In the present study, we used a LEK and population genetics approach to better understand the demographic trajectories of the leatherback and the olive ridley turtles. Our combined approach allowed uncovering an undescribed lineage for the leatherback turtle, which may be in high risk of extinction. Thus, our study highlights the importance of considering local knowledge in the study of biodiversity and corroborating this information with appropriate methods.
We registered lower levels of genetic diversity in terms of polymorphic sites, π, and Hd for L. olivacea compared to those of other studies25. In contrast, we found high genetic diversity levels for D. coriacea. This was not expected given that the population of D. coriacea is currently decreasing while that of L. olivacea is increasing3,8. The values of nucleotide diversity for the leatherback turtle (π = 0.07917) were 72 times higher than the value for L. olivacea (π = 0.00109) and higher than those reported by Dutton et al.22 (π = 0.0015), Vargas et al.24 (π = 0.0011), and Dutton et al.23 (π = 0.0006–0.0032). These high levels of genetic diversity were clearly caused by nucleotide differentiation between the two groups of haplotypes (124 polymorphic sites). When we considered each lineage separately, the nucleotide diversity values were consistent with those obtained for the olive ridley turtle and with those that have been previously reported for the leatherback22,23,24.
The haplotype diversity value for lineage B (π = 0.714) is very similar to the value reported by Dutton et al.22 (π = 0.712) for Mexiquillo beach in the Mexican Pacific. The nucleotide diversity of Lineage A (0.00034) was an order of magnitude lower than that of lineage B (0.00176), which concurs with what has been reported previously22,23,24. These values suggest that lineage A may be less diverse than lineage B, and thus individuals from lineage A may be more vulnerable than leatherback turtles from lineage B.
The latter is further supported by interviewees reporting one morphotype that is rarely seeing.
The high and surprising genetic diversity of D. coriacea was associated with the presence of two lineages that are genetically distinct and reflect different evolutionary histories and demographic changes. The interviewees recognized the presence of a leatherback turtle variant, which was supported by the genetic information generated in this study. The results suggest the presence of a different species or subspecies of leatherback turtle on the beaches of Oaxaca that diverged ~ 13.5 Mya. Nonetheless, further morphological analyses are necessary to confirm that this lineage constitutes a distinct species, along with genetic analyses based on nuclear molecular markers such as microsatellite loci or single nucleotide polymorphisms to test for reproductive isolation between the two lineages. In addition, one interviewee mentioned differences in the number of eggs laid for each morphotype, as well in the size of the eggs and the hardness of the eggshell. These observations have to be considered in future study designs in order to be confirmed. On the other hand, the values of Hd (0.584) and π (0.00109) for L. olivacea are lower than those that have been previously reported in other locations of the Mexican Pacific (Hd = 0.6048 and π = 0.0022 in Ceuta, Sinaloa; Hd = 0.6190 and π = 0.0023 in El Verde, Sinaloa; and Hd = 0.6800 and π = 0.0029 in Ixtapilla, Michoacán)25. The lower levels of genetic diversity found in this study could be related to the exploitation of this species in the region. Even if the populations of this species are currently increasing, the recovery of genetic diversity after a bottleneck would depend on the magnitude and duration of a bottleneck26. As marine turtles have long generation times, their effective population size, and their genetic variation, would take a longer time to recover.
Fu’s Fs values for leatherback lineage A (0.177) and lineage B (0.671) are consistent with a signal from a population bottleneck. In contrast, the olive ridley turtle value (− 1.75) agrees with those expected for a species experiencing a recent population expansion. and is further supported by a starlike shape in the haplotype network. Tajima’s D on the other hand was not significant for either leatherback lineage or for the olive ridley turtle. Nevertheless, Tajima’s D is not as sensitive to demographic changes as Fu’s Fs, thus a positive value (even if not significant) for both leatherback lineages supports the existence of a historic bottleneck, while the negative value for the olive ridley turtle suggests population expansion27.
In the case of the haplotype networks for the leatherback turtle, large genetic distance was observed between lineage A (reported for the first time in this study) and lineage B (previously reported with data collected in different areas of the world; Fig. 2). Both lineages are distributed on Barra de la Cruz and Cahuitán. The lack of resolution in the haplotype network within each of the two leatherback lineages (observed loops) may be related to the high mutation rate for the mtDNA D-loop and the small sample size given that we are describing two different lineages. We recommend increasing the sample size of both variants, including samples from Central America, and to use other molecular markers for better resolution. The haplotype network of the olive ridley turtle showed a star-like shape, which, as previously mentioned, has been related to population expansions.
The dates of divergence from the molecular clock were consistent with those previously reported by Duchene et al.20. The molecular clock analysis suggests that leatherback lineages A and B diverged approximately 13.5 Mya (Fig. 3). This event is older than the divergence times of other sister species, such as those in the Lepidochelys genera, which occurred ~ 4.75 Mya, and those of C. mydas and N. depressus, which separated ~ 7.95 Mya. However, this divergence time is similar to that of the Caretta and Lepidochelys genera (14.55 Mya). The molecular clock analysis also revealed that the haplotypes of lineage B in the eastern Pacific are closely related to those found in the Italian (Ity), South Korean, and Colombian Caribbean (CC) areas, whereas lineage A has evolved as an independent group (Fig. 3). Nevertheless, since mtDNA is maternally inherited, we cannot rule out the possibility of gene flow or introgression between these groups until nuclear genetic markers are analyzed. The Eastern Tropical Pacific extends from the Gulf of California to Peru, and encompasses four biogeographic regions: the Cortez Province, the Mexican Province, the Panamic Province and the Galapagos Province28. Our sampling area falls within the Gulf of Tehuantepec, which represent the southern limit of the Mexican Province. Future studies should delineate the distribution of each lineage and determine if the two lineages are sympatric south of the Gulf of Tehuantepec or they are allopatric, and the coast of Oaxaca constitutes an area of range overlap.
In his “Natural History of the Reptiles of Bermuda,” Garman10 mentioned that leatherback turtles from different oceans were morphologically distinct. Garman described two different species of leatherback turtle, Sphargis coriacea (Sphargis now correspond to Dermochelys) and S. schlegelii, although no descriptions were given for S. schlegelii. Later, one S. schlegelii specimen was included in a herpetological list of the Isthmus of Tehuantepec11, which is in the same region as our study area.
In 1899, Philippi12 proposed a new species of Sphargis nominated S. angusta. The specimen was collected near Tocopilla, Chile. The description of this new species indicated that the shell was narrower and darker than that of S. coriacea with spots that were yellowish or hardly visible. The description also indicates that the hind flippers were proportionally smaller than those of S. coriacea while the neck and tail were longer and the back flippers pointier. The reptile collection of the National Museum of Natural History of Chile has embalmed specimens of D. angusta and D. coriacea that were preserved in 191629. Until the first half of the twentieth century, D. schlegelii was described as a valid species30,31.
Pritchard13 compiled taxonomic descriptions of both species and considered that there was not enough information to continue referring to them as separate species and thus advised that they should collectively be classified as D. coriacea. More recently, Eckert et al.32 noticed the taxonomic ambiguities that have arisen since 1884 and concluded that no author had made a sufficiently valid claim to either confirm or reject the existence of two different species or subspecies of leatherback turtle. Some authors argue that the morphological differences may have been the result of adaptations to environmental conditions or simple variations between populations22,32. Future research should search for preserved specimens in herpetological scientific collections, so that morphological and genetic data can be correlated, and to confirm the taxonomic status of this taxon.
The results of the genetic diversity analysis, haplotype network, molecular clock analysis, and interviews suggest the existence of two variants of leatherback turtles. However, it is still too early to define if they are two separate species. The morphological characteristics described by local people of both variants agree with older accounts that describe what were considered to be two different species of leatherback turtle in the past (Table 1)10,11,12. Thus, it is urgent to conduct a taxonomic reassessment of the leatherback turtle that includes both population genetics analyses and morphological descriptions of each specimen. Sufficiently robust sampling efforts are needed that cover large areas to determine whether we are currently dealing with one highly endangered leatherback turtle species or two different species. Nuclear genetic analyses are also needed to understand whether both groups have remained separate over 13.5 million years or there has been secondary contact or introgression.
Regarding current population trends and population recovery of these turtles, the threat that the Chilean swordfish fishery5 potentially poses to the entire eastern tropical Pacific leatherback turtle population appears to be only one component of a complex network of factors that currently influence the potential recovery of these sea turtle lineages. The results of our study clearly highlight the need to invest more time and resources to study the ecology, demography, evolutionary history and population genetics of these sea turtles. With better information, we will be able to infer the evolutionary trajectories of both leatherback lineages with more confidence and the potential capacity of the olive ridley turtle to increase its population size even with apparently low levels of genetic diversity.
Conservation genetics/genomics have been largely recognized as a key tool to improve conservation strategies33,34,35 along with local ecological knowledge15,17,36. However, combining local ecological knowledge with genetics appears to be an untapped tool to better understand the dynamics that are impeding the recovery of highly endangered marine animals such as marine turtles. However, global studies have also found similar agreement. For example, Dao et al.21 identified that local people understand the potential genetic connectivity of marine species through ocean currents.
Our study shows, how genetic analyses coupled with LEK, can help us speed our understanding of the complex drivers that are behind the recovery of vulnerable species such as the leatherback turtle in the eastern Pacific. We are aware there is still work to be done to determine if there are indeed two separate species of leatherback turtle. Nonetheless, the information collected in our study indicates that the conservation strategies aimed at the recovery of this species in the Eastern Pacific Ocean will benefit from understanding its evolutionary trajectory and the observations made by local people.