Subsections:
Dynamics and Range Expansion of a Re-established Muskox Population
Seasonal Strategies of Muskoxen: Distribution, Habitats, and Activity Patterns

Muskoxen (Ovibos moschatus) disappeared from Alaska in the late 1800s, but returned to the Arctic National Wildlife Refuge when animals were reestablished into areas of former range in 1969-1970 (Klein 1988). Released at Barter Island (Kaktovik) and the Kavik River, muskoxen initially moved into regions that encompassed the 1002 Area on the coastal plain of the Arctic Refuge. From 1974 to 1986 the muskox population grew rapidly. By 1987, however, numbers declined in the regions that they had first occupied (Reynolds 1998a).

Petroleum exploration and development could occur in muskox habitat in the 1002 Area of the Arctic Refuge. Status of the muskox population and factors related to trends in local abundance need to be determined if changes resulting from natural processes are to be separated from those that might result if industrial development is permitted in the Arctic Refuge.

We developed a study with the following objectives to understand the dynamics of the muskox population in and near the 1002 Area of the Arctic Refuge: 1) determine abundance and rates of population increase, production, and survival; 2) document changes in population distribution over time; and 3) evaluate factors associated with changes in the number of muskoxen.

Numbers of muskoxen seen during annual censuses in 1982-2001, combined with data from earlier studies, were used to estimate animal abundance and population trends of muskoxen in the 1002 Area (regions first occupied) and adjacent areas to the east and west (regions occupied later) (Reynolds 1998a). Rates of successful calf production (defined as calves:100 females >2 years old present in late June), survival of calves and yearlings, and long term reproduction patterns by marked female muskoxen were determined from annual sex and age composition counts made from the ground in 1983-2001.

Radio-collared adult muskoxen were relocated 6 times/year from 1982 to 1994 to determine seasonal and annual variability in population distribution and to document adult mortalities. Locations were determined from the air with a global positioning system (GPS) or were plotted on 1:63,360-scale maps. The adaptive-kernel technique within the computer program CALHOME (Kie et al.1996) was used to delineate the size and locations of regions used by mixed-sex groups of muskoxen in 1969-1981, 1982-1985, 1986-1989, and 1990-1993. Core areas of use (70% adaptive-kernel contour) and total range (95% contour) were used to document changes in distribution and range expansion over time.

Locations of muskoxen seen during seasonal muskox surveys in the Arctic Refuge, as well as locations of mixed-sex groups seen during other studies, were used to document the continued expansion of the population distribution from 1994-2001. Reconstructed models of the population estimating maximum and predicted population growth were used to evaluate how changes in calf production, survival, and emigration affected local abundance of muskoxen.

The Arctic Refuge’s reestablished population of muskoxen grew slowly for a few years and then increased rapidly for more than a decade (Reynolds 1998a). Between 1977 and 1981, the population grew at r = 0.24, a rate approaching maximum growth. In 1986, 368 muskoxen were counted in the 1002 Area, but after 1986, numbers of muskoxen declined (Fig.7.1). The rate of increase slowed to 0.14 in 1982-1986 and to <0.00 in 1987-1995.

Figure 7.1. Number of muskoxen observed in the regions first occupied after reintroduction — the 1002 Area, Arctic National Wildlife Refuge, Alaska, USA, during spring censuses, 1982-2001

The rate of population growth (r = 0.14) in 1972-1996 (Reynolds 1998a) was similar to rates recorded for other expanding populations of muskoxen in Alaska and Canada (Spencer and Lensink 1970, Gunn et al. 1991). An introduced population of muskoxen in Greenland was still irrupting 25 years after release, but those animals were in an area of abundant high-quality forage with no predation and low snowfall (Olesen 1993).

In 1996-2001, numbers of muskoxen counted in the 1002 Area ranged from 168 to 212 (P. E. Reynolds, U.S. Fish and Wildlife Service, unpublished data) and indicate that muskox abundance is still declining slowly (Fig. 7.1).

Factors affecting changes in the number of muskoxen in the 1002 Area of the Arctic Refuge included changes in rates of successful reproduction and survival as well as changes in animal distribution.

Calf production was the only source of increase in this reestablished population; a lack of adjacent populations of muskoxen precluded immigration. Growth of the population during its most rapid increase was due to high rates of reproduction and survival of muskoxen in newly occupied habitats. In the 1002 Area, indices of successful calf production reached a maximum between 1977 and 1980 (87 calves:100 females >2 years old) during years when successful reproduction by some 2-year-old females and 100% reproduction among females >2 years old occurred in some groups (Jingfors and Klein 1982).

Calf production declined between 1983 and 2001 (Fig. 7.2). In 1983-1986, as the rate of population increase began to slow, calf production in the 1002 Area averaged 61 calves:100 females >2 years old compared with 49 in 1987-1990, 41 in 1991-1994, and 28 in 1995-1999 (Reynolds 1999). In June 2000 and 2001, very few calves (<5 calves per 100 females >2 years old) were seen. Because calves were counted several weeks after birth, we could not determine if changes in the production of muskox calves were due to lower fecundity or increased neonatal mortalities.

Figure 7.2. Changes in rates of successful production of muskox calves in the regions first occupied after reintroduction – the 1002 Area, Arctic National Wildlife Refuge, Alaska, 1983-2001. Successful calf production was measured by counting the number of calves per 100 females >2 years of age in late June.

Reproductive patterns of radio-collared females in the 1002 Area showed similar trends. Mean reproductive intervals (number of years between successful reproductive events in a 3-year period) increased significantly (r2 = 0.95, n = 6, P = 0.0009) between 1982 and 1999. By 1991-1993, most marked females successfully reproduced at intervals of 2 to 3 years, rather than every year (Reynolds 2001). Percentages of marked females without calves for 3 or more consecutive years were 0% in 1982-1987, 15% in 1988-1990, and 25% in 1994-1996 (Reynolds 2001). In summer, body weights of 8 lactating muskoxen (mean = 223 kg, range = 188-254 kg) were not different (t = 2.2, df = 10, P = 0.167) from 8 non-lactating females (mean = 196 kg, range = 136-254 kg) and were similar to weights of female muskoxen in other wild and captive populations (Reynolds and Reynolds 1999).

Unlike calf production, which declined in 1983-1995 (r2= 0.75, P < 0.001), calf and yearling survival did not decline over time (calf survival: r2 = 0.01, P = 0.71; yearling survival: r2 = 0.01, P = 0.82). Annual variability in young animal survival followed the same annual trends as calf production and was related to snow depth and the length of the snow season (Reynolds 1998a).

Between 1983 and 1999 the percentages of radio-collared muskoxen dying each year were variable but showed an increased trend (Reynolds 1999). Sources of mortality included kills by predators, including humans.

Legal hunting of muskoxen in the Arctic Refuge began in 1982. Over time, the number of permits issued each year increased from 5 males only to 12 males and 3 females. The season was expanded from 2 to 8.5 months. An average of 7 muskoxen was killed in 1985-1989 compared with an average of 10 in 1995-1999. About 3% of the muskox population in the Arctic Refuge was harvested annually from 1990-1999 (P. E. Reynolds, U.S. Fish and Wildlife Service, unpublished data).

Kills or scavenging of muskoxen by grizzly bears (Ursus arctos) in and near the Arctic Refuge increased significantly between 1986 and 2001 (b = 0.504, df = 18, P < 0.001) (Fig. 7. 3). Known kills of muskoxen by grizzly bears ranged from 0-2 deaths per year before 1993, 1-4 deaths per year in 1993-1996, and 5-10 deaths per year in 1997-2001 (Reynolds et al. 2002).

Figure 7.3. Number of muskoxen killed or scavenged by grizzly bears from April 1982 through June 2001 in northeastern Alaska, USA. (from Reynolds et al. 2002)

Forty-seven deaths of adult or sub-adult muskoxen from known grizzly bear predation occurred between 1982-2001. Of these, 28 muskoxen died during 10 incidents of multiple kills in which bears killed more than one muskox from a group. Most of these kills (79%) took place between May 1998 and June 2001 (Reynolds et al. 2002).

Grizzly bears likely also killed muskox calves and caused other mortalities of young calves that were deserted during predation events. Multiple kills of calves were observed in Canada (Clarkson and Liepins 1993). The increase in kills by grizzly bears suggests that predation may have been one factor that resulted in very low numbers of calves in late June of 2000 and 2001. Deep snow and a prolonged winter season in 2000 and 2001 also likely contributed to the low numbers of calves seen in those years and may have exacerbated the number of predation events (Reynolds et al. 2002).

Shifts in distribution and emigration also affected numbers of muskoxen in the 1002 Area of the Arctic Refuge. Following their release in 1969 and 1970, most muskoxen became associated with 1 of 3 mixed-sex groups in 3 regions of the Arctic Refuge. The regions first occupied were located between the Canning and Aichilik Rivers within the boundaries of the 1002 Area (Fig. 7.4a) (Reynolds 1998a). After 1986, muskoxen in mixed-sex groups colonized new regions east and west of the 1002 Area (Fig. 7.4c,d) (Reynolds 1998a).

Figure 7.4. Range expansion of muskoxen in mixed-sex groups in and near the Arctic National Wildlife Refuge, Alaska, 1969-1993. Total ranges were defined by 95% adaptive kernel contours. Core areas were defined by 70% adaptive kernel contours. (from Reynolds 1998a)

In 1995, about 800 muskoxen were counted in the entire range of the population (Table 7.1), which had expanded westward to the Itkillik River, Alaska, and eastward to the Babbage River in northern Yukon Territory, Canada (Reynolds 1998a). In 1998-2001, mixed-sex groups of muskoxen continued to expand their range west to the Colville River, southwest along the Sagavanirktok River, and south and east of the Babbage River in northwestern Canada. During these years, <700 muskoxen were counted throughout the total range of the population (Table 7.1) (P. E. Reynolds, U.S. Fish and Wildlife Service, unpublished data, E. A. Lenart, Alaska Department of Fish and Game, unpublished data, and D. A. Cooley, Yukon Renewable Resources, unpublished data).

Table 7.1. Number of muskoxen seen in different regions in northeastern Alaska, USA, and northwestern Canada in 1982-2000 during pre-calving surveys. GMU 26B and 26C are State of Alaska game management units. The muskox population originated from animals released adjacent to the Arctic National Wildlife Refuge, Alaska, in 1969 and 1970. The muskoxen began to disperse into new regions east and west of the Arctic Refuge by 1986 (Reynolds 1998a).

Differences between the observed and predicted abundance in the 1002 Area, based on reconstructed population projections, suggest that changes in muskox calf production and animal survival caused most of the decline in the rate of population growth (Reynolds 1999). Density dependent factors as well as annual variability in snowfall and increasing rates of predation all likely influenced observed changes in calf production and animal survival.

Emigration of mixed-sex groups of muskoxen also reduced the number of muskoxen in the 1002 Area (Reynolds 1998a, 1999). In 2000 and 2001, the additional emigration of mixed-sex groups containing marked animals and the low rates of successful calf production (<5 calves per 100 females > 2 years old) contributed to the declining trend in numbers of muskoxen in the 1002 Area (P. E. Reynolds, U.S. Fish and Wildlife Service, unpublished data).

Although muskoxen are continuing to expand into their former range in northern Alaska and northwestern Canada, numbers of muskoxen in the 1002 Area are not likely to increase from their present level of <250 animals in the near future.

If exploration and extraction of petroleum resources are permitted in the Arctic Refuge coastal plain, associated industrial activities could further reduce the number of muskoxen in the 1002 Area either through induced dispersal or decreased productivity and survival. Muskoxen are year-round residents of the 1002 area, which heightens their vulnerability. In addition, their small numbers make it less likely that the muskoxen can recover from perturbations.

Status and distribution of muskoxen in and near the 1002 Area should continue to be monitored to document future trends.

Seasonal shifts in distribution, habitat use, and activity are means by which animals maximize energy intake and avoid conditions that risk their survival. The muskox is an energetically conservative species (Klein 1992) and its seasonal habitat use and energy budgets influence its reproduction and survival (White et al. 1989). Limited forage availability and energy constraints in winter as well as potential cumulative effects of disturbance contribute to its susceptibility.

As year-round residents of the coastal plain of the Arctic Refuge, muskoxen are vulnerable to human activities in both winter and summer. Information is needed about their seasonal patterns of distribution and activity to evaluate and minimize potential effects associated with oil and gas exploration and development proposed for the Arctic Refuge’s 1002 Area.

Our study to determine seasonal patterns of muskoxen on the coastal plain of the Arctic Refuge had the following objectives: 1) compare distribution and habitat use of muskoxen in different seasons, and 2) determine seasonal movements and activity patterns of muskoxen.

In the Arctic Refuge, snow is present from 8-9 months each year (September - May). Five seasons were defined for muskoxen based on ecological and biological conditions: calving (late March to mid-June), summer (late June to mid-September), early winter (late September to mid-November), mid-winter (late November to mid-January), and late winter (late January to mid-March).

To identify population distribution in different seasons, 19-25 radio-collared muskoxen were monitored and 4 to 6 radio-relocation surveys were flown each year from 1982 to 1995. Locations of groups of muskoxen, both marked and unmarked, were determined using a global positioning system or were plotted on 1:63,360 scale maps.

Seasonal distribution and movement rates were determined from 15 female muskoxen fitted with satellite collars (ultra-high frequency platform transmitter terminals that were relocated by satellite) (Reynolds 1989). Three to 5 animals carrying satellite collars were monitored yearly from October 1986 through March 1992. These collars transmitted information about animal location and activity every second or third day for 6 hr/day (Reynolds 1998b).

Seasonal distribution of the population and seasonal home ranges of satellite-collared muskoxen were delineated with an adaptive-kernel technique and the program CALHOME (Kie et al. 1996). Seasonal differences in population distribution were compared as the overlap of core areas (70% contour), distances between core-area centers, and core-area sizes.

Mean movement rates (km/day) for each season and each month were calculated from distances moved by satellite-collared muskoxen. Distances were calculated between consecutive locations at 40-55 hr intervals. Mean activity indices for each season and each month were derived. Activity counts from 5 satellite-collared muskoxen with >10 days of activity counts per month were used to estimate mean activity (Reynolds 1998b).

Land-cover and terrain types, extracted from a land-cover map derived from Landsat-Thematic Mapper data (Jorgenson et al. 1994), were used to determine seasonal differences in habitat use at a landscape scale. Selection ratios of 6 land-cover classes and 5 terrain types were based on proportions present in core areas (habitats used) divided by proportions in the entire study area (habitats available) (Reynolds 1998b).

The average size of core areas used by muskoxen carrying satellite collars was significantly larger (P < 0.05) in summer (223 km2) than in calving season or the 3 winter seasons (27-70 km2) (Reynolds 1998b). The size of core areas was highly variable in summer, but means differed by almost an order of magnitude between summer and other seasons. The minimum size of core areas used in summer was >4 times larger than minimum core areas occupied in winter or calving.

Muskoxen were conservative in their daily movements throughout the year. Most (95%, n = 2314) movements made by satellite-collared muskoxen were <5 km/day (Reynolds 1998b). Of these, 46% were <1 km/day. Moderate movements of 5-10 km/day took place primarily between June and September (77 of 108). Only 1 (<1%) moderate movement of 5-10 km/day was recorded between January and April. Movement rates >10 km/day, resulting in relatively long moves, were rare (18 of 2314); 16 (89%) movements >10 km/day occurred in July.

Mean daily movements in summer (2.6 km/day) were greater (P < 0.05) than in other seasons (1.1–1.4 km/day) (Reynolds 1998b). Mean rates of movement were significantly higher (P < 0.05) in July than in other months (Fig. 7.5). Activity counts/minute from satellite-collared muskoxen were also greater in summer (P < 0.001) than in other seasons. Activity counts differed among months (P = 0.001) and were highest in July and lowest in April during the onset of the calving season.

Figure 7.5. Seasonal changes in rates of movement and activity counts of satellite-collared female muskoxen in and near the Arctic National Wildlife Refuge, Alaska, 1986-1992. (from Reynolds 1998b)

Seasonal home ranges occupied by females with satellite-collars overlapped less (P = 0.01) between calving and summer than between early winter and mid-winter and between mid-winter and late winter (Reynolds 1998b). This reflected the sedentary nature and small home range size of the muskoxen in winter. Distances between seasonal home ranges were also small. The distribution of the population of muskoxen occupying the coastal plain of the Arctic Refuge showed little change between seasons.

At a landscape scale, muskoxen used riparian cover along river corridors, floodplains, and foothills in all seasons. Moist sedge was selected in late winter and calving; tussock tundra was avoided in late winter. Wet sedge was used in proportion to availability in summer and early winter but avoided in other seasons. Upland shrub was selected only during the calving season and avoided in other seasons. Bare cover (including bare ground, water, and ice) was selected in all seasons except spring. Mountain terrain was avoided in all seasons (Reynolds 1998b).

Ground-based studies (Wilson 1992) provided more information at regional and local scales (see next subsection on winter habitat use). Locations of mixed-sex groups of muskoxen during summer and winter surveys demonstrated the importance of river corridors and adjacent uplands to this population (Fig. 7.6).

Figure 7.6. Locations of mixed-sex groups of muskoxen seen during winter and summer surveys in the Arctic National Wildlife Refuge, Alaska, USA, 1982-1999.

The small seasonal shifts in distribution and low movement rates observed in this study confirmed that muskoxen are energetically conservative throughout the year (Jingfors 1980, Thing et al. 1987) and that they have a high fidelity to geographic regions. Seasonal changes in movements, activity, and habitat use were related to availability of forage and the energy budgets of muskoxen.

In winter, snow limits forage availability and habitat selection (Jingfors 1980). In late winter, muskoxen selected feeding sites with soft shallow snow (Biddlecomb 1992, Wilson 1992). These sites were frequently narrow windblown bluffs adjacent to rivers where snow accumulation was low (Nellemann and Reynolds 1997). By mid- to late winter, riparian willows and wet-sedge communities may be unavailable to muskoxen as snow depths increase (Wilson 1992, Evans et al. 1989).

Winter forage of muskoxen is of low quality (Staaland and Olesen 1992). Graminoids were a dominant component of the late winter diet of muskoxen in northeastern Alaska (O’Brien 1988, Biddlecomb 1992, Wilson 1992). Muskoxen, however, can digest low quality graminoids efficiently and may have a fasting metabolic rate lower than other ruminants (Adamczewski et al. 1994, Lawler 2001). In winter, muskoxen conserve energy by reducing movements and activity, decreasing the size of use areas, and concentrating in a few habitats where forage is not covered with deep snow.

Unlike caribou that calve in early June when nutritious vegetation is emerging, most muskoxen give birth several weeks before high quality green forage is available. In Alaska, Dall’s sheep (Ovis dalli) (Rachlow and Bowyer 1994) and moose (Alces alces) (Bowyer et al. 1998) have similar calving strategies to muskoxen. To reproduce successfully, female muskoxen must be in good body condition at calving time to fuel the high cost of lactation. Their energy-conserving strategy of restricting movements and activity and selecting habitats with low snowcover allows female muskoxen to maintain body condition throughout the winter and spring (Thing et al. 1987).

Most muskoxen in the Arctic Refuge give birth in April and May and lactate under conditions of poor quality forage and harsh weather. Their increased use of foothill terrain and upland shrub during the calving season reflected shifts into areas where snowcover was shallow or blown free and their energetic costs of foraging were lower. Muskoxen with young calves also may avoid flooded riparian areas during calving and post-calving periods. Movement rates of muskoxen carrying satellite collars reached a yearly low in April at the onset of the calving season.

During the snow-free summer when food quality and quantity are high, muskoxen increase their movement and activity, occupy larger areas, and use diverse habitats as they forage on a variety of high-quality vegetation (Robus 1981, O’Brien 1988). They track the changing plant phenology in local areas to obtain high quality forage and rapidly regain body weight lost during winter, pregnancy, and early lactation. Muskoxen that fail to regain body weight are less likely to breed or successfully reproduce (White et al. 1997) and are less likely to survive a severe winter.

In our study we found that movement rates and activity increased in June as plants began to leaf out and were highest in July as live plant biomass peaked (Chapin 1983). Movement rates and activity of muskoxen carrying satellite-collars began to decline in August as plant senescence and rut occurred (Reynolds 1998b).

Seasonal strategies that emphasize energy intake in summer and energy conservation in winter, combined with physical adaptations for cold weather and the ability to process low quality forage, permit muskoxen to survive year-round in locations seasonally avoided by most other animals. Muskoxen are present during all seasons in the potential oil exploration and development area of the Arctic Refuge.

This study did not quantify the effects of petroleum development on muskoxen. But human activities that increase energetic costs to muskoxen in winter or decrease foraging opportunities in summer have the greatest probability to affect the muskox population. Riparian habitats frequently used by muskoxen are also likely to be used as sites for gravel and water extraction and winter road construction if exploration and development of petroleum resources occur on the coastal plain of the Arctic Refuge.

Exploratory and construction activities in northern Alaska often take place in winter. Muskoxen are particularly vulnerable to disturbance in winter because of limited habitat, the length of the arctic winter, and their need to conserve energy throughout the winter including the calving season. The average size of muskox groups is larger in winter than in mid-summer (Reynolds 1993). Large groups of animals often are more easily disturbed than small groups because large groups contain more individuals responsive to perturbations.

Effects of human activities on muskoxen are likely related to the scale of the activity and the availability of alternate habitats that can be used if animals are displaced. Muskoxen that expanded westward from the Arctic Refuge use the wide Sagavanirktok River valley in summer despite the presence of the Dalton Highway and the trans-Alaska oil pipeline. Habitats available to muskoxen in the Arctic Refuge, however, are geographically constricted: The coastal plain is narrower because the mountains of the Brooks Range are closer to the Beaufort Sea (Fig. 1.1).

If undisturbed, muskoxen generally stay in relatively small areas throughout the winter. Avoidance by industry of these areas used by muskoxen could reduce the probability of disturbance and displacement of muskoxen. Minimizing human activities in areas occupied by muskoxen from mid-winter through the calving season could reduce the likelihood of disturbance during the period when energy conservation is critical to survival. Locating permanent facilities away from river corridors, flood plains, and adjacent uplands could also help to reduce potential effects of industrial development on muskoxen.

(continued to Part 2)

| Home | Section 1 - Introduction | Section 2 - Land Cover | Section 3 - Porcupine Caribou Herd |
| Section 4 - Central Arctic Caribou Herd | Section 5 - Forage Quantity and Quality | Section 6 - Predators |
| Section 7 - Muskoxen | Section 8 - Polar Bears | Section 9 - Snow Geese | Acknowledgements |