Alabama ranks first in the nation regarding turtle biodiversity, and most of that diversity can be found in its aquatic habitats (Guyer et al. 2015; Scott et al. 2018). Alabama's subtropical climate and incredible number of rivers and streams have resulted in diversification in several chelonian lineages (Guyer et al. 2015). Even though there is high diversity, no sustained research has examined the potential effects of urbanization on turtle populations in the state. Past studies have shown that human influences from urbanization have negative impacts on freshwater turtle populations, which include high numbers of female-biased mortalities from roadkills (Aresco 2005; Gibbs and Steen 2005; Steen et al. 2006) and a higher prevalence of human-subsidized nest predators (e.g., raccoons, Procyon lotor) (Marchand and Litvaitis 2004; Browne and Hecnar 2007). However, certain urbanized habitats (e.g., golf courses) have been shown to seemingly benefit some turtle populations (Harden et al. 2009; Winchell and Gibbs 2016). For instance, nest predations near golf course ponds were not significantly higher than more rural ponds (Foley et al. 2012). Further, residential lawns and road banks can provide suitable nesting habitat (Marchand and Litvaitis 2004). Nevertheless, responses to urbanization can be species-specific (Eskew et al. 2010), so generalizations may be misleading depending on what species is being considered.

Because turtles are long-lived species (Gibbons 1987), long-term field studies are needed to examine various aspects of their life history and ecology (Tinkle 1979; Congdon et al. 1993). Recent studies have utilized data sets collected from such long-term field studies to examine questions of conservation and reproductive behaviors. The frequency of propeller damage in different species of freshwater turtles was tracked over 30 yrs, and, although within-species frequencies did not significantly differ among years, the authors observed among-species differences that reflected the species' different behaviors (Smith et al. 2018). Juvenile growth rate was shown to have long-lasting positive influences on not only age and size at maturity but also long-term reproductive output (Congdon et al. 2018). Additionally, over 20 yrs of data were used to determine that painted turtles (Chrysemys picta) display senescence in the wild (Warner et al. 2016), and this same data set also detected a shift in nesting phenology due to a warming climate (Janzen et al. 2018). Finally, an almost 50% decline in a population of spotted turtles (Clemmys guttata) was observed over a 30-yr period, even though the population resides within a protected area (Howell et al. 2019).

The Urban Turtle Project was initiated in 2018 to collect long-term demographic and ecological data from turtle populations inhabiting various urban waterways of the Birmingham, Alabama, metropolitan area. These data are critical in establishing a baseline for future comparisons. In addition, the project aims to provide opportunities to citizen scientists to participate in the fieldwork to increase public awareness of local and state turtle diversity. Herein, some of the early findings from this effort are described, and some insights into the benefits and drawbacks of developing a citizen science project are provided.

METHODS

Sampling for turtles occurred at 11 urban sites in the Cahaba River system, including 6 sites on the Cahaba River and 5 sites on Cahaba River tributaries (Pinchgut Creek and Shades Creek). Sampling also occurred at 1 site in Valley Creek, a tributary of the Black Warrior River (Fig. 1). The Birmingham metropolitan area lies in both watersheds, which are separated by Red Mountain. While all sites were sampled during the larger volunteer weekends occurring in May 2018, October 2018, July 2019, and October 2019, some sites in the Cahaba River were sampled more often from March/April to October in 2018 and 2019. Turtles were collected using baited hoop nets (Memphis Net and Twine, Memphis, TN) and collapsible minnow traps (Promar, Gardena, CA), inverted hoop nets for basking turtles (Lindeman 2014), and opportunistic hand captures. Once captured, morphological measurements were collected, and this included straight-line carapace length (SCL), straight-line carapace width (SCW), plastron length (PL), and shell height (SH). In addition, head width (HW), defined as widest point of the head, was collected on Alabama map turtle (Graptemys pulchra) individuals. Measurements were taken using a 40-cm Haglöf Mantax tree caliper or a 15-cm SPI dial caliper (Forestry Suppliers Inc., Jackson, MS), depending on the size of turtle. After measurements were collected, a cable tie tag with molded plastic containing a unique number (Hallprint Fish Tags, Hindmarth Valley, Australia) was inserted through a hole drilled in a marginal scute on the right side of the tail in hard-shelled turtles (Coleman 2011). For softshell turtles, a unique number was applied to underside of the carapace with a cordless, battery-operated tattoo wand (Inkinator, KW Cages, Santee, CA; Munscher et al. 2015a). Sex was determined when possible based on presence or absence of secondary sexual characteristics (e.g., tail size and the degree to which the cloaca extends beyond the edge of the carapace, and long foreclaws). Once the processing was completed, turtles were released at their capture location.

View Figure

• Download as Powerpoint
• Download Figure

To assess clutch size in adult female G. pulchra, captured females were palpated in the inguinal area to determine if the females were gravid (Donini et al. 2017). If gravid, they were transported to local veterinary offices (Alford Avenue Animal Hospital and Riverview Animal Clinic, Birmingham, Alabama) for radiographic imaging to determine clutch size (Gibbons and Greene 1979). Settings for the radiographic imaging were 30.0 mAs (milliampere-seconds), 300 mA (milliamperage), and 50 peak kV.

A chi-square goodness-of-fit test compared expected and observed sex ratio of G. pulchra captures. A 1-way analysis of variance tested for differences in HW between adult G. pulchra males and females with SCLs that overlapped with measured SCLs of adult males. Statistical analyses were run in Microsoft Excel, and alpha was set at 0.05 for all statistical procedures.

RESULTS

Over the 4 sampling weekends in 2018 and 2019, 52 citizen scientists volunteered over 200 hrs in the field. Of the 52 volunteers, 16 attended multiple sampling weekends. In addition to assisting in sampling, volunteers also interacted with various passersby that inquired about the activities of the project group.

To date, a total of 232 turtles have been captured with 192 of those being tagged and 19 being too small to tag (Table 1). Most turtles were captured in the Cahaba River (n = 169) and its tributaries of Pinchgut (n = 19) and Shades (n = 28) creeks. An additional 17 turtles were captured in the Black Warrior River tributary of Valley Creek. The discrepancy between watersheds did not reflect a true difference in overall turtle abundance between the sites but rather a difference in sampling effort as mentioned above.

Table 1. Number of total tagged, total not tagged, and recaptures by species. Turtles were not tagged if deemed too small for shell tag.

View Fullscreen

Of the 192 tagged turtles, 14 were recaptured at least once. The species encountered the most was the Alabama map turtle (G. pulchra; n = 70, recaptures = 13), and all of the Alabama map turtles were captured at 4 Cahaba River sites. The second and third most encountered species were the pond slider (Trachemys scripta; n = 57, recaptures = 3) and stripe-necked musk turtle (Sternotherus peltifer; n = 24, recaptures = 2). Overall, 10 species were encountered, including a female Barbour's map turtle (Graptemys barbouri), a nonnative species to this area. No flattened musk turtles were detected (Sternotherus depressus), which is a species that could have been encountered at the Valley Creek site.

Although they did not represent a large number of captures, 6 alligator snapping turtles (Macrochelys temminckii) were sampled, all from 3 Cahaba River sites. The mean SCL was 19.1 ± 10.5 cm SD (range, 4.8–39 cm). All of these captures were juvenile and subadult individuals, suggesting a reproducing population in this urban setting.

Of the 70 G. pulchra individuals captured (excluding recaptures), a female-biased sex ratio in captures was observed: 42 were females, 21 were males, and 7 were juveniles (Table 2; n = 63, χ² goodness-of-fit = 7.0, df = 1, p = 0.008). The female mean SCL and PL were 15.0 ± 6.9 cm SD (range, 6.64–27.1 cm) and 13.3 ± 5.9 cm SD (range, 5.85–24.7 cm), respectively. The male mean SCL and PL were 9.2 ± 1.2 cm SD (range, 7.04–12.16 cm) and 8.3 ± 1.1 cm SD (range, 6.37–10.6 cm), respectively. When compared with the head widths of males (mean HW = 1.64 cm), females that displayed SCLs between 6.6 and 11 cm (which coincided with the SCLs of measured males) had significantly wider heads (mean HW = 1.94 cm; F_1,26 = 8.97, p = 0.006) (Fig. 2).

Table 2. Mean ± SD straight-line carapace length (SCL) and mean ± SD straight-line plastron length (PL) and the respective ranges of juvenile, male, and female Alabama map turtles (Graptemys pulchra).

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

Six captured G. pulchra females were determined to be gravid and subsequently radiographed (Table 3). The mean clutch size was 9.2 ± 2.8 eggs SD (range, 4–13 eggs). The earliest clutch was detected in mid-May, and the latest clutch in late June.

Table 3. Clutch sizes based on radiographs obtained of gravid female Alabama map turtles (Graptemys pulchra). Mean clutch size was 9.2 ± 2.8 eggs SD.

View Fullscreen

DISCUSSION

Alabama Aquatic Biodiversity and Citizen Science. — Aquatic ecosystems are experiencing a high degree of threats (habitat destruction, pollution, invasive species, resource overexploitation, urbanization, etc.) on a local, regional, and global scale (Gangloff et al. 2016). These threats have negatively impacted aquatic ecosystems of the southeastern United States, systems that support an impressive amount of freshwater biodiversity (Lydeard and Mayden 1995; Elkins et al. 2019). For instance, Alabama leads the nation in biodiversity of freshwater fish, crayfish, and mussels, but 48 species of freshwater mussels in the state are federally listed (Duncan 2013). As stated above, Alabama also hosts the highest diversity of turtle species; however, a number of the state's aquatic chelonian species has some level of protection or conservation concern. Two species are considered federally Threatened/ Endangered (Pseudemys alabamensis and S. depressus), and several additional species are considered Species of Greatest Conservation Need (Graptemys barbouri, Graptemys ernsti, Malaclemys terrapin pileata, and Sternotherus carinatus) or Moderate Conservation Concern (M. temminckii, Deirochelys reticularia reticularia, Graptemys nigrinoda delticola, Graptemys ouachitensis ouachitensis, G. pulchra, Kinosternon baurii, Apalone calvata, Apalone ferox, and Apalone mutica; Alabama Department of Conservation and Natural Resources, Division of Wildlife and Freshwater Fisheries 2015).

Increasing public awareness is one strategy to combat these threats to biodiversity, and the number of conservation-focused citizen science projects has greatly increased over recent decades to varying degrees (Rotman et al. 2014). Personal interests, potential professional benefits, and a sense of being involved in a worthwhile endeavor were identified to be initial reasons for citizen scientists to engage a project (Rotman et al. 2014). However, to be an engaged long-term participant, one needed to feel trust with the scientist(s) running the project, have meaningful communication with project leaders, receive proper recognition, and in some cases, receive relevant mentor-ship (Rotman et al. 2014). During the volunteer weekends for this project, participants were directly involved with setting and checking traps, performing visual surveys, and collecting data. It was encouraging that several attendees returned for additional sampling weekends, and this was indicative that they valued their experience enough to volunteer at additional events. In fact, 7 volunteers attended 3 or more sampling weekends. During sampling weekends, one task that further engaged participants was interacting with the public. They interacted with various passersby that inquired about the activities of the project group, thereby communicating the goals and importance of the project to other members of the public.

Citizen science efforts can also document information that may have eluded scientists. For example, during the study, an adult female Barbour's map turtle (G. barbouri) was captured in the Cahaba River. However, the natural range of the species is only in the Apalachicola and Chattahoochee river systems of eastern Alabama, western Georgia, and panhandle Florida. It seems likely that this individual represented a released pet. The female was gravid with 10 eggs, but it was not released. It was donated to a local environmental education center (Turtle Point Science Center, Flomaton, Alabama). This study was also the first to verify a reproducing, urban population of M. temminckii in the Cahaba River. Although healthy populations of both M. temminckii and G. pulchra exist in the state, they are considered by the state to be of Moderate Conservation Concern (Alabama Department of Conservation and Natural Resources, Division of Wildlife and Freshwater Fisheries 2015) due to historical overharvesting and habitat degradation and the potential for current poaching (Guyer et al. 2015). Even though the range of S. depressus, a federally protected species, is restricted to permanent streams in the upper reaches of the Black Warrior River system (inclusive of Valley Creek; Guyer et al. 2015), none were detected in our Valley Creek sampling. The absence of this species is not surprising because of their widespread and catastrophic decline throughout their range over the last 50 yrs (Scott and Rissler 2015). However, their absence could be attributed to the larger environmental issues threatening freshwater systems in our study area, including sedimentation and point and nonpoint source pollution.

A methodological weakness of the current study was absence of reported catch per unit effort (CPUE) data. CPUE can indicate relative population abundance (Sutherland 2000) and can provide a useful geographical and temporal comparisons between populations (Selman and Jones 2017; Huntzinger et al. 2019; Van Dyke et al. 2019). However, CPUE can be difficult to estimate in visual surveys or snorkel surveys with the use of citizen scientists because effort may not be standardized in these surveys. Nevertheless, CPUE in trapping effort will be calculated in the future for critical population comparisons.

The inspiration for the Urban Turtle Project came from the work of the North American Freshwater Turtle Research Group, which surveys turtle populations in various freshwater springs and spring-fed streams in Florida and Texas (Munscher et al. 2013; https://turtlesurvival.org/category/blog-naftrg/). These researchers have been surveying some their sites for over 15 yrs, and their long-term data sets are providing important insight into the ecology and life history of various freshwater species. Seemingly healthy populations of peninsula cooters (Pseudemys peninsularis), Florida red-bellied cooters (Pseudemys nelsoni), and loggerhead musk turtles (Sternotherus minor) from Blue Spring State Park in Volusia County, Florida, were observed (Riedle et al. 2016); however, their multiyear data indicated that different factors, such as habitat degradation and competition from various native and nonnative species, may be resulting in lower recruitment and shifts to other habitat (Riedle et al. 2016). Walde et al. (2016) reported record sizes in common snapping turtles (Chelydra serpentina) that were caught over a 15-yr period and hypothesized that the optimal growing conditions provided by freshwater springs supported this record growth. Impressive growth rates in P. peninsularis (30.38–74.28 mm/yr) and P. nelsoni (19.96–42.15 mm/yr) were documented in a freshwater spring community in central Florida (Munscher et al. 2015b). Although these 2 species can be relatively common in their range, little life history information has been gathered on these historically understudied species (Lovich and Ennen 2013; Munscher et al. 2015b).

Graptemys pulchra Ecology. — Graptemys pulchra is one of the most poorly studied species in the United States (Lovich and Ennen 2013), and Lovich et al. (2014) stated that research is needed in all areas to better understand it. One area of information that is lacking is understanding population demographics including size class distributions and sex ratios. This project amassed one of the largest demographic and ecological data sets on G. pulchra, and it provided insights that changed our understanding of the species.

This study's captures of G. pulchra were highly female-biased, but it is unclear if the populations in the Cahaba River are indeed this biased. Most G. pulchra turtles were captured with basking traps, and this method may be more conducive to trapping females. Males may have basked more in areas where the basking traps used in this study could not be set. Lindeman (2016) observed a slightly male-biased sex ratio (males, n = 23; females, n = 19) in captures from the Alabama River, and unbaited fykenets and basking traps were the main trapping methods. However, the author did not state if either method was more successful at trapping a particular sex (Lindeman 2016). More research is needed to verify if the Cahaba River population sex ratio is reflected in the observed capture sex ratio.

Lindeman (2016) reported a mean adult female PL of 19.96 cm, with a range of 18.8–20.7 cm, and the largest previously reported PL (22.2 cm) for this species was by Lindeman (2008). Lovich et al. (2014) reported the maximum carapace length of 27.3 cm, but did not include a maximum PL. In the current study, an adult female from the Cahaba River was measured to be 24.7 cm PL; 6 adult females were also captured with PL measuring over 22.2 cm. The largest recorded SCL in this study was a recapture that measured at 27.9 cm. Like other Graptemys species, G. pulchra display dramatic sexual size dimorphism, with adult females growing to a much larger size than adult males, leading to differences in feeding ecology (Lindeman 2013). With their larger heads, adult females can feed on a larger variety of prey, including mollusks. The major molluscan prey observed in fecal samples from adult females were invasive Corbicula clams (A.T.C., unpubl. data, 2020), and the same has been documented in other populations (Lindeman 2016). The much smaller adult G. pulchra males have been reported to also feed on Corbicula clams as well as aquatic insects (Lindeman 2016). In this population, juvenile females displayed significantly wider heads than adult males, suggesting that differences in feeding ecology may begin in the juvenile stage for females. Interestingly, although the abundance of native freshwater mussels has drastically decreased in this region, this species has been successful in exploiting the invasive Corbicula clams. These invasive clams have been shown to negatively affect native freshwater mussels (Ferreira-Rodríguez et al. 2018).

Little is also known about the reproductive ecology of G. pulchra. Mount (1975) reported a range in clutch sizes of 4–6 eggs in this species (location not reported), and Lindeman (2020) reported a mean of 5.4 eggs and a range of 4–7 eggs from radiographed G. pulchra females inhabiting the Alabama River. A larger mean clutch size (9.2 ± 2.8 eggs SD) and a wider range (4–13 eggs) were observed in the current study. Many of this study's radiographed females were near record size and larger than those radiographed by Lindeman (2020; PL range 18.5–20.7 cm), thus they could be more fecund than smaller females. Nevertheless, there are limitations in using only palpation and radiography as some reproductively active females could have been misidentified (Donini et al. 2017). Yet, radiography is a common method to obtain clutch size data in turtles (e.g., Averill-Murray et al. 2018; McKnight et al. 2018), and obtaining additional clutch size data is important to better understand the true potential reproductive output in G. pulchra females.

Conclusions and Future Directions. — Chelonians are a highly threatened group, with approximately 61% of the 356 recognized species being extinct or possibly facing extinction in the near future (Lovich et al. 2018). This potential loss of so many species could have serious impacts on various ecosystem services they provide, yet the current decline in turtle species is not properly appreciated and understood by the public (Lovich et al. 2018). Without studying these species in their different habitats, their complete ecological roles will remain unknown. For instance, alligator snapping turtles represent apex predators feeding on a variety of different prey, and they may also play a role in seed dispersal and germination in riparian habitats (Lovich et al. 2018). A potentially healthy, urban population of M. temminckii was recently documented in the populous Houston, Texas, metropolitan area, and this population could provide opportunities to study this species' means of survival under the pressures of urbanization (Munscher et al. 2020). Although this project in the Birmingham metropolitan area has established an initial baseline for these urban populations, more work must be done to investigate these populations. By involving the citizen scientists in hands-on field research as is being done through the Urban Turtle Project, the awareness of participants regarding turtle conservation can be increased and spread throughout their diverse communities, while scientists can document information to help conserve these imperiled species.