Migration Behavior of White-Tailed Deer under Varying ...people.trentu.ca/~jebowman/SabineetalJWM.pdf · MIGRATION BEHAVIOR OF WHITE-TAILED DEER UNDER VARYING WINTER CLIMATE REGIMES

MIGRATION BEHAVIOR OF WHITE-TAILED DEER UNDER VARYING WINTER CLIMATE REGIMES IN NEW BRUNSWICK DWAYNE L. SABINE,1 New Brunswick Cooperative Fish and Wildlife Research Unit, University of New Brunswick, P.O. Box

44555, Fredericton, NB E3B 6C2, Canada SHAWN F. MORRISON, New Brunswick Cooperative Fish and Wildlife Research Unit, University of New Brunswick, P.O. Box

44555, Fredericton, NB E3B 6C2, Canada HEATHER A. WHITLAW,2 New Brunswick Cooperative Fish and Wildlife Research Unit, University of New Brunswick, P.O. Box

44555, Fredericton, NB E3B 6C2, Canada WARREN B. BALLARD,34 New Brunswick Cooperative Fish and Wildlife Research Unit, University of New Brunswick, P.O. Box

44555, Fredericton, NB E3B 6C2, Canada GRAHAM J. FORBES, New Brunswick Cooperative Fish and Wildlife Research Unit, University of New Brunswick, P.O. Box

44555, Fredericton, NB E3B 6C2, Canada JEFF BOWMAN,5 New Brunswick Cooperative Fish and Wildlife Research Unit, University of New Brunswick, P.O. Box 44555,

Fredericton, NB E3B 6C2, Canada

Abstract: White-tailed deer (Odocoileus virginianus) exhibit a variety of migration strategies across northern portions of their range. Factors reported as being responsible for migration initiation have shown no consistent pattern. We monitored 186 radiocollared white-tailed deer from 1994 to 1998 in 2 areas of New Brunswick: a southern area with moderate and variable winter climate and a northern area with consistently severe winter climate. We determined that deer in the south contained a large proportion of conditional migrators (individuals that may or may not migrate to winter range in a given year, and may or may not remain until spring), whereas deer in the north con- sisted almost entirely of obligate migrators (those that annually migrate to winter range for the duration of winter). Occurrence of conditional migration appeared to be a function of climate variability, although distribution of the behavior among individual deer was influenced by migration distance. Initiation of autumn migration in the south was related to snow depth for most deer and represented a response to the proximate cue of the onset of

limiting conditions. Autumn migration in the north appeared to be a response to seasonal cues, and the direct influence of snow depth was reduced. Initiation of spring migration in the 2 study areas showed a similar pattern. Migration distance may represent a factor influencing distribution of migrational cues among individual deer within a population. The effect of winter climate variability on deer migration behavior may account for the disparity in behavior reported in the literature. The differences in migration behavior have implications for deer management surveys in northern areas where deer yarding occurs. Managers have assumed that deer observed during winter surveys were on winter range, but this may not be a reasonable assumption in areas with variable winter climates.

JOURNAL OF WILDLIFE MANAGEMENT 66(3):718-728

Key words: deeryard, migration, movement, New Brunswick, Odocoileus virginianus, white-tailed deer, winter.

The annual migration of white-tailed deer between summer and winter ranges is a well-known

phenomenon throughout northern portions of their range (Severinghaus and Cheatum 1956). This migration generally is thought to be an adap- tive response to the presence of snow cover, since

1 Present address: New Brunswick Department of Natural Resources and Energy, Fish and Wildlife Branch, P.O. Box 6000, Fredericton, NB E3B 5H1, Canada.

2 Present address: Department of Range, Wildlife, and Fisheries Management, Texas Tech University, P.O. Box 42125, Lubbock, TX 79409, USA.

3 Present address: Department of Range, Wildlife, and Fisheries Management, Texas Tech University, P.O. Box 42125, Lubbock, TX 79409, USA.

4 E-mail: [email protected] 5 Ontario Ministry of Natural Resources, 300 Water

Street, Peterborough, ON K9J 8M5, Canada.

winter ranges typically are in areas that ameliorate winter climate conditions (e.g., Verme 1973, Huot

1974, Drolet 1976). Furthermore, deer tend to con-

gregate on winter range, forming communal trails that further lessen the impact of snow cover on

mobility. In northern areas subject to cold winter

temperatures but low snowfall, white-tailed deer

may respond to winter conditions by temporarily shifting habitat use within ranges, but seldom

migrate (Sparrowe and Springer 1970, Kucera

1976, Beier and McCullough 1990). Likewise, white-tailed deer in southern portions of their

range do not exhibit seasonal migration (Marchin- ton and Hirth 1984, Gaudette and Stauffer 1988).

Although the presence of snow in winter as the ultimate cause of seasonal migration behavior in northern deer generally is accepted, the factors associated with initiation of migration to and from winter ranges are complex. Reported pri-

718

MIGRATION BEHAVIOR OF DEER * Sabine et al. 719

mary stimuli for migration initiation include fluc- tuation in snow depth (Westover 1972, Drolet 1976, Tierson et al. 1985), changing temperatures (Verme 1973, Kearney and Gilbert 1976, Swanson 1993), and complex sets of stimuli, possibly including photoperiod and vegetation changes with snow depth and temperature (Kucera 1992, Nelson 1995, Nicholson et al. 1997). The issue

may be further complicated by the tendency of some deer to migrate to staging areas prior to final entry into a yard (e.g., Swanson 1993). Stim- uli directing initial migrations to staging areas and final movements into yards may differ.

Migration behavior can differ among individual deer within a population. Migratory and nonmi-

gratory deer can co-occur (e.g., Nelson and Mech 1991, Brown 1992, Van Deelen et al. 1998). Behav- ior may differ within groups of migratory deer. Deer normally migrate once, for the duration of winter (i.e., obligate migration). However, some

migratory deer migrate inconsistently or temporarily (i.e., facultative migration or conditional migration; Nelson 1995, Nicholson et al. 1997). Extreme

variability in yearly snow depth has been suggested as the primary cause for mixed-migration respons- es within populations (Nicholson et al. 1997).

Management of deer wintering areas often is a central facet of deer population management in northern areas. Clarification of the factors influ-

encing migration to winter ranges would allow for better management of these migratory populations. Although various proximate causes of

migration behavior by white-tailed deer in northern regions have been reported, a central theme

may tie these differing factors together. We studied the seasonal migration patterns of deer win-

tering under different prevailing climate regimes in New Brunswick, Canada: a continental climate

experiencing severe and consistent winter conditions in the north, and a maritime climate experiencing moderate and variable winter conditions in the south. Our objectives were to determine (1) whether deer in these 2 areas differed with respect to migration strategies, and (2) whether factors influencing migration initiation differed between the 2 areas.

STUDY AREA

Our study areas were located in southern (Canaan area: 45?55'N, 65?40'W) and northern (Odell Deeryard: 46?45'N, 67?30'W) New Brunswick (Fig. 1). The Canaan study site was

comprised of 3 separate deer wintering areas mea-

suring 5,658, 712, and 2,307 ha, arranged on an

Fig. 1. Location of Canaan and Odell, New Brunswick, Cana- da, study areas in relation to the Bay of Fundy.

east-west axis spanning approximately 28 km

(Sabine 1999). Land ownership primarily was small private holdings, resulting in a pattern of small (<20 ha) forest openings and partial cuts often near agricultural lands. Forest cover of the

wintering areas primarily was mixed stands of

trembling aspen (Populus tremuloides), largetooth aspen (Populus grandidentata), red maple (Acer rubrum), white birch (Betula papyrifera), red spruce (Picea rubens), and white spruce (Picea glauca), with scattered stands of cedar and hemlock. Topogra- phy ranged from flat to strongly ridged. Areas used as summer range by deer radiocollared at Canaan were located directly north of their respec- tive wintering areas in an area dominated by large, industrial forest holdings. Forest cover of the summer ranges differed from that of wintering areas

by absence of agriculture lands and stands of cedar and hemlock, and by presence of large (20-100 ha) clearcuts and young plantations of black

spruce (Picea mariana) and jack pine (Pinus banksiana). Topography was flat to gently rolling.

The Odell Deeryard (1,690 ha) was situated 175 km northwest of Canaan. Forest cover at Odell

primarily was composed of white spruce, eastern hemlock (Tsuga canadensis), eastern white-cedar

(Thuja occidentalis), and balsam fir (Abies bal- samea), with interspersed small openings (<3 ha) and stands of balsam poplar (Populus balsamifera) and white birch (Whitlaw et al. 1998). Ownership mainly was large, industrial forest holdings. Topography of the area was flat to rolling. The area used as summer range by radiocollared

J. Wildl. Manage. 66(3):2002

720 MIGRATION BEHAVIOR OF DEER * Sabine et al.

Table 1. Pearson correlation values between Julian date, mean daily temperature, and snow depth in New Brunswick, Cana- da, 1994-1998.

Mean daily Snow Season temperature depth n

Autumn Canaan -0.547 0.484 276 Odell -0.635 0.616 460

Spring Canaan 0.717 -0.612 235 Odell 0.790 -0.852 446

Odell deer was situated immediately south and southeast of the yard. Topography and ownership patterns were similar to those in the yard. Forest cover on summer range differed from that on the

yard by a greater proportion of hardwood stands and by occurrences of large (20-100 ha) clearcuts and young plantations of white spruce, black

spruce, and jack pine. The climate of southern New Brunswick is

strongly influenced by close proximity of the Bay of Fundy, resulting in a distinct maritime influence that moderates climatic extremes. Winter storms often bring rain as well as snow, resulting in a pattern of increasing and decreasing snow depths throughout winter (Phillips 1990). Annu- al mean yarding period was 50 days in the Canaan area (New Brunswick Department of Natural Re- sources and Energy 1991). In contrast, northern New Brunswick experiences a continental climate with a persistent period of snow cover in winter (Phillips 1990). The annual mean yarding period, defined as number of days with open area snow depths >50 cm, was 90 days in the Odell area (New Brunswick Department of Natural Re- sources and Energy 1991). Snow and temperature variables were strongly correlated withJulian date at Odell during both autumn and spring sea- sons (Table 1).

METHODS

Daily maximum and minimum air temperatures were recorded at the Environment Canada climate station at the J. D. Irving Limited, Sussex Tree Nursery climate station at Sussex, New Brunswick, 12 km southeast of Canaan, and at Aroostook, New Brunswick, 21 km west of Odell. Daily temperatures used in analyses were calculated as the median between daily maximum and minimum temperatures. Open-area snow depths were recorded as mean depth from a straight line series of 6 stakes situated 10 m apart in a clearcut at Canaan

(Sabine 1999) and as single measurements from the Aroostook station for Odell. Sinking depths were not measured because they do not correlate closely to deer behavior (Pauley et al. 1993) and are imprecise, varying considerably both within and among days (Potvin and Huot 1983).

Pearson correlation coefficients were calculated forJulian date, mean daily temperature, and snow

depth during mid-autumn-early winter (1 Nov-31

Jan) and late winter-mid-spring (15 Feb-15 May) periods. Correlation coefficients for the 2 study areas and for the 2 periods were compared as described by Zar (1984).

We captured and radiocollared (Lotek Engi- neering, Newmarket, Ontario, Canada) deer between 1994 and 1996. Deer were captured by several methods, including Clover traps (Clover 1956), rocket nets, darting using ketamine hydrochloride and xylazine hydrochloride (Jes- sup et al. 1983, Mech et al. 1985) or Telazol? (tiletamine hydrochloride and zolazepam hydrochloride) and xylazine hydrochloride, and by helicopter net-gunning (Helicopter Wildlife Management, Salt Lake City, Utah, USA). Most deer were captured on winter ranges at Canaan. All deer were captured on winter ranges at Odell. Deer were aged by examining tooth wear and replacement, and classified by sex. Juveniles were defined as deer 8-11 months old at capture. Adults were defined as those deer captured at 218 months of age. Additional details on methods are described by Ballard et al. (1998).

Deer locations were estimated by triangulation with >3 azimuths (White and Garrott 1990). Azimuths were measured from known locations with either a hand-held, 2-element H antenna or a 3-element Yagi antenna. Fixed-wing aircraft were used to relocate deer whose signals could not be obtained from the ground. Location estimates were generated in either New Brunswick Stereographic Grid (Canaan) or UTM (Odell) format using the software package LOCATE II (Pacer, Truro, Nova Scotia, Canada). Locations were determined daily when logistically feasible during winter (15 Dec-15 Apr) and opportunisti- cally the remainder of the year.

We classified deer as migratory if they engaged in seasonal movements between non-overlapping summer and winter home ranges. Obligate migrators were those deer that migrated to winter range in early winter and remained until spring. Deer were considered conditional migrators if they failed to migrate during any 1 winter, or if they migrated to winter range only briefly



and returned to summer range following midwinter thaws (Nelson 1995). Delayed or reluctant

migrators, or those deer that migrated later than other deer and long after snow accumulations reached limiting depths, were also classified as conditional migrators. Migration from summer to winter range was classified as autumn migration, regardless of calendar date. Likewise, migration back to summer range was classified as spring migration regardless of date. Migration dates were calculated as the median between date of last observation on a seasonal range a deer

migrated from, and either first date of observation on the seasonal range it migrated to or the first date it could not be located on the seasonal

range from which it migrated. Difference in pro- portional migration behavior between study sites was assessed with a chi-square test.

The software package CALHOME was used to

generate 95% minimum convex polygon home

range estimates for individual deer. Summer and winter centers of activity (COA) for individual deer were calculated as mean easting and nor-

thing based on the New Brunswick Stereograph- ic Grid or UTM coordinates of locations on summer and winter home ranges (Hayne 1949). Migration distance was calculated as straight-line distance between these COA on winter and summer ranges. Differences in migration distances between or among groups of deer were tested with t-tests or single-factor analyses of variance followed by Tukey tests. Distance distributions were positively skewed, so distance data were

square-root-transformed prior to analysis. Sex and age have been shown to have little impact on

migration timing or distance in white-tailed deer

(Van Deelen et al. 1998), hence, deer were

pooled when examining these factors.

RESULTS

Julian date was more strongly correlated with mean daily temperature during spring than dur-

ing autumn at both Canaan (Z= 3.2, P= 0.001) and Odell (Z= 4.8, P= 0.001; Table 1). Likewise,

Julian date and snow depth were more strongly correlated during spring than autumn at both Canaan (Z= 2.1, P= 0.039) and Odell (Z= 8.2, P

0.001). Julian date was more strongly correlated with mean temperature at Odell compared to Canaan during spring (Z= 2.1, P= 0.036), but not

during autumn (Z = 1.8, P = 0.076). Julian date and snow depth also were more strongly correlated at Odell than at Canaan during both autumn

(Z= 2.5, P= 0.013) and spring (Z= 6.8, P= 0.001).

Migration Strategies We captured and radiocollared 63 deer and relo-

cated them 2,588 times in the Canaan area. Sea- sonal movement patterns were ascertained for 54 of these deer (42 adult F, 2juvenile F, 9 adult M, 1

juvenile M). Fifteen deer (28%) were nonmigratory (i.e., residents) and had home ranges within or near the 3 wintering areas. Of 39 migratory deer, 26 survived through 2 migration periods and were evaluated for conditional versus obligate migration patterns. Sixteen deer (62%) were conditional migrators. Mean straight-line migration distance for conditional migrators (13.7 + 1.7 [SE] km) at Canaan was shorter than that of obligate migrators (17.9 ? 1.6 [SE] km; t= -2.031, P= 0.054).

We captured and radiocollared 123 deer and relocated them 2,611 times in the Odell deeryard. Seasonal movement patterns could be determined for 51 of these deer (27 adult F, 7 juvenile F, 8 adult M, 9 juvenile M). All were migratory. Thir-

ty-six of these migratory deer survived through >2

migration periods and were evaluated for conditional versus obligate migration patterns. Condi- tional migrators accounted for 11% (4 of 36), sig- nificantly less than at Canaan (X2 = 15.336, P =

0.001). Mean migration distance for conditional

migrators (6.8 + 2.1 [SE] km) was shorter than that of obligate migrators (20.2 + 1.8 [SE] km) at Odell (t = -3.714, P= 0.019). Mean migration distance for conditional migrators at Odell also was less than that for conditional migrators at Canaan (t = -2.283, P= 0.035), while migration distance of obligate migrators did not differ between the 2 study areas (t = 0.621, P= 0.540).

Autumn Migration Staging behavior was not noted at Canaan or

Odell. Radiocollared deer moved directly from 1 seasonal home range to another when migrating. Migration duration appeared to be brief, since several instances occurred of deer being located on winter range only 8-10 hours after being located on summer range at Canaan. Only on rare occasions were deer located in transit between ranges.

Autumn migration by most deer at Canaan coincided with accumulation of snow above lim-

iting depths (40 cm) during all 3 years (Fig. 2). Seventy-six percent (n = 35 of 46) of initial 1995-1997 autumn migration events occurred within 2 weeks of first snow depths reaching >40 cm. Timing of first occurrence of such depths ranged from 26 December in 1995 to 19 February in 1997, resulting in a wide range of peak migration periods (Table 2).



1994-95

Fall migration: n = 4 Spring migration: n = 33

-.sk-. _ __ l'.' ,' I .I

40 1995-96

80 -

Fall migration: n = 23 60 - Spring migration: n = 23

40-

20- ,- AA ..%

0 . .

',:,"v- : ,. 1",!Iv, B I:1

-20 -

-40

80 -

60-

40-

20 -

0

-20 -

-V% . %, % A &: 11 1 1 11lr

l V. I

. %..

I .-

1996-97

Fall migration: n = 19

:^/' v.." A. '.A /- * .11 , , ..

I

-40

o 0 w <r>

0 z eo

w (. IL

0) a

cu

z .-)

m 1U. LU LL

04

a. CL

<: c~

Fig. 2. Mean daily air temperatures (?C, dashed lines), snow depths (cm, solid lines), weekly autumn migration events (percent of seasonal total, open bars), and weekly spring migration events (percent of seasonal total, solid bars) at Canaan, New Brunswick, Canada, 1994-1997. The Y-axis is shared by all 4 variables. An additional spring migration event during 1994-1995, representing 3.0% of the total that year, occurred between 25 May and 12 Jul 1995 (not shown). An additional spring migration event during 1995-1996, representing 4.3% of the total that year, occurred between 7 Jun and 28 Aug 1996 (not shown).

Autumn migration patterns for the remaining 11 deer at Canaan were mixed. Of the remaining initial migration events, 9% (n = 4 of 46) occurred in mid- to late winter several weeks after snow depths had reached limiting conditions and showed no obvious pattern with respect to climate conditions. Another 11% (n = 5 of 46) coincided with initial snow accumulation, although snow depths reached only 10 cm. The remaining 4% (n = 2 of 46) of migrations occurred in mid-

December 1996, a period of no snow cover and of relatively warm mean daily temperatures (approx. 0 ?C). Migration distances for these early migrators were greater than for those deer migrating later that winter when limiting snow depths accu- mulated in late February 1997 (t = 3.502, P= 0.004).

Due to the high incidence of conditional migration at Canaan, an additional 6 repeat autumn migrations occurred during winter 1995-1996 by deer that undertook either 2 (n = 4 deer) or 3 (n = 1 deer) trips to winter ranges and back. Of these 6 repeat migrations, 4 occurred within 2 weeks following snow accumulations (6-36 cm, mean = 15 cm) subsequent to periods of no snow cover resulting from midwinter thaws.

During winter 1997-1998, migration dates of individual deer at Canaan were not precisely determined. However, the overall migration pattern appeared consistent with previous years. The wintering area was visited on 4 December 1997, at which time snow depths were 30-35 cm and none of the remaining 13 radiocollared deer had migrated. A subsequent visit on 17 December 1997-several days after a snowfall that increased snow depths to 55-60 cm-revealed that 9 radiocollared deer (69%) had migrated to winter ranges. The remaining 4 deer were present near their summer ranges, at a large cutting operation for a new highway right-of-way.

Mean autumn migration dates at Odell occurred 1 to 2 months earlier than at Canaan (Table 3). During 1995-1998, 51% (n= 40 of 79) of autumn

migrations at Odell occurred during the 2 weeks following initial snow accumulation (2-20 cm), which closely coincided with the first 7-day period of mean temperatures 0 ?C (Fig. 3). Some inconsistency occurred among years, as peak migration period coincided with initial snowfall only in 3 of 4 years (Table 2). Initial snow accumulation consistently occurred during the last week of November in all 4 years. Most of the remaining migrations (23%, n = 18 of 79) were associated with a period of several days in 1998 of minimum daily temperatures dropping below -13 ?C, which occurred several weeks after initial snow accumulation (25-30 cm). These accounted for most (78%) migrations in 1998. An additional 11% (n = 9 of 79) of deer at Odell migrated following the first snow depths >40 cm. Of the remaining migrations, 3% (n = 2 of 79) were associated with several consecutive days of mean temperatures 0 ?C with no snow cover, while the remainder showed no association with changes in snow depth or temperature and occurred either

100

80 -

60 -

40 -

20 -

0

-20-

.................... **.....* .**


.* I


Table 2. Peak autumn migration periods for white-tailed deer and climate variables in New Brunswick, Canada, 1995-1998.

Peak migration Migration Snow Temperature Area Winter na periodb %c First snowd >40 cm <0 oCe

Canaan 1994-1995 4 22 Jan 50.0 10 Dec 12 Jan 12 Dec 1995-1996 23 24 Dec 82.6 3 Dec 26 Dec 29 Nov 1996-1997 19 23 Feb 57.9 7 Jan 19 Feb 24 Nov

Odell 1994-1995 12 27 Nov 58.3 24 Nov 7 Jan 23 Nov 1995-1996 17 23 Nov 64.7 28 Nov 26 Dec 23 Nov 1996-1997 27 21 Nov 63.0 20 Nov 29 Jan 12 Nov 1997-1998 23 10 Dec 78.3 24 Nov 26 Dec 13 Nov

a Number of deer monitored. b Start date of 2-wk period containing the highest percentage of migration events. c Percentage of migration events occurring during peak migration period. d Date of initial occurrence of snow accumulation persisting for >7 days. e Date of initial occurrence of mean daily temperatures <0 ?C persisting for >7 days.

in early-late autumn (4%, n = 3 of 79) or in midwinter (9%, n = 7 of 79).

Spring Migration Timing of 95% (n = 53 of 56) of spring migra-

tions at Canaan during 1995-1996 coincided with the disappearance of snow cover in open areas

(Fig. 2). Peak migration periods began 0 to 5 days before or after bare ground appeared in both

years (Table 4). During winter 1995-1996, snow cover disappeared on 3 separate occasions: late

January, late February, and late March. All resulted in spring migration events, causing a wide

range of migration dates (Table 5). Of the

remaining migrations, 2% (n = 1 of 56) occurred

very early in winter 1995-1996 when a deer moved permanently back to summer range on 7 January following a rapid decrease in snow depth from 50 to 21 cm. The other 2 (4%) represent 1 deer migrating to summer range at a very late date for 2 consecutive years: between 25 May and 12 July 1995, and between 7 June and 28 August

Table 3. Autumn migration dates for white-tailed deer in New Brunswick, Canada, 1995-1998.

SE Area Winter na Mean (days)

Canaan 1994-1995 4 12 Feb 11.7 1995-1996 23 2Jan 2.8 1996-1997 19 6Feb 6.1

Odell 1994-1995 12 7 Dec 7.6 1995-1996 17 10 Dec 4.6 1996-1997 27 8Dec 3.9 1997-1998 23 20 Dec 3.8

Range Range (days)

22 Jan-11 Mar 48 21 Dec-19 Feb 60 12 Dec-4 Mar 82 27 Oct-25 Jan 90 23 Nov-10 Jan 48 18 Nov-31 Jan 74

8 Nov-12 Feb 96

a Number of deer monitored.

1996. Migration dates of radiocollared deer at Canaan were not determined in winter 1996-1997, but snow cover (12 cm) persisted when deer were last located on 12 April. All deer remained on winter range at this time, 2 weeks later than deer had initiated spring migration in the previous 2 years.

An additional 6 repeat spring migrations occurred by deer that undertook multiple trips to winter range and back in winter 1995-1996. All of these repeat migrations occurred immediately following disappearance of snow cover in open areas in late February and late March, concurrent with initial spring migrations by other radiocollared deer.

At Odell during 1994-1999, 88% (n = 102 of 116) of spring migrations also coincided with disappearance of snow cover (Fig. 3). Peak migration periods occurred within 1-7 days after snow depths reached 0 cm (Table 4). Of the remaining migrations, 12% (n= 14 of 116) occurred during periods of rapidly decreasing snow depths 1-2 weeks prior to the main pulse of migration. The mean and range of spring migration dates varied little over the 5 years of study (Table 5).

DISCUSSION

Migration Strategies Deer in the 2 study areas exhibited a mixture of

migration strategies. Both migratory and nonmigratory deer were present at Canaan, with conditional migration accounting for most migrators. All deer at Odell were migratory, and conditional migration was uncommon. Nicholson et al. (1997) suggested that variability in climate, particularly in snow depths, ultimately may be



1993-94

n = 20

' SJ; .. ^ r. A Y . ... A_ v i

-.. . .. .... ,t - "I .. : .. I;, .:-

.

1994-95


'&\.:-fr N y, - 1 .A

1995-96


.1 .99,.. .9

1996-97

Fall migration: n = 27 oprnng migrauun; n - zo

Di o..,.., ) , ' \... *.

I1 '4~~ ~ ' j-

1997-98

Fall migration: n =23

Spring migration, n=17

1% /fn - - o a -U'^^

=0 z~

s.v o zm pr a: 0 0 - I. X WI 0 I1I

NI a A -4 re N N

Fig. 3. Mean daily air temperatures (?C, dashed lines), snow depths (cm, solid lines), weekly autumn migration events (percent of seasonal total, open bars), and weekly spring migration events (percent of seasonal total, solid bars) at Odell, New Brunswick, Canada, 1993-1998. The Y-axis is shared by all 4 variables.

responsible for development of a mixed strategy of residents and migrators among a mule deer (0. hemionus) population in montane California. Our results appear to support this hypothesis for white-tailed deer. The proximity of the Canaan area to the Bay of Fundy resulted in a moderating maritime influence on climate and high variabil-

ity in snow depth, while at Odell, winter climate was more consistently severe. However, occurrence of nonmigratory deer only in or near the

wintering areas at Canaan suggests that this strat-

egy was viable only under certain conditions within the context of these variable climate conditions. We suggest that 2 factors are important. Resident deer occupying home ranges within

wintering areas would benefit from the positive aspects of increased deer densities during winter, primarily the development of a communal trail

system at deep snow depths. However, this benefit to resident deer in wintering areas would be balanced by increased browsing pressure at

greater deer densities. Nonetheless, during fre-

quent periods of low snow depths occurring under a variable climate regime, both residents and migrants within a wintering area would be able to access the browse-rich open forest types. This would act to reduce the impact of browsing pressure, although it would still be greater than for deer remaining dispersed on summer range under similar conditions.

Another potential factor explaining the occurrence of nonmigratory deer only near wintering areas at Canaan was the distribution of cedar and hemlock, both of which were preferred forest- cover species during severe winter conditions

(e.g., Habeck 1960, Armstrong et al. 1983, Verme and Johnston 1986). These species were largely absent on summer range of migratory deer; hence, nonmigration from summer range may be a poor strategy for deer at Canaan given that severe winters periodically occur. However, areas of mature spruce (Picea spp.), which provided acceptable, although less-preferred winter cover, existed on summer range so this factor may be of

secondary importance. Behavior corresponding with conditional

migration has been reported for white-tailed deer in New York (Cook and Hamilton 1942), New Brunswick (Drolet 1976), Minnesota (Nel- son 1995), and Michigan (Van Deelen et al. 1998), and for mule deer in Idaho (Brown 1992) and California (Nicholson et al. 1997). These cases of conditional migration share 2 characteristics: they usually occurred when snow depths

100

80

60

40

20

0

-20

100 '

80

60

40

20

0

-20

-40 1

100 -

80 -

60 -

40 -

20 -

0

-20 -

-40

100 ?

80 -

60 -

40 -

20 -

0

.20 -

-40 1

160 -

120 -

80 -

40 -

40 -40

_, *-_ .. . . .r.; _._ .t. --W .. 4,,) . . . . . . . - . 1 r

-40 ?

.. v.v e Al.. ts. . . Av ...... -.r .

* w- t W .S Rv y= v * - - -.


, ; :

"

.a1 .I, I S

. I

"I,r

I


Table 4. Peak spring migration periods for white-tailed deer and climate variables in New Brunswick, Canada, 1994-1998.

Peak migration Migration Snow Snow Temperature Area Winter na periodb %c <40 cm 0 cm >0 oCd

Canaan 1994-1995 33 4 Apr 66.7 14 Jan 29 Mar 14 Mar 1995-1996 23 28 Mar 39.1 31 Dec 29 Mar 5 Apr

Odell 1993-1994 20 23 Apr 95 14 Apr 17 Apr 9 Apr 1994-1995 26 21 Apr 92.3 1 Apr 20 Apr 12 Apr 1995-1996 27 8 Apr 74.1 21 Mar 2 Apr 11 Apr 1996-1997 26 17 Apr 84.6 5 Apr 22 Apr 16 Apr 1997-1998 17 6 Apr 94.1 29 Mar 13 Apr 27 Mar

a Number of deer monitored. b Start date of 2-week period containing the highest percentage of migration events. c Percentage of migration events occurring during peak migration period. d Date of initial occurrence of mean daily temperatures <0 ?C persisting for >7 days.

were below normal, and they were undertaken by a small proportion of the total deer population. These characteristics were consistent with deer behavior at Odell. However, behavior by deer at Canaan differed in that most were conditional

migrators. Variable snow cover has been suggested as the cause of conditional migration behavior in mule deer (Nicholson et al. 1997). Our results

support this hypothesis for white-tailed deer where conditional migration was a very minor

strategy in which winters were consistently severe, but prevalent where winter conditions were variable. Nonetheless, climate variability cannot ex-

plain the distribution of conditional migration among animals of a given area.

Nelson (1995) suggested several possible rea- sons for variation in prevalence of conditional

migration among deer of adjacent wintering areas: differences in hunting mortality on summer range, lower population size and density, and shorter migration distance. Two of these pro-

Table 5. Spring migration dates for white-tailed deer in New Brunswick, Canada, 1994-1998.

SE Range Area Winter na Mean (days) Range (days)

Canaan 1994-1995 33 10Apr 1.3 27Mar-24Apr 28 1995-1996 23 11 Mar 7.6 7 Jan-19 Apr 103

Odell 1993-1994 20 29 Apr 0.9 16Apr-5 May 19 1994-1995 26 29 Apr 0.9 12Apr-9May 27 1995-1996 27 17 Apr 1.3 2Apr-3 May 31 1996-1997 26 26 Apr 1.6 2 Apr-1 May 29 1997-1998 17 12Apr 1.1 6Apr-25Apr 19

a Number of deer monitored. One deer that migrated on an undetermined date between 25 May and 12 Jul 1995, and between 7 Jun and 28 Aug 1996 was not included in summa- ry statistics for spring migration at Canaan.

posed causes, hunting mortality and population size-density, appear inconsistent with our data from the Canaan study area. Although sample sizes were insufficient to accurately assess hunt-

ing mortality separately for the 3 subpopulations of deer in the area, all summer ranges were near each other on public land and likely were ex-

posed to equal levels of hunting pressure. Differ- ences in population size also seem to be an un-

likely cause because conditional migrators constituted most migratory deer in the wintering areas at Canaan that contained both the highest and lowest numbers of deer. However, migration distances of conditional migrants were shorter than those of obligate migrants. A similar pattern was observed at Odell. We suggest that this pattern developed in response to the cost of delaying departure from summer ranges. By delaying departure, deer risk being trapped on summer

range without access to suitable cover by sudden winter storms bringing snow depths limiting to deer movements. Deer attempting migration under such conditions would face greater costs with respect to time and energy expenditure (Parker et al. 1984), with cost increasing with

migration distance. This might translate into de- creased survival given higher mortality rates re-

ported for deer during migration (Nelson and Mech 1991). Further reinforcing this hypothesis is the pattern of migration distance for northern versus southern conditional migrants. Condi- tional migrants at Odell occupied summer

ranges less than half the distance from winter

ranges than those at Canaan, where winters were less severe. Thus, conditional migration is a viable behavior primarily under variable winter climate conditions, and is especially so for those individual deer whose summer ranges are situat-



ed such that the potential cost of adopting the behavior is lowest.

Autumn Migration Autumn migration for deer at Canaan appeared

to be driven primarily by snow depth. Peak migration periods coincided with accumulation of 40 cm of snow in all 4 years of study. The wide variation in timing of such snow accumulations

among years corresponded with wide variation in

migration date. Deer at Odell also appeared to

migrate in response to changes in snow depth, although depths associated with migration initiation were much lower and generally were the initial snow accumulations of winter. These periods closely coincided with the initiation of mean tem-

peratures consistently below freezing, showed a

high consistency with respect to date of occurrence among years, and exhibited a higher degree of correlation with date compared with those at Canaan. Thus, it appears that autumn

migration initiation is determined primarily by the proximate stimulus of limiting snow depth occurrence at Canaan, but by seasonal stimuli (possibly snow, temperature, and/or photoperiod) at Odell.

Previous studies of deer migration have reported increasing snow depth (Westover 1972, Drolet 1976, Tierson et al. 1985) and decreasing tem-

perature (Verme 1973, Kearney and Gilbert 1976, Swanson 1993) as the primary stimuli for winter

migration. Other studies suggest that a complex set of stimuli may influence migration, possibly including changes in photoperiod and vegetation characteristics with snow depth and temperature, and that the responsible stimulus may vary greatly among individuals within a local population or among years (Kucera 1992, Nelson 1995, Nicholson et al. 1997). Snow also may have an additive effect on migration, with major snowfall

initiating migration earlier than would otherwise occur in its absence (Nelson 1995). Whether these factors represent proximate or seasonal stimuli is difficult to determine in the absence of detailed climatological information for the areas studied. While changes in vegetation and photoperiod, as well as general declines in temperature, certainly are driven by seasonal change and often coincide, snow accumulation is more variable in timing and extent. Nonetheless, snow accumulation can represent a seasonal stimulus in areas with relatively consistent winter climate conditions. The situation in New Brunswick illus- trates this pattern quite well, where accumulation

of snow was more consistent in timing among years in the north (Odell), but varied greatly in the south (Canaan).

Given this pattern, we suggest that as with conditional migration, autumn migration timing appears to be related to climate variability. In

regions with consistently severe winter conditions, seasonal stimuli, which may include changes in snow depth depending on prevailing climate

regimes, drive autumn migration. They serve to cue deer to migrate with the approach of winter, which will almost certainly result in limiting snow

depths. In areas where winters are only moder- ately severe, and particularly where climate conditions are variable, limiting snow depths do not

necessarily occur. As a result, risk of delaying departure from summer range is less, and the

potential benefit of remaining on summer range, such as increased food abundance and reduced

competition, is greater. Hence, under these conditions deer are more likely to migrate in

response to proximate stimuli, particularly the onset of limiting snow depths.

Although autumn migration timing may be related to climate variability, the distribution of individuals responding to either proximate or seasonal stimuli within a deer population cannot be explained by this factor, as climate variability generally is consistent across local areas. In addition, the often coincident occurrence of various climatic and environmental factors creates diffi- culties in determining whether deer are responding to proximate or seasonal stimuli. However, at Canaan in 1997, the onset of limiting snow depths occurred late enough to allow us to easily identi-

fy seasonal migrators. These deer faced longer migrations than those responding to the proximate stimulus of snow depth. This suggests that autumn migration initiation may correspond to occurrence of conditional migration in that the risk of delaying departure from summer range, which increases with migration distance, might exert significant influence on response stimuli.

Spring Migration Spring migration for deer at both Canaan and

Odell appeared to be driven by snow depth. Peak migration periods coincided with disappearance of snow in all cases. Initiation of spring migration showed a great degree of consistency relative to autumn migration. Migration tended to occur in a distinct pulse as bare ground appeared, resulting in a lower range of migration dates relative to autumn for all years except in 1996 at Canaan,



when 2 midwinter thaws in addition to spring thaw resulted in 3 separate migration pulses. Pre- vious studies have indicated that temperature (Hamerstrom and Blake 1939, Swanson 1993), snow depth (Rongstad and Tester 1969, Westover 1972, Verme 1973), or a combination of snow and

temperature (Drolet 1976, Nelson 1995) may trig- ger spring migration. Although deer appeared to be superficially responding to snow depth at both Canaan and Odell, snow depth, temperature, and date were strongly correlated at both study areas, although more so at Odell. This indicates that these factors may interact, and that spring migration generally may be considered a

response to seasonal stimuli. However, the ten-

dency of deer at Canaan to initiate spring migration in response to midwinter disappearance of snow cover indicates that proximate cues also have a strong influence on initiation of spring migration under a variable climate regime.

Differences in winter climate variability, through its influence on snow cover, appeared to have a dramatic effect on occurrence of conditional

migration and initiation of both autumn and

spring migration in New Brunswick. This factor also may be responsible for the wide disparity in

migratory deer behavior reported in the literature, particularly with respect to factors deter-

mining migration initiation. A close examination of winter climate patterns in other regions may reveal that the co-occurrence of high climate vari-

ability with both conditional migration and

response to proximate stimuli found in New Brunswick is part of a general pattern.

MANAGEMENT IMPLICATIONS In northern areas, management of deer winter-

ing areas depends on knowledge of their location, extent, and numbers of deer using them. This knowledge often is acquired during aerial

surveys, assuming that deer observed during winter were on winter range. This may not always be a reasonable assumption in areas with variable winter climate. We predict that the high frequen- cy of conditional migration in southern boreal forests reduces the effectiveness of deer population surveys during winter. Surveys of wintering areas when snow depths are <40 cm will underes- timate deer numbers in yards or fail to identify yards because some deer remain on summer

ranges at these times. Snow depths >40 cm occur

infrequently in southern areas; thus, effective sur-

veys of wintering areas might be impossible, or

perhaps possible only briefly, in most winters. A

survey method sensitive to dispersed deer over

large areas would be more appropriate in most winters in these areas. However, surveys undertaken in areas where winters are consistently severe and deer migrate to yards largely indepen- dent of proximate conditions, would be more effective if designed to detect deer in clumped distributions. Hence, survey methods should reflect expected deer distributions given the local climate regime and migration patterns.

ACKNOWLEDGMENTS

Funding for this study was provided by the

Fundy Model Forest, Fraser Papers Inc., New Brunswick Department of Natural Resources and

Energy, J. D. Irving Limited, New Brunswick

Cooperative Agreement on Forest Development, and New Brunswick Cooperative Fish and Wild- life Research Unit. We wish to acknowledge the technical and logistic contributions of A. Boer, A. Chown, T. Dilworth, K. Finnemore, J. Gilbert, R.

Jenkins, D. Keppie, S. Lusk, W. Parker, T. Petti-

grew, G. Redmond, M. Roberts, M. Sabine, J. Simpson, G. Sisson, M. Sullivan, L. Vietinghoff, S.

Young, and numerous assistants.

LITERATURE CITED ARMSTRONG, E., D. EULER, AND G. RACEY. 1983. Winter

bed-site selection by white-tailed deer in central Ontario. Journal of Wildlife Management 47:880-884.

BALLARD, W. B., H. A. WHITLAW, D. L. SABINE, R. A. JENK- INS, S. J. YOUNG, AND G. F. FORBES. 1998. White-tailed deer, Odocoileus virginianus, capture techniques in yarding and non-yarding populations in New Brunswick. Canadian Field-Naturalist 112:254-261.

BEIER, P., AND D. R. MCCULLOUGH. 1990. Factors influencing white-tailed deer activity patterns and habitat use. Wildlife Monographs 109.

BROWN, C. G. 1992. Movement and migration patterns of mule deer in southeastern Idaho. Journal of Wild- life Management 56:246-253.

CLOVER, M. R. 1956. Single-gate deer trap. California Fish and Game 42:199-201.

COOK, D. B., AND W.J. HAMILTON,JR. 1942. Winter habits of white-tailed deer in central New York. Journal of Wildlife Management 6:287-291.

DROLET, C. A. 1976. Distribution and movements of white-tailed deer in southern New Brunswick in relation to environmental factors. Canadian Field-Natu- ralist 90:123-136.

GAUDETTE, M. T., AND D. F. STAUFFER. 1988. Assessing habitat of white-tailed deer in southwestern Virginia. Wildlife Society Bulletin 16:284-290.

HABECK,J. R. 1960. Winter deer activity in the white cedar swamps of northern Wisconsin. Ecology 41:327-333.

HAMERSTROM, F. N., AND J. BLAKE. 1939. Winter movements and winter foods of white-tailed deer in central Wisconsin. Journal of Mammalogy 20:206-215.

HAYNE, D. W. 1949. Calculation of size of home range. Journal of Mammalogy 30:1-18.



HUOT, J. 1974. Winter habitat of white-tailed deer at Thirty-one Mile Lake, Quebec. Canadian Field-Natu- ralist 88:293-301.

JESSUP, D. A., W. E. CLARK, P. A. GULLETT, AND K. R.

JONES. 1983. Immobilization of male deer with ketamine and xylazine, and reversal of immobilization with yohimbine. Journal of the American Veterinary Medical Association 38:1339-1341.

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.

KUCERA, E. 1976. Effects of winter conditions on the white-tailed deer of Delta Marsh, Manitoba. Canadian Journal of Zoology 54:1307-1313.

KUCERA, T. E. 1992. Influences of sex and weather on migration of mule deer in California. Great Basin Naturalist 52:122-130.

MARCHINTON, R. L., AND D. H. HIRTH. 1984. Behavior. Pages 129-168 in L. K. Halls, editor. White-tailed deer ecology and management. Stackpole Books, Harris- burg, Pennsylvania, USA.

MECH, L. D., G. D. DELGIUDICE, P. D. KARNS, AND U. S. SEAL. 1985. Yohimbine hydrochloride as an antago- nist to xylazine hydrochloride-ketamine hydrochloride immobilization of white-tailed deer. Journal of Wildlife Diseases 21:405-410.

NELSON, M. E. 1995. Winter range arrival and departure of white-tailed deer in northeastern Minnesota. Cana- dian Journal of Zoology 73:1069-1076.

, AND L. D. MECH. 1991. White-tailed deer movements and wolf predation risk. Canadian Journal of Zoology 69:2696-2699.

NEW BRUNSWICK DEPARTMENT OF NATURAL RESOURCES AND ENERGY. 1991. Forest land habitat management program progress report. New Brunswick Depart- ment of Natural Resources and Energy, Fish and Wildlife Branch, Fredericton, Canada.

NICHOLSON, M. C., R. T. BOWYER, AND J. G. KIE. 1997. Habitat selection and survival of mule deer: tradeoffs associated with migration. Journal of Mammalogy 78:483-504.

PARKER, K. L., C. T. ROBBINS, AND T. A. HANLEY. 1984. Energy expenditures for locomotion by mule deer and elk. Journal of Wildlife Management 48:474-488.

PAULEY, G. R.,J. M. PEEK, AND P. ZAGER. 1993. Predicting white-tailed deer habitat use in northern Idaho. Jour- nal of Wildlife Management 57:904-913.

PHILLIPS, D. 1990. The climates of Canada. Minister of Supply and Services Canada, Ottawa, Ontario, Canada.

POTVIN, F., AND J. HUOT. 1983. Estimating carrying capacity of a white-tailed deer wintering area in

Quebec.Journal of Wildlife Management 47:463-475. RONGSTAD, O.J., ANDJ. R. TESTER. 1969. Movements and

habitat use of white-tailed deer in Minnesota. Journal of Wildlife Management 33:366-379.

SABINE, D. L. 1999. Migration behavior and winter habitat use of white-tailed deer under moderate winter climate conditions. Thesis, University of New Brunswick, Fredericton, Canada.

SEVERINGHAUS, C. A., 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. Stackpole Books, Harrisburg, Pennsylvania, USA.

SPARROWE, R. D., AND P. E SPRINGER. 1970. Seasonal activity patterns of white-tailed deer in eastern South Dakota. Journal of Wildlife Management 34:420-431.

SWANSON, L. A. 1993. Migratory behavior and management of a white-tailed deer herd in Algonquin region, central Ontario. Thesis, University of Waterloo, Ontario, Canada.

TIERSON, W. C., G. F MATFIELD, R. W. SAGE,JR., AND D. E. BEHREND. 1985. Seasonal movements and home range of white-tailed deer in the Adirondacks. Journal of Wildlife Management 49:760-769.

VAN DEELEN, T. R., H. CAMPA, III, M. HAMADY, AND J. B. HAUFLER. 1998. Migration and seasonal range dynam- ics of deer using adjacent deeryards in northern Michi- gan. Journal of Wildlife Management 62:205-213.

VERME, L. J. 1973. Movements of white-tailed deer in upper Michigan. Journal of Wildlife Management 37:545-552.

, AND W. F. JOHNSTON. 1986. Regeneration of northern white cedar deeryards in upper Michigan. Journal of Wildlife Management 50:307-313.

WESTOVER, A. J. 1972. The use of a hemlock-hardwood winter yard by white-tailed deer in northern Michi- gan. Huron Mountain Wildlife Foundation Occasion- al Papers 1, Three Rivers, Michigan, USA.

WHITE, G. C., AND R. A. GARROT. 1990. Analysis of radio-tracking data. Academic Press, London, United Kingdom.

WHITLAW, H. A., W. B. BALLARD, D. L. SABINE, S. J. YOUNG, R. A. JENKINS, AND G.J. FORBES. 1998. Survival and cause-specific mortality rates of adult white-tailed deer in New Brunswick. Journal of Wildlife Manage- ment 62:1335-1341.

ZAR, J. H. 1984. Biostatistical analysis. Prentice-Hall, Englewood Cliffs, NewJersey, USA.

Received 10 November 2000. Accepted 30January 2002. Associate Editor: Krausman.


Migration Behavior of White-Tailed Deer under Varying ...people.trentu.ca/~jebowman/SabineetalJWM.pdf · MIGRATION BEHAVIOR OF WHITE-TAILED DEER UNDER VARYING WINTER CLIMATE REGIMES

Documents