Peary caribou , muskoxen and Banks Island forage : Assessing seasonal diet similarities

Peary caribou (Rangifer tarandus pearyi) and muskoxen {Ovibos moschatus) on Banks Island had considerable similarity in their annual diets, with monthly similarities ranging from 17.8-73.3%. Diet similarity was more pronounced in areas of high muskox density {ca. 1.65/km2) than in areas of low muskox density {ca. 0.4/km2). Willow (Salix arctica) and sedge (Carex aquatilis and Eriophorum spp.) represented >80% of the monthly diet of muskoxen. The caribou diet was more diverse, and was dominated by sedge, willow, Dryas integrifolia, and Oxytropis maydelliana, Lichen use was rare, likely as a consequence of low availability on Banks Island. Lichen standing crop was estimated at 2.96 g/m2. The differences in muskox diet between high and low density areas could not be explained by differences in forage distribution or standing crop. We discuss diet similarities of caribou and muskoxen and potential consequences for the current Peary caribou population in relation to winter weather conditions and increasing muskox density.

The decrease in caribou numbers was attributed to a variety of factors including severe winter weather, predation, harvest, inter-island movements, and competition with muskoxen as reviewed by Nagy et al. (1996).The actual cause or causes of the decline remain unknown.However, in Rangifer, 17 (1), 1997 order to manage the recovery of the Peary caribou population on Banks Island, it is important to assess dietary overlap and the potential for food competition between the 2 species given the current animal numbers and forage abundance and distribution.
Caribou and muskoxen are the only ungulate species successfully occupying arctic tundra environments.Caribou and muskoxen have different morphological adaptations which have presumably enabled them to utilize fotage resources with little overlap (Klein, 1992).Muskoxen represent the classic grazer (Hofmann, 1989).With a large body size and gut capacity, they are capable of processing large amounts of low quality forage.Caribou are representative of a mixed feeder type, and are inrer-mediate between roughage feeders and concentrate selectors (Hofmann, 1989)-Their smaller body size and smaller gut capacity, combined with a higher fasting metabolic fate than muskoxen (Tyler & Blix, 1990), require them to pursue a more selective feeding strategy.
In contrast to muskoxen, caribou meet theif nutritional tequirements by a relatively rapid rate of passage of highly digestible forage (Klein, 1992).
Lichen is an important winter food for barrenground caribou on the mainland, but in the high arctic islands which support low lichen biomass, caribou use other forages, usually willow and graminoids (Reimers etal., 1980;Klein, 1992).In west Greenland, where lichen biomass was apparently depleted by overgrazing (Staaland & Olesen, 1992), both muskox and caribou diets were dominated by monocots.
In areas inhabited by muskoxen only, willow became an important summer diet component (Thing et al, 1987).Therefore, both animals demonstrate the ability to utilize a variety of forages when availability dictates.
Muskoxen can clearly make good use of high protein, low-fibre foods (White et al, 1984;Adamczewski et al, 1994) and even though they show many attributes of classic grazers they can be quite selective in their feeding (Oakes et al, 1992).Despite their relatively wide muzzles muskoxen ate remarkably adept at finding the leafy portions of forage and rejecting larger stems (J.Adamczewski, pers. comm.).
Reconsideration of data on muskox and caribou ecology implies that competition for food may occur where muskox concentrations are high (McKendrick, 1981), and that overlapping winter diets may adversely affect caribou numbers (C. Olesen, unpubl.data).
In this paper we report pteliminary findings on the Peary caribou and muskox diers, monitored on a monthly basis, the current forage disrribution and standing crop of the 4 major terrestrial habitats in areas of high and low muskox density on Banks Island, and compare our findings with previous work.

Study Area
Banks Island is the most western island in the Canadian Arctic Archipelago and covers approxima- Ptecipitation is low with an annual mean of 9 cm (Zoltai et al, 1980).Sachs Harbour is the only permanent settlement on the Island.Zoltai et al. (1980) provided a general overview of the geology and glacial history.
There are 4 major terresttial habitats: i) wet sedge  Faecal samples were thawed, air dried for 24 hours, oven dried at 60°C for 48 hours, and ground through a 1 mm screen with a centrifugal mill.
Subsamples (1 g) were forwarded to the Composition Analysis Laboratory, Ft.Collins, Colorado for analysis.Diet composition was determined by analyzing plant fragments (Sparkes & Malechek, 1968) according to Hansen et al. (1976).The microhistological technique has inherent limits, such as an inability to separate some species, and a limited percent of identifiable fragments in the slides (Johnson et al, 1983;Barker, 1986).We deemed this method suitable for this study, since differing proportions of forage classes, not changes in individual species composition wete of importance, and this met- ).There were traces of unidentifiable forb material in 6 samples, (4 caribou and 2 muskoxen, ranging from 0.33 -0.89%) which were included in the other category, and traces of unidentifiable grass material in 51 muskox samples, (ranging from 0.46 -1.93%) which were included in the grass category.We present results from samples collected prior to September, 1994.We used the Renkonen index (Renkonen, 1938;Krebs, 1989)  Occurrence data were collected during each clipping episode and again in early August.We lumped fotages into the same 8 classes as above.We compared the occurrence of forages in similar habitats between low and high density muskox areas using a X 2 contingency analysis.

Studies on various
Canadian arctic islands in the 1960's and 1970's suggested that caribou and muskoxen coexisted with little resource overlap (Tener, 1965;Kevan, 1974;Thomas & Edmonds, 1984).Studies were conducted in the early 1970's to compare diet composition of the 2 ungulates on Banks Island (Wilkinson et al., 1976;Shank et al., 1978) and the Patty Islands (Patker, 1978)  Knowledge of forage availability is an integral requirement for documenting competition for food (Klein & Staaland, 1984;Gunn, 1990), yet data pertaining to forage availability was conspicuously absent from previous studies.Therefore , Wilkinson et al.*s (1976) conclusions about the lack of competition between muskoxen and caribou on Banks Island are not surprising.Their study was conducted in summer, when forage quality and availability are highesr.

Currently, the caribou population on Banks
Island is 16-fold smaller and the muskox population 16-fold larger than in 1972, thetefore previous results and conclusions are likely to differ from ours.
Although Wilkinson et al. (1976) and Shank et al. (1978) tuled out competition for food between muskoxen and caribou, their results demonstrated substantial similatity (PS) in diet between the two species during borh wintet (March) and summer (August) (Table 1.).PS in the diets found in the 1970's may be inflated somewhat in relation to our findings because the diet was only partitioned into 5 components versus our 7 components.Our findings indicated: i) considerable similarity in diet between the 2 ungulates regardless of muskox density, ii) a noticeable absence of lichen in the diet of caribou, iii) a noticeable absence of grasses in the diet of both catibou and muskoxen, and iv) sedges and willows wete found throughout all habitats, but grasses were found mostly in upland habitats.
The diet of caribou was more diverse rhan that of muskoxen, being dominated by 4 major forage groups, sedge, willow, rose/saxifrage, and legume.
There was a distinct seasonal shift in proportions of willow, rose/saxifrage, and legume.Willow use was greatest during June through August, presumably when new growth leaves and stems are high in crude protein and energy content.Rose/saxifrage and legume, as well as sedge, predominated from October to May.Sedge and rose/saxifrage occur in all three upland habitats and wet sedge meadows on Banks Island, whereas legumes occur only on the upland habitats.The gtass component of the diet reported in the eatly 1970's (Wilkinson et al., 1976;Shank et al, 1978), was noticeably absent in the 1990's.Rose/saxifrage, legumes, and other forbs 14 appear to have teplaced this component of the diet.
The lack of lichen in the diet was consistent with previous findings (Shank et al, 1978), and is likely related to low availability.Lattet & Nagy (1996) found similar percenrages of lichen in the rumen contents and faecal material of mainland barrenground caribou during wintet indicating that the proportion of lichen in the diet, determined from the analyses of faecal plant fragments, was not significantly influenced by high lichen digestibility during winter.The 2.96 g/m 2 we report is almost 5-fold lower than the 14 g/m 2 reported on Coats Island (Ouellet et al, 1996), an island considered to have a low standing crop of lichen.Larter & J. Nagy, unpubl. data).
The occurrence of willow in both the wintet (March) and summer (August) diets of muskoxen in the high density area was 2-3 times greater than that found by Wilkinson et al. (1976) and Shank et al. (1978) in the 1970's.This difference cannot be attributed to the difference in technique used to determine diet.The macroscopic technique used by Wilkinson et al. (1976) and Shank et al. (1978)  where the diet of muskoxen is almost exclusively sedge (Raillard, 1992).Increasing competition for sedges may have resulted in an inctease in rhe use of other forages, like willow.Smith (1996) tely 70,000 km 2 .The climate is Arctic Maritime along coastal ateas where weather stations are located, tending toward Arctic Desert inland.Winters are long and cold; summers are short and cool.10

(
n=10) collected from 22 groups of muskoxen was similar to the mean sedge component determined by lumping the individual samples across groups, 62.9 vs. 64.2%respectively.Also, the SD associated with the mean of 70 individual samples (34.31) was higher than any SD associated with one of the 22 groups (average SD = 7.98, range SD 0.042 -32.3 with 16 groups having SD < 7.98).Therefore, we assumed that composite samples of fresh (< 4 hour old) muskox faeces from a number of individuals were representative of a group of muskoxen, and that by sampling groups instead of individuals we would be able to get a better measure of the used the composite sample from a muskox group as the sample unit.Initially the sample unit consisted of samples pooled from 5 pellet groups.This was reduced to 3 pellet groups/muskox group.We collected muskox faecal samples from mixed sex and age groups during all months except January and September in the high density area.The number of groups sampled each month ranged from 1-10, representing between 5 and 152 animals.In the low density atea, samples were collected during April, June, July, August, and October.The number of groups sampled each month ranged from 1-7 representing between 30 and 89 animals.The location, habitat the samples were collected from, and group size was recorded.We present mean monthly diet composition of muskoxen with groups as the sample unit, weighted by the numbet of individuals in a group.June to August data are pooled across years.
hod had been used in previous work on Banks Island.We used the following fotage classes: sedge (Cyperaceae), willow (S. arctica), grass (Gramineae), rose/saxifrage (Rosaceae and Saxifragaceae), legume Rangifer, 17 (1), 1997 (Leguminosae), lichen {Cetraria spp., Cladonia spp., Cladina spp., Peltigera spp., and Thamnolia subuliformis), and other (other forbs, moss, and Equisetum spp. to compare monthly percent diet similarity (PS) between Peary caribou and muskoxen in both high and low density areas.Forage availability was assessed by 2 measures, standing crop and the presence/absence of forages in plots.Twelve 0.125 m 2 plots (Wein & Rencz, 1976) were clipped along 2 or 3 permanent transects in each habitat at both camps at 3 times during the snow free period: mid-June, mid-July, and mid/late-August.Transect lengths ranged from 90 to 450 m depending upon habitat type.Plot locations were assigned systematically without replacement based upon total transect length in each habitat.We clipped the following forages at ground level: sedge, grass, legume (Oxytropis spp., Hedysarum mackenzii., and Astragalus alpinus), ericad (Cassiope tetragona), rose/saxifrage (Dryas integrifolia, Saxifraga spp.), and other forbs.Lichen was plucked from the substrate.Only currenr annual growth of willows was clipped.In the laboratory, sedge, grass, and Cassiope tetragona samples were separated into their live and dead components; willow was separated into leaf/bud and stem components.All samples were oven dried at 60°C for 48 hours and weighed to 0.1 mg.We used ANOVA to determine whether there were habitat, sampling time (June, July, August), area (high vs. low muskox density), and year effects on the standing crop of each forage.Because there were no year effects, we pooled data across years.Since growth rate of lichen is low, we calculated the mean standing crop from all the plots clipped in upland habitats (UB, HT, and SB) pooled across sampling time and year.
was the dominant component during June to August.Legume and rose/saxifrage were dominant components duting December to April.The sedge component remained relatively constant at ca. 25% each month.Lichen and grass use was negligible.The annual diet of muskoxen was dominated by sedge and willow in both high and low density areas (Fig.2).There was a high percent similarity (PS>89) of muskox diets between high and low density areas with the exception of a larger proportion of willow in the June and legume in the July diet in the high density area which reduced PS to 52 and 76, respectively (

(
Carex spp.) which were abundant in wet lowland habitats.Caribou summer and winter diers were both dominated by willow (Salix spp.), forbs, grasses and sedges which were abundant in drier upland habitats, but the proportion of willow in the winter diet was reduced.
The muskox diet was dominated by seasonally varying proportions of sedge and willow regardless of muskox density.Standing crop and occurrence of sedges in wet sedge meadows was similar between areas, and therefore could not explain the latger proportions of sedge in the diet of animals in the low density area.The increased occurrence of legumes in the diet of animals in the high density atea during July may be related to availability.Legume standing crop and occurrence in upland barren and hummock tundra habitats was greater in rhose habitats present in the high density area.July is the peak in available etude protein of legumes (N. demonstrated that wet meadows subjected to grazing by a Rangifer, 17 (1), 1997 high density of muskoxen had decreased net aboveground primary productivity, and that over-compensation of plant growth did not occur.Increasing use of willows, especially during winter, by a rapidly increasing population of muskoxen is of immediate concern.During winter, willows are dormant and many years of growth can be removed, possibly more than could be replaced during one growing season.Willow twigs of >4 years growth have been found in the rumen of adult muskoxen during April (N.Latter & J. Nagy, unpubl.data).Continued cropping of most previous years twigs and buds may stress willow plants beyond tecovery and increase plant mortality.Reductions in new growth of willows during June to August may have serious consequences for caribou who utilize them as a primary food source during this time of lactation and body growth.The elevated winter use of willows may be somewhat highet than normal for our March and May data because the high density muskox area dara were collected in 1993.During winter 1992-1993, Banks Island had more snowfall than in subsequent winters, resulting in a deeper and denser snow cover of wet sedge meadows and upland battens during late winter.Muskox crater sites wete rare in wet sedge meadows, with the majority in upland habitats (N.Larter & J. Nagy, unpubl.data).These snow conditions may have forced muskoxen to feed more in uplands, where there is a higher proportion of willow forage, and therefore may have biased our data.Regardless, data collected in November, December, February and April of the following wintet showed much higher willow use than that found in the 1970's.Snow depth and density in wet sedge meadows and upland barrens was lower and nearer normal during the 1993-1994 winter (N.Larter & J. Nagy, unpubl.data).Much of the traditional wintering area of caribou overlaps the high density muskox area.Any dietary overlap with muskoxen may be magnified should harsh snow conditions occur Similar habitats provide relatively similar standing crops of eight major forage classes with the possible exception of legumes and Dryas integrifolia which are more prominent, both in occurrence and standing crop measures, in the high density muskox areas.However, because we have not completed habitat mapping in these ateas, these data must be treated with caution tegarding an absolute measure of forage availability.Higher frequency occurrence of forages does not necessarily indicate increased standing crop.Rangifer, 17 (1), 1997 Although caribou and muskoxen have different morphological and physiological adaptations which enable them to utilize forage resources with little overlap, our interpretation of data reported by Wilkinson et al. (1976), and Shank et al. (1978) suggests dietary overlap of forage classes occurred in the early 1970's on Banks Island.Dietary overlap between caribou and muskoxen has since become evident during 11 months of the year, is greater in areas of high muskox density, and may increase during winters with elevated snow depth and density.Currently, data cannot disprove or prove that forage competition has occurred or is occurring.However, given: i) increased willow utilization by a rapidly increasing muskox population, which may increase during harsh winters, and ii) the four-fold increase in density of a relatively sedentary potential competitor for food in traditional caribou wintering areas berween 1985 and 1994, the potential impact of muskoxen on the caribou wintet range and availability of willows may well be a factor limiting the recovery of the Peary caribou population.

Table 1 .
Table 1).From October to Match willow represented ca.20-48% of the monthly diet.Willow use peaked in May at ca. 70%.The Renkonen index of percent similarity (PS) Monthly PS of caribou and muskoxen (from both high and low density areas) ranged from 18 to 73-Monthly PS was generally >30.0 for most months and greatest in the high muskox density (Table1).Sedge (Carex spp.andEriophorum spp.) and willow made up substantial portions of the annual diet of both caribou and muskoxen (Figs.1 and 2). of the monthly diets of caribou and muskoxen in high density areas (C/MH), caribou and muskoxen in low density areas (C/ML), and muskoxen in high and low density areas (MH/ML) determined from this study.The PS of the sea-

Table 2 .
The percent frequency occurrence of various forages in the 3 upland habitats located in high and low density Rangifer, 17 (1), 1997