Blanding’s turtle (Emydoidea blandingii) select populations COSEWIC assessment and status report 2016: appendix 1

Appendix 1. The decline in relative abundance of Blanding’s Turtles (emydoidea blandingii) in Ontario attributable to historic wetland conversion

A report in support of the COSEWIC Update Status Report and Status Assessment in progress.

AU: Matthew G. Keevil

Submitted to Teresa Piraino

16 September 2015.

Purpose and scope

The purpose of this analysis is to provide an estimate of the relative decline of Blanding’s Turtles across Ontario caused by habitat loss over three intervals between ~1800 and 2002. Proportional declines are calculated as wetland loss across ecoregions weighted by density estimates obtained from available mark-recapture data sets within each region. Other causes of abundance decline such as fragmentation by roads, upland habitat loss, and subsidized predation, which collectively cause declines or extirpation within habitats, were not considered.

Methods

The provincial scale proportional change in abundance due to habitat loss was calculated from the following equation which quantifies density-weighted habitat loss:

Equation quantifying density-weighted habitat loss for the Blanding’s Turtle.
Density-weighted habitat loss formula
Long description for Figure 1

Equation quantifying density-weighted habitat loss for the Blanding’s Turtle.

Where DR is mean density estimated from study sites in ecoregion R, QRt is the quantity of habitat in ecoregion R at time t, and is the ratio of habitat area in R at times t and t - 1. R represents one of three ecoregions: Lake Erie-Lake Ontario (Ecoregion 7E), Lake Simcoe-Rideau (6E), and Georgian Bay (5E) (OMNRF 2007). This model breaks the Ontario distribution of Blanding’s Turtles into three regions in order to accommodate available data on ecoregional differences in population density, historical wetland loss, and total wetland area. Spatial analyses were performed using ArcGIS 9.3.

Ratio of habitat area  formula

Range extent of Blanding’s turtles in Ontario

The range extent of Blanding’s Turtles was digitized based on records reported in the Ontario Reptile and Amphibian Atlas (Ontario Nature 2015). Blanding’s Turtles are distributed across much of southern and central Ontario except for Bruce and Grey counties and parts of southeastern Ontario. A limitation of these distributional data is that historic range extent may be underestimated if Blanding’s Turtles were extirpated before records were collected. All subsequent analyses are based on spatial data clipped to the Blanding’s Turtle range extent. A proportion of the northwestern limit of the Blanding’s Turtle distribution extended into the Lake Temagami Ecoregion (4E) and this was merged into the Georgian Bay Ecoregion for this analysis.

Figure 1. Ecoregions (OMNRF 2007) clipped to the distribution of Blanding’s Turtles in Ontario digitized from Ontario Reptile and Amphibian Atlas records.
Ecoregions (OMNRF 2007) clipped to the distribution of Blanding’s Turtles in Ontario
Photo: © Ontario Nature 2015
Long description for Figure 1

Map of ecoregions clipped to the distribution of the Blanding’s Turtle in Ontario. Carte des écorégions rattachées à l’aire de répartition de la tortue mouchetée en Ontario.

Estimating wetland conversion as an indicator of relative habitat loss (ht/ht-1)

Wetland conversion has been estimated between four time periods (~1800, 1967, 1982, 2002) for geographic townships corresponding to the Mixwood Plains Ecozone (Lake Simcoe-Rideau and Lake Erie-Lake Ontario Ecoregions) and presented in the Southern Ontario Wetland Conversion Analysis Final Report (Ducks Unlimited Canada 2010). These data were used to estimate total wetland conversion within the Blanding’s Turtle distribution for these two ecoregions. Township wetland conversion statistics were allocated to ecoregions based on centroid location from pre-amalgamation 1977 geographic townships (OMNRF 2013). Wetland area conversion estimates (as absolute areas) were then summed across ecoregions in over each time period. Estimates of wetland conversion were not available for most municipalities outside the Lake Simcoe-Rideau and Lake Erie-Lake Ontario ecoregions. Therefore, habitat loss in the Georgian Bay Ecoregion was not quantified in this analysis.

Figure 2. Wetland loss by township across the Lake Ontario/Lake Erie and Lake Simcoe/Rideau ecoregions within the range of Blanding’s Turtles in Ontario from pre-settlement (~1800) to 2002. Wetland loss data from Southern Ontario Wetland Conversion Analysis Final Report.
Wetland loss by township across the Lake Ontario/Lake Erie and Lake Simcoe/Rideau ecoregions within the range of Blanding’s Turtles in Ontario
Photo: © Ducks Unlimited Canada 2010
Long description for Figure 2

Map illustrating wetland loss by township across the Lake Ontario-Lake Erie and Lake Simcoe-Rideau ecoregions within the range of the Blanding’s Turtle in Ontario from pre-settlement (about 1800) to 2002.

Estimating relative habitat quantity (q)

Total potential habitat was quantified in order to account for relative differences between ecoregions due to differences in total land area and proportion of wetlands. These habitat area estimates are not intended to quantify the absolute amount of suitable habitat for Blanding’s Turtles. Wetland area was estimated using land cover layers instead of tabulated from the Southern Ontario Wetland Conversion Analysis (Ducks Unlimited Canada 2010) because the latter was unavailable for the Georgian Bay ecoregion. Two land cover data sources were used: SOLRIS (OMNRF 2008) in southern Ontario and the Provincial Land Cover Database (Spectranalysis Inc. 2004). Where these datasets spatially overlap, SOLRIS data were used. Habitat quantity was calculated as the total area of all wetland land cover classes (Ontario Land Cover classes 15-23; SOLRIS classes 50, 55, 59, and 63) within the Blanding’s Turtle range extent across each ecoregion.

Table 1. Area of wetland land cover classes within the range extent of Blanding’s Turtles in Ontario broken down by ecoregion. 4086 km 2 from the Lake Temagami Ecoregion that was within the range extent of Blanding’s Turtles was merged with the Georgian Bay Ecoregion before analysis.
Ecoregion Wetland area (km2)
SOLRIS
Wetland area (km2)
Prov. Land Cover
Wetland area (km2)
total
Geographic area (km2)
Simcoe/Rideau 7417 51 7468 50947
L. Erie/L. Ont. 1428 0 1428 21899
Georgian Bay 377 1685 2062 68478

Estimating relative population density (d)

Mark-recapture abundance estimates and study area data were available from eight study sites in Ontario (Appendix 1 Table), two in the L.Erie/L. Ontario Ecoregion, five in the L. Simcoe/Rideau Ecoregion, and one in the Georgian Bay Ecoregion. Mean density estimates were 0.78, 0.29, and 0.12 adults/ha respectively. Differences in historic densities between ecoregions may have been greater than for recent densities because populations in the more productive southern parts of Ontario are also subjected to greater disturbances such as road mortality and higher levels of subsidized nest predation. This could have caused estimated historic relative abundance to be lower than true values where wetland loss has been most extensive. In turn, this would cause an underestimation of total declines.

Data limitations:

  • Great Lakes coastal wetland losses were not available in the wetland conversion analysis (Ducks Unlimited Canada 2010).
  • Wetlands smaller than 10 ha were not available from the wetland conversion analysis (Ducks Unlimited Canada 2010).
  • Land cover and wetland conversion data were only available up to 2002.
  • Wetland conversion data is only available for the Mixedwood Plains Ecozone portion of the Ontario distribution of Blanding’s Turtles.
  • Density estimates are based on a small number of sites. Recent relative densities may be biased indicators of historic relative densities.

Results

Table 2. Decline in density-weighted Blanding’s Turtle habitat across the range extent in Ontario. Recent wetland land cover reflects data gathered up to 2000-2002 in SOLRIS (OMNRF 2008) and The Provincial Land Cover Dataset (Spectranalysis Inc. 2004). Proportional wetland loss and density weighted declines are reported relative to the start and end of corresponding time intervals. Therefore, the loss of wetlands in the Simcoe/Rideau portion of the Blanding’s Turtle range between 1967 and 2002 is 3.4% of the wetlands that remained in 1967. An assumption of no wetland loss in the Georgian Bay Ecoregion was necessary because of a lack of data on wetland conversion.
Ecoregion Density (adults/ha) Recent wetland land cover (ha) Wetland loss between 2002 and 1800 Wetland loss between 2002 and 1967 Wetland loss between 2002 and 1982 Provincial decline in density-weighted wetlands between 2002 and 1800 Provincial decline in density-weighted wetlands between 2002 and 1967 Provincial decline in density-weighted wetlands between 2002 and 1982
L.Erie/L.Ont. 0.78 142800 92.1% 34.3% 26.2% 65% 13% 11%
Simcoe/Rideau 0.29 746800 58.5% 3.4% 4.5% blank blank blank
Georgian Bay 0.12 206200 0.0% 0.0% 0.0% blank blank blank
Table 3. Trends in the distribution of density-weighted habitat of Blanding’s Turtles across ecoregions and time periods scaled as a proportion of total remaining wetlands. Data sources and assumptions are as described for Table 2.
Ecoregion 1800 1967 1982 2002
L.Erie/L.Ont. 72.2% 40.7% 37.7% 31.8%
Lake Simcoe/Rideau 26.6% 53.3% 56.1% 61.1%
Georgian Bay 1.3% 6.0% 6.2% 7.1%
Figure 3. Stacked area plot of the estimated proportions of density-weighted Blanding’s Turtle habitat in Ontario partitioned by ecoregion and plotted against year (See Tables 2, 3).
Stacked area plot of the estimated proportions of density-weighted Blanding’s Turtle habitat in Ontario
Long description for Figure 3

Chart (stacked area plot) illustrating the estimated proportions of density-weighted Blanding’s Turtle habitat in Ontario partitioned by ecoregion and plotted against year.

The generation time of Blanding’s Turtles in Ontario has been estimated as >40 years (COSEWIC 2005). Times when habitat losses were quantified correspond to 5.05 (1800), 0.875 (1967), 0.5 (1982), and 0 (2002) generations before the most recent measurement year or 5.375, 1.2, 0.825, and 0.325 generations before present (2015). This analysis estimates a minimum decline in Blanding’s Turtle abundance of 13% within the past 1.2 generations due to wetland loss in addition to unmeasured habitat losses after 2002 and probable within-habitat declines over all periods caused by road mortality, subsidized predation, poaching, upland habitat loss and other chronic threats.

A large gap in measured wetland loss occurs between the pre-settlement period and 1967. If a constant rate of wetland loss occurred during this interval is provisionally assumed and calculated separately for each ecoregion, then the proportion of density weighted habitat would have been twice as great in 1936 (~1.98 generations before present) as it was in 2002. This corresponds to a minimum 50% decline across Ontario within the last two Blanding’s Turtle generations. However, an approximately constant rate of habitat loss during this interval is unlikely and therefore this extrapolation must be interpreted with extreme caution.

References

COSEWIC 2005. COSEWIC assessment and update status report on the Blanding’s Turtle Emydoidea blandingii in Canada. Committee on the Status of Endangered Wildlife in Canada. Ottawa.

Ducks Unlimited Canada. 2010. Southern Ontario Wetland Conversion Analysis Final Report. Ducks Unlimited. Barrie, ON.

OMNRF. 2013. Geographic Township Boundaries. Ontario Ministry of Natural Resources and Forests. Available as spatial data from Land Information Ontario

OMNRF. 2008. Southern Ontario Land Resource Information System (2000-2002). v.1.2. Ontario Ministry of Natural Resources and Forests. Available as spatial data from Land Information Ontario.

OMNRF. 2007. Ecological Land Classification of Ontario. Ontario Ministry of Natural Resources and Forests. Available as spatial data from Land Information Ontario.

Ontario Nature. 2015. Ontario Reptile and Amphibian Atlas Accessed 2015/08/08

Spectranalysis Inc. 2004. Introduction to the Ontario Land Cover Data Base, Second Edition (2000): Outline of Production Methodology and Description of 27 Land Cover Classes. Report to Ontario Ministry of Natural Resources. Available as spatial data from Land Information Ontario.

Appendix 1 Table. Blanding’s Turtle adult densities from mark-recapture studies in Ontario (provided by Teresa Piraino).
Subpopulation Ecoregion Size of study site (ha) Adult population estimate Density (adults/ha) Study period (# of seasons) Sampling effort Survey methods Sources
SW Ontario 1 Erie/Ontario 3300 690
Schnabel Method (modified closed-capture model)
0.21 9 seasons (2000-2001; 2008-2014) >650 person-days Hand/dip net captures while conducting visual surveys in wetlands and nesting areas Gillingwater and Brooks 2001; Davy unpub. data
SW Ontario 2 Erie/Ontario 607 818
based on an estimated 341 ±214 adult females (Jolly-Seber method in program JOLLY using model A and mark-recapture data from 2003-2006) and an average sex ratio of 1.4M:1F.
1.35 21 seasons (1973; 1979; 1980; 1982; 1992-1994; 2003-2016) >680 person-days;
~500 trap days
Hand/dip net captures while conducting visual surveys in wetlands and nesting areas; hoop net traps Weller 1973; Hubbs 1979; Purves 1980; Ashenden 1983; Saumure 1995; Gillingwater and Piraino 2004, 2007; Piraino and Gillingwater 2005, 2006; Gillingwater 2009, 2013; Enneson 2010
SW Ontario 3 Simcoe 800 138
Lincoln Index where N=MC/R (using data from 2010-2011)
0.17 5 seasons (2010-2014) 1317 person-hours;
2200 trap days
Hand/dip net captures while conducting radio-telemetry surveys, road mortality surveys and visual surveys in wetlands and nesting areas; hoop net traps Caverhill et al. 2011; Toronto Zoo unpub. data
SE Ontario 2 Simcoe 690 99
(95% CI: 89-124)
0.14 4 seasons (2010-2013) 5300 person-hours;
2360 trap days
Hand/dip net captures while conducting radio-telemetry surveys, road mortality surveys and visual surveys in wetlands; hoop net traps Dillon Consulting Ltd. 2014
SE Ontario 3 Simcoe 900 114
(95% CI: 103-136)
Closed capture model in MARK
0.13 3 seasons (2007-2009) Wetlands surveyed every day from April-Sept every season. Hoop traps were also set all season. Hand/dip net captures while conducting radio-telemetry surveys and visual surveys in wetlands; hoop net traps Millar 2009, unpub. data; Millar and Blouin-Demers 2012
SE Ontario 4 Simcoe 238 85
(95% CI: 53-206)
Schnabel Method (modified closed-capture model)
0.36 5 seasons (2010-2014) ~68 person-days;
~54 traps days
Hand/dip net captures while conducting visual surveys in wetlands; hoop net and basking traps Middleton 2014; Ontario Nature unpub. data
SC Ontario 1 Georgian Bay 340 41 (95% CI: 39-50) 0.12 5 seasons (2006-2008; 2009-2010) Wetlands surveyed several days between April-May every season. Nightly nest site patrols from 7-11pm for 3-4 weeks/season. Several incidental captures during telemetry and at communal hibernacula. Hand/dip net captures while conducting visual surveys in wetlands and nesting areas Edge et al. 2009, 2010, unpub. data; Paterson et al. 2014, unpub. data
SC Ontario 3 Simcoe 90 57
Lincoln Peterson
N=n1*n2/m2
0.63 2 seasons (2013-2014) 134 person-hours (2013);
?? person-hours (2014)
Hand/dip net captures while conducting visual surveys in wetlands Sheppard 2013, 2014, unpub. data
Report a problem or mistake on this page
Please select all that apply:

Thank you for your help!

You will not receive a reply. For enquiries, contact us.

Date modified: