Demographics of Sympatric Musk Turtles: The Loggerhead Musk Turtle (Sternotherus minor) and Eastern Musk Turtle (Sternotherus odoratus) in a Florida Spring Ecosystem
Abstract
The loggerhead musk turtle (Sternotherus minor) and eastern musk turtle (Sternotherus odoratus) are widespread species considered common throughout much of their respective ranges. Florida populations of S. minor and S. odoratus inhabit a variety of habitats, including freshwater springs and spring runs. We sampled populations of S. minor and S. odoratus from May 2000 through July 2015 as part of a long-term freshwater turtle assemblage study in a state-protected spring-run complex in central Florida. For each species, we 1) generated population estimates, 2) quantified survivorship and recruitment, and 3) calculated sex ratios, biomass, and density. Population estimates were 3417 for S. minor and 1977 for S. odoratus. Sex ratios were approximately 1:1 for both species, and density and biomass estimates were higher than most previous studies. Annual apparent survivorship was higher in male S. minor than in females, but the opposite was true for S. odoratus. Lambda values (k) suggest stable to slowly increasing populations for both species. Population growth rates for S. odoratus were equally influenced by both adult female mortality and mortality of first-year turtles, while those for S. minor were influenced by only adult female mortality. The lower fecundity values for S. odoratus may increase the importance of younger age classes in maintaining populations. Healthy populations of common species are important in maintaining overall community functions, as they often provide the base for many ecosystems. Therefore, it is prudent to promote conservation of common species to protect ecosystem structures and services.
Long-term studies are important in understanding environmental effects on population trends and demographics (Franklin 1989; Seigel and Dodd 2000). However, long-term studies on long-lived species are rare, as they are difficult to maintain (Tinkle 1979; Congdon et al. 1994). Many species outlive the life cycle of a typical grant, graduate student study, and sometimes the career of the primary researcher, making it difficult to maintain the personnel and equipment required to continue the research (Franklin 1989; Seigel and Dodd 2000). This is particularly true of chelonians, with some species having documented maximum life spans of more than 100 yrs (Gibbons 1987).
Compared with any other taxonomic groups, freshwater turtles are drastically understudied (Lovich and Ennen 2013). An estimated 61% of the 356 recognized species worldwide are considered threatened, endangered, or already functionally extinct (Turtle Taxonomy Working Group 2017). Many of these species are imperiled through some combination of anthropogenic issues, such as unsustainable harvest for food or the pet trade, habitat destruction, degradation and alteration, and climate change (Gibbons et al. 2000; Lovich et al. 2018). Additionally, for many long-lived species, changes in population structure and abundance can take years to manifest, making it difficult to ascertain the causes of population fluctuations (Gibbons 1997; Bjorndal et al. 2005; Hrycyshyn 2007). Long-term mark–recapture studies allow for population metrics such as survivorship, growth and maturation, migration, and fecundity to be calculated over the life span of the organism (Hrycyshyn 2007). Fluctuations in these demographic vital statistics over time can then be used to evaluate trends in population trajectories (Williams et al. 2001). Kinosternids (mud and musk turtles) rank as one of the least studied groups of turtles in North America (Lovich and Ennen 2013). In particular, there is a paucity of data concerning vital population demographics for species within the genus Sternotherus (musk turtles; Lovich and Ennen 2013).
The loggerhead musk turtle (Sternotherus minor) and eastern musk turtle (Sternotherus odoratus) are widespread, common species within their respective ranges (Ernst and Lovich 2009; Krysko et al. 2011; Turtle Taxonomy Working Group 2017). Sternotherus minor ranges from eastern Tennessee and southwestern Virginia in the north, across most of Georgia into central Florida in the southeast and west across most of Alabama to eastern Mississippi (Zappalorti and Iverson 2006; Ernst and Lovich 2009; Krysko et al. 2011). Sternotherus odoratus has one of the largest ranges of any North American turtle (Ernst and Lovich 2009). This species ranges from southeastern Maine along the East Coast south into all of Florida, west to the hill country of Texas, and north through the Midwest into Wisconsin, Michigan, and southern Quebec and Ontario, Canada (Iverson and Meshaka 2006; Ernst and Lovich 2009).
Florida populations of S. minor and S. odoratus inhabit a variety of ecosystems (Iverson and Meshaka 2006; Zappalorti and Iverson 2006), including freshwater springs and spring run systems (Berry 1975; Iverson and Meshaka 2006; Riedle et al. 2016). Habitat preferences for both species are similar and can overlap (Iverson and Meshaka 2006; Zappalorti and Iverson 2006; Krysko et al. 2011); however, they seldom co-occur at the same locality in abundance (Berry 1975; Chapin and Meylan 2011; Riedle et al. 2016). Sternotherus minor has an affinity for lotic habitats, whereas S. odoratus prefers more lentic habitats (Tinkle 1958; Iverson and Meshaka 2006; Zappalorti and Iverson 2006). When they are found sympatrically at a locality, they are rarely found at similar densities (Tinkle 1958; Meylan et al. 1992; Iverson and Meshaka 2006; Zappalorti and Iverson 2006; Chapin and Meylan 2011; Riedle et al. 2016).
We sampled populations of S. minor and S. odoratus at a state park during a 15-yr period (May 2000 through July 2015) as part of a long-term multispecies monitoring study of the freshwater turtle assemblage (Munscher et al. 2015a, 2015b; Walde et al. 2016). The objective of this study was to provide baseline data to better understand the population dynamics of these 2 species. For each species, we were interested in 1) generating population estimates, 2) quantifying annual survivorship and recruitment, and 3) estimating sex ratios, biomass, and density. Our primary goal was to gain a better understanding of how protected populations of these 2 perceived common species may function over a long period of time.
METHODS
Study Area and Design
Wekiwa Springs State Park (WSSP), Apopka, Florida (Orange and Seminole counties; Fig. 1), was purchased by the state of Florida in 1969. The main spring, which expels approximately 164 million liters of water a day, has been used as a recreational area since 1941 (Philpott 2008; Stamm 2008). Rock Springs Run State Preserve, purchased by the state of Florida in 1983, borders WSSP and creates a protected area of more than 16,187 ha (Philpott 2008). The area represents a typical central Florida spring with wet lowlands surrounding the springs, adjacent to dry sandhill uplands maintained by frequent prescribed fires (Philpott 2008; Stamm 2008). At WSSP, the area adjacent to the spring has been modified with concrete walls with ladders and steps to facilitate swimming and general recreational use. This small swimming area drains into the main lagoon, which is used for canoeing and fishing but not swimming. The main lagoon then drains into the Wekiwa Springs Run. The entire study area consists of approximately 2.67 ha of protected water habitat (Munscher et al. 2015a). Wekiwa Springs Run joins Rock Springs Run to become the headwaters of the Wekiva River. The survey area is made up of the main lagoon and Wekiwa Springs Run, with turtles occasionally being captured in the swimming area. We collected data from May 2000 until July 2015 on S. minor and S. odoratus. Surveys were conducted twice annually, with the exception 2001 (1 sample) and 2005 (3 samples). Previous descriptions of the study area and methodologies area available from earlier publications from this study (Munscher et al. 2013, 2015a, 2015b; Walde et al. 2016)
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World's Turtle and Tortoise Journal 19, 1; 10.2744/CCB-1442.1
Turtle Capture
The present project is a citizen science/volunteer project (Munscher et al. 2013) with all participants being volunteers. For each sampling period, a variable number of snorkelers hand-captured turtles from ∼ 0800 to 1600–1900 hrs, depending on the time of year and weather conditions. We placed all captured turtles in labeled bins within canoes that indicated the capture area within 100-m sections and brought them to a central location for data processing. We also placed hoop traps with leaders to target other turtles in the system (Walde et al. 2016) but never caught Sternotherus in these traps. Attempts to use baited collapsible crab traps (Promar, Gardena, CA), as shown to be highly successful for Sternotherus in previous studies by Munscher et al. (2017), resulted in failure due to American alligators (Alligator mississippiensis) destroying the traps to access the bait.
Data Collection
Maximum straight-line measurements of carapace length (CL), plastron length, carapace width, and shell height were recorded to the nearest millimeter. Turtles were then sexed based on secondary sexual characteristics, notably tail length and girth as described in Ernst and Lovich (2009). Unique features and physical anomalies, such as damage, scars, or coloration, were noted for each turtle and aided in confirming individual identity. All turtles were weighed using Ohaus top-loading digital scales (Ohaus Corp, Parsippany, NJ). Turtles were then released back into the spring run at their approximate capture locations.
Marking Method
Turtles were marked using a variation of the technique described by Cagle (1939) and a variation of Cagle's method of marking the plastron to denote “series”/thousands (J. Iverson, pers. comm., August 2014). Beginning in 2009, we used passive integrated transponder (PIT) tags as a secondary identification method for turtles with a CL greater than 70 mm. A PIT tag was injected under the turtle's right bridge (Buhlman and Tuberville 1998; Runyan and Meylan 2005). Capture and handling protocols were approved by the Florida Department of Environmental Protection, Division of Recreation and Parks (District III, Apopka), the Florida Fish and Wildlife Conservation Commission, and the Institutional Animal Care and Use Committees at the University of North Florida (Jacksonville) and conform to the animal use guidelines of the American Society of Ichthyologists and Herpetologists (2001), the Herpetologists League, and the Society for the Study of Amphibians and Reptiles.
Data Analysis
We calculated population abundance for adults using the POPAN parameterization of Jolly-Seber models (Jolly 1965; Seber 1965) in Program MARK (White and Burnham 1999). We calculated the biomass for each species and sex based on the mean mass of that sex, multiplied by the sex-specific density estimates of the study period (Congdon et al. 1986). We calculated annual apparent survival (Φ) and recapture rates (p) using open population Cormack-Jolly-Seber models (CJS; Lebreton et al. 1992) in the program MARK (White and Burnham 1999). To test for differences in Φ and p between the sexes, we generated CJS models to test whether Φ or p differed based on sex, time, or a sex–time interaction. We also generated CJS models to test whether Φ or p differed between the 2 species of Sternotherus, as a function of time, or through a species–time interaction. We based model selection for all analyses on corrected Akaike information criterion values, with lower values denoting greater parsimony (Burnham and Anderson 2002).
Apparent survivorship is typically calculated by using encounter rates to calculate the probability of an individual leaving a population. If the encounter rates are reversed, one can estimate the probability of an individual entering a population, providing a rough estimate of a population growth rate (λ; Pradel 1996). To avoid confusion with other analyses, we refer to these values as Pradel's λ in this article. Pradel's λ differs from traditional estimates of λ, as no fecundity values are included in its calculation, so is not necessarily equivalent to a true population growth rate. Mark estimated Pradel's λ in conjunction with the CJS model described above.
Finally, we developed population viability models in the program Vortex (Lacy and Pollak 2014). We constructed default scenarios using population and survivorship estimates and sex ratios calculated from mark–recapture data. Reproductive data were taken from summaries in Iverson and Meshaka (2006) and Zappalorti and Iverson (2006). For S. minor, age at first reproduction was set at 5 yrs for females and at 4 yrs for males, maximum number of clutches annually at 4, and total number of eggs per year at 20. Egg mortality was set at 85% based on reviews in Iverson (1991a, 1991b). For S. odoratus, age at first reproduction was set at 4 yrs for females and at 3 yrs for males, maximum number of clutches annually at 3, and total number of eggs per year at 12. We then ran a series of sensitivity tests for 4 parameters: percent of females breeding/year, hatchling mortality, first-year mortality, and adult female mortality. Values for each parameter ranged from 0% to 100% and increased by 5% with each iteration.
RESULTS
We captured and individually marked 1347 S. minor (2011 total captures) and 628 S. odoratus (852 total captures; Fig. 2a–b). Female S. minor had significantly greater plastron length and shell heights compared with males, while all other variables were similar (Table 1). Sternotherus minor did not exhibit sexual size dimorphism with regard to mass, while females were heavier than males in S. odoratus (Table 1). All morphometric variables were significantly different between male and female S. odoratus (Table 1).
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World's Turtle and Tortoise Journal 19, 1; 10.2744/CCB-1442.1
Population estimates were 3417 for S. minor and 1977 for S. odoratus (Table 2). Density estimates were 1279 turtles/ha for S. minor and 740 turtles/ha for S. odoratus. Biomass calculations were 56.3 and 62.4 kg/ha for male and female S. minor, respectively, for a total of 118.7 kg/ha. For S. odoratus, biomass estimates were 18.3 and 20.2 kg/ha for males and females, respectively, for a total of 38.5 kg/ha.
Apparent survivorship was higher in male S. minor than in females, but the opposite was true for S. odoratus, with low recapture rates observed for both species (Table 2). In the most parsimonious model for S. minor, survivorship and recapture rates differed between the sexes and sampling periods (Table 3). Variation among the top models for S. odoratus was slight (Table 4), with little difference between models, suggesting that apparent survivorship and recapture rates were constant or varied by group. Using the model averaging function in MARK, weighted averages and standard errors for apparent survivorship in S. odoratus were 0.76 ± 0.03 for females and 0.74 ± 0.03 for males. Weighted averages and standard errors for recapture rates were 0.12 ± 0.01 for females and 0.13 ± 0.01 for males. These values varied little from those reported for each sex in Table 2, suggesting that variation in those parameters by sex is well supported.
In comparing apparent survival and recapture rates between species, the most parsimonious model was one where survivorship varied by time and recapture rates by group and time (Table 5). There was little variation between models and, when averaged, weighted average plus standard error 0.45 ± 89.95 for survivorship and 0.26 ± 37.35 for recapture rate. This suggests a high degree of uncertainty between models. This uncertainty may be influenced by variation in encounters between sampling periods.
Pradel's λ values for both species suggest stable to slightly increasing populations (Table 6). When reproductive characteristics were introduced in Vortex, some differences were observed between the species. For S. odoratus, λ was slightly larger than 1.0, and net reproductive rate (R0, the average number of age class zero offspring produced by an average newborn organism during its lifetime) was larger than 1.0, meaning each turtle replaced itself and contributed to an additional individual (Table 7). For S. minor, both λ and R0 were slightly lower than 1, suggesting a stable to slightly declining population.
Population growth rate for S. minor was most influenced by adult female mortality within sensitivity tests (Fig. 3a). Population growth rate for S. odoratus was equally influenced by both adult female mortality and mortality of year 1 turtles (Fig. 3b). The lower fecundity value for S. odoratus may increase the importance of younger age classes in population maintenance.
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World's Turtle and Tortoise Journal 19, 1; 10.2744/CCB-1442.1
DISCUSSION
Few studies report demographics for both study species where they co-occur, as they rarely occur in sympatry (Iverson and Meshaka 2006; Zappalorti and Iverson 2006). Florida springs represent some of the few sites where the 2 species are found living sympatrically. At Rainbow Run Springs, a habitat similar to WSSP, Meylan et al. (1992) calculated population estimates for both species—1270 for S. minor and 1056 for S. odoratus—in their 10-ha study area. These population estimates have been among the highest estimates for each species in Florida. In comparison, at WSSP, a much smaller system (2.67-ha site), we calculated population estimates of 3417 for S. minor and 1977 for S. odoratus. A survey of the second-largest freshwater spring in Florida, Ichetucknee Springs, reported one of the highest population estimates for S. minor at 5070 turtles (or 461/ha) for an 11-ha study site (Chapin and Meylan 2011). The authors admitted that theirs was a conservative estimate due to their low recaptures rates, stating that the population within Ichetucknee Springs could surpass 18,000 turtles. In contrast, Chapin and Meylan (2011) observed only a small population of S. odoratus (estimate of 104) at this study site. The authors attributed this to the robust population of S. minor present in the system. A few interesting points can be made when comparing results from Chapin and Meylan (2011) to this present study. The 2 study sites differ drastically in overall size (11 ha vs. 2.67 ha), site makeup (spring-fed river vs. lagoon and spring run), water flow (first-magnitude vs. second-magnitude spring), and vegetation composition (eelgrass vs. hydrilla; Scott et al. 2004; Munscher et al. 2015b). The Ichetucknee Springs habitat is more suited for S. minor, as the species is known to prefer spring run systems with flowing water (Zappalorti and Iverson 2006). In contrast, the majority of the WSSP study site is a large lagoon with slower water flow, which allows for a more dynamic habitat that can be used by both species of Sternotherus.
Additional population comparisons exist within Florida but not from sympatry. Bancroft et al. (1983) found that S. odoratus was the most abundant turtle species in their study site, Lake Conway. They noted observations of 3008 individuals during their study but were unable to calculate any true population estimates. In Volusia Blue Springs, a spring-fed run system located within the same watershed (St Johns River) as WSSP, Riedle et al. (2016) calculated a population estimate of 252 ± 7 individuals for S. minor, or 133/ha. In a study period of 8 yrs, the authors captured only 5 S. odoratus. In a small pond habitat located in Okaloosa County, Steen et al. (2012) estimated a population of 27.5 S. minor individuals. In all of these instances 1 species of Sternotherus was dominant in the study system, whereas at WSSP, both species of Sternotherus appear to have robust populations.
Further examples of S. odoratus population estimates exist outside of Florida. A relatively high estimate of 534 ± 45 individuals was obtained from an urban lake and creek system in Virginia (Mitchell 1988). Similarly, in a small lake system (13.76 ha) in Alabama, Dodd (1989) surveyed a population of S. odoratus and found a population of 148.5 turtles/ha, equating to approximately 2043 individuals. In contrast, in a large Texas freshwater spring system (10.1 ha), Munscher et al. (2019) estimated a population > 14,000 individuals at a density of 1690/ha. While this estimate seems quite large, recapture rates were very low (0.08–0.10). Ultimately, S. odoratus appears to be a successful generalist species, but the presence of other kinosternids may influence habitat use and abundance (Mahmoud 1969; Berry 1975; Riedle et al. 2015; Munscher et al. 2019).
Wildlife managers often rely on population density estimates in order to make recommendations for species management, yet accurately estimating this parameter remains difficult due to a limited number of long-term studies (Gibbons 1997). The densities of Sternotherus species reported for this study fall into the mid- to high range of other studies reporting densities. Density estimates for S. minor have been estimated as high as 2857 turtles/ha at Emerald Springs Boil, Bay County, Florida (Cox and Marion 1979). At Rainbow Run, density estimates for the species were lower, at 127 turtles/ha (Meylan et al. 1992). A study of a small pond system in the panhandle of Florida estimated a population density of 35–63 turtles/ha (Steen et al. 2012). Looking across the range of values presented by Zappalorti and Iverson (2006), the density estimates for the present study (1279 turtles/ha for S. minor and 740 turtles/ha for S. odoratus) are higher with only 1 exception within the state of Florida, that being the 2857 turtles/ha from Emerald Springs (Iverson and Meshaka 2006). Sternotherus odoratus is known to be locally abundant throughout its range (Zappalorti and Iverson 2006). A few examples within Florida show a variety of density estimates, ranging from 106 turtles/ha in the Rainbow River in Marion County (Meylan et al. 1992) to 700 turtles/ha in a small pond near Gainesville in Alachua County (Iverson 1982). Most of the estimates for this species come from outside of Florida in nonspring habitats. In Alabama, Dodd (1989) calculated a density of 149 turtles/ha. In Virginia, Mitchell (1988) calculated a density of 194 turtles/ha, while Ford (1999) calculated a density of 174 turtles/ha in a Missouri river. The high densities observed during our study suggest that the habitat at WSSP is well suited for both Sternotherus species, which could have cascade effects on the system, as it relates to influence on the environment and ecosystem function, which can be better shown by biomass.
Estimates of turtle biomass are important factors for calculating the energy and nutrient flow that passes through a population (Iverson 1982). Musk turtle species are highly omnivorous and have a wide variety of prey sources, including mollusks, gastropods, vegetation, and carrion (Ernst and Lovich 2009; Morrison et al. 2017, 2019). Biomass is also a necessary metric when estimating prey base for predators. Turtles are sources of food for many organisms across their life history. Eggs, juveniles, and even adults of some species provide food for predators such as mammalian mesopredators, including raccoons (Procyon lotor), opossums (Didelphis virginiana), nine-banded armadillo (Dasypus novemcinctus), red foxes (Vulpes vulpes), and river otters (Lontra canadensis), as well as American alligator (Alligator mississippiensis), and numerous fish species, including bowfin (Amia calva), and even other turtle species like the alligator snapping turtle and many other species (Brooks et al. 1991; Bondavalli and Ulanowicz 1999; Elsey 2006; Munscher et al. 2012). Not surprisingly, the more we learn about turtle populations, the more we realize how much their populations contribute to ecosystem energetics, composition, and overall function. This is partially due to their lower food intake requirements compared with similarly sized endotherms; as such, turtles can achieve higher population and biomass densities than endotherms in the same ecosystem (Iverson 1982; Hrycyshyn 2007). One of the most underreported metrics in turtle studies is biomass (Iverson 1982), as it is often not measured at all or has been estimated incorrectly, substituting average body mass from closely related species rather than actual body mass found in study populations (Iverson 1982; Congdon et al. 1986; Hrycyshyn 2007), with corresponding over- or underestimation of true biomass.
In the present study, S. minor had a geometric mean biomass of 118.7 kg/ha, and S. odoratus had a geometric mean biomass of 38.5 kg/ha with the sexes pooled. In comparison, the biomass of the S. minor population at Emerald Springs Boil (which was calculated to have the highest density of the species known) in Marion County, Florida, was calculated to be 45.7 kg/ha (Iverson 1982; Hrycyshyn 2007). The biomass for a small population of S. minor at Volusia Blue Springs was calculated to be 16.4 kg/ha (Riedle et al. 2016). In Rainbow Run, Meylan et al. (1992) found a biomass of 12.5 kg/ha for S. minor and 6.1 kg/ha for S. odoratus x both substantially lower than our biomass estimates at WSSP.
The WSSP system is regarded as a high-quality habitat by the Florida Department of Environmental Protection and the National Wild and Scenic River System (Wekiva River System Advisory Management Committee 2012). The Wekiva River is 1 of only 2 wild and scenic rivers within Florida. The Wild and Scenic Rivers Act identifies a number of Outstandingly Remarkable Values (ORVs) that are used in designating a river as wild and scenic, thus providing legal protection. The Wekiwa River has been assigned ORVs including scenic, recreational, geologic, fish and wildlife, historic and cultural, and water quality and quantity. The Wekiva River is 1 of only 2 wild and scenic rivers in Florida, and management addresses these ORVs (Wekiva River System Advisory Management Committee 2012). We speculate that the differences we are seeing regarding density and biomass values within this system compared with other Florida spring systems is due to the availability of food (Munscher et al. 2015b; Riedle et al. 2016). Other spring system habitats have degraded through water loss, water quality, and the presence of invasive species (Toth and Fortich 2002; Saint Johns River Water Management District 2006; Kennedy et al. 2009; Riedle et al. 2016).
We calculated the sex ratios for both species of musk turtle to ascertain the health and stability of each population. Our calculated sex ratios were 1.08:1 male/female for S. minor and 1.31:1 for S. odoratus. Comparing our data with other Florida studies, Meylan et al. (1992) found a sex ratio of 0.92 males/female for S. minor and a significantly male-biased estimate for S. odoratus at 2.63 males/female at Rainbow Run. Cox et al. (1991) found a sex ratio of 1.14 males/female for S. minor in northwestern Florida springs, and Riedle et al. (2016) calculated a significantly skewed male/female ratio of 2.13:1 at Volusia Blue Springs. Numerous studies have reported a variety of sex ratios for S. odoratus, as this wide-ranging species is known from a variety of habitats and ecoregions; these estimates range from significantly female biased to significantly male biased (Gibbons 1990; Iverson and Meshaka 2006; Ernst and Lovich 2009). This variation in sex ratio is most likely due to regional variation in both biotic and abiotic factors acting on each population.
Ultimately, long-term persistence of populations will depend on annual apparent survival of each population. Annual apparent survival probabilities are different from true survival probabilities in that mortality and emigration are confounded and therefore likely to be biased low (White and Burnham 1999). To our knowledge, the survival estimate provided for this study is only the second to be calculated for S. odoratus in Florida, albeit from the same ecosystem (Iverson and Meshaka 2006; Zappalorti and Iverson 2006; Hrycyshyn 2007).
Only a few studies have examined annual survival probabilities for Sternotherus and other kinosternid species. At Volusia Blue Springs, Riedle et al. (2016) reported annual apparent survival of a small population of S. minor, finding an overall annual survival probability of 0.97 when the sexes were pooled. This estimate is much higher than the 0.81 estimated for the present study. A major factor that could contribute to this skew is that Volusia Blue Springs is a limestone-based spring run that has little to no aquatic vegetation due to the presence of grazing West Indian manatees (Trichechus manatus; Riedle et al. 2016). Sternotherus are easier to find at Blue Springs than at WSSP due to the abundance of invasive aquatic vegetation at WSSP that at times overtakes the study area. Mitchell (1988) studied S. odoratus in a Virginia pond for several years and found an overall annual survival probability of 0.78 for males and 0.845 for females, a similar trend to this study. The male survival probability estimated by Mitchell (1988) fell within the lower 95% confidence interval for our estimate of male S. odoratus survival, and the female survival probabilities did not vary greatly.
WSSP has robust Sternotherus populations that reside within the lagoon and spring run system. In fact, it seems like an ideal habitat with high densities and biomass; however, this protected system is not without anthropogenic impacts. From 2000 to 2015, the ecosystem has seen decreases in both water quality and quantity due to loss of recharge area (due to more impervious surfaces) and high nitrate levels (due to in-ground antiquated septic tanks) in the watershed (Saint Johns River Water Management District 2006; Hrycyshyn 2007; Tucker et al. 2014). The water coming out the spring during the time period of this study has spent roughly 20 yrs underground, and the nitrates are from lawn fertilizers and septic tanks in place prior to the 1980s (Toth and Fortich 2002; Kennedy et al. 2009). These high nitrate levels could explain why invasive plants have taken root and increased in abundance at WSSP (Kennedy et al. 2009).
Hydrilla was first observed in small quantities in 2001 and within a year had overtaken native vegetation and choked off a significant part of the study site lagoon and spread into the WSSP spring run (Hrycyshyn 2007). WSSP staff and the Florida Department of Environmental Protection employed several methods over the next 15 yrs to combat this invasive species (V. Oros, pers. comm., March 2015). One such treatment was the use of an aquatic barrier that effectively split the lagoon in two, with one side having the spring flow cut off in which treatment could be applied through the use of a photosynthetic inhibitor. This treatment occurred periodically over the next 14 yrs. Manual removal was also used in areas of the spring and lagoon. By 2015, the removal efforts were a success, with hydrilla being effectively removed from the system.
Although we were not able to quantify effects of the treatments, anecdotal information suggests that the turtles were not impacted in any observable negative way. However, shortly after the plant's removal, 2 negative situations presented themselves: H. verticillata provided abundant food and cover from predators such as American alligators and also made it far more difficult for turtles to be hand captured (Munscher et al. 2015b; Adler et al. 2018). We observed that the presence of hydrilla did make capture of Sternotherus difficult in certain areas of the study site during periods of plant abundance. The second issue that could be more problematic in the following years is the accumulation of dead plant matter and detritus associated with the dead hydrilla and storm runoff. Wekiwa Springs is a second-class spring that apparently does not have the necessary flow rate needed to push most suspended organic materials downstream. As of late 2015, there was a noticeable muck layer at the bottom of the study site lagoon (Wetland Solutions Inc 2007; Munscher et al. 2015b); only continued sampling will determine what effect, if any, this muck layer has on the Sternotherus in this system.
WSSP is a protected natural area in the midst of the Orlando metropolitan area urban sprawl that happens to have a thriving musk turtle community. This community is unique in that both species of musk turtles seem to have large, stable populations. Our data can serve as a baseline for other, less studied populations of the same species or similar species. Continued monitoring of this community will document how decreased spring outflow, sustained high nitrate levels, the removal of a dominant invasive species, and the subsequent increase in organic layer at the bottom of the lagoon impact these species. In the event that either species begins to show declines, our data can serve as a basis for restoration efforts.
Tribute to Peter C.H. Pritchard
For all of us, growing up with a passion for turtles and turtle conservation was inflamed by those giants in the field who came before us. Dr Peter Pritchard was the giant among giants. He embodied what a true conservationist should strive for, and that is conservation through common understanding and education. Peter was able to find common ground with anyone who was willing to listen and through mutual education establish lasting environmental changes that benefited the species that we love. He influenced generations of colleagues and students, showing them that you can make conservation and research your life's work, which can really make a difference. All the authors of this article have had interactions with Peter over the years. His brilliance and humbleness had a lasting impact on all of us and will continue to do so as we move forward his message through our turtle conservation efforts. Our deepest gratitude and respect to Dr Pritchard for helping to light the torch so that turtles can continue to survive and inspire love and affection for generations to come.
Aerial photograph from 2005 of Wekiwa Springs State Park, Orange and Seminole counties, Florida. The study site includes the public lagoon and main lagoon northeast of it, the connecting run between the lagoons, and 1.1 km of spring run habitat.
Size distribution of (a) Sternotherus minor and (b) Sternotherus odoratus showing maximum carapace length at time of initial capture of captured at Wekiwa Springs State Park, Florida, from 2000 to 2015. Note that there are 561 male, 607 female, and 179 juvenile turtles individually marked S. minor. There are 344 male, 261 female, and 23 juvenile S. odoratus individually marked.
Sensitivity plots for Sternotherus minor (a) and Sternotherus odoratus (b) at Wekiwa Springs State Park, Florida.
Contributor Notes
Handling Editor: Jeffrey A. Seminoff
