Home Range Characteristics and Overwintering Ecology of the Stripe-Necked Musk Turtle (Sternotherus minor peltifer) in Middle Tennessee
Abstract
Little is known about the movement behavior of the stripe-necked musk turtle, Sternotherus minor peltifer. Using radiotelemetry, we calculated mean (± SD) home range length, which was 341.4 ± 90.3 m, with home range length not differing between the sexes (males, 335 ± 194 m; females, 346 ± 79.5 m). Sternotherus m. peltifer were active in every month of the year but decreased their movement distance and frequency between December and March; during nonwinter and winter periods, individuals used limestone bluffs most often.
Knowledge of annual movement patterns is critical for understanding the basic ecology of turtles (e.g., home ranges, habitat utilization, and overwintering; Gibbons et al. 1990). Many factors can influence movement behaviors, including habitat quality (Cagle 1944; Mahmoud 1969; Plummer and Shirer 1975; Duda et al. 1999), seasonality (Pluto and Bellis 1988; Jones 1996; Cadi et al. 2008), or mate searching (Morreale et al. 1984). Furthermore, within a species, differences in movement patterns between males and females can emerge due to sex-specific reproductive activities, such as nesting movements for females (Litzgus and Mousseau 2004) and mate searches for males (Morreale et al. 1984).
Sternotherus minor peltifer (stripe-necked musk turtle; Smith and Glass 1947) is a common chelonian found in small to medium-sized rivers throughout the southeastern United States (Ernst and Lovich 2009) and whose movement ecology is poorly understood. Sternotherus m. peltifer is characterized as a “bottom walker” (Berry and Shine 1980), which could possibly restrict its dispersal ability within a river. Likewise, terrestrial movements in S. m. peltifer are rare and associated mostly with nesting (Tinkle 1958) and basking (Ernst and Lovich 2009). Ennen and Scott (2008) reported the diel movements of S. m. peltifer but not home range characteristics. While diel movement reveals daily activity patterns (e.g., foraging), annual movement elucidates seasonal behaviors and habitat utilization, particularly during reproductively active and winter periods.
Because seasonal home range characteristics and movement data are lacking for S. m. peltifer, we studied a population of this subspecies in the Tennessee River drainage of western Middle Tennessee. Our overall goals were to describe 1) home range and habitat characteristics, 2) habitat use and movements while overwintering, and 3) seasonal sex-specific movement patterns and behaviors.
METHODS
Study Sites
We studied 2 access points 4.1 km apart along Whiteoak Creek (Rushing Bluff, lat 36°14′58″N, long 87°48′45″W; Gander Branch, lat 36°13′30″N, long 87°46′18″W). The Whiteoak Creek population of S. m. peltifer is 120 km north of known populations in Alabama (Mount 1975). Whiteoak Creek is a small, sinuous tributary (width < 4 m in some places) of Kentucky Lake (the impounded Tennessee River) and drains parts of Houston, Humphreys, and Dickson counties in Middle Tennessee. The 2 stream sections used in this study did not have any tributaries other than small springs, which no study specimen utilized. For a more detailed site description, see Ennen and Scott (2008).
Data Collection
Fieldwork was conducted from 19 May 2004 through 27 April 2005. Turtles were hand captured by wading, snorkeling, or canoeing. In the laboratory, we affixed VHF radio transmitters (Wildlife Materials International, Murphysboro, IL; model: SOPB-2190) with PC-Superepoxy® adhesive (Protective Coating Company, Allentown, PA) to the carapace of 14 individuals (6 males, 8 females) and allowed the epoxy to cure overnight (∼ 12 hrs). Once individuals were fitted with transmitters, they were released at their capture sites. We attempted to locate each turtle weekly.
We used a Trimble™ GPS receiver and a Trimble Survey Controller™ (Trimble Navigation Ltd, Sunnyvale, CA) to construct digital site maps to analyze home range characteristics. ArcGIS 9 (ESRI, Redlands, CA) was used to calculate home range size (ha) and length (m). Linear home range estimates accounted for sinuosity of the creek and did not use straight-line distances. We found that the 100% minimum convex polygon generally included habitats that S. m. peltifer never traversed (e.g., pastures and other terrestrial habitats) due to the sinuosity of Whiteoak Creek; therefore, we calculated the area (ha) of aquatic habitats where S. m. peltifer occurred. We use only linear home range length in our statistical comparisons but report home range size in Table 1 for comparisons to other turtles. To determine frequency of movement, we defined a movement as any weekly relocation (i.e., triangulation) that was not in the exact same location as the week before.
At each radio fix (i.e., triangulation), we collected creek depth (m) and generically characterized habitat, which was used to descriptively assess the hibernacula and habitat used within a home range. The habitat types included these categories: submerged limestone bluffs, vegetated rocky outcrops, bank with exposed root system, woody debris (including fallen trees, snags, and logjams), unconsolidated bank cover, patches of water willow (Justicia sp.), and open water. We calculated percent usage of each habitat type within each period for individual turtles. We measured depth (m) with a weighted nylon rope at the point of triangulation.
Statistical Analyses
All data were tested for normality and the appropriate parametric or nonparametric tests were performed. We conducted a Wilcoxon rank sum test (WRS) to compare mean homerange lengths between genders. We used Spearman's ρ correlations to assess the relationships between 1) turtle mass and homerange length and 2) homerange length and the length of time an individual was tracked. Individuals were tracked for varying lengths of time (i.e., weeks) because all individuals were not simultaneously outfitted with transmitters and transmitter failure did not occur simultaneously. Because of our low sample size with regard to both turtles and number of relocations within certain months, we categorized movement, depth, and habitat data into nonwinter (April–November) and winter (December–March) periods.
We determined differences between frequencies of movement (i.e., percentage of total relocations that were recorded as movements), depth, and mean distance traveled by using 3-factor analyses of variance (ANOVA) with individual turtles nested within sex (random factor), sex and month as fixed effects, and an interaction term (i.e., sex by period) as a fixed effect. We assessed the potential shift in cover type between winter and nonwinter periods by conducting individual WRS tests and used a Bonferroni correction for multiple comparisons on the same data set. All statistical analyses were completed using JMP 7.0.1 (SAS Institute Inc. 2007) using an alpha of 0.05. We report means and standard deviations when appropriate.
RESULTS
Home Range
Sternotherus m. peltifer linear home ranges were highly variable (7.8–1283.2 m); however, no relationship was found between length of time tracked and individual home range length (ρ = 0.08, p = 0.78). Mean linear home range size for S. m. peltifer (n = 14) was 341.4 ± 338.0 m (Table 1) and did not differ between males and females (335.2 ± 475.7 vs. 346.0 ± 224.7 m; WRS: z = −0.968, p = 0.333). Also, home ranges overlapped within and between the sexes. Eleven of 14 turtles (4 males and 7 females) moved at least 100 m on 1 or more occasions between weekly relocation points. Of the 3 turtles that never traversed a distance of 100 m or greater, 2 were males and 1 was a female. Throughout the study, 33.3% of relocation points were at the prior week's relocation point for that individual. There was no correlation between mass (164.8 ± 40.3 g, max = 254 g, min = 105 g) and home range length (ρ = 0.27, p = 0.35). We visually located 1 male while tracking during this study and did not observe any terrestrial activity.
Seasonal and Overwintering Behavior
Turtles displayed varying degrees of movement throughout the annual cycle. Seven turtles (4 males and 3 females, a combination of individuals from both access points) showed no activity throughout the winter, while 4 (1 male and 3 females) were active. The overall mean distance moved for all turtles varied throughout the study and differed between periods (nonwinter: n = 14, 32.2 ± 20.5; winter: n = 11, 2.0 ± 2.11; ANOVA, F1,10 = 21.50, p = 0.0009; Fig. 1) but not between the sexes (F1,10.03 = 0.05, p = 0.83), nor was there a sex-by-period interaction (F1,10 = 0.20, p = 0.77). Likewise, the frequency of movement differed between periods (F1,9.4 = 71.23, p < 0.001) but not between the sexes (F1,9.9 = 0.32, p = 0.58), nor was there a sex-by-period interaction (F1,9.4 = 1.40, p = 0.27).
Two males moved outside of their summer and fall activity areas to overwinter, while all other individuals remained within their summer and fall activity areas. Also, 3 males traversed downstream distances of 184–200 m to overwinter. Four turtles (3 males and 1 female) overwintered within a 37-m section of a submerged limestone bluff and occupied the same location but never simultaneously. Because of transmitter failures in the winter months, only 5 of the 7 turtles that were inactive during the winter were observed as they emerged from dormancy.
Citation: Chelonian Conservation and Biology 12, 1; 10.2744/CCB-1026.1
Habitat Use
Habitat use varied among individuals and shifted throughout the study. All but 1 individual used multiple (2–4) habitats within their respective home ranges. The most frequently used habitat was submerged limestone bluffs during both winter (57.1%) and nonwinter periods (39.6%; Table 2). Although not used as often as limestone bluffs and outcrops overall, woody debris was used in all but 1 home range. Use of habitats by S. m. peltifer during the winter period was more restricted than during the nonwinter period, and winter habitats did not include any water willow patches or open water (Table 2). Unlike nonwinter habitat usage, winter habitats were restricted mainly to areas beneath limestone bluffs (57.1 ± 51.4%) or vegetated rocky outcrops (17.5 ± 37.0%), with minimal use of the other features (Table 2). Woody debris and bank with exposed root systems were the only habitat to be used significantly more often in the nonwinter period than in the winter period (WRS: z = −2.7 to −3.9, p = 0.007–0.0001). Submerged limestone bluffs were the only habitat to be used significantly more often in the winter period than the nonwinter period (WRS: z = 3.5, p = 0.0005). The mean depth of turtle relocation points ranged between 0.2 and 3.6 m and averaged 0.88 ± 0.33 m. There was no difference in mean depth between periods (F1,12.6 = 1.15, p = 0.30) or between the sexes (F1,11.88 = 4.13, p = 0.07) and no interaction between sex and month (F1,11.88 = 0.06, p = 0.81).
DISCUSSION
No previous study has reported home range length of S. m. peltifer. In Whiteoak Creek, the linear home range of S. m. peltifer averaged 341.4 ± 338.0 m. Sternotherus odoratus and S. depressus have been the subjects of several published studies assessing home range size or length, which we use as a comparison. For example, the home range of S. m. peltifer in Whiteoak Creek is larger (x¯ = 0.518 ± 0.62 ha) than that reported for S. depressus (0.008 ha; Dodd et al. 1988) and S. odoratus in Oklahoma (males 0.024 ha, females 0.049 ha; Mahmoud 1969) but smaller than reported for S. odoratus in Pennsylvania (1.18 ha; Ernst 1986) and Michigan (2.8 ± 0.61 and 1.5 ± 0.14 ha; Rowe et al. 2009). However, only Dodd et al. (1988) and Rowe et al. (2009) used radiotelemetry to estimate home range, and Dodd et al. (1988) reported a home range from only 1 individual. Therefore, the difference between home range sizes could be related not only to species but to methodology as well.
Site fidelity is the proclivity of an individual to frequent a certain location (Switzer 1993) and is not an uncommon phenomenon in turtles (e.g., S. depressus, Dodd 1986; Macrochelys temminckii, Harrel et al. 1996; Graptemys flavimaculata, Jones 1996). Evolution should select for individuals with a proclivity to inhabit high-quality or at least suitable habitat more frequently because these habitats produce more resources that ensure reproductive success (Switzer 1993). As for S. m. peltifer, site fidelity and proclivity for a certain location was shown in the daily movement patterns (Ennen and Scott 2008), and we found similar behaviors. For example, approximately 33% of relocations were at the previous relocation point. Another behavior supporting site fidelity in S. m. peltifer is long-distance movements (> 50 m) from a location for a short amount of time (< 1 wk) with a return to the specific site occupied prior to the long-distance movement. Perhaps these long-distance movements, also reported for S. depressus (Dodd 1986), play a vital role in resource acquisition, mate searching, or gene flow within a river system.
Seasonal and Overwintering Behavior
In some turtle species, overwintering consumes more than half of the annual cycle and is therefore a significant portion of a turtle's life history (Ultsch 2006). In Whiteoak Creek, S. m. peltifer showed activity in every month of the study. However, from December to March, distance and frequency of movement decreased. This winter period is similar to previous reports for S. minor, S. odoratus, and S. carinatus, which also are relatively inactive during December–February in Oklahoma (Mahmoud 1969; Ernst and Lovich 2009). However, in Pennsylvania with colder winters than Middle Tennessee, S. odoratus is inactive for a 5-m period extending from November to March, with 220 d of activity (Ernst 1986). Although S. m. peltifer movement decreases notably during the winter months, some individuals remain active but to a lesser degree than during warmer months. Plummer and Burnley (1997) found that Apalone spinifera in Arkansas were not exclusively sedentary in winter; rather, some individuals moved to a new hibernaculum an average of 1.8 times over the course of winter. Perhaps some S. m. peltifer in Whiteoak Creek are similarly relocating to new hibernacula.
Communal overwintering is an uncommon phenomenon in turtles (Ultsch 2006). There are, however, several species (e.g., Clemmys guttata, Litzgus et al. 1999; Actinemys marmorata, Holland 1994) that have displayed this behavior, including S. m. peltifer in this study. In Whiteoak Creek, several S. m. peltifer overwinter in close proximity to one another. This phenomenon of communal overwintering was also reported within Sternotherus. In Ohio, 450 hibernating S. odoratus were discovered within a 13.72-m stretch of a canal (Thomas and Trautman 1937).
Sternotherus m. peltifer used fewer habitats during the winter period compared to the nonwinter period. The species' predominant use of limestone bluffs during winter and the aggregation of individuals suggest that use of this area by S. m. peltifer could be a preference rather than random use. Similar behaviors were reported for Chelydra serpentina, which displays aggregation and site fidelity to overwintering habitats (Brown and Brooks 1994). Interestingly, S. m. peltifer used woody debris less often during the winter than the nonwinter period. Perhaps its ephemeral nature (i.e., being swept away during high flows) makes woody debris less suitable for overwintering than more stationary habitats (e.g., limestone bluffs and outcrops). Alternatively, woody debris potentially contains a food source that is exploited only during nonwinter periods.
Mean distance traveled per period (bars) and mean frequency of movement (line, representing the mean percentage of movement per period; see text for full explanation) by Sternotherus minor peltifer tracked in Whiteoak Creek, Tennessee. Error bars represent ± 1 SD.
