Telemetry

Twenty-three female and 12 male Blanding's turtles were captured while on land and in aquatic habitats using baited traps and by hand. Turtles were transported to the laboratory, where we assigned individual identification codes and recorded their straight-line carapace length, age, sex, weight, reproductive status, and date and capture location. Individuals with a carapace length of 165 mm or greater were considered adults. We used fast-drying binary epoxy putty to attach radio transmitters (Advanced Telemetry Systems, Isanti, MN; radios and mounting material weighed ∼ 65 g) midway between the dorsal line and marginal scutes of the carapace and released turtles at or near (within ∼ 25 m) their point of capture. Turtles were typically located 1–2 times a week during the active season (from April to June through late October 1999) and once per month after ice formed on wetland surfaces (from early November 1999 through March 2000). Locations were obtained by triangulation from the closest areas feasible to radioed turtles with 3-element, directional antennas and handheld receivers manufactured by Advanced Telemetry Systems or observation in an attempt to minimize error associated with obtaining locations via compass azimuth bearing triangulation. Initially, each turtle's locations were plotted onto tracings of US Geological Survey Digital Orthophoto Quarter Quadrangle maps in the field. Locations were later converted to X/Y coordinates in ArcView GIS (Environmental Systems Research Institute Inc [ESRI] 1999) and plotted on remotely sensed land cover data representing Navigation Pool 5 and its surrounding area (Fig. 1).

Movements and Home Ranges

Movements and home ranges of 16 females and 8 males with a minimum of 25 radio locations recorded during the activity season and at least 2 locations during the overwintering period were evaluated. Data were analyzed using ArcView GIS with the Minnesota Department of Natural Resources ArcView tools, Xtools, Spatial Analyst, and Animal Movement Analysis extensions (Hooge and Eichenlaub 1997; ESRI 1999). Aquatic activity areas were delineated as areas within a wetland where turtles remained (within 75 m for 10 or more days during the active season). Movements were categorized as those 1) made between aquatic activity areas, 2) made immediately prior to and after nesting (females only), and 3) made to and from overwintering sites (after the first of September and prior to the first of May, respectively; Piepgras and Lang 2000).

We calculated home range size estimates in hectares using 2 frequently used methods to assess home ranges: the minimum convex polygon (MCP; Mohr 1947; Hooge and Eichenlaub 1997) and the bivariate normal density kernel with a 95% probability (BNK; Silverman 1986; Worton 1989; Hooge and Eichenlaub 1997). While the MCP and BNK tend to overestimate home range sizes by incorporating unused terrestrial habitats located between aquatic activity areas (Jennrich and Turner 1969; Anderson 1982; Worton 1989, 1995; Naef-Daenzer 1993; Piepgras 1998; Seaman et al. 1999; Piepgras and Lang 2000), we used them for comparing our results to those of other Blanding's turtle studies. Our field observations suggested that home range size estimates at Weaver Dunes were exaggerated using the MCP and BNK methods; therefore, we calculated home range with a third method, the poly-buff (PB) method, which combines the areas within minimum convex polygons delineated around each individual's aquatic locations with the areas of 40-m-wide buffers overlaid on top of the track of each observed terrestrial movement path (20 m on each side; Hamernick 2001). The PB method is similar to the cluster analysis method (Edmonds 1998; Carter et al. 1999) and the grid-summation method utilized by Piepgras and Lang (2000) in that they attempt to reduce the exaggerated overestimation of home range sizes, eliminating tracts of terrestrial and aquatic areas that were not used during movements. We also calculated home range length (the maximum straight-line distance a turtle could move within the estimated home range) in meters by measuring the straight-line distance between the 2 furthest locations we recorded for each of the 24 Blanding's turtle adults.

Turtle movements and home ranges were compared with the statistical software packages SPSS (Version 9; SPSS 1999) and SAS (Version 7; SAS Institute 1998). We used Mann-Whitney U-test and Wilcoxon 2-sample tests to compare male and female movement frequencies, to make paired comparisons of home range sizes calculated by the different methods, and to examine differences in PB home range size among inhabitants of the 2 main study area wetland complexes. All statistical examinations were considered significant with alpha ≤ 0.05, and all descriptive statistics were reported as means and standard error (SE) of the means.

Movements

Radio-tracked Blanding's turtles were located from 10 to 62 times (mean = 34) throughout the active season and overwintering period of 1999–2000. Movements were examined for 24 adults (16 females, 8 males). In general, females initiated movements between activity areas earlier in the active season than did males. Males had almost twice the number of activity areas (mean = 3.13, SE = 0.40; Table 1) than did the 16 females (mean = 1.75, SE = 0.19; Table 1). Males moved between activity areas significantly more often (mean = 2.13, SE = 0.40) during the active season than did females (mean = 0.75, SE = 0.19, Z = –2.60, p = 0.0093; Fig. 3). During the summer, all males were observed to have at least 2 activity areas, while only 9 of 16 (56%) females had more than 1 summer activity area.

Table 1 Average number of activity areas, average number of interactivity area movements, overwintering movements as a percent of total movements, and average distances of overwintering movements between activity areas, from last activity area to nest site, and traveled overland from last activity area to nest site and then to next aquatic area after nesting; data are means (SE).

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

Turtles occupying the OZRW study area (3 males, 6 females) exhibited fewer interactivity area movements (mean = 1.11, SE = 0.35) than did those radio-tracked in WB (4 males, 6 females; mean = 1.50, SE = 0.40); however, differences were not found to be statistically significant (Z = 0.57, p = 0.57).

Females at Weaver Dunes moved to nesting areas in the late afternoon and early evening hours, and nested between 1800 and 2400 hrs. Among the 16 radio-tracked females, we examined nesting movements of 8 individuals with the most complete pre- and postnesting records. The mean distance of nesting migrations was 1458.25 m (n = 8, SE = 246.46; Table 1). Mean straight-line distances from the edge of the last wetland occupied by a female to nesting areas averaged 489.23 m (SE = 84.44; Table 1), and the durations of entire nesting migrations were 1–9 d.

Home Range Sizes

Sizes of home range of 8 male and 16 female Blanding's turtles were highly variable regardless of how they were calculated (Table 2). For males, the mean home range size calculated with the PB method averaged 56.89 ha (SE = 34.39), which was half the mean home range size calculated with the MCP and BNK methods (Table 2). The mean home range size of females estimated with the PB method was 18.89 ha (SE = 4.82; Table 2) and was half the mean home range size calculated with the MCP and BNK methods. Home range lengths of males and females averaged 1794.0 m (SE = 547.93) and 1472.0 m (SE = 191.35), respectively (Table 2).

Table 2 A summary of estimated home range sizes (ha) and lengths (m) of adult Blanding's turtles at Weaver Dunes calculated using 3 methods; data are means (SE).

View Fullscreen

Home range sizes of males and females pooled by primary wetland complex (OZRW: 3 males and 6 females vs. WB: 4 males and 6 females) in a paired comparison were significantly different (Z = –2.12, p = 0.034; Table 3).

Table 3 Mean home ranges calculated with the poly-buff method for adult Blanding's turtles in the 2 distinct wetland complexes (Old Zumbro River Wetlands [OZRW] and Weaver Bottoms [WB]) of the Weaver Dunes study area; data are means (SE).

View Fullscreen

Movements

Seasonal movement patterns of male and female Blanding's turtles were substantially different compared to each other at Weaver Dunes. Male turtles moved more often between clusters of aquatic activity (activity areas), and those movements were distributed throughout the warmer months (June–August). In contrast, females moved earlier and over longer distances in the active season than did males, and movements were presumed to be related primarily to prenest staging in aquatic areas near nesting sites. At other locations, male Blanding's turtles also moved more often than females and over shorter distances (Rowe and Moll 1991; Piepgras and Lang 2000), and male Blanding's turtles in New Hampshire increased activity attributed to mate searching later in the active season (August and September; Innes et al. 2008). Mate number is considered a major component of male reproductive success in Blanding's turtles (McGuire et al. 2011, 2013), and in freshwater turtles, male movements are related primarily to searching for mates (Morreale et al. 1984; Gibbons et al. 1990; Brown and Brooks 1993). However, movements of both sexes of Blanding's turtles in southeastern Michigan influenced reproductive paternity of clutches and the reproductive success of males (McGuire et al. 2013), and researchers observed no differences in movement frequency between sexes in Blanding's turtles in Ontario, Canada (Edge et al. 2010). While the males we followed presumably moved more often than their female counterparts, particularly in late summer, for mate searching reasons we did not record mating behavior as part of our radio-tracking data. Several other factors, such as food (Ross and Anderson 1990), water levels (Rubin et al. 2001), habitat thermal properties (Sajwaj and Lang 2000), and habitat quality, may have also been key contributors to the increases in late summer male movements, and further study of these factors is necessary to determine the movement disparity reasons between the sexes at the Weaver Dunes.

Terrestrial movements of females at Weaver Dunes were associated primarily with prenesting movements and nesting migrations, and once those movements occurred, the majority of females did not make other terrestrial movements throughout the active season. A study of Blanding's turtles in Ontario (Edge et al. 2010) noted that increased numbers of interwetland prenesting movements by females would be expected if suitable nest sites were limited, which may partially explain why females made fewer interactivity area movements at Weaver Dunes, where suitable nesting sites are abundant.

Home Range

The average home ranges of males from all 3 methods of calculation were 1.6 (MCP), 2.1 (BNK), and 3.0 (PB) times larger than those for females (Table 2) and 1.6 and 4.5 (PB) times larger for males than females from OZRW and WB, respectively (Table 3). However, small sample sizes coupled with large individual variation in home range sizes (combined data from OZRW and WB) may have resulted in the lack of significant differences in average home range sizes between the sexes, as a statistical difference was detected in movement patterns. Similar home range sizes of male and female Blanding's turtles were observed in several studies conducted in other areas over the past 30 yrs (Ross and Anderson 1990; Rowe and Moll 1991; Piepgras and Lang 2000; Schuler and Thiel 2008; Edge et al. 2010; Anthonysamy 2012; Hawkins 2016). Whereas a large body of research suggests that male and female Blanding's turtles exhibit similar geo-spatial requirements, among-year variation in their spatial biology (Schuler and Thiel 2008) may obscure differences or similarities in home ranges (and movement patterns) of male and females. Multiple-year studies (2 or more years) with large sample sizes and highly accurate transmitters that regularly record locations with minimal human interference (e.g., recaptures) are required to determine whether home ranges of males and females are different or the same in similar habitats over the same period of time.

Although we did not observe significant differences in home ranges of males and females, we did observe substantially smaller home range sizes for adult Blanding's turtles that occupied the wetland complex with proportionally more habitat that is typically preferred by the species (OZRW) compared with individuals that occupied lower-quality, highly fragmented, highly disturbed aquatic habitats (WB). The statistically significant differences in the 2 distinctly different areas support the prediction that habitat fragmentation and lower habitat productivity result in adults having larger home ranges than turtles in the less fragmented, more productive adjacent habitats. At Camp Ripley, Minnesota, aquatic habitats are highly fragmented, and home ranges of Blanding's turtles were larger than found at other locations (Piepgras and Lang 2000; Table 4).

A factor that influenced the differences in home ranges between the 2 wetland complexes is the large landscape alteration that occurred when the US Army Corps of Engineers implemented the lock and dam system during the 1930s. Prior to inundation, a major portion of WB was a wetlands meadow. On impoundment, WB developed emergent vegetation initially but then underwent a large-scale reduction in the amount of shallow aquatic habitat areas with emergent vegetation over a relatively short time. WB largely lacked emergent vegetation in shallow aquatic habitats between 1970 and 2010 (Fig. 2a–b). The 40 yrs during which WB had degraded to a large shallow windswept lake with the majority of emergent vegetation-containing productive areas being restricted to shorelines probably resulted in increases in 1) density of turtles in the shoreline area, 2) competition for resources, 3) movements among patches, and 4) emigration to other locations. Blanding's turtles may have remained in the same general area of WB because they exhibit strong fidelity to a resident wetland, a trait that suggests high costs associated with relocation (Congdon et al. 2011). Both remaining in and emigrating out of WB are associated with increased risk from predators and other causes of injury or death.

As the WB wetland complex experienced the loss of vegetation cover and increased the open water areas of lower productivity (Fig. 2b), changes in the location, connectivity, and abundance of preferred habitats of Blanding's turtles also resulted. In combination, the 4 decades of reduced vegetation may have reduced the population of Blanding's and other turtles substantially. During the 2005–2006 growing seasons, drawdowns were conducted in Navigation Pool 5 (Weaver Bottoms Habitat Project 2010) that resulted in a substantial increase of new areas of vegetation (Fig. 2c). If the gains in new vegetation are not relatively permanent or if another major landscape alteration event occurs locally, Blanding's turtles will again face the need to expand their home ranges to include preferred aquatic habitats. Blanding's turtles are long lived, reaching ages between 75 and 85 yrs or more (Brecke and Moriarty 1989; J.D.C., unpubl. data, from the E.S. George Reserve in southeastern Michigan), and females do not mature until they are 14 to more than 20 yrs of age (Congdon et al. 1993; Herman et al. 2003). The coevolved suite of life history traits of Blanding's turtles makes population stability sensitive to any increases in adult mortality (Congdon et al. 1993, 2011). Increasing risks of injury or death associated with increased duration or distance of movements (e.g., encounters with predators, farm equipment, or automobiles while conducting overland movements) coupled with decreased fecundity associated with reduced levels of resources can result in population declines, including in the uniquely large Blanding's turtle population at Weaver Dunes. Holistic management planning and efforts that reduce the occurrence of large-scale aquatic vegetation loss in WB and adjacent habitat areas and that even restore emergent vegetation through additional pool drawdowns will help ensure the long-term viability of Blanding's turtle at Weaver Dunes.

Tribute to Peter C.H. Pritchard

We would like to pay special tribute to Dr Peter C.H. Pritchard for his contribution to public awareness of turtles and tortoises. His global “home range” as an advocate of chelonian conservation was effective and inspiring.