Robust estimates of abundance are required for the proper management of imperiled or at-risk species (Williams et al. 2002); estimates of population size and trends are often used to determine whether a species deserves listing as threatened or endangered on state and federal levels. However, estimating populations is not easy, especially for species that are difficult to survey, like some reptiles (Olivier et al. 2010; Durso et al. 2011; Erb et al. 2015). One such species is the diamondback terrapin (Malaclemys terrapin), whose population status is poorly known throughout much of Florida.

The diamondback terrapin is the only turtle species endemic to coastal marshes of North America (Ernst et al. 1994), where it has a relatively high trophic position and exerts strong top-down control in the salt marsh by controlling snail populations, which otherwise might overgraze habitat (Hurd et al. 1979; Levesque 2000; Silliman and Bertness 2002). This species ranges along coastlines from Massachusetts to South Texas and occupies multiple habitat types, including estuaries, salt marshes, islands, creeks, rivers, and sandy beaches (Ernst and Lovich 2009). Malaclemys terrapin faces threats such as road mortality, habitat loss, nest predation, boat strikes, and bycatch mortality in blue crab (Callinectes sapidus) traps (Butler et al. 2006; Roosenburg and Kennedy 2018) that have caused population declines throughout its range (Seigel and Gibbons 1995; Dorcas et al. 2007). All states throughout its range classify M. terrapin as a Species of Greatest Conservation Need (US Geological Survey 2021). Three states (Massachusetts, North Carolina, and Mississippi) have listed the terrapin as threatened or of special concern, and many other states prohibit harvest (Rhode Island, Connecticut, New York, New Jersey, Maryland, Virginia, Alabama, and Texas) or have harvest restrictions (Delaware, South Carolina, Florida, and Louisiana; Roosenburg and Kennedy 2018).

Florida is important to the conservation of M. terrapin because its coastline represents 20% of the species' range (Butler et al. 2006). Five of the seven recognized subspecies (Hartsell and Ernst 2004) inhabit Florida: Carolina diamondback terrapin (M. t. centrata), Mississippi diamondback terrapin (M. t. pileata), eastern Florida diamondback terrapin (M. t. tequesta), mangrove diamondback terrapin (M. t. rhizophorarum), and ornate diamondback terrapin (M. t. macrospilota). The latter 3 subspecies are endemic to Florida. Studies of M. terrapin populations in Florida have been conducted on the Atlantic coast (Seigel 1984; Butler 2002; Butler et al. 2004), in south Florida (Wood 1992; Baldwin et al. 2005; Hart and McIvor 2008), and on the Gulf coast (Boykin 2004; Butler and Heinrich 2013). However, little information exists on populations inhabiting long stretches of Florida's coastline (Butler et al. 2006). Further, population assessments of M. terrapin in Florida have been generally limited to the peninsula, resulting in data gaps for populations throughout much of the Florida panhandle (Butler et al. 2006; Butler and Heinrich 2013).

Malaclemys terrapin macrospilota inhabits estuaries, islands, tidal creeks, salt marshes, and mangrove islands from Monroe County to Walton County, Florida (Butler et al. 2006). Although the subspecies likely faces significant threats, it receives little protection beyond a personal possession limit of 2 and prohibition of sale (which is the case for turtles of all species taken from the wild in Florida; Florida Fish and Wildlife Conservation Commission 2020). Consequently, an urgent need exists to gather population data on M. t. macrospilota.

We conducted an intensive capture–mark–recapture study of M. t. macrospilota during the summer of 2013 on a series of adjacent islands in the eastern panhandle of Florida. Our main objectives were to estimate population size, determine population structure, and examine body sizes of M. t. macrospilota in the study region. Results from our study could serve as a baseline for future studies and help determine the subspecies' listing status in Florida.

METHODS

Study Site. — Our study site was a series of small sand bar islands located in the Gulf of Mexico along the eastern panhandle of Florida that are sometimes submerged by extreme tides and storms. Malaclemys terrapin commands high prices in the pet trade, so island names are not included here. Our study site consisted of 3 adjacent islands, which were ca. 7.0, 0.9, and 0.6 ha in area and vegetated by salt marsh and coastal grassland, according to the Florida Natural Areas Inventory (2010) habitat classification scheme. Submerged vegetation surrounding the islands is dominated by seagrass (Thalassia spp.) beds and other aquatic plants that washes ashore as tidal wrack. The tidal wrack often sets atop saltmarsh cordgrass (Spartina alterniflora), which provides tented shaded areas that presumably provide cover or thermal refugia for M. t. macrospilota.

Sampling. — We hand-captured M. t. macrospilota while walking the perimeter of each island's intertidal zone weekly from 2 May through 30 July 2013. Surveys for M. t. macrospilota at each island were performed between 1000 and 1730 hrs. The entire intertidal area of each island was sampled on each visit. We targeted primarily areas of saltmarsh cordgrass under suspended tidal wrack, but we also captured terrapins in patches of open saltmarsh cordgrass.

For each terrapin, we measured midline carapace length (CL), carapace width (CW), shell height (HT), and plastron length (PL) to the nearest mm with 50-cm aluminum calipers (Haglof®, Langsele, Sweden). We measured turtle mass to the nearest 10 g with 1000- or 2500-g spring scales (Pesola®, Baar, Switzerland). We marked each individual by drilling holes in its marginal scutes using a standard numbering system (Cagle 1939). All turtles > 80 mm CL were also marked by implanting a PIT (passive integrated transponder) tag (Biomark®, Boise, ID) in the manor similar to Buhlmann and Tuberville (1998). We determined sex from the position of the cloacal opening, which is located posterior to the carapace in males (Lovich and Gibbons 1990). In addition, we palpated females for eggs to determine PL at maturity.

Data Analysis. — We used survey data to compile a detection history for each individual. We used Program MARK (White and Burnham 1999) to estimate the size of the superpopulation (N; the total number of animals in the study site during the study period) with the POPAN formulation of the Jolly–Seber model (Schwarz and Arnason 1996) using the R package RMark (version 2.2.6; Laake 2013). We constructed 6 models that differed by capture probability (p), apparent survival (ϕ), and probability of entrance (pent). We hypothesized that ϕ, p, and pent varied with time (t; survey #) or group (g; male, female, juvenile). Our simplest model had constant p, ϕ, pent, and N, whereas our most complex model possessed time-dependent p, ϕ, and pent. We used Akaike's information criterion (AIC; Akaike 1973, 1974) to select the most parsimonious model and considered models with lower corrected AIC values (AIC_c) to be more parsimonious (Burnham and Anderson 2002). We calculated population density by dividing N by the area (ha) of each island combined. In addition, we used a chi-square (χ²) test to determine whether the adult sex ratio deviated from 1:1. To examine differences in mean body size between sexes, we used 2-sample t-tests and calculated a sexual-dimorphism index by dividing the mean body size of the larger sex (females) by that of the smaller sex (males) and subtracting 1 (Lovich and Gibbons 1992). All statistical analyses were conducted in R (version 3.5.2; R Development Core Team 2018) implemented in R Studio (version 1.1.463; RStudio Team 2016).

RESULTS

We captured 420 terrapins (86 of which were recaptures) on the 3 islands. Forty-seven (14%) terrapins were adult females, 178 (53%) were adult males, and 109 (33%) were juveniles. All sizes were present except for hatchlings and small juveniles (Fig. 1). For a summary of morphological measurements, see Table 1. The sex ratio was skewed 4:1 males:females, which differed significantly from 1:1 (χ² = 76.27, n = 225, df = 1, p < 0.0005). As expected, analyses revealed that females were larger than males in CL, PL, HT, CW, and mass (Table 2). Interestingly, the smallest mature female had a PL of only 113 mm, and the smallest mature male had a PL of only 64 mm.

View Figure

• Download as Powerpoint
• Download Figure

Table 1. Standard morphological measurements taken from 334 Malaclemys terrapin macrospilota captured on 3 islands in the northern Gulf of Mexico during the summer of 2013. Data are presented as mean ± SE with ranges in parentheses. CL = midline carapace length; PL = plastron length; CW = maximum carapace width; HT = maximum carapace height.

View Fullscreen

Table 2. Mean morphological measurements, sexual-dimorphism index (SDI) calculations, and significance test for male and female Malaclemys terrapin macrospilota from 3 islands in the northern Gulf of Mexico.

View Fullscreen

Our most parsimonious model had group-determined ϕ, pent, and N, with time-dependent p (Table 3). Our population model estimated that N consisted of 1282 (867–1905 95% CI; Table 4) individuals. Juveniles and males were more abundant than were females (Table 4). In addition, the model indicated that the number of males, females, and juveniles all would decrease in the superpopulation over time; this trend was more marked for females (Fig. 2). Density was estimated at 150 (102–224 95% CI) terrapins/ha, but males and juveniles attained higher densities than did adult females (Table 4). The estimate of pent was essentially zero for females but slightly higher for males (0.03) and juveniles (0.05).

Table 3. Model selection table showing the most parsimonious population model for abundance of Malaclemys terrapin macrospilota from islands in the northern Gulf of Mexico. Models were constructed to allow apparent survival (ϕ), probability of capture (p), and population entrance (pent) to vary by group (g), time (t), or remain constant (.). The parameter N varied by group or remained constant. AIC_c is Akaike's information criterion corrected for small sample size.

View Fullscreen

Table 4. Superpopulation size (N) and density estimates for Malaclemys terrapin macrospilota on 3 islands in the northern Gulf of Mexico. Densities calculated for the entire population and by group (juvenile, male, female) with 95% CI.

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

DISCUSSION

Few population studies have been conducted on M. t. macrospilota, and ours is the first study to report population size and structure for the subspecies in Florida's northern Gulf of Mexico region. Our population model suggested this is the largest known insular population of M. t. macrospilota, with 1282 individuals estimated inhabiting the island complex. An earlier study of M. t. macrospilota in the Gulf of Mexico, which used similar capture techniques to this study reported much lower population estimates, 255 for one island and 141 for another (Boykin 2004). Our abundance estimates are also higher than those for other subspecies of M. terrapin. For example, Seigel (1984) reported estimates of 404 and 213 from 2 sites on the Atlantic coast of Florida. Hart (2005) estimated a population of > 1500 individuals, but this estimate was for a broad area of ca. 400 ha. Our population size findings may help in assessing whether this subspecies meets the criteria for state listing as threatened. Our estimated density of 150 turtles/ha is one of the highest reported. This high density leads us to believe that our study site has had little impact from major anthropogenic threats. One of the greatest threats to terrapin populations is crab traps (Roosenburg et al. 1997; Hoyle and Gibbons 2000; Grosse et al. 2009). The eastern panhandle region has the lowest crab-trapping effort in Florida because there are so few legal-size blue crabs there (Florida Fish and Wildlife Marine Information System 2016), and our results suggest that this is not a significant source of mortality in this area. In addition, threats that exist at some locations may be less important at our study site. For example, road mortality is a problem in some places (Butler et al. 2006; Grosse et al. 2011; Crawford et al. 2014), but there are no roads on these islands. In addition, these islands appear to be free of mammalian predators, which can devastate terrapins and their nests (Butler et al. 2004, 2006; Munscher et al. 2012). However, sea level rise (Titus et al. 2009) and illegal collection for the pet trade (Enge 2005; Commission for Environmental Cooperation 2019) are impending threats to all M. t. macrospilota populations.

Several explanations are possible for our findings that males and juveniles are far more abundant than are females. Females may come to these islands to nest and leave soon after oviposition. High densities and resource competition may result in females using different foraging areas away from the islands. Most of the observed prey items in the island complex were salt-marsh periwinkles (Littorina irrorata) and small crabs, such as fiddler crabs (Uca spp.). Female terrapins have larger heads, which allow them to consume larger prey items than males (Ernst and Lovich 2009). Females could forage outside of the intertidal zone in deeper water, where they would be more likely to encounter larger prey.

Our model also suggested that population size changes in all groups over time, possibly indicating emigration from the islands near the end of summer. This pattern was especially evident for females, and this population change indicates that this insular population is likely an aggregation of turtles that emigrate and immigrate seasonally. In addition, our model estimated some entrance into the population by males and juveniles, which could indicate that terrapins continue to immigrate to the islands during the summer. Terrapins may be present on the islands throughout the year, but this aggregation likely reaches its greatest size in the spring and early summer. To better interpret this result, however, requires further study that incorporates seasonal variability. Surveys we have conducted on these islands at other times of the year indicated that most terrapins leave in winter, but their overwintering locations are not known. Salt-marsh habitat is scarce along the mainland opposite these islands; the closest substantial salt-marsh habitat is in the mouth of a river > 2.5 km away. An earlier pilot study used acoustic telemetry to examine turtle movements and suggested that some island terrapins moved into that river, but the sample size was too small to allow conclusions regarding the population (T.M.T., unpubl. data, 2015).

Our study found a high percentage of juveniles and males, which is inconsistent with other studies of M. terrapin (Seigel 1984; Hart 1999; Butler 2000; Gibbons et al. 2001; Boykin 2004; Estep 2005; Sheridan 2010; Haskett 2011; Mealey et al. 2014). Our sampling technique may explain this. Other studies have applied capture techniques that require nets, which might have excluded terrapins of smaller size classes because of the nets' large mesh size or because they sampled habitats that lacked juveniles (Hart 1999; Gibbons et al. 2001). Sex ratios reported for M. terrapin populations are highly variable. Our observed male-skewed sex ratio agrees with that of many other studies (Cagle 1952; Hurd et al. 1979; Bishop 1983; Lovich and Gibbons 1990; Butler 2002; Estep 2005), but others have reported female-skewed sex ratios (Hurd et al. 1979; Seigel 1984; Roosenburg et al. 1997; Boykin 1999, 2004; Butler 2000; Sheridan 2010). The abundance of juveniles and males in our study may indicate lower rates of mortality from crab traps, which in other populations might have negatively affected smaller size classes (Roosenburg et al. 1997; Hoyle and Gibbons 2000; Dorcas et al. 2007).

Overall, we observed smaller mean PLs for both males and females than those reported in other studies. For example, our mean PL for adult females was 143 mm, which is ∼ 6 mm smaller than the smallest mean reported elsewhere (i.e., Mealey et al. 2014). Our mean PL for males was 88 mm, whereas the smallest mean PL reported elsewhere was 101 mm (Mealey et al. 2014). Although we did not conduct resource-availability surveys, we suspect our study site had ideal conditions for growth, and terrapins there might be reaching adulthood at smaller sizes than those elsewhere. For example, Hildebrand (1932) observed sexually mature females as small as 120 mm PL in captivity under ideal growth conditions. Alternatively, terrapins could be smaller because of their high densities in our site (i.e., density dependence). Regardless, the small size of adults in our population has conservation implications pertaining to bycatch-reduction devices (BRDs), which are used to restrict the entrance of crab traps and reduce catch of animals other than crabs. Smaller adult terrapins can enter crab traps more easily than others; the mean shell height of our adult males was 41.5 mm, smaller even than the most restrictive BRD (45 mm in height). Although commercial crab fishing is not of significant concern in this area, recreational crab traps are used near our study site. Recreational crab traps may pose a greater risk to terrapin populations than commercial traps because they are often set near areas inhabited by terrapins and are more likely to go unattended or be abandoned (Roosenburg 1992; Hoyle and Gibbons 2000). Just a few abandoned crab traps could drastically reduce M. terrapin populations (Roosenburg 1992; Grosse et al. 2009). Although the Florida Fish and Wildlife Conservation Commission (FWC) was petitioned in January 2020 to adopt or amend regulations to require BRDs on all commercial and recreational blue crab traps in state waters (Bennett et al. 2020), currently BRDs are not required on crab traps in Florida. We recommend that BRDs be placed on crab traps near M. terrapin aggregations like the one found in our study.

CONCLUSIONS

Our findings show that a large population or aggregation of M. t. macrospilota inhabits an island complex in the eastern panhandle of Florida. This is the first such population assessment conducted in the Florida panhandle, and though our population estimate is relatively large, surveys in other parts of the subspecies' range and the collection of population trend data will be needed to more solidly evaluate its status. Our study site, however, might represent an unusual situation in the region, and it should not be inferred that population densities on other islands are similar. Our study was a short-term population assessment, and single, short-term studies rarely produce definitive results (Nichols et al. 2019). Therefore, we recommend long-term population monitoring at this site and others in the region, which will allow researchers to estimate long-term population trends that will be useful in managing this subspecies.