Dispersal in a declining caribou `meta-population`?

advertisement
Dispersal in a declining caribou ‘meta-population’?
March 31, 2008
Harry van Oort1, Bruce N. McLellan2, and Robert Serrouya3
1. Columbia Mountain Caribou Project
RPO 3, P.O. Box 9158, Revelstoke, B.C., V0E 3K0, Canada
email: hvanoort@gmail.com
tel: 250-837-0820
2. British Columbia Ministry of Forests
RPO 3, P.O. Box 9158, Revelstoke, B.C., V0E 3K0, Canada
Email: rserrouya@telus.net
Tel: 250-837-7613
3. British Columbia Ministry of Forests
RPO 3, P.O. Box 9158, Revelstoke, B.C., V0E 3K0, Canada
Email: Bruce.McLellan@gov.bc.ca
Tel: 250-837-7613
1. Abstract
Dispersal behaviours are known to influence the risk of extinction for local populations
and meta-populations. The distribution of mountain caribou, Rangifer tarandus caribou
(Gmelin 1788) has retracted and is currently fragmented into 18 local populations. Most
local populations of mountain caribou are declining. We analyzed a long-term dataset of
radiolocation records to determine parameters of mountain caribou dispersal. Among 252
animals followed for 2 or more years, 8.3 % of the animals moved to new mating grounds
at a distance greater than one home range from their initial breeding home range that we
observed. Annual home ranges differed among age groups, and suggest that 2-year old
caribou are more likely to disperse. Analysis of summer home range switching supported
this notion, with all natal dispersal occurring among 2-year old caribou, but not 1-year
old or 3-year old caribou. The average annual rate of breeding dispersal was 2.9 %, but
dispersal rates were found to differ among sub-populations, with Columbia North caribou
having the largest annual dispersal rate of 13.5 %. Dispersal distances were generally
short relative to this species body size and home range size. There were only 2
documented cases (0.8 % of animals) where dispersal resulted in potential gene flow
among local populations. These results suggest that mountain caribou are not prone to
dispersing often or far, and are therefore susceptible to negative issues associated with
poor dispersal among local populations.
2
2. Introduction
The movement of organisms between breeding locations is common in many plant and
animal taxa. Animals often disperse as juveniles, and sometimes switch breeding
locations later in life (natal and breeding dispersal respectively). Despite potential costs,
dispersal behaviours are thought to be adaptive (Johnson and Gains 1990). Within
species, the probability of dispersing could be altered by many factors including the
animal’s sex (Greenwood 1980), the inheritance of innate exploratory behaviours (Drent
et al. 2003), social factors (Hestbeck 1982; Nilsson 1989), body condition (Dufty and
Belthoff 2001), and population density (Matthysen 2005). Interspecific differences in
dispersal strategies vary greatly, but remain poorly understood; however some
interspecific relationships are known. Dispersal behaviours appear to vary with mating
system (Greenwood 1980), diet (Sutherland et al. 2000), body size (Sutherland et al.
2000), and home range size (Bowman et al. 2002). Beyond these generalizations,
theoretical and empirical knowledge of the evolution of dispersal strategies among
species remains poor. Presumably, species adopt a dispersal strategy that best balances
the costs and benefits of dispersal given their life history requirements, social
organization, and the nature of their environment.
The dispersal strategy adopted by a species may partially explain differences among
species in their sensitivity to environmental change, especially after populations become
fragmented. Variability in dispersal among populations can alter the rate at which
populations change size (Hanski 2001) and share genes (Saccheri et al. 1998), and
ultimately affect the risk of extinction (Simberloff 1988; Stacey and Taper 1992; Doak
and Mills 1994). Dispersal behaviour affects gene flow between populations with
consequences for the genetic viability of populations (Saccheri et al. 1998; Perrin and
Goudet 2001). In small populations with low immigration rates, genetic drift and
consanguineous matings can lead to an increased homozygous inheritance of deleterious
alleles; as a result, inbred progeny may have compromised immune systems, higher
incidence of parasites, and reduced reproductive potential (Simberloff 1988; Keller and
Waller 2002). Inbreeding has been documented to occur within local populations and
may threaten the integrity of a meta-population; for example, small local populations
3
were characterized by inbreeding depression in adders, Vipera berus, and toads, Bufo
bufo (Madsen et al. 1996; Hitchings and Beebee 1998), and this effect has been shown to
have large influence on the local extinctions within a meta-population of butterflies
(Saccheri et al. 1998). Dispersal behaviour also plays a role in maintaining metapopulations by buffering unpredictable decline of local populations via source-sink
mechanisms, provided that population regulatory factors vary in time and space
(Simberloff 1988; Hanski et al. 1995). Hence, species that are prone to dispersal are
better equipped to survive the population fragmentation.
Because dispersal strategies vary among species, and are related to their risk of
extinction, this behaviour has relevance to conservation of species at risk, especially
when their distribution has become fragmented into a multiple local populations. The
objective of this study was to appraise existing data for evidence of dispersal in mountain
caribou, Rangifer tarandus caribou (Gmelin 1788). Mountain caribou, an ecotype of
woodland caribou (Heard and Vagt 1998), are currently declining (Wittmer et al. 2005a;
McLellan et al. 2006). Much of their historic distribution is no longer occupied, and their
current distribution is now fragmented into 18 sub-populations (Wittmer et al. 2005a;
Apps and McLellan 2006). Between, and within most subpopulation boundaries, there is
rugged geographic relief in the form of alpine peaks and large valleys. Additionally,
deforestation has altered the habitat available to mountain caribou throughout much of
their distribution (Apps and McLellan 2006; Wittmer et al. 2007). Population decline has
been greatest in the southern sub-populations to the point where negative densitydependent ‘Allee’ processes are likely taking effect (Chourchamp et al. 1999; Wittmer et
al. 2005a; McLellan et al. 2006). These southern populations are more isolated than the
central and northern sub-populations, which are generally packed close together.
Populations appear to be strongly regulated by predation (Wittmer et al. 2005b), and
predator abundance is known to vary among local populations, and appears to have
changed over time (Wittmer et al. 2005a; Mowat in prep). Dispersal behaviour, or lackthereof, is likely to play a role in the extinction risks of local populations and of this
caribou meta-population as a whole; yet, despite the relevance of dispersal behaviour to
conservation, accounts of dispersal in mountain caribou are illusive. In this study, we
4
quantify dispersal behaviour of mountain caribou via analyses of a 24-year dataset of
radiolocations.
3. Study area
Caribou telemetry data were collected from throughout the entire distribution of
mountain caribou. Mountain caribou generally occur in the mountainous interior wet belt
of British Columbia (Wittmer et al. 2005a). Their distribution extends from the
McGregor range in the Rocky Mountains - approximately 80 km to the northeast of
Prince George (55N), south to the international border with the United States near
Kootenay lake (49N). In-between these geographic endpoints, mountain caribou are
found in the Cariboo, Monashee, Selkirk, Purcell and Rocky Mountains.
Throughout their distribution, mountain caribou typically utilize rolling plateau or rugged
mountainous terrain characterized by deep winter snow pack. Mountain caribou select
areas with intact tracts of old cedar-hemlock forest alongside alpine habitat (Apps and
McLellan 2006). Migration between seasonal ranges is typical (Apps et al. 2001). In
summer and late winter, mountain caribou generally utilize sub-alpine parkland habitat
characterized by patchy stands of subalpine fir (Abies lasiocarpa (Hook) Nutt.) and
Engelmann spruce (Picea engelmannii Parry ex Engelm.). In early winter and spring
many caribou move to lower elevation old growth forests of cedar (Thuja plicata Donn
ex D. Don) and Pacific hemlock (Tsuga heterophylla (Raf.) Sarg.).
4. Methods
Between 1984 and 2008, on-going telemetry research on mountain caribou occurred
throughout their geographic range. Data from these studies were used previously to
describe the sub-population structure of mountain caribou throughout their distribution
(Wittmer et al. 2005a), for modelling the dispersion of caribou populations across their
distribution (Apps and McLellan 2006), and to investigate factors that may regulate these
populations (Wittmer et al. 2005b; Wittmer et al. 2007).
Using a helicopter, caribou were captured with a net-gun in late winter when caribou
5
frequent high-elevation open parkland habitat. Among all successfully collared caribou (n
= 518), 60 (12%) were male, but age was not known for all animals. A small proportion
of these animals were captured and collared as calves (i.e., when almost 1 year old; n =
26), and all others were either yearlings (when almost 2 years old) or adults. There were
43 animals classified as yearlings, and 137 were classified as adults. The remaining 312
animals were unclassified, but were known to be either adults or yearlings. Captured
caribou were fitted with VHF- or GPS-collars. VHF collars operated for up to 6 years and
locations were usually determined every 16 days. GPS collars generally lasted up 2 years.
Smaller expandable VHF-collars were used for calves; these lasted up to 2 years. Some
study animals were re-collared and followed for up to a maximum of 11 years. A
proportion of animals (n = 103) were relocated from other areas to the South Selkirk subpopulation. A small dataset was available for examining concurrent locations of calves
and their mothers (n = 10); the majority (7) of these cases provided data for less than one
year, but were still adequate for examining concurrent mother-yearling summer home
ranges.
It was unclear from the literature at what age offspring become independent from their
mothers, so we examined the concurrent mother-calf location data to make this known.
Location data for all days in which both mother and calf locations were collected were
compiled, and the distance between these locations was measured in kilometres. These
data were averaged for each caribou season (5 seasons per year; Apps et al. 2001;
Wittmer et al. 2006) – the average distance between cows and their calves within a
caribou season. All cases began in the calf’s 4th season – late winter. These data were
examined graphically to provide evidence of when calves move away from their mothers,
signifying their independence and the initiation of the natal dispersal phase (Labonté et
al. 1998).
Telemetry location data were examined for evidence of dispersal by testing several
spatial predictions constructed around dispersal behaviour. Dispersal involves unusually
large mobility, and should increase the maximum distance that caribou move from
capture, and the size of their annual home range. In organisms that have a well-defined
6
natal dispersal period, measures of home range size and distance travelled should differ
among age classes. To test for evidence of natal dispersal, we first examined the effects
of age at capture and sex on the tendency toward mobility by comparing first annual
home range size in the year following capture. Considering all caribou locations, we
examined the effects of age at capture, sex, and the number of locations on the maximum
distance moved from capture location. For this study, we defined home ranges as
minimum convex polygons fitted around all annual or seasonal locations for an animal.
We tested for differences among means using general linear models (GLM).
Dispersal of organisms should also result in geographically distinct home ranges.
Dispersal distance and dispersal rates were calculated from a sub-dataset of caribou
locations determined in the summer season. By restricting the dataset to one season,
migration to and from seasonal ranges allowed us to define temporally discrete home
ranges across years. Summer home ranges were examined because this time period
includes their mating period, and therefore bears relevance to the genetic structure of
caribou populations; however, other attributes also make the analysis of summer home
ranges desirable. Summer is a relatively long mountain caribou season, allowing greater
opportunity for accumulating telemetry locations (11 June to 21 October; Wittmer et al.
2006). This means that our determination of a seasonal home range will incur less errors
from inadequate numbers of telemetry locations. Mountain caribou also show highest site
fidelity during the summer (Wittmer et al. 2006), thereby providing greater analytical
power to detect home range switching in this season compared with others.
In almost all cases, we lacked knowledge of the true natal summer home range because
all animals were captured after their first summer. However, for the calves observed
concurrently with their mothers (n = 10), we assumed that the mothers’ concurrent home
range approximates the calf’s natal home range. Otherwise, we used the initial summer
home range delineated for a study animal after first capture to serve as its place of origin,
and examined the data for evidence of summer home range switching (breeding
dispersal). Dispersal was defined to occur when animals used summer home ranges that
did not overlap with their either their mother’s concurrent summer home range, or their
7
initial summer home range, provided that the minimum intervening distance to the edge
of the two summer home range polygons was greater than one average summer home
range length. This proviso was included to reduce errors associated with underestimating
home range size due to limited numbers of locations to delineate summer home range
polygons for each animal. Additionally, the probability of reproducing with a different
mate presumably increases when an animal switches to a home range more than one
home range away (males and females have similar home range size in mountain caribou –
this study). There was an average of 8.6 VHF locations per summer season for all
animals monitored for more than one summer (n = 312). The length of an average home
range (7.85 km) was estimated from a hypothetical square home range with an area equal
to the mean area encompassed by the first summer home ranges from all caribou that
were no trans-located (n = 348, x = 6160 ha).
Consecutive summer home ranges were classified as a dispersal event, or not, for each
observed animal until it was no longer observed, or until the animal was recorded to
disperse. Dispersal rate was calculated as the number of dispersal events divided by the
total number of dispersal opportunities; for example, an animal followed for 9 years
provided 8 opportunities to observe dispersal from the initial summer home range.
Dispersal rates were compared among age groups, sex, and sub-populations using Chisquare goodness of fit procedure (1999; 1999).
Dispersal distance was calculated as the distance between observed summer home range
centres (averaged UTM coordinates). For animals that were followed more than 2
summers, we report the maximum dispersal distance for each animal calculated from the
initial summer home range.
All spatial computing was performed using geo-referenced mapping software (ESRI
ArcView 3.3). SYSTAT 8.0 was used for statistical analyses. In all GLMs we used type
III sums of squares. An alpha level of 0.05 was used throughout all statistical tests.
8
5. Results
Of the 10 calves concurrently telemetered with their mothers, 3 were male. The true
parent-offspring relationship was supported by microsatelite allele sharing for 3 of these
cow-calf cases, for which we had genetic data at 10 loci. The concurrent distances
measured between the collared calf locations and their mothers’ locations showed that
calves leave their mothers in their first spring just prior to entering their ‘yearling’ (2nd)
year (Fig. 1).
Trans-located animals had larger first annual home ranges than native animals (translocated x = 88 625 ha, native x = 24 695 ha, F1, 366 = 69.5, P < 0.0005). We therefore
excluded trans-located animal data from further analyses unless stated otherwise. Among
the animals monitored for at least one year, the first annual home range size was
compared using a multivariate GLM. The initial model included age at capture, sex, sexage interaction term, and the number of locations used to infer home range size. After
running a backward stepwise procedure (with exclusion rule: P > 0.15), only one variable
was retained: age at capture (F2, 105 = 4.73, P = 0.011). A post-hoc comparison showed
that caribou caught as yearlings used larger post-capture (i.e., 2-year old) annual home
ranges than animals caught as calves or adults (Fisher’s LSD test, P = 0.029, and P =
0.003 respectively); but animals caught as calves did not have larger home ranges than
adults (P = 0.83). Neither age, sex, or the age-sex interaction were significantly
predictive of the maximum distance moved away from capture location (backward
stepwise GLM: all terms NS).
There were 252 native animals that we observed for more than one summer season for
which we constructed minimum convex polygons to delineate summer home ranges. In
75% of the animals (188/252), all summer home range polygons overlapped with their
initial summer home range. There were 21 cases (8.3%) where caribou were recorded to
disperse to a different home range separated from the initial home range by a distance
that was greater than one home range length (maximum intervening distance = 39 km;
Fig. 2; Table 1). These data, and data from mother-calf home ranges are examined within
age groups below.
9
Among the caribou monitored concurrently with their mothers (3 males, 7 females), none
were observed to disperse in their yearling summer. Six animals were occasionally
located with their mothers during their yearling summer. Two animals (1 male, 1 female)
used summer home ranges that did not overlap their mother’s home range, but only the
female calf (#149) used a summer home range that was greater than one average summer
home range length from her mothers’ concurrent home range. However, location data
from her mother preceding calf capture allowed us to determine the true natal home range
of this female, and showed that it was the mother that had moved; hence, this was a case
of breeding dispersal for the mother. There were 4 other cases where the mother’s home
range was known in the yearling’s calf year, but none of these other cases involved
breeding dispersal. Correcting for the one known case of maternal breeding dispersal (by
substituting the dispersed mother’s concurrent home range data with her previous year’s
home range data), the median dispersal distance between mother and calf home range
centres was 5.8 km, and the maximum dispersal distance was 11.0 km (Fig. 3).
The first documented dispersal occurred when caribou were two years old and had
switched to a new summer home range for their third summer season: 2 cases (1 male, 1
female) out of 9 opportunities (2 male, 7 female). The median and maximum dispersal
distance between each juvenile’s 1st and 2nd summer home range centres was 3.6 km and
14.0 km respectively (Fig. 3).
There were 12 caribou originally captured as yearlings that were observed for both their
3rd and 4th summers (3 males, 9 females). None of these caribou were found to have
dispersed to a new home range when 3 years old. Median and maximum dispersal
distance between 3rd and 4th summer home range centres was 2.9 km and 8.1 km
respectively (Fig. 3).
The annual rate of breeding dispersal among animals classified as adults was estimated to
be 5% (n = 73; 11 cases of dispersal in 208 annual opportunities). Breaking proportions
of dispersal events and annual opportunities down by sex, there were 2 cases of male
10
dispersal in 27 opportunities; and 9 cases of female dispersal in 181 opportunities (χ2
=0.015, P = 0.90).
In the above results, only 2-year old caribou appeared to behave differently. Although the
data were few, both male and female yearling data gave the suggestion of an elevated
dispersal phase at this age (with rates of 50% and 16% respectively). There was no
indication that 3 year olds had increased levels of dispersal compared to adults. Hence,
we pooled all data for animals 3 years or older (n = 244 caribou), including data from
animals not classified by age, which were definitely not caught earlier than at the late
yearling stage (e.g., after 2-year old dispersal has occurred). Recalculating annual
dispersal rates with this larger dataset also found no sex differences (df = 1, χ2 = 0.73, P >
0.79) and an average annual dispersal rate of 2.9 % (SE = 0.75). The proportion of
dispersal events that occurred were not evenly distributed among the 18 sub-populations
(χ2 = 45.8, df = 17, P = 0.0002; Table 2), with the Columbia North sub-population
having a notably large incidence of dispersal – a total rate of 13.5 %.
Among all locations, 83.5% were located within the sub-population ranges defined by
Wittmer et al. (2005a). A small proportion of non-relocated (2.6%) and relocated caribou
(3.8%) were located within more than one sub-population range (Table 3). Among the 21
non-relocated caribou that dispersed to a new summer home range, 2 animals dispersed
across sub-population boundaries. Caribou 80 was caught in the late winter of 2003 as an
adult in the North Columbia sub-population and in 2004 dispersed approximately 19 km
north to a new summer home range that was mostly located within the Groundhog subpopulation range, but still partially within the North Columbia range. She adopted the
new summer home range for three years in succession. Caribou 139 was captured as a
yearling in 2002 and spent 3 consecutive summers on one summer home range in the
Columbia North range before dispersing in 2005 approximately 39 km north to a new
summer home range overlapping the Columbia North/Groundhog boundary.
11
6. Discussion
Dispersal is important with regards to the extinction risk of local populations and metapopulations (Simberloff 1988; Hanski 2001). During the decline of mountain caribou,
their distribution has been fragmented into 18 subpopulations (Wittmer et al. 2005a;
Apps and McLellan 2006); some of these local populations are relatively isolated and
small (Wittmer et al. 2005a; McLellan et al. 2006). In this study, we appraised a large
dataset of mountain caribou radio-locations to determine their tendency to disperse within
and among sub-population ranges. From the available data, 8.3 % of the animals were
witnessed to move to new mating grounds at a distance greater than one home range from
their initial breeding home range and the average annual rate of breeding dispersal was
2.9 %. Support for a phase of greater natal dispersal rates suggested that natal dispersal
occurs when animals are 2 years old. Dispersal rates were found to differ among subpopulations with Columbia North caribou having the largest annual dispersal rate of 14
%. There were only 2 documented cases (0.8 % of animals) where a dispersal event
resulted in potential gene flow among sub-populations. Our results suggest that mountain
caribou are generally poor dispersers; this may contribute the risk of extinction for local
populations and for the meta-population as a whole. We discuss these results in greater
detail below.
The dispersal rates of mountain caribou in our study were generally low compared with
other ungulates (Table 4). However, our dataset is composed largely of female caribou
collared as adults, so we must be cautious about how we interpret our results, especially
when looking at dispersal rates. In many ungulates, dispersal occurs primarily via
juvenile males; hence, datasets that do not represent young males well, like ours, stand to
underestimate the true dispersal rates if care is not taken. We deal with this problem by
considering dispersal distances and by considering dispersal rates within sex/age groups.
Ungulates typically begin natal dispersal as yearlings, with many yearlings using a
different breeding home range then where they were born; this is true for species that are
smaller than caribou such as the white-tailed deer (Odocoileus virginianus), and for
larger species such as the moose (Alces alces; Table 4). In dispersal studies from other
12
ungulate species, dispersal rates of yearlings is reported to be between 14 % and 58 % for
females at this age, and between 21 % and 64 % for males (Table 4). In 10 yearling cases
observed in this study (3 male, 7 female), there was no dispersal observed when caribou
were 1 year old. Although this is not a large sample, it gives the impression that young
mountain caribou yearlings are not strongly driven to disperse after they become
independent.
Our data suggested that a natal dispersal phase occurred when caribou moved to their
third summer home range as 2-year olds. Annual home ranges were larger during this
period. Among female 2-year olds, the dispersal rate was estimated at 14 %. One out of 2
monitored males dispersed at this age. The female rate is generally low compared with
other female ungulate species during natal dispersal. Given that we did not observe many
2-year old male cases, we cannot comment on dispersal rates with a high degree of
certainty; however, unless 2-year old males are extremely prone to dispersing, when
considering all other comparisons within the various juvenile age groups, our data give
the impression that mountain caribou disperse at low rates (Table 4).
The annual rate of breeding dispersal was found to be relatively normal in mountain
caribou compared with other studies (Table 4); in short, like other species, breeding
dispersal generally does not provide a high among of movement in mountain caribou.
However, this generalization may not apply to all populations, which we discuss in
greater detail below.
Another way to compare the dispersal behaviour mountain caribou with other species is
to apply models produced by recent meta-analyses. These models predict median and
maximum dispersal distances for species based on body size or home range size.
Sutherland et al. (2000) produced models suggesting that body size and diet were
predictive of median and maximum natal dispersal distance. Applying the Sutherland
model for herbivorous mammals to mountain caribou using a body mass of 129 kg for
female mountain caribou (McTaggart-Cowan and Guiguet 1978) predicted a median natal
dispersal distance of 19.7 km and a maximum dispersal distance of 77.9 km. The mass of
13
a large bull (272 kg; (McTaggart-Cowan and Guiguet 1978) produced median and
maximum natal dispersal distance estimates of 29.5 km and 126.6 km respectively.
Bowman et al. (2002), produced a pair of complimentary models for predicting median
and maximum natal dispersal distance based on annual home range size for mammals;
based on the linear dimension of an average annual home range (15.71 km), median natal
dispersal distance was predicted to be 110 km, and maximum natal dispersal distance was
predicted to be 628 km. These estimates above predict normal natal dispersal distances of
mammals, yet the maximum natal dispersal distance found among young caribou was
only 14 km, which is 71 % of the smallest predicted median distance. Breeding dispersal
distances (Table 1) were between 4 % and 74 % of the predicted natal dispersal values.
Hence, in terms of dispersal distance, our data suggest that mountain caribou have poorly
developed dispersal behaviour or abilities compared with other mammal species.
Our data provide robust data on breeding dispersal of females, and reasonable data on
breeding dispersal of males; however the limited data on juvenile caribou provide a more
limited analysis of natal dispersal. From these data, we generally found that mountain
caribou do not show strong tendencies to disperse often (in the comparisons that could be
made), or far. Interspecific differences in dispersal strategies have not been explored
theoretically or empirically in much detail. It is possible that innate dispersal behaviours
are lost in migratory herding organisms such as the barren ground caribou. Hence, low
dispersal rates could be an ancestral trait for mountain caribou. It is also possible that
mountain caribou are poor dispersers because they are constrained by the geography of
the landscape.
Breeding dispersal rates differed among populations, with the rate for one sub-population
– Columbia North – being 14 %, which is unusually large. We have not investigated
circumstances that may have contributed towards this finding. Dispersal rates could be
higher in this population if densities are unusually high or low compared with other
populations (Matthysen 2005). Another possibility is that dispersal rates in the Columbia
North are higher because of some landscape attribute; for example, habitat patch size,
patchiness, or connectivity (Coulon et al. 2004). In the former hypothesis, dispersal rates
14
may have changed over time with changing density, which would not necessarily be true
for the latter hypothesis. It is also possible that some ecological factor (food quality,
predation pressure, disturbance by human activity) was more variable in the Columbia
North sub-population, causing caribou to switch summer home ranges during the study
period.
Mountain caribou are in strong decline, and their ever-retracting distribution has been
accompanied by increasing population fragmentation. Many of the sub-populations are
divided by considerable geographic barriers such as steep glaciated mountains, reservoirs,
and busy motor-vehicle highways. The Monashee sub-population – last censused to be
only 7 remaining individuals (McLellan et al. 2006) – is isolated by at least 20 km from
the closest neighbouring sub-population. In meta-populations such as this, the ability of a
species to disperse becomes an important attribute with regards to the meta-population’s
extinction risk (Simberloff 1988; Saccheri et al. 1998; Hanski 2001). Our study utilized
the available data to address the question: do mountain caribou have well-developed
dispersal behaviours. While the data were not altogether ideal for this question, we feel
that we have uncovered enough evidence to suggest that mountain caribou are not
endowed with a great degree of dispersal behaviour – both in terms of dispersal rates and
distances. Furthermore, we uncovered very little evidence of potential gene flow among
sub-populations. Hence, we predict that the smaller local populations are likely to have
very low genetic variability and could potentially be expressing symptoms of inbreeding.
Future research should endeavour to explore genetic variability within local populations.
In this study, we considered dispersal at the scale of 1 home range and at the scale of
among sub-populations. It is possible that sub-populations are segregated into multiple
breeding populations, or segments (Smith and Anderson 2001), that utilize similar early
winter or spring habitats but different breeding grounds. We suggest that the analysis
reported here is expanded to consider dispersal at an intermediate scale by subdividing
caribou into breeding ranges according to their summer locations, so that dispersal among
true breeding grounds can be ascertained. Finally, we suggest that a variety of analysis
15
should be performed to understand why the Columbia North sub-population accounts for
such a large proportion of the dispersal detected in this study.
7. Acknowledgements
Funding for this research was provided by the Okanagan Innovative Forestry Society
(project 4776005, telemetry to monitor dispersal events), Federated CooP (project
4772003, collar purchases for continued dispersal monitoring) the Kamloops TSA
licensees and SIMPCW Development Corporation, B.C. Species at Risk Coordination
Office, B.C. Forest Sciences Program, B.C. Ministry of Environment, and the B.C.
Ministry of Forests and Range. We thank K. Furk, T. Kinley, D. Seip, and G. Watts for
help compiling and collecting much of the data.
8. Reference List
Apps, C.D. and McLellan, B.N. 2006. Factors influencing the dispersion and
fragmentation of endangered mountain caribou populations. Biological Conservation
130: 84-97.
Apps, C.D., McLellan, B.N., Kinley, T.A., and Flaa, J.P. 2001. Scale-dependent habitat
selection by mountain caribou, Columbia Mountains, British Columbia. The Journal of
Wildlife Management 65: 65-77.
Bowman, J., Jaeger, J.A.G., and Fahrig, L. 2002. Dispersal distance of mammals is
proportional to home range size. Ecology 83: 2049–2055.
Chourchamp, F., Clutton-Brock, T., and Grenfell, B. 1999. Trends in Ecology and
Evolution 14: 405-410.
Coulon, A., Cosson, J.F., Angibault, J.M., Cargnelutti, B., Galan, M., Morellet, N., Petit,
E., Aulagnier, S., and Hewison, J.M. 2004. Landsacpe connectivity influences gene flow
in a roe deer population inhabiting a fragmented landsacpe: an individual-based
approach. Molecular Ecology 13: 2841-2850.
Doak, D.F. and Mills, L.S. 1994. A useful role for theory in conservation. Ecology 75:
615-626.
Drent, P.J., van Oers, K., and van Noordwijk, A.J. 2003. Realized Heritability of
Personalities in the Great Tit (Parus major). Proceedings of the Royal Society of London,
16
Series B 270: 45-51.
Dufty, A.M. and Belthoff, J.R. 2001. Proximate mechanisms of natal dispersal: the role
of body condition and hormones. In Dispersal. Oxford University Press, New York. pp.
217-229.
Greenwood, P.J. 1980. Mating systems, philopatry and dispersal in birds and mammals.
Animal Behaviour 28: 1140-1162.
Hanski, I. 2001. Population dynamic consequences of dispersal in local populations and
in metapopulations. In Dispersal. Oxford University Press, New York. pp. 283-298.
Hanski, I., Pakkala, T., Kuussaari, M., and Guangchun, L. 1995. Metapopulation
persistence of an endangered butterfly in a fragmented landscape. Oikos 72: 21-28.
Heard, D.C. and Vagt, K.L. 1998. Caribou in British Columbia: a 1996 status report.
Rangifer 10: 117-123.
Hestbeck, J.B. 1982. Population Regulation of Cyclic Mammals: The Social Fence
Hypothesis. Oikos 39: 157-163.
Hitchings, S.P. and Beebee, T.J.C. 1998. Loss of genetic diversity and fitness in common
toad (Bufo bufo ) populations isolated by inimical habitat. Jounral of evolutionary biology
11: 269-283.
Jacques, C.N. and Jenks, J.A. 2007. Dispersal of yearling pronghorns in Western South
Dakota. The Journal of Wildlife Management 71: 177-182.
Johnson, M.L. and Gains, M.S. 1990. Evolution of dispersal: theoretical models and
empirical tests using birds and mammals. Annual Review of Ecology and Systematics
21: 449-480.
Keller, L.F. and Waller, D.M. 2002. Inbreeding effects in wild populations. Trends in
Ecology and Evolution 17: 230-241.
Labonté, J., Ouellet, J.-P., Courtois, R., and Bélisle, F. 1998. Moose dispersal and its role
in the maintenance of harvested populations. The Journal of Wildlife Management 62:
225-235.
Madsen, T., Stille, B., and Shine, R. 1996. Inbreeding depression in an isolated
population of adders Vipera berus. Biological Conservation 75: 113-118.
Matthysen, E. 2005. Density-dependent dispersal in birds and mammals. Ecography 28:
403-/416.
McLellan, Bruce, Serrouya, Robert, and Flaa, John. Population censuses of caribou the
north Columbai Mountains. 2006.
17
McTaggart Cowan, I. and Guiguet, C.J. 1978. The Mammals of British Columbia.
British Columbia Provincial Museum, Victoria, B. C.
Mowat, G. (In prep). Large carnivore population review in the Kootenay region.
Nelson, M.E. 1993. Natal Dispersal and Gene Flow in White-Tailed Deer in Northeastern
Minnesota. Journal of Mammalogy 74: 316-322.
Nilsson, J.-A. 1989. Causes and Consequences of Natal Dispersal in the Marsh Tit,
Parus palustris. The Journal of Animal Ecology 58: 619-636.
Ochiai, K. and Susaki, K. 2007. Causes of natal dispersal in a monogamous ungulate, the
Japanese serow, Capricornis crispus. Animal Behaviour 73: 125-131.
Perrin, N. and Goudet, J. 2001. Inbreeding, kinship, and the evolution of natal dispersal.
In Dispersal. Oxford University Press, New York. pp. 123-142.
Porter, W.F., Underwood, H.B., and Woodard, J.L. 2004. Movement behavior, dispersal,
and the potential for locatlized management of deer in a suburban environment. Journal
of wildlife management 68: 247-256.
Rosenberry, C.S., Lancia, R.A., and Conner, M.C. 1999. Population effects of whitetailed deer dispersal. Wildlife Society Bulletin 27: 858-864.
Saccheri, I., Kuussaari, M., Kankare, M., Vikman, P., Fortelius, W., and Hanski, I. 1998.
Inbreeding and extinction in a butterfly metapopulation. Nature 392: 491-494.
Simberloff, D. 1988. The Contribution of Population and Community Biology to
Conservation Science. Annual Review of Ecology and Systematics 19: 473-511.
Smith, B.L. and Anderson, S.H. 2001. Does dispersal help regulate the Jackson elk herd?
Wildlife Society Bulletin 29: 331-341.
Smith, B. L. and Robbins, R. L. Migrations and management of the Jackson elk herd.
1994. Washington, D. C., U. S. A., United States National Biological Service.
Stacey, P.B. and Taper, M. 1992. Environmental Variation and the Persistence of Small
Populations. Ecological Applications 2: 18-29.
Sutherland, G.D., Harestad, A.S., Price, K., and Lertzmann, K.P. 2000. Scaling of Natal
Dispersal Distances in Terrestrial Birds and Mammals . Conservation Ecology 4.
Wittmer, H.U., McLellan, B.N., and Hovey, F.W. 2006. Factors influencing variation in
site fidelity of woodland caribou (Rangifer tarandus caribou ) in souteastern British
Columbia. Canadian Journal of Zoology 84: 537-545.
Wittmer, H.U., McLellan, B.N., Seip, D.R., Young, J.A., Kinley, T.A., Watts, G.S., and
Hamilton, D. 2005a. Population dynamics of the endangered mountain ecotype of
18
woodland caribou (Rangifer tarandus caribou) in British Columbia, Canada. Canadian
Journal of Zoology 83: 407-418.
Wittmer, H.U., McLellan, B.N., Serrouya, R., and Apps, C.D. 2007. Changes in
landscape composition influence the decline of a threatened woodland caribou
population. Journal of Animal Ecology 76: 568-579.
Wittmer, H.U., Sinclair, A.R.E., and McLellan, B.N. 2005b. The role of predation in the
decline and extirpation of woodland caribou. Oecologia 144: 257-267.
Zar, J.H. 1999. Biostatistical analysis. Prentice Hall, Upper Saddle River, New Jersey.
19
Table 1. Dispersal data for animals followed for 2 or more summers partitioned by sex and age at capture. Dispersal was defined to
occur when summer home range polygons were separated by at least one average summer home range length (7.85 km) from the
initial summer home range. Dispersal distance was calculated as the distance between averaged summer coordinates, and the median
was calculated from the maximum dispersal distances recorded for every animal.
Proportion that dispersed further than 1 home range (%)
Dispersal distance (km)
Calves
Yearlings
Adults
Unclassified
Total
Median
Maximum
♀
1/7 (14)
1/8 (12)
9/63 (14)
6/145 (4)
17/222 (8)
5.6
49.0
♂
1/2 (50)
1/3 (33)
2/10 (20)
0/15 (0)
4/30 (13)
7.0
58.0
Total
2/8 (25)
2/11 (18)
11/73 (15)
6/160 (4)
21/252 (8)
5.9
58.0
20
Table 2. The total number of cases where caribou (greater than 2 years of age) dispersed
to a new summer home ranges separated by more than one average summer home range
length (7.85 km) from their initial summer home range are listed for each sub-population.
The total number of opportunities equals the total number of discrete summer home
ranges defined following the initial summer home range within each sub-population.
Sub-population
Hart Ranges
North Cariboo
Narrow Lake
George Mountain
Barkerville
Allen Creek
Wells Gray
Groundhog
Columbia North
Central Rockies
Columbia South
Frisby-Boulder
Duncan
Monashee
Nakusp
Purcells North
Purcells South
South Selkirks
TOTAL
No. caribou
17
8
2
2
10
3
83
5
36
5
17
9
3
2
24
5
8
5
Total opportunities
38
19
4
3
29
6
208
13
96
20
63
20
11
6
52
8
15
21
No. dispersed
1
0
0
0
0
0
3
0
13
0
0
0
0
0
2
0
0
0
244
632
19
21
Table 3. Proportion of animals located in multiple sub-population ranges.
Age at Capture
No. moved among ranges
Total
%
Calf
0
26
0.0
Yearling
1
17
5.9
Adult
6
112
5.3
Unclassified
3
276
1.1
Calf
0
1
0.0
Yearling
1
27
3.7
Adult
2
30
6.7
Unclassified
1
46
2.2
Native
Trans-located
22
Table 4. Dispersal rates found in this study compared with rates found in other studies. Scale denotes the type of dispersal reported:
non-overlapping is the smallest scale for dispersal criteria, where home ranges simply do not overlap; 1 home range is similar methods
used by this study, where home ranges need to be separated by a distance of a home range length; inter-segment is the largest scale of
dispersal in this table, and denotes studies where animals disperse to new breeding areas within a sub-population.
Species
age
Caribou
1
Elk
1
Pronghorn
1
Moose
1
White-tailed deer
1
Caribou
1
Elk
1
Pronghorn
1
Moose
1
White-tailed deer
1
White-tailed deer
1
White-tailed deer
1
Caribou
1
Caribou
2
Elk
2
White-tailed deer
2
Caribou
2
Elk
2
White-tailed deer
2
Caribou
2
Caribou
3
Elk
3
Caribou
3
Elk
3
Serow
2 to 4
Serow
2 to 4
Sex
Female
Female
Female
Female
Female
Male
Male
Male
Male
Male
Male
Both
Both
Female
Female
Female
Male
Male
Male
Both
Female
Female
Male
Male
Female
Male
No. Disp
0
6
10
5
7
0
8
9
5
28
38
No. Could have
dispersed
7
42
17
17
35
3
39
17
16
44
54
0
1
5
0
1
1
2
2
0
1
0
0
14
16
10
7
35
17
2
20
22
9
9
21
3
4
16
17
%
0.0
14.3
58.8
29.4
20.0
0.0
20.5
52.9
31.3
63.6
70.4
14.3
0.0
14.3
14.3
0.0
50.0
5.0
9.1
22.0
0.0
4.8
0.0
0.0
87.5
94.1
23
Scale
1 Home Range
Inter-segment
Non-overlapping
1 Home Range
1 Home Range
1 Home Range
Inter-segment
Non-overlapping
1 Home Range
1 Home Range
1 Home Range
Non-overlapping
1 Home Range
1 Home Range
Inter-segment
1 Home Range
1 Home Range
Inter-segment
1 Home Range
1 Home Range
1 Home Range
Inter-segment
1 Home Range
Inter-segment
Non-overlapping
Non-overlapping
Reference
This study
(Smith and Anderson 2001), Table 1
(Jacques and Jenks 2007), Table 1
(Labonté et al. 1998)1998, p. 231
(Nelson 1993), p. 318
This study
Smith and Andersson 2001, Table 1
Jacques and Jenks 2007, Table 1
Labonte et al. 1998, p. 231
Nelson 1993, p. 318
(Rosenberry et al. 1999), p. 8
(Porter et al. 2004), p. 250
This study
This study
Smith and Andersson 2001, Table 1
Nelson 1993, p. 318
This study
Smith and Andersson 2001, Table 1
Nelson 1993, p. 318
This study
This study
Smith and Andersson 2001, Table 1
This study
Smith and Andersson 2001, Table 1
(Ochiai and Susaki 2007), Table 2
Ochiai and Susaki 2007, Table 2
Table 4 Continued.
Species
Caribou
age
adults
Sex
Both
No. followed
244
Elk
White-tailed deer
White-tailed deer
adults
adults
adults
Both
Both
Both
48
59
12
Annual % Scale
2.9
1 Home Range
2.0
8.3
0.0
Inter-segment
Non-overlapping
1 Home Range
24
Reference
This study
(Smith and Robbins 1994) 1994,
p.14
Porter et al. 2004, p. 250
Nelson 1993, p. 318
35
Average Distance (km)
30
25
20
15
10
5
0
Late 4Winter
(Capture)
5
Spring
6
Calving
7
Summer
Early8 Winter
9
Season
Fig. 1. The average distance between calves and their mothers telemetered
simultaneously. All 10 calves were collared in their first late winter season. It appears
that they generally begin to disperse away from their mother’s side in their first spring,
when many mountain caribou move to lower elevation habitats.
25
250
Frequency
200
150
100
50
0
0
5
10
15
20
25
30
35
40
Distance (km)
Fig. 2. The minimum distances between initial and the furthest subsequent summer
home range for caribou followed for 2 or more summers are plotted as a stacked
histogram (n = 252). Of these, 188 (75%) always used overlapping summer home ranges
(in white), and 21 animals (8.3%; in black) were recorded to disperse to a nonoverlapping home range in subsequent years separated by a distance that was greater than
one average summer home range length (7.85 km). The remaining caribou were recorded
to use non-overlapping summer home ranges separated by less than 7.85 km (in grey).
26
60
Dispersal distance (km)
50
40
30
20
10
0
0
10
9
11
8
6
4
3
3
1
2
3
4
5
6
7
8
9
Age (years)
Fig. 3. Dispersal distance, as measured between summer home range centres, plotted for
different age groups. At age 1, comparisons are made between concurrent maternal home
range and calf home range centres. In all other ages, the first summer home range
observed (either as 1 or 2 year old) is used as a place of origin. Sample sizes are noted
below each box plot. In these data, there were 4 cases of dispersal: 2 2-year olds, 1 4-year
old, and 1 5-year old.
27
Download