Green Turtles Nesting at Their Northern Range Limit in the United States Represent a Distinct Subpopulation
Abstract
Green turtle nesting has been recorded in North Carolina, since 1980, but how these nesting females fit into the broader regional context genetically has not been determined. Genetic tagging through microsatellite genotyping of clutches laid in northern South Carolina, North Carolina, and Delaware from 2010 through 2014 identified at least 52 individual nesting females. The mitochondrial control region haplotype frequencies of these individuals were significantly different from all northern Greater Caribbean subpopulations, including those in Florida, suggesting that these northern US females represent an incipient subpopulation that warrants distinct management unit status.
Greater Caribbean green turtles suffered severe population declines, including the extirpation of some major rookeries, with the advent of European exploration and subsequent colonization (McClenachan et al. 2006). Florida nesting beaches host the vast majority of green turtle nesting in the southeastern United States, and this nesting aggregation came perilously close to being extirpated during the 20th century due to decades of overharvest (Parsons 1962). During biennial “low years” in the 1980s and 1990s, only tens to hundreds of nests were recorded statewide (Witherington et al. 2006). With protection afforded by the US Endangered Species Act, Florida green turtle nesting has rebounded dramatically over the previous few decades. Index survey data from Florida rookeries indicate an 80-fold increase in nest counts since surveys were standardized in 1989, with a new statewide record of 53,102 nests in 2017 (Florida Fish and Wildlife Research Institute Index Nesting Beach Survey and Statewide Nesting Beach Survey program data). Mitochondrial DNA analyses suggested that green turtles nesting on the Atlantic coast of Florida represent at least 2 subpopulations, also known as management units, with an apparent genetic break in the vicinity of the St. Lucie Inlet (Shamblin et al. 2015).
The demographics and genetics of green turtles nesting on the Atlantic coast of the United States north of Florida have not been characterized. Green turtle nesting in Georgia and throughout much of South Carolina has been considered rare and sporadic prior to increases in nesting activity over the past few years (M. Dodd and M. Pate, pers. comm.). However, low but consistent green turtle nesting has been recorded in North Carolina since 1980 (Schwartz et al. 1981; Woodson and Webster 1999). Here we used a genetic tagging approach to generate preliminary data on the number of unique females nesting at the northern range limit on the Atlantic coast of the United States. We also analyzed mitochondrial DNA from these individuals to assess demographic connectivity with other nesting populations in the northern Greater Caribbean region.
Methods
Beach surveyors recorded green turtle clutches laid in northern South Carolina (Cape Romain and the Grand Strand) and in North Carolina (hereafter NUSA) during the 2010 through 2014 nesting seasons (Table 1; Fig. 1). They obtained genetics samples from a majority of these clutches, either as a single egg collected during the morning survey following oviposition or as dead hatchling tissue and/or an eggshell from an undeveloped egg collected during nest evaluation inventories following hatchling emergence. We conducted sample processing and DNA extractions as previously described for loggerhead turtle eggs (Shamblin et al. 2011b). We genotyped DNA samples at 15 microsatellite loci originally isolated from loggerhead turtles (Shamblin et al. 2007) as previously described (Shamblin et al. 2017a). These loci have been used for a long-term genetic capture-recapture loggerhead turtle project and provided species confirmation. Three of the markers (CcP2G10, CcP5C08, and CcP7C06) failed to amplify or had inconsistent products; 4 additional loci (CcP1G03, CcP1H11, CcP7C04, and CcP7G11) were monomorphic or only weakly polymorphic in green turtles. We used the latter only to confirm species identity. The remaining 8 loci were highly polymorphic in green turtles (Table 2), with a combined nonexclusion probability of identity of 2.99 × 10−12 based on allele frequencies of the individuals identified in this study.
Citation: Chelonian Conservation and Biology 17, 2; 10.2744/CCB-1332.1
We used a matching approach to assign clutches to individual females as previously described for loggerhead turtle genetic tagging based on nest samples (Shamblin et al. 2017a). However, because we used only 8 loci for green turtle individual identification, we tolerated only a single allele mismatch (due to either allele dropout or embryonic DNA contamination) in considering 2 samples as representing the same nesting female. Any samples mismatching at more than 1 locus were regenotyped to clarify. Samples for which genotypes failed to match any others were subjected to a second round of DNA extraction and genotyping to confirm the uniqueness of the individual female. Eggshells or dead hatchling tissue collected after hatchling emergence and any eggshells collected at oviposition that contained obvious paternal contamination (3 or more alleles present) were assigned via parentage analysis. Dead hatchlings that could not be assigned to maternal genotypes via parentage analysis were tentatively assigned nonmaternal identification codes provided that genotypes from multiple extractions perfectly matched. Eggshells collected at oviposition that contained paternal contamination and could not be assigned to known females via parentage analysis were not assigned but were retained in the database for future parentage analyses.
Following individual identification, we amplified and sequenced the mitochondrial control region of a representative sample for each individual female using amplification primers LCM15382 and H950 and sequencing primers LCM15382 and Cm1820 as previously described (Shamblin et al. 2015). We assigned each individual an 817-base pair (bp) haplotype based on standard nomenclature for Atlantic green turtles (Archie Carr Center for Sea Turtle Research mtDNA sequence database: https://accstr.ufl.edu/resources/mtdna-sequences). We compared these NUSA haplotype frequencies to those from all other proposed management units in the northern Greater Caribbean region (Fig. 2): Western Bay of Campeche (Tamaulipas/Veracruz mainland), Mexico (WBCMX); Eastern Bay of Campeche (Campeche/Yucatán mainland), Mexico (EBCMX); Cayo Arcas, Campeche, Mexico (CAMX); Scorpion Reef, Yucatán, Mexico (SRMX); Quintana Roo, Mexico (QRMX); southwestern Cuba (SWCB); southeastern Florida and the Keys (SOFL); and central eastern Florida (CEFL) (Pérez-Ríos 2008; Millán-Aguilar 2009; Ruiz-Urquiola et al. 2010; Shamblin et al. 2015). We tested for population structure by calculating pairwise FST values and exact tests of population differentiation using the program Arlequin (Excoffier and Lischer 2010). SWCB data were available only as 490-bp haplotypes, so pairwise comparisons involving Cuba were restricted to 490-bp haplotype frequencies. As an additional test, we compared NUSA haplotype frequencies with published 490-bp haplotype frequencies for juveniles foraging in the adjacent inshore waters in North Carolina (Bass et al. 2006); p-values for pairwise comparisons were corrected for multiple comparisons using a false discovery rate approach (Benjamini and Yekutieli 2001).
Citation: Chelonian Conservation and Biology 17, 2; 10.2744/CCB-1332.1
Results and Discussion
Surveyors recorded 138 green turtle nests during the 2010 through 2014 nesting seasons and collected genetics samples from 123 of these (Table 1). Microsatellite genotyping and assignment analyses identified a minimum of 52 individual females nesting in the region over this 5-yr period (Table 1). Allele frequency analyses of the maternal genotypes (excluding 2 hatchling samples tentatively assigned to unique mothers) indicated that locus CcP5F01 deviated significantly from Hardy-Weinberg expectations due to a heterozygote deficiency (Table 2). The substantial null allele estimate for this locus suggests a priming problem (possibly due to a mutation in the primer-binding site) that would make this marker unsuitable for population analyses. However, this issue should not affect inferences of identity or parentage, as these null alleles should behave consistently in closely related individuals.
Seven 817-bp mitochondrial control region haplotypes were present among NUSA nesting females (Table 1). Pairwise FST values and exact tests of population differentiation between NUSA and all genetically characterized nesting populations in the northern Greater Caribbean region were significant (Table 3). Despite complete haplotype sharing between NUSA and the Florida nesting aggregation, the significant frequency differences suggest that NUSA represents a demographically isolated subpopulation. Haplotype frequencies for the North Carolina juvenile foraging aggregation were significantly different from all nesting populations except NUSA (Table 3).
Green turtles and loggerhead turtles nesting on the Atlantic coast of Florida share a similar population structure with 2 common haplotypes dominating (Shamblin et al. 2012, 2015). In both species, frequencies of these 2 common haplotypes are essentially inverted between southeastern and central eastern Florida. For loggerhead turtles, northern US nesting areas represent a distinct subpopulation where the dominant haplotype in the central eastern Florida management unit is nearly fixed (Shamblin et al. 2011a, 2012). This pattern was hypothesized to have arisen via northerly stepping-stone colonization as the thermal profile of the beaches in the region became warm enough to support embryonic development (Bowen et al. 1993). However, green turtle nesting densities on the Atlantic coast of the United States north of Florida and the genetics results presented here are not consistent with a pattern of stepping-stone colonization. Although the precise origins of the NUSA nesting females are obscured by poor marker resolution, the lack of differentiation between the NUSA rookery sample and the juvenile aggregation foraging in adjacent waters suggests a possible link. Phylogeographic analyses of loggerhead and green turtle populations in the Atlantic and Mediterranean basins suggest that the eastern Mediterranean was colonized by northwestern Atlantic lineages of each species (Bowen et al. 1992, 1993). Both species undergo a surface-pelagic stage during which neonates departing from northwestern Atlantic beaches enter the Gulf Stream and are transported into the eastern Atlantic (Bolten et al. 1998; Monzón-Argüello et al. 2010). Although contemporary entry into the western Mediterranean by Atlantic green turtles is not common (Carreras et al. 2014), historically this may have been more likely when Bermuda supported a large nesting population “downstream” of all others in the Greater Caribbean region and proximal to the Gulf Stream (Shamblin et al. 2015). By contrast, juvenile loggerhead turtles of northwestern Atlantic origin are common in western Mediterranean foraging areas (Carreras et al. 2011). Despite this extensive surface pelagic dispersal, juveniles of both species typically recruit to foraging areas in the vicinity of their natal regions as adults (Bowen et al. 2004; Bass et al. 2006).
Although regional natal homing may be the rule for female marine turtles, strict natal philopatry at the population scale over evolutionary time would be detrimental (Bowen and Karl 2007). Recent analyses of sporadic loggerhead turtle nests in the western Mediterranean basin have indicated that many of these events likely represent colonization by Atlantic individuals that were present as juveniles in adjacent foraging habitats (Carreras et al. 2018). We propose a similar hypothesis to explain the colonization of nesting habitats in the northern latitudes of the Atlantic coast of the United States. Many high-latitude foraging sites in the North Atlantic appear to serve as developmental habitats that currently support juvenile green turtles but not adults (Epperly et al. 1995; Morreale and Standora 1998; Barco and Swingle 2014). As an example, juvenile green turtles appear to depart Bermudan waters as they reach 60–65-cm straight carapace length for foraging sites farther south in the Gulf of Mexico and Caribbean Sea (Meylan et al. 2011). Juvenile green turtles have regularly occupied the inshore waters of the Pamlico-Albemarle estuarine complex since at least the 1800s (True 1884). Their relative abundance in North Carolina estuarine waters has increased in the past few decades as evidenced from bycatch and stranding data (Epperly et al. 2007; Byrd et al. 2011; North Carolina Wildlife Resources Commission, unpubl. data). Previous mixed stock analyses based on 490-bp data suggested that these juveniles were most likely from Quintana Roo and Florida (Bass et al. 2006). Therefore, juvenile foraging aggregations “downstream” of natal populations likely facilitate colonization of adjacent novel nesting habitats as a result of less strict natal homing (Carreras et al. 2018).
Whether the recent increase in green turtle nesting across this study area reflects recruitment from a small number of founder females or continued straying from other nesting populations, likely via the adjacent juvenile foraging aggregation, remains to be determined. The 8 microsatellite loci used for individual identification in the present study were insufficient to resolve relatedness among the nesting females, so distinguishing between these scenarios will require a deeper genotyping effort. A combination of relatedness analysis of reconstructed maternal genotypes and mitochondrial DNA sequencing suggested that recent green turtle nesting on the main Hawaiian Islands represented recruitment from a small number of founders from the French Frigate Shoals nesting aggregation (Frey et al. 2013). Taking a similar approach with the NUSA females may further resolve their origins, particularly with application of additional mitochondrial markers to overcome poor marker resolution. For example, mitogenomic sequencing revealed fixed differences between the CM-A1.1 lineages nesting in Tamaulipas, Mexico, and Florida (Shamblin et al. 2017b), but baseline data for this marker are needed from Quintana Roo and Cuban rookeries for more robust context.
Our results demonstrate the utility of genetic tools for making population inferences where traditional tagging studies would be impractical due to low nesting densities and the expansive nesting habitat. Population genetic studies often focus on the largest populations for obvious reasons, but genetic analyses of smaller, peripheral populations have offered valuable insight into demographic processes that are often difficult to infer in large populations (Frey et al. 2013, 2014; Carreras et al. 2018). Colonization of novel nesting habitats is one compensatory mechanism that may help marine turtles adapt to altered thermal and precipitation regimes driven by global climate change. Sampling these northern nesting habitats now, near the inception of the rookery, provides a valuable baseline against which future genetic data can be compared and should ultimately facilitate the characterization of the demographics underlying the colonization process.
Northernmost recorded nesting habitats for green turtles along the Atlantic coast of the United States. The map was created using the maptool function at http://seaturtle.org.
Locations of green turtle management units (indicated by stars) in the northern Greater Caribbean region and the inshore North Carolina juvenile foraging aggregation (NC, indicated by the triangle). WBCMX: western Bay of Campeche, Mexico; EBCMX: eastern Bay of Campeche, Mexico; CAMX: Cayo Arcas, Mexico; SRMX: Scorpion Reef, Mexico; QRMX: Quintana Roo, Mexico; SWCB: southwestern Cuba; SOFL: southern Florida, USA; CEFL: central eastern Florida, USA; NUSA: northern South Carolina and North Carolina, USA. The map was created using the maptool function at http://seaturtle.org.
Contributor Notes
Handling Editor: Jeffrey A. Seminoff
