Endangered and Threatened Wildlife; 12-Month Finding on a Petition To Identify the Northwest Atlantic Leatherback Turtle as a Distinct Population Segment and List It as Threatened Under the Endangered Species Act
Federal RegisterAug 10, 2020
Ask Donna
What actually matters in this document.
Text
DEPARTMENT OF INTERIOR
Fish and Wildlife Service
50 Part 17
DEPARTMENT OF COMMERCE
National Oceanic and Atmospheric Administration
50 CFR Parts 223 and 224
[Docket No. 200717-0190]
RIN 0648-XF748
Endangered and Threatened Wildlife; 12-Month Finding on a Petition To Identify the Northwest Atlantic Leatherback Turtle as a Distinct Population Segment and List It as Threatened Under the Endangered Species Act
AGENCY:
National Marine Fisheries Service (NMFS), National Oceanic and Atmospheric Administration (NOAA), Commerce; U.S. Fish and Wildlife Service (USFWS), Interior.
ACTION:
Notification of 12-month petition finding.
SUMMARY:
We, NMFS and USFWS, announce a 12-month finding on a petition to identify the Northwest Atlantic population of the leatherback turtle (
Dermochelys coriacea
) as a distinct population segment (DPS) and list it as threatened under the Endangered Species Act (ESA). In response to the petition, we completed a comprehensive status review of the species, which also constitutes the 5-year review of the species, to determine potential DPSs following the Policy Regarding the Recognition of Distinct Vertebrate Population Segments Under the ESA and to perform extinction risk analyses. Based on the best scientific and commercial data available, including the Status Review Report, and after taking into account efforts made to protect the species, we conclude that seven populations would meet the discreteness and significance criteria for recognition as DPSs, including the Northwest Atlantic population. However, even if we were to list them separately, all seven DPSs would meet the definition for endangered species (
i.e.,
they are in danger of extinction throughout all or a significant portion of their range). The species is already listed as endangered throughout its range. We have determined that the listing of DPSs is not warranted, and therefore we do not propose any changes to the existing global listing.
DATES:
This finding was made on August 10, 2020.
ADDRESSES:
The Status Review Report are available on NMFS' website at
https://www.fisheries.noaa.gov/species/leatherback-turtle.
FOR FURTHER INFORMATION CONTACT:
Jennifer Schultz, NMFS Office of Protected Resources, (301) 427-8443,
jennifer.schultz@noaa.gov.
Persons who use a Telecommunications Device for the Deaf (TDD) may call the Federal Information Relay Service (FIRS) at 1-800-877-8339, 24 hours a day and 7 days a week.
SUPPLEMENTARY INFORMATION:
Background
The leatherback turtle species as a whole was listed as an endangered species (one determined to be threatened with worldwide extinction) (35 FR 8491; June 2, 1970), under the Endangered Species Conservation Act of 1969, the precursor statute to the ESA (16 U.S.C. 1531
et seq.
). When the ESA was enacted in 1973, it specifically provided for continuity with the lists previously in effect under the Endangered Species Conservation Act. Section 4(c)(3) of the ESA directed that species on the lists of endangered foreign or native wildlife at the time the ESA took effect would be deemed “endangered species” under the ESA without interruption.
See
39 FR 1444 (January 9, 1974) (explaining transition provisions); 39 FR 1158, 1172 (January 4, 1974) (setting out the final list of “endangered foreign wildlife,” including “Turtle, Leatherback” at 50 CFR 17.11).
On September 20, 2017, the Blue Water Fishermen's Association petitioned NMFS and USFWS (together, the Services) to identify the Northwest (NW) Atlantic leatherback turtle population as a DPS and to list it as threatened under the ESA. On December 6, 2017, NMFS published a “positive” 90-day finding in the
Federal Register
(82 FR 57565) announcing the determination that the petition presented substantial information indicating that the petitioned action may be warranted. At that time, NMFS also solicited information on leatherback turtles and announced that it would commence, jointly with USFWS, a status review of the entire listed species, pursuant to ESA section 4(b)(3)(A) and 50 CFR 424.14. The resulting Status Review Report includes all information used to evaluate the petitioned actions and explains the process followed by the Status Review Team (
i.e.,
the Team). The following summarizes that information; for additional details, please see the Status Review Report (see
ADDRESSES
).
ESA Statutory, Regulatory, and Policy Provisions and Evaluation Framework
Under the ESA, the term “species” includes any subspecies of fish or wildlife or plants, and any DPS of any vertebrate fish or wildlife which interbreeds when mature (16 U.S.C. 1532(16)). The Services adopted a joint policy clarifying their interpretation of the phrase “distinct population segment” for the purposes of listing, delisting, and reclassifying a species under the ESA (“Policy Regarding the Recognition of Distinct Vertebrate Population Segments Under the Endangered Species Act,” 61 FR 4722 (Feb. 7, 1996; “DPS Policy”). The DPS Policy stipulates two elements that must be considered: (1) Discreteness of the population segment in relation to the remainder of the species to which it belongs; and (2) the significance of the population segment to the species to which it belongs.
Section 3 of the ESA defines an endangered species as any species which is in danger of extinction throughout all or a significant portion of its range and a threatened species as one which is likely to become an endangered species within the foreseeable future throughout all or a significant portion of its range (16 U.S.C. 1532(6) and (20)). Thus, we interpret an “endangered species” to be one that is presently in danger of extinction. A “threatened species,” on the other hand, is not presently in danger of extinction, but is likely to become so within the foreseeable future (that is, within a specified later time). In other words, the primary statutory difference between a threatened and endangered species is the timing of when a species may be in danger of extinction, either presently (endangered) or within the foreseeable future (threatened). The ESA uses the term “foreseeable future” to refer to the time over which identified threats are likely to impact the biological status of the species. The duration of the “foreseeable future” in any circumstance is inherently fact-specific and depends on the particular kinds of threats, the life-history characteristics, and the specific habitat requirements for the species under consideration. The existence of threats to a species and the species' response to such threats are not, in general, equally predictable or foreseeable. Hence, in some cases, the ability to foresee a threat to a species is greater than the ability to foresee the species' exact response, or the timeframe of such a response, to that
threat. For purposes of making this 12-month finding, the relevant consideration is whether the species' population response (
i.e.,
abundance, productivity, spatial distribution, diversity) is foreseeable, not merely whether the emergence of a threat is foreseeable. The foreseeable future extends only as far as we are able to reliably predict the species' population response to threats.
Pursuant to the ESA and our implementing regulations, we determine whether a species is threatened or endangered based on any one or a combination of the following ESA section 4(a)(1) factors or threats (16 U.S.C. 1533(a)(1), 50 CFR 424.11(c)):
1. The present or threatened destruction, modification, or curtailment of its habitat or range;
2. Overutilization for commercial, recreational, scientific, or educational purposes;
3. Disease or predation;
4. Inadequacy of existing regulatory mechanisms; or
5. Other natural or manmade factors affecting its continued existence, which could include but are not limited to: Fisheries bycatch; vessel strikes; pollution (including marine debris and plastics, contaminants, oil and gas activities, and derelict fishing gear); natural disasters; climate change; and oceanographic regime shifts.
Section 4(b)(1)(A) of the ESA requires us to make listing determinations based solely on the best scientific and commercial data available after conducting a review of the status of the species and after taking into account efforts being made by any State or foreign nation or political subdivision thereof to protect the species' existence (16 U.S.C. 1533(b)(1)(A)).
Approach to the Status Review
The Services convened a team of NMFS and USFWS biologists (
i.e.,
the Team) to gather and review the best available scientific and commercial data on the leatherback turtle, assess the discreteness and significance of populations by applying the DPS Policy, evaluate the extinction risk of any population segments that meet the DPS criteria, and document all findings in a report (
i.e.,
the Status Review Report). Although the petitioner requested evaluation only of the NW Atlantic leatherback population, we instructed the Team to perform a comprehensive status review to identify and evaluate the status of all potential DPSs.
The Team compiled information on leatherback turtle life history, biology, ecology, demographic factors, and threats. This included the information received in the petition and in response to the
Federal Register
request associated with the 90-day finding (82 FR 57565; December 6, 2017). The Team also requested leatherback nesting data from beach monitoring programs. To evaluate recent abundance and trends, unpublished nesting beach monitoring datasets were often the best available data (
i.e.,
most recent and relevant). The Team assessed these data in terms of standardization (
i.e.,
the use of standardized methodology), consistency (
i.e.,
consecutive seasonal data collection), and duration of data collection (
i.e.,
the number of years that data were collected). When evaluating threats, peer-reviewed information, specifically primary research with large sample sizes and long-term sampling duration, was often the best available data. In some locations, reports from governments or non-governmental organizations and expert opinion constituted the best available information. The Team also addressed the source and magnitude of any uncertainty and the impact on its conclusions.
The Team evaluated the discreteness and significance of each population and provided their evaluation of whether each population would meet the criteria of the DPS Policy. The DPS Policy states that a population of a vertebrate species may be considered discrete if it satisfies one of the following conditions: (1) It is markedly separated from other populations of the same taxon as a consequence of physical, physiological, ecological, or behavioral factors (quantitative measures of genetic or morphological discontinuity may provide evidence of this separation); or (2) it is delimited by international governmental boundaries within which differences in control of exploitation, management of habitat, conservation status, or regulatory mechanisms exist that are significant in light of section 4(a)(1)(D) of the ESA (61 FR 4722, February 7, 1996). While the Team used the term “DPS” in describing and discussing populations that they concluded meet the requirements of discreteness and significance, it is important to note that the DPS term is used throughout the Status Review Report for ease of reference only. A DPS is formally recognized under the ESA only upon a listing action by the Services, and the Services cannot delegate authority to take formal listing actions to status review teams. The information compiled by the Team must be reviewed by the Services, which retain responsibility for making the listing determination after complying with all the requirements of Section 4 of the ESA and considering agency policies. Because we ultimately conclude for the reasons discussed in this finding that it would not be appropriate to disaggregate the existing global listing into DPSs, references in the Status Review Report (and in this finding when we are reviewing the information presented by the Team) must be understood as references to potential or hypothetical DPSs only.
The Team evaluated significance in terms of the importance of the population segment to the overall welfare of the species, such as: (1) Persistence of the population segment in an unusual or unique ecological setting; (2) evidence that loss of the population segment would result in a significant gap in the range of the taxon; (3) evidence that the DPS represents the only surviving natural occurrence of a taxon that may be more abundant elsewhere as an introduced population outside its historic range; or (4) evidence that the population segment differs markedly from other populations of the species in its genetic characteristics.
For each population segment that the Team determined would meet the criteria of the DPS Policy (which the Team and we refer to as a “DPS” for ease of reference), the Team performed an extinction risk analysis, which involved the evaluation of demographic factors and threats. Demographic factors reflect the impact that operative threats have had on the species. In some cases those threats or the impacts from the threats are continuing in nature. The demographic factors included abundance, productivity, spatial distribution, and diversity. Because sea turtles spend the majority of their lives at sea, where they are spread across vast distances, it is difficult to estimate total abundance. However, the number of nesting females can be counted directly, or estimated indirectly by counting the number of nests on beaches, during a nesting season. Females nest more than once in a season (
i.e.,
clutch frequency, which is the average number of nests per season) and do not nest every season (
i.e.,
remigration interval, which is the average number of years between successive nesting seasons). To calculate the index of nesting female abundance at a nesting beach, the Team summed the total number of nests over the most recent remigration interval (
i.e.,
a run-sum) and divided this number by the clutch frequency. The Team performed these calculations only if available data were recent (
i.e.,
last year of the remigration interval occurred in 2014 or more recently), consistent
(
i.e.,
seasonal data collected for each year of the remigration interval), and collected in a standardized manner (
i.e.,
data collection methods remained the same over the remigration interval), as further detailed in the Status Review Report. To provide a total index of nesting female abundance for each DPS, we summed the indices of nesting female abundance for all monitored beaches used by that DPS. The total index of nesting female abundance for each DPS is an index (rather than a census) because not all nesting beaches met these criteria. However, the nesting beaches that were not included were generally unmonitored or not recently monitored because they host few nesting females. Even where data were not sufficient to allow for a calculation of the index of nesting female abundance, the Team provided all available data to ensure the analysis would be as robust as possible.
The Team evaluated the productivity for each DPS by evaluating nesting trends (through trend analyses or bar graphs) and productivity metrics. Where available data allowed it, they estimated the long-term trend for individual beaches using a Bayesian state-space model of stochastic exponential population growth (Boyd
et al.
2017), where the rate parameter describes the annual percent change in observed nest counts (or female counts where applicable) over the period of data collection. This is further explained in the Status Review Report. To reflect current trends over approximately three remigration intervals, the criteria for trend analyses were as follows: Nesting data (
i.e.,
nest or nesting female counts) consistently collected over nine or more years in a standardized manner (for that site), with the most recent data collection in 2014 or later and with a minimum average number of nests of 50 annually. The Team reported the median trend, along with the standard deviation (sd), 95 percent credible interval (CI), and an “f statistic” which is the proportion of the posterior distribution with the same sign as the median (
i.e.,
the confidence that the trend is positive or negative). When the data did not meet the criteria for performing trend analyses, the Team provided bar graphs and/or historical data in the Status Review Report. Based on the trend analysis (where possible) and the best available historical data, the Team characterized the nesting trend for each DPS as decreasing, stable, or increasing. The Team also evaluated the following productivity metrics (if available): Average size of nesting female; nesting female survivorship; remigration interval; clutch size; clutch frequency; internesting interval; incubation period; hatching success (the proportion of eggs in a nest that produce live hatchlings); and sex ratio. Each of these metrics contributes to the growth rate, or reproductive potential, of the population.
For each DPS, the Team evaluated spatial distribution, which included the number and location of nesting beaches and foraging areas, as well as spatial structure (
i.e.,
whether the DPS exists as a single population or several subpopulations connected by metapopulation dynamics). The Team also evaluated diversity, which like spatial distribution, is a measure of resilience. In general, diverse populations with broad spatial distributions and metapopulation dynamics are more resilient to threats and environmental changes than less diverse populations with narrow distributions.
For each DPS, the Team next evaluated each of the ESA Section 4(a)(1) factors (or “threats”) as listed above (16 U.S.C. 1533(a)(1), 50 CFR 424.11(c)). For each threat, the Team used the best available information to describe the threat, identify which life stages are affected, and describe the impact to the DPS with as much specificity as the best available information allowed to link the threat to the demographic factor it affected. The best available data often allow only for qualitative assessment. For each DPS, the Team identified the primary threat(s) to its continued existence, as well as other threats. The Team considered the impact of each threat individually, with the primary threat(s) given the greatest weight, and all threats cumulatively, to determine the extinction risk. To assess confidence in the extinction risk determination, the Team identified any sources of uncertainty and the impact of uncertainty on the conclusions. They analyzed all threats assuming the DPS had lost ESA protections going forward because a DPS would not receive such protections if it was not listed under the ESA. For example, a DPS would not have benefits of section 9 take prohibitions or section 7 consultations on actions that may affect the DPS.
The Team performed an extinction risk assessment for each of the seven DPSs by evaluating the demographic factors and threats, as described above. Then, the Team voted, based on the best available data, on whether the extinction risk of each DPS was high, moderate, or low, following the definitions included in NMFS' internal guidance document, “Guidance on Responding to Petitions and Conducting Status Reviews under the Endangered Species Act, Section II” (
i.e.,
NMFS' Guidance; November 9, 2017) and in the Status Review Report.
After the Team completed its draft Status Review Report, the Services met to review and discuss that document and conservation efforts. The Services based our status determinations of the DPSs on the best scientific and commercial data available (as compiled and reflected in the Status Review Report) and after taking into account efforts by States and foreign nation, or any political subdivision thereof, to protect the species as mandated by the statute.
DPS Analysis
The following is a summary of the DPS analysis conducted by the Team. For a detailed description of the Team's analyses of discreteness and significance, please see the Status Review Report. As a starting point, the Team considered seven leatherback populations that were previously identified as regional management units (RMUs) by Wallace
et al.
(2010) and recognized as subpopulations under the International Union for Conservation of Nature (IUCN) Red List (
https://www.iucnredlist.org/species/6494/43526147
). The Team found that seven leatherback populations met the discreteness and significance criteria per the DPS Policy and identified the following potential DPSs: Northwest (NW) Atlantic; Southwest (SW) Atlantic; Southeast (SE) Atlantic; SW Indian; Northeast (NE) Indian; West Pacific; and East Pacific.
Discreteness
The Team evaluated all populations for discreteness and determined that each showed marked separation from the others as a consequence of behavioral and physical factors. Behavioral factors, especially returning to waters off a turtle's natal beach to breed, have prevented interbreeding, resulting in reproductive isolation, as indicated by genetic discontinuity.
Although some populations use the same foraging areas, tagging and telemetry studies also demonstrate the discreteness of the populations at nesting beaches. Physical factors, such as land masses, ocean currents, and other oceanographic features, have established and reinforced barriers to gene flow among the seven populations.
Genetic data provide the most compelling evidence for discreteness among the seven populations. The most recent and comprehensive global analysis of published and unpublished
mitochondrial deoxynucleic acid (mtDNA) sequence data (
i.e.,
28 haplotypes, which are unique sequences of mtDNA) evaluated samples collected from 21 nesting sites representing key regions from all ocean basins (Dutton
et al.
2007; Dutton
et al.
2013; Shanker
et al.
2011; Dutton and Shanker 2015); analyzing the evolutionary relationship of these data revealed three distinct haplogroups (
i.e.,
similar haplotypes that cluster together, relative to other haplotypes) that are geographically segregated across the Atlantic, Indian, and Pacific Oceans (Dutton, unpublished data; NMFS and USFWS 2020). Early mtDNA analyses indicated strong genetic discontinuity, globally (F
ST
= 0.415,
P
<0.001) and within ocean basins (F
ST
= 0.203 to 0.253,
P
<0.001; Dutton
et al.
1999). Wallace
et al.
(2010) combined these and other genetic data with nesting, flipper tagging, and satellite telemetry data to identify seven leatherback RMUs, which provided the starting point for our identification of discrete populations.
From this starting point, the Team then evaluated more recent genetic data. Subsequent genetic analyses confirmed genetic discontinuity among the NW, SW, and SE Atlantic populations (Wallace
et al.
2010; Dutton
et al.
2013; Carreras
et al.
2013; Molfetti
et al.
2013; Vargas
et al.
2017). Elevated genetic differentiation at nuclear DNA (F
ST
= 0.211−0.86) indicates that males, like females, likely return to the waters off their natal beaches to mate and that male-mediated gene flow may not be as pronounced as previously thought (Dutton
et al.
2013; see Jensen
et al.
2013). Nuclear (F
ST
>0.126,
P
<0.001; Dutton
et al.
2013) and mtDNA (F
ST
>0.061,
P
= 0.05−0.001; Dutton
et al.
2013; F
ST
>0.061,
P
<0.01; Vargas
et al.
2017) analyses indicate genetic discontinuity between the Atlantic populations and the SW Indian population. Preliminary mtDNA results for leatherback turtles nesting at Little Andaman Island, India (Shanker
et al.
2011; Dutton and Shanker 2015), indicate that this population is closely related to the extinct Malaysian population, with which it shares common haplotypes. It is markedly different from the South African nesting population, as well as those in the West Pacific population (Dutton
et al.
2007, 2013 and unpublished). Samples from extant and extirpated nesting aggregations of the NE Indian population (Shanker
et al.
2011; Dutton and Shanker 2015; Dutton
et al.
unpublished data) are genetically differentiated from the SW Indian population (F
ST
= 0.415,
P
<0.003; Dutton
et al.
1999) and the West Pacific population (X
2
= 49.346,
P
= 0.002; Dutton
et al.
2007). There is genetic discontinuity between the West and East Pacific populations, as demonstrated by significant genetic differentiation between the samples from Solomon Islands in the western Pacific and Mexico or Costa Rica in the eastern Pacific (F
ST
= 0.270 and 0.331,
P
<0.001; Dutton
et al.
1999). Genetic discontinuity among all seven populations provides evidence for marked separation from the others and thus discreteness of each population.
Tagging and telemetry studies confirm marked separation of the seven populations because nesting sites remain distant and isolated. Nesting females of one population have not been tracked to, or observed on, beaches used by another population, even though telemetry data indicate shared use of foraging areas by different populations.
Telemetry studies demonstrate that females nesting on NW Atlantic beaches move throughout most of the North Atlantic from the Equator to about 50° N latitude (Ferraroli
et al.
2004; Hays
et al.
2004; James
et al.
2005a; James
et al.
2005b; 2005c; Eckert 2006a; Eckert
et al.
2006b; Hays
et al.
2006; Doyle
et al.
2008; Evans 2008; Dodge
et al.
2014; Fossette
et al.
2014; Aleksa 2017; Aleksa
et al.
2018). Turtles originating from beaches of the NW Atlantic appear to mix at foraging areas throughout the North Atlantic Ocean (Fossette
et al.
2014), but their movements rarely extend into waters south of the Equator. Tagging studies further support the connectivity within and among nesting beaches and foraging areas of the North Atlantic Ocean (Troëng
et al.
2004; Bräutigam and Eckert 2006; Chacón-Chaverri and Eckert 2007; Turtle Expert Working Group (TEWG) 2007; Sönmez
et al.
2008; Dutton
et al.
2013b; Horrocks
et al.
2016), but turtles tagged in the North Atlantic Ocean have never been found on nesting beaches in Brazil (SW Atlantic population) or Africa (SE Atlantic population). In the South Atlantic Ocean, post-nesting females tracked from nesting beaches in Gabon and Brazil use the same foraging areas, including waters off SW Africa, in the south equatorial Atlantic and off SE Brazil and Uruguay (Almeida
et al.
2011; Witt
et al.
2011). Turtles incidentally captured in fisheries off South America (Billes
et al.
2006, López-Mendilaharsu
et al.
2009) also demonstrate that turtles originating from the SW and SE Atlantic Ocean beaches share foraging areas. Despite such mixing at foraging areas, there is no evidence for the shared use of nesting beaches. Genetic data indicate that turtles return to their natal beaches to nest on opposite sides of the Atlantic Ocean (Dutton
et al.
2013; Vargas
et al.
2017), and no tag recoveries contradict these data.
In the Indian Ocean, telemetry studies have been conducted at South African nesting beaches in the SW Indian Ocean (Hughes
et al.
1998; Luschi
et al.
2006; Robinson
et al.
2016) and at Andaman Islands nesting beaches in the NE Indian Ocean (Namboothri
et al.
2012; Swaminathan
et al.
2019). South African nesting females showed diverse movements that were highly influenced by complex oceanographic currents and features that lead them to foraging destinations in the South Atlantic Ocean, SW Indian Ocean, and Mozambique Channel (Hughes
et al.
1998, Luschi
et al.
2006, Robinson
et al.
2016). About half of the 10 post-nesting females tagged at the Andaman Islands moved westward: Two individuals reached the Mozambique Channel; the other half moved southeastward, past the Indonesian islands of Sumatra and Java, with one leatherback reaching an apparent foraging ground off NW Australia before transmissions stopped (Namboothri
et al.
2012; Swaminathan
et al.
2019). Despite overlap in one foraging area (
i.e.,
reaching the Mozambique Channel), tagging data do not indicate movement between the distant nesting beaches.
Within the Pacific Ocean, nearly all turtles tracked from East Pacific nesting beaches moved southward across the Equator to forage in open-ocean waters of the SE Pacific Ocean or in the coastal waters of Central America, Peru, and Chile. The movements of post-nesting females from the West Pacific Ocean are dependent on the season in which they nest, with winter-nesting females predominantly tracked into the Southern Hemisphere and summer-nesting females foraging in diverse coastal and oceanic ecosystems throughout the northern Indo-Pacific region (Benson
et al.
2011). Telemetry data indicate little or no overlap with foraging destinations utilized by nesting females of the East and West Pacific populations (Bailey
et al.
2012; Benson
et al.
2011). However, a genetic study of bycaught turtles off the coast of Chile and Peru indicated that 15 percent of leatherback turtles originated from West Pacific nesting beaches (Donoso and Dutton 2010), suggesting that foraging overlap may be more prevalent than estimated by telemetry data. Still, there is no genetic evidence for contemporary interbreeding between the two populations (Dutton
et al.
2007), and telemetry and tagging data do not indicate movement between the distant
nesting beaches. Thus, flipper tagging and satellite telemetry data support the marked separation, and thus discreteness, of the seven populations at their nesting beaches.
Physical factors likely shape and reinforce the behavior patterns that result in reproductive isolation. Though the species has a global range, with foraging areas extending into high latitudes, nesting mainly occurs on tropical or subtropical beaches. Post-hatchling dispersal is determined by the ocean currents they encounter off nesting beaches. While adults move throughout tropical and temperate waters irrespective of ocean currents, both males and females return to the waters off their natal nesting beach to mate. This natal homing is somewhat flexible, (Dutton
et al.
2013; Jensen
et al.
2013), creating reproductive isolation only among distant nesting sites, which may also be physically separated from one another by land masses and oceanographic barriers to gene flow. For example, leatherback turtles in the Atlantic Ocean are physically separated from those in the Pacific Ocean by the Americas. Though leatherback turtles have greater cold tolerance than other sea turtles, they do not appear to venture into latitudes greater than 47° S or 71° N (Eggleston 1971; Eckert
et al.
2012). Therefore, the low latitude and cold waters of the Cape Horn Current likely prevent movement between the Atlantic and Pacific Oceans. Within ocean basins, nesting beaches of the discrete populations are separated by long distances of uninterrupted deep water (
e.g.,
the East Pacific Barrier and the mid-Atlantic Barrier). While leatherback turtles clearly cross these open-ocean barriers to reach distant foraging areas, they do not appear to do so for nesting and breeding, but rather return to their natal region to breed and nest (Barragan
et al.
1998; Dutton
et al.
1999; Barragan and Dutton 2000; Dutton
et al.
2013). Within ocean basins, currents shape post-hatchlings' movement patterns, which they may retain as adults (
e.g.,
Fossette
et al.
2010; Benson
et al.
2011). The NW Atlantic leatherback population appears to be physically separated from the SE and SW Atlantic populations by the current systems of the South and North Atlantic Gyres, respectively. NW Atlantic leatherback nesting beaches are adjacent to northward moving currents (
e.g.,
Gulf Stream). Leatherback hatchlings from these nesting beaches, therefore, are transported northward, remaining in the North Atlantic Ocean. Those that survive return to their nesting beaches as adults, completing their life stages within the North Atlantic (Fossette
et al.
2010; Chambault
et al.
2017). The SE and SW Atlantic populations are similarly retained in the South Atlantic Ocean by the South Atlantic Gyre and the Benguela Current, which flows northward along the SE coast of Africa, restricting movement into the Indian Ocean. Within the Indian Ocean, the Somali Current runs between the nesting beaches of the SW and NE Indian populations. The NE Indian and West Pacific populations likely became isolated as a result of exposed land barriers between Indonesia, New Guinea, and the Philippines as a result of low sea levels within the past 6,000 years (Barber
et al.
2000). Seasonal monsoons may also play a contemporary role by altering current directions and hatchling dispersal patterns (Benson
et al.
2011; Gaspar
et al.
2012). Thus, physical factors have likely helped to shape, or at least reinforce, the reproductive isolation among distant nesting beaches.
Based on these data, the Team concluded that the seven populations demonstrate discreteness, or marked separation from each other, due to behavioral and physical factors. These are the NW Atlantic, SW Atlantic, SE Atlantic, SW Indian, NE Indian, West Pacific, and East Pacific populations.
Significance
Each of the discrete populations is significant to the species because the loss of any one would result in a significant gap (
i.e.,
a half or quarter of an ocean basin) in the range of the species. Several populations also persist in unique ecological settings. Each population likely possesses unique genetic characteristics and local adaptations as a result of thousands of years of reproductive isolation, but none have yet been identified because all genetic studies have involved neutral markers. Therefore, the Team did not rely on evidence of unique genetic characteristics and local adaptations for its significance finding.
A loss of the NW Atlantic population would result in a gap (
i.e.,
the entire North Atlantic Ocean) of the nesting and foraging range of the species. If the NW Atlantic population were extirpated, it is unlikely that leatherback turtles from other populations would recolonize the North Atlantic Ocean in an ecological time frame (
i.e.,
tens to hundreds of years), leaving a significant gap in the range of the species. Extirpation of this population would also significantly reduce the genetic diversity of the species, as reflected by the possession of several unique haplotypes. Leatherback turtles of the NW Atlantic Ocean also occur in a unique ecological setting; this is the only DPS that regularly forages at high latitudes. Sightings have been documented as far north as Norway and Iceland (Brongersma 1972; Goff and Lien 1988; Carriol and Vader 2002; McMahon and Hayes 2006; Eckert
et al.
2012). Such high latitude foraging is likely facilitated by the warm Gulf Stream, which meets cold water currents to create highly productive foraging areas. The Team concluded that the NW Atlantic population is biologically significant to the species.
In the SW Atlantic Ocean, leatherback turtles only nest in a small area of the coastline of Brazil. All other nesting in South America occurs above the Equator or on the Pacific Coast. Therefore, the loss of this population would result in a gap of the nesting range of the species (
i.e.,
the SW Atlantic coast). Although SE Atlantic leatherback turtles forage off the coasts of Brazil, Argentina, and Uruguay, they do not breed there. Rather, they return to the waters off western Africa to mate (Vargas
et al.
2017). Therefore, if the SW Atlantic population were extirpated, it is unlikely that leatherback turtles from other populations would recolonize this region, leaving a significant gap in the nesting range of the species. The extirpation of this population would also significantly reduce the genetic diversity of the species, as reflected by the possession of unique haplotypes and high genetic diversity, despite the small population size (Vargas
et al.
2017). The SW Atlantic population is biologically significant to the species.
Leatherback turtles of the SE Atlantic population nest in West Africa and forage in the South Atlantic Ocean. This population is much more abundant than the SW Atlantic population, which also forages in the South Atlantic Ocean. Therefore, the loss of this population would result in a gap of the nesting range of the species (
i.e.,
western Africa) and a significant reduction in the abundance of leatherback turtles foraging throughout the South Atlantic Ocean. The extirpation of this population would also significantly reduce the genetic diversity of the species, as reflected by the possession of unique haplotypes. The Team concluded that the SE Atlantic population is biologically significant to the species.
In the SW Indian Ocean, leatherback turtles only nest in a small area along the South African and Mozambican coastlines. No other leatherback turtles nest in eastern Africa or in other areas throughout the entire western Indian Ocean. Therefore, the loss of this population would result in a gap of the
nesting range of the species (
i.e.,
the SW Indian Ocean). The SW Indian population also occurs in a unique ecological setting: It is the only population to nest on temperate beaches. The warm Agulhas Current, adjacent to the nesting beaches, likely facilitates their high-latitude nesting. The Team concluded that the SW Indian population is biologically significant to the species.
Leatherback turtles nest in small numbers in the NE Indian Ocean. These nesting sites are separated from other Indian Ocean nesting sites by at least 5,000 km. Although western Pacific nesting sites are closer, males and females return to the waters off their natal beaches to breed, preventing interbreeding among NE Indian and West Pacific populations. Therefore, the loss of this population would result in a gap of the nesting range of the species (
i.e.,
the NE Indian Ocean). The extirpation of this population would also significantly reduce the genetic diversity of the species, as reflected by the possession of unique haplotypes. The Team concluded that the NE Indian population is biologically significant to the species.
West Pacific leatherback turtles nest in small numbers primarily in four nations of the West Pacific Ocean. These nesting sites are separated from East Pacific nesting sites by over 10,000 km. Though NE Indian nesting sites are closer in distance, male and female philopatry prevents interbreeding. Therefore, the loss of this population would result in a gap of the nesting range of the species (
i.e.,
the West Pacific Ocean). The loss of this population would also result in a gap of the foraging range of the species (
i.e.,
the North Pacific Ocean). The extirpation of this population would also significantly reduce the genetic diversity of the species, as reflected by the possession of unique haplotypes. The West Pacific population is ecologically unique in two ways: It is the only population to forage in both hemispheres; and it nests year-round, with nesting peaks in the summer and winter. The Team concluded that the West Pacific population is biologically significant to the species.
Leatherback turtles nesting on eastern Pacific coastlines also forage in the East Pacific Ocean. A loss of this population would result in a gap of the nesting range of the species (
i.e.,
the East Pacific Ocean). Though West Pacific leatherback turtles may forage off the coasts of Peru and Chile, they do not breed there (Donoso and Dutton 2010). Therefore, if the East Pacific population were extirpated, it is unlikely that leatherback turtles from other populations would recolonize this region, leaving a significant gap in the nesting range of the species. The extirpation of this population would also significantly reduce the genetic diversity of the species, as the population possess several unique haplotypes. The East Pacific population is unique in having the smallest nesting female size, clutch size, and egg size of all populations, possibly reflecting unique foraging conditions that are subject to oceanographic regime shifts (
e.g.,
the El Niño Southern Oscillation, or ENSO). The Team concluded that the East Pacific population is biologically significant to the species.
DPS Summary
The Team found that seven populations met the definition for discreteness. These populations are markedly separated as a result of the behavioral factors of movement (as demonstrated by satellite telemetry and flipper tagging studies) and philopatry, which has led to reproductive isolation (as demonstrated by genetic discontinuity). They are also physically separated by land masses, oceanographic features, and currents. The Team found these seven populations to be significant to the species because the loss of any one of them would result in a significant gap in the range of the species as well as a significant loss of genetic diversity, reducing the evolutionary potential of the species. Some populations also occur in a unique ecological setting. Thus, after reviewing the best available information, the Team identified the following populations as potential DPSs: NW Atlantic, SW Atlantic, SE Atlantic, SW Indian, NE Indian, West Pacific, and East Pacific. The Team defined the potential DPSs as leatherback turtles originating from nesting beaches within the boundaries for each DPS. The range of each DPS, which also includes foraging areas, thus extends beyond the nesting boundaries for most DPSs, and may overlap extensively with the range of another DPS. The boundaries are based on the best available genetic, telemetry, and observational data. When such data were not available, the Team used information on possible barriers to gene flow, such as oceanographic features. For ease of use, the Team applied political boundaries when this did not conflict with biological or oceanographic data. Additional information on the boundaries is available in the following sections, which summarize the extinction risk analysis for each DPS, and in the Status Review Report.
NW Atlantic DPS
The Team defined the NW Atlantic DPS as leatherback turtles originating from the NW Atlantic Ocean, south of 71° N, east of the Americas, and west of Europe and northern Africa; the southern boundary is a diagonal line between 5.377° S, 35.321° W and 16.063° N, 16.51° W. The northern boundary reflects a straight latitudinal line based on the northernmost documented occurrence of leatherback turtles (Brongersma 1972; Goff and Lien 1988; Carriol and Vader 2002; McMahon and Hayes 2006; Eckert
et al.
2012). The southern boundary is a diagonal line between the elbow of Brazil, where the Brazilian current begins and likely restricts the nesting range of this DPS, and the northern boundary of Senegal. The boundary between Senegal and Mauritania was chosen because the SE Atlantic DPS does not appear to nest above this boundary (Fretey
et al.
2007).
The range of this DPS (
i.e.,
all areas of occurrence) extends throughout the North Atlantic Ocean, including the Caribbean Sea, Gulf of Mexico (GOM), and Mediterranean Sea. Available data indicate that the NW Atlantic DPS occurs (at varying levels of frequency) in the waters of the following nations or territories: Albania, Algeria, Anguilla, Antigua and Barbuda, Aruba, Azores, Bahamas, Barbados, Belize, Bermuda, Bonaire, Bosnia and Herzegovina, Brazil, British Virgin Islands, Canada, Cape Verde, Cayman Islands, Colombia, Costa Rica, Croatia, Cuba, Curaçao, Cyprus, Denmark, Dominica, Dominican Republic, Egypt, France, French Guiana, Greece, Greenland, Grenada, Guadeloupe, Guatemala, Guyana, Haiti, Honduras, Iceland, Ireland, Israel, Italy, Jamaica, Lebanon, Libya, Madeira, Malta, Martinique, Mauritania, Mexico, Montenegro, Montserrat, Morocco, Netherlands Antilles, Nicaragua, Norway, Panama, Portugal, Slovenia, Spain, St. Barthelemy, St. Eustatius, St. Kitts and Nevis, St. Lucia, St. Maarten, St. Pierre and Miquelon, St. Martin, St. Vincent and the Grenadines, Suriname, Sweden, Syria, Trinidad and Tobago, Tunisia, Turkey, Turks and Caicos Islands, United Kingdom, United States (including Puerto Rico and the U.S. Virgin Islands (USVI), Venezuela, and Western Sahara.
All nesting in this DPS occurs in the NW Atlantic Ocean, concentrated from the southeast United States throughout the Wider Caribbean Region (Dow
et al.
2007). Leatherback nesting in the NW Atlantic can be grouped into several broad geographical areas, including the
U.S. mainland (primarily Florida), North Caribbean (including USVI and Puerto Rico), West Caribbean (Honduras to Colombia), and Southern Caribbean/Guianas (Venezuela to French Guiana; TEWG 2007). The largest nesting aggregations occur in Trinidad, French Guiana, and Panama. The northern-most confirmed nesting occurs in North Carolina, but there has been a crawl recorded as far north as Assateague Island National Seashore, Maryland (Rabon
et al.
2003). No nesting occurs in the Mediterranean Sea (Casale and Margaritoulis 2010).
Nesting occurs on unobstructed, high-energy beaches with either a deep water oceanic approach or a shallow water approach with mud banks, but without coral or rock formations (TEWG 2007). The main characteristics of leatherback nesting beaches include coarse-grained sand; steep, sloping littoral zone; obstacle-free approach; proximity to deep water; and oceanic currents along the coast (Hendrickson and Balasingam 1966 in Eckert
et al.
2015). During the nesting season, adult females and males inhabit the waters off nesting beaches. During a nesting season, females generally stay within about 100 km of their nesting beaches, remaining close to the coast on the continental shelf, and engaging in shallow dives (Eckert
et al.
2012). Intra-seasonal movement of greater than 100 km also occurs, especially between French Guiana and Suriname (Fossette
et al.
2007; Georges
et al.
2007), Panama and Costa Rica (Chacón-Chaverri and Eckert 2007), and among Caribbean nesting beaches, including those on Trinidad (Brautigam and Eckert 2006; Georges
et al.
2007; Horrocks
et al.
2016). Adult males migrate from temperate foraging areas in the North Atlantic Ocean to waters off nesting beaches, typically arriving before the nesting season and remaining for the majority of the season (James
et al.
2005b; Doyle
et al.
2008; Dodge
et al.
2014).
Foraging areas of the NW Atlantic DPS include coastal and pelagic waters of the North Atlantic Ocean (Eckert
et al.
2012; Saba 2013; Shillinger and Bailey 2015). These waters include the GOM, North Central Atlantic Ocean, northwestern Atlantic shelf waters of the United States and Canada, waters along the southeastern U.S. coast, the Mediterranean Sea, and the northeastern Atlantic shelf waters of Europe and northwestern Africa (TEWG 2007). Some post-nesting females also remain in tropical waters to forage (Fossette
et al.
2010). This DPS is mostly commonly associated with open-ocean and coastal shelf foraging areas off Nova Scotia (Canada), northeastern United States, GOM, northwestern Europe, and northwestern Africa (James
et al.
2005a, 2006b, 2007; Eckert 2006; Eckert
et al.
2006; Fossette
et al.
2010a; Fossette
et al.
2010b; Dodge
et al.
2014; Stewart
et al.
2016; Aleksa
et al.
2018). Fossette
et al.
(2014) analyzed available satellite telemetry data from 1995 to 2010 on post-nesting females (n = 93) as well as males (n = 4), females (n = 8), and a juvenile (n = 1) from foraging grounds throughout the Atlantic Ocean. They found widespread use of the North Atlantic Ocean (Fossette
et al.
2014). High-use areas mainly occurred in the central (25 to 50° N, 50 to 30° W) and eastern Atlantic Ocean, in particular in the waters offshore Western Europe, around Cape Verde (year-round) and the Azores (October to March; Fossette
et al.
2014). Fossette
et al.
(2014) found that seasonal high-use areas also occurred along the eastern U.S. coast (April to June and October to December) and off Canada (July to December). The GOM is also a high-use foraging area, with a peak in the northeast GOM during August and September (Aleksa
et al.
2018). Overall, leatherback turtles of the North Atlantic population appear to have a diverse array of foraging habitat available.
Abundance
The total index of nesting female abundance for the NW Atlantic DPS is 20,659 females. The nesting beaches with the greatest abundance have been included in this index, and most beaches with an unquantified number of nests likely host few nesting females. We based this index on 24 nesting aggregations in 10 nations: Trinidad and Tobago (n = 11,324), French Guiana (n = 2,519), Panama (n = 2,251), United States (n = 1,694), Costa Rica (n = 1,306), Suriname (n = 698), Grenada (n = 499), Venezuela (n = 215), Guyana (n = 76), and Nicaragua (n = 10). With the possible exception of Colombia, our total index does not include 31 unquantified but likely small nesting aggregations for which data are not available. It also does not include outdated data published by Dow
et al.
(2007), which includes binned crawls, categorized as less than 25, 25 to 100, 100 to 500, 500 to 1000, or unknown abundance. Crawls or emergences (measured as females or tracks on beaches) include both successful egg-laying and unsuccessful nesting, so the number of crawls represents approximately two to 10 times the number of nests (Dow
et al.
2007). Because the Dow
et al.
data, which are more than 10 years old and do not provide the number of actual nests, may not be representative of recent nesting trends, we did not include them in our total index. To calculate the indices of nesting female abundance, we added the number of nests over the last 3 years (representing the most recent remigration interval; Eckert
et al.
2012) and divided by the clutch frequency (site-specific values or, when such values were not available, the average of the site-specific values,
i.e.,
5.5 clutches per season).
Our total index of nesting female abundance is based on the best available data for this DPS. It is the most robust estimate of nesting females at this time because it only includes available nesting data from recently and consistently monitored nesting beaches. Our total index does not include data from beaches where we were unable to quantify the number of nesting females, either due to the lack of recent or available nesting data or because only crawl data were reported (often on smaller nesting beaches). Scattered nesting may occur on beaches throughout the region, but because these beaches are not monitored, or have not been recently monitored, recent data are not available.
Nesting in the NW Atlantic DPS is characterized by many small nesting beaches. Large nesting aggregations are rare; only about 10 leatherback nesting beaches in the Wider Caribbean Region (about two percent of the DPS's total nesting sites) host more than 1,000 crawls annually (Dow Piniak and Eckert 2011). Only one site, Grande Riviere in Trinidad, hosts more than 5,000 nesting females, representing 29 percent of the total index of nesting female abundance. Relatively large nesting aggregations are also found in Matura (Trinidad), Chiriqui Beach (Panama), and Cayenne and Remire Montjoly (French Guiana). In contrast, most known nesting beaches support a small nesting female abundance; 71 percent of the total nesting sites record annual crawls of less than 100 (Dow Piniak and Eckert 2011). The number of nesting females is unquantified at 31 beaches (
i.e.,
the majority of nesting sites for the DPS). However, for the reasons identified above, most of those sites have small abundance levels as inferred from the numbers of crawls estimated by Dow
et al.
(2007). Therefore, our total index of nesting female abundance represents the most robust estimate allowed by the best available data and includes the majority of nesting females because the largest nesting aggregations were included. The data regarding additional nesting aggregations are not sufficiently recent, specific, or reliable for inclusion, and the contribution of these nesting
aggregations to the total index is expected to be small.
Our total index of nesting female abundance is similar in comparison to other published estimates. TEWG (2007) estimated the abundance of NW Atlantic leatherback turtles using nesting data from 2004 and 2005. At that time, the number of adult females (equating to total index of nesting female abundance in our analysis) was estimated to be approximately 18,700 (range 10,000 to 31,000). While a wide range was provided, the point estimate in TEWG (2007) is similar to, albeit slightly lower than, our total index of 20,659 nesting females. The most recent, published IUCN Red List assessment for the NW Atlantic Ocean subpopulation estimated a total of 20,000 mature individuals (The NW Atlantic Working Group 2019). Our total index, which only includes nesting females, exceeds their estimate, likely due to our use of a 3-year remigration interval, which has increased at some locations in recent years (
e.g.,
4.5 years at St. Croix; K.R. Stewart, The Ocean Foundation and C. Lombard, USFWS, pers. comm., 2019).
We conclude that the total index of nesting females for the NW Atlantic DPS is 20,659 females. The nesting beaches with the greatest abundance have been included in our total index, and most beaches with an unquantified number of nests likely host few nesting females. Current nesting female abundance is not at a level where stochastic or environmental changes would have catastrophic impacts, but the abundance at several nesting sites with previously high density has declined drastically. However, as we discuss below, a declining nest trend and several existing threats will likely continue to reduce this abundance.
Productivity
The NW Atlantic DPS exhibits decreasing nest trends at nesting aggregations with the greatest indices of nesting female abundance. Though some nesting aggregations indicate increasing trends, most of the largest ones demonstrate declining nest trends. We evaluated nest trends by using nest count data consistently collected using a standardized approach for at least 9 years, with the last year of data in 2014 or more recently and with an average of more than 50 nests annually. When data did not meet these criteria, we evaluated bar graphs provided in the Status Review Report to consider all available data. Thus, these data are representative of the DPS because they include the largest nesting aggregations. With the possible exception of Colombia, nesting aggregations for which data are not available are likely small. Significant declines have been observed at nesting beaches with the greatest historical or current nesting female abundance, most notably in Trinidad and Tobago (Grande Riviere, Fishing Pond, and Tobago), Suriname, French Guiana (Awala-Yalimapo), Florida, and Costa Rica (Tortuguero). Therefore, these nest trends represent the best available data for this DPS.
In Trinidad and Tobago, trends in annual nest counts were largely negative between 2009 and 2017, the years for which data were available. For Trinidad, we analyzed trends for three separately monitored beaches, including Grande Riviere, Matura, and Fishing Pond. The long-term trend was negative for Grande Riviere (median = −6.9 percent; sd = 17.4 percent; 95 percent CI = −43.8 to 26.9 percent; f = 0.682; mean annual nests = 13,272), positive for Matura (median = 1.8 percent; sd = 15.1 percent; 95 percent CI = −29.2 to 33.0 percent; f = 0.561; mean annual nests = 7,359), and negative for Fishing Pond (median = −19.3 percent; sd = 15.1 percent; 95 percent CI = −49.8 to 12.0 percent; f = 0.916; mean annual nests = 3,892). For Tobago, the median trend was −0.9 percent annually (sd = 11.3 percent; 95 percent CI = −25.0 to 21.5 percent; f = 0.540; mean annual nests = 452).
For French Guiana, we analyzed nest count data from 2002 to 2017 for Awala-Yalimapo beach in the west and data from 1999 to 2017 for Cayenne and Remire Montjoly beaches in the east. There was a steep decline at Awala-Yalimapo, with a median trend of −19.4 percent annually (sd = 12.2 percent; 95 percent CI = −43.2 to 6.0 percent; f = 0.942; mean annual nests = 3,200). In contrast to Awala-Yalimapo, nest counts at Cayenne and Remire Montjoly increased by 2.8 percent annually (sd = 12.9 percent; 95 percent CI = −24.9 to 27.9 percent; f = 0.596; mean annual nests = 3,498). In addition, leatherback nesting occurred on remote beaches in western French Guiana until 2013 (
e.g.,
a high of 4670 nests was found in 2003, with 1,270 mean annual nests from 2002 to 2013), but we were unable to analyze trends because monitoring on these remote beaches has been reduced since approximately 2010 due to significant beach erosion and the disappearance of some previously monitored beaches.
Suriname, Grenada, and Panama each had a single time series sufficient for trend analysis. For Suriname, we combined datasets from two beaches, Galibi and Braamspunt, which were monitored between 2001 and 2017. Total nests in Suriname declined by −14.6 percent annually (sd = 9.6 percent; 95 percent CI = −36.4 to 4.5 percent; f = 0.953; mean annual nests = 4,586). In Grenada, data on the number of nesting tracks were collected on Levera beach between 2002 and 2018. There was a 7.1 percent annual increase in tracks at Levera during that period (sd = 8.7 percent; 95 percent CI = −10.5 to 25.3 percent; f = 0.827; mean annual tracks = 895). In Panama, the nest counts at Chiriqui beach increased by 0.8 percent annually (sd = 7.0 percent; 95 percent CI = −14.1 to 14.6 percent; f = 0.557; mean annual nests = 4,463) between 2004 and 2017.
In Costa Rica, the four beaches for which we had sufficient data to analyze annual nest count trends mostly exhibited declining trends. Tortuguero experienced the steepest decrease, with a median trend of −10.9 percent annually (sd = 4.2 percent; 95 percent CI = −19.5 to 2.2 percent) for data collected between 1995 and 2017. Nest counts decreased by −3.8 percent annually at Pacuare beach (sd = 9.3 percent; 95 percent CI = −22.6 to 16.9 percent) between 2004 and 2017, but increased by 1.8 percent annually (sd = 6.0 percent; 95 percent CI = −10.8 to 14.2 percent) at the nearby Pacuare Nature Reserve between 1991 and 2017. Nest counts at Estacion la Tortuga deceased slightly, with a median trend of −0.5 percent annually (sd = 7.0 percent; 95 percent CI = −15.7 to 13.1 percent) between 2002 and 2017.
For the United States, we analyzed annual nest count trends for Florida (statewide data collected between 2008 and 2017), three beaches in Puerto Rico, including Culebra (1984 to 2017), Luquillo-Fajardo (1996 to 2017), and Maunabo (1999 to 2017), and Sandy Point National Wildlife Refuge in St. Croix, USVI (1982 to 2017). The median trend for Florida was a decline of −2.1 percent annually (sd = 13.0 percent; 95 percent CI = −28.3 to 25.5 percent; f = 0.582; mean annual nests = 1,288). Culebra nests decreased by −3.7 percent annually (sd = 5.3 percent; 95 percent CI = −14.9 to 6.8 percent; f = 0.791; mean annual nests = 153), while nests increased by 15.9 percent annually at Luquillo-Fajardo (sd = 5.5 percent; 95 percent CI = −7.1 to 15.3 percent; f = 0.805; mean annual nests = 283) and by 7.7 percent annually at Maunabo (sd = 4.9 percent; 95 percent CI = −2.7 to 17.4 percent; f = 0.945; mean annual nests = 161). In St. Croix, nests increased by 1.7 percent annually (sd = 4.6 percent; 95 percent CI = −7.8 to 10.7 percent; f = 0.660; mean annual nests = 399).
These trend data are similar to other recent findings, adding further
confidence in declining trends at multiple large nesting aggregations. Because of concerns about declining nest counts throughout the region, the National Fish and Wildlife Foundation (NFWF) convened a NW Atlantic Leatherback Working Group (
i.e.,
the Working Group) to assess recent nesting data and complete a region-wide trend analysis (NW Atlantic Leatherback Working Group 2018). The trend analyses conducted by the Working Group used leatherback nesting data from 23 sites from 14 different nations with at least 10 years of data with consistent within-site methodology, analyzing data for three time periods: 1990 to 2017, 1998 to 2017, and 2008 to 2017. Our approach to trend analyses was similar to that used by the Working Group in that both approaches involved Bayesian analyses of data meeting set criteria. However, the Team decided against aggregating the data over the DPS due to incongruity of data collection methods, collection dates and duration, and reporting. Despite these differences, the overall conclusion was the same—an overall declining nest trend.
The Working Group found that regional, abundance-weighted trends were negative for all three time periods and became more negative in the more recent time series (NW Atlantic Leatherback Working Group 2018). Specifically, overall nesting trends decreased at −4.21 percent annually from 1990 to 2017 and at −5.37 percent annually from 1998 to 2017, with the most notable decrease (−9.32 percent annually) occurring during the most recent time frame of 2008 to 2017. While site-level trends showed variation within and among sites and across the time periods, overall the sites also reflected the same regional pattern: More negative trends were apparent during the most recent time frame. Seven sites had significant positive nesting trends from 1990 to 2017, but no sites exhibited significant positive trends from 2008 to 2017. The significant decline observed at Awala-Yalimapo, French Guiana (−12.95 percent annually from 1990 to 2017, −19.05 percent annually from 1998 to 2017, and −31.26 percent annually from 2008 to 2017), drove the regional results, but similar significant declines were found at other nesting beaches for the longer time period, including: St. Kitts and Nevis (−12.43 percent annually), Tortuguero, Costa Rica (−10.42 percent annually), Suriname (−5.14 percent annually), and Culebra, Puerto Rico (−4.61 percent annually). It should be noted that the other nesting beach in French Guiana (Cayenne) demonstrated an increasing trend (7.44 percent annually from 1990 to 2017 and 8.19 percent annually from 1998 to 2017). However, it exhibited a decreasing trend (−14.21 percent annually) from 2008 to 2017. While nesting increased over time at Cayenne, this increase has apparently not resulted from females shifting from Awala-Yalimapo, as turtles that nest at Cayenne are genetically distinct (Molfetti
et al.
2013) and females tagged in Awala-Yalimapo are not seen in Cayenne or vice versa (NW Atlantic Leatherback Working Group 2018).
These modeling results demonstrate that there has been a decline in NW Atlantic nesting from 1990 to 2017, with the most significant decreases occurring from 2008 to 2017. Some nesting beaches demonstrated positive trends for the longer time period. However, none showed significant increases over the most recent time period. The cause for the decline is uncertain, but the Working Group identified anthropogenic sources (
e.g.,
fisheries bycatch), habitat losses, and changes in life history parameters (such as remigration interval) as potential drivers of the regional decline. While these results were taken into consideration by the Team when evaluating the extinction risk of the NW Atlantic DPS, the Team also performed its own trend analysis of the data provided to the Team so that the trends were calculated in a manner consistent with other DPSs. Regardless, both trend analyses conclude that the NW Atlantic DPS is experiencing a significant decline in nesting.
In-water abundance studies of leatherback turtles are rare. Archibald and James (2016) assessed the relative abundance of turtles at a foraging area off Nova Scotia, Canada, from 2002 to 2015. This study evaluated opportunistic sightings per unit effort and found a mean density of 9.8 turtles per 100 km
2
, representing the highest in-water density of leatherback turtles reported to date. Archibald and James (2016) concluded that the relative abundance of foraging leatherback turtles off Canada exhibited high inter-annual variability but, overall, showed a stable trend from 2002 to 2015. The authors reported that (at that time) these results were consistent with the stable or, in some cases, increasing trends reported for contributing NW Atlantic nesting beaches over the last decade (Dutton
et al.
2005; Girondot
et al.
2007; Fossette
et al.
2008; McGowan
et al.
2008; Stewart
et al.
2011; Rivas
et al.
2015). While there were no indications of a decreasing trend, the results should be interpreted with caution because of the small study area, opportunistic data collection, availability bias variance, and lack of understanding of the relative density outside the study area (Archibald and James 2016).
Despite the declining trend in nesting, productivity parameters for the DPS are similar to the species' averages (though some may be declining, as we discuss below). While there is some variation, most productivity parameters are relatively consistent throughout the DPS. The overall survival rate for nesting females is relatively high at 85 percent (Pfaller
et al.
2018), with mean estimates of 0.70 to 0.99 in French Guiana (Rivalan
et al.
2005, 2008), 0.89 in St. Croix (Dutton
et al.
2005), and 0.89 to 0.96 on the Atlantic coast of Florida (Stewart
et al.
2007, 2014). Remigration intervals range from 1 to 11 years (Schulz 1975; Boulon
et al.
1996; Chevalier and Girondot 1998; Hilterman and Goverse 2007; Eckert
et al.
2012; Stewart
et al.
2014; Rivas
et al.
2016; Garner
et al.
2017). In St. Croix and St. Kitts, the median remigration interval appears to be increasing (4.5 years; K.R. Stewart, The Ocean Foundation and C. Lombard, USFWS, pers. 2019; K.M. Stewart, Ross University School of Veterinary Medicine and St. Kitts Sea Turtle Monitoring Network, pers. comm., 2019). Averaging all available data, the mean remigration interval for the DPS is 2.7 years, rounded to 3 years for use in our calculation of the index of nesting female abundance. Average clutch frequency per nesting season ranges from 3.6 to 8.3 throughout the region, with an overall mean of 5.5 nests per season, interspersed with 9 to 10 day internesting intervals (Eckert
et al.
2015; Garner
et al.
2017). Recent records indicate that nesting females deposit 80 to 88 eggs per clutch. However, an early study by Carr and Ogren (1959) reported only 67 eggs per clutch. Hatching success is highly variable for nests that remain
in situ,
even for those that are viable and do not experience significant inundation or predation, with estimates as low as 8.9 percent in Costa Rica (Troëng
et al.
2007) and 10.6 percent in Suriname (Hilterman and Goverse 2007) and as high as 93.4 percent in Florida (Perrault
et al.
2012). Overall, hatching success is estimated at approximately 50 percent (Eckert
et al.
2012). Hatchling sex ratios often exhibit a female bias, but less so than for other sea turtle species, with estimated production of anywhere from 30 to 100 percent females in Suriname, Tobago, Colombia, and Costa Rica (Mrosovsky
et al.
1984; Dutton
et al.
1985; Godfrey
et al.
1996; Leslie
et al.
1996; Mickelson and
Downie 2010; Patiño-Martínez
et al.
2012). However, the proportion of females documented in foraging individuals and strandings ranges from 57 to 70 percent (Murphy
et al.
2006; James
et al.
2007; TEWG 2007), and the ratio of females to males during an individual breeding season is thought to be closer to 1:1 (Stewart and Dutton 2014).
We conclude that the DPS exhibits a declining nest trend. In addition, there are indications of decreased productivity within the DPS. In St. Croix, one of the most thoroughly monitored nesting beaches in this DPS, the data from 1981 to 2010 indicate that hatching success and clutch frequency are declining and remigration intervals are increasing (Garner
et al.
2017). Overall, we have a high degree of confidence in the decreasing nest trend and productivity metrics for this DPS, due to the large amount of data available from the largest nesting aggregations. We acknowledge that data are not available from all nesting beaches, but the data that we have relied upon is the best available and meets established standards. The declining trends reflect reduced nesting female abundance. In addition, longer remigration intervals and/or reduced clutch frequencies may play a role in this decline. The decline reflects a reduction in productivity that places the DPS at risk given the magnitude and duration of the decreasing trend.
Spatial Distribution
The DPS has a broad spatial distribution for both foraging and nesting. There is significant genetic population structure, with subpopulations connected via various levels of gene flow and metapopulation dynamics. Tagging and telemetry studies indicate considerable mixing of leatherback turtles among nesting beaches and at multiple foraging areas throughout the North Atlantic Ocean.
Nesting is widespread throughout the NW Atlantic beaches, occurring primarily as scattered, small aggregations throughout the Wider Caribbean, but with larger concentrations of nesting activity at certain sites in Trinidad, French Guiana, Suriname, Trinidad, Colombia, Panama, Costa Rica, Puerto Rico, St. Croix, and Florida (Horrocks
et al.
2016).
Genetic sampling in the NW Atlantic DPS has been generally extensive with good coverage of large populations in this region. However, sampling from some smaller Caribbean nesting aggregations is absent, and there are gaps in sampling or analysis for nesting sites along the coasts of South and Central America (
e.g.,
Guyana, Venezuela, Colombia, and Panama). A comprehensive survey of genetic population structure in the Atlantic Ocean included large sample sizes from five nesting populations representative of the DPS and analysis of longer mtDNA sequences in combination with an array of 17 nuclear microsatellite DNA loci (Roden and Dutton 2011; Dutton
et al.
2013). The microsatellite data revealed fine-scale genetic differentiation among neighboring subpopulations (Dutton
et al.
2013): Trinidad, French Guiana/Suriname, Florida, Costa Rica, and St. Croix. The mtDNA data failed to find significant differentiation between Florida and Costa Rica or between Trinidad and French Guiana/Suriname. However, Dutton
et al.
(2013) show that the mtDNA sequence variation had relatively low statistical power to detect fine scale structure compared to the microsatellite DNA loci. The mtDNA homogeneity between Costa Rica and Florida, with differentiation demonstrated at nuclear DNA loci, suggests that Costa Rica may be the source of founders for the Florida population via one or multiple recent colonization events, likely indicating historic connectivity rather than ongoing demographic connectivity (Dutton
et al.
2013). Likewise the French Guiana/Suriname and Trinidad populations were undifferentiated with mtDNA likely indicating historic connectivity. However, microsatellite DNA reveal fine-scale genetic structure that is consistent with tagging studies demonstrating a lack of nesting female movement between the two nesting aggregations (TEWG 2007). Significant genetic differentiation has also been reported for Martinique and Guadeloupe and the mainland French Guiana rookery (Molfetti
et al.
2013). St. Croix likely represents a broader Northern Caribbean subpopulation of the NW Atlantic population that includes multiple neighboring island nesting aggregations in the USVI and Puerto Rico. However, sampling and analysis would be required to determine extent of fine scale structuring (NMFS unpublished data; Dutton
et al.
2013). The Costa Rica (Tortuguero and Gandoca) and Guiana (French Guiana and Suriname) nesting aggregations are distinct subpopulations based on microsatellite and mtDNA results (Dutton
et al.
2013), but information on tag returns indicates movement of nesting females between adjacent beaches of Panama, Colombia, Venezuela and Guyana. Therefore, these nesting aggregations have “fuzzy” boundaries, likely a result of flexible natal homing. Nesting females use beaches up to 400 km apart between nesting seasons (Troëng
et al.
2004; Chacón-Chaverri and Eckert 2007) and up to 463 km apart within the same nesting season (Stewart
et al.
2014). Additional sampling of the remaining nesting sites will be required to determine the extent of fine-scale structuring within the NW Atlantic DPS. However, the available science indicates significant substructure within the DPS.
Tagging studies indicate individual movement and gene flow among nesting aggregations. This is facilitated by the species' flexible natal homing,
i.e.,
philopatry to a region, rather than a specific beach. In adjacent nesting sites in French Guiana and Suriname, five to six percent of nesting females were observed to shift from one site to the other within a season (TEWG 2007), while Schulz (1971) reported this proportion to be slightly higher at 8.5 percent. In contrast, 35 percent of nesting females in Gandoca, Costa Rica, were estimated to nest at sites other than the study site during an individual season (Chacón-Chaverri and Eckert 2007). The predisposition of nesting females to stray within a nesting season may be influenced by the proximity of alternative nesting sites within a range of approximately 200 km (Horrocks
et al.
2016). However, even within a given nesting season, females have been observed to move as far as 369 km (Grenada), 463.5 km (Florida), and 532 km (Dominica) from their original location (Horrocks
et al.
2016). Among nesting seasons, interchange between nesting locations also appears to be frequent and wide-ranging, with maximum distance separating two nesting sites for an individual female recorded as 1,849 km over an 8-year span (Horrocks
et al.
2016).
Genetic studies have revealed that turtles from different nesting aggregations use the same foraging areas. Analyzing 684 longline bycatch samples from across the NW Atlantic in a mixed stock analysis and microsatellite assignment, Stewart
et al.
(2016) found that leatherback turtles from Costa Rica were caught in a higher proportion in the GOM (43 percent) compared to the Northeast Distant fishing zone, an area in the northwestern Atlantic Ocean (6 percent), while turtles from Trinidad and French Guiana comprised 54 percent of bycatch in the GOM and 93 percent in the Northeast Distant fishing zone. A study of turtles foraging off Nova Scotia, Canada, similarly assigned most (82 percent) of the 288 sampled turtles to Trinidad (n = 164) and French
Guiana (n = 72), with 15 percent (n = 44) from Costa Rica, and the remainder from St. Croix (n = 7) and Florida (n = 1; Stewart
et al.
2013). These proportions generally represent the relative population sizes for these breeding populations. Microsatellite DNA assignment of wild captured or stranded males (n = 122) throughout the NW Atlantic and Mediterranean found that all males originated from NW Atlantic nesting aggregations (Trinidad: 55 percent, French Guiana: 31 percent, and Costa Rica: 14 percent; Roden
et al.
2017). No turtles were identified from St. Croix or Florida. One turtle that stranded in Turkey was assigned to French Guiana, while strandings in France were assigned to Trinidad or French Guiana (Roden
et al.
2017).
The mixing of nesting aggregations at foraging areas is also supported by several tagging and/or satellite telemetry projects, conducted in U.S. waters (Murphy
et al.
2006; LPRC 2014; Dodge
et al.
2014, 2015; Aleksa
et al.
2018), Canada (James
et al.
2005a, 2005b, 2005c, 2006b, 2007; Bond and James 2017), Atlantic Europe and Mediterranean (Doyle
et al.
2008; Sonmez
et al.
2008), and on nesting beaches of various nations (Hildebrand 1987; Hays
et al.
2004; Ferraroli
et al.
2004; Eckert 2006; Eckert
et al.
2006; Hays
et al.
2006; TEWG 2007; Sonmez
et al.
2008; Evans
et al.
2008; Fossette
et al.
2010a, 2010b; Richardson
et al.
2012; Bailey
et al.
2012; Stewart
et al.
2014; Fossette
et al.
2014; Horrocks
et al.
2016; Chambault
et al.
2017). For instance, turtles from Nova Scotian foraging grounds were tracked to nesting areas off Colombia, Trinidad, Guyana, and French Guiana (Bond and James 2017). The reverse has also been demonstrated: some leatherback turtles from the western Atlantic undertake annual migrations to Canadian waters to forage (James
et al.
2005c), exemplified by post-nesting adults tracked to the waters off Nova Scotia from a variety of nesting locations, including French Guiana and Trinidad (Fossette
et al.
2014), Costa Rica, Panama (Evans
et al.
2008), and Anguilla (Richardson
et al.
2012). The eastern and western GOM also provide foraging areas for this DPS (Aleksa
et al.
2018), as observed from tracks of post-nesting turtles from Florida (Hildebrand 1987), Costa Rica (Tortuguero, Gandoca), and Panama (Chiriquí Beach; Evans
et al.
2008; Evans
et al.
2012). Evans
et al.
(2008) suggested that the GOM may represent a significant foraging ground for leatherback turtles from the Caribbean coast of Central America.
High use foraging areas may be identified through available telemetry data, but the migration routes to those areas may vary. Ferraroli
et al.
(2004) tracked leatherback turtles from French Guiana and found turtles dispersed widely throughout the North Atlantic but mostly followed two dispersion patterns: (1) Moving north to the Gulf Stream area, where they started following the general ocean circulation; and (2) traveling east, swimming mostly against the North Equatorial Current. Fossette
et al.
(2014) found a relatively broad migratory corridor when turtles departed their nesting sites in French Guiana/Suriname, and their movements overlapped with turtles from Grenada and Trinidad. Fossette
et al.
(2010a, 2010b) found that turtles tracked from nesting beaches in French Guiana, Suriname, and Grenada and turtles caught in waters off Nova Scotia and Ireland displayed three distinct migration strategies: (1) Heading northwest to fertile foraging areas off the Gulf of Maine, Canada, and GOM; (2) crossing the North Atlantic Ocean to areas off western Europe and Africa; and (3) residing between northern and equatorial waters. Essentially, tagging data coupled with satellite telemetry data indicate that leatherback turtles of the NW Atlantic DPS use the entire North Atlantic Ocean for foraging and migration (TEWG 2007).
Although adults forage at multiple areas throughout the North Atlantic Ocean (Fossette
et al.
2014), the range of juvenile leatherback turtles may be more restricted. Using an active movement model, Lalire and Gaspar (2019) found that most juveniles originating from nesting beaches in French Guiana and Suriname cross the Atlantic Ocean at mid-latitudes with north-south seasonal migrations; after several years, they reach the coasts of Europe and North Africa. Eckert (2002) reviewed the records of nearly 100 sightings of juvenile (less than 100 cm curved carapace length (CCL)) leatherback turtles and determined they are generally found in waters warmer than 26 °C, suggesting that the first portion of their life is spent in tropical and subtropical waters. After exceeding 100 cm CCL, distribution extends into cooler waters (as low as 8 °C), which is considered to be the primary habitat for the species (Eckert 2002).
The wide distribution of nesting and foraging areas likely buffers the DPS against local catastrophes or environmental changes. The fine-scale population structure, with movement of individuals and genes among nesting aggregations, indicates that the DPS has the capacity to withstand other catastrophic events.
Diversity
The NW Atlantic DPS exhibits spatial diversity, as demonstrated by insular and continental nesting, multiple diverse foraging areas, and moderate genetic diversity. The DPS nests along both continental and insular coastlines. Nesting beach habitat also shows considerable diversity, ranging from coarse-grained, sandy beaches to silty, ephemeral shorelines whose dynamics are influenced by estuarine input. The breadth and, in some cases, transiency, of suitable nesting habitat in the western North Atlantic may contribute to consistent, low-level flexibility in natal homing, both within and among reproductive seasons (Bräutigam and Eckert 2006), and this flexibility is thought to surpass that of other sea turtle species (TEWG 2007).
This DPS exhibits some temporal variation in nesting. Nesting generally begins in March or April, peaks in May or June, and ends in July or August (Eckert
et al.
2012). In French Guiana, a second small nesting peak was documented in Awala-Yalimapo during December and January. However, the number of nests deposited during that time frame decreased from 700 in 1986/1987 to 40 in 1992/1993, and now only a small number of individuals are observed to nest during that time (Girondot
et al.
2007). Some evidence indicates that the timing of nesting may be modulated by environmental characteristics distant from the nesting beach, such as water temperatures at foraging grounds (Neeman
et al.
2015).
The foraging strategies are also diverse, with turtles using coastal and pelagic waters throughout the entire North Atlantic Ocean (Fossette
et al.
2014). Foraging habitats include temperate waters of the GOM, North Central Atlantic Ocean, northwestern shelf (United States and Canada), southeastern U.S. coast, the Mediterranean Sea, and northeastern shelf (Europe; TEWG 2007). Some post-nesting females also remain in tropical waters (Fossette
et al.
2010). Overall, leatherback turtles in the North Atlantic Ocean appear to have a diverse array of foraging habitat available.
Genetic diversity of the DPS is moderate, with six mtDNA haplotypes (Dutton
et al.
2013). In St. Croix, a unique haplotype occurs at high frequency. The Florida and Costa Rica nesting aggregations each possess one unique, low frequency haplotype.
Based upon this information, we conclude that nesting location and habitat are diverse, providing some level of resilience against short-term spatial and temporal changes in the
environment. However, high-abundance nesting occurs only at a few locations (
e.g.,
Trinidad, French Guiana, and Panama). The foraging diversity likely provides resilience against local reductions in prey availability or catastrophic events, such as oil spills, by limiting exposure to a limited proportion of the total population. Moderate genetic diversity may provide the DPS with the raw material necessary for adapting to long-term environmental changes, such as cyclic or directional changes in ocean environments due to natural and human causes (McElhany
et al.
2000; NMFS 2017). We conclude that such diversity provides some level of resilience to threats for this DPS.
Present or Threatened Destruction, Modification, or Curtailment of Habitat or Range
Destruction and modification of leatherback turtle nesting habitat results from a variety of activities including coastal development and construction; beach erosion and inundation; placement of erosion control and nearshore shoreline stabilization structures and other barriers to nesting; beachfront lighting; vehicular and pedestrian traffic; beach sand placement; sand extraction; removal of native vegetation; and planting of non-native vegetation (Lutcavage
et al.
1997; Bouchard
et al.
1998; USFWS 1999; Dow
et al.
2007; Eckert
et al.
2012; NMFS and USFWS 2013). As a result, most nesting beaches are severely degraded by such activities that continue to cause adverse impacts throughout the range of the DPS.
Coastal Development and Construction
In many areas, nesting habitat is under constant threat from coastal development and construction (Dow
et al.
2007; Crespo and Diez 2016; Flores and Diez 2016). Coastal development impacts include construction of buildings and pilings on the beach; increased erosion; artificial lighting; pollution; recreational beach equipment and other obstacles on the beach; beach driving; increased human disturbance; and mechanized beach cleaning (Lutcavage
et al.
1997; USFWS 1999; Hernandez
et al.
2007; Dow
et al.
2007; Trinidad and Tobago Forestry Division
et al.
2010; Flores and Diez 2016). Driftwood found on nesting beaches also has the potential to alter nesting beach habitat and obstruct nesting females and hatchlings, as seen in Gandoca, Costa Rica (Chacón-Chaverri and Eckert 2007). These threats impact nesting habitat by reducing the amount and quality of suitable beaches, preventing or deterring nesting females from using optimal locations, destroying nests, eggs, and hatchlings, and preventing hatchlings from successfully reaching the ocean (USFWS 1999; Chacón-Chaverri and Eckert 2007; Hernandez
et al.
2007; Witherington
et al.
2014). Development involving the construction of tall buildings and clearing of vegetation can also alter sand temperatures and skew sex ratios (Gledhill 2007).
Development occurs to varying extents throughout the range of the DPS, but most leatherback nesting occurs in proximity to some coastal development. The Florida shoreline is extensively developed outside wildlife refuges (Witherington
et al.
2011). In Grenada, nearly 20 percent of all nests surveyed from 2001 to 2005 occurred in an area affected by development, resulting in ongoing run-off onto nesting beaches (Maison
et al.
2010). In Trinidad, increasing rural and commercial beachfront development is a concern, especially on the east coast where the main nesting beaches are located (Trinidad and Tobago Forestry Division
et al.
2010), including Grande Riviere, the largest nesting aggregation of this DPS. Likewise, several Tobago beaches are densely developed for commercial tourism, resulting in reduced turtle access to potential nesting sites due to buildings, umbrellas, and other recreational equipment (Trinidad and Tobago Forestry Division
et al.
2010). Development in Puerto Rico, in particular Playa Grande-El Paraiso (
i.e.,
Dorado Beach, which is considered to be the most important nesting beach in Puerto Rico), is also a notable concern (Crespo and Diez 2016; Flores and Diez 2016). There, ecosystems continue to be threatened by coastal development, even though the coastal zone is protected by the Maritime-Terrestrial Zone designation (
i.e.,
Coastal Public Trust Lands; Flores and Diez 2016).
Coastal development likely influences leatherback nest placement and subsequent nest success, which is the percentage of nesting attempts (
i.e.,
emergences onto the beach) that result in eggs being deposited. On Margarita Island, Venezuela, Hernandez
et al.
(2007) found that leatherback nesting aggregated towards the portions of the beach with fewer risk factors, such as light pollution and concentrations of beach furniture. This change in nesting behavior resulted in females nesting in less optimum areas (
e.g.,
areas with lower hatching success), thus affecting the reproductive potential of leatherback turtles in this region.
The magnitude of development is also changing in some areas, where nest placement and success may be affected in the future. For instance, the area around Cayenne, French Guiana, is undergoing increased urbanization and recreational use (Fossette
et al.
2008). In recent years, nesting has increased at Cayenne and eastern beaches compared to the western Awala-Yalimapo beaches (Réserve Naturelle de l'Amana data in Berzins 2018 and KWATA data in Berzins 2018). As such, more nesting in French Guiana is exposed to coastal development and the associated threats, and these threats are likely to continue and increase.
Beach Erosion and Inundation
While erosion is often intensified due to anthropogenic influences, natural features in some areas result in high erosion rates and unstable beaches, thus affecting leatherback nesting. For instance, the Maroni River influence in the Guianas (French Guiana especially) has resulted in highly dynamic and unstable beaches, with shifting mudflats making nesting habitat unsuitable (Crossland 2003; Goverse and Hilterman 2003; Fossette
et al.
2008). Beaches are often created and lost along the coast of French Guiana (Kelle
et al.
2007). For example, remote beaches in western French Guiana experience significant beach erosion and several disappeared, reducing or preventing monitoring (and likely nesting). In Suriname, Braamspunt Beach at the mouth of the Suriname River is moving west, out of the established Wia Wia Nature Reserve and may disappear in the next several years (M. Hiwat, WWF, pers. comm., 2018). This is significant in that Braamspunt is currently the main nesting beach in Suriname. The second highest nesting area in Suriname, Galibi Beach, is also experiencing significant erosion and becoming narrower. Similar beach erosion is occurring in Guyana, as well as in Trinidad and Tobago (Reichart
et al.
2003; Trinidad and Tobago Forestry Division
et al.
2010). At some Trinidad and Tobago nesting sites (
e.g.,
Fishing Pond, Matura, Grande Riviere, and Great Courland Bay), rivers emerge onto nesting beaches and create additional erosion during the nesting season (Godley
et al.
1993; Lee Lum 2005), intensifying nest loss (up to 35 percent of nests; Trinidad and Tobago Forestry Division
et al.
2010).
Seasonal erosion also occurs at most Caribbean nesting beaches. A survey of Wider Caribbean Regions found that erosion/accretion was the highest threat to nesting habitat (Dow
et al.
2007). For example, at Playa Gandoca, Costa Rica, erosion from strong coastal drift currents is thought to be one of the largest obstacles to hatching success, destroying greater than 10 percent of all
nests laid in some years (Chacón-Chaverri and Eckert 2007). In 2006 and 2007, coastal erosion and inundation accounted for 33 to 42 percent of nest loss in southern Panama and 29 to 48 percent on Caribbean Colombia beaches (Patiño-Martínez
et al.
2008).
Inundation of nests is also a concern. Leatherback turtles generally nest closer to the water than other sea turtles (Caut
et al.
2010). If nests are laid too close to the high tide line, they are subjected to erosion and inundation, which can result in egg mortality from suffocation or curtailed embryonic development (Chacón-Chaverri and Eckert 2007; Caut
et al.
2010). This inundation phenomenon occurs on multiple nesting beaches and is particularly of concern in areas with high tidal influence and dynamic coastlines. On Krofajapasi beach in Suriname, 31.6 percent of nests laid by females were below the spring high tide level and determined to be “doomed” clutches (Dutton and Whitmore 1983). Similarly, in Gandoca, Costa Rica, 37 percent of nests from 1990 to 2004 were laid in the low tide zone and would have been inundated if not relocated (Chacón-Chaverri and Eckert 2007). In St. Croix, 43 percent of the nests (with a range of 25 to 68 percent) were considered to be “doomed” each season (McDonald-Dutton
et al.
2001), but beginning in 1983, all doomed clutches were relocated to improve hatching success (Dutton
et al.
2005). Without intervention, these nests would likely have been lost. On Awala-Yalimapo, French Guiana, 27 of 89 nests were overlapped by tide at least once during the incubation period, and the hatching success was on average significantly lower in overwashed nests (Caut
et al.
2010). Observed mortality was 100 percent in the intertidal zone at sites along the coasts of Panama and Colombia, with an overall nest loss by erosion and inundation ranging from 16 to 48 percent among three major nesting sites (Patiño-Martínez
et al.
2008). While levels of inundation and resulting declines in hatching success have been noted at multiple sites throughout the range of the NW Atlantic DPS, the specific impacts of inundation may be variable. Hilterman and Goverse (2007) noted that leatherback nests can tolerate relatively high levels of inundation, so hatching may still be successful despite proximity to the tide line. Because of this, and because it may affect natural sex ratios (Mrosovsky and Yntema 1980), the relocation of nests susceptible to inundation was abandoned in 2002 in Suriname (Hilterman and Goverse 2007); only nests directly threatened by beach erosion are relocated, under certain circumstances. Other nations still relocate nests to reduce the impacts of erosion. However, as mentioned, such practices may result in cooler nests and affect sex ratios (Spanier 2008). While eggs relocated to hatcheries could have been lost under natural circumstances, due to coastal erosion and inundation in some areas (Dutton and Whitmore 1983, Chacón-Chaverri and Eckert 2007), hatching success in relocated nests is often lower than
in situ
nests (Revuelta
et al.
2014; Valentin-Gamazo
et al.
2018; Florida Department of Environmental Protection unpublished data 2018).
Such naturally dynamic areas make it difficult to protect nesting beach habitat and accurately assess leatherback nesting trends. This is particularly noteworthy given that nesting females use high energy, erosion-prone beaches, which often result in high nest loss (Chacón-Chaverri and Eckert 2007; TEWG 2007; Spanier 2008; Trinidad and Tobago Forestry Division
et al.
2010). However, leatherback turtles in the Guianas seem to have adapted to this constant geomorphological change of beaches. When new beaches develop, they may be colonized within months by nesting females, who take advantage of the fresh, clean sand (or seashells, in Guyana) and absence of entangling or deep-rooted beach vegetation (TEWG 2007).
Nest site selection by leatherback turtles is still poorly understood (Maison
et al.
2010), but nesting females may be changing their nesting patterns due to erosion. Spanier (2008) found that nesting females at Playa Gandoca, Costa Rica, appear to actively select nest sites that are not undergoing extensive erosion, with slope considered to be the cue for site selection. A similar result was found on Grande Riviere, Trinidad, with a nesting shift from east to west throughout the season as an apparent response to erosion on the eastern end of the nesting beach (Lee Lum 2005). Further, Maison
et al.
(2010) studied nest placement in Grenada and discovered that leatherback turtles seemed to respond to the accretion of the north facing beach and erosion of the east facing beach in 2005 by nesting more often on the north facing beach. If erosion is increasing in existing nesting locations, nesting may occur in areas with lower success rates, thus affecting productivity. In addition, leatherback nests are deeper than those of other sea turtles; water content and salinity typically increase with depth, leading to a decrease in sea turtle hatching success (Foley
et al.
2006).
Erosion Control, Nearshore Shoreline Stabilization Structures, and Other Barriers
A widespread strategy to reduce coastal erosion is to construct erosion control structures. However, these structures reduce the amount of available nesting habitat. Also, when beachfront development occurs, the site is often engineered to protect the property from erosion. This type of shoreline engineering, collectively referred to as beach armoring, includes sea walls, rock revetments, riprap, sandbag installations, groins and jetties. Beach armoring can result in permanent loss of a nesting beach through accelerated erosion and prevention of natural beach/dune accretion. These impacts can prevent or hamper nesting females from accessing suitable nesting sites (USFWS 1999). Clutches deposited seaward of these structures may be inundated at high tide or washed out entirely by increased wave action near the base of the erosion control structures. As these structures fail and break apart, they spread debris on the beach, thus creating additional impacts to hatchlings and nesting females.
In the southeastern United States, numerous erosion control structures that create barriers to nesting have been constructed. In Florida, the total amount of existing and potential future armoring along the coastline is approximately 24 percent (164 miles; FDEP, pers. comm., 2018). This assessment of armoring does not include other structures that are a barrier to sea turtle nesting, such as dune crossovers, cabanas, sand fences, and recreational equipment. Additionally, jetties have been placed at many ocean inlets in the United States to keep transported sand from closing the inlet channel. The installation of jetties resulted in lower loggerhead and green turtle nesting density updrift and downdrift of the inlets, leading researchers to propose that beach instability from both erosion and accretion may discourage turtle nesting (Witherington
et al.
2005). Leatherback nesting near jetties and inlets is low, possibly reflecting their avoidance of such areas. There are some efforts, such as the Coastal Construction Control Line Program, that provide protection for Florida's beaches and dunes while allowing for continued use of private property. However, armoring structures on and adjacent to the nesting beach continue to be permitted and constructed on the nesting beaches of Florida, as in other nations where the DPS nests.
Due to erosion, beach nourishment is a frequent activity in some developed
areas, and many beaches are on a periodic nourishment schedule. Beach nourishment may result in direct burial and disturbance to nesting females, if conducted during the nesting season. It may also result in changes in sand density, beach hardness, beach moisture content, beach slope, sand color, sand grain size, sand grain shape, and sand grain mineral content, if the placed sand is dissimilar from the original beach sand (Nelson and Dickerson 1988; USFWS 1999). These changes can affect nest site selection, digging behavior, incubation temperature (and hence sex ratios), gas exchange parameters within incubating nests, hydric environment of the nest, hatching success and hatchling emerging success (Lutcavage
et al.
1997; Steinitz
et al.
1998; Ernest and Martin 1999; USFWS 1999; Rumbold
et al.
2001; Brock
et al.
2009). On severely eroded sections of beach, where little or no suitable nesting habitat previously existed, beach nourishment has been found to result in increased nesting (Ernest and Martin 1999). However, on most beaches in the southeastern United States, nesting success typically declines for the first year or two following nourishment, even though more nesting habitat is available for turtles (Trindell
et al.
1998; Ernest and Martin 1999; Herren 1999; Brock
et al.
2009). Further, nourishment projects result in heavy machinery, pipelines, increased human activity and artificial lighting on the project beach, further affecting nesting females and beach habitat. Overall, the impacts of beach nourishment to this DPS are not as widespread as other threats to nesting habitat, as Dow
et al.
(2007) found that only four nations (Anguilla, Cuba, Mexico, and United States) reported frequent or occasional beach nourishment.
Artificial Lighting
Coastal development also contributes to habitat degradation by increasing light pollution, which can result in hatchling and nesting female disorientation, altering behavior and leading to mortality. In Florida, from 2013 to 2017, a total of 341 leatherback nests (representing the whole or majority of hatchlings in the nest) and five nesting females were disoriented (FWC unpublished data 2018). Artificial lighting ranked as the third highest threat to nesting/hatching turtles in the Wider Caribbean Region (Dow
et al.
2007). For example, urban development is significant in Puerto Rico, with light pollution (as well as coastal erosion and deforestation) occurring near leatherback nesting beaches (Crespo and Diez 2016). Fortunately, some of the major nesting beaches in this DPS are located in comparatively remote areas, and large-scale development is currently less of an issue there (Trinidad and Tobago Forestry Division
et al.
2010; NMFS and USFWS 2013). That said, even within the same country, light pollution is variable. Fossette
et al.
(2008) reported that in French Guiana, light pollution from residential areas is a problem at Cayenne Beach, but it is not an issue at Awala-Yalimapo. Similarly, lighting is not a significant problem on nesting beaches in Trinidad, but is a concern in Tobago (Trinidad and Tobago Forestry Division
et al.
2010). With the risk of increased development in some of these relatively remote areas, additional light pollution is anticipated, and disorientation of hatchlings and adults from such lighting may become a bigger problem. In Costa Rica, beachfront lighting is increasing and may become problematic at Gandoca Beach (Chacón-Chaverri and Eckert 2007) and Tortuguero (de Haro and Troëng 2006).
Light pollution has been managed to some extent (Witherington
et al.
2014). Lighting in Florida is regulated by multiple rules and regulations including Florida statutes, the Florida Building Code, and local lighting ordinances (Witherington
et al.
2014). In addition, the Florida Department of Transportation and local governments have adopted lighting-design standards. A total of 82 municipalities in Florida have adopted lighting ordinances to minimize the impact of lighting on adjacent sea turtle nesting beaches (Witherington
et al.
2014). However, compliance and enforcement is lacking in some areas. Further, lighting away from areas covered by beachfront ordinances is unregulated, resulting in urban glow. Although outreach and conservation programs control the impacts of lighting in some other locations, such as Costa Rica, Mexico, and Puerto Rico (Lutcavage
et al.
1997; Crespo and Diez 2016), a majority of nations do not have regulations in place.
Sand Extraction
Extracting sand from nesting beaches for construction projects has a detrimental effect on the amount of available nesting beach habitat and also accelerates erosion (resulting in the aforementioned associated impacts). Sand mining occurs in most Wider Caribbean nations to varying extent and frequency (Dow
et al.
2007). In particular, beach sand mining has been extensive at Matura Bay and Blanchisseuse in Trinidad (Trinidad and Tobago Forestry Division
et al.
2010). Some nations regulate sand mining: In St. Lucia, the Conservation and Management Act of 2014 requires a certificate of environmental approval for projects removing sand from nesting beaches.
Removal of Native Vegetation
In some nations, upland deforestation and the resultant deposition of debris and garbage can destroy or modify nesting beaches. The debris can block access of gravid (pregnant) females and fatally trap emergent hatchlings (Chacón-Chaverri and Eckert 2007). The accumulation of logs reduces the amount of available nesting habitat, possibly forcing leatherback females to nest in suboptimal locations (TEWG 2007). Deforestation due to coastal development is a notable concern in Puerto Rico (Crespo and Diez 2016).
Vehicular Traffic
Beach driving also occurs in most nations throughout the range of this DPS (Chacón-Chaverri and Eckert 2007; Dow
et al.
2007; Trinidad and Tobago Forestry Division
et al.
2010). In the United States, vehicular driving is allowed on certain beaches in Florida (
e.g.,
Duval, St. Johns, and Volusia Counties). Beach driving reduces the quality of nesting habitat in several ways. Vehicle ruts on the beach can prevent or impede hatchlings from reaching the ocean following emergence from the nest (Mann 1977; Hosier
et al.
1981; Cox
et al.
1994; Hughes and Caine 1994). Sand compaction by vehicles hinders nest construction and hatchling emergence from nests (Mann 1977; Gledhill 2007). Vehicle lights and vehicle movement on the beach after dark can deter females from nesting and disorient hatchlings. Additionally, vehicle traffic contributes to erosion, especially during high tides or on narrow beaches where driving is concentrated on the high beach and foredune.
Vegetation
Beach vegetation (native and non-native) can affect turtle nesting productivity by obstructing nest construction and potentially drying the sand (resulting in egg chamber collapse). Vegetation can form impenetrable root mats that can invade and desiccate eggs and affect developing embryos, impede hatchling emergence, and trap hatchlings (Conrad
et al.
2011). Non-native vegetation has invaded many coastal areas and often outcompetes native plant species (USFWS 1999). The occurrence of exotic vegetation (or loss of native vegetation) was recognized as a medium-ranked threat in many Wider Caribbean nations
(Dow
et al.
2007). The Australian pine (
Casuarina equisetifolia
) is particularly harmful to sea turtles (USFWS 1999). Australian pines cause excessive shading of the beach that would not otherwise occur. Studies of loggerhead turtles in Florida suggest that nests laid in shaded areas are subjected to lower incubation temperatures, which may alter the natural hatchling sex ratio (Marcus and Maley 1987; Schmelz and Mezich 1988). Fallen Australian pines limit access to suitable nest sites and can entrap nesting females (Reardon and Mansfield 1997). The shallow root network of these pines can interfere with nest construction (Schmelz and Mezich 1988). Dense stands of Australian pine have overtaken many coastal areas throughout central and south Florida.
While non-native vegetation can affect nesting habitat throughout the range of the DPS, native vegetation can also affect productivity. For instance, at Sandy Point, St. Croix, changing erosion-accretion cycles led to native
Ipomoea pes-caprae,
a creeping vine, extending into the nesting area in some years. Nesting females at Sandy Point typically avoided nesting in vegetation, resulting in more nests laid near the high-tide line (Conrad
et al.
2011). As a result,
Ipomoea pes-caprae
decreased nest productivity by reducing leatherback hatching and emergence (percentage of hatchlings that emerge from the nest) success rates (Conrad
et al.
2011).
Mitigations to Habitat Modification
Nesting habitat disruptions are minimized in some areas. Several areas in the NW Atlantic DPS range are under U.S. Federal ownership as National Wildlife Refuges in Florida (Archie Carr and Hobe Sound), Puerto Rico (Culebra and Vieques) and St. Croix (Sandy Point). Beaches in some Wider Caribbean countries are also protected. In Trinidad, Matura and Fishing Pond beaches were declared Prohibited Areas in 1990, and the nesting beach at Grande Riviere in 1997. In 1998, the Amana Nature Reserve, which includes Awala-Yalimapo beach and a 30 m wide marine fringe, was established in French Guiana. In Suriname, the Wia Wia Nature Reserve was implemented in 1961 (amended and enlarged in 1966 to protect sea turtles), and in 1969, the Marowijne beaches were declared a sanctuary (the Galibi Nature Reserve; Schulz 1971). In addition, Tortuguero National Park, Costa Rica, was established in 1976 to protect nesting habitat (Bjorndal
et al.,
1999). Terrestrial habitat in these areas is therefore protected from the above threats to some extent. USFWS and NMFS also designated as critical habitat for leatherback turtles the nesting beaches at Sandy Point, St. Croix (43 FR 43688; September 26, 1978) and surrounding marine waters (44 FR 17710; March 23, 1979), which benefits the turtles in this DPS. However, if ESA protections did not continue (
i.e.,
if this species were no longer listed), these protections would be lost.
Marine Habitat Modifications
In the marine environment, habitat threats include anthropogenic noise and offshore lighting. We discuss other threats to marine habitat and prey (
e.g.,
marine pollution, oil exploration, and climate change) in later sections. Anthropogenic noise impacts the marine habitat of the DPS. Dow Piniak
et al.
(2012) measured hearing sensitivity of leatherback hatchlings. They found that hatchlings are able to detect sounds underwater and in air, responding to stimuli between 50 and 1200 Hz in water and 50 and 1600 Hz in air, with maximum sensitivity between 100 and 400 Hz in water and 50 and 400 Hz in air. This sensitivity range overlaps with the frequencies and levels produced by many anthropogenic sources used in the North Atlantic, including seismic airgun arrays, drilling, low frequency sonar, shipping, pile driving, and operating wind turbines. These noise sources may affect leatherback turtles' marine habitat and subsequently impact distribution and behavior. Offshore artificial lighting occurs in some marine waters of this DPS (Dow
et al.
2007) but is less of a threat than beachfront lighting throughout the range of the DPS.
Summary
We conclude that nesting females, hatchlings, and eggs are exposed to the loss and modification of nesting habitat, especially as a result of coastal development and armoring, erosion, and artificial lighting. These threats impact the DPS by reducing nesting and hatching success, thus, lowering the productivity of the DPS. Based on the information presented above, we conclude that habitat reduction and modification pose a threat to the NW Atlantic DPS.
Overutilization for Commercial, Recreational, Scientific, or Educational Purposes
Overutilization is a threat to the NW Atlantic DPS, mostly due to poaching of turtles and eggs in certain nations. Legal harvest of turtles and eggs also occurs in some nations.
While the vast majority of nations within the range of the NW Atlantic DPS protect leatherback turtles from harvest, it is legal in some Caribbean and Central American nations (Brautigam and Eckert 2006; Dow
et al.
2007; Richardson
et al.
2013; Horrocks
et al.
2016). For example, the harvest of leatherback turtles over 20 pounds is allowed in Montserrat and Dominica from October 1 to May 31; Saint Lucia allows leatherback turtles over 65 pounds to be taken from October 2 to February 27; and St. Kitts and Nevis allows take of leatherback turtles over 350 pounds from October 2 to February 27 (Montserrat Turtles Act 2002; Bräutigam and Eckert 2006). In some nations, commercial use is prohibited, but traditional use is allowed, which can still diminish protection. In Colombia, subsistence fishing of sea turtles is permitted, and indigenous use is allowed in Honduras. Traditional or cultural use is permitted in Belize with prior approval (Bräutigam and Eckert 2006). However, regular leatherback nesting does not occur in Belize, and its occurrence in surrounding waters is infrequent, reducing the impact of such mortality. Legal harvest throughout the range of this DPS is not monitored, and the precise magnitude of this threat is not clear. However, we conclude that legal harvest of turtles is significant because, when it occurs, nesting turtles are targeted, removing the most important individuals from the population. More often, leatherback eggs, rather than turtle meat, are harvested (TEWG 2007; Patiño-Martínez
et al.
2008), reducing productivity in the DPS.
Poaching of turtles and eggs occurs throughout the NW Atlantic DPS, and Dow
et al.
(2007) ranked it as a threat for all turtle species on the beaches in the Wider Caribbean Region. In Panama, interviews with locals revealed that the development of a new way for cooking leatherback turtle meat has resulted in a recent increase of its consumption in Changuinola, Bocas del Toro Province (CITES Secretariat 2019). Adult turtles are killed in Panama and on remote beaches in Trinidad and Tobago (Troëng
et al.
2002; Ordoñez
et al.
2007; Trinidad and Tobago Forestry Division
et al.
2010). Most poaching, however, targets eggs, and the level often is determined by how much monitoring and activity to deter poachers occur on the nesting beaches. Some of the highest levels of egg poaching occur throughout Costa Rica (Troëng
et al.
2004). Troëng
et al.
(2007) found that, at a minimum, between 13 to 21.5 percent of nests between 2000 and 2005 were illegally
collected at Tortuguero. Poaching of leatherback nests was higher outside Tortuguero National Park (minimum 33 percent) than within the National Park (minimum 9 percent) in 2005 (de Haro and Troëng 2006). At Pacuare Playa, Costa Rica, 55 percent of nests were poached in 2012 (Fonseca and Chacón 2012) and 42 percent were poached in 2017, which was the lowest level since Latin American Sea Turtles (LAST) started to monitor in 2012 (LAST 2017). Poaching at Gandoca Beach has decreased over time (previously 100 percent of nests were poached), but rates still averaged 15.5 percent annually from 1990 to 2004 (Chacón-Chaverri and Eckert 2007). In the Dominican Republic, poaching is also high. Revuelta
et al.
(2012) determined the poaching of clutches in Jaragua National Park and Saona Island ranged from 0 to 100 percent from 2006 to 2010, with averages of 19 percent on western Jaragua National Park beaches, 71 percent on eastern Jaragua National Park beaches, and 74 percent on Saona. Poaching also occurs at relatively high levels in Colombia (
e.g.,
22 to 31 percent of clutches at Playona in 2006 and 2007; Patiño-Martínez
et al.
2008) and, to some extent, in most other Caribbean nations (
e.g.,
Guyana and Grenada). Poaching is likely more prevalent, and occurs at higher levels, on unmonitored or unprotected beaches (Dow
et al.
2007; TEWG 2007; Troëng
et al.
2007; Trinidad and Tobago Forestry Division
et al.
2010; K. Charles, Oceans Spirits Inc., pers. comm., 2018).
Poaching has been significantly reduced at some nesting beaches. In Suriname, high levels of egg poaching (at least 26 percent of nests) occurred in the late 1990s, but due to better monitoring and enforcement, that level has been significantly reduced (Hilterman and Goverse 2007; M. Hiwat, WWF, pers. comm., 2018). Poaching was also a major problem in Trinidad, but levels have been reduced with more people monitoring the beach (Trinidad and Tobago Forestry Division
et al.
2010). The Marine Turtle Conservation Act of 2004 (MTCA) funds activities in Panama in an attempt to reduce poaching. At Chiriqui Beach, Panama, intense monitoring efforts have attempted to reduce poaching. However, of the monitored nests, 29 leatherback nests (0.7 percent) were still poached in 2017 (Sea Turtle Conservancy 2017). Further, poaching in Panama outside the monitored areas still occurs, with the clandestine sale of eggs widespread (Brautigam and Eckert 2006). In St. Croix, almost 100 percent of nests were lost to poaching prior to 1981 (Garner
et al.
2017). However, the establishment of the USFWS Sandy Point National Wildlife Refuge has reduced egg poaching to 0 to 1.8 percent annually as a result of nightly patrols (Garner
et al.
2017).
Poaching of eggs is widespread throughout the Caribbean, especially on beaches of Costa Rica, Dominican Republic, and Colombia. The total number of individuals affected by poaching cannot be quantified at this time. However, we conclude that many eggs and some adults are affected by illegal poaching at nesting beaches. Adults and eggs are also exposed to legal harvest in some nations. The legal and illegal harvest of nesting females reduces both abundance (through loss of nesting females) and productivity (through loss of reproductive potential), resulting in a high impact to the DPS. Legal and illegal egg harvest reduces productivity only. Thus, we conclude that overutilization poses a threat to the DPS.
Disease or Predation
For the NW Atlantic DPS, information on diseases is limited, but predation is a well-documented threat.
Much of the available information on disease in leatherback turtles was obtained by necropsy of stranded large juvenile and adult turtles; the health implications of various conditions reported in this species are incompletely understood. Solitary large intestinal diverticulitis of unknown etiology was found in 31 subadult and adult leatherback turtles stranded in U.S. waters (Stacy
et al.
2015). All lesions were chronic and unrelated to the cause of death in all cases, although risk of perforation and other complications are possible. Adrenal gland protozoal parasites were found in 17 leatherback turtles in North American waters examined from 2001 to 2014; it is not currently known whether parasitism affects adrenal function (Ferguson
et al.
2016). In addition, leatherback turtles are hosts for several trematode parasites (flatworms), known species of which also occur in hard-shelled sea turtles (Manfredi
et al.
1996, Greiner
et al.
2013). In general, trematodes are frequently encountered without any apparent clinical effect on the turtle host but can affect some heavily parasitized individuals. With regard to other types of potential disease-causing organisms, there are a small number of reports of bacterial infections in stranded individuals (Poppi
et al.
2012; Donnelly
et al.
2016). A variety of other bacteria have been documented in nesting females on beaches in Costa Rica (Santoro
et al.
2008) and St. Kitts (Dutton
et al.
2013); the majority of identified bacterial species may be considered as potential or opportunistic pathogens for sea turtles. A putative case of fibropapilloma, a virus-associated tumor-causing disease in sea turtles, has been reported in a leatherback; this disease is considered very rare in the species (Huerta
et al.
2002).
An in-water health assessment was performed on 12 turtles directly caught at-sea and seven turtles bycaught in fishing gear in the NW Atlantic Ocean (Innis
et al.
2010). Most were determined to be in good health, but several exhibited evidence of past injuries. The blood chemistry of entangled turtles indicated stress, seawater intake, and reduced food consumption associated with entanglement. In addition, Perrault
et al.
(2012) examined baseline blood chemistry metrics (
i.e.,
plasma protein electrophoresis, hematology, and plasma biochemistry) as indicators of health for nesting females in Florida. They found that multiple measures of maternal health significantly correlated with leatherback hatching and emergence success (the percentage of hatchlings that emerge from the nest).
From these data, we estimate that the exposure of eggs, juveniles, and adults to disease is low. The impact of disease cannot be quantified at this time as we have no documentation of any deaths or reductions in productivity directly related to disease. However, disease may compound the effects of or have synergistic effects with other threats to the species and related physiologic derangements. We conclude that disease, alone or in combination with other threats, is likely a threat to the DPS.
Throughout the range of the DPS, predation is a threat to leatherback eggs, hatchlings, and adults. Eckert
et al.
(2012) provides an exhaustive list of the documented predators for each life stage and area. For eggs in the NW Atlantic DPS, predators include ants (
Dorylus spininodis
), fly larvae (
Diptera
spp.), locust larvae (
Acrididae
spp.), mole crickets (
Scapteriscus didactylus
), ghost crabs (
Ocypode quadratus
), vultures (Cathartidae), dogs (
Canis familiaris
), cattle (
Bos taurus;
due to trampling), armadillo (Dasypodidae), opossum (
Didelphis marsupialis
), coati (
Nasua
spp.), and raccoons (
Procyon lotor
); see Eckert
et al.
2012).
In particular, dog predation of eggs occurs in many areas (
e.g.,
Colombia, French Guiana, Guyana, Panama, Puerto Rico, and Trinidad and Tobago). In Trinidad, where the largest nesting aggregation occurs, feral dogs are
considered to be the primary threat to eggs, even above poaching and coastal erosion (Trinidad and Tobago Forestry Division
et al.
2010). On Chiriqui Beach, Panama, 54 percent of the monitored leatherback nests were depredated by dogs in 2003 and approximately eight percent in 2004 (Ordoñez
et al.
2007). Such predation may been reduced as a result of protection efforts funded by the MTCA. In Playa California, Maunabo, Puerto Rico, more than 30 percent of the leatherback nests were depredated by stray dogs in 2012 (Crespo and Diez 2016). A public outreach project in Puerto Rico was established in 2013 to reduce this impact. Puerto Rico is a U.S. territory; if ESA protections were removed, it is likely that predation rates would be higher.
Egg predation by other species is also a notable concern in some areas. On Gandoca Beach, Costa Rica, dipteran larvae infestation exceeded 75 percent of nests in 2005 and 2006 (Gautreau
et al.
2008). In French Guiana, on average, mole crickets preyed on 18 percent of all eggs (Maros
et al.
2003). These threats are likely to continue, as no predator screening typically occurs in Wider Caribbean nations due to the potential for increased poaching as well as logistical difficulties in these areas of high density nesting. Nest loss to predators was found to be the seventh ranked threat to turtles (all species, not specific to leatherback turtles) on nesting beaches in the Wider Caribbean Region, and have been noted to frequently occur in Honduras, Mexico, Panama, Puerto Rico, and Venezuela (Dow
et al.
2007).
Hatchlings are preyed upon by a wide variety of species, including mole crickets, ghost crabs, horse-eye jack fish (
Caranx latus
), gray snapper (
Lutjanus griseus
), tarpon (
Megalops atlanticus
), vultures, hawks (Accipitridae), gulls (
Larus
spp.), night heron (
Nyctanassa violacea
), frigate birds (Fregatidae), dogs, mongoose (
Atilax paludinosus
), coati, and raccoons (Eckert
et al.
2012). Again, dogs are a serious threat to leatherback hatchlings in some areas, and especially in Puerto Rico (Crespo and Diez 2016).
There are few documented predators to subadults and adult leatherback turtles, presumably because of their large size and pelagic behavior. Predation by sharks (Elasmobranchii) and killer whales (
Orcinus orca
) has been reported in Barbados and St. Vincent, respectively (Caldwell and Caldwell 1969; Horrocks 1989). Sharks have also been reported to prey on nesting females off St. Croix, USVI (DeLand 2017; Scarfo
et al.
2019). Over the past 6 years, researchers at Sandy Point have observed an apparent increase in injuries to leatherback turtles (K. Stewart, NMFS, pers. comm., 2019). These injuries, many of them consistent with shark predation, affect up to 70 percent of all nesting females at the beach (Scarfo
et al.
2019). While some turtles probably survive these encounters, it is unknown how many encounters result in mortality or reduced nesting effort. Jaguars (
Panthera onca
) prey on nesting females in some areas, including Suriname, French Guiana, Guyana, and Costa Rica (see Eckert
et al.
2012). While three nesting females were killed by jaguars at Tortuguero, Costa Rica, from 1998 to 2005, this mortality is only considered to be a minor threat and is therefore unlikely to cause a population decline on its own (Troëng
et al.
2007). Archibald and James (2018) examined 228 leatherback turtles for injuries off Atlantic Canada and on Matura, Trinidad, and found 15.7 percent of turtles exhibited injuries of suspected predatory origin.
Predation on early life stages is natural; however, at high rates, it reduces the viability of the DPS (see the Status Review). Predation primarily reduces productivity via reduced egg and hatching success and the loss of hatchlings. Predation on nesting females reduces abundance and productivity. We conclude that predation is a threat to the NW Atlantic DPS.
Inadequacy of Existing Regulatory Mechanisms
Many regulatory mechanisms (including state, Federal and international) have been promulgated to protect leatherback turtles, eggs, and nesting habitat throughout the range of the NW Atlantic DPS. We reviewed the objectives of each regulation and to what extent they adequately address the targeted threat (
i.e.,
the threat that the regulation was intended to address). The effectiveness of many international regulations was evaluated by Hykle (2002), who found that international instruments often do not realize their full potential, either because they do not include all key countries, do not specifically address sea turtle conservation, are handicapped by the lack of a sovereign authority that promotes enforcement, or are not legally binding.
National regulatory mechanisms are described in full in the Status Review Report. Although these regulatory mechanisms provide some protection to the species, most inadequately reduce the threat they were designed to address, generally as a result of poor implementation or incomplete enforcement. Specifically, existing regulatory mechanisms continue to be inadequate to control impacts to nesting beach habitat and overutilization (harvest of turtles and eggs) for this DPS. In addition, regulatory mechanisms are inadequate to reduce several other threats including bycatch in fishing gear, vessel strikes, and marine debris. Despite existing regulatory mechanisms, bycatch from fisheries (discussed in detail along with existing regulatory mechanisms in the Fisheries Bycatch section), incomplete nesting habitat protection, and poaching remain major threats to the DPS.
Fisheries Bycatch
Fisheries bycatch is the primary threat to the NW Atlantic DPS. Bycatch occurs throughout the range of the DPS, affecting juveniles, subadults, and adults.
Finkbeiner
et al.
(2011) analyzed sea turtle bycatch across all commercial U.S. fisheries from 1990 to 2007. They examined sea turtle bycatch reduction based on the year a particular fishery implemented bycatch reduction measures. Prior to implementing bycatch reduction measures, approximately 3,800 leatherback interactions, of which 2,300 were lethal, occurred in U.S. Atlantic Ocean and GOM commercial fisheries annually. After bycatch reduction measures were implemented, 1,400 leatherback turtles, 40 of those dead, were estimated to be taken annually in the Atlantic Ocean. The Atlantic/GOM pelagic longline fishery was responsible for the most annual interactions (n = 900) and mortality events (n = 17) in the Atlantic Ocean, followed by the southeast Atlantic/GOM shrimp trawl fishery (Finkbeiner
et al.
2011). These estimates represent minimum numbers of actual bycatch and mortality. Because the observer coverage for these fisheries is low (so some bycatch may not be observed and observed effort may not be a true representation of actual fleet effort), not all fisheries are observed and thus some are not included in these estimates. Interactions are difficult to observe if gear modifications are in place, and so the methods used are conservative (Finkbeiner
et al.
2011).
In the Wider Caribbean Region, reports of leatherback bycatch in fisheries are common. In a survey of Caribbean nations, Dow
et al.
(2007) ranked fisheries bycatch among the highest in-water threat to sea turtles. Many fisheries in less industrialized nations are coastal and small-scale, but these fisheries are reported to have significant ecological impacts due to their high bycatch discards and impacts
to the marine environment (Shester and Micheli 2011). Of particular concern are leatherback bycatch in artisanal nearshore and offshore gillnet, longline and trawl fisheries (Barrios-Garrido and Montiel-Villalobos 2016). Information on fisheries bycatch is collected mostly from stranding records but also from fisher surveys (Moncada
et al.
2003; Delamare 2005; Madarie 2006, 2010, 2012) and observations of nesting females. Hilterman and Goverse (2007) recorded fisheries related injuries on nesting females in Suriname. In 2002, 16.9 percent of the nesting females had fisheries- related injuries; in 2003, at least 18.3 percent had such injuries; and in 2005, 9 percent (Hilterman and Goverse 2007). From 2000 to 2003, an average of 28 leatherback turtles stranded on the Suriname survey beaches. Although no cause of death was immediately apparent, Hilterman and Goverse (2007) indicated that the mortalities were fisheries-related, based upon the fisheries that occur offshore with high bycatch and documented fisheries-related injuries on nesting leatherback turtles at the same time. On the western oceanic nesting beaches of French Guiana, injuries consistent with fisheries interactions (
e.g.,
scars, wounds) were recorded on 8.4 percent (n = 1,259) of nesting females in 2003 (Morisson
et al.
2003). In Venezuela, 55 percent of strandings from 2001 to 2007 (n = 57) exhibited evidence of fisheries interactions (Barrios-Garrido and Montiel-Villalobos 2016). Most recently, an injury assessment of 228 leatherback turtles from two foraging areas off the Atlantic coast of Canada and Trinidad nesting beaches found 19 percent of turtles exhibited injuries indicative of entanglement in lines or nets, and 17 percent showed evidence of hooks; 62 percent of turtles assessed exhibited a minimum of one external injury (Archibald and James 2018).
Fisheries bycatch also occur in the Mediterranean and eastern North Atlantic Ocean. Casale
et al.
(2003) analyzed 411 records of leatherback turtles in the Mediterranean, of which 152 were collected from Italy. Most of these records were from fishery captures (n = 170) or found in unknown circumstances (n = 127). Of those reported by fishermen, set or drift nets had the highest number of interactions (29.4 percent), followed by unknown fishing equipment (22.9 percent), longlines (20.6 percent), unspecified nets (12.9 percent), other fishing equipment (9.4 percent), and trawls (4.7 percent). The main fisheries affecting turtles in the Mediterranean (all turtle species, not just leatherback turtles) are Spanish and Italian surface longlines, North Adriatic Italian trawls, Tunisian trawls, Turkish trawls, Moroccan driftnets, and Italian driftnets (Camiñas 2004). The same types of fishing gear from other nations also affect turtles, but the bycatch numbers are lower (Camiñas 2004). Stranding records from Portugal from 1978 to 2013 found that 49 of 275 leatherback turtles exhibited evidence of fishery interactions (the cause of stranding could not be determined in most cases due to decomposition state; Nicolau
et al.
2016). Multifilament nets accounted for approximately 41 percent of the strandings, followed by monofilament nets, traps/pots, and longlines. Coastal artisanal fisheries were recognized as a particular threat in Portugal.
Based upon these summary reports and stranding assessments, it is clear that fisheries have a large impact on the NW Atlantic DPS. In the following paragraphs, we review information on specific gear interactions, including the following fisheries: Gillnet, longline, trawl, pot/trap, and other.
Gillnet Fisheries
Gillnet fisheries are common throughout the range of this DPS. Due to the nature of the gear and fishing practices (
e.g.,
relatively long soak times), bycatch in gillnets is among the highest source of direct sea turtle mortality (Upite
et al.
2013; Wallace
et al.
2013; Upite
et al.
2018). Upite
et al.
(2018) evaluated observed fishery interactions and post-interaction mortality and determined a 79 percent sea turtle mortality rate for Northeast and Mid-Atlantic gillnet gear from 2011 to 2015. Wallace
et al.
(2013) calculated leatherback bycatch in gillnets throughout the NW Atlantic Ocean of 0.015 turtles/set, with a 21 percent median mortality rate (not considering post-interaction mortality). This gear was classified as having a relatively high bycatch impact on the NW Atlantic leatherback population. Small scale fisheries are of particular concern, given the magnitude of bycatch, nearshore distribution, and limited monitoring (Lewison
et al.
2015). When nets are used in waters off nesting beaches, where leatherback turtles mate, nesting females and mature males are often captured and killed.
The largest documented bycatch of leatherback turtles in gillnet gear occurs off the coast of Trinidad. Lee Lum (2006) estimated that more than 3,000 leatherback turtles were captured by coastal surface gillnets off Trinidad annually, with an approximate 30 percent mortality rate. These captures involved adult turtles, occurring off the north and east coasts of Trinidad during January to August,
i.e.,
the breeding and nesting season, when nesting females and adult males occur in the waters off nesting beaches (Lee Lum 2006). Gilman
et al.
(2010) extrapolated leatherback bycatch estimates (Lee Lum 2006; Gearhart and Eckert 2007) to the entire Trinidad Spanish mackerel and king mackerel surface gillnet fishery, and estimated that almost 7,000 turtles were captured in 2000. Additionally, Eckert
et al.
(2013) worked with drift gillnet fishermen to identify leatherback bycatch hot spots off the north and east coasts of Trinidad (where the nesting beaches are), with capture probability increasing from March to July and a secondary peak in October.
Whereas most of the documented leatherback bycatch off Trinidad occurs in surface drift gillnet fisheries, bottom set gillnet fishing also captures leatherback turtles (Gass 2006; S. Eckert, WIDECAST, pers. comm., 2018). The magnitude of effort and turtle bycatch in this fishery are lower than for surface nets, but mortality rates are higher (approximately 70 percent; Gass 2006). As such, the bottom set gillnet fishery is thought to have a comparable level of mortality to the drift gillnet fishery (approximately 500 to 1,000 leatherback turtles annually; Gass 2006; S. Eckert, WIDECAST, pers. comm., 2018). The Sea Turtle Recovery Action Plan for the Republic of Trinidad and Tobago noted that drowning in gillnets is that nation's most significant cause of sea turtle mortality (Trinidad and Tobago Forestry Division
et al.
2010). Bond and James (2017) tracked a female from Canadian waters to a nesting beach off Trinidad, but the turtle was confirmed dead, entangled in coastal fishing gear, just prior to the date of her first predicted nesting event. Venezuelan fishers have also been seen hauling leatherback turtles from Trinidad waters into their boats (Brautigam and Eckert 2006). Together, drift and bottom-set gillnets off the Trinidad beaches, which host the largest nesting aggregation in the DPS, are estimated to kill well over 1,000 leatherback turtles annually, and they thus pose a large threat to the DPS.
High levels of gillnet bycatch occur in other Caribbean and South American nations, also off major nesting beaches. In French Guiana, bycatch was confirmed to be high in the Maroni estuary (Chevalier 2001; Girondot 2015). In 2003, 26 leatherback turtles were caught in coastal gillnets and released off the Cayenne and Montjoly nesting sites (Gratiot
et al.
2003 in TEWG 2007). Delamare (2005) conducted fishermen interviews and estimated an average of 1,149 leatherback captures in 2004 and
2005 by bottom-set or drifting gillnets in French Guiana. No estimate of mortality was provided, but it is likely similar to Trinidad fisheries,
i.e.,
70 and 30 percent, respectively. In Suriname, a World Wildlife Fund survey of fishermen estimated leatherback bycatch in drifting gillnets at 584 in 2006, 174 in 2010, and 424 in 2012 (Madarie 2006; Madarie 2010; Madarie 2012). Most of the turtles were captured alive. In Colombia, 10 to 40 leatherback turtles are killed annually by gillnets (Patiño-Martínez
et al.
2008). Longline and driftnet gillnet fisheries in Moroccan waters off the northwestern Africa coast capture approximately 100 leatherback turtles annually (Benhardouze
et al.
2012).
Although not at as high a rate as in the Caribbean (based upon observed interactions), gillnet bycatch occurs in U.S. and Canadian waters. Although South Carolina, Georgia, Florida, Louisiana, and Texas have prohibited gillnets in their State waters, active gillnet fisheries remain in other states and U.S. Federal waters. No cumulative estimates of leatherback bycatch in gillnet fisheries in U.S. waters are available due to the limited observed interactions. However, from 2003 to 2017, fishery observers recorded lethal and non-lethal bycatch in fixed sink, drift sink, and drift floating gillnets throughout the U.S. Atlantic Exclusive Economic Zone (EEZ) and GOM (NMFS unpublished data). From 2012 to 2016, 27 leatherback turtles (coefficient of variation = 0.71, 95 percent CI over all years: 0-68) were bycaught with 21 mortalities in sink gillnet gear in the Georges Bank and Mid-Atlantic regions (Murray 2018). From 1989 to 1998, U.S. drift pelagic gillnets captured 54 leatherback turtles, but that gear is no longer used. Hamelin
et al.
(2017) reviewed leatherback entanglement records reported by Canada in Atlantic Canadian waters between 1998 and 2014. Gillnets, mainly targeting groundfish, were involved in 24 of 205 entanglements (11.7 percent), particularly in Newfoundland and Labrador (n = 15). Often, gillnet entanglements involve the vertical lines associated with gear (M. James, DFO, pers. comm., 2019).
Gillnet bycatch occurs in the eastern North Atlantic Ocean and in the Mediterranean Sea. As in other areas, sea turtles have the potential to interact with set gillnets and drift gillnets. The United Nations (UN) established a worldwide moratorium on drift gillnet fishing effective in 1992; the General Fisheries Commission for the Mediterranean prohibited driftnet fishing in 1997; a total ban on driftnet fishing by the European Union fleet in the Mediterranean went into effect in 2002; and the International Commission for the Conservation of Atlantic Tunas (ICCAT) banned driftnets in 2003. Nevertheless, unregulated driftnetting continued to occur in some areas (
e.g.,
the Mediterranean Sea and off Europe; Pierpoint 2000; Camiñas 2004). In the Atlantic Ocean, leatherback bycatch has been reported from NE Atlantic tuna driftnet fisheries by English, French and Irish vessels (Pierpoint 2000). Of 20 leatherback turtles found in nets in British and Irish waters (1980 to 2000), eight were caught in the NE Atlantic tuna driftnet fishery (with 25 percent mortality) and one was caught in a hake gillnet (Pierpoint 2000).
Historically, driftnet fishing in the Mediterranean Sea caught large numbers of sea turtles. And today an estimated 600 illegal driftnet vessels operate in the Mediterranean, including fleets based in Algeria, France, Italy, Morocco, and Turkey (Environmental Justice Foundation 2007). Out of 411 records of leatherback turtles (stranded, captured, sighted, or found in unknown circumstances) in the Mediterranean Sea, 170 turtles were captured by fishermen, of which 29.4 percent were caught by set or drift nets (Casale
et al.
2003). Driftnets and gillnets in Greece, Israel, Italy, Tunisia and Turkey have reported documented leatherback interactions, and occasional leatherback bycatch occurs in Croatian artisanal gillnet fisheries (Camiñas 2004; Ergene and Ukar 2017). In particular, Karaa
et al.
(2013) reviewed 36 leatherback bycatch records from Tunisia fisheries in the Gulf of Gabes, and found that gillnets are the dominant threat to leatherback turtles in the region. A similar result (
e.g.,
gillnets being a high threat to leatherback turtles in the area) was found in the Adriatic Sea (Lazar
et al.
2012). The first leatherback recorded on the Aegean coast of Turkey was caught in a gillnet (Taskavak
et al.
1998). Further, a review by Casale (2008) found that leatherback turtles are taken in the drift gillnet fishery in Spain at a rate of 0.065 turtles/day-boat.
Throughout the range of the NW Atlantic DPS, effective gillnet bycatch reduction measures have not been required, but measures to reduce leatherback bycatch have been discussed in some areas (
e.g.,
Trinidad; Eckert 2013). If nations have a closed season for fishing, at least in the nesting season (
e.g.,
Suriname; Madarie 2006), nesting females are afforded some level of protection from gillnet bycatch. Some nations have prohibited gillnet gear; St. Barthelemy does not allow trammel nets in its territorial waters and St. Lucia prohibits fishing within 100 meters of shore to protect nesting turtles. There are gillnet and trammel net restrictions in Curacao (Ministry of Health, Environment, and Nature 2014, UN Environment Programme 2017). In the United States, gillnets with stretched mesh seven inches and larger are prohibited at certain times off North Carolina and Virginia to protect sea turtles (50 CFR 223.206(d)(8); 71 FR 24776, April 26, 2006). While no gear modifications are currently required for U.S. gillnet fisheries, Federal U.S. fisheries are subject to section 7 of the ESA, 16 U.S.C. 1536(a)(2), and through formal consultations on specific fisheries, measures may be required to minimize the impact of incidental take in gillnets (NMFS 2013). Regardless of some of these protective measures, gillnet bycatch (especially off nesting beaches) results in the loss of thousands of mature individuals annually.
Longline Fisheries
Leatherback turtles are known to interact with longline fishing gear, most commonly pelagic longlines (Lewison
et al.
2004; Zollett 2009; Wallace
et al.
2010; Wallace
et al.
2013). There is significant concern over the effects of pelagic longline fishing, which extends globally throughout temperate and tropical waters, including several high pressure fishing areas in the North Atlantic Ocean (Fossette
et al.
2014; Gray and Diaz 2017). In international waters, numerous flag states have high seas longline fisheries that frequently catch leatherback turtles (Lewison
et al.
2004). Individuals are found entangled and hooked in this gear, mostly by the flippers (Witzell and Cramer 1995; Coelho
et al.
2015; Huang 2015). Leatherback bycatch in longlines throughout the NW Atlantic Ocean was calculated at 0.062 turtles per set, classifying the gear as a relatively low bycatch impact relative to other sea turtle populations (Wallace
et al.
2013; Lewison
et al.
2015). However, because longline fisheries are widespread across leatherbacks' distribution and use millions of hooks each year, they pose a large threat to the NW Atlantic DPS and are estimated to kill thousands of individuals (mature and immature) annually.
Pelagic longline fishing is widespread throughout the range of the DPS and involves a number of nations, so an accurate estimate of total bycatch is difficult to obtain. In the Atlantic Ocean from 2002 to 2013, the largest longline fishing fleets belonged to Taiwan, Japan, Spain, Belize, and China, with the Taiwanese fleet comprising the largest distant-water longline effort throughout
the region (Angel 2014; Huang 2015). In an assessment of the impact of ICCAT fisheries on sea turtles, Gray and Diaz (2017) estimated leatherback interactions with pelagic longlines in the ICCAT area from 2012 to 2014 (15 to 16 fleets). Using a combination of published and assigned sea turtle bycatch rates as a function of estimated fishing effort submitted to ICCAT by its members, Gray and Diaz (2017) found a high degree of overlap in the central North Atlantic Ocean and equatorial waters (some of which are outside this DPS). Within the NW Atlantic region, an estimated 7,138 leatherback interactions occurred in 2012, 6,036 in 2013 and 4,991 in 2014 (Gray and Diaz 2017). Applying a reasonable estimated mortality rate of 21.4 percent, as seen in other high seas pelagic longline gear (Huang 2015), results in an average annual estimated mortality of 1,296 leatherback turtles from 2012 to 2014. However, this is likely an underestimate of total mortality, as the high seas mortality rate in Huang (2015) was based upon the disposition of the turtle when boarded and therefore did not account for post-interaction mortality; 240 of 459 leatherback turtles caught from 2002 to 2013 were alive and 121 were of unknown status (Huang 2015). Angel
et al.
(2014) conducted a risk assessment of turtles from the impacts of tuna fishing in the ICCAT region and found the NW Atlantic RMU (which is comparable to the NW Atlantic DPS; Wallace
et al.
2010) has high-moderate vulnerability to longline gear, with as many as 270 million longline hooks annually from 2000 to 2009. In particular, Fossette
et al.
(2014) analyzed leatherback satellite tracks (converted to densities) overlaid with longline fishing effort from 1995 to 2009 in the Atlantic Ocean. In the North Atlantic Ocean, a total of four seasonal high-susceptibility areas were identified: one in the central northern Atlantic in international waters, one along the east coast of the United States, and one each in the Canary and Cape Verdean basins (Fossette
et al.
2014). These areas partly occurred in the EEZs of eight nations (Cape Verde, Gambia, Guinea Bissau, Mauritania, Senegal, Spain/Canaries, United States, and Western Sahara). Given the species' flexible diving behavior, it is reasonable to expect that turtles are likely to encounter pelagic longlines throughout the Atlantic Ocean, regardless of whether they are engaged in foraging or migratory behavior (Fossette
et al.
2014).
Bycatch in U.S. Atlantic and GOM pelagic longlines has been extensively studied in the last decade. Current estimates of leatherback interactions with the U.S. Atlantic pelagic longline fishery are lower than previous years. In the late 1990s and early 2000s, estimates of Atlantic U.S. pelagic longline bycatch were around 1000 leatherback turtles annually (NMFS 2001; Yeung 2001; NMFS 2018), with bycatch rates of about 0.15 to 0.5 turtles per 1000 hooks (Watson
et al.
2005). In 2005, after the United States required pelagic longline gear modifications (50 CFR 635.21), the fleet was estimated to have interacted with 351 leatherback turtles outside experimental fishing operations (Walsh and Garrison 2006). NMFS (2018) estimated 239 leatherback interactions in the U.S. Atlantic pelagic longline fishery in 2011, 596 in 2012, 363 in 2013, 268 in 2014, 299 in 2015, and 339 in 2016. The majority of interactions occurred in the GOM, Mid-Atlantic Bight, Northeast Coastal, and Northeast Distant areas (NMFS 2018). The post-interaction mortality estimate for the most recently available 3-year period (2013 to 2015) for leatherback turtles is 30.13 percent (L. Desfosse, NMFS, pers. comm., 2018). Based on the average leatherback interaction estimate for the entire U.S. pelagic longline fleet from 2011 to 2016 (351), the estimated average annual mortality for the U.S. pelagic longline fishery is 106 leatherback turtles.
Leatherback interactions also occur in Canadian pelagic longline fisheries. From summer to fall, primarily on the Scotian Shelf, encounters with leatherback turtles have been documented in the large pelagic longline fishery since 2001 (DFO 2012). With observer coverage ranging from 5 to 30 percent since 2001, there were 102 reported interactions with pelagic longlines from 2001 to 2005, and 36 from 2006 to 2010 (DFO 2012). Mortality rates are estimated to be in the range of 21 to 49 percent, resulting in an estimated mortality of 13 to 44 leatherback turtles annually. Based on an analysis of Canadian observer data from 2002 to 2010, the bycatch rate in this fishery is estimated to have declined from 120-190 leatherback turtles annually from 2002 to 2006 to 60-90 leatherback turtles annually from 2006 to 2010, largely as a result of gear modifications (Hanke
et al.
2012).
In the Mediterranean Sea, longlining is prevalent. Drifting longlines targeting swordfish (
Xiphias gladius
), albacore (
Thunnus alalunga
), and bluefin tuna (
T. thynnus
) are considered to be the most dangerous fishing gear for turtles in the Mediterranean Sea (Lucchetti and Sala 2010). Drifting longlines (mainly for albacore tuna) in Spain, Italy, Greece, and Albania have documented leatherback interactions (Camiñas 2004). In the western Mediterranean, swordfish longlines appeared to be responsible for most of the leatherback bycatch and entanglements (Camiñas 1998; Camiñas 2004). Casale
et al.
(2003) reviewed bycatch rates for longline fisheries targeting swordfish and estimated the average Mediterranean longline bycatch rates at 0.0025 leatherback turtles/1000 hooks, with a maximum rate of 0.0510 leatherback turtles/1000 hooks in the Tyrrhenian Sea of Italy (Casale
et al.
2003; Casale and Margaritoulis 2010). Of 170 leatherback fishery captures in fisheries from the Mediterranean Sea, approximately 35 involved longlines (Casale
et al.
2003). While leatherback turtles are encountered in Mediterranean longlines, loggerheads are the most common species caught; only 0.1 percent of turtles captured during an observer program in Spain, Italy and Greece were leatherback turtles (3 out of 2,370 observed turtles; Laurent
et al.
2001). However, given the extensive longline effort in the Mediterranean Sea (Casale 2008), leatherback bycatch in the Mediterranean is still a concern. Lewison
et al.
(2004) estimated a range of 250 to 10,000 leatherback turtles bycaught in the Mediterranean in 2000, with 6 percent observer coverage.
Longline bycatch of leatherback turtles in the range of the NW Atlantic DPS also occurs in waters off Cape Verde (Melo and Melo 2013; Coelho
et al.
2015), Morocco (Benhardouze
et al.
2012), and Brazil (Pacheco
et al.
2011). Given the wide distribution of both pelagic longline gear and leatherback turtles, bycatch of individuals in longline gear can occur wherever and whenever the gear and sea turtle distribution overlap.
Large circle hooks (non-offset) have been found to reduce leatherback bycatch by as much as 55 percent compared to traditional J-style hooks (Andraka
et al.
2013; Coelho
et al.
2015). While the vessels of certain nations may employ large circle hooks, there are no obligations for international longline fleets to adopt such bycatch mitigation measures (Richardson
et al.
2013). In 2005, an ICCAT resolution encouraged circle hook research (ICCAT 2005), but no legally binding measure to require circle hooks exists (Gilman 2011). Without the widespread use of non-offset circle hooks, it is likely that the high bycatch rates of leatherback turtles in pelagic longline gear will continue throughout the North Atlantic high seas fisheries.
Since 2004, the United States has issued regulations that require
modifications to pelagic longline gear in the U.S. Atlantic and GOM to reduce the bycatch and post-interaction mortality of sea turtles; these regulations (50 CFR 635.21(c)(2)) specify hook type and size (18/0 or 16/0 circle hooks depending on the area), bait type, use of turtle disentangling equipment and handling guidelines. Swimmer
et al.
(2017) recently analyzed pelagic longline interactions before (1992 to 2001) and after (mid-2004 to 2015) these regulations were promulgated. Throughout the study period, 844 leatherback turtles were captured. Overall, turtle bycatch was highest in the Northeast Distant statistical reporting area (0.3 turtles/1000 hooks), followed by the Northeast Coastal, GOM, and Caribbean areas. Bycatch rates were higher for years prior to 2004; after the regulations, Atlantic leatherback bycatch rates declined by 40 percent (0.13 to 0.078 turtles/1000 hooks). Within the Northeast Distant area alone, where additional restrictions include a large circle hook (18/0) and limited use of squid bait, rates declined by 64 percent (0.44 to 0.16 turtles/1000 hooks; Swimmer
et al.
2017). Gilman and Huang (2017) found similar results: Fish versus squid bait lowered catch rates of leatherback turtles, and wider circle hooks reduced leatherback catch rates relative to narrower J and tuna hooks. Capture probabilities are lowest when using a combination of circle hook and fish bait.
Efforts have been made to reduce interactions in Canadian waters as well. Circle hook use has been recommended in the swordfish-directed Canadian longline fleet since 2003, whereas corrodible circle hooks have been required in the pelagic longline fishery since 2012 (DFO 2013; C. MacDonald, DFO, pers. comm., 2019). There is no mandatory hook size restriction for the Canadian longline fleet, but license holders almost exclusively use 16/0 circle hooks (C. MacDonald, DFO, pers. comm., 2019). De-hooking and line-cutting kits are required on swordfish longline fishery vessels (C. MacDonald, DFO, pers. comm. 2019).
Some fishing fleets in the Atlantic Ocean (
e.g.,
U.S., Canadian, ICCAT vessels) use large circle hooks and modified bait, but these measures are not required in all areas (Watson
et al.
2005; Gilman
et al.
2007; Gilman 2011). Some nations in the Wider Caribbean Region have implemented circle hook provisions; in Belize, the high seas fishing fleet adopted the use of circle hooks on 10 percent of the fleet and are required to report capture of sea turtles by longlines (Belize Fisheries Department 2017). Because the measures are not widely required, the number of vessels that do not employ bycatch reduction measures is likely higher than the number of vessels that do, and so we conclude on the basis of the best available information that leatherback bycatch in pelagic longline fisheries is still a significant threat (Lewison
et al.
2015).
Leatherback interactions with bottom longlines also occur. Directed shark fisheries using bottom longlines in the Atlantic Ocean and GOM may capture or entangle leatherback turtles (NMFS 2012), and the GOM reef fishery is also anticipated to take leatherback turtles (NMFS 2011). On February 7, 2007, NMFS published a rule that required commercial shark bottom longline vessels to carry the same dehooking equipment as the pelagic longline vessels; this rule was promulgated to reduce post-interaction mortality (72 FR 5633).
The Canadian east coast groundfish longline fishery targets a wide variety of groundfish species, including cod, haddock, pollock and white hake. Observer coverage has ranged from 2 to 30 percent depending on area, and there have been no reported interactions of leatherback turtles in the observer database since 2001 (DFO 2012). However, there have been three reports from Quebec logbooks and 10 reports of interactions with groundfish longline gear to non-governmental groups (DFO 2012). This indicates that the risk of interactions in this gear may be higher than documented through the observer program.
Bottom longlines are also used in the Mediterranean Sea (Casale 2008). While there have not been any documented leatherback captures from this gear type, loggerheads have been caught at high rates in Tunisia, Libya, Greece, Turkey, Egypt, Morocco, and Italy (Casale 2008), and interactions with leatherback turtles are possible.
Commercial pelagic longline fisheries do not operate in some Caribbean nations, such as in Panama where effort is limited to vessels under six tons (Executive Decree 486, December 28, 2010). However, other Caribbean nations allow commercial pelagic longline fishing, and many find leatherback turtles with longline hooks (Réserve Naturelle de l'Amana data in Berzins, Office National de la Chasse et de la Faune Sauvage, pers. comm., 2018 and KWATA data in Berzins 2018). While no longlines exist in the Caribbean Dutch nations of Bonaire, St. Eustatius and Saba, there are efforts to introduce circle hooks into the trolling fishery (Ministry of Economic Affairs 2014). We consider longline bycatch to be a widespread threat to this DPS, likely resulting in the loss of thousands of individuals annually.
Trawl Fisheries
Leatherback turtles may interact with bottom and midwater trawl gear throughout the North Atlantic Ocean. The highest reported trawl bycatch of leatherback turtles of the NW Atlantic DPS is likely from the southeastern U.S. shrimp trawl fishery. Epperly
et al.
(2003) anticipated an average of 80 leatherback mortalities a year in shrimp trawl interactions, dropping to an estimate of 26 leatherback mortalities in 2009 due to reduction in fishing effort (Memo from Dr. B. Ponwith, SEFSC, to Dr. R. Crabtree, SERO, January 5, 2011). The 2014 NMFS Southeast U.S. Shrimp Fishery Biological Opinion estimated 167 annual leatherback captures (144 mortalities) in the Atlantic Ocean and GOM shrimp otter trawl fishery, with an additional 34 captures in try nets (single nets testing for shrimp concentrations; NMFS 2014). The majority of these interactions were in the GOM. However, a more recent study of the GOM and southeastern U.S. Atlantic coast shrimp otter trawl fishery found fewer leatherback captures: From 2007 to 2017, only 3 leatherback turtles were reported in the observer data (with coverage levels around 2 percent of nominal days at sea; Babcock
et al.
2018).
In the mid-Atlantic and northeastern U.S. waters, observers reported 9 leatherback captures in bottom otter trawl gear and 5 captures in midwater trawls from 1993 to 2017 (NMFS unpublished data 2018). In the Wider Caribbean Region, leatherback turtles are reported captured in trawls in French Guiana (Ferraroli
et al.
2004; TEWG 2007), Guyana (Reichart
et al.
2003), Suriname (Madarie 2010), Trinidad (Forestry Division
et al.
2010), and Venezuela (Alio
et al.
2010).
Since 1980, there were eight reports of leatherback turtles incidentally captured by trawl gear in British and Irish waters (Pierpoint 2000). In the Mediterranean Sea, leatherback bycatch in bottom trawls off Tunisia (Caminas 2004) and Egypt (Casale 2008) has also been reported.
Trawl bycatch reduction measures (
e.g.,
turtle excluder devices (TEDs) are in place in some nations. The southeastern U.S. shrimp fishery has required TEDs since the early 1990s. However, TEDs that were initially required for use in the U.S. Atlantic Ocean and GOM shrimp fisheries were less effective for leatherback turtles as compared to smaller, hard-shelled turtle species, because the TED openings were
too small to allow leatherback turtles to escape. To address this problem, NMFS issued a final rule on February 21, 2003, to amend the TED reg
This text is long and has been trimmed here. Open the source document for the complete record.
This is a copy of a public record, reproduced as it was published. It is not legal advice, and it may not be the version a court would rely on. Check the official source before you cite it.