ALCES VOL. 45, 2009 MCLAREN et al. - HABITAT SELECTION IN GROS MORNE

HOW MOOSE SELECT FORESTED HABITAT IN GROS MORNE NATIONAL PARK,

Brian. E. McLaren1, S. Taylor2 and S. H. Luke1

1Lakehead University, Faculty of Forestry and the Forest Environment, 955 Oliver Road, Thunder Bay, ON, P7B 5E1; 2Gros Morne National Park, P.O. Box 130, Rocky Harbour, NL, Canada A0K 4N0.

ABSTRACT: Current policy does not allow moose (Alces alces) to be hunted in Na- tional Parks in Newfoundland and Labrador; combined with the extirpation of wolves (Canis lupus), this policy creates a situation where introduced moose (A. a. americana) are relatively predator-free in Gros Morne National Park. Forested areas of this park are frequently disturbed by defoliating insects resulting in extensive young conifer forest; increasingly, more areas are identified as failing to regenerate to normal tree densities or “not sufficiently restocked” (NSR). We used data from GPS-collared moose that occupy areas of the park where limited timber cutting is allowed for domestic purposes and a very detailed and current forest inventory exists; such areas are still dominated by insect and wind disturb- ance, including a large designation of NSR forest. We hoped to determine whether moose are found preferentially in disturbed forest versus other landscape patches during summer or winter, during day or night, and under certain temperature conditions. Variability in habitat availability and habitat use by moose appears to preclude forest management options directed at specific habitat types.

ALCES VOL. 45: 125-135 (2009)

Key words: Alces alces, absence of predators, Gros Morne National Park, moose, Newfoundland, overabundance, population dynamics, resource selection function.

When moose (Alces alces) face fewer functional response of human predation and predators they can reach higher densities (Pe- delayed density dependence in hunter kill have terson et al. 2003). Moose (A. a. americana) been implicated (Ferguson and Messier 1996). were first introduced to central Newfoundland, Relationships with their food source during Canada in 1878 with the release of a male and extreme population peaks were first described female from Nova Scotia (Pimlott 1953). A by Bergerud and Manuel (1968); declines in second release of 2 males and 2 females from population following such peaks identified by New Brunswick into western Newfoundland Mercer and McLaren (2002) suggest that food followed in 1904. Gray wolves (Canis lupus), availability is often the only limiting factor for their only potential predator, were extirpated moose in Newfoundland. in 1932 (Pimlott 1959). Therefore, for most Moose are not hunted in the National of their occupation of the island province, Parks in Newfoundland, creating a situation Newfoundland moose were preyed on only by where populations are predator-free in 1,805 human hunters, with black bear (Ursus ameri- km2 of western Newfoundland (Gros Morne cana) predating only calves. Consequently, National Park) and 404 km2 of eastern New- their density averages about tenfold higher than foundland (). The in other parts of their range in North America resulting extreme moose demographics in (Crête and Daigle 1999). Newfoundland both parks (McLaren et al. 2000), the fact moose experience large population fluctua- that moose are not native to Newfoundland, tions even where heavily hunted in central and the potential for their high densities to areas of the island, an effect that the limited alter natural ecosystem processes (McLaren

125 HABITAT SELECTION IN GROS MORNE - MCLAREN et al. ALCES VOL. 45, 2009 et al. 2004), add up to policy and management cutting is allowed for domestic purposes (home challenges for Parks Canada (Corbett 1995). and boat construction and fuelwood) and, In Canada, the response to hyperabundant therefore, a very detailed and current forest species in protecting the ecological integrity inventory exists. The sample size and inven- within and by means of national parks follows tory data quality are sufficient to determine a definition that very explicitly requires relat- whether moose are found preferentially in ing the reasons for hyperabundance to human disturbed forest versus other landscape patches impacts (PCA 2000). Examples of acceptable during summer or winter, during day or night, reasons to control large herbivores include and under certain temperature conditions. We artificial introductions and the loss of key explored these expected differences against a predator species like carnivores, both of which null hypothesis that selection within the home have occurred in Newfoundland. However, to range varies among individual moose to an institute any control program, Parks Canada extent that general prediction about their use also demands confirmation that the reasons of landscape patches is not possible. for hyperabundance are well understood and the program is conducted under an adaptive STUDY AREA management framework where the original Gros Morne National Park is situated on assumptions are subject to review. Moreover, the Gulf of St. Lawrence in Newfoundland, control is better justified when “ecological in- encompassing parts of the Northern Peninsula, tegrity” is compromised by the large herbivore; Long Range Barrens, and Western Newfound- according to the Canada National Parks Act, land Forest ecoregions (Damman 1983). ecological integrity means “…a condition that Moose likely invaded the area now protected is determined to be characteristic of its natural by GMNP in 1925 and became common by the region and likely to persist, including…rates 1950s, with modest population increases until of change and supporting processes.” It is in the 1970s. Surveys in the late 1970s indicated this context that we embarked on a review of that after GMNP became established (and hu- how moose exploit areas of forest disturbance man hunting was excluded), moose population in Gros Morne National Park (GMNP). density increased much more rapidly (GMNP, It has been hypothesized that the Gros unpublished data). Increases first occurred in Morne moose population expanded as its upland areas, but by the 1980s moose density range expanded into new habitat created by increased throughout the park (Connor et al. insect, wind, and timber cutting disturbances 2000). In a 1998 survey, approximately 8,000 in GMNP (Connor et al. 2000). We argue that animals occurred in <1,000 km2 of potential forecasts for forest regeneration and moose habitat (GMNP unpublished data). Some habitat in GMNP depend on how much the survey units in 2007 and 2008 had densities cause for high moose densities can be ascribed >15 moose/km2 in lowland areas, where the to forest disturbance, and how confident we average density remains ~ 4 moose/km2. are that moose generally select and occupy Gros Morne National Park includes an disturbed areas. Therefore, we matched the allowance for domestic timber cutting in 12 forest inventory and disturbance database to “cutblocks” (193 km2), but excludes 6 adjacent information from a series of collars with global community enclaves (140 km2). The cutblocks positioning system (GPS) capability placed on used by moose in this study were located in the adult female moose in 1997-1998 (McLaren et Coastal Plain sub-region of the Northern Pen- al. 2000). While the sample size is small (N insula ecoregion, east of the flat coastal areas = 4), we focus on a subset of collared moose of the Gulf and including the western slopes that occupy areas of the park where timber of the . At elevations

126 ALCES VOL. 45, 2009 MCLAREN et al. - HABITAT SELECTION IN GROS MORNE

<425 m, these areas experience a cool boreal information was delineated using data on climate with a relatively long growing season colour, 1:12,500 scale aerial photographs of 110-150 days (Damman 1983). Forests taken in 1995. This information included consist mostly of balsam fir (Abies balsamea), age, height, and crown density estimates, and with some spruce (Picea spp.) and a mix of the approximate species composition of each pioneer (mostly Betula spp.) and tolerant forested stand >0.5 ha; the disturbance type (mostly Acer spp. and Sorbus spp.) hardwoods; and (if known) were tagged to disturbed a more detailed description is found in Connor forest stands identified on these photographs. et al. (2000). Insect outbreaks, primarily of If disturbed forest was interpreted as failing spruce budworm (Choristoneura fumiferana) to regenerate to normal tree densities at the in 1977 and hemlock looper (Lambdina fiscel- time of interpretation, based on absence of laria) in 1969, 1988 and 1996 affected a large crown closure, stands in this category were area of the forest: 7,550 ha in total (2,800 tagged “not sufficiently restocked” (NSR). ha in the cutblocks) with individual areas In addition, non-forest vegetation types such of canopy disturbance up to 49 ha in extent as barren, bog, scrub, residential areas and (GMNP unpublished data). A study just east water were mapped. of GMNP on the northern peninsula concluded In 2004, black-and-white, 1:10,000 scale that insects caused defoliation and death of aerial photographs were acquired to update the forest canopy in >60% of the landscape, and classify more accurately the forest dis- plus an additional canopy break-up and gap turbances limited to the cutblocks. Forest regeneration in >10% of the forest (McCarthy stands that were labelled in the 1995 inven- and Weetman 2007). In contrast, domestic tory as disturbed, NSR, or regenerating forest timber cutting entails manual tree removal were re-evaluated to determine their status. If from small patches (maximum 2.2 ha); to an regeneration had partly or completely failed, extent within the higher areas of this range, they were labelled (or re-labelled) NSR; stands ‘high-grading’ may occur where large, domi- with sufficient regeneration were labelled with nant trees are preferentially cut leaving smaller the appropriate regenerating forest label from trees in place. At the local level then, cutting, the original age of disturbance. Approximately wind disturbance, and insect outbreaks create half of these stands were labelled in the “0-20 very similar forest structures. The disturbance year” age category and half in the “20-40 year” types also frequently occur in combination; category. New disturbances, such as recent for example, cutting has occurred in 340 ha of cutting or insect outbreaks (~ 1996-2004) were defoliated forest following insect outbreaks in also delineated in 2004. the cutblocks, and windfall is frequently as- sociated with defoliation. Timber cutting, on Moose Locations the other hand, is a relatively minor contribu- We used location data collected from 4 tor to new disturbances in GMNP, amounting adult, female moose collared and monitored to approximately 20 ha per year, equivalent 26 June 1997-13 October 1998. Collars were to just over 1% of the total forested areas set to record differentially correctable GPS within the cutblocks since records began in locations every third hour (Lotek Engineering the mid-1990s. Inc., Newmarket, Ontario, Canada). Details of the collaring and of accuracy testing of METHODS the location data are found in McLaren et Forest Inventory al. (2000). Only locations with 3D accuracy A new, detailed forest inventory for and differential correction were used in this GMNP was completed in 1997. Forest stand study; these locations had a 95% accuracy

127 HABITAT SELECTION IN GROS MORNE - MCLAREN et al. ALCES VOL. 45, 2009 of <25 m. using 100% minimum convex polygons. Size A night location dataset was created ap- of seasonal ranges was compared across all 4 proximately equal in size to a daytime loca- moose by t-test. Areas outside the cutblocks tion dataset for each moose by first choosing and large areas of open water were then ex- from the 3D locations recorded closest to cluded from each of the seasonal ranges. The 0300 and 1500 hr each monitoring day. The remaining area was divided into a) young NSR night location datasets were supplemented in forest ≤ 20 since disturbance, b) older 13-21% of cases with locations recorded at NSR forest, c) young (regenerating) conifer approximately 0000 hr when no 3D location forest ≤ 20 years since disturbance, d) older occurred at 0300 hr, and the daytime datasets conifer forest, e) mixed forest and deciduous were supplemented similarly in 18-25% of forest of all ages, f) scrub forest, and g) barren cases with locations recorded at approximately (non-forested) areas. The first 2 categories 1200 hr. The resulting datasets covered 91- were defined according to information in both 97% of possible night locations and 96-99% the original forest inventory (1995) and the of possible day locations during a 351-402 day update (2004); disturbances that were labelled monitoring period, depending on the moose. NSR in either database (or both databases) The monitoring periods were further were re-categorized as “young” or “older” divided into summer and winter seasons fol- NSR forest. Recognizing that the moose lo- lowing a method developed by Vander Wal cations were recorded during 1997-1998, we (2005). For each moose, cumulative distance classed “young” NSR based either on a date moved was calculated in ArcView version 9 in the inventory indicating ≤20 years since (ESRI, Redlands, California) beginning with disturbance by 1997, or on a forest age class the first 3D location on 1 January 1998, and in the 1995 inventory of “0-20 years.” Areas ending with either 26 June 1998 (2 cases) or of NSR first identified in the 2004 update 13 October 1998 (2 cases). The time series were counted as NSR habitat only if they were were completed with data beginning on 26 classed as “0-20 years” of age (or older) at that June 1997 or 13 October 1997 and ending 31 time, because most insect outbreaks occurred December 1997. Winter was defined as the before 1997; new areas of NSR in the update period when slope of cumulative distance over were not counted as NSR areas in the habitat cumulative time was less than the annual mean analysis if they were assigned to a date after slope; summer was when movement within 1998. The reason 2 of the habitat categories the home range was faster. Each time series refer to “young” forest is that forage for moose was closest matched to a logistic curve, using in Newfoundland has long been estimated regression with statistical software (SPSS, to be highest in regenerating balsam fir for- version 16). Summer fell between 8-28 est at 8-10 years of age (Parker and Morton April and 17 September-18 October, ranging 1978), where the amount of browsing by from 158-186 days in length, depending on moose increases with the fraction of balsam the moose. For 3 of the 4 moose, 2 nearly fir among trees <3 m in height (Thompson complete summer seasons (1997 and 1998) 1988). Thus, “young” forest is a special and could be quantified in the GPS database. The desired habitat for foraging by moose, while fourth moose was sufficiently monitored only older forest may serve largely as cover, but during the 1998 summer season. not foraging habitat. Habitat selection was tested for all 7 of Data Analysis the categories in 4 separate resource selection Summer (1997 and 1998) and winter functions (RSF), one for each moose. RSFs (1997-1998) home ranges were calculated are statistical tools that describe the relative

128 ALCES VOL. 45, 2009 MCLAREN et al. - HABITAT SELECTION IN GROS MORNE probability of occurrence of animals based on (>8º C) collar temperatures were recorded their response to their habitat (Boyce 2006). (limited by data availability). The temperature In this study, each observed moose location thresholds were chosen as close as possible was linked to 10 random locations within a to the temperatures at which heat stress may 700-m radius circle, an area that encompassed begin for moose, −5º C in winter and 15º C in approximately 50% of the distances between 2 summer (Renecker and Hudson 1986), while successive locations during any season for all also dividing the 3D locations approximately moose. Thus, the observed moose locations evenly among cold, warm, and cool catego- were considered “selected” among random ries. Young NSR forest was considered the areas within 700 m, a distance chosen to as- best choice for a standard against which to sume that moose selected the habitat class compare habitat selection in the other 6 cat- recorded by the GPS collar as their location egories, both because it serves as a theoretical among random locations to which they would target for foraging by moose, and because it have been capable of moving, but were not provides the highest actual likelihood among found. In other words, locations are tracked the other young forest categories that it was relative to the patches of habitat immediately visited by moose due to its NSR designation available to a moose at any given time. and the likelihood that regeneration failure is The RSF typically distinguishes among linked to moose overabundance. observed locations and available habitats using log-linear modelling, for this study constructed RESULTS in SPSS (version 16). Comparisons among Seasonal range sizes varied from 135- resulting RSF models were made using a 1,692 ha (Table 1), and were larger (P = 0.002) combination of the significance of change in in winter (1,200 ± 257 ha, N = 4) than in sum- deviance in an analysis of deviance table (Man- mer (419 ± 101 ha, N = 7). Range composition ley et al. 2002) and the Akaike’s Information also varied considerably among individual Criterion (AIC, Burnham and Anderson 2002). moose (Table 1); 3-86% of a seasonal range Relative habitat selection was calculated, set- occurred outside a cutblock. In 3 of 4 cases, ting the probability of selection of young NSR the proportion of range outside cutblocks was forest at 1.000. Reporting relative rather than greater in winter than in summer. For Moose absolute habitat selection probabilities follows 15 and 16, old and young NSR forest made up the recommendation of Arthur et al. (1996), more than one-third of all range area within for situations when availability of habitats is the cutblocks, whereas scrub forest filled a not constant and comparisons among habitats dominant proportion of every range for Moose may be affected by the choice of which habi- 22 and 25. By contrast, non-forested (barren) tats to include. areas comprised anywhere from 1-33%, mixed Variation in habitat selection among indi- and deciduous forest 4-33%, and older conifer vidual moose was compared to 1) variation in forest 2-32% of ranges. habitat selection between summer and winter Habitat selection was significant and var- and between day and night in each season, 2) ied significantly for all tested factors (Table variation in habitat selection when the collar 2). Variation in habitat selection among indi- recorded cold (≤0º C) versus warm (>0º C) vidual moose was significant and larger than winter temperatures, 3) variation in habitat variation in habitat selection by season, by selection when the collar recorded cool (≤8º temperature, by time of day, or by year for C) versus warm (>8º C) summer temperatures, the summer season. and 4) variation in habitat selection between summers (1997 and 1998) only when warm

129 HABITAT SELECTION IN GROS MORNE - MCLAREN et al. ALCES VOL. 45, 2009 - collars set to record Global Positioning System (GPS) locations every third hour; records began on 26 June 1997 and were censored at the end of the second summer, 26 September 1998. Range size and composition were calculated from Minimum Convex Polygons (MCPs) encompassing 100% of 3D differentially corrected loca tions. Sample sizes in subsequent analyses are shown as the number of locations, divided into categories of day and night, and cool- or cold- versus warm-temperature dashed line indicates a censored category in comparative analyses due to low sample size. A locations. able 1. Characteristics of seasonal ranges for adult female moose captured (Moose 15, 16, 22 and 25) in Gros Morne National Park (GMNP). Monitoring occurred from occurred Monitoring (GMNP). Park National Morne Gros in 25) and 22 16, 15, (Moose captured moose female adult for ranges seasonal of Characteristics 1. able T

130 ALCES VOL. 45, 2009 MCLAREN et al. - HABITAT SELECTION IN GROS MORNE constructed from one log-linear model for each moose from n locations and 10n additional random locations. Model significance was calculated by comparison of the difference in deviance (∆ D) between successive models with the χ2 distribution for df degrees of freedom. Akaike's Information Criteria (AICs) are shown for each model step. able 2. Analysis of deviance table for models of habitat use. N is the number of collared, adult female moose using cutblocks. Resource selection functions (RSFs) were (RSFs) functions selection Resource cutblocks. using moose female adult collared, of number the is N use. habitat of models for table deviance of Analysis 2. able T

131 HABITAT SELECTION IN GROS MORNE - MCLAREN et al. ALCES VOL. 45, 2009

Young NSR forest was the most selected variation in selection among individual moose among all habitat types, although less selected (Table 2). For 3 moose (16, 22, and 25), mixed by Moose 22 (Table 3). Among winter, day- and deciduous forest was either the most se- time locations for Moose 15, young NSR lected habitat type or the second most selected forest was selected more than any other throughout winter, next to young NSF forest; habitat type. The same pattern, although not however, Moose 15 selected significantly less significant, held at all times except on summer mixed and deciduous forest and significantly nights when scrub forest was slightly, but not more older coniferous forest in colder periods significantly, more selected than young NSR during winter, relative to young NSF forest forest. For Moose 25 the same pattern held (Table 4). On the other hand, coniferous for- for winter day-time locations, with young est was significantly less selected in colder NSR forest selected more often than 4 of the periods by Moose 22 and Moose 25, relative 6 other habitat types. In contrast, Moose 22 to young NSR forest. During warmer periods selected only barren areas significantly less in winter, selection among moose was more than young NSR forest, while mixed forest constant when conifer forest was less selected and scrub forest were generally more selected than young NSR forest. There was a smaller, than young NSR forest. Moose 25 had the but significant difference in summer habitat largest shift in selection between winter and selection during cool versus warm periods for summer; both older conifer forest and scrub the 3 moose with sufficient records for this forest were less selected than young NSR forest comparison. As in winter, Moose 15 selected on winter days, but more selected on summer significantly less mixed and deciduous forest days. In contrast to young NSR forest, older than young NSR forest during the cooler pe- NSR forest did not appear to be selected. riod; the same pattern held for Moose 25 for There was a significant difference in win- the summer period. Moose 25 selected scrub ter habitat selection during cold versus warm forest significantly more than NSR forest on periods for all moose. Conclusions about these warm summer days, but significantly less scrub differences were difficult, due to the higher forest was selected on cool summer days.

Table 3. Relative selection probabilities when comparing habitat use by season and time of day. Rela- tive selection was assessed using 1.000 as the probability of selection for young (≤20 year-old) NSR forest. Probabilities were derived from parameters in RSFs constructed for individual moose for which habitat selection varied significantly (Table 2a). Resource selection varied for the other fac- tors, season and time of day, for moose 15 and 22, but not for moose 25. For these conclusions, ∆ D is compared to the χ2 distribution for 6 degrees of freedom. Asterisks indicate cases where prob- abilities are derived from significant parameter estimates (P <0.05).

132 ALCES VOL. 45, 2009 MCLAREN et al. - HABITAT SELECTION IN GROS MORNE

Table 4. Relative selection probabilities comparing habitat use on cold and warm days in winter (∆ D1), cool and warm days in summer (∆ D2), and for warm summer days, 2 monitoring seasons in 1997 and 1998 (∆ D3). Other definitions and calculations are as inTable 3.

The difference in summer habitat selection from conclusions made from VHF-collared between 1997 and 1998 was significant for moose in the same area (McLaren et al. 2000). only 1 (Moose 16) of the 3 moose with a long For example, migrations did not occur from enough monitoring record to allow compari- cutblocks during the early months of summer son. In this case, the least selected habitat in among the 4 moose reported in this study. Even 1997 was mixed and deciduous forest, while though local movement rates increase in sum- in 1998 it was either older conifer forest or mer months, moose residing in the cutblocks barren areas. Except for scrub forest on warm may be less prone to long-distance migrations summer days, young NSR forest was the most if food can be found in young forest, even in selected habitat type throughout the summer in young NSR forest. If this condition describes both years, while older NSR forest was among the current situation, it likely also explains the least selected habitat types. smaller summer home ranges, while ranges sufficiently larger to include older coniferous DISCUSSION forest in colder periods, for example, were For GMNP we conclude that selection likely required during winter. The observation within the home range varies among individ- of higher daily movement rates in summer ual moose to an extent that general prediction versus winter, the only uniform pattern among about use of landscape patches by even 4 the moose in this study, is consistent with moose, let alone the population, is not pos- Vander Wal’s (2005) conclusion that winter sible. This conclusion is not different from is a period of energy conservation and limited that made recently by Gillingham and Parker daily movement. (2008) for northern British Columbia, and it Variation in snow conditions in the Coastal may be a general caution about the interpreta- Plain subregion of Newfoundland is probably tion of RSFs, as well as guidance for managers the source for variation in habitat selection by in the management of moose moose in winter. Except for warmer summer We also garnered additional information days, young NSR forest appears generally the from the GPS-collared moose that differed most selected habitat type in the cutblocks.

133 HABITAT SELECTION IN GROS MORNE - MCLAREN et al. ALCES VOL. 45, 2009

However, overall high variability in habitat in the cutblocks and is responsible for the selection during winter and summer suggests detailed aerial photo updates that allowed this that conclusions drawn to assist moose or study to be undertaken. The original fieldwork moose habitat management in GMNP will not involving GPS collars was coordinated under be general ones. One possibile generalization the direction of Shane Mahoney (Government is that in 20 years, when young NSR forest of Newfoundland and Labrador), and Douglas converts to older NSR forest, foraging op- Anions and Christopher McCarthy (Parks portunities for moose may be depleted. Even Canada), and with the assistance of pilots with modestly high densities, moose habitat Baxter Slade (Newfoundland Helicopters) and in Newfoundland demonstrably declines over Hughie Day (Canadian Helicopters). a few decades (Mercer and McLaren 2002). However, this study is unable to show that REFERENCES moose are food limited in GMNP. Variability Ar t h u r , S. M., B. F. J. Ma n l y , L.L. McDo n - in habitat availability and habitat use by moose a l d , and G. W. Ga r n e r . 1996. Assess- appears to preclude forest management options ing habitat selection when availability directed at specific habitat types. changes. Ecology 77: 215-227. Concern with high-density moose popula- Be r g e r u d , A. T., and F. Ma n u e l . 1968. Moose tions in national parks arises from the possibil- damage to balsam fir-white birch forests in ity that forests will not regenerate naturally. central Newfoundland. Journal of Wildlife A study in Terra Nova National Park detailed Management 32: 729-746. that even after moose are removed from the Bo y c e , M. S. 2006. Scale for resource selec- ecosystem, alteration of forest dynamics by tion functions. Diversity & Distributions hyperabundant moose can persist at least 2 12: 269-276. decades (McLaren et al. 2009). Only circum- Bu r n h am , K. P., and D. R. An d e r s o n . 2002. stantial evidence of altered forest composition Model Selection and Multi-Model Infer- in GMNP was found by Connor et al. (2000). ence: A Practical Information-Theoretic However, in extreme cases in GMNP, persistent Approach. Springer-Verlag, New York, invasive plants, including coltsfoot (Tussilago New York, USA. farfara) and Canada thistle (Cirsium arvense), Co n n o r , K. J., W. B. Ba l l a r d , T. Di l w o r t h , S. as well as native grasses (e.g., Calamagrostis Ma h o n e y , and D. An i o n s . 2000. Changes canadensis), become widespread problems as in structure of a boreal forest community a result of moose browsing and trampling in following intense herbivory by moose. disturbed areas (Rose and Hermanutz 2004). Alces 36: 111-132. We recommend continued monitoring in Co r b e t t , G. N. 1995. Review of the history GMNP to evaluate impacts by moose on for- and present status of moose in the national ested habitats relative to forest regeneration parks of the Atlantic region: management and composition. implications? Alces 31: 255-268. Cr ê t e , M., and C. Da i g l e . 1999. Management ACKNOWLEDGEMENTS of indigenous North American deer at the We are grateful to the Inland Fish and end of the 20th century in relation to large Wildlife Division of the Government of predators and primary production. Acta Newfoundland and Labrador for contributing Veterinaria Hungarica 47: 1-16. to the effort to resolve questions of moose Damma n , A. W. H. 1983. An ecological overabundance in GMNP. We are also very subdivision of the island of Newfound- grateful to Carson Wentzell, GMNP Resource land. Pages 163-205 in G. R. South (ed.) Conservation, who leads the forest inventory Biogeography and Ecology of the Island

134 ALCES VOL. 45, 2009 MCLAREN et al. - HABITAT SELECTION IN GROS MORNE

of Newfoundland. W. Junk Publishers, Pa r k e r , G. R., and L. D. Mo r t o n . 1978. The Boston, MA, USA. estimation of winter forage and its use by Fe r g u s o n , S. H., and F. Me ss i e r . 1996. Can moose on clearcuts in northcentral New- human predation of moose cause popula- foundland. Journal of Range Management tion cycles? Alces 32: 149-161. 31: 300-304. Gi l l i n g h am , M. P., and K. L. Pa r k e r . 2008. PCA (Pa r ks C a n ada A g e n c y ). 2000. Manag- The importance of individual variation ing hyperabundant species. Pages 5-12 and in defining habitat selection by moose 5-13 in Panel of the Ecological Integrity in northern British Columbia. Alces 44: of Canada’s National Parks. Unimpaired 7-20. for Future Generations? Protecting Eco- Ma n l y , B. F. J., L. L. McDo n a l d , D. L. Th o m - logical Integrity with Canada’s National as , T. L. McDo n a l d , and W. P. Er i c ks o n . Parks. Volume II. Setting a New Direc- 2002. Resource Selection by Animals, tion for Canada’s National Parks. Ottawa, second edition. Kluwer Academic Pub- Ontario, Canada. lishers, Dordrecht, The Netherlands. Pe t e r s o n , R. O., J. A. Vu c e t i c h , R. E. Pag e , McCa r t h y , J. W., and G. We e t ma n . 2007. and A. Ch o u i n a r d . 2003. Temporal and Stand structure and development of an spatial aspects of predator-prey dynamics. insect-mediated boreal forest landscape. Alces 39: 215-232. Forest Ecological Management 241: Pi m l o t t , D. H. 1959. Reproduction and 101-114. productivity of Newfoundland moose. McLa r e n , B., L., He r ma n u t z , J. Go ss e , B. Journal of Wildlife Management 23: Co l l e t , and C. Kas i m o s . 2009. Broad- 381-401. leaf competition interferes with balsam ______. 1953. Newfoundland moose. fir regeneration following experimental Transactions North American Wildlife removal of moose. Forest Ecological Conference 18: 563-581. Management 257: 1395-1404. Re n e c k e r , L. A., and R. J. Hu ds o n . 1986. ______, B. A. Ro b e r t s , N. Dja n -Ch é ka r , and Seasonal energy expenditure and thermo- K. P. Le w i s . 2004. Effects of overabundant regulatory response of moose. Canadian moose on the Newfoundland landscape. Journal of Zoology 64: 322-327. Alces 40: 45-59. Ro s e , M., and L. He r ma n u t z . 2004. Are ______, C. McCa r t h y , and S. P. Ma h o n e y . boreal ecosystems susceptible to alien 2000. Extreme moose migrations in Gros plant invasion? Evidence from protected Morne National Park, Newfoundland. areas. Oecologia 139: 467-4776. Alces 36: 217-232. Th o mps o n , I. D. 1988. Moose damage to pre- Me r c e r , W. E., and B. E. McLa r e n . 2002. commercially thinned balsam fir stands in Evidence of carrying capacity effects in Newfoundland. Alces 24: 56-61. Newfoundland moose. Alces 38: 123- Va n d e r Wa l , E. J. 2005. Core areas of habi- 141. tat use: the influence of spatial scale of analysis on interpreting summer habitat selection by moose (Alces alces). M. Sc. Thesis, Lakehead University, Thunder Bay, Ontario, Canada.

135