of forage development were variable in 20-40 year-old Douglas-fir forests on xeric uplands. Conifers, deciduous shrubs, and forbs were more abundant on poorly stocked Douglas-fir forests, which developed on poor sites, than on more productive sites (Table 2.1). Greater sunlight and abundant low branches of Douglas-fir resulted in greater availability of forage in poorly stocked forests.

Thinning practices produced negligible forage benefits for elk in 20 year-old Douglas-fir forests (Table 2.1). Conifers were more available in thinned than in unthinned forests; however there were only slight differences in availability of deciduous shrubs, evergreen shrubs and grasses. Understory responses to thinning, if any, would have persisted only a short time until canopies reduced by thinning once again closed over. It appeared that herbaceous forages important to elk had already declined by the time stands were thinned at approximately 18-20 years of age. Additionally, heavy accumulation of slash in thinned forests may have reduced sunlight and hindered forage.

Forage Quality

IVDDM and CP contents of forages followed the same seasonal pattern; each tended to be highest during November and April and lowest during midwinter (Table 2.2). Grasses, forbs, and deciduous shrubs increased in nutritive value between mid-winter and April during spring green-up. Grasses and forbs contained the highest level of IVDDM and CP throughout winter, reflecting low lignin contents and high proportions of cell contents (Cook 1972). Aquatic forbs, such as water parsley (Oenanthe sarmentosa and American veronica (Veronica americana) were succulent all winter and provided a limited, yet highly nutritious, winter forage on hydric sites. Evergreen shrubs, conifers, and deciduous shrubs contained low levels of IVDDM and CP during winter. Among browse species, trailing blackberry (Rubus ursinus) was the most nutritious. Swordfern (Polystichum munitum) was high in CP and low in IVDDM. In contrast, horsetail rush (Equisetum arvense), here considered a fern, was highly digestible during spring green-up.

Diet Composition and Quality

Seasonal differences in diet selection by elk reflected seasonal changes in forage availability. Elk consumed more forbs, grasses and deciduous shrubs during autumn, when a variety of foods were available, than during winter (Fig. 2.2). A wide variety of forbs and grasses were eaten during fall (Appendix III). Abscised leaves of red alder and black cottonwood (Populus trichocarpa) were the most prevalent deciduous shrubs in the fall diets (5-13% of diet). Other important shrubs included willows (Salix spp.), huckleberry (Vaccinium spp.), trailing blackberry, and salmonberry.

Figure 2.2. Mean percentages (n = 4, ±SE) of major force classes in the season diets of elk inhabiting unmanged (____) and managed (----) forest ecosystems in the White River, Nov 1986-April 1987. Different letters between seasons indicate significant dietary differences (LSD test, P < significant site X season interaction (ANOVA, P < 0.05).

Elk consumed more conifers and evergreen shrubs during winter than during fall and spring (Fig. 2.2), particularly when snow covered low-lying forages during January. Important evergreens included Pacific yaw (Taxus brevifolia), western red cedar, western hemlock, trailing twinflower, and salal (Appendix III). Elk switched from eating evergreens to eating a variety of forbs and grasses as soon as green grasses and forbs were available during spring (Fig. 2.2, Appendix III).

Relative preferences of elk for the major forage classes were ranked each season using Ivlev's RPI (Fig. 2.3). Although there was considerable seasonal variation in forage preference, the following general ranking of preference was evident: grasses and forbs > ferns > deciduous shrubs > evergreen shrubs > conifers. Relative preferences of forage classes were not correlated with either average IVDDM or CP of forage classes (P > 0.05).

Figure 2.3. Seasonal Relative Preference Indices (RPI) of forage classes in the White River, Nov 1986-April 1987.

Diets of elk differed between old-growth, unmanaged forests of MORA and nearby managed forests (Fig. 2.2). Differences appeared to be related to differences in forage availabilities. For example, grasses and forbs were more abundant in seral than in climax stages of forest succession (Table 2.1), and they were more prevalent in the diets of elk in managed than in unmanaged forests (Fig. 2.2). Evergreen shrubs were more abundant in climax than in seral communities, and they were more prevalent in the diets of elk in unmanaged forests of MORA than in adjacent managed forests. Additionally, abscised leaves of alder and cottonwood were more prevalent in the fall diets of elk in MORA than in the managed forests. Perhaps greater abundance of preferred grasses and forbs in the seral forest resulted in less use of deciduous leaves.

Seasonal and site differences in forage selection produced similar differences in dietary quality. Diets of elk an both study sites were highest in IVDDM and CP during April; they were intermediate during November, and they were lowest during January and February when elk ate mainly evergreen browse (Fig. 2.4). Highly nutritious spring diets reflected a high proportion of nutritious grasses and forbs in the diets.

Figure 2.4. Mean percentages (n = 4, ±SE) of in vitro digestible dry matter (IVDDM) and crude protein (CP) in the seasonal diets of elk inhabiting unmanged (____) and managed (----) forest ecosystems in the White River, Nov 1986-April 1987. Different letters between seasons indicate significant dietary differences (LSD test, P < 0.05). * indicates a significant site difference and ** indicates significant site X season interaction (ANOVA, P < 0.05).

Nutrient quality of elk diets was consistently greater in managed forests adjacent to MORA than in unmanaged forests within MORA (Fig. 2.4). The magnitude of that site difference varied seasonally (significant site x season interaction) for both IVDDM and CP. For both IVDDM and CP, dietary quality increased earlier in spring and more rapidly in the managed forest than in MORA. Seral communities in the managed forest provided a greater abundance of nutritious herbaceous forage earlier in the spring than did old-growth forests.

Relative Forage Values of Plant Communities

Seasonal Forage Value Indices (FVI) of plant communities are presented in Fig. 2.5. Each vegetation class exhibited the same seasonal trend; FVI decreased from November to January, increased following snowmelt in February and increased further following green-up beginning in February.

Figure 2.5. Seasonal Forage Value Indices (FVI) of 14 plant communities on xeric and mesic sites along the White River, Nov. 1986-April 1987. See text for discussion of FVI.

Overall, forage values were greater on mesic bottomlands along the White River than on xeric uplands within each age-class (Fig. 2.5). High forage values among mesic communities was related to abundant grasses and forbs, particularly during November and April. True forage values of mesic communities during April may have been greater than was actually measured because intensive grazing by elk reduced standing crop of many highly preferred grasses send forbs.

Forage values decreased with stand-age in both the xeric and mesic seres; however, the reduction in forage value was more pronounced in the xeric sere (Fig. 2.5). By 20 years of age overstory closure had reduced understory forage values in each of the xeric vegetation classes. Forage values remained high in 20-40 year-old red alder, grass-sedge and black cottonwood communities in the mesic sere. In both seres, mature forest communities contained lower forage values than did seral communities. Midseral Douglas-fir forest (~120 years) provided negligible forage values for elk.

Discussion

Post-logging successional patterns in the White River drainage are comparable to patterns previously reported for xeric uplands and mesic bottomlands in the Cascade Mountains (Long and Turner 1975, Hanley 1984). Succession on xeric uplands produced the now-classic picture of forage development that has emerged from comparable studies in the Pacific Northwest; i.e., clearcutting produces a pulse of forage that persists for approximately 8-20 years until the developing overstory shades out understory vegetation (Long and Turner 1975, Wallmo and Schoen 1980, Hanley 1984). Once overstory crown closure is complete, understories are often nearly devoid of elk forages. Understory development after complete crown closure was a long-term process marked by gradual thinning of forest overstories, increased lighting of the forest floor, and gradual increase of shrubs, primarily salal. Descriptions of succession from throughout the Cascade Mountains suggest that this pattern applies widely to montane forests in the Douglas-fir zone.

Successional patterns of forage development were variable on mesic bottomlands of the White River. Divergent successional pathways in riparian forests were probably related to complex patterns of alluvial deposition and soil moisture. Grass sedge communities became established following logging in low-lying wet sites, whereas second-growth alder communities became established in better-drained bottomlands. Overall, successional pathways on bottomlands were distinct from those of uplands in containing a long-lived seral stage characterized by high productivity of grasses, forbs and deciduous shrubs.

Seasonal nutritive qualities of forages and dietary selections of elk were similar to broad patterns reported throughout the Pacific Northwest. Our findings upheld the general belief that grasses and forbs are preferred winter forages of elk, but that elk often consume less-digestible browses whenever availability of preferred forages limits rate of forage intake. In the White River, elk ate more herbaceous forage during autumn and spring than during winter, following trends in availability. Elk also consumed more grasses and forbs in managed forests, where seral bottomland communities provided abundant herbage, than in MORA. Use of evergreen browse, in contrast, was greatest in unmanaged forests where seral forest communities were limited, and in managed forests following a deep snow that restricted availability of herbaceous forage. Abscised leaves of alder and cottonwood also were important alternate forages of elk during autumn, especially in MORA where grasses and forbs were comparatively scarce. Alder also comprised 22-50% of the autumn of diets of elk and deer in Olympic National Park (Leslie et al. 1984), which suggests that deciduous leaves may be of greater importance to cervids during autumn than was previously believed.

Dietary CP levels in the White River drainage (5-8%) were within the range of values previously reported for Roosevelt elk during winter in the Pacific Northwest (7-8%, Janz 1983; 8.3%, Leslie et al. 1984; 8% (November), Merrill et al. 1987). Most investigators have considered those levels adequate for maintaining body weight over winter, assuming forage intake, digestible energy, and protein digestibility is adequate (Janz 1983, Leslie et al. 1984). Few studies, however, have examined those assumptions critically. Dry-matter intake may be severely constrained during winter by low turnover rates associated with high lignin-cutin contents, cell structural characteristics of browse (Spalinger et al. 1986), or by limited ability of ruminants to detoxify or absorb phenolic compounds in browse (Robbins et al. 1987). Additionally, endogenous protein may be catabolized to help mitigate the effects of energy deficiencies. Furthermore, digestibility of protein is reduced depending upon concentrations of insoluble protein-binding phenolic compounds in the diet (Robbins et al. 1987). For example, Robbins et al. (1987) reported that conifer rations containing 6-8% CP contained only 0-3% CP that was digestible by deer. These findings suggest that dietary protein in the White River, as elsewhere in the Pacific Northwest, may be deficient.

Dietary IVDDM of elk during winter (35-40%) was above dietary levels reported for elk inhabiting nearby coastal regions (31-34%, Janz 1983; 25%, Leslie et al. 1984). Those differences may reflect different inoculum sources used in in-vitro digestion trials (Milchunas and Baker 1982) and differences in forage sampling methods, as well as real regional differences. Our results, however, agree with previous studies concluding that IVDDM of winter diets was insufficient for elk to meet daily energy requirements and animals could be expected to decline in body weight ever winter. Daily energy requirements of ruminants have been approximated as 150 kcal/kg^0.75/day (Robbins 1983). An average 250-kg cow elk, therefore, would require a daily intake of 5.1-5.9 kg DM/day, assuming a gross energy content of forage equal to 4.5 kcal/kg, a metabolizable energy coefficient of 0.85 (Hobbs et al. 1982), and assuming percent IVDDM of forage ranging from 35-40% (this study). Merrill et al. (1987) estimated that "average" elk were capable of consuming only 3.2 kg DM/day during November near Mount St. Helens, Washington. That estimate was derived from activity budgets, forage availability and foraging rates of elk on Mount Saint Helens, but it provides a general indication of relative deficiencies of winter diets in the White River.

Seasonal and site variability of dietary nutrients were related to availability of herbaceous forages. Dietary levels of IVDDM and CP were both greater in a managed forest, where herbaceous forage was comparatively abundant, than in an unmanaged forest ecosystem. In addition, dietary levels of IVDDM and CP decreased in early winter in both sites as availability of herbaceous forage decreased, and it increased markedly during spring green-up. Green-up and associated nutritional benefits, however, occurred approximately 2-3 weeks earlier in the managed than the unmanaged forests.

It is tempting to equate site differences in dietary qualities solely to influences of past forest management activities; however, geographic differences between study sites also played a role. Elk winter range within MORA was approximately 300m higher in elevation than winter range outside the park. After one January storm, snow depth averaged only 5 cm deeper on MORA winter range than on lower-elevation managed ranges. Accumulations resulting from that storm, however, persisted approximately two weeks longer within MORA than at lower elevations outside the park. Unfortunately, we cannot determine the relative influences of elevation and forest management activities on dietary nutrient qualities of elk inhabiting these two winter ranges.

Cause of nutritional differences notwithstanding, enhanced nutrition of elk wintering outside MORA could have important consequences for populations of elk summering within the park. Improved winter nutrition would result in reduced weight-loss and mortality of cow elk, reduced pre-natal and neo-natal losses of calves (Thorpe et al. 1976), greater fall weights of cows, and increased ovulation and pregnancy rates (Trainer 1971). These results were reflected in observations of higher calf:cow ratios and rates of population growth of elk inhabiting lower-elevation managed ranges outside MORA than of elk wintering within the park (unpubl. data, K. Cooper).

Relative foraging values of old-growth and immature forests are at the core of contemporary issues regarding management of remaining old-growth forests in the Pacific Northwest (Schoen et al. 1981, Bunnell 1985). Many wildlife managers believe that forage values of old-growth forests have been underestimated in the past relative to values of seral forests. Measurable benefits of old-growth forests have included greater standing crop biomass of forage (Bunnell and Jones 1984), reduced snow accumulation and greater availability of forage (Jenkins et al. 1990; Harestad et al. 1982), increased nutritional value of browse (Hanley et al. 1987, Happe et al., in prep.), and greater availability of arboreal lichens and litterfall in old-growth than in immature forests (Stevenson and Rochelle 1984).

Forage values measured along a successional sequence of plant communities in the White River, however, suggested that many seral forests provided forage values superior to those of old-aged forests during mild winters. Admittedly, we did not measure standing crops of arboreal lichens and litter, which Stevenson and Rochelle (1984) reported often exceeded biomass of ground-rooted vegetation in old-aged forests on Vancouver Island, BC. Nor did we compare forage values during prolonged periods following deep snow. Additionally, we measured forage values based on available biomass, forage digestibility, and forage preferences of elk, rather than based on just forage biomass. Previous studies reviewed by Bunnell and Jones (1984:413) indicated that old-age forests contained 2-20X as much biomass of key winter forages of black-tailed deer (Odocoileus hemionus columbianus), primarily salal, than did immature forests. We, too, measured 2-25X as much evergreen shrubs in old-age than in immature upland forests of the Douglas-fir zone. We chose, however, to discount the importance of salal and Oregongrape in estimations of forage values because they were of such low preference by elk. Forage values of plant communities, one should expect, would differ for black-tailed deer and elk because of interspecific differences of foraging strategies and forage preferences (Hanley 1984).

Comparisons of forage values between xeric and mesic seres in the White River drainage revealed that forage values are influenced by a variety of unmeasured factors besides stand-age. Forage values computed for communities within a single age class often varied as much as between age-classes. For example, during November forage values of communities 11-20, 21-40 and >200 years old, differed by as much as 2.5X, 4.2X, and 2.0X within an age-class, respectively (Fig. 2.5). We suggest, therefore, that standard notices of post-logging forage succession are too simple to warrant widespread application in forest management. Clearly, additional systematically conducted research is needed on successional trends of elk forages in a variety of riparian and upland forest associations throughout the range of elk.

The data presented above leave little question that clearcuts provided a significant increase in forage values for elk during mild winters. Clearly, however, this should not be construed to mean that clearcuts are better elk habitat than old-age forests. In both a managed and an unmanaged forest, elk foraged principally upon conifers and evergreen shrubs following a winter storm. Evaluation of Fig. 2.5 reveals that 12" of snow virtually negated forage values of many immature forests, which equalized forage values of different aged stands. Deeper or more prolonged snow than was observed in this study would result in further increases of forage values in old-age relative to immature forests. Previous studies of big game-forest management relationships have revealed that snow intercepting capabilities of old-age forests may increase accessibility of browse (Harestad et al. 1982), reduce energy costs of foraging (Parker et al. 1984), and enhance abilities of elk to conserve energy during nutritionally restrictive winters. Additional behavioral preferences of elk for old-age forest communities are likely to occur, but have been difficult to determine.

It seems clear from the results of this and comparable studies of elk-habitat relations that a mosaic of immature and old-age forest habitats is optimum for elk. Excessive harvesting of old-aged forests may increase ecological carrying capacity (sensu Caughley 1976) for elk during successive mild winters, but would likely result in greater density-dependent and -independent winter losses of elk during severe winters. Optimum proportions of forest ages, therefore, will vary as functions of prevailing snow characteristics, site-specific successional patterns of forage development, and big game management objectives.