Population demography of Peary caribou and muskox on Banks Island , N . W . T . , 1982-1992

The Peary caribou population (excluding calves) o n Banks Island declined from 6 970 (± 1133) i n 1982 to 897 (± 151) i n 1991. The 1992 estimate was 1 005 (± 133). Percent calves i n the population varied among years (range 3.2¬ 31.1%). M e a n group sizes dropped from 5.0 i n 1985 to a l o w o f 2.0 i n 1989 and 1991. M e d i a n group sizes were significantly larger between 1982 and 1987 than between 1989 and 1992 (P<0.001). Large post-calving aggregations prevalent on the N W portion o f Banks Island i n the early 1980's were absent by the 1990's. T h e muskox population (excluding calves), increased from 29 168 (± 2104) i n 1985 to 52 959 (± 2240) i n 1992. Percent calves i n the population var i ed among years, (range 11.8-17.1%). These values may be underestimates, because calves are small and muskox groups sometimes form defensive circles i n response to aircraft. M u s k o x e n were distributed throughout the island during all surveys; however, the greatest increase i n density occurred i n the southern parts o f the island. K e y words: Pea ry ca r i bou , m u s k o x , B a n k s Island, p o p u l a t i o n d e m o g r a p h y


Introduction
During the past decade, caribou numbers decreased.Severe winter weather is believed to be the major cause of declining caribou numbers (Parker et ah, 1975;Gunn, 1992).Annual die-offs of 60-300 caribou were recorded following the winters of 1987-88, 1988-89, and 1990-91  Scat analysis indicates that both muskox and caribou are taken by wolves (Canis lupus); however, data on wolf numbers and predation rates are lacking.Inuvialuit in Sachs Harbour harvest wolves for subsistence use.Fewer than 10 wolves were harvested annually from 1988-89 to 1991-92; however, 50 wolves were harvested in winter 1992-93 (Larter & Clarkson, 1994).
This paper summarizes all Banks Island survey data for caribou and muskox collected during the 10-year period 1982-92.We document changes in population size and distribution of the two arctic ungulate populations and, where methodologies permit, we compare changes in group size and the proportion of calves in each population.

Study Area
Banks Island is the most western island in the Canadian Arctic Archipelago which covers an area of approximately 70 000 km 2 (Fig. 1).The climate is Arctic Maritime along coastal areas tending toward Arctic Desert inland.Winters are long and cold; summers are short and cool.Precipitation is low, with an annual mean of 9 cm (Zoltai et ah, 1980).
Rangifer, Special Issue No. 9, 1996 Sachs Harbour is the only permanent settlement on the Island (71° 59' N., 125° 17' W.).A general overview of the geology and glacial history can be found in Zoltai et al. (1980).
Coverage varied considerably, lowest being in when some blocks were surveyed with parallel lines 20 km apart (McLean et al, 1986), and highest in 1992 when some blocks were surveyed with lines 2.5 km apart (Nagy, J. & Fraser, P., unpubl.data).
Surveys were conducted in June and early July, except in 1992 when the survey was in late August.
The entire island was surveyed, except in when an area of high muskox density was omitted.
As a result, we believe the 1982 muskox population estimate is likely an underestimate.We estimated caribou and muskox numbers each year, except 1987 when muskoxen were not surveyed.The number of calves (both species) were counted on transect and their percentage in the population was calculated.

Verification of Original Survey Data
We verified the area of each stratum flown by digitizing 1:250,000 scale maps of Banks Island, entering the coordinates of the stratum boundaries, and calculating the resulting areas using a Lambert Azimuthal Equal Area projection in the Spans Geographical Information System (GIS).The accuracy of the Spans GIS was verified by entering the coordinates of an area of known size and then comparing it to the GIS area estimate.These resulting areas were then used to recalculate the survey estimates using the original observations and the Jolly (1969) method for unequal sized sampling units (Norton-Griffiths, 1978).The island-wide estimates of both caribou and muskox are reported.Because we recalculated survey areas, the estimates we report may differ from those found in the original survey reports.

Post-survey Stratification
In order to examine in more detail the changes in muskox density, especially in relation to harvest, we reanalyzed the 1985 and 1991 survey data.We placed the original observations from these surveys into the 8 strata that were used for the 1989 and 1992 surveys (Fig. 2).We recalculated the estimates for each of the 8 strata and the island-wide population, and report both values for 1985, 1989, 1991, and 1992.

Growth Rate
Total estimates for the island population and for each stratum were compared with corresponding data from each subsequent year using two-tailed Student's £-tests.For caribou, we compared all estimates derived from 1982 to 1992.For muskoxen, we compared estimates derived for each of the 8 strata from 1985 to 1991.For total island population comparisons we also included the 1992 estimate.
We calculated the instantaneous rate of population growth (r) for island and strata populations between surveys following Caughley (1977).

Group Size and Distribution
We used the Kruskal-Wallis (K-W) test (Gibbons, 1985) to assess differences in group size for caribou among years (1982, 1985, 1987, 1989, and 1991).
When K-W tests indicated significant (P<0.05)results, multiple comparisons (Gibbons, 1985) were conducted on mean class ranks to identify significantly different population subsets.We used an overall significance level of 0.25 for multiple comparisons.Median and mean group sizes are reported.
The post-calving distribution of caribou and muskoxen observed during the 1982 to 1991 surveys was plotted for visual comparisons.

Verification of Original Survey Data
The accuracy of the Spans GIS measurement was ±0.026%.In general, the difference between areas reported in the original surveys and those measured using the Spans GIS ranged between ± 8% for strata flown for caribou and from -7 to ± 6% for strata flown for muskoxen.The most notable differences were found for strata B, C and Thomsen River (Fig. 2).In 1985 the areas for strata B and Thomsen River were underestimated by 10 and 29% respectively, while stratum C was overestimated by 35%.
These errors continued through 1992.

Caribou
The Peary caribou population (excluding calves) decreased from 6 970 ( ± 1133) in 1982 to 897 ( ± 151) in 1991 (Fig. 3a).The corresponding densities decreased from 0.11 to 0.01 caribou/km 2 (Table 1).Rangifer, Special Issue No. 9, 1996 Concomitant with the population decline and decreasing group sizes, was the disappearance of post-calving aggregations that had been prevalent on the northwest portion of Banks Island in the early 1980's (Fig. 4).
Consistent positive growth occurred in strata A, B, Egg and Masik, whereas strata C, D, Parker and Thomsen showed both positive and negative growth during periods between surveys.All strata except for C however, showed an overall trend for positive growth during the period 1985 to 1991 (Table 3).
The proportion of calves in the total muskox population was less variable than in caribou: 11.8% in 1985, 14.5% in 1989, 14.0% in 1991, and 17.1% in 1992.Because calves are small and muskox groups tend to form defensive circles in response to the aircraft, the number of calves and therefore their proportion in the population are likely underestimated.
Muskoxen were distributed throughout the island during all surveys during 1982-1991 (Fig. 5); however the greatest increases in density occurred in strata A, D, Egg and Masik (Table 3).High, but stable, densities of muskoxen were observed in the Thomsen River strata during the period 1985 to 1991 (Table 3).Densities were variable, but generally low, in the remaining strata.

Caribou
Based on current and historical data, we cannot attribute the decline of the Peary caribou population on Banks Island to any one cause.Our review indicates that some of the apparent decline in cari-Ratlgifer, Special Issue No. 9, 1996 218 Rangifer, Special Issue No. 9, 1996 A loss of 60 animals due to weather caused winter mortality was documented (Fraser et al, 1992), accounting for an additional 3% of the animals lost.
However, the loss of 1342 non-calf caribou (76%) cannot be explained.
During both periods it is unlikely that the unexplained losses were a result of unreported harvest or winter mortality.The annual subsistence harvest by residents of Sachs Harbour ranged from 211-354 adult caribou during the late 1980's (Fabijan, unpubl.data), and was similar to that reported during 1962-72 (Urquhart, 1973).The traditional winter range is located in southwest Banks Island (Urquhart, 1973).Therefore, substantial die-offs on the winter range would have been detected by residents or others travelling in the area.
The current impact of wolf predation is unknown.The vulnerability of the primary prey relative to an alternate prey determines the impact of predation (Messier, 1994).In some areas of western North America, moose populations are the main prey for wolf populations which can remain at levels high enough to deplete caribou populations (Rangifer tarandus) populations.This scenario has been used to explain declines in the hunted Nelchina caribou population (Bergerud & Ballard, 1988), and various woodland caribou populations (Bergerud & Elliot, 1986;Edmonds, 1988).Seip (1992) documented that a woodland caribou population that lived apart from wolves and moose during summer, was therefore less vulnerable to wolf predation, had a lower adult mortality rate and higher calf survival than a similar woodland caribou population that lived with wolves and moose.
Whether or not one or both populations were hunted was not indicated.Larter et al. (1994) documented increasing wolf predation on moose caused by an increasing wood bison population which provided a substantial alternate prey source that maintained high wolf populations capable of depleting the more vulnerable prey population.Wolf predation has a greater impact on moose populations when the moose population is declining (Gasaway et al, 1983).
Peary caribou on Banks Island may be in a situation similar to that described for some hunted moose and caribou populations, where a high biomass of muskoxen supports an increasing wolf population.Wolf numbers were reduced on Banks Island during poisoning programs conducted during the 1950's (Zoltai et al, 1980) predation may be higher.
The impact of disease in unknown.There have been no reported incidences of harvested animals appearing diseased, but studies have not been conducted to directly address the issue.
Inter-island movement of Peary caribou may have contributed to declining numbers.Miller & Gunn (1978, 1980)  Rangifer, Special Issue No. 9, 1996 (Manning & Macpherson, 1958).This movement appears to have occurred when available forage was scarce (Urquhart, 1973).They noted that some of the animals returned later in the winter only to starve along with caribou that remained on the island.
During this time period Peary caribou were observed on the mainland coast near Baillie Island approximately 160 km south of Banks Island (Carpenter, A. & Wolki, F., pers. comm.).The caribou were weak and covered in ice.
These observations suggest that Peary caribou may normally move between Banks and Victoria Island as part of an inter-island population as suggested by Miller (1985, cited in McLean, 1992).
However, large scale movements from Banks Island may be driven by reduced forage availability during episodes of severe winter weather.This would have a net negative effect on the island population if these were one way movements.
Whether or not range conditions have deteriorated during 1982-1992 is unknown.Lichen has never been abundant (Wilkinson & Shank, 1974;Zoltai et al, 1980).Current research is examining range conditions in areas of high and low muskox density on and adjacent to the core caribou winter range.Preliminary indications are that lichen biomass is low throughout the range.
Although, competition between muskox and caribou has been considered unlikely in the past (Kevan, 1974;Wilkinson et al, 1976;Parker, 1978;Vincent & Gunn, 1981;Biddlecomb, 1992), reconsideration of caribou and muskox ecology and recent research indicates they may in fact compete for forage resources especially during winter or when muskox densities are high (McKendrick, 1981, White et al, 1981;Klein & Staaland, 1984).It is generally agreed that lichen is an important winter food for caribou, but in areas supporting low lichen biomass caribou switch to other forages, either willow or monocots (Reimers et al, 1980;Klein, 1992;Staaland & Olesen, 1992).Muskox can clearly make use of high protein foods (White et al, 1984), and even though they show many attributes of a classic grazer can be quite selective in their fee-   1985-1989, 1989-1991, and 1985-1991  Approximately half of the muskoxen were harvested from each strata.Because age of first reproduction has increased, harvesting has not created an increased numerical response of muskoxen in these areas.
Peary caribou (Rangifer tarandus pearyii) and muskox (Ovibos moschatus) are an important subsistence food source for the residents of Sachs Harbour, Banks Island.These ungulate populations have been syste- when freezing rains occurred.In response to decreasing caribou numbers, the Sachs Harbour Hunter's and Trapper's Committee established a quota of 150 animals in 1990 and 30 males in 1991.In 1992, the quota was increased to 36 males to allow each family to harvest one caribou.Over the same period muskox numbers increased.The residents of Sachs Harbour harvest muskoxen to support a local subsistence economy and conduct guided sport hunts for non-resident hunters.In addition, a commercial harvest program has been in place since 1981.During the quota years 1980-81 to 1989-90, the average commercial take was 124 muskoxen/year (range 0 to 260).The Rangifer, Special Issue No. 9, 213-222 first large-scale harvest was conducted during 1990¬ 91 when 494 muskoxen were taken.During the quota year 1991-92, 2 031 muskoxen were taken.The annual quota has been 5 000 muskoxen since 1991.

Fig. 3 .
Fig. 3.Estimated population size (+ SE) for (a) Peary caribou and (b) muskox from the originally flown surveys.The arrows in (a) point to winters which had freezing rains.The numbers above the arrows correspond to the recorded number of dead caribou.The 1982 value for muskox (b) is likely an underestimate.

Fig. 5 .
Fig. 5. Changes in distribution of Banks Island muskoxen from 1985 to 1991.The hatched areas on the 1982 map were not surveyed.See fig. 2 for strata boundaries.
ding(Oakes et al, 1992).Thing et al. (1987) showed willow as a major component of the summer diet of muskox in Greenland.Because of low lichen biomass on Banks Island, it is not surprising that willows, sedges, and forbs make up most of the caribou diet.The muskox diet is dominated by sedge, but in high density areas willow becomes an important dietary component in late winter(Larter, N. & Nagy, J., unpubl.data).The area between the Kellet and Big rivers and centered around the Egg River was identified as the core of the caribou winter range on Banks Island migratory, and seasonally utilize to a various extent forage consumed by caribou, there exists the potential for food competition in the core caribou winter range.If this is the case, some caribou may be displaced to less preferred winter ranges or may move to other areas on or off Banks Island in search of more adequate winter ranges.Competition for food cannot be ruled out as a factor contributing to the caribou population decline.The decline was likely caused by the cumulative effects of a combination of factors including human harvest, severe winters, wolf predation, inter-island movement and possibly competition from the expanding muskox population.Muskoxen The island muskox population increased during the period 1985 to 1992.Recent data indicate that reproduction has been delayed from 2-year old to 3-year old females (Nagy, J., et al, unpubl.data) possibly indicating some density dependent feedback mechanism.However, these data are limited to animals inhabiting the Egg and Masik strata.These two strata had continuous positive population growth, had the highest or among the highest densities of muskoxen since 1985 (Table 3), and are the two areas where commercial muskox harvests have been conducted.Age of first reproduction may not be increasing in other strata where muskoxen densities are lower.Population increase and consistent calf proportions continued in both the Egg and Masik strata even though 2 525 muskoxen were harvested from these areas between 1990 and prior to the 1992 survey.

Table 1 .
Population estimates (excluding calves, +SE), with corresponding density, median and mean group sizes, and the percent calves of Banks Island caribou from 1982-1992.

Table 2 .
Estimates and densities of muskox based upon original transect flight lines and strata, 1985 to 1992.

Table 4 .
The instantaneous rates of growth (r) for each stratum, and the total island for the periods Urquhart, 1973).From July 1985 to July 1991 muskox densities on the southern half of Banks Island increased from 0.19/km 2 to 0.65/km 2 .In the Egg River strata, which overlaps in excess of 60% of the core caribou winter range, muskox densities increased from 0.5/km 2 to 1.55/km 2 during that period (Table4).Because muskoxen are non-