4104.p65 ALCES VOL. 40, 2004 COBB ET AL. - SYMPATRIC MOOSE AND DEER 169 RELATIVE SPATIAL DISTRIBUTIONS AND HABITAT USE PATTERNS OF SYMPATRIC MOOSE AND WHITE-TAILED DEER IN VOYAGEURS NATIONAL PARK, MINNESOTA McCrea A. Cobb1,2, Peter J.P. Gogan3, Karin D. Kozie4, Edward M. Olexa3, Rick L. Lawrence5, and William T. Route6 1Department of Ecology, Montana State University, Bozeman, MT 59717, USA; 3USGS – Northern Rocky Mountain Science Center, Forestry Sciences Laboratory, Montana State University, Bozeman, MT 59717, USA; 4Voyageurs National Park, 3131 Highway 53 South, International Falls, MN 56649, USA; 5 Department of Land Resources and Environmental Sciences, Montana State University, Bozeman, MT 59717, USA; 6National Park Service, Great Lakes Network Office, 2800 Lake Shore Drive, Ashland, WI 54806, USA ABSTRACT: We examined the distribution and home range characteristics of moose (Alces alces) and white-tailed deer (Odocoileus virginianus) at Voyageurs National Park, Minnesota. Pellet count transects revealed low densities of moose and higher densities of white-tailed deer, and provided evidence of partial spatial segregation between moose and white-tailed deer possibly due to habitat heterogeneity. There was limited interspecific overlap in the relatively large annual home ranges of radio-collared moose and white-tailed deer. Both moose and white-tailed deer exhibited significant selection for spruce (Picea spp.) and balsam fir (Abies balsamea) vegetation types at the home range scale. White-tailed deer significantly selected a 12-20 m canopy height over all others while moose significantly selected 5-11 m and 21-30 m canopy heights over the 12-20 m canopy height. Moose significantly selected open/discontinuous canopy cover and white-tailed deer selected both closed/continuous and open/discontinuous canopy covers over dispersed/ sparse canopy cover. Differential habitat selection between moose and white-tailed deer at Voyageurs National Park might be related to the differences between these species' abilities to cope with a northern mid-continental climate. Spatial segregation between moose and white-tailed deer at Voyageurs National Park may allow moose to persist despite the presence of meningeal worm (Parelaphostrongylus tenuis) in white-tailed deer. ALCES VOL. 40: 169-191 (2004) Key words: Alces alces, compositional analysis, ecology, home range, meningeal worm, moose, Odocoileus virginianus, Parelaphostrongylus tenuis, pellet groups, sympatric, white- tailed deer 2Present address: Department of Environmental Science, Policy, and Management, University of California, Berkeley, CA 94720-3114, USA Moose inhabit a circumpolar region of northern boreal forests dominated by spruce (Picea spp.), pine (Pinus spp.), and fir (Abies spp.). The range of moose in North America has expanded since 1955 (Peterson 1955) while numbers throughout the range increased from approximately 940,000 to 975,000 between 1960 and 1990 (Karns 1998). Moose numbers in Minnesota in- creased eight fold from approximately 1,500 animals in 1960 to 12,000 in 1990 (Karns 1998). Numbers of moose in northern Min- nesota may have peaked prior to the 1990 estimates as moose abundance in adjacent Ontario started to decline in the mid-1980s (Thompson and Euler 1987). Moose at SYMPATRIC MOOSE AND DEER – COBB ET AL. ALCES VOL. 40, 2004 170 Voyageurs National Park (VNP), Minne- sota, are at the southern periphery of the species’ North American range and within a low density range between two high moose density ranges to the northeast and north- west (Fuller 1986). In the last 10 years, the northwestern moose population has dropped dramatically and now only numbers a few hundred animals (M. Lenarz, Minnesota Department of Natural Resources, personal communication). White-tailed deer (WTD) expanded their distribution into northern Minnesota around 1900 and were common in the area by the 1920s (Petraborg and Burcalow 1965). WTD densities in north- eastern Minnesota were estimated at be- tween 6 and 8/km2 in the late 1930s (Olson 1938, Petraborg and Burcalow 1965). Moose and WTD are sympatric across a relatively narrow band of North America, and the species’ habitat use patterns within this band are not well understood. The two species are thought to have occurred sympatrically in the area that is now VNP at least since the early 1930s (Cole 1987, Gogan et al. 1997). Fluctuations in WTD numbers might be due to changes in habitat and changes in moose population levels have been attributed to changes in vegeta- tive types and seral stages (Mech and Karns 1978, Cole 1987). Following the 1971 Little Sioux Fire in adjacent Superior National Forest, Minnesota, moose densities in- creased to five times their previous number (Neu et al. 1974, Peek et al. 1976). Both moose and WTD were found to consume similar browse after a fire in northern Min- nesota (Irwin 1975). Active fire suppres- sion within VNP has limited recent natural disturbances, and together with logging re- strictions, affected the vegetation composi- tion. The low frequency of wildland fires since the establishment of VNP could be a factor contributing to relatively low densi- ties of moose. In the absence of specific information on moose and WTD distribu- tions and habitat use patterns at VNP, the relationship of each species to vegetative conditions remains unclear. Parasite mediated competition between moose and WTD might be responsible for recent declines in moose numbers. Menin- geal worm, a parasite that is characteristi- cally benign in WTD but fatal in moose, has been attributed for moose declines in Min- nesota and elsewhere (Karns 1967, Prescott 1974). The extent to which meningeal worm impacts moose abundance at VNP is an unresolved issue. Spatial separation and differential habitat selection between moose and WTD may allow moose to persist in the presence of infected WTD (Gilbert 1974). WTD in Nova Scotia were excluded from some habitats at high elevation by snow depth, providing moose with refuges from WTD during the winter season (Telfer 1967). A “refugium” between moose and WTD in Ontario was identified as a possible factor allowing moose to persist in the presence of sympatric populations of meningeal worm- infected WTD (Kearney and Gilbert 1976). Questions however have been raised con- cerning the validity of the refugia hypoth- esis. The purported benefits of seasonal refugia for moose in warmer months when the potential infection rate is highest might not exist because moose and WTD habitat use overlapped during other times of the year (Nudds 1990). Even partial refugia from infected WTD however may provide moose with a relative advantage (Whitlaw and Lankester 1994). In the absence of specific information on moose and WTD distributions and habi- tat use patterns at VNP, the relationship of the abundance of each species to vegeta- tive conditions and the potential for moose refugia from meningeal worm infection re- mains unclear. This study was initiated to determine the relative spatial distribution and home range characteristics of moose and WTD at VNP, and to examine the ALCES VOL. 40, 2004 COBB ET AL. - SYMPATRIC MOOSE AND DEER 171 influence of habitat types on these distribu- tions, and to assess overlapping use pat- terns of the two species. STUDY AREA VNP encompasses 882 km2 on the southern portion of the Canadian Shield along the U.S.-Canada border. VNP is made up of a central landmass largely sur- rounded by lakes, called the Kabetogama Peninsula, and adjacent lands. Approxi- mately 40% of VNP is covered by 4 large lakes. There is little overall elevation change with a maximum topographic relief of 80-90 m (Johnson and Sales 1995). Adja- cent areas in Minnesota include lands ad- ministered by the state (Kabetogama State Forest), the federal government (Superior National Forest), and privately owned lands. Adjacent areas of Ontario are mainly Pro- vincial Crown Lands. The study area bound- ary was defined as the area encompassed within the GIS vegetation coverage (USGS 2001) of the VNP region (Fig. 1). # International Falls # Rainy Lake # Namakan Lake # Crane Lake # Kabetogama Lake # Sandy Point LakeUnited States Canada # Kabetogama Peninsula N EW S 5 0 5 10 Kilometers # # # Minnesota Intern ational Falls Duluth Twin Cities Water VNP Boundary U.S.- Canada Border Habitat Cover Boundary Fig. 1. Location of Voyageurs National Park, northern Minnesota. Climate The climate is characterized as cold winters and cool summers. Temperature extremes during the study period ranged from 35oC (August 1, 1989) to -39oC (De- cember 30, 1990, National Weather Serv- ice, International Falls, MN). Average an- nual snowfall is 160 cm, with the most snowfall occurring during January (31 cm). The winter of 1988 – 1989 was a high snow year with 266 cm. Snowfall in the winter of 1989 – 1990 was 155 cm, close to the long- term mean, while snowfall in the winter 1991 – 1992 was 247 cm, considerably higher than the mean. The first significant winter snowfall usually occurs in early No- vember, and the last significant snow usu- ally occurs in early April (National Weather Service, International Falls, MN). The North Atlantic Oscillation (NAO) index (Lamb and Peppler 1987, Hurrell 1995) showed that winter temperatures were colder than average during the study period, with the winters of 1988 – 1989 and 1989 – 1990 SYMPATRIC MOOSE AND DEER – COBB ET AL. ALCES VOL. 40, 2004 172 being particularly cold. Vegetation VNP lies on the boundary between southern boreal forest and northern hard- wood forest types (Pastor and Mladenoff 1992). Northern hardwood forests are domi- nated by red pine (Pinus resinosa), white pine (P. strobus), red maple (Acer rubrum), and black ash (Fraxinus nigra) (Kurmis et al. 1986). Southern boreal forest types are characterized by a mosaic of secondary growth jack pine (P. banksiana), white spruce (Picea glauca), quaking aspen (Populus tremuloides), paper birch (Betula p a p y r i f e r a ) , a n d b a l s a m f i r ( A b i e s balsamea) (Kurmis et al. 1986). The soil in the region is thin and sandy (Ohmann and Ream 1971). Varying sources and levels of distur- bance have created spatial heterogeneity in the vegetation of the VNP region. Logging has been an important influence on the current spatial distribution of vegetation across much of VNP. Parts of VNP, including approximately 25% of the Kabetogama Peninsula, were extensively logged between 1910 and 1930 (Crowley and Cole 1995). The combined impacts of these harvests decreased the abundances of white spruce, balsam fir, white pine, and red pine on the Kabetogama Peninsula. The relative abundance of aspen conse- quently increased to higher levels post-har- vest. While logging practices within VNP ceased with the inception of the park in 1975, the majority of forested lands adja- cent to the park have continued to be man- aged for timber harvest. Fire suppression efforts began in 1911 and have since limited major fires in the park region to 1917 – 1918, 1923, and 1936. Fires burned substan- tial portions of the Kabetogama Peninsula in 1923 and 1936, adding to the mosaic of vegetative cover in VNP (Fig. 2). Human and naturally caused wildfires within the park since 1936 have been relatively small (<2.0 km2). Wildlife Woodland caribou (Rangifer tarandus) and moose are thought to have been the most common ungulates in the VNP region in pre-historic times (Cole 1987). WTD expanded northward into the region in the late 1890s and were reported to be common in the region by the 1920s (Petraborg and Burcalow 1965). Woodland caribou were extirpated from the region by the 1940s (Gogan et al. 1997). Moose possibly de- clined in numbers from the establishment of the park in the mid – 1970s through the mid – 1980s (Cole 1987). Approximately 60 – 100 moose were estimated to inhabit VNP at a mean density of 0.23/km2 in the early 1990s (Whitlaw and Lankester 1994, Gogan et al. 1997). Estimated densities of WTD in and immediately adjacent to VNP ranged from 1.5/km2 to 11.5/km2 from 1975 (R. O. Peterson, Michigan Technological Univer- sity, unpublished report, 1976) through 1992 (Whitlaw and Lankester 1994, Gogan et al. 1997). The reasons for the recent varia- tions in moose and WTD population levels are unknown. METHODS GIS Vegetation Coverage Habitat availability of the study area was determined using a Geographic Infor- mation System (GIS) vegetation coverage created by interpreting 1:15,840-scale color infrared (CIR) aerial photographs taken in 1995 and 1996 (USGS 2001). The entire coverage consisted of 156,886 ha, of which VNP comprised 88,244 ha (56%) of the total. A total of 40 vegetation cover types defined the ground features within the project area. Each vegetation cover type was further classified by canopy height and canopy density. For this study, we consoli- dated the original 40 GIS vegetation cover ALCES VOL. 40, 2004 COBB ET AL. - SYMPATRIC MOOSE AND DEER 173 types into 8 classes based on functional groups to facilitate analysis and alleviate the problem of missing habitat types during compositional analysis (Table 1). Pellet Group Transects We applied a 1 – km2 grid to a 1:50,000- scale map of the park and randomly se- lected 32 cells as pellet group transect sam- pling units. Two parallel transect lines were established in a north-south orientation with at least 100 m separation between transect lines within most sampling units. Sampling units containing >50% water cover (11 of 32, or 34%) were limited to one transect line resulting in a total of 53 transect lines within the 32 sampling units. Each transect line was 800 m long and consisted of 4 plots (22 m by 3.6 m) at 200 m intervals. A survey chain (20 m) was used to measure distance traveled while surveying. Transect lines were sampled once in late May of 1989 and again in the late May 5 0 5 10 Kilometers N EW S Burned (1923 and 1936) Logging (1950s and 1960s) Fig. 2. Historical fire and logging locations within Voyageurs National Park, Minnesota. of 1991, after snow melt and prior to the onset of new vegetative growth. All pellet groups within each plot above the previous fall’s leaf litter were identified to species and tallied. We used the total number of moose and WTD pellet groups observed along each transect line in our analysis. We sampled 16 paired and 14 single transect lines in 1989, and 19 paired and 12 single lines in 1991. Some sampling units (4 of 32, or 12.5%) were only visited one year due to a lack of personnel and access problems (private land, terrain). To insure that the detection probability was even between sampling units containing 1 and 2 transect lines, we randomly selected a single transect line from sampling units containing 2 transect lines for analysis. Numbers of moose and WTD pellet groups detected within each sampling unit by sampling year were entered into a GIS for interpretation. For moose, sampling units were stratified into those with pellet SYMPATRIC MOOSE AND DEER – COBB ET AL. ALCES VOL. 40, 2004 174 Table 1. Description of vegetation classes used in compositional analysis. Consolidated vegetation classes were created by combining original GIS habitat classifications from the USGS/NPS vegetation coverage of the Voyageurs National Park region. Cons olidat ed Clas ses O riginal G IS H abit at Class ificat ion Shrubland A lliance Beaked H az el/Serviceberry shrubland alliance Bog Birch/Willow sat urat ed shrubland alliance Leat herleaf s at urat ed dw arf shrubland alliance Red O s ier D ogw ood/Willow seas onally flooded s hrubland alliance Sp eckled A lder s easonally flooded shrubland alliance N ort hern Whit e Cedar/Red M ap le s at urat ed forest alliance N ort hern Whit e Cedar/Yellow Birch fores t alliance N ort hern Whit e Cedar fores t alliance N ort hern Whit e Cedar sat urat ed fores t alliance T amarack s at urat ed forest alliance Black A sh/Red M ap le Black A sh/Red M ap le sat urat ed fores t alliance Jack P ine Jack P ine/Lichen nonvascular alliance Jack P ine forest alliance Jack P ine, Red P ine w oodland alliance M osaic (Jack P ine forest alliance and Q uaking A s p en/P ap er Birch fores t alliance) Red/W hit e P ine Red P ine forest alliance Whit e/Red P ine and Q uaking A s p en fores t alliance Whit e P ine forest alliance Sp ruce/Bals am F ir Black Sp ruce/Q uaking A sp en fores t alliance A N D /O R W hit e Sp ruce/Bals am F ir/A s p en fores t alliance Black Sp ruce forest alliance Black Sp ruce s at urat ed forest alliance Whit e Sp ruce/Bals am F ir forest alliance C ons olidat ed C las s es O riginal G IS H abit at C las s ificat ion B ur O ak B ur O ak/O ak (W hit e, N ort hern P in, B lack) w oodland alliance B ur O ak fores t alliance A s p en/B irch P ap er B irch fores t alliance Q uaking A s p en/P ap er B irch fores t alliance T rembling A s p en t emp orarily flooded fores t alliance Q uaking A s p en w oodland alliance H erbaceous A lliance C anada B luejoint s eas onally flooded herbaceous alliance C at t ail/B ulrus h s em ip ermanent ly flooded herbaceous alliance C ommon R eed s emip ermanent ly flooded herbaceous alliance F ew -s eeded/W iregras s Sedge s at urat ed herbaceous alliance H ards t em/Soft s t em B ulrus h s emip ermanent ly flooded herbaceous alliance M os aic (1 s at urat ed dw arf s hrubland alliance and 3 w et land herbaceous alliances ) M os aic/C omp lex (5 w et land herbaceous alliances ) M os aic/C omp lex (7 w et land herbaceous alliances ) P ondw eed/H ornw ort /W at erw eed p ermanent ly flooded herbaceous alliance P overt y G ras s herbaceous alliance Yellow /W hit e W at er Lily p ermanent ly flooded herbaceous alliance W ild R ice s emip erm anent ly flooded herbaceous alliance ALCES VOL. 40, 2004 COBB ET AL. - SYMPATRIC MOOSE AND DEER 175 groups “present” and those without pellet groups “absent”. For WTD, sampling units were stratified on the basis of abundance into low (1 – 21 pellet groups) and high (>21 – 42 groups) use areas. We determined the average percent composition of vegetation types, canopy densities, and canopy heights for sampling units within each stratum of moose and WTD from the modified GIS vegetation map of VNP. We performed t- tests to examine differences between habi- tat proportions in absent vs. present stratum of moose sampling units and high vs. low stratum of WTD sampling units. Capture and Radio Telemetry We fitted 10 moose (3 bulls, 7 cows) with motion-sensing radio telemetry collars on the Kabetogama Peninsula between Feb- ruary 26 and March 2, 1989. Each moose was immobilized and sedated with a mixture of carfentantil and xylazine hydrochloride via a barbed syringe fired from a helicopter. The immobilizing drugs were reversed with a hand injection of naltrexone. We fitted motion-sensing radio telemetry collars on 20 white-tailed deer (9 bucks, 11 does) within VNP between January 24 and March 9, 1989. Thirteen WTD were captured within the Moose Bay-Black Bay region, 5 along the Daley Brook snowmobile trail, and 2 on or adjacent to Cutover Island. WTD were captured in collapsible clover traps (Clover 1956) and immobilized using a pole-mounted syringe with a mixture of ketamine hydrochloride (Ketaset) and xylazine hydrochloride. The immobilizing drugs were reversed with an intravenous hand injection of talozoline. Instrumented moose and WTD were relocated via aerial radio telemetry at approximately 10-day intervals from January 24, 1989 to May 16, 1991. Relocations were attempted on all animals throughout the study period unless there was a mechanical failure in the radio collar or the animal was confirmed dead. We calculated moose and WTD home ranges rather than use individual point relocations in habitat analyses to account for errors associated with radio telemetry point relocations (Kernohan et al. 1998). We created 90% adaptive kernel home ranges for individual moose and WTD using Home Range Extension (Rodgers and Carr 1998) in ArcView 3.2 (ESRI 2000). The data were standardized by dividing each value of x and y by its respective standard deviation (Seaman and Powell 1996). We calculated the smoothing factor (h) indi- vidually for each animal using the biased cross-validation (BCV) method (Sain et al. 1994). We limited our calculation of annual home ranges to those animals for which we secured a minimum of 30 relocations since kernel home range estimates suffer from inaccuracies and inflated sizes when small numbers of animal locations are used (Sea- man et al. 1999). All moose and 15 WTD were used in the home range analysis. One collared WTD fawn, approximately 9 months old at capture, was initially associated with a collared doe. This animal was included in the analysis because it established its own individual home range soon after capture, and therefore its locations were independ- ent from its dam. We tested for significant differences between the mean sizes of moose and WTD home ranges and between male and female WTD mean home range sizes in VNP. Samples of male moose (n = 3) were inadequate to test for sexual differences in moose home range size. Habitat Availability We determined available habitats sepa- rately for moose and WTD as 2 extended 100% minimum convex polygons (MCP) containing either all moose or all white- tailed deer radio-telemetry locations within our study area (Fig. 3). We widened each MCP by 1.2 km for moose and 0.5 km for SYMPATRIC MOOSE AND DEER – COBB ET AL. ALCES VOL. 40, 2004 176 Deer Buffered MCP Moose Buffered MCP Habitat Cover Boundary 5 0 5 10 Kilometers N EW S Fig. 3. Buffered minimum convex polygons of moose and WTD radiotelemetry locations, depicting areas used to delineate habitat availability for compositional analysis. WTD to encompass the entire adaptive kernel home ranges of all study animals and calculated the percent composition of each habitat type available to moose and WTD using the extended MCPs. We clipped the modified GIS habitat coverage to individual moose and WTD adaptive kernel home ranges using ArcView 3.2 Patch Analyst extension (Rempel and Carr 2003). Lakes, ponds, and streams were excluded from habitat use or habitat selection calculations, however, vegetation adjacent to bodies of water were considered in the analysis. In addition to the 8 vegetation types, we exam- ined the height and density of canopy cov- ers within areas available to moose and WTD. Habitat Selection We compared habitat use to habitat availability using compositional analysis (Aebischer et al. 1993). Analysis was performed using the BYCOMP program (Ott and Hovey 1997) within the SAS work- ing environment (SAS Institute Inc. 2000). This program first determined whether habi- tat use differed from random using Wilks’ Lambda ( ) statistics in multivariate analy- sis of variance (MANOVA). If habitat use was nonrandom, habitats were ranked in order of preference and levels of signifi- cance between ranks were determined us- ing a t-test. Only home ranges located entirely within the study area (GIS vegetation coverage extent) were included in habitat analyses. One migratory moose and one migratory WTD did not meet this criterion and were not used. Small sample sizes precluded seasonal and sexual habitat selection analy- sis. A minimum of 10 animals per group (season or sex) is needed to produce reli- able results using compositional analysis (Aebischer et al. 1993). Our data would not have met these standards when partitioned into groups by season or sex. Compositional λ ALCES VOL. 40, 2004 COBB ET AL. - SYMPATRIC MOOSE AND DEER 177 analysis required that each animal use all habitat types (Aebischer et al. 1993). When proportional habitat use was estimated to be zero for moose and WTD, we replaced these values with 0.001. Substituting a value smaller than the smallest recorded nonzero value produced results that were robust relative to the substituted value (Aebischer et al. 1993). RESULTS Pellet Group Counts A total of 1,674 deer pellet groups (820 in 1989, 854 in 1991) and 45 moose pellet groups (30 in 1989, 15 in 1991) were enu- merated over all line transect surveys. WTD pellet groups were more abundant than moose pellet groups in all sampling units. Twenty-two of the 32 total sampling units (68.8%) contained no moose pellet groups. All sampling units contained WTD pellet groups at varying abundances ( x = 16.8, SD = 12.0). Moose and WTD pellet groups occur- rence varied spatially across VNP (Fig. 4). Moose pellet groups were present in low (0 - 13) abundances ( x = 1.9, SD = 1.6) in sampling units in the central and eastern regions of the Kabetogama Peninsula and absent from all other sampling units. WTD pellet groups occurred at high (>21 - 42 pellet groups) abundances in sampling units in the central and western regions of Kabetogama Peninsula and in the south- eastern corner of the park, and low densi- ties (0 - 20 pellet groups) in sampling units on the eastern end of the Kabetogama Peninsula. Only 3 sampling units (9.4%) contained moose pellet groups and high numbers of WTD pellet groups. These sampling units were located in the central region of the Kabetogama Peninsula and on the western periphery of moose pellet group distribution. The abundance of moose and WTD pellet groups varied with the average per- cent composition of sample unit habitat types. Sampling units with high abundances of WTD pellet groups contained signifi- cantly more spruce/balsam fir habitat than did sampling units with low abundances of WTD pellet groups (t = 2.79, P = 0.02, Table 2a). Sampling units with moose pellet groups contained significantly less closed/continu- ous canopy cover (t = 2.25, P = 0.04) and significantly more open/discontinuous (t = 2.21, P = 0.04) than did sampling units lacking moose pellet groups (Table 2b). Sampling units with moose pellets also con- tained significantly less 12 – 20 m canopy cover (t = 2.37, P = 0.04) than did sampling units lacking moose pellets (Table 2c). Home Range Moose and WTD were relocated by fixed-wing aircraft on a 10-day mean inter- val (min = 1, max = 119, SD = 14) for a period of 842 days. With outliers removed, 10 moose were relocated 786 times and 20 WTD were relocated 1,032 times. Each moose was relocated an average of 79 times (min = 30, max = 96, SD = 22). Each WTD was relocated an average of 52 times (min = 6, max = 76, SD = 27). Three (33%) radio-collared moose and 8 (40%) radio- collared WTD died during the study. Moose home range averaged 48 km2 (min = 29, max = 141, SD = 33.5). One male moose had an especially large home range of 141 km2 because of seasonal migratory behavior. Excluding this animal, the aver- age annual moose home range size was 37 km2. The average annual WTD home range was 9 km2 (min = 2, max = 49, SD = 4.32). One female WTD that exhibited seasonal migratory behavior had an especially large home range (49 km2). Excluding this ani- mal, the average annual WTD home range was 6 km2. The average moose home range was significantly larger than the average WTD home range (t = 4.16, P < 0.01, Fig. 5). SYMPATRIC MOOSE AND DEER – COBB ET AL. ALCES VOL. 40, 2004 178 Moose Pellet Groups Absent Present N EW S 5 0 5 Kilometers W TD Pellet Groups Low (1 - 21) High (>21 - 42) 5 0 5 Kilometers N EW S Fig. 4. (a) Presence or absence of moose pellet groups and (b) abundance of WTD pellet groups in sampling units at Voyageurs National Park, Minnesota, based on pellet count transects conducted in May 1989 and 1991. (a) (b) ALCES VOL. 40, 2004 COBB ET AL. - SYMPATRIC MOOSE AND DEER 179 at least one other WTD, and all moose home ranges overlapped with at least one other moose. Five of 10 moose home ranges overlapped with WTD home ranges, al- though the overlapping areas were rela- tively small. The total area of overlapping home ranges of instrumented moose and WTD was 6 km2. This area of overlapping home ranges encompassed 2.5% of all moose home range area and 6.0% of all WTD home range area, and was located in the central Kabetogama Peninsula (Fig. 6). Habitat Availability and Use Available moose and WTD habitats were largely similar in terms of vegetation types, canopy height, and canopy density (Table 3). However, 58% of combined moose home ranges vs. 9% of WTD com- bined home ranges had been burned or Table 2. Percent composition of (a) vegetation types, (b) canopy densities, and (c) canopy heights found in moose sampling units that contained pellet groups (present) and that contained no pellets (absent), and white-tailed deer (WTD) sampling units that contained high (27-40) and low (0-12)abundances of pellet groups. Present High Low (a) Vegetation Type Aspen/Birch 23 20.1 26.7 Black Ash/Red Maple 2.7 2 1.4 Spruce/Balsam Fir 14.1 34.4* 12.0* Bur Oak 19.1 6.1 9.6 Herbaceous Alliance 8 6.3 8.8 Jack Pine 11.9 12.8 20.9 Red/White Pine 15.8 16.5 17.3 Shrubland Alliance 5.4 2 3.3 (b) Canopy Density Dispersed/Sparse (10-25%) 0.4 0 0.6 Open/Discontinuous (25-60%) 41.2* 28.3 29.7 Closed/Continuous (60-100%) 58.4* 71.7 70.2 (c) Canopy Heights Open 7.7 5.9 9 <0.5 m 3.2 1.3 1.8 0.5 – 5 m 6.3 5 3.2 5 – 12 m 38.4 29.9 23.7 12 – 20 m 34.6* 49.1 50.6 20 – 30 m 9.8 8.8 11.610.7 11.1 26.5 21 47.6 55.6* 1.6 1.1 4.4 2.9 9.3 8.3 72.5 74.8* 0.4 0 27.1 25.2* 17.6 17.5 3.5 2 9.1 8.1 15.5 20.8 20.4 20 8.9 4.9 23 25.4 2 1.2 Available Absent Moose WTD 0 10 20 30 40 50 60 70 80 moose deer H o m e R an g e A re a (k m 2) Fig. 5. Moose and WTD home range areas (km2), with 95% confidence intervals, at Voyageurs National Park, Minnesota. The average male WTD home range (8 km2) was larger than the average female home range (4 km2), however, the differ- ence was not statistically significant (t = 0.30, P = 0.77), even with the single female migratory WTD removed (t = 1.33, P = 0.21). All WTD home ranges overlapped with * indicates a significant difference ( = 0.05).α SYMPATRIC MOOSE AND DEER – COBB ET AL. ALCES VOL. 40, 2004 180 logged within the last 55 years. Individual moose and WTD 90% kernel home ranges were largely similar in vegeta- tion types with spruce/balsam fir and aspen/ birch types making up > 50% of the vegeta- tion type for both ungulates (Table 4). Jack pine was slightly more abundant than the herbaceous alliance in moose home ranges, whereas herbaceous alliance was the third most abundant vegetation type in WTD kernel home ranges. Moose and WTD showed similar rankings of abundance of canopy densities in their home ranges, how- ever moose home ranges contained less closed/continuous and more open/discon- tinuous canopy densities than WTD. The 2 species differed in canopy height use with almost 50% of the 5 – 12 m height class available to moose and over 50% of the 12 – 20 m height class available to WTD. Habitat Selection Vegetation types within moose home ranges differed significantly from available N EW S Species Overlap Deer Home Ranges Moose Home Ranges 5 0 5 10 Kilometers Fig. 6. Home range overlap of instrumented moose and WTD in and adjacent to Voyageurs National Park, Minnesota. vegetation types ( = 0.01, P = 0.04). Moose showed a significant preference for spruce/ balsam fir over all other types except the shrubland alliance (t = 2.12, P = 0.07) and bur oak (t = 2.22, P = 0.07) (Table 5a). The shrubland alliance, aspen/birch, herba- ceous alliance, bur oak, and red/white pine types all tied for second in preference and did not differ significantly in preference from one another. Moose exhibited signifi- cant nonrandom use of canopy densities ( = 0.30, P = 0.02). Moose significantly se- lected open/discontinuous canopy cover over all others, but exhibited no significant differ- ence in preference between closed/con- tinuous and dispersed/sparse canopies (Ta- ble 5b). Moose exhibited nonrandom use of canopy heights ( = 0.09, P = 0.03) and showed a significant preference for 5 – 12 m canopy cover over open habitat and 12 – 20 m canopy cover (Table 5c). There was no significant difference in preference be- tween 5 – 12 m and 20 – 30 m canopy height (t = 1.02, P = 0.34), or between 5 – 12 m and λ λ λ ALCES VOL. 40, 2004 COBB ET AL. - SYMPATRIC MOOSE AND DEER 181 (t = 5.61, P < 0.01) and open/discontinuous canopies (t = 4.86, P < 0.01) over dispersed/ sparse canopy (Table 6b). Selection for closed/continuous canopy over open/dis- continuous canopy was not significant at P < 0.05 (t = 1.60, P = 0.12). WTD exhibited nonrandom use of canopy heights ( = 0.11, P < 0.01) with a highly significant (P < 0.02) preference for 12 – 20 m canopy over all others (Table 6c). There was no evidence of significant preference for any other canopy height. DISCUSSION Distribution Pellet group sampling provided evidence that WTD were more widespread than moose at VNP. WTD pellet groups occurred in the high stratum toward the western and cen- tral portions of the Kabetogama Peninsula 0.5 – 5 m (t = 2.39, P = 0.06) or <0.5 m (t = 2.25, P = 0.06) canopy heights. Vegetation types within WTD home ranges differed from available habitats but not significantly at P-value < 0.05 ( = 0.25, P = 0.09). WTD significantly selected spruce/bal- sam fir over all other vegetation types ex- cept aspen/birch (t = 2.03, P = 0.06) (Table 6a). Aspen/birch was significantly selected over all remaining vegetation types except herbaceous alliance (t = 1.83, P = 0.10). Jack pine and bur oak tied for lowest in WTD preference at the home range scale. WTD exhibited significant nonrandom use of canopy densities ( = 0.27, P < 0.01), significantly selecting closed/continuous Table 3. Percent composition of available moose and white-tailed deer (WTD) (a) vegetation types, (b) canopy densities, and (c) canopy heights based on expanded minimum convex polygons of all moose and all WTD locations, respectively. M oose WT D (a) Veget at ion T y p e A sp en/Birch 24.1 28 Black A sh/Red M ap le 0.9 3.2 Sp ruce/Balsam F ir 26.8 26.6 Bur O ak 7.6 2.7 H erbaceous A lliance 11.2 13.3 Jack P ine 18.1 6.4 Red/Whit e P ine 5.1 6.1 Shrubland A lliance 6.2 13.8 (b) Canop y D ensit y D isp ersed/Sp arse (10-25% ) 0.4 0.3 O p en/D iscont inuous (25-60% ) 29.3 22.7 Closed/Cont inuous (60-100%) 70.3 77 (c) Canop y H eight s O p en 11.4 13.6 <0.5 m 2.2 2 0.5 – 5 m 6.3 9.8 5 – 12 m 34.3 33.1 12 – 20 m 44 39.3 20 – 30 m 1.8 2.1 Moose WTD (a) Vegetation Type Aspen/Birch 22.6 28.2 Black Ash/Red Maple 0.5 2.5 Spruce/Balsam Fir 36.9 40.5 Bur Oak 7.2 2.1 Herbaceous Alliance 10 11 Jack Pine 11.6 3.1 Red/White Pine 4.4 5.5 Shrubland Alliance 6.8 7.3 (b) Canopy Density Dispersed/Sparse (10-25%) 0.3 0.1 Open/Discontinuous (25-60%) 37.4 20.4 Closed/Continuous (60-100%) 62.3 79.5 (c) Canopy Heights Open 10.1 11.1 <0.5 m 2.1 1.5 0.5 – 5 m 6.9 5.8 5 – 12 m 45.6 24 12 – 20 m 33.3 54.8 20 – 30 m 2.1 2.9 Table 4. Average percentage habitat use within moose and white-tailed deer (WTD) 90% adap- tive kernel home ranges. λ λ λ SYMPATRIC MOOSE AND DEER – COBB ET AL. ALCES VOL. 40, 2004 182 and along the periphery of the southeastern portion of VNP while moose pellet groups were restricted to the central/eastern re- gion of the Kabetogama Peninsula. The distribution of moose based upon our pellet group sampling and home range calcula- tions were similar, and both agreed with the distribution of moose determined during aerial censuses (Gogan et al. 1997). The Table 5. Simplified ranking matrices for moose based on comparing proportional (a) vegetation type, (b) canopy density, and (c) canopy height use within 90% adaptive kernel home range to proportions available within the available area (extended MCPs). Habitat classes are ranked from most preferred to least preferred. Habitat classes that differ significantly in preference from random at P = 0.05 are indicated by either a “+++” or “—”. Habitat classes that differ in preference from random at P = 0.10 are indicated by either a “++” or “—”. Habitat classes that differ in preference from random at P > 0.10 are indicated by either a “+” or “-”. (a) Vegetation Type R A N K Spruce/B alsam Fir Shrubland A lliance A spen/B irch H erbaceous A lliance B ur O ak R ed/W hite Pine Jack Pine B lack A sh/R ed M aple Spruce/Balsam Fir 1 . ++ +++ +++ ++ +++ +++ +++ Shrubland Alliance 2 -- . + + + ++ +++ +++ Aspen/Birch 3 --- - . + + + +++ +++ Herbaceous Alliance 4 --- - - . + + +++ +++ Bur Oak 5 -- - - - . + ++ ++ Red/White Pine 6 --- -- - - - . +++ +++ Jack Pine 7 --- --- --- --- -- --- . + Black Ash/Red Maple 8 --- --- --- --- -- --- - . (b) Canopy Density R A N K O pen/ D iscontinuous (25-60% ) C losed/ C ontinuous (60-100% ) D ispersed/ Sparse (10-25% ) Open/Discontinuous (25-60%) 1 . +++ +++ Closed/Continuous (60-100%) 2 --- . ++ Dispersed/Sparse (10-25%) 3 --- -- . (c) Canopy Heights R A N K 5-12 m 20-30 m 0.5-5 m O pen <0.5 m 12-20 m 5 – 12 m 1 . + ++ +++ ++ +++ 20 – 30 m 2 - . + + + +++ 0.5 – 5 m 3 -- - . + + + Open 4 --- - - . + + <0.5 m 5 -- - - - . + 12 – 20 m 6 --- --- - - - . capture and instrumenting of moose and WTD was completed prior to the establish- ment of the pellet group sampling transects and was therefore not dependent on our sampling of pellet groups. Our calculated home ranges of instrumented WTD were largely coincident with the distribution of pellet group units that we assigned to the high WTD stratum. Trapping locations for ALCES VOL. 40, 2004 COBB ET AL. - SYMPATRIC MOOSE AND DEER 183 the species in the contiguous United States. Annual home ranges of WTD at VNP (x = 5.74 km2, n = 15) were much larger than those in northeastern Minnesota (MCP, 0.8 km2 summer, 0.4 km2 winter) (Nelson and Mech 1981). In general, the home ranges of WTD at the northern limits of the species distribution are larger than those in the s o u t h e r n p e r i p h e r y o f t h e i r r a n g e (Severinghaus and Cheatum 1956). There are a number of causes for vary- ing home range sizes between locations. Home range size might be dictated directly by an animal’s energetics (McNab 1963). Following this theory, animals of the same species in more productive habitats have smaller home ranges than those in poor habitats, as the latter require greater areas to secure the resources required for sur- vival. Other factors possibly influencing moose and WTD home range sizes at VNP include reproductive activity, relative distri- bution and diversity of suitable habitats, and species density (Leptich and Gilbert 1989, Beier and McCullough 1990, Ballard et al. 1991). Sexual differences in home range size have been reported for moose and WTD. Males of both species usually occupy larger home ranges than females (Carlsen and Farmes 1957, Ballard et al. 1991) although no difference between male and female home range sizes has been observed in some areas (Phillips et al. 1973, Taylor and Ballard 1979, Hauge and Keith 1981). There was no significant difference between male and female WTD home range sizes at VNP. Our relatively small WTD sample caused large confidence intervals and strong outlier effects. Moose and WTD in northern regions typically undergo significant seasonal home range shifts (Messier and Barrette 1985, Van Deelen et al. 1998). Yarding behavior by WTD is common in northern regions (Telfer 1967, Rongstad and Tester 1969, WTD were based upon our observations of high concentrations of deer in winter and not on the distribution of high densities of pellet groups based upon our pellet group sampling in May. While WTD pellet groups occurred in all sampling units, only 3 sampling units assigned to the high WTD stratum also contained moose pellet groups. This pat- tern is indicative of differing habitat use patterns between the 2 species. The in- verse relationship between moose and white- tailed deer distributions at VNP is consist- ent with observations in adjacent Ontario, where moose reached their highest densi- ties in areas where white-tailed deer were <4/km2 (Whitlaw and Lankester 1994). Moose densities in Ontario were inversely related to the mean intensity of meningeal worm larvae in white-tailed deer pellet groups (Whitlaw and Lankester 1994). There is currently no information on the relative spatial distribution of meningeal worm larvae in WTD pellet groups at VNP. Home Range Moose 90% adaptive kernel home ranges in VNP (x = 47.7 km2, n = 10) were among the largest recorded in the contigu- ous United States, and were much larger than those found in northwestern Minne- sota (x male = 3.1 km 2, x female = 3.6 km 2, n = 26) (Phillips et al.1973) and northwestern Ontario (x = 14.0 km2, n = 1) (Addison et al. 1980). The differences may be greater than these comparisons of size alone suggest since the other studies used methods that generally produce larger home range esti- mates than does the adaptive kernel method used here. Moose home ranges at VNP were substantially smaller than those in Alaska, where home range sizes (MCP) are between 120 km2 and 350 km2 (Gravogel 1984). WTD adaptive kernel home ranges at VNP were larger than most recorded for SYMPATRIC MOOSE AND DEER – COBB ET AL. ALCES VOL. 40, 2004 184 Table 6. Simplified ranking matrices for white-tailed deer (WTD) based on comparing proportional (a) vegetation type, (b) canopy density, and (c) canopy height use within 90% adaptive kernel home range to proportions available within the available area (extended MCPs). Habitat classes are ranked from most preferred to least preferred. Habitat classes that differ significantly in preference from random at P = 0.05 are indicated by either a “+++” or “—”. Habitat classes that differ in preference from random at P = 0.10 are indicated by either a “++” or “—”. Habitat classes that differ in preference from random at P > 0.10 are indicated by either a “+” or “-”. (a) Vegetation Type R A N K Spruce/B alsam Fir A spen/B irch H erbaceous A lliance B lack A sh/R ed M aple Shrubland A lliance R ed/W hite Pine Jack Pine B ur O ak Spruce/Balsam Fir 1 . ++ +++ +++ +++ +++ +++ +++ Aspen/Birch 2 -- . ++ +++ +++ +++ +++ +++ Herbaceous Alliance 3 --- -- . + +++ ++ +++ +++ Black Ash/Red Maple 4 --- --- - . +++ ++ +++ +++ Shrubland Alliance 5 --- --- --- --- . + + ++ Red/White Pine 6 --- --- -- -- - . + + Jack Pine 7 --- --- --- --- - - . + Bur Oak 8 --- --- --- --- -- - - . (b) Canopy Density R A N K C losed/ C ontinuous (60-100% ) O pen/ D iscontinuous (25-60% ) D ispersed/ Sparse (10-25% ) Closed/Continuous (60-100%) 1 . + +++ Open/Discontinuous (25-60%) 2 - . +++ Dispersed/Sparse (10-25%) 3 --- --- . (c) Canopy Heights R A N K 12 – 20 m O pen 5 – 12 m 0.5 – 5 m 20 – 30 m <0.5 m 12 – 20 m 1 . +++ +++ +++ +++ +++ Open 2 --- . + ++ + ++ 5 – 12 m 3 --- - . + + + 0.5 – 5 m 4 --- -- - . + + 20 – 30 m 5 --- - - - . + <0.5 m 6 --- -- - - - . Nelson 1998), and by moose in eastern Canada (Proulx 1983). Approximately 80% of white-tailed deer in nearby Superior National Forest, Minnesota, exhibit migra- tory behavior (Nelson 1998). Migratory behavior has traditionally been thought of as an adaptive response to the presence of snow (Townsend and Smith 1933) or an anti-predator response (Nelson and Mech 1991). None of the WTD on the Kabetogama Peninsula exhibited migratory behavior typi- cal of yarding based on the size and shape of their annual home ranges. One male moose extended its range into adjacent ALCES VOL. 40, 2004 COBB ET AL. - SYMPATRIC MOOSE AND DEER 185 Ontario in summer. Two white-tailed deer captured and radio marked in the vicinity of D a l e y B a y s e a s o n a l l y m i g r a t e d during warmer months beyond the bounda- ries of the study area and returned to Daley Bay during the winter. Vegetation in the vicinity of Daley Bay is predominately north- ern white cedar swamp (classified as shrubland alliance in this study), which is typically associated with white-tailed deer wintering yards (Crawford 1982). WTD may not exhibit migratory behavior in VNP because of an abundance of winter habitat. Forest maturation in adjacent Su- perior National Forest during the early 1970s provided white-tailed deer with abundant winter cover, and may have allowed the population to disperse into smaller groups rather then exhibit common yarding behavior (Wetzel et al. 1975). Winter cover, such as balsam fir, was abundant in VNP during this study due to a long-term absence of major wildfires and limited timber harvest in VNP since the 1930s, particularly in the area utilized by instrumented WTD. The poten- tial abundance of preferred winter cover may be a reason for the non-migratory behavior of WTD at VNP. There was minimal overlap between instrumented moose and white-tailed deer home ranges. Home range overlap be- tween the two species was limited to a small region on the central Kabetogama Penin- sula. Concurrent pellet group sampling showed the same region of the Kabetogama Peninsula to be the only area within VNP where moose pellet groups occurred within the high-density stratum for WTD pellet groups. Pellet group sampling showed that WTD occur throughout VNP at varying densities and that on the Kabetogama Pe- ninsula, moose are generally limited to the central and eastern area. Habitat Selection Our pellet group surveys provided in- formation on winter habitat use only, while the radio telemetry data provided informa- tion on the habitat characteristics of year- round home ranges. The pellet group sur- veys indicate that WTD preferred the spruce/balsam fir vegetation type over all others but did not select for canopy height or density. In contrast, moose showed no preference for any vegetation type but did prefer the open/discontinuous canopy den- sities, and avoided the dispersed/sparse canopy densities and 12 – 20 m canopy height. The year-round telemetry data re- veal a different pattern with WTD and moose selecting for vegetative type, canopy density, and canopy height. We attribute this difference to differences in sampling intensity and portions of the annual cycle embraced by each sampling technique. Moose distributions appeared to be re- lated by the distribution of canopy heights and canopy densities in VNP. Lower canopy heights and canopy densities were more prevalent in sampling units containing moose pellets. The most common canopy height in sampling units containing moose was 5 – 12 m, but the most common canopy height in sampling units containing high abundances of white-tailed deer pellets was 12 – 20 m. Lower discontinuous vegetation might pro- vide moose with more accessible forage in the winter. Both moose and WTD at VNP exhib- ited a strong preference for spruce/balsam fir habitat types. This type satisfies differ- ent needs of each herbivore species. Moose habitat use generally is dictated by food abundance rather than shelter (Telfer 1970) and balsam fir is an important source of forage for moose in boreal forests, espe- cially during the winter season (Irwin 1975, Peek et al. 1976, Ludewig and Bowyer 1985, Allen et al. 1987). However, balsam fir would also provide moose at VNP with a refuge from deep snow during winter: spruce and balsam fir were among the dominant SYMPATRIC MOOSE AND DEER – COBB ET AL. ALCES VOL. 40, 2004 186 overstory species in moose winter yards in southern Quebec (Proulx 1983). WTD are less adapted to harsh winter conditions than moose and select winter habitat based on thermal protection and shelter, rather than forage preference (Telfer and Kelsall 1979). Spruce and balsam fir are considered poor quality forage for WTD (Crawford 1982, Blouch 1984) but offer WTD ideal winter protection. WTD in northeastern Minne- sota used balsam fir dominated stands fre- quently in late winter (Wetzel et al. 1975). The unusually harsh winter conditions dur- ing our study might have caused white- tailed deer to utilize winter shelter habitats such as the spruce/balsam fir type for longer periods and at higher levels than usual. Monthly snowfall exceeded 50 cm in three consecutive months during the winters of 1988 – 1989 and 1990 – 1991, and remained into early spring following these winters. WTD movements become restricted at snow depths of approximately 30 cm and con- fined at snow depths > 50 cm (Telfer 1970). The aspen/birch vegetation type ranked second in WTD preference. Both species are preferred WTD forage in boreal forests (Cairns and Telfer 1980, Ludewig and Bowyer 1985). WTD in adjacent Ontario annually used aspen and birch more than other vegetation types (Kearney and Gil- bert 1976). The aspen/birch type does not offer WTD as much thermal cover and snow protection as the spruce/balsam fir type. The shrubland alliance, aspen/birch, and herbaceous alliance vegetation types all ranked second in moose preference. These types offer moose important forage spe- cies. Moose in northern Minnesota forage extensively on aspen and birch (Peek et al. 1976). The shrubland alliance includes northern red osier dogwood (Cornus stolonifera) and willows (Salix spp.) while the herbaceous alliance type contains hy- drophilic plant species. Moose eat both the woody and hydrophilic plants during the early spring and fall seasons in boreal for- ests (Peek et al. 1976, Jordan 1987). Aquatic habitats also may provide moose with an escape from biting insects during the early summer season (Ritcey and Verbeek 1969). Moose and WTD selected different canopy densities and heights at VNP. Moose exhibited a significant preference for open/ discontinuous canopy density to all other canopy densities and for canopy heights of 5 – 12 m and 20 – 30 m. WTD preferred closed/discontinuous canopy to open/dis- continuous canopy densities and significantly preferred 12 – 20 m canopy height to all others. Differences between moose and WTD in body size may be an important factor in the differences in canopy prefer- ences between these species. The amount of energy that ungulates expend in moving increases linearly with increasing snow depth, until breast height, at which it in- creases exponentially (Parker et al. 1984). Dense canopy cover displaces snow and causes structural changes to snow that in- fluence the energy ungulates need to move and forage (Kirchoff and Schoen 1987). WTD are more restricted by snow depth and cold temperatures than are moose and therefore utilize vegetative types that allow movement during periods of deep snow (Telfer 1970). Moose are less affected by snow conditions than WTD and therefore would be more likely to select for forage availability rather than thermal and snow protection during winter. Moose winter yards in southern Quebec were in slightly closed canopy forest (41-80%) with tree heights of 9.1 – 18.3 m (Proulx 1983). However, these yards tended to be on slopes that reduced the energetic cost of moving t h r o u g h d e e p s n o w ( P r o u l x 1 9 8 3 ) . Moose selection of 20-30 m canopy cover was unexpected, however there are a few possible explanations for these obser- vations. Low abundances of 20-30 m canopy ALCES VOL. 40, 2004 COBB ET AL. - SYMPATRIC MOOSE AND DEER 187 class were clustered throughout VNP, com- posed primarily of red/white pine overstory habitat. Although moose typically do not exhibit a high preference for white and red pine (Peek et al. 1976), as they did not in this study, understory vegetation in red/white pine habitat included aspen and birch that are palatable forage for moose. Heat stress avoidance is another possible explanation for moose selecting high canopy cover (Kelsall and Telfer 1974). Moose at VNP are at the southern periphery of the species’ distribution. Southern populations of moose select forested upland sites during the sum- mer season, possibly to reduce energy ex- penditure and enable them to forage for longer periods per day (Miller and Litvaitis 1992). During warm periods, the overhead canopy of a 20-30 m coniferous canopy could shade moose and reduce ambient temperatures better than other habitats. Meningeal Worm Refuge We did not find a single sampling unit that was free of WTD pellets. It is there- fore highly unlikely that moose had a com- plete refuge from meningeal worm-infected WTD within VNP. Lower densities of WTD within moose range however might reduce the rate of meningeal worm trans- mission to moose, and thereby increase moose survival. This partial refuge could be enough to allow moose to survive in the presence of WTD. Moose are able to maintain low population levels in the pres- ence of meningeal worm if WTD numbers do not exceed 4.6/km2 (Karns 1967). Moose, WTD, and meningeal worm have existed sympatrically in Ontario since the early 1980s, and their interaction does not appear to be negatively affecting moose population numbers (Whitlaw and Lankester 1994). The prevalence of meningeal worm in WTD in northwestern Ontario was simi- lar to VNP based upon the presence of larvae in feces and adults in the cranium (Gogan et al. 1997). Our pellet group surveys and radio te- lemetry data reveal evidence of spatial seg- regation and resource partitioning between moose and WTD at VNP. This spatial separation and differences in habitat pref- erences between moose and white-tailed deer at VNP may reduce the prevalence of meningeal worm infection in moose by pro- viding moose with a partial refuge from meningeal worm-infected WTD (Gilbert 1974). The high prevalence of meningeal worm in WTD at VNP (Gogan et al. 1997) indicates that the parasite and its gastropod intermediate hosts are sufficiently abun- dant for transmission of the parasite to moose. The parasite is highly pathogenic to moose, and therefore moose are not ex- pected to persist in its presence at VNP unless they are somehow isolated from in- fection. Moose densities in adjacent On- tario were lowest in areas with the highest mean intensity of meningeal worm larvae in WTD pellet groups (Whitlaw and Lankester 1994). The potential for meningeal worm transmission to moose may be even greater at VNP than Ontario since WTD densities and the prevalence of meningeal worm are higher at VNP than in most regions in Ontario (Whitlaw and Lankester 1994, Gogan et al. 1997). Partial habitat separa- tion between moose and WTD at VNP may be a factor allowing moose to persist de- spite a high prevalence of meningeal worm. REFERENCES AEBISCHER N. J., P. A.ROBERTSON, and R. E. KENWARD. 1993. Compositional analy- sis of habitat use from animal radio- tracking data. Ecology 74: 1313 – 1325. ADDISON, R. B., J. C. WILLIAMSON, B. P. SA U N D E R S , and D. F R A S E R. 1980. Radiotracking of moose in the boreal forest of northwestern Ontario. Cana- dian Field Naturalist 94: 269 – 276. ALLEN, A. W., P. A. JORDAN, and J. W. SYMPATRIC MOOSE AND DEER – COBB ET AL. ALCES VOL. 40, 2004 188 TERRELL. 1987. Habitat suitability index models: moose, Lake Superior region. Fish and Wildlife Service. Biological Report 82 (10.155). BALLARD, W. B., J. S. WHITMAN, and D. J. REED. 1991. Population dynamics of moose in south-central Alaska. Wild- life Monographs 114. BEIER, P., and D. R. MCCULLOUGH. 1990. Factors influencing white-tailed deer activity patterns and habitat use. Wild- life Monographs 54. BLOUCH, R. I. 1984. Northern Great Lakes states and Ontario forests. Pages 391- 410 in L.K. Halls, editor. White-tailed D e e r : E c o l o g y a n d M a n a g e m e n t . Stackpole Books, Harrisburg, Pennsyl- vania, USA. CAIRNS, A. L., and E. S. TELFER. 1980. Habitat use by 4 sympatric ungulates in boreal mixed wood forest. Journal of Wildlife Management 44: 849 – 857. CARLSEN, J. C., and R. E. FARMES. 1957. Movements of white-tailed deer tagged in Minnesota. Journal of Wildlife Man- agement 21: 397 – 401. CLOVER, M. R. 1956. Single-gate deer trap. California Fish and Game 42: 199-201. COLE, G. F. 1987. Changes in interacting species with disturbance. Environmen- tal Management 11: 257 – 264. CRAWFORD, H. S. 1982. Seasonal food selection and digestibility by tame white- tailed deer in central Maine. Journal of Wildlife Management 46: 974 – 982. CROWLEY, K. F., and K. L. COLE. 1995. Patterns of temporal and spatial change in the vegetation of Voyageurs National Park. Bulletin of the Ecological Society of America 77: 57. (ESRI) ENVIRONMENTAL SYSTEMS RESEARCH I N S T I T U T E . 2 0 0 0 . A r c V i e w 3 . 2 . Redlands, California, USA. FULLER, T. K. 1986. Observations of moose, Alces alces, in peripheral range in north-central Minnesota. Canadian Field Naturalist 100: 359 – 362. GILBERT, F. F. 1974. Parelaphostrong- ylus tenuis in Maine: prevalence in moose. Journal of Wildlife Manage- ment 30: 200 – 202. GOGAN, P. J. P., K. D. KOZIE, E. M. OLEXA, and N. S. DUNCAN. 1997. Ecological status of moose and white tailed deer at Voyageurs National Park, Minnesota. Alces 33: 187 – 201. GRAUVOGEL, C. A. 1984. Seward Peninsula moose population identity study. Fed- eral Aid for Wildlife Restoration Final Report. Alaska Department of Fish and Game, Juneau, Alaska, USA. HAUGE, T. M., and L. B. KEITH. 1981. Dynamics of moose populations in north- eastern Alberta. Journal of Wildlife Management 45: 573 – 597. HURRELL, J. W. 1995. Decadal trends in the North Atlantic Oscillation: regional tem- peratures and precipitation. Science 269: 676 – 679. IRWIN, L. L. 1975. Deer-moose relation- ships on a burn in northeastern Minne- sota. Journal of Wildlife Management 39: 653 – 662. JOHNSON, C., and J. SALES. 1995. Influence of geomorphology on the vegetation of Voyageurs National Park. Poster. Tenth Annual U.S. Landscape Ecology Sym- posium. International Association of Landscape Ecology, University of Min- nesota. JORDAN, P. A. 1987. Aquatic foraging and the sodium ecology of moose: a review. Swedish Wildlife Research Supplement 1:119-137. KARNS, P. D. 1967. Parelaphostrongylus tenuis in deer in Minnesota and implica- tions for moose. Journal of Wildlife Management 31: 299 – 303. _____. 1998. Population distribution, den- sity and trends. Pages 125 – 139 in A.W. Franzmann and C.C. Schwartz , editors. Ecology and Management of t h e N o r t h A m e r i c a n M o o s e . ALCES VOL. 40, 2004 COBB ET AL. - SYMPATRIC MOOSE AND DEER 189 Smithsonian Institution Press, Wash- ington D.C., USA. KEARNEY, S. R., and F. F. GILBERT. 1976. Habitat use by white-tailed deer and moose on sympatric range. Journal of Wildlife Management 40: 645 – 657. KELSALL, J. P., and E. S. TELFER. 1974. Biogeography of moose with particular reference to western North America. Naturaliste Canadien 101: 117 – 130. KERNOHAN, B. J., J. J. MILLSPAUGH, J. A. JENKS, and D. E. NAUGLE. 1998. Use of an adaptive kernel home-range estima- tor in a GIS environment to calculate habitat use. Journal of Environmental Management 53: 83 – 89. KIRCHOFF, M. D., and J. W. SCHOEN. 1987. Forest cover and snow: implications for deer habitat in southeast Alaska. Jour- nal of Wildlife Management 51: 28 – 33. KURMIS, V., S. L. WEBB, and L. C. MERRIAM, JR. 1986. Plant communities of Voyageurs National Park, Minnesota, U.S.A. Canadian Journal of Botany 64: 531 – 540. LAMB, P. J., and R. A. PEPPLER. 1987. North Atlantic oscillation: concept and application. Bulletin of the American Meteorological Society 68: 1218 – 1225. LEPTICH, D. J., and J. R. GILBERT. 1989. Summer home range and habitat use of moose in northern Maine. Journal of Wildlife Management 53: 880 – 884. LUDEWIG, H. A., and T. BOWYER. 1985. Overlap in winter diets of sympatric moose and white-tailed deer in Maine. Journal of Mammalogy 66: 390 – 392. MCNAB, B. K. 1963. Bioenergetics and the determination of home range. Ameri- can Naturalist 97: 130 – 140. MECH, L. D., and P. D. KARNS. 1978. Role of the wolf in a deer decline in the Superior National Forest. North Cen- tral Forest Experimental Station. U.S. Forest Service Paper NC-148. St. Paul, Minnesota, USA. MESSIER, F., and C. BARRETTE. 1985. The efficiency of yarding behaviour by white- tailed deer as an anti-predator strategy. Canadian Journal of Zoology 63: 785 – 789. MILLER, B. K., and J. A. LITVAITIS. 1992. Habitat selection by moose in a boreal forest ecotone. Acta Theriologica 37: 41 – 50. NELSON, M. E. 1998. Development of mi- gratory behavior in northern white-tailed deer. Canadian Journal of Zoology 76: 426 – 432. _____, and L. D. MECH. 1981. Deer social organization and wolf predation in north- eastern Minnesota. Wildlife Mono- graphs 77. _____, and _____. 1991. Wolf predation risk associated with white-tailed deer movements. Canadian Journal of Zool- ogy 69: 2696 – 2699. NEU, C. W., C. R. BYERS, and J. M. PEEK. 1974. A technique for analysis of utili- zation-availability data. Journal of Wild- life Management 38: 541 – 545. NUDDS, T. D. 1990. Retroductive logic in retrospect: the ecological effects of meningeal worm. Journal of Wildlife Management 56: 617 – 619. OHMANN, L. F., and R. R. REAM. 1971. Wilderness ecology: virgin plant com- munities of the Boundary Waters Ca- noe Area. North Central Forest Ex- perimental Station. U.S. Forest Service Resource Paper NC – 63. St. Paul, Minnesota, USA. OLSON, H. F. 1938. Deer tagging and population studies in Minnesota. Trans- actions of the North American Wildlife and Natural Resources Conference 3: 280 – 286. OTT, P., and F. HOVEY. 1997. BYCOMP. SAS version 1.0. British Columbia Forest Service, Victoria, British Co- lumbia, Canada. SYMPATRIC MOOSE AND DEER – COBB ET AL. ALCES VOL. 40, 2004 190 PARKER, K. L., C. T. ROBBINS, and T. A. HANLEY. 1984. Energy expenditure for locomotion by mule deer and elk. Jour- nal of Wildlife Management 48: 474 – 488. PASTOR, J. R., and D. J. MLADENOFF. 1992. The southern boreal-northern hardwood forest border. Pages 216-229 in H.H. Shugart, R. Leemans, and G.B. Bonan, editors. A System Analysis of the Glo- bal Boreal Forest. Cambridge Univer- sity Press, New York, USA. PEEK, J. M., P. L. URICH, and R. J. MACKIE. 1976. Moose habitat selection and re- lationships to forest management in northwestern Minnesota. Wildlife Monographs 48. PETERSON, R. L. 1955. North American Moose. University of Toronto Press, Toronto, Ontario, Canada. PETRABORG, W. H., and D. W. BURCALOW. 1965. The white-tailed deer in Minne- sota. Pages 11 – 48 in J.B. Moyle, editor. Big Game in Minnesota. Min- nesota Department of Conservation, Technical Bulletin No. 9. PHILLIPS, R. L., W. E. BERG, and D. B. SINIFF. 1973. Moose movement pat- terns and range use in northwestern Minnesota. Journal of Wildlife Man- agement 37: 266 – 278. PRESCOTT, W. H. 1974. Interrelationships of moose and deer of the genus Odocoileus. Naturaliste Canadien 101: 493 – 504. PROULX, G. 1983. Characteristics of moose (Alces alces) winter yards on different exposures and slopes in southern Que- bec. Canadian Journal of Zoology 61: 112 – 118. REMPEL, R. S., and A. P. CARR. 2003. Patch Analyst for ArcView 3.2. Ontario Min- istry of Natural Resources, Centre for Northern Forest Ecosystem Research, Thunder Bay, Ontario, Canada. RITCEY, R. W., and N. A. M. VERBEEK. 1969.Observations of moose feeding on aquatics in Bowron Lake Park, Brit- ish Columbia. Canadian Field Natural- ist 83: 339 – 343. RODGERS, A. R., and A. P. CARR. 1998. HRE: the Home Range Extension for ArcView. Ontario Ministry of Natural Resources, Centre for Northern Forest Ecosystem Research, Thunder Bay, Ontario, Canada. RONGSTAD, O. J., and J. R. TESTER. 1969. Movements and habitat use of white- tailed deer in Minnesota. Journal of Wildlife Management 33: 366 – 379. SAIN, S. R., K. A. BAGGERLY, and D. W. SCOTT. 1994. Cross-validation of multivariate densities. Journal of the American Statistical Association 89: 807 – 817. SAS INSTITUTE INC. 2000. SAS Institute Inc. Cary, North Carolina, USA. SEAMAN, D. E., J. J. MILLSPAUGH, B. J. KE R N O H A N, G. C. B R U N D I G E, K. J. RAEDEKE, and R. A. GITZEN. 1999. Ef- fects of sample size on kernel home range estimates. Journal of Wildlife Management 63: 739 – 747. _____, and R.A. POWELL. 1996. An evalu- ation of the accuracy of kernel density estimators for home range analysis. Ecology 77: 2075 – 2085. SEVERINGHAUS, C. W., and E. L. CHEATUM. 1956. Life and times of the white-tailed deer. Pages 57 – 186 in W.P. Taylor, editor. The Deer of North America. The Stackpole Company. Harrisburg, Pennsylvania, USA. TAYLOR, K. P., and W. B. BALLARD. 1979. Moose movement and habitat use along the Susitna River near Devil’s Canyon. Proceedings of the North American Moose Conference and Workshop 15: 169 – 186. TELFER, E. S. 1967. Comparison of a deer yard and a moose yard in Nova Scotia. ALCES VOL. 40, 2004 COBB ET AL. - SYMPATRIC MOOSE AND DEER 191 Canadian Journal of Zoology 45: 485 – 490. _____. 1970. Winter habitat selection by moose and white-tailed deer. Journal of Wildlife Management 34: 553 – 559. _____, and J. P. KELSALL. 1979. Studies of morphological parameters affecting ungulate locomotion in snow. Canadian Journal of Zoology 57: 2153-2159. THOMPSON, I. D., and D. L. EULER. 1987. Moose habitat in Ontario: a decade of change in perception. Swedish Wildlife Research Supplement 1: 181 – 193. TOWNSEND, M. T., and M. W. SMITH. 1933. T h e w h i t e - t a i l e d d e e r o f t h e Adirondacks. Roosevelt Wildlife Bulle- tin 6: 161 – 325. (USGS) U.S. GEOLOGICAL SURVEY. 2001. Vegetation Spatial Database Coverage for Voyageurs National Park. Vegeta- tion Mapping Project, Upper Midwest Environmental Sciences Center, La Crosse, Wisconsin, USA. VAN DEELEN, T. R., H. CAMPA, M. HAMADY, and J. B. HAUFLER. 1998. Migration and seasonal range dynamics of deer using adjacent deeryards in northern Michigan. Journal of Wildlife Manage- ment 62: 205 – 213. WETZEL, J. F., J. R. WAMBAUGH, and J. M. PEEK. 1975. Appraisal of white-tailed deer winter habitats in northeastern Min- nesota. Journal of Wildlife Manage- ment 39: 59 – 66. WHITLAW, H. A., and M. W. LANKESTER. 1994. The co-occurrence of moose, white-tailed deer, and Parelapho- strongylus tenuis in Ontario. Cana- dian Journal of Zoology 72: 819 – 825.