Blanding's turtle COSEWIC assessment and status report: chapter 9
Population sizes and trends
There were 1908 sightings reported to the Ontario Herpetofaunal Summary (OHS) database from 1881 to June 6, 2002 (Ontario Herpetofaunal Summary 2004). In Québec, a total of 38 sightings is recorded in NatureServe and the Centre de données sur le patrimoine naturel du Québec (2005), and a total of 100 sightings have been reported to the QAAR (Québec Atlas of Amphibians and Reptiles data bank 2005). In Ontario and Québec, there is very little other published information on the Blanding’s Turtle; however, there is ongoing research being conducted by Bob Johnson (Toronto Zoo) on four very small remnant populations in the Toronto area; by Scott Gillingwater (Upper Thames River Conservation Authority) on populations in Rondeau P.P. and at Big Creek N.W.A.; by Ben Porchuk (Wilds of Pelee Island) on Pelee Island, and in Québec by Daniel St-Hilaire (Société de la faune et des parcs du Québec). Other biologists throughout Ontario and Québec have reported occurrences within their jurisdictions (e.g. Browne 2003).
Most sightings of Blanding’s Turtles are of adults, and not juveniles or hatchlings (Ontario Herpetofaunal Summary 2004; NatureServe 2004; Québec Atlas of Amphibians and Reptiles data bank 2005), which is a concern for long-term population stability (Congdon and van Loben Sels 1991; Power et al. 1994; Heppell et al. 1996; Morrison 1996; Congdon et al. 2001; Browne 2003). However, it seems that infrequent observations of hatchlings have been the case as long as people have been studying this species, so it is not clear if this rarity is characteristic of abnormally low recruitment or of a typical stable population.
The only detailed, long-term ecological study of Blanding’s Turtle in Canada is being conducted on the Nova Scotia populations. The Kejimkujik N.P. population has been studied since 1969, with intensive work occurring since 1987. In 1996, researchers began looking for Blanding’s Turtles outside the park by soliciting the public to report sightings and by systematically trapping and surveying new areas for turtles. The result was the identification of two additional populations: McGowan Lake (intensively studied since 1996) and Pleasant River (intensively studied since 2002). This research has focused on both adult and juvenile distribution, habitat use, and demography; individuals in all age groups have been marked and tracked over time (Tom Herman, Jennifer McNeil, pers. comm. Jan 24, 2005).
In the United States, a long-term study has been conducted at the E.S. George Reserve (part of the University of Michigan) in southeast Michigan, beginning in 1954, and currently being continued by J.D. Congdon (see references). In this report, we have assumed that the Ontario/Québec Blanding’s Turtles have life-history traits within the range expressed at the Nova Scotia and Michigan study sites. In general, the more northern turtles of Nova Scotia have later maturity and lower annual reproductive output than turtles in southeast Michigan. These differences are probably a consequence of shorter and cooler active seasons in populations existing at higher latitudes.
It is difficult to estimate the abundance of Blanding’s Turtles in Ontario/Québec, as there has been very little work on abundance or population trends in this region. In the OHS database, there were 1248 sightings of Blanding’s Turtles from 1984 to 1994, and from 1995 to 2002 there were 163 Blanding’s Turtle sightings reported. This accounts for 10.3% and 8.9% of all Turtle sightings reported during those time periods respectively (Michael J. Oldham, pers. comm. Oct. 13, 2004). It is probable that the decrease in the number of sightings does not correspond to a reduction in the population, but is simply a reduction in the total number of turtle sightings being reported. In Quebec, the Blanding’s Turtle is quite rare, and the populations are isolated. It has been reported that in the Gatineau Park, densities may be as low as less than one per km2 (McMurray 1984).
An overall estimate of the number of adult Blanding’s Turtles in the Great Lakes/St. Lawrence population necessarily must be crude. It doesn’t appear that any population exceeds 1000 mature individuals, and although some exceed 100 adults and the Big Creek N.W.A. may have 600 adults, most populations are much smaller. The great majority of reports of Blanding’s Turtles to the O.H.S. are of fewer than 5 individuals. There are approximately 150 Element Occurrences (EO) in Ontario (Austen and Oldham 2001) and many of these sites are small and have few adults observed (Austen and Oldham 2001). To achieve a total population estimate of 10 000 adults for the Great Lakes/St. Lawrence population requires an average of 65 adults per EO. This average seems high given the observations in the O.H.S. database. Therefore, a maximum Great Lakes/St. Lawrence population estimate of 10 000 adults is not unreasonable. This seems like a substantial number, but given the life history of the species, as described elsewhere throughout this report, these numbers may represent primarily older cohorts that are declining from increased mortality and very low recruitment.
In a study in 2001-2002, Browne (2003) captured and marked 95 Blanding’s Turtles in Point Pelee National Park. She concluded that a larger mean body size in her turtles compared to the mean size in an earlier study (Rivard and Smith 1973 cited in Browne 2003) meant that the mean age of the population was older in 2003. However, the long-term study at ESGR in Michigan found no support for the notion that adult body size in this species correlates with age (Congdon and Van Loben Sels 1991; Congdon et al. 1993, 2001). Browne (2003) also concluded that observed rates of mortality of adult Blanding’s Turtles on the roads in and around Point Pelee N.P. could cause population declines. She reported nest predation at 70%, and concluded that this rate of loss would also lead to declines in the Blanding’s Turtle populations (Browne 2003). Using a model (Ramas simulation) and admittedly limited data, Browne found that if one extra (beyond natural mortality) adult female is killed by a vehicle every two years, and if nest mortality is >32% annually, the population would slowly decline to extinction (Browne 2003, p. 72-74).
In Big Creek N.W.A., 429 adult Blanding’s turtles have been individually marked (Scott Gillingwater, unpublished data). This population is by far the largest documented in Canada, and most others are likely much smaller. The Big Creek population has been noted by Saumure (1997) to be male-dominated. A Z-score comparing the observed ratio of males (55.5%) (Gillingwater unpublished data) to the expected sex ratio of 1:1, indicates that this population is significantly male-dominated (p<0.05). A finite population correction factor was calculated using the upper confidence interval (N=1326) for the population size estimated by Saumure (1997) and gave a significant bias from the expected 1:1 male to female ratio (p<0.05). A significant difference from the expected 1:1 sex ratio is present even if the finite population correction factor is calculated with a total population size of 5000 (p<0.05) (Chris Edge pers. comm.) . A male-biased population could be the result of road mortality affecting nesting females more than their male counterparts, as this species often nests on the gravel shoulders of roadways (Saumure 1995 , 1997; Standing et al. 1999 ). Females likely suffer higher road mortality than do males at Big Creek N.W.A., which has a major highway through the wetland where many turtles are killed each year (Ashley and Robinson 1996).
Recent studies on Snapping Turtles (Chelydra serpentina) in the USA have concluded that this species and Painted Turtles (Chrysemys picta) develop male-biased sex ratios and skewed age (adult biased) distributions toward adults (Marchand and Litvaitis 2004; Gibbs and Shriver 2002; Steen and Gibbs 2004; Tucker and Lamer 2004). In one paper, collecting of females nesting on the roadside plus female-biased mortality were cited as the cause of male-biased sex ratios (80-85% male) (Tucker and Lamer 2004). This latter study was conducted on Snapping Turtles that were being taken for food or eggs, but these results would apply to Blanding’s turtles being collected for the pet trade and being killed by vehicles.
In Nova Scotia, approximately 250 individual adult Blanding’s Turtles have been encountered since 1969 (Herman et al. 2003).Initial population size estimates calculated for turtles in the Kejimkujik subpopulation, using data from 1969 to 1988, resulted in an estimate of 132 adults (Herman et al. 1995). However, these estimates were based on limited capture-mark-recapture data. Recently, Jolly-Seber estimates based on more extensive and more long-term data indicate that the number of adults in Kejimkujik N.P. is only about 66 (Tom Herman, pers. comm. E-mails April 28, 30, 2005). The subpopulation at McGowan Lake is estimated to contain 79 adults (95% CI: 59.9-116.5), based on capture-mark-recapture data from 1996 to 2001 (McNeil 2002). No population estimates have been calculated yet for the Pleasant River sub-population; however, this sub-population is believed to be the largest in the province. Sixty-five adults have been marked in this population; 57 of these have been marked in three years of intensive sampling (Caverhill 2003; Caverhill in progress; as cited by Tom Herman and Jennifer McNeil, pers. comm. Jan. 24, 2005). In the most recent estimates, the total population of adults in Nova Scotia is: Kejimkujik N.P.= 66; McGowan Lake=79; and Pleasant River=65-100 =210-245. (Tom Herman, pers. comm. April 28, 2005) (Nova Scotia information supplied by Tom Herman and Jennifer McNeil, pers. comm. Jan. 24, 2005).
Congdon et al. (1993) used Euler’s equation to predict the survivorship of juveniles and hatchlings necessary to sustain a population at the University of Michigan’s E.S. George Reserve (ESGR). The calculations were based on data collected on Blanding’s Turtle populations from 1975 to 1986, and also in 1991. These data indicate that from 1976-1984 the mean annual nest survivorship was 43.8%, with the value falling to 3.3% during 1985 and 1991, giving an overall mean of 26% annually. From the same data, mean annual adult survivorship was reported at 96%. According to Euler’s equation, a stable population requires the annual juvenile survivorship to be in excess of 72%. This calculation is set with an age of maturity of 14 years, the lowest speculated for the Blanding’s Turtle at ESGR.
The age of maturity for Blanding’s Turtles in Ontario/Québec should be considered to be closer to 20 years, and possibly exceeding 25 years for the northern portion of the species’ range (Ron Brooks, pers. comm.). In the northern portion of the species’ range, turtles experience shorter active seasons and cooler temperatures, which is indicative of a later age of maturity. If the age of maturity increases to 20 years (from 14 years), the necessary annual juvenile survivorship, predicted by Congdon et al. (1993), is increased to 85%. If age at maturity increases to 25 years, as appears possible in Nova Scotia and likely for Ontario/Québec, then the annual juvenile survivorship needs to approach 90% to maintain the population at a stable level.
Average annual nest survivorship in Ontario could be as low as 5% for areas such as Rondeau P.P. (Scott Gillingwater, pers. comm. Feb 16, 2005). The average annual nest survivorship for the Great Lakes/St. Lawrence population is most likely in the range of <1% for metropolitan areas to 15% for a pristine environment, with an overall average of 3-4% (Congdon et al. 1993, Herman et al. 2003). The Euler’s equation predicts that a drop in annual nest survivorship to 5% corresponds with a necessary increase in annual juvenile survivorship of 10%, for the population to remain stable.
More closely representing the Great Lakes / St. Lawrence population would be a reduction in nest survivorship to 15% for a pristine environment. Congdon et al. (2000) investigated the percentage of eggs to produce viable hatchlings; this value was estimated at 17.6% for the population on the ESGR in Michigan. This could be balanced by increasing adult survivorship by 1.5%, to 97.5%, which would then increase the generation time to 40 years. Adult survivorship is the most difficult to increase of all the life stages, so this does not appear to be a viable option to stabilize the population. Embryo survivorship to hatching of 17.6% may be an overestimate for the northern portion of the Blanding’s Turtles range. Eggs in the northern portion of the species range are more susceptible to environmental changes, due to a decrease in the length of the active season as a result of cooler temperatures.
A long-term study of Painted Turtles in Algonquin Park, Ontario, has estimated annual survivorship of adults at 98-99%, and of juveniles greater than 5 years of age at 90-95% (Samson 2003). The Blanding’s Turtle is not nearly as common or widespread as the Painted Turtle. The reasons for these differences are not known, but it is possible that the extreme delay in age at maturity in Blanding’s Turtles (Painted Turtles mature at 5-14 years in Ontario, maturing earlier as one goes south) is a significant reason for its lower abundance. Blanding’s Turtles also reproduce less often (less than one clutch per year) than the Painted Turtle (1-2 clutches per year).
In Nova Scotia, a recent population viability analysis identified an alarming decline in the Kejimkujik N.P. subpopulation (Herman et al. 2004). A deterministic stage based matrix was constructed using the following average annual survivorships calculated from life-history data collected from the population: adult 94% (confidence interval 85%-100%), large juvenile (10-18.49cm) 89%, small juvenile (5-9.99cm carapace length) 69%, hatchling 12%, and egg 60% (based on the current program of screening nests against predators). Despite the apparently high adult and juvenile survivorships, the model indicated that without intervention, the Kejimkujik N.P. population would continue to decline. Although the model is most susceptible to changes in adult mortality, it is difficult to increase survivorship in this life stage. Modeling the effect of different management regimes (one year headstarting, two year headstarting, laboratory incubation of eggs) indicated that enhancing the survival of early life-stages also has the capacity to stabilize the Kejimkujik N.P. population (Herman et al. 2004) (Nova Scotia information supplied by Tom Herman and Jennifer McNeil, pers. comm. Jan. 24, 2005).
In the Greater Chicago Metropolitan Area (GCMA), Ruben et al. (2001) examined populations that had been separated by urban sprawl for evidence of genetic differentiation or loss. They compared the differentiation in the GCMA population to other populations that were not physically separated, such as the ESGR and Kejimkujik N.P. populations. Results indicated no genetic differentiation among Chicago populations, but significant loss of variability compared to the population on the ESGR, which is considered to be panmictic, and with the Kejimkujik N.P. population, which is not panmictic (Mockford et al. 2005).
In Ontario, there is little potential for rescue effect except perhaps along the upper St. Lawrence River, in the Thousand Islands area. However, there is no evidence that turtles are crossing over in this region and indeed it appears that extreme eastern Ontario is one of the areas from which Blanding’s Turtles have been extirpated, or were never present (Figs.2, 3; Ontario Herpetofaunal Summary 2004; Michael Oldham, pers. comm. Oct. 13, 2004). A similar situation exists in southwest Ontario along the St. Clair and Detroit Rivers; a rescue effect might be possible, but again there is no evidence that the Blanding’s Turtle lives on the Ontario side of the St. Clair River. Blanding’s Turtles do occur on the Canadian shorelines of Lake St. Clair and the Detroit River, so it is possible that turtles could enter Canada in these regions. However, it seems more likely that any successful migration would be in the opposite direction. Essentially, there seems to be no potential for rescue effect and certainly no evidence offering support. Along the Ottawa River, exchange of individuals between Ontario and Québec populations could possibly be limited due to the increase in the breadth of the Ottawa River due to significant damning for hydro electricity. Populations do however exist on both sides of the Ottawa River, and in order to support this hypothesis of a significant barrier, more information is needed.
Genetic evidence from the NS population indicates significant spatial structure among the three known subpopulations, with no evidence of recent bottlenecks. Estimates of gene flow are very low (1.8 – 5.8 individuals per generation), despite proximity (15-25 km) of the three subpopulation centers (Mockford et al. 2005) (Nova Scotia information supplied by Tom Herman and Jennifer McNeil, pers. comm. Jan. 24, 2005).
Page details
- Date modified: