Caribou recovery and coexistence with introduced feral reindeer on the Nuussuaq Peninsula ( 70-71 ° N ) , West Greenland

The small native caribou population (Rangifer tarandus groenlandicus) of Nuussuaq Peninsula was supplemented in 1968 with 10 semi-domestic reindeer (Rangifer tarandus tarandus). Hunt ing was prohibited in the early 1990s, but resumed wi th a quota of 100 animals i n 1996 after the population was estimated to be around 400. Despite local crit icism that herd size had increased, managers kept the estimate unchanged and permitted similar quotas for the next 5 years. To ascertain current status of the population, a late winter ground survey for m in imum count, recruitment and distribution was done in A p r i l 2002 employing local hunters. Data collected included group size, location and animal sex/age. Only two age classes were used; calf (<1 year) and "adult" (>1 year). The 2002 ground survey observed 1164 individuals and a calf percentage of approximately 30%. The bu l l to cow ratio was 0.32. This data d id not allow a cal¬ culation of population size, because areas where maximum animal numbers were expected were preferentially sampled. Spatial segregation of these two subspecies is suggested, given the observed and unexpected dissimilar behavior, phenotype and spatial distribution. If true, then by 2002 feral reindeer had established a successful population, while native caribou had recovered to number several hundred. Genetic sampling is necessary to examine this hypothesis. A t current late winter recruitment rates animal density could increase rapidly making both range expansion and genetic mix ing likely i n future. Since the total non-ice covered area available is about 6000 k m 2 , greater caribou/reindeer densities may not be compatible wi th sustainable range use. Harvest quotas were increased in 2002 and 2003, and may reduce densities and preserve caribou range for the future.


Introduction
The Nuussuaq Peninsula (70-71°N; 49-55°W) in West Greenland is isolated by topography and distance from other Rangifer tarandus populations in Greenland.Prior to 1968 only native wild caribou (Rangifer t. groenlandicus) were present and just 15-25 individuals were presumed remaining (Meldgaard, 1986).Therefore 10 semi-domestic reindeer (Rangifer t. tarandus) were introduced in 1968.These ani¬ mals originated from the Greenland Godthåbsfjord's Itivnera/Kapisillit reindeer herding district (64.2°N; 50.5°W), which was established in 1952 with a ship¬ ment of reindeer from northern Norway, (Cuyler, 1999).The 10 semi-domestic reindeer were released into the eastern portion of the Nuussuaq Peninsula in the vicinity of the lake, Boyes S0 (Fig. 1).
These semi-domestic reindeer were never intended for animal husbandry, but were to become feral, possibly mix with the native caribou, and ultimately provide the region's hunters with a meat resource for future harvests.Semi-domestic reindeer typically rut and calve about two or three weeks earlier than their caribou counterparts.Because the latitude of the Nuussuaq Peninsula matches the extreme northern limit for semi-domestic reindeer in Scandinavia, the success of this introduction was uncertain.Still, neither wolves (Canis lupus) nor any other potential predator have existed in West Greenland for several hundred years (Dawes et al., 1986).Harvesting by local hunters, however, was unregulated by quotas until 1987 (Fig. 2).
In 1991, following an aerial survey, which observed only 44 animals, the decision was made to pro¬ hibit hunting for 10 years (Greenland Institute of Natural Resources [GN] file 28.63.02/15Nuussuaq).Another survey in 1995 observed 161 animals, which included 11% calves (age <1 year), and population size was estimated to be 400 (Ydemann & Pedersen, 1999).Thus, despite the decision to prohibit hunting until 2001, a quota of 100 animals was permitted in 1996.Harvesting continued with little variation in

Study area
The Nuussuaq Peninsula (70-71°N; 49-55°W) roughly follows an eastwest orientation, with the Greenland Ice Cap forming its eastern border.It encompasses an area of about 7150 km 2 .If the numerous alpine ice caps and glaciers are deleted, however, the remaining area is about 6000 km 2 and most is above 200 meters elevation.

Minimum count ground survey
This ground survey sought a minimum count of R. tarandus and the percentage of calves from the previ¬ ous spring 2001.The minimum count is the number of animals observed and was not intended to be a population size estimate.This data did not allow a calculation of population size, because areas where maximum numbers were expected were preferen¬ tially sampled.Minimum counts, if repeated, could Natural Resources.Six snowmobiles were used, and the total distance of survey routes was about 1000 kilometers.
The observers drove together as a group and counted all animals seen using 10x30 Leica binoculars or 60x Leica telescopes, which made it possible to observe animals even from several kilometers distance.Since caribou/reindeer can be expected to move in the terrain, to prevent the possibility of "double-counting" animals, routes taken were sur¬ veyed once only and as quickly as weather and snow conditions permitted.Animals were sexed and aged.Sex was determined by presence or absence of a vulva and/or urine patch on the rump.This reliably indi¬ cated a female on both adults and calves.No other method was 100% certain, e.g., antler size, shape, presence/absence were not used.There were two age categories, calf (<1 year) and adult (>1 year).Age was determined by body size.Calves of both sexes were smaller than all others.
Snow depth and type were noted.Area of survey coverage was calculated using MapInfo Professional Version 7.0 (Copyright©1985-2002 MapInfo Cor¬ poration) regions.Significance testing and predict¬ ability were tested using 2 tailed t-tests (2 sample assuming unequal variances).

Variations in snow and ice cover
In the western third of the peninsula a ground ice layer occurred, and this was overlain by 70 cm of hard packed snow.The thick ground layer of ice was ice ended abruptly at the western most edge of the middle third.Snow in the middle third was often a deep (60-90 cm) loose powder, with no ground layer of ice.The situation was similar in the eastern third, except snow conditions also included extremely windblown areas of either stone-hard snow or rocky expanses blown bare of snow.South facing aspects had a thinner snow layer and relatively large patches of bare ground.

Status and distribution
In late April 2002 animals were observed over the entire peninsula, however, the western third was scarcely used (Figs. 3 and 4).The greatest concen¬ trations were in the middle third and specifically the eastern third of the peninsula.In both, valley bottoms were the primary habitat, followed by val¬ ley slopes or elevated plateaus.All animals in the middle third were observed at elevations above 200 meters and reaching to about 500-600 meters.Most in the eastern third occurred below 200 meters.Few animals were observed at elevations over 600 meters.Most of the animals and feeding craters were on south facing slopes and plains, where snow was thin and large patches of bare ground were available.Macro-lichens preferred by caribou/reindeer (Cladina spp., Cetraria spp.) were present on the south-facing slopes and being grazed.
The ground survey area coverage was ca.25% of the 6000 km 2 of non-ice covered terrain and observed a minimum count of 1164 individuals, which includ¬ ed 324 calves (age <1 year; born in 2001), for an overall calf percentage of 27.8% (Table 1).However, 87 of the animals could not be aged due to within group animal movement, which confounded identification, combined with too great a distance between observer and animals.All 87 were in an area where observations included primarily females with calves.Given the known numbers of adults ( 753), the true late winter calf percentage was 30.1% and calf recruitment was 43 calves per 100 adults, or similarly, 73 calves per 100 cows (443).Given the known number of bulls (143) and cows (443), there were just over 3 females for every bull for a bull to cow ratio of 0.32.Average group size was small in the middle third of the peninsula (Table 2), but large in the eastern third, which included tightly cohesive groups of 63, 206 and 212 animals.Similarly, density was greatest in the eastern third.Body condition was subjectively ascertained as prime.All animals appeared well rounded and not a single animal showing ribs was observed.

Phenotypic variation
Average group size and density were not the only notable differences between the animals inhabiting the middle versus the eastern third of the peninsula.Adult females in the eastern third possessed antlers, and even many female calves were antlered (Table 3).Whether single or paired, antlers were well formed, with no raggedy remnants of velvet.Antlers among females in the middle third were not common.Fur¬ ther, the animals in the middle third were always a pale gray-white color.In contrast, the animals in the eastern third were typically dark, with some being a deep brown.Further, calves in the middle third were noticeably smaller than those observed in the eastern third.The latter were almost adult in size and often possessed paired antlers with several tines.energy expenditure.
Although most of the Nuussuaq Peninsula is above 200 meters elevations, most animals were observed in valley bottoms.South facing slopes and plateaus also received high utilization, likely due to the thinner snow layer and abundance of bare patches.Obtaining food from these locations would require a minimum of energy expenditure.Whether the vegetation at these locations differed from others, e.g., north-facing slopes, is currently unknown, but differences are suspected.

Status
Body condition was subjectively ascertained as prime.Although late winter, animals appeared round and fat, when some individuals in other Greenland populations of Rangifer tarandus may clearly show backbone and ribs (C.Cuyler, pers. obs.).
During the ground survey in late April 2002, observers counted a minimum of 1164 animals.A population estimate was not possible.The late winter calf percentage was high at ca. 30%, and the recruitment was 73 calves per 100 cows.Since these calves were almost 1-year of age, at which time their mortal¬ ity rate becomes similar to mature animals, this was a high rate for an ungulate population.Studies from North America and Scandinavia report late winter recruitments of 20, 22 and 41 calves per 100 cows (Dzus, 1999;Parker, 1972;Fancy et al., 1994, respectively), but these populations have predators.The Southhampton Island Herd, like Greenland, has no predators, and late winter recruitment varies between 22 and 77 calves per 100 cows and is likely dependent on climate (Heard & Ouellet, 1994).
The sex ratio of bulls to cows, 0.32, was low, and likely the result of selective hunting of bulls (estimated to be 90% of the harvest) (Loison et al., 2000).Persistent maleskewed harvesting carries risks for the population dynamics and genetics of a population (Ryman et al., 1981;Ginsberg & Milner-Gulland, 1994).On the Nuussuaq Peninsula, however, there is oppor¬ tunity for crossbreeding between caribou and rein¬ deer, which would likely tend to increase genetic variation in the popula¬ tion.Still, the Nuussuaq population would benefit if the preference for males were reduced and harvest¬ ing females encouraged.A female dominated popula¬ tion with good body con¬ dition and a high recruit¬ ment rate can rapidly become too large for the range to support, poten¬ tially resulting in long¬ term degradation of the range and ultimately a population crash.(Cuyler & Linnell, 2001;Cuyler et al., 2002Cuyler et al., , 2003)) Almost three times the number of animals was observed in 2002 than had been assumed present on the peninsula, and the late winter calf percentage was excellent.The few remaining native caribou from 1968 appear to have recovered.Also, the 1968 introduction of 10 semi-domestic reindeer has successfully established a feral population, which is currently providing a meat resource for locals.Greenland managers increased the harvest quota in 2002 to 350 animals and to 400 animals in 2003.

Future surveys
Random transects used elsewhere in Greenland are not recommended for the Nuussuaq Peninsula.The feral reindeer exhibit tight group cohesion.This results in an extremely clumped distribution of large numbers of animals.The native caribou, although more evenly spread throughout the terrain in small groups, were still more common in valley bottoms.Both distributions would likely make population estimates highly inaccurate if based on aerial surveys using random transects.An index of abundance is possible if the ground survey for min¬ imum count is repeated, preferably annually, using the same routes and time period.Although vegetation maps, terrain models or NDVI maps are not yet available for this region, in future such maps could be used to stratify the census area.Then, an index of caribou density could be calculated for the various habitat types encountered on the survey route.Further, an objective documentation of snow and ice cover will better allow comparison of count data from different years.

Fig. 4 .
Fig. 4. Survey area coverage and the presence or absence areas surveyed on the Nuussuaq Peninsula in April

Table 1 .
Late winter R. tarandus herd structure obser-: four local commercial hunters and one caused by an unusual and heavy December 2001 hunting officer from the Uummannaq Municipal-rainstorm, which was followed by below freezing ity, and two staff from the Greenland Institute of temperatures for the rest of the winter.The ground Fig. 3. Locations of Rangifer tarandus observations on the Nuussuaq Peninsula, West Greenland, April 2002. participants

Table 2 .
Rangifer tarandus group size and density on the Nuussuaq Peninsula, West Greenland, based on ground survey minimum counts of late April 2002.

Table 3 .
Phenotypic and behavioral differences in Rangifer tarandus observed between the animals inhabiting the middle and eastern third's of the Nuussuaq Peninsula, West Greenland, April 2002.