Thaidene Nene caribou report: January 2011

Transcription

Migratory tundra caribou seasonal and annual distribution relative to Thaidene Nene,
a national park reserve proposal in the East Arm of Great Slave Lake and Artillery Lake area,
Northwest Territories
Submitted in fulfillment of Contract No. 45284055
Parks Canada, Ft. Smith, NWT
Prepared by
1
Anne Gunn , Kim G. Poole2, and J. Wierzchowski3
January 2011
1
368 Roland Road, Salt Spring Island, BC V8K 1V1 gunnan@telus.net
2
Aurora Wildlife Research, 1918 Shannon Point Rd., Nelson, BC V1L 6K1 kpoole@aurorawildlife.com
3
Geomar Consulting Ltd., Stuben 136, Stuben, 7434 Bernstein, Austria jack@abcminerals.com
2
Caribou distribution near Thaidene Nene
EXECUTIVE SUMMARY
Migratory tundra caribou (Rangifer tarandus groenlandicus) from 3 herds seasonally use the area around
the East Arm of Great Slave Lake and Artillery Lake, Northwest Territories. Initially proposed in 1970, work
is continuing on developing a national park reserve in this area. To aid this effort we conducted spatial
analyses of caribou movements and distribution in this proposed reserve (called Thaidene Nene) to
examine how well the withdrawal area captures fall and pre-calving caribou migratory routes and winter
habitat use. Analyses were based on satellite-collared cows from the Bathurst and Ahiak herds, and on the
literature for the Beverly herd. Collar data may under-estimate the caribou’s use of Thaidene Nene as bulls
often winter further south than cows and sub-adults.
Some Bathurst caribou have wintered at least in part in Thaidene Nene in 9 of the past 14 winters, although
in only 1 of the past 4 years after 2005-06. When present, about 52% of collared Bathurst caribou cows
used Thaidene Nene at some time during the winter. In contrast, part of the Ahiak herd has wintered within
Thaidene Nene for 8 of the past 9 years, including every winter since 2004-05. When present,
approximately 66% of collared Ahiak cows used Thaidene Nene. Since 1996, the Bathurst herd has overall
used approximately 82% of Thaidene Nene, concentrated in the south-western portion of the area. The
Ahiak herd utilized all of Thaidene Nene at some part of the past 9 years, with greater concentration of
wintering in the north-eastern two-thirds. The proportion of the annual winter range of the Bathurst herd
that occurred within Thaidene Nene varied from 0-72% ( x = 12%), and for the Ahiak herd from 3-31% ( x =
15%). When caribou from either herd wintered within Thaidene Nene, they were present for roughly 4–5
months, from December to April.
Most crossings of the withdrawal boundary by Bathurst animals occurred north and southeast of Reliance,
with the greatest concentration of boundary crossing north of Reliance over a 30 km stretch on both sides
of treeline. The highest density of individual caribou migration pathways paralleled the treeline lying 20–30
km outside Thaidene Nene’s north-western boundary north of McLeod Bay. Most Ahiak caribou crossed the
northern and eastern portions of the boundary. Thaidene Nene included a larger proportion of the Ahiak
herd’s migration corridors - the eastern half of Thaidene Nene was within moderate to high density
migratory pathways. Movements of the Ahiak herd showed greater overlap (correlation year-to-year)
compared with the Bathurst herd.
During migration and winter areas of burns and lakes were avoided. Selection for terrain ruggedness during
winter differed between herds, with Bathurst caribou showing greater selection for more rugged terrain,
while Ahiak caribou showed equally strong avoidance. Habitat selection during migration was more muted
than during winter. Both herds made high use (but not selection) of hillsides during migration. Selection for
rugged terrain was greater for Bathurst caribou during migration, and was positive for Ahiak caribou during
migration, a reversal from avoidance during winter movement.
The distribution of the Beverly herd was mapped in the 1980s. Caribou tended to move into Thaidene Nene
from the south and southeast in early winter. Their pre-calving migration routes included the Snowdrift
River valley. The Beverly Herd declined between 1994 and 2007 and too few cows were fitted with satellitecollars to support analyses of their distribution relative to Thaidene Nene.
Thaidene Nene appears to largely capture fall and pre-calving movements and winter use of Ahiak caribou.
Recent use of Thaidene Nene by Bathurst caribou has been lower as the herd has declined sharply in size.
As herds decline, the southern boundary of their winter range contracts northward. This suggests that as
the Bathurst herd recovers, increased use of Thaidene Nene is likely. Parks Canada may wish to consider
changes to the boundary of Thaidene Nene to better capture medium to high-density caribou migratory
pathways. Further habitat selection analyses should be explored, and incorporation of Traditional
Knowledge of historic caribou movement and hunting areas should also be considered to complement our
science-based approach.
INTRODUCTION
Migratory tundra caribou (Rangifer tarandus groenlandicus) from 3 herds seasonally use the area
around the East Arm of Great Slave Lake and Artillery Lake, Northwest Territories. Caribou from the
Bathurst, Beverly and Ahiak herds typically leave the tundra in late summer and early fall and migrate to
winter in forested areas. Caribou tend to follow similar migratory pathways, although annual and
longer-term variation does occur. In spring, caribou return to the tundra led by cows on pre-calving
migration to their traditional calving ground. Caribou are an important component of the ecosystem
during their seasonal migrations and while on their wintering ranges in the boreal forest. Through their
foraging and role in nutrient cycling, caribou influence vegetation and also support a guild of predators,
parasites and scavengers. Migratory tundra caribou are fundamentally important in the history and
culture of the region’s aboriginal peoples. The Denesoline have a long history of hunting caribou in the
area, and have considerable information on the distribution of caribou and the timing of their arrival in
the area during the fall (Parlee et al. 2005, SENES Consultants Ltd. 2006).
The East Arm of Great Slave Lake and Artillery Lake area is within the Northwestern Boreal Uplands.
That ecoregion is one of the natural regions within Canada currently unrepresented in the national parks
system (Mondor 2006). Initially proposed in 1970, but delayed for various reasons, work is continuing on
developing a national park reserve in this area (Fig. 1; Mondor 2006).
National parks have a goal to maintain ecological integrity over the long term, which requires that parks
encompass habitat and connectivity needs for viable populations of wildlife. For migratory species such
as caribou and at least some of their predators, park boundaries should include sufficient migratory
corridors within natural ranges of variability. Elsewhere in North America, the role of national parks for
migratory mammals is well-recognized (Berger 2004).
Spatial analyses of caribou movements and distribution in Thaidene Nene (a national park reserve
proposal in the East Arm and Artillery Lake area of the Northwest Territories) would allow Parks Canada
to examine how well the withdrawal area captures fall and pre-calving caribou migratory routes and
winter habitat use. Specifically, such analyses would describe the spatial and temporal representation of
how the migration routes intersect withdrawal boundaries and whether intensively used annual routes
cross or parallel proposed boundaries. Analyses would also allow Parks Canada to examine the annual
variation of migration corridors and winter ranges that lie within the area of interest and whether there
are spatial patterns such as a tendency to use specific areas or landforms such as valleys or lakes.
Thaidene Nene (Fig. 1) is the area withdrawn from development to assess the feasibility of creating a
national park reserve in the East Arm and Artillery Lake region. The withdrawal area represents the area
of interest for Thaidene Nene; final boundaries will be selected through consultation and negotiations
(C. Zimmer, Parks Canada, pers. comm. 2011). Boundary options will be developed once the feasibility
assessment has been completed (C. Zimmer, Parks Canada, pers. comm. 2011).
Caribou groups tend to be smaller in winter although terrain (numbers and size of lakes) plays in a role
in how caribou are dispersed over the landscape, which influences group size. During pre-calving
migration, groups tend to coalesce as they move north. To date, analyses of how well collared cows
represent either a group or the overall seasonal range are few because it requires an independent
measure of distribution. Gunn et al. (In press) reported that during a low coverage aerial survey of the
Bathurst herd in March 2000, most collared cows were within or close to concentrations of caribou, and
3 collared cows were in areas with scattered caribou and many caribou tracks. The 3 eastern
concentrations of caribou were not represented by a satellite-collared cow. A statistical model indicated
that 15–52 collared individuals were necessary to detect 50% probability of representative distribution
in a herd of 350,000–525,000 caribou (Otto et al. 2003).
4
The available data for our analyses were satellite telemetry covering 1996-2010, as well as aerial survey
data periodically available from surveys associated with mapping fall, winter and spring distribution
during sex and age composition surveys. The largest sample sizes of collared cow caribou are from the
Bathurst herd dating back to 1996, with increased sampling of the Ahiak herd since 2005, and a limited
number of collars from the Beverly herd. To assess longer-term use, we reference Zalatan et al.’s (2006)
data on hoof scar frequency (as aged through dendrochronology) to indicate the timescale for caribou
use. We were unable to obtain reference to archaeological material and historic accounts of caribou
migrations as compiled by the Lutsel K’e Dene First Nation.
STUDY AREA
The study area for this project was the proposed Thaidene Nene in the East Arm and Artillery Lake
region of the Northwest Territories (here-after the ‘Thaidene Nene‘; Fig. 1) with an added buffer of out
to ~100 km, to ensure that Thaidene Nene was examined within the larger context as related to
seasonal migrations into and out of the region. The region lies within the Canadian Shield, and consists
of a rolling, rugged upland with a dominance of bedrock outcrops and a large number of lakes and rivers
(Barr and Begrand 1980). Glacial erosion and deposition has influenced the landscape.
Figure 1. Thaidene Nene, Northwest Territories.
5
Black spruce (Picea mariana) forest dominates the vegetation with white spruce (P. glauca), jack pine
(Pinus banksiana) and white birch (Betula papyrifera). The forests are closed-canopy forest of spruce
and jack pine with an understory of moss, lichens, and shrubs. North of Great Slave Lake, the black
spruce forests become more open until the clumps of trees are scattered with an understorey of dwarf
shrubs, lichens and moss. Fire is a strong influence on vegetation within treeline, creating a mosaic of
different aged stands.
The current proposed land withdrawal totals 33,525 km2, approximately 29,560 km2 of land, and 3,965
km2 of waters of the East Arm of Great Slave Lake (Mondor 2006). “Criteria for defining the area of
conservation interest included protection of watersheds, lands used for harvesting, protection of sites of
spiritual and cultural importance, connectivity to the Thelon Game Sanctuary, inclusion of whole
lake/river systems where possible, and inclusion of Bathurst and Beverly Caribou wintering grounds”
(Mondor 2006:11).
METHODS AND MATERIALS
Data acquired
We obtained Bathurst, Ahiak and Beverly caribou collar location data from the GNWT Environment and
Natural Resources (ENR) Wildlife Management Information System (WMIS) covering the period April
1996 to June 2010. All collars were on adult females. No caribou collared by Government of Nunavut
were included in the dataset. Collar data were screened to remove caribou not associated with the
calving grounds of these herds (Appendix A), and we merged the data into a single dataset.
The dataset included satellite collars with a duty cycle that varied over time and among seasons from
daily locations to a location every 7 days. Since November 2008 the dataset also included global
positioning system (GPS) collars with a satellite uplink, which obtained from 3 to 8 locations daily. GPS
collar locations were reduced to daily locations to better match the satellite collar data.
We obtained the following background data for the project from Parks Canada:
Digital elevation model (DEM; 1:250,000 scale with 50 m pixels).
Land cover classification (typed circa 2000; EOSD: Earth Observation for Sustainable
Development of Forests; Wulder et al. 2003)
Land features (lakes, eskers, sand, etc.)
Current Thaidene Nene withdrawal boundary
Hillshading background
Additionally, we obtained fire history, updated to 2009, from ENR. To categorize these fires, we labelled
all fires since 1982 as “recent”, and fires that occurred in 1982 or earlier as “older”.
We did not have access to the detailed Denesoline knowledge of caribou for our report, although we
recognize its complimentary nature and usefulness for understanding the significance of Thaidene Nene
for migratory tundra caribou.
During preparation of the land cover data for analysis, we realised that some classifications appeared
inaccurate. Examples included continuous forest cover in areas well north and east of treeline, and a
stitch line of sudden changes in land cover classification that appears to align with landsat imagery. A
USGS land cover classification (typed circa 2000; Homer et al. 2004) provided apparently more accurate
classification of cover above treeline, but suffered from inaccurate depiction of vegetation cover below
treeline, and a relatively crude/coarse coverage with simplified cover types. We therefore focussed our
6
exploratory examination patterns of habitat use on fire history, land, lakes, topography, and terrain
ruggedness rather than the vegetation or land cover classifications.
Analyses
a) Analyses of telemetry data for caribou migration corridors and distribution
We used several approaches to investigate caribou distribution relative to the proposed boundaries of
Thaidene Nene. Firstly, we used daily movement rates to determine the timing of seasonal use of
Thaidene Nene. This was based on the assumption that caribou during fall, pre-calving migration and
winter use are selecting different habitat attributes that determine their use of an area.
We also examined caribou distribution to describe among-year variation and differences among herds
using Thaidene Nene. For these analyses of caribou distribution and movements relative to Thaidene
Nene, we defined a ‘year’ as starting in August of a given calendar year, and ending in May the following
calendar year.
Daily movement rates: We estimated the daily rate of movement primarily to allow us to determine
timing of the seasonal changes from fall migration to winter and to pre-calving migration. To describe
the daily rate of caribou movements, we calculated distances between consecutive locations using a
program that calculates Euclidian distances between radiolocations while taking into account the Earth’s
curvature (Geomar Consulting Ltd., unpubl. data). The program was written in MapBasic (Version 4.5)
programming language and used MapInfo (Version 6.0) libraries.
The satellite collar data provided locations on various days of any given month; therefore, we grouped
locations into time brackets of 5 days to calculate daily movement rates. Although this decreased
temporal resolution, it allowed us to use all satellite locations in the analysis. We classified records into
5-day time brackets (days 1-5 of a month, days 6-10, etc.); locations separated by >8 days were deleted.
Records split equally between the adjacent 5-day periods were removed from the analysis of mean daily
movements, resulting in removal of <1% of sequential records. We used SPSS and Microsoft Excel to
calculate and chart mean daily movement rates on an annual basis (August-May) for both the Bathurst
and Ahiak herds.
Migration frequency of crossing the Thaidene Nene boundary and migration corridors density: We
summarized the individual pathways of collared caribou to describe the migratory corridors (routes)
relative to the Thaidene Nene boundary and to be able compare migratory corridors among years.
Boundary crossing: To describe how migration routes intersect with Thaidene Nene boundaries, we
calculated the cumulative frequency of crossings of movement trajectories (vectors based on sequential
[consecutive] telemetry of individual caribou) with the Thaidene Nene boundary. Analyses were
conducted on 1-km sections, and summarized by 10-km section. Kilometre 0 was established at Lutsel
K’e, and segmentation was run counter-clockwise. Results were displayed for each herd on a map of the
boundary with frequency of crossing shown by 10-km segment, and descriptively in a table to quantify
relative movement among 5 sections (western, southern, eastern, northern and north-western) and
between herds. Within the table relative crossings within each section of the boundary were calculated
by summing the number of crossings for each section, determining the proportion of crossings within
each section for each herd, and dividing by the number of 10-km segments within that section.
Migration corridor density: To describe the density of animals moving throughout the landscape we
generated annual and cumulative maps of migration corridor density using a moving window technique
to quantify the density of lines joining sequential locations (km/km2). Migration density was calculated
7
using a kernel of 2.5 km that moved horizontally at 1 km step across the study area, then dropped down
1 km and started again throughout the study area (Fig. 2). At each step the program cut a set of
trajectories (lines joining sequential caribou locations) found within a circle of a 2.5 km radius,
calculated total trajectory length, and divided that by the circle's area (19.6 km2), then generated a file
into which the coordinates and the results were written. Annual maps of density were rescaled equally
for all years to account for sample size differences; this allowed comparison among years without bias
towards years with large number of collared animals. The study area for annual analysis was reduced to
a 10 km buffer around Thaidene Nene because of computer limitations in calculation, but was extended
to a 100 km buffer for cumulative analysis.
Figure 2. Kernel step and radius concepts in the moving window technique to calculate the migration
corridor density. For each 1 km step within each circle (called a kernel), the density of
movement trajectories is calculated and assigned to the circle’s central node (blue circle – blue
centroid, green circle, green centroid, etc.).
Comparison of migration corridors between years: For each herd, we re-classified the yearly density
maps were into 3-class density maps (Table 1) and statistically tested the degree of overlap among
movement years using the Kappa Index of Agreement (KIA). The KIA is an index that describes the
similarity between the 2 images, scaling the agreement between 0 (no agreement) and 1 (perfect
overlap). It is calculated using procedures developed from the Chi-square analysis of standard
contingency tables. Further information is available at
http://www.spatialanalysisonline.com/output/html/Crosstabulatedgriddata.html.
Table 1. The density brackets used to develop 3-class density maps for the Bathurst and Ahiak herds
(km/km2).
Herd
Class 1
Class 2
Class 3
Bathurst 0.5 to 2.0 2.0 to 3.5 >3.5
Ahiak
0.5 to 2.5 2.5 to 4.5 >4.5
8
Seasonal range use and high use areas of Thaidene Nene: We defined winter as the period between a
drop in daily movement distances (the end of fall migration) and an increase in movements (the
beginning of pre-calving migration). In most cases movement rates dropped in mid-November to midDecember (although for some years, the Bathurst herd movement rates did not diminish until midJanuary), and increased again in early to mid-April. We developed 90% and 50% (to represent core
areas) fixed kernel polygons of winter ranges. Polygons were calculated using standard settings (href,
volume density, 120 raster resolution) from the Rodgers and Carr (1998) Home Range extension for
ArcView 3.2a. We summarized annual winter distribution for the Bathurst and Ahiak herds in relation to
Thaidene Nene. Years with <5 caribou were not used for winter range mapping because of the
unreliable kernels produced and low representation of herd movements.
We examined shifts in winter range distribution of the Bathurst herd, since this herd was known to have
declined over the past 2 decades (Boulanger et al., in press). We established the 60th parallel as the base
line, and measured the perpendicular distance to the southern edge of the 90% winter range each year.
We applied a linear trend to these data using Microsoft Excel.
Caribou distribution relative to habitat attributes in Thaidene Nene: We examined winter ranges and
migration corridors relative to terrain (topography, ruggedness) and the 2 forest fire age categories.
Topography: We used a geostatistical analysis tool (TOPO analysis; IDRISI raster GIS software) to extract
and categorize terrain features that could influence caribou. We modified the digital elevation model
(DEM) before running the TOPO analysis because the raw DEM data proved was poor (noisy) and the
resolution was coarse (elevation accuracy was 25 m and the positional accuracy was 150 m). In addition,
many individual map sheets differed in elevation along the stitch lines. Thus the DEM data required
time-consuming filtering (Fourier transformations, and use of a Gaussian Frequency Domain Filter with a
cut-off frequency of 46.7; J. Wierzchowski, unpubl. data). The resulting TOPO coverage was of relatively
poor quality, as the attempts to remove significant errors and smooth the “stepped” character from the
DEM were not completely successful.
We used the DEM to generate a description of 11 topographic features (Klimanek 2009), which we
collapsed to 5 categories:
Ridge (ridge and peak): contiguous lines of cells that occur when the cells are higher than the
cells on either side; or a surface location that is higher in elevation than all surrounding cells.
Flat: surface with no slope.
Ravine (ravine, valley, and pit): contiguous lines of cells that occur when the cells are lower than
the cells on either side; or a surface location that is lower than all surrounding cells.
Hillside: convex, concave, saddle, slope and inflection hillsides.
Hillside unknown: unclassified hillsides.
Ruggedness: Caribou likely select their surroundings and migration corridors partly based on the
ruggedness or complexity of the topography. In addition to the TOPO analysis described above, we
estimated ruggedness of the terrain based on elevational change and aspect (Clevenger et al. 2002). We
calculated an index of terrain ruggedness (TR) using the following equation:
TR = (A*B)/W_aver(f*A+B), where
A - Gradient change index
9
B - Aspect variability index
W_aver(f*A+B) - weighted average of A + B
f - user-defined weight: set to 0.5 (no weighting)
The Gradient change captures the elevation gradient within the kernel, calculated using the Relative
Richness formula:
R = n/nmax*100, where:
n = number of different 10-meter elevation classes present in the kernel (350 X 350 meters and
1000 X 1000 meters)
nmax = maximum number of classes in entire image.
The aspect variability index captures the frequency of change of 9 cardinal aspects (0-45°, 45-90°, etc.,
plus Flat). Aspect variability was expressed as the number of different pairs of aspects found within the
applied kernels.
Habitat selection analysis: We compared habitat use to availability to describe winter habitat selection
within Thaidene Nene. We estimated use from the Ahiak and Bathurst satellite collar datasets using only
locations contained within a 50 km buffer around Thaidene Nene and obtained between August and
May the following year. Locations falling on the largest lakes (Great Slave Lake, Artillery) were removed
from the analysis. When sampling the land cover map, the date on which a given location was obtained
was taken into account; if a given caribou location fell within a burn, we developed a program that
compared the date of that burn to the date of the caribou location, and if the location was older than
the burn, a value depicting land was substituted for the date of the burn.
Availability was defined by a generated random sample with sample size equal to that for the Bathurst
and Ahiak herds. The same restrictions as with the “use” points (50 km buffer, no points within large
lakes, burn history compared with the timing of use) were imposed. Separate random samples were
generated for each herd. For analysis of winter habitat selection, only available locations contained
within the cumulative 90% winter range kernel were considered.
To examine habitat selection during migration, we used the output from the density of migration
pathways analysis (see Results). To quantify use for each herd we placed random locations within the
movement density coverage divided into 3 classes, with locations in the highest class weighted by 3, the
middle class weighted by 2, and the lowest class unweighted. Available habitat was determined by
placing random locations in the whole study area (Thaidene Nene plus 10 km buffer; large lakes
removed) equal to the number of use locations.
We attempted a Resource Selection Function analysis but as explained in the Results, we replaced the
approach with univariate analyses. We examined use and selection across variables using Ivlev's
electivity ratio, which scales the use/availability ratio between 1 and –1; 1 being maximum selection and
–1 being maximum avoidance (Krebs 1989:394).
Satellite locations provide accuracy ranging from 350-1,000 m, while GPS collars generally provide <20 m
accuracy. Satellite location quality was not provided for all locations used in this analysis. Given the
spatial resolution of the base data and considering this was an exploratory analysis, we elected to use
the habitat variable at the point location, rather than examine habitat within an arbitrary buffer around
each location.
10
RESULTS
Our analyses were based on 91 satellite-collared cows from the Bathurst herd and 69 cows from the
Ahiak herd. The number of collars varied between herds and among years (Table 2), and increased
especially for Ahiak herd in winter 2007-08. Although a few Beverly caribou have been collared, the
sample size was too low for analyses. Our reliance on satellite-collared cows allowed us to attribute the
distribution and movements to individual herds and to describe in greater detail caribou cow’s use of
Thaidene Nene. However, reliance on the satellite-collared cows has 2 limitations. In some years, the
numbered of collared cows was low and secondly, it may under-estimate the caribou’s use of Thaidene
Nene as caribou bulls often winter further south than cows and sub-adults.
Table 2. Dates and sample sizes for Bathurst and Ahiak collared caribou on their overall winter ranges
and relative to Thaidene Nene (Park). Sample sizes (e.g., 17 (28)) refer to number of collared
caribou within the Thaidene Nene withdrawal boundary and a 10 km buffer of the boundary
and total number of collars for that year.
Bathurst
Year
1996-97
1999-00
Winter range
Park to NW of
Yellowknife
S–SE of Park
Mainly N–NW of
YK; some in Park
Gameti to S of Park
2000-01
Park and to SE
2001-02
N–NE of Yellowknife
2002-03
2003-04
N of Yellowknife
N of YK to Park
2004-05
Park to S of Great
Bear L
Park
1997-98
1998-99
2005-06
2006-07
2007-08
2008-09
2009-10
N of Yellowknife to
Gameti
Great Bear L to
Park
NW of Yellowknife
N of Yellowknife to
Gameti
Ahiak
Winter season
dates
26 Dec–20 Apr
n
5 (10)
20 Dec–15 Apr
5 Jan–25 Apr
6 (7)
7 (24)
??10 Jan–15
Apr
??20 Dec–25
Apr
20 Dec–5 May
9 (14)
Winter range
Winter season
dates
n
21 Nov–15 Apr
3 (4)
26 Dec–15 Apr
Poor
sample/data
Poor
sample/data
1 Dec–10 Apr
2 (10)
0 (2)
26 Nov–5 Apr
15 (23)
11 Dec–10 Apr
25 (27)
26 Dec–10 Apr
1 Jan–5 Apr
30 (48)
17 (28)
13 (13)
5 (12)
2 (12)
6 (11)
N and E of Park to
Sask. border
NE and SE of Park
N–NW of Thelon
Game Sanc.
Park to Sask. border
15 Nov–15 Apr
15 Jan–10
May
~15 Dec–25
Apr
11 Jan–25 Apr
17 (20)
21 Dec–30 Apr
0 (14)
6 Dec–30 Apr
5 (19)
N, in and S and SE of
Park
NW, N, in and S of
Park
Park and E and S
6 Jan–30 Apr
1 Dec–??
0 (22)
0 (9)
Park and N and NE
Park and E and S
2 (15)
1 (2)
7 (8)
Full dataset for Bathurst herd not available post 1 Jan 2010, but no collared caribou were near Thaidene Nene (ENR weekly
maps).
11
Using the collared caribou gives us 3 approaches to describe caribou use of Thaidene Nene. Firstly, the
number of individual collared cows allowed us to assess the relative proportion of the herd (at least
cows and sub-adults) that use Thaidene Nene. Secondly, we estimated the density of migratory
pathways relative to Thaidene Nene boundaries and thirdly, we described how the collared cows are
distributed within Thaidene Nene.
The proportion of the herds that wintered within Thaidene Nene (based on the proportion of the
collared caribou) varied since 1996 (Table 2). Some Bathurst caribou have wintered at least in part in
Thaidene Nene in 9 of 14 winters (Table 2). When present, about 52% of collared Bathurst caribou used
Thaidene Nene at some time during the winter. Bathurst caribou did not use Thaidene Nene in 3 of the
past 4 years after 2005-06, which was when the decline in herd size accelerated (Boulanger et al., in
press). The decline in the Bathurst herd’s use of Thaidene Nene likely reflects a contraction of the
southern boundary of the herd’s winter range as its abundance declined. This is supported by a trend of
increased distance wintering north of the 60th parallel with time (Fig. 3).
In contrast, based on collar data, part of the Ahiak herd has wintered within Thaidene Nene 8 of the past
9 years, including every winter since 2004-05 (Table 2). When present, approximately 66% of collared
Ahiak cows used Thaidene Nene.
Distance from 60̊ N (km)
500
400
300
200
100
0
-100
Figure 3. Distance north of 60 degrees latitude from the southern edge of the 90% winter range for
the Bathurst herd, 1996-2010.
We separated the period of fall or pre-calving migration from winter based on when the caribou slowed
down their daily movement. The timing and duration of the winter season of reduced rate of daily
movements varied annually and between the 2 herds (Fig. 4). The mean length of time of reduced
movements during winter was similar between the Bathurst ( x = 122 days, range 95–151 days, n = 13
years) and Ahiak herds ( x = 119 days, range 94–145 days, n = 7 years). Thus, although location of
caribou winter range varied among years (Table 2), when caribou from either herd wintered within
Thaidene Nene, they were present for roughly 4–5 months, from December to April. In other words,
when caribou movements slowed down at the end of fall migration, if they were inside Thaidene Nene
they tended to stay within Thaidene Nene.
12
20
Ahiak
18
16
14
2001-02
12
Km/day
2002-03
2003-04
10
2004-05
2005-06
8
2006-07
2007-08
2008-09
6
2009-10
4
2
1-5
6-10
11-15
16-20
21-25
26-31
1-5
6-10
11-15
16-20
21-25
26-30
1-5
6-10
11-15
16-20
21-25
26-31
1-5
6-10
11-15
16-20
21-25
26-30
1-5
6-10
11-15
16-20
21-25
26-31
1-5
6-10
11-15
16-20
21-25
26-31
1-5
6-10
11-15
16-20
21-25
26-28
1-5
6-10
11-15
16-20
21-25
26-31
1-5
6-10
11-15
16-20
21-25
26-30
1-5
6-10
11-15
16-20
21-25
26-31
0
August
September
October
November
December
January
February
March
April
May
Figure 4. Daily movement rates of satellite-collared cows A) (upper) Bathurst herd, for the period 1
August – 31 May 1996-2010, and B) (lower) Ahiak herd for 1 August – 31 May 2001-2010.
13
Frequency and location of migration corridors relative to the Thaidene Nene boundary: Our analyses
revealed different patterns of migration through the Thaidene Nene boundary for Bathurst and Ahiak
caribou. Most crossings of the boundary by Bathurst animals occurred north and southeast of Reliance,
with the greatest concentration north of Reliance over a 30 km stretch on both sides of treeline (Fig. 5).
Most Ahiak caribou crossed the northern and eastern portions of the boundary, with the greatest
frequency on the eastern boundary (Fig. 5). Frequency of crossing by collared Ahiak caribou (535
crossings) was greater than for Bathurst caribou (354 crossings). Each herd appears to cross the
boundary in different areas, with greater use of the north-western and southern sections by Bathurst
caribou, and greater use of the eastern and northern sections by Ahiak caribou (Table 3, Fig. 5).
Table 3. Index of relative crossing of the boundary of Thaidene Nene by section by Bathurst and Ahiak
caribou scaled to number of collars (Bathurst n = 354; Ahiak n = 535) and the number of 10-km
segments within each section.
Section of Thaidene Nene boundary
Western
Southern
Eastern
Northern
North-western
Bathurst
0.75
1.07
0.36
0.56
2.15
Ahiak
0.05
0.60
2.13
1.18
0.05
Migration corridors density relative to the Thaidene Nene boundary: Our analyses revealed different
patterns of migration between the Bathurst and Ahiak herds in the vicinity of and within Thaidene Nene.
For the Bathurst herd, the highest density of individual caribou migration pathways (km/km2) paralleled
the treeline lying 20–30 km outside Thaidene Nene’s north-western boundary north of McLeod Bay (Fig.
6). Thaidene Nene boundaries included a larger proportion of the Ahiak herd’s migration corridors - the
eastern half of Thaidene Nene was within moderate to high density migratory pathways. The highest
density concentration straddled the Thaidene Nene boundary in the vicinity of Mary Frances Lake (east
of Artillery Lake), and occurred within Thaidene Nene east and southeast of Artillery Lake.
The density of migration pathways for fall to pre-calving reflected at the broader scale where the
caribou were migrating from during late summer and rut, as well as the calving ground that they were
migrating toward during pre-calving migration. Within those broader scales, the mapped densities
strongly suggested that the large lakes (Great Slave and Artillery lakes) influenced the orientation of the
pathways. The Bathurst herd may also have been influenced by the more rugged terrain along the
northern coast of the East Arm .The influence of terrain and other habitat attributes is addressed
through univarate analyses of caribou locations further on in the Results.
The density of the migration pathways crossing Thaidene Nene’s northern and southern boundary
indicates that the Bathurst herd in some years migrates through Thaidene Nene in comparison to the
Ahiak herd, which largely resides within Thaidene Nene. During the short time period when distribution
data are available for the Ahiak herd (2001-10), Thaidene Nene is at the eastern limits of its winter
distribution at the latitude of Great Slave Lake.
14
Bathurst
Ahiak
Figure 5. Boundary crossing frequencies for the A) (upper) Bathurst herd relative to the boundary of
Thaidene Nene, 1996-2010, and B) (lower) Ahiak herd, 2001-2010.
15
Bathurst
Ahiak
Figure 6. Density of caribou individual migration pathways (km/km2), A) (upper) Bathurst herd, for the
period 1 August – 31 May 1996-2010, and B) (lower) Ahiak herd, for the period 1 August – 31
May 2001-2010.
16
Our KIA analysis indicated that movements of the Ahiak herd showed greater overlap (correlation yearto-year) compared with the Bathurst herd (Table 4). The lack of repetition between the annual density
of migratory pathways for the Bathurst herd, except in 2000 and 2005, may reflect the variable
proportion of the Bathurst herd that wintered on the south-eastern part of its winter range as the
winter range’s southern extent was contracting to the north (Fig. 3). In 2000 and 2005, the migratory
density pathways were at a higher density and were more widely distributed within Thaidene Nene
(Appendix B).
Table 4. Comparison of the annual similarity (k) in the density classes of collared caribou pathways
(km/km2) for the Bathurst and Ahiak herds.
Year
1997
1998
Reference in columns
1999
2000
2003
1997
1.000
0.038
0.173
0.447
0.038
0.310
0.193
1998
1999
2000
2003
2005
2007
0.079
0.199
0.195
0.169
0.129
0.209
1.000
0.055
0.047
0.000
0.045
0.032
0.101
1.000
0.076
0.000
0.112
0.131
0.226
0.198
1.000
0.302
0.318
0.469
0.000
0.000
0.029
1.000
0.062
0.015
0.228
0.310
0.333
0.673
1.000
0.434
0.063
0.139
0.189
0.060
0.167
1.000
Bathurst
Reference
in rows
Ahiak
Reference
in rows
Year
2005
2006
2007
2008
2009
2005
1.000
0.607
0.677
0.651
0.760
Value of κ
≤0.20
0.21 - 0.40
0.41-0.60
0.61-0.80
0.81-1.00
Reference in columns
2006
2007
2008
0.106
0.101
0.102
1.000
0.382
0.737
0.323
1.000
0.335
0.658
0.353
1.000
0.596
0.609
0.597
Strength of
agreement
Poor
Fair
Moderate
Good
Very good
2005
2007
2009
0.138
0.575
0.497
0.513
1.000
17
Winter range distribution within Thaidene Nene: Since 1996, the Bathurst herd has overall used
approximately 82% of Thaidene Nene (as calculated by cumulative 90% winter ranges), with
concentrated use only in some years in the south-western portion of Thaidene Nene (Fig. 7). On the
other hand, the Ahiak herd utilized all of Thaidene Nene at some portions of the past 9 years, with
greater concentration of wintering in the north-eastern two-thirds (Fig. 8).
The proportion of the annual winter range of the Bathurst herd that occurred within Thaidene Nene
varied from 0-72%, and for the Ahiak herd from 3-31% (Table 5). Mean proportions of winter range
within Thaidene Nene was similar between years, but variation among years was greater for the
Bathurst herd. For years with sufficient sample sizes for both herds, overlap in wintering areas were
observed in 3 of 6 years, primarily within the southern half of Thaidene Nene (Table 6).
Table 5. Proportion (%) of annual winter range for the Bathurst and Ahiak herds that occur within
Thaidene Nene, 1996–2010. Ninety percent and Thaidene Nene (In Park) areas in km2.
Year
1996-97
1997-98
1998-99
1999-00
2000-01
2001-02
2002-03
2003-04
2004-05
2005-06
2006-07
2007-08
2008-09
2009-10a
Mean
90% (km2)
90837
111766
46787
127014
105591
50920
104354
74034
104744
25653
45507
176777
29837
Bathurst
In Park (km2)
3981
14405
2001
8687
22346
117
0
17596
6670
18524
0
27118
0
%
4%
13%
4%
7%
21%
0%
0%
24%
6%
72%
0%
15%
0%
0%
12%
90% (km2)
Ahiak
In Park (km2)
%
243714
258996
32594
10931
13%
4%
140019
154072
57586
251021
80940
4746
24187
17905
23380
23381
3%
16%
31%
9%
29%
15%
a
We were unable to build a 90% winter range for the Bathurst herd for 2009-10, but no collared Bathurst caribou were
within Thaidene Nene during that winter (ENR maps).
Table 6. Overlap between wintering ranges of the Bathurst and Ahiak herds, 2002–2010.
Year
2002-03
2005-06
2006-07
2007-08
2008-09
2009-10
Overlap
None
3%; SE of Reliance
None
13%; southern half of Park
None
3%; NE of Yellowknife
Overlap calculated as Overlap/(area of Bathurst + area of Ahiak)
18
Bathurst 90%
Bathurst 50%
Figure 7. Cumulative winter range polygons for the Bathurst herd, 1996-2009; 90% (upper figure) and
50% (lower) fixed kernel polygons.
19
Ahiak 90%
Ahiak 50%
Figure 8. Cumulative winter range polygons for the Ahiak herd, 2001-2009; 90% (upper figure) and
50% (lower) fixed kernel polygons.
20
Habitat selection
Our initial intent was to run a Resource Selection Function analysis to quantify the relationship between
caribou habitat selection and topography, fire history, and land cover classes for caribou winter
locations and movement data. Given inaccuracies with land cover data available, we attempted the
analysis using only broad land cover (water, land, old burns, and recent burns), topographic features,
and terrain ruggedness. The resultant preliminary models using winter locations for the Ahiak herd
(larger sample sizes of the 2 herds) resulted in weak models that fit the data poorly (area under the ROC
curve equals 0.635 [0.50 would result from random fit]) and accounted for very low portion of variation
in the data (R2 = 0.09). These models contributed little to understanding caribou selection of habitat
within the study area. Therefore, we examined selection across these variables as an exploratory first
run at examining this issue.
Winter habitat selection analyses were run for each herd separately (Figs. 9, 10, 11, 12). During winter
both Bathurst and Ahiak caribou avoided water (lakes) and burns of both ages, and showed limited
selection for topographic features (Figs. 9, 11). However, Bathurst caribou showed greater selection for
more rugged terrain at both 350 m and 1 km scales, while Ahiak caribou showed equally strong
avoidance for rugged terrain (Figs. 10, 12).
As expected, habitat selection during migration was more muted than during winter (Figs. 13, 14, 15,
16). Both herds made high use (but not selection) of hillsides during migration (Figs. 13, 15). Selection
for rugged terrain was greater for Bathurst caribou during migration, and was positive for Ahiak caribou
during migration (Figs. 14, 16), a reversal from avoidance during winter and periods of reduced
movement.
21
0.8
1.0
0.7
0.4
0.0
0.3
Ivlev's selectivity index
0.5
Used
Proportion
0.6
Used
Available
Selection
0.2
0.1
0
-1.0
Water
Land
Old Burn
New Burn
Cover type
0.8
1.0
0.7
Proportion
0.6
0.5
0.4
0.0
0.3
Available
Selection
0.2
0.1
0
-1.0
Ridge
Flat
Ravine
Hillside
Hillside:unkwn
Topo type
Figure 9. Bathurst caribou winter selection of broad land cover classes (upper figure) and topographic
features (lower) in the East Arm area, 1996-2010.
22
0.25
1.0
0.15
0.0
0.10
Proportion
0.20
Used
Available
Selection
0.05
0.00
-1.0
Ruggedness index (350 m)
0.25
1
Proportion
0.20
0.15
0
Used
0.10
Available
Selection
0.05
0.00
-1
Ruggedness index (1 km)
Figure 10. Bathurst caribou winter selection of terrain ruggedness at the 350 m (upper figure) and 1
km (lower) scales in the East Arm area, 1996-2010.
23
0.8
1.0
0.7
0.4
0.0
0.3
0.5
Used
Proportion
0.6
Used
Available
Selection
0.2
0.1
0
-1.0
Water
Land
Old Burn
New Burn
Cover type
0.8
1.0
0.7
Proportion
0.6
0.5
0.4
0.0
0.3
Available
Selection
0.2
0.1
0
-1.0
Ridge
Flat
Ravine
Hillside
Hillside:unkwn
Topo type
Figure 11. Ahiak caribou winter selection of broad land cover classes (upper figure) and topographic
features (lower) in the East Arm area, 2001-2010.
24
0.25
1.0
0.15
0.0
0.10
Proportion
0.20
Used
Available
Selection
0.05
0.00
-1.0
0.25
1
Proportion
0.20
0.15
0
Used
0.10
Available
Selection
0.05
0.00
-1
Figure 12. Ahiak caribou winter selection of terrain ruggedness at the 350 m (upper figure) and 1 km
(lower) scales in the East Arm area, 2001-2010.
25
0.8
1.0
0.7
0.4
0.0
0.3
0.5
Used
Proportion
0.6
Used
Available
Selection
0.2
0.1
0
-1.0
Water
Land
Old Burn
New Burn
Cover type
0.7
1.0
0.6
Proportion
0.5
0.4
0.0
0.3
0.2
Available
Selection
0.1
0
-1.0
Ridge
Flat
Ravine
Hillside
Hillside:unkwn
Topo type
Figure 13. Bathurst caribou selection of broad land cover classes (upper figure) and topographic
features (lower) during migration, East Arm area, 1996-2010.
26
0.25
1.0
0.15
0.0
0.10
Proportion
0.20
Used
Available
Selection
0.05
0.00
-1.0
0.25
1
Proportion
0.20
0.15
0
Used
0.10
Available
Selection
0.05
0.00
-1
Figure 14. Bathurst caribou selection of terrain ruggedness at the 350 m (upper figure) and 1 km
(lower) scales during migration in the East Arm area, 1996-2010.
27
0.8
1.0
0.7
0.4
0.0
0.3
0.5
Used
Proportion
0.6
Used
Available
Selection
0.2
0.1
0
-1.0
Water
Land
Old Burn
New Burn
Cover type
0.7
1.0
0.6
Proportion
0.5
0.4
0.0
0.3
0.2
Available
Selection
0.1
0
-1.0
Ridge
Flat
Ravine
Hillside
Hillside:unkwn
Topo type
Figure 15. Ahiak caribou selection of broad land cover classes (upper figure) and topographic features
(lower) during migration, East Arm area, 2001-2010.
28
0.25
1.0
0.15
0.0
0.10
Proportion
0.20
Used
Available
Selection
0.05
0.00
-1.0
0.25
1
Proportion
0.20
0.15
0
Used
0.10
Available
Selection
0.05
0.00
-1
Figure 16. Ahiak caribou selection of terrain ruggedness at the 350 m (upper figure) and 1 km (lower)
scales during migration in the East Arm area, 2001-2010.
29
Discussion
We analysed the distribution of satellite-collared caribou to describe how migratory tundra caribou use
Thaidene Nene. The percentage of the Bathurst and Ahiak herds wintering within Thaidene Nene (based
on the proportion of the collared caribou) annually varies. Our period of analysis started for the Bathurst
herd in 1996 and for the Ahiak herd in 2001. Since 1996 some Bathurst caribou cows wintered at least in
part in Thaidene Nene in 9 of 14 winters with fewer caribou in the most recent years. In contrast, Ahiak
herd cows have wintered within Thaidene Nene during 8 of the past 9 years.
When present, approximately 52% of the collared Bathurst cows compared to 66% of collared Ahiak
cows used Thaidene Nene. Thaidene Nene comprised an average of 12-15% of the annual overall winter
range of both herds, with much higher variability in use by Bathurst animals. We suspect that the
difference between the 2 herds may in part reflect the accelerated decline of the Bathurst herd between
2006 and 2009 (Boulanger et al., in press). The Bathurst herd’s use of Thaidene Nene likely reflects a
contraction of the southern boundary of the herd’s winter range as its abundance declined. Possibly, for
the same reason of declining abundance and contracting southern distribution, we found that spatial
patterns such as a tendency to use specific areas (overlap in annual distribution) varied between the 2
herds.
The duration of when the caribou wintered within Thaidene Nene annually varied and was
approximately 4 to 5 months when present. We separated fall and pre-calving migration from winter
based on when the caribou slowed down their daily movement and we mapped the density of the
migratory pathways. The Bathurst herd showed less intensive use (annual overlap) of migratory
pathways in contrast to the greater annual overlap of the Ahiak herd’s migratory pathways.
How the migration routes during 1996-2010 intersect Thaidene Nene boundaries varies between the 2
herds. For the Bathurst herd, the highest density of individual caribou migration pathways (km/km2)
parallels the treeline lying 10–15 km outside Thaidene Nene’s north-western boundary. Thaidene Nene
boundaries include a larger proportion of the Ahiak herd’s migration corridors - the eastern half of
Thaidene Nene is within the moderate to high density migratory pathways. The highest density
concentration straddles the Thaidene Nene boundary in the vicinity of Mary Frances Lake (east of
Artillery Lake) which also suggests that an extension of the Thaidene Nene boundary toward the
western boundary of the Thelon Game Sanctuary would increase the proportion of high density
migratory pathways within Thaidene Nene. For the timescale of this study for the Ahiak herd, extending
the boundary all the way to the Thelon Game Sanctuary would not add high density migration routes.
However, this is based on analysis of only 9 years of telemetry. Integrating information from Traditional
Knowledge, caribou trails on the landscape and chronology of the frequency of hoof scars on spruce
roots (cf Zalatan et al. 2006) sampled across the area to the Thelon Game Sanctuary boundary would
provide a longer-term basis whether to recommend extension of the Thaidene Nene to the Thelon
Game Sanctuary. Within the Sanctuary caribou migratory routes have been dated to have been used for
8,000 years (Gordon 2005). Providing protection to caribou ranges from Great Slave Lake northeast
through the Thelon Game Sanctuary would be far-sighted leadership. Migratory tundra caribou are the
last of the world’s migratory terrestrial large mammal species with mostly uninterrupted migratory
pathways, and protected areas will be a key to their future conservation (Berger 2004). In this context of
protecting migratory species and connectivity of protected areas, we note that during selection of collar
data for this analysis, we found 2 collared caribou that calved northwest of Wager Bay in the northern
Kivalliq region used Thaidene Nene during the winter of 2004-05, a straight-line distance of >1,000 km.
The movement was unusual in relation to freezing rain and heavy snowfall along the coast of Hudson
Bay in November 2004 (M. Campbell, Government of Nunavut, pers. comm. 2005). This points to
30
interesting landscape-scale linkages between national parks, in this case Ukkusiksalik National Park
around Wager Bay and Thaidene Nene.
We looked at whether the caribou had a tendency to use specific landforms such as valleys or lakes. We
did not detect strong patterns partly because our analyses were limited by the number of collars, and
the resolution of the landform classes was relatively coarse.
While we documented how the extent to which Thaidene Nene represents fall migration, winter and
pre-calving migration distribution of the 2 herds varies among years, we note limitations to conclusions
about caribou use of Thaidene Nene. Our analyses were limited to cows from 2 herds (Bathurst and
Ahiak) and prime-aged bulls often are distributed further south than cows and younger caribou.
Additionally, our analyses were based on data for a relatively short period of time (1996-2010).
The distribution of caribou at the scale of Thaidene Nene is less well-documented before 1996 when
satellite collars were deployed on caribou of the Bathurst herd (Gunn et al. 2001). Previous to 1996,
descriptions of caribou winter ranges came from reports of hunters and pilots, and aerial surveys to map
caribou distribution. Returns of ear-tags of caribou marked at summer river crossings had revealed
overlap in winter ranges for neighbouring herds (Heard 1984). The extent of overlap became more
evident with application of satellite telemetry. We found that for years with sufficient sample sizes for
the Bathurst and Ahiak herds, overlap in wintering areas were observed in 3 of 6 years, primarily within
the southern half of Thaidene Nene. Our results are similar to the patterns observed elsewhere.
Migratory tundra caribou characteristically shift their winter distribution among years, and winter
ranges often overlap between neighbouring herds (Schmelzer and Otto 2003, Bergerud et al. 2008).
Until the use of satellite collars, descriptions of caribou winter distribution from aerial surveys could not
always be attributed to a specific herd unless the surveys were continued into May to track the direction
of pre-calving migration. The surveys were not conducted annually as they were flown for specific
projects and survey coverage of the range also varied.
The Beverly herd was a relatively large herd known for thousands of years of pre-calving migration
across the Thelon River to the calving grounds near Beverly Lake (Gordon 1996). The herd’s summer
ranges included the Thelon Game Sanctuary and the tundra south of the Sanctuary. The generalized fall
pattern was the herd migrating south of the Thelon Game Sanctuary typically across the treeline toward
northern Saskatchewan (see map at http://www.arctic-caribou.com/parttwo/pdf/fr.pdf). The fall
movements and rut generally were east of Great Slave Lake but Thomas et al. (1998) reported that later
in winter, the Beverly herd tended to move west, which brought the caribou into the area of the East
Arm.
Thomas et al. (1998) used aerial surveys to track winter movements of the Beverly herd from 1982 to
1987. The emphasis was on mapping distribution every 4-6 weeks from October to May. The maps show
hand-drawn limits to distribution at a relative coarse-scale (about 3 cm = 100 km) and includes the flight
lines. The information suggests that for 4 of 6 winters between 1980 and 1986, caribou from the Beverly
herd would have been distributed within the area of Thaidene Nene (Table 7). Thomas et al.’s (1998)
survey area was northern Saskatchewan as far east as the Manitoba border, and north into southern
Northwest Territories to the East Arm of Great Slave Lake. Thomas et al. (1998) related the overall
direction and rate of movements to snow conditions. They noted that in most years caribou moved from
east to west across the winter range as snow depths tended to be deeper on the eastern ranges. Snow
depths early in winter when <30 cm had a negligible effect on movements but main concentrations of
caribou later in winter were in areas with <70 cm snow (Thomas et al. 1998).
In summary, in the 1980s the Beverly herd tended during fall migration to cross the treeline and use
areas around Wholdaia Lake before moving northwest to the Snowdrift River valley - Nonacho Lake area
31
approaching Thaidene Nene from the south and southeast in early winter (December)(Fig. 17). The cows
would begin pre-calving migration in late April into early May and their routes included the Snowdrift
River valley heading northeast; major movements also crossed through the Whitefish and Lynx lakes
area (Thomas et al. 1998).
Figure 17. Location of lakes discussed in relation to movements by caribou herds.
Monitoring of the Beverly herd and its distribution was infrequent after the 1980s. Between 1994 and
2009 the herd abundance sharply declined about 99% based on counts of caribou on the traditional
Beverly calving grounds (Adamczewski et al. 2009). The decline is supported by observations of a longterm guide and outfitter on the Thelon River who reported few calves after 2000 and a decline in
caribou sightings by 2005 (A. Hall, pers. comm. 2010). Further supporting evidence is from the frequency
of hoof scars on spruce roots across caribou late summer trails at the treeline, which shows a peak in
the mid-1990s followed by a sharp drop after 2000 (Zalatan et al. 2006).
The Ahiak herd was well-known to the Inuit as it calved and summered along the Queen Maud Gulf
coast (Gavin 1945, Nunavut Atlas 1992). Relatively little monitoring for abundance or distribution was
undertaken except for calving ground surveys in 1986 and 1996 that suggested the herd increasing in
size (Gunn et al. 2000). Little was reported on its fall, winter and pre-calving movements until satellite
telemetry was undertaken in 1996.
32
The first indications of seasonal movements by Ahiak caribou were based on a small number of collars
from 1996 to 1998 (Gunn et al. 2000). The 3 collared cows remained on the tundra during the rut and
wintered in an area from Dubawnt Lake northwest toward Contwoyto Lake, except 1 cow in 1997
moved south to the treeline at Rennie Lake. This movement may have been a precursor toward
wintering within the trees, as in 2001 cows were collared at Nonacho Lake and just north of the
Saskatchewan border. Although those locations had been considered typical of the Beverly winter
range, collared cows of the Bathurst herd (based on calving locations) had migrated south of Great Slave
Lake in fall 2000. Nine more cows were collared in April 2001 between the Saskatchewan border and
Nonacho Lake; 5 went to the Ahiak’s calving grounds east of Bathurst Inlet, 4 went to the Bathurst
calving ground, and 1 cow went to the Beverly calving ground. This was the first evidence for overlap in
winter distribution among those 3 herds and the information also indicated that pre-calving migration
from the Bathurst and Ahiak herd were within the Thaidene Nene area. Subsequent to the winter of
2000-01, collared cows from the Ahiak herd wintered on the tundra in an arc paralleling the treeline
from Contwoyto Lake southeast to Aylmer Lake but the annual number of collared cows at less than 5
was very low.
After 2005, the number of collars increased in the Ahiak herd (Gunn et al., in prep.; GNWT unpubl. data).
The 2006-2010 collar information indicates fall migration which brings the Ahiak cows to rut along the
treeline from Whitefish Lake northwest toward Artillery Lake and the western boundary of the Thelon
Game Sanctuary. This positions Ahiak cows to enter Thaidene Nene from the east and northeast in the
fall, in contrast to the 1980s migrations of the Beverly herd.
The Bathurst herd’s fall and winter range has been monitored through satellite telemetry since 1996.
Prior to then, reconnaissance surveys were flown in late winter to map distribution to determine calf
survival during 1987-95. Williams and Fournier (1996) report on flightlines and areas of caribou
concentration (Table 7). In 2 years, Williams and Fournier (1996) reported caribou concentrations in the
vicinity of the East Arm, for 2 years no sightings, and for the remaining 5 years the flightlines did not
extend to the East Arm. The distribution of the Bathurst herd before 1996 relative to the East Arm area
is that the Bathurst herd did occasionally cross from the north into the area.
There is no overall landscape management plan for the Bathurst, Beverly or Ahiak herds that describes a
role for Thaidene Nene in protecting fall and pre-calving migration routes as well as winter ranges for all
3 herds. Although there is management planning for the Bathurst and Beverly herd (Wek’eezhi
Renewable Resource Board; Beverly and Qamanirjuaq Caribou Management Board), it currently does
not provide direction to integrate different approaches (levels) of habitat protection and land use
planning, especially mineral exploration and development activities. Thaidene Nene would be a
particularly significant component as it provides protection to part of the annual ranges for 3 herds.
The extent that the Thaidene Nene would play a role in protecting the herd’s habitats will change with
the phase of their abundance cycle. As abundance increases, caribou extend their wintering areas
further south, which would increase use of Thaidene Nene for foraging and travelling during migration
and as wintering range. Our objectives for this report did not include to address the role of Thaidene
Nene in the wider context of landscape planning for those herds, although it is an important question
given that in 7 of 14 years for the Bathurst herd and for 7 of 9 years the Ahiak herd, half or more of the
collared cows wintered within Thaidene Nene (Table 2).
Our analyses have demonstrated the value of the telemetry data both for demonstrating the timing and
areas used within Thaidene Nene, but also for establishing on an annual basis the use relative to the
herd’s overall ranges. Telemetry, while a useful part of a monitoring project, is only useful if it includes
not only data collection but also prompt analysis relative to the design of the monitoring program.
33
Sample size relative to the objective for the application of the data is a key part. Also, the telemetry data
need to be integrated with environmental conditions especially as there are strong trends in
environmental conditions across the Bathurst annual range (Gunn and Poole 2010). Two determinants
of winter range use are forest fires and winter conditions – the rate and timing of fall to winter snow
conditions including wind speed and temperature as well as snowfall.
Table 7. Summary of winter distribution for the Beverly herd (Thomas et al. 1998) and for the Bathurst
herd (Williams and Fournier 1996), 1980-1995.
Winter
Beverly
1980 Jan
Concentrations along 60°N south of Nonacho Lake
1981-82
Concentrations along 60°N south of Nonacho Lake
1982 Dec
Mid Oct Wholdaia Lake,– north to Snowdrift River
1983 Jan Mar
Lutsel K’e north Snowdrift River inland
1983 Dec
Lutsel K’e north Snowdrift River across East Arm
and to south end Artillery Lake; moved in from
east (Beaverhill Dubawnt Lakes)
1984 Jan-Mar
Nonacho to Gordon Lake – all of East Arm Park
but included overlap with Bathurst as May
migration split to Bathurst Inlet and Beverly Lake
1984 Oct-Nov
Ennadai to Rennie Lake then west to Wholdaia
Lake south of 62°N
1985 March
Caribou along Snowdrift River and Nonacho Lake
from northwest
1985 Dec
Caribou moving south across Snowdrift River
1986 Jan - Mar
Snowdrift River to Nonacho Lake and Lutsel K’e
1986 Dec
Caribou moving southeast from north of East Arm
to Nonacho Lake
Bathurst
April No recorded concentrations
1987 Feb-Mar
High densities north of East Arm
1988 Mar
Scattered groups south of Artillery Lake
1989 Mar
No flight lines East Arm
1990 Feb - Mar
No recorded sightings
1991 Feb - Mar
No recorded sightings
1992 Mar
1993 Feb - Mar
1994 Mar
1995 Feb - Mar
34
Recommendations
Thaidene Nene appears to largely capture fall and pre-calving movements and winter use of Ahiak
caribou. Recent use of Thaidene Nene by Bathurst caribou has been lower, but at increased population
levels may be greater in the future. The highest density of migration movements by Bathurst caribou
since 1996 has occurred 20-30 km north of the withdrawal boundary north of McLeod Bay (Fig. 6), but it
may not be feasible for Thaidene Nene to expand to capture these areas. We make the following
recommendations for consideration by Parks Canada:
1. To capture medium to high density migratory pathways of the Ahiak herd, Parks Canada should
consider extending the north-eastern boundary of Thaidene Nene by approximately 25 km (Fig. 18). This
would add approximately 2,100 km2 to Thaidene Nene (6%); smaller portions could be selected,
ensuring that the northern, high-density portion of that area is targeted.
2. Further examination of patterns of habitat selection by tundra caribou inhabiting Thaidene Nene
could be explored using a land cover classification and dataset that is most appropriate and accurate for
the area. The land cover classification used in the Taltson Hydro EIS could be examined.
3. Parks Canada should incorporate Traditional Knowledge of historic caribou movement and hunting
areas into their background documentation and decision-making for Thaidene Nene. We understand
that those data may be available and would complement our science-based approach. Additionally, to
assess longer-term use, other approaches are useful such hoof scar frequency (as aged through
dendrochronology; Zalatan et al. 2006) and archaeological.
Ahiak
Figure 18. Suggested addition (red polygon) to Thaidene Nene to capture moderate to high density
migration areas used by the Ahiak herd.
35
Literature cited
Adamczewski, J., J. Boulanger, B. Croft, D. Cluff, B. Elkin, J. Nishi, A. Kelly, A. D’Hont, and C. Nicolson.
2009. Decline in the Bathurst caribou herd 2006-2009: a technical evaluation of field data and
modeling. 17 Dec 2009 draft, Environment and Natural Resources, Yellowknife, NT.
Barr, W., and R. Begrand. 1980. Regional Analysis of the Northwestern Boreal Uplands (Natural Region
17). Institute of Northern Studies, University of Saskatchewan, Saskatoon.
Berger, J. 2004 The longest mile: how to sustain long distance migration in mammals. Conservation
Biology 18:320–332.
Bergerud, A.T., S.N. Luttich, and L. Camps. 2008. The return of the caribou to Ungava. McGill-Queen’s
University Press, Montreal, Quebec. 586 pp.
Boulanger, J., A. Gunn, J. Adamczweski, and B. Croft. In Press. Exploration of the decline of the Bathurst
caribou herd using a data-driven demographic model. Journal of Wildlife Management.
Clevenger, A.P., J. Wierzchowski, B. Chruszcz, and K. Gunson. 2002. GIS-generated expert based models
for identifying wildlife habitat linkages and mitigation passage planning. Conservation Biology
16:503–514.
Gavin, A. 1945. Notes on mammals observed in the Perry River District, Queen Maud Sea. Journal of
Mammalogy 26:226–230.
Gordon, B. 1996. People of sunlight; people of starlight: barrenland archaeology in the Northwest
Territories of Canada. Archaeological Survey of Canada, Mercury Series Paper 154. Canadian
Museum of Civilization. 219 pp.
Gordon, B.C. 2005. 8000 years of caribou and human seasonal migration in the Canadian Barrenlands.
Rangifer, Special Issue No. 16:155–162.
Gunn, A., and K.G. Poole. 2010. Environmental trends across the range of the Bathurst caribou herd and
timing of the arrival of cows on their calving ground 1996–2009. Unpublished report for Northwest
Territories Department of Environment and Natural Resources, Yellowknife, NWT
Gunn, A., B. Fournier, and J. Nishi. 2000. Abundance and distribution of the Queen Maud Gulf caribou
herd, 1986-98. Northwest Territories Department of Resources, Wildlife and Economic Development.
File Report No. 126. 76 pp.
Gunn, A., J. Dragon, and J. Boulanger. 2001. Seasonal movements of satellite-collared caribou from the
Bathurst herd. Final Report to the West Kitikmeot Slave Study Society, Yellowknife, NWT. 80 pp.
Gunn, A. In Prep. Satellite collaring and calf survival in the Bathurst herd of barren-ground caribou 2003–
2005. Department of Environment and Natural Resources, Yellowknife, NT. Manuscript Report.
Gunn, A., A. D’Hont, J. Williams, and J. Boulanger. In Press. Satellite-collaring in the Bathurst herd of
barren-ground caribou 1996-2005. Department of Environment and Natural Resources Manuscript
Report.
Heard, D. C. 1984. Hunting patterns and the distribution of the Beverly, Bathurst and Kaminuriak caribou
herds based on the return of tags by hunters. Northwest Territories. Wildlife Service. Yellowknife, NT.
NWT Wildlife Service. File Report 49.
Homer, C., C. Huang, L. Yang, B. Wylie, and M. Coan. 2004. Development of a 2001 National landcover
database for the United States. Photogrammetric Engineering and Remote Sensing 70:829–840.
36
Klimanek, I.M. 2009. Geoinformation support of derived mapping based on digital terrain model.
Mendel University of Agriculture and Forestry Brno, Czech Republic, Faculty of Forestry and Wood
Technology, Department of Geoinformation Technologies.
Krebs, C.J. 1989. Ecological methodology. Harper and Row, New York, USA.
Mondor, C. 2006. Preliminary area of interest for a national park in the East Arm of Great Slave Lake,
Northwest Territories. Parks Canada.
Nunavut Atlas. 1992. Edited by R. Riewe. Published by the Canadian Circumpolar Institute, Edmonton
and the Tungavik Federation of Nunavut. 259 pp.
Otto, R.D., N. P. P. Simon, S. Couturier, and I. Schmelzer. 2003. Evaluation of satellite collar sample size
requirements for mitigation of low-level military jet disturbance of the George River caribou herd.
Proceedings of the Ninth North American Caribou Workshop, Kuujjuaq, Quebec, Rangifer Special
Issue 14:297–302.
Parlee, B., M. Manseau, and Lutsel K’e Dene First Nation. 2005. Using traditional knowledge to adapt to
ecological change: Denésoliné monitoring of caribou movements. Arctic 58:26–37.
Rodgers, A.R., and A.P. Carr. 1998. HRE: The Home Range Extension for ArcView. Ontario Ministry of
Natural Resources, Centre for Northern Forest Ecosystem Research, Thunder Bay, Ontario.
Schmelzer, I., and R. Otto. 2003. Winter range drift in the George River Caribou Herd: a response to
summer forage limitation? Rangifer, Special Issue No. 14:113–122.
SENES Consultants Ltd. 2006. Thaydene Nene state of knowledge report. Lutsel K’e Dene First Nation,
Lutsel K’e, Northwest Territories.
Thomas, D.C., H.P.L. Kiliaan, and T.W.P. Trottier. 1998. Fire-caribou relationships (III) Movement
patterns of the Beverly herd in relation to burns and snow. Canadian Wildlife Service Technical
Report Series No. 311. 176 pp.
Williams, T.M., and B. Fournier. 1996. Summary of spring classification surveys of the Bathurst caribou
herd 1985-1995. Manuscript report No. 92. Dept. of Resources, Wildlife, and Economic Development,
Yellowknife, NT.
Wulder, M.A., J.A. Dechka, M.A. Gillis, J.E. Luther, R.J. Hall, A. Beaudoin, and S.E. Franklin. 2003.
Operational mapping of the land cover of the forested area of Canada with Landsat data: EOSD land
cover program. Forestry Chronicle 79:1075–1083.
Zalatan, R., A. Gunn, and G.H.R. Henry. 2006. Long-term abundance patterns of barren-ground caribou
using trampling scars on roots of Picea mariana in the Northwest Territories, Canada. Arctic,
Antarctic, and Alpine Research 38:624–630.
37
Appendix A: Caribou collar dataset management
Pulled together Yellowknife and Ft. Smith collar datasets from ENR.
Removed Nunavut collars (BLxxxxx and Wager Bay)
Removed collars that calved in the Bluenose-East calving grounds
Requested from ENR the latest collars to Jun 2010 for this project, and received only those collar
locations within 100 km of the boundary.
Note: A few AH collars from 1997-99 from Yellowknife dataset (AG) were not included, as they did not
go near the park.
Beverly collar 138 removed (not near boundary).
Note 317 and 357 calved both east of Chantrey Inlet; left in as AH.
Collars that switched CGs (BA to BNE, or BA-AH) were labelled as where calved (or removed for BNE)
Caribou ID
175
190
201
203
138
117
167
182
192
209
210
211
214
<20 locations
<20 locations
<20 locations
<20 locations
Beverly collar 2001-04
One year in BN-E
One year in BN-E
One year in BN-E
One year in BN-E
BN-E
BN-E
BN-E
BN-E
Deleted
Deleted
Deleted
Deleted
Deleted, not near boundary
Removed BNE locations
Deleted
Deleted
Deleted
Deleted
38
Appendix B: Density of caribou individual migration pathways (km/km2) by year, Bathurst and Ahiak
herds, Thaidene Nene. N = how many different collars contributed to the particular density map, and
the total is the overall number of individual collars in that particular movement year (August to May).
39
Bathurst
40
Ahiak
41
Ahiak

Thaidene Nene caribou report: January 2011

Transcription

Similar documents

Brochure FR 2009.qxd - Explo

to print this coupon

Imraan`s Article

ANWR FLoat - Ouzel Expeditions, Inc.

7000 Head Herd Dispersal

e-parentline - Sacred Hearts Academy

Musselwhite Mine Environmental Working Committee

Home on the Tundra

Water Quality and Biology of Streams Draining Abandoned and