COSEWIC assessment and update status report on the woodland caribou Rangifer tarandus caribou in Canada
- Assessment Summary
- Executive Summary
- COSEWIC Mandate, Membership and Definitions
- Lists of Figures, Tables and Appendices
- Species Information
- Population Sizes and Trends
- Limiting Factors and Threats
- Special Significance of the Populations
- Existing Protection or Other Status
- Summary of Status Report
- Technical Summary
- Acknowledgements and Biographical Summary of Authors
- Authorities Consulted, Collections Examined and Literature Cited
Population Sizes and Trends
Accurate (estimate vs. true number) and precise (degree of confidence) estimates are exceedingly difficult to obtain for forest-dwelling caribou. The proportion visible relative to those hidden by trees is difficult to estimate. Observed age and sex ratios may be biased because visible caribou are not representative of the population (Hatler 1986). In alpine/tundra and upper subalpine cover types, minimum estimates are obtained from attempts at total counts. Ratios of observed-to-expected radio collared or painted caribou are used to account for caribou missed on surveys. In forests and peatlands, sample surveys are conducted where standard errors and confidence limits of estimates can be calculated. Confidence limits at 80% probability may be adequate for caribou numbers. Many estimates are guesses based on extent of known distribution, density estimates, occasional sightings, and track counts.
Several problems confound any analysis of caribou numbers and trends. When estimates are obtained by sampling procedures, the confidence limits are wide even at 90% probability. For example, standard errors for 11 surveys in Yukon have averaged 16.5% (R. Farnell pers. comm. 2000), which equates to a 90% confidence interval of about 28%. In insular Newfoundland, the 90% confidence interval averaged 58%, 29%, and 19% of the estimate obtained respectively by strip transect (n = 5), stratified random block (n = 5), and mark-resight survey designs (n = 6) (Mahoney 2000).
Accuracy remains unknown unless sightability indices are estimated for each survey or search intensity is high using a helicopter over relatively open canopies. Even when caribou are visible on the tundra, many counts are inaccurate and imprecise (Thomas 1998). For example, the 90% confidence interval for caribou population estimates in the NWT and Quebec averaged 57% (n = 9) and 32% (n = 5), respectively. Regular counting of all woodland caribou populations is cost prohibitive.
Williams and Heard (1986) summarized the status of 32 of 43 local populations of woodland caribou in the mountains and plains. Trends were 7 increasing, 16 stable, and 9 decreasing (Table 2). Those totals excluded a declining Gaspésie population and insular Newfoundland, where all 11 local populations were increasing. A few years later, reviews appeared on the status of woodland caribou in western Canada (Edmonds 1991) and North America (Ferguson and Gauthier 1992). The latter tallied 71 populations of forest-dwelling woodland caribou and trends for 18 of 46 located in mountains and plains (same exclusions as above) were 5 increasing, 8 stable, and 5 decreasing.
In 1996, Mallory and Hillis (1998) concluded that "populations of woodland or forest ecotypes were declining and threatened throughout the circumpolar region, possibly due to the interaction of human disturbance and predation." In reviews by provinces/territory in 1996, the estimated status of 25 local populations of forest-dwelling woodland caribou in mountain ranges was 3 increasing, 8 stable, and 7 decreasing (Table 2). Corresponding ratios in 2000/2 are 4:15:3, 0:13:12, and 1:6:12 for the Northern Mountain, Southern Mountain, and Boreal populations, respectively.
Short-term changes in numbers can vary considerably due to weather and predators. Consequently, the IUCN has adopted 10 years or three generations, whichever is longer, as a time frame to evaluate changes in numbers. A generation is defined as the average age of parents (IUCN 1994, COSEWIC 2000b) but should be the average or median age of females of breeding age. It can be obtained from a life table but none exist for forest-dwelling caribou. The most reliable life tables are for female caribou in the George River and Beverly populations of migratory, forest-tundra caribou (Messier et al. 1988, Thomas and Barry 1990). Survival of 50% of females >2 years old was just over 7 and 5 years, respectively. Therefore, three generations of caribou is rounded to 20 years and one generation is therefore 6.7 years. A 20-year span should be adopted as a standard for all populations.
There is need for standard criteria of what constitutes an increase, stability, and a decrease in local populations. It is difficult to suggest criteria because it depends on the accuracy and precision of data and the interval between estimates. For periods shorter than 20 years, stable could be defined as an average annual change in numbers of less than 2%. The term decline should be reserved for a decrease in numbers of 20% or more over 20 years. For periods shorter than 20 years, a decrease should not be inferred on a prorated average of 1% (rounded) change per year. A temporal scaling is required. An average decrease of more than 3% per year over 10 years or 10% per year over 5 years may be suitable where data are reliable. Caribou populations fluctuate widely in numbers because recruitment can be low for several years and relatively high in following years. Caribou populations can grow at up to 11-15% per year and decrease faster.
Northern Mountain Population (NMP)
The 2001 estimate for the NMP is 44 000 (Table 1). It accounts for about 24% of all forest-dwelling caribou in Canada. Stratified random-quadrat designs and total counts account for 61% of estimates. All populations contain more than 100 caribou and 20 of 36 contain more than 500 (Table 4). Most (72%) of the estimates were made after 1996. Trends in numbers were 4 increasing, 15 stable, 3 decreasing, and 14 unknown (Table 3). Fluctuations were large in both directions, with Ibex up 63% in 8 years and Chisana down 78% in 10 years. Range sizes > 5000 km2 predominate (Table 5). Densities averaged 11.3 caribou per 100 km2 and range from 3.0 to 26.9 (Table 6).
In 1991, the Yukon contained about 19% of the Canadian woodland caribou population as defined by Ferguson and Gauthier (1992). Inventory and long-term monitoring of 22 populations suggested that most were about stable (Farnell et al. 1998). Trends in numbers were eight stable, four increasing, one decreasing, and one stable or decreasing. The apparent increase from 21 000 in the early 1990s (Ferguson and Gauthier 1992) to 28 000-35 000 in 1997 (Farnell et al. 1998) and to 43 150-48 150 in 2001 stems mostly from improved survey methods and better estimates than from growth in numbers. Three populations contained only 180-200 caribou. Trend in numbers in 2001 were 4 increasing, 9 stable, and 2 decreasing (Appendix 1a).
Five populations of woodland caribou in the Mackenzie Mountains of the NWT are shared with and tallied by the Yukon: Bonnet Plume, Redstone, Nahanni, Clear Creek (new), and La Biche (Farnell et al. 1998, Gullickson 2000, Farnell pers. comm. 2001).
In B.C., woodland caribou in the Northern Mountain NEA occur in 16 populations with an estimated total of about 11 000, unchanged from 1996 (Table 2). Seven local populations are about stable, one is increasing, one is declining, and seven unknowns. In 1996, seven were stable, one was increasing, and two were decreasing (Heard and Vagt 1998).
Southern Mountain Population (SMP)
Numbers in the SMP are estimated at 7200 (Table 1). This COSEWIC population accounts for about 3.9% of all forest-dwelling caribou in Canada. Counts extrapolated from marked individuals and total-count surveys account for 77% of estimates. Twenty-eight of 30 local populations number fewer than 500 caribou, with eight comprised of 50 caribou or fewer (Appendix 1b, Table 4). Most (77%) of the estimates were made since 2000. Trends in numbers for 25 of 30 local populations are 0 increasing, 13 about stable, and 12 decreasing (Table 3). For 90% of populations, the surveyors expressed high (37%) or moderate (53%) confidence in the trend. Nineteen (63%) of the caribou ranges were relatively small at <5000 km2 (Table 5). Densities average 8.3, 5.9, and 3.0 caribou per 100 km2 in the west central, north central, and southern metapopulations in B.C. and ranged from 4.9-16.4 per 100 km2 in Alberta (Table 6). However, six small local populations (<51 caribou each) in the southern metapopulation in B.C. occur at densities of only 0.3-2.3 caribou per 100 km.2 The Southern Purcells population may become extinct within 10 years (Kinley and Apps 2001). Only four adult females were found in 2000. Its peculiar male-dominated sex ratio may be caused by mountain lion (Puma concolor) predation in fragmented habitats. The Banff population also is reduced to a few caribou.
From 1997 to 2002, the average annual rate of decline of SMP caribou in B.C. was 2.47%. That rate would result in 39.3% decline over 20 years (I. Hatter, unpubl. data, 2002).
Boreal Population (BP)
The estimated number of forest-dwelling caribou is 33 000 (Table 1), comprising 18% of the total for Canada. However, numbers and trends for most forest-dwelling populations are poorly known over most of the Boreal NEA. For example, the NWT estimate of 4000-6400 was based on estimated extent of occurrence and a density of 2 caribou/100 km2 in one area intensively surveyed and estimated densities of 1 and 3 caribou per 100 km2 in other areas (A. Gunn pers. comm. 2001). The previous estimate in 1992 was 2000-5000 woodland caribou (Ferguson and Gauthier 1992). In northeastern B.C., 725 caribou are estimated in the Boreal NEA (I. Hatter pers. comm. 2000), a density of 1.4 per 100 km2. The previous estimate in 1996 was 750 (Heard and Vagt 1998).
In Alberta, an estimated 3285 forest-dwelling caribou occur in 12 local populations in the Boreal Plain ecozone. Most populations are decreasing (Table 3) based on data for female mortality and calf:cow ratios (Dzus 2001). In 1996, numbers appeared to be about stable or slightly decreasing (Stuart-Smith et al. 1997, Edmonds 1998). Stability occurs when there is replacement of adult females that die by an equal number of female yearlings. Whether data from radio-collared caribou are representative of the populations is not clear. The growth curves are sensitive to an overestimation of mortality of adult females or underestimation of survival of female calves. Survival of adult females could be underestimated because of age structure differences between sampled caribou and the population. Calf survival could be underestimated if some calves have separated from radiocollared cows by the time of surveys in March. Sex ratio equality cannot be assumed as male calves generally have higher mortality than females. Some male calves were unaccounted for in another study (Stuart-Smith et al. 1997). All but 3 of 28 local populations contain fewer than 500 caribou (Table 4). Most local populations occupy ranges larger than 5000 km2 (Table 5). Thus, average densities are only 3.3 caribou per 100 km2 and range from 1.8 to 13.1 (Table 6).
Woodland caribou populations apparently decreased south of the Shield in Saskatchewan after pulp harvesting operations began in the mid-1960s (Trottier 1988a, Rock 1992). They continue to decrease slowly based on recruitment and mortality data and reduced area of occupation (Rettie et al. 1998). Former estimates of 2500 caribou (Ferguson and Gauthier 1992, Rettie et al. 1998) are now revised upwards to 5000, a result of surveys in the Boreal Shield and a larger mapped extent of occurrence (Godwin and Thorpe 2000). Mean densities in favourable habitat in two ecoregions of the Boreal Shield are estimated at 3.5 and 3.0 per 100 km2 compared with 2.8 on the Boreal Plain (Godwin and Thorpe 2000). Numbers include large areas that were not surveyed and where estimated densities of 0.7/100 km2 and 0.3/100 km2 were guesses.
Thirteen local populations in Manitoba were estimated to contain 1840 to 3125 caribou, rounded to 2000 and 2500 by Rebizant et al. (2000). No estimate is available for the Nelson-Hayes rivers population, which overlaps with the Pen Island population (Abraham and Thompson 1998). The Pen Island and Cape Churchill populations are the forest-tundra ecotype and are excluded from this review. Caribou numbers are likely to decrease in local populations at Kississing-Naosap Lakes, Wabowden, Atikaki-Berens, and Owl-Flintstone Lakes unless strategies are developed to reduce impacts of development (R. Larche pers. comm. 1997). Densities average 1.1 to 1.8 caribou per 100 km2 and range from 0.5 to 4.3 (Table 6).
Improved estimates for caribou numbers along the Hudson Plain (Hudson and James Bay lowlands) increased the estimate for caribou in Ontario to over 20 000 in 1996 (Cumming 1998). Exclusion of the forest-tundra ecotype reduces the number to about 5000 of the forest-dwelling ecotype relevant to this review (Harris 1999). There are about 500 caribou in remnant populations south of the line of semi-continuous distribution: Slate Islands, Pic Island, Pukaskwa National Park, Caramat, Flanders Township, and Hagarty Road (Darby et al. 1989, Cumming 1998). Small numbers of caribou were translocated in the early 1980s from the Slate Islands to Michipicoten, Montreal, and Bowman islands (Darby et al. 1989). Caribou persisted on the first two islands to 1989 and all but one disappeared from Bowman Island (Bergerud and Mercer 1989).
In 1990, the total population in Ontario was estimated at 11 000, excluding the Pen Island population of 4 000 caribou that ranged mostly in the taiga (Abraham and Thompson 1998). The boreal population was 6000-6700 in the 1970s and 1980s excluding 4800 in the Pen Island population and 3500-5600 in northeastern Ontario (Ferguson and Gauthier 1992). However, the northeastern total may contain some boreal caribou. Many estimates are essentially guesses.
In the boreal forest of Quebec south of 490N, there are two isolated, sedentary caribou populations, the Val D'Or (40-90 individuals) and the Grands Jardins. The introduced Grands Jardins population has fluctuated from a low of 38 animals in 1978 to a high of 126 in 1992 (Banville 1998) and 103 in 1998. Farther north, between 490N and 550N, there are several sedentary populations totalling less than 10 000 caribou (Couturier 1996). They are located in the northern boreal forest and southern taiga. Perhaps the best known is the Lac Joseph population, estimated to number 1025 in year 2000. This population, shared with Labrador, is subject to range loss and fragmentation (R. Otto pers. comm. 2000). Densities in Labrador average 1.3 caribou per 100 km2 and range from 0.4 to 2.1 (Table 6). The status of other local populations is poorly known. Since about 1981, in winter thousands of caribou from the expanding Leaf River and George River populations have invaded ranges of sedentary caribou (Brown et al. 1986, Messier et al. 1988, Couturier 1996) and disrupted them (Schaefer et al. 1999).
Since about 1980, local populations in Labrador have declined in number by 80% (Red Wine Mountains and Lac Joseph) and 75% (Mealy Mountains) (R. Otto pers. comm. 2000). Causes were overhunting, predation, or both (Mahoney and Schaefer 1996). However, the Lac Joseph population now is increasing (Appendix 1c). The Red Wine population, which occurs in the taiga, declined presumably from wolf predation and losses when members migrated with George River caribou (Schaefer et al. 1999).
Newfoundland Population (NP)
Populations of woodland caribou on Newfoundland continue to grow and the total population in 2001 is estimated at 100 000 (Table 1), up from about 25 600 in 1979 (Bergerud 1980) (Table 2). Newfoundland caribou account for 54% of forest-dwelling caribou in Canada. Numbers were estimated for 15 natural populations and 12 introduced ones (Appendix 1d). There are another 10 introduced populations whose status is not known. The population declined sharply from 40 000 in 1900 to 1 000-2 000 in 1930 (Bergerud 1971). That decrease may have been caused by introduction of a parasite with reindeer (Ball et al. 2001). Densities now average 150 caribou per 100 km2 and range from 11 to 634 (Table 6).
Atlantic (Gaspésie) Population
This small, isolated population declined in number from 500-1000 in the 1950s to about 200 in the 1970s (Crête et al. 1994). It then stabilized at about 200 in 1991 (Ferguson and Gauthier 1992), and 200 to 250 caribou in 1993 and 1996 (RENEW 1994, Crête et al. 1994, Boileau 1996, Couturier 1996). In 1992 and 1993, both subpopulations in the Gaspésie Conservation Park contained over 30 calves per 100 females (Crête et al. 1994). Densities are 20-25 caribou per 100 km2 (Table 6). A recent report suggests that the population is decreasing (Fournier 2001).
- Date Modified: