Latin American Journal of Aquatic Research ISSN X

Transcription

1

2 Latin American Journal of Aquatic Research ISSN X CHIEF EDITOR Sergio Palma Pontificia Universidad Católica de Valparaíso, Chile CO-EDITOR Patricio Arana Pontifícia Universidad Católica de Valparaíso, Chile ASSOCIATE EDITORS Leonardo A. Abitia-Cárdenas Instituto Politécnico Nacional, La Paz, México Álvaro J. Almeida-Bicudo Universidade Federal do Paraná, Brasil Eduardo Ballester Universidade Federal do Paraná, Brasil Claudia S. Bremec Universidad Nacional de Mar del Plata, Argentina Sergio Contreras Universidad Católica de la Santísima Concepción, Chile Patricio Dantagnan Universidad Católica de Temuco, Chile José Gallardo Pontifícia Universidad Católica de Valparaíso, Chile Mariel Gullian Klanian Universidad Marista de Mérida México José Luis Iriarte Universidad Austral de Chile, Chile César Lodeiros-Seijo Universidad de Oriente, Venezuela Luis M. Pardo Universidad Austral de Chile, Chile Jesús T. Ponce-Palafox Universidad Autónoma de Nayarit, México Nelson Silva Pontificia Universidad Católica de Valparaíso, Chile Fernando Vega-Villasante Universidad de Guadalajara, México Cristian Aldea Universidad de Magallanes, Chile Antônio Olinto Ávila-da-Silva Instituto de Pesca, Brasil Sandra Bravo Universidad Austral de Chile, Chile Loretto Contreras Universidad Andrés Bello, Chile Amilcar Cupul Universidad de Guadalajara, Puerto Vallarta, México Enrique Dupré Universidad Católica del Norte, Chile Diego Giberto Instituto de Investigación y Desarrollo Pesquero, Argentina Crisantema Hernández Centro de Investigación en Alimentación y Desarrollo México Maurício Laterça-Martins Universidade Federal de Santa Catarina, Brasil Beatriz E. Modenutti Universidad Nacional del Comahue, Argentina Guido Plaza Pontificia Universidad Católica de Valparaíso, Chile Erich Rudolph Universidad de Los Lagos, Chile Jorge Urbán Ramírez Universidad Autónoma de Baja California del Sur, México Ingo Wehrtmann Universidad de Costa Rica, Costa Rica Escuela de Ciencias del Mar, Pontificia Universidad Católica de Valparaíso Valparaíso, Chile - lajar@pucv.cl

3 LATIN AMERICAN JOURNAL OF AQUATIC RESEARCH Lat. Am. J. Aquat. Res., 45(3) 2017 Preface CONTENTS Ximena Velez-Zuazo, Jeffrey C. Mangel, Jeffrey A. Seminoff, Bryan P. Wallace & Joanna Alfaro-Shigueto Filling the gaps in sea turtle research and conservation in the region where it began: Latin America Research Articles Ana R. Patrício, Ximena Vélez-Zuazo, Robert P. van Dam & Carlos E. Diez Genetic composition and origin of juvenile green turtles foraging at Culebra, Puerto Rico, as revealed by mtdna Michael J. Liles, Alexander R. Gaos, Allan D. Bolaños, Wilfredo A. Lopez, Randall Arauz, Velkiss Gadea, José Urteaga, Ingrid L. Yañez, Carlos M. Pacheco, Jeffrey A. Seminoff & Markus J. Peterson Survival on the rocks: high bycatch in lobster gillnet fisheries threatens hawksbill turtles on rocky reefs along the Eastern Pacific coast of Central America Rocío Álvarez-Varas, Juan Contardo, Maike Heidemeyer, Lina Forero-Rozo, Beatriz Brito, Valentina Cortés, María José Brain, Sofía Pereira & Juliana A. Vianna Ecology, health and genetic characterization of the southernmost green turtle (Chelonia mydas) aggregation in the Eastern Pacific: implications for local conservation strategies Tania Zuñiga-Marroquin & Alejandro Espinosa de los Monteros Genetic characterization of the Critically Endangered hawksbill turtle (Eretmochelys imbricata) from the Mexican Pacific region Pilar Santidrián-Tomillo, Nathan J. Robinson, Luis Gabriel Fonseca, Wagner Quirós-Pereira, Randall Arauz, Madeleine Beange, Rotney Piedra, Elizabeth Vélez, Frank V. Paladino, James R. Spotila & Bryan P. Wallace Secondary nesting beaches for leatherback turtles on the Pacific coast of Costa Rica Alexander R. Gaos, Michael J. Liles, Velkiss Gadea, Alejandro Peña de Niz, Felipe Vallejo, Cristina Miranda, Jodie Jessica Darquea, Ana Henriquez, Eduardo Altamirano, Alejandra Rivera, Sofía Chavarría, David Melero, José Urteaga, Carlos Mario Pacheco Didiher Chácon, Carolina LeMarie, Joanna Alfaro-Shigueto, Jeffrey C. Mangel, Ingrid L. Yañez & Jeffrey A. Seminoff Living on the Edge: Hawksbill turtle nesting and conservation along the Eastern Pacific Rim Astrid Jiménez, Sergio Pingo, Joanna Alfaro-Shigueto, Jeffrey C. Mangel & Yuri Hooker Feeding ecology of the green turtle Chelonia mydas in northern Peru Israel Llamas, Eric E. Flores, Marino E. Abrego, Jeffrey A. Seminoff, Catherine E. Hart Rodrigo Donadi, Bernardo Peña, Gerardo Alvarez, Wilfredo Poveda, Diego F. Amorocho & Alexander Gaos Distribution, size range and growth rates of hawksbill turtles at a major foraging ground in the eastern Pacific Ocean Sergio Pingo, Astrid Jiménez, Joanna Alfaro-Shigueto & Jeffrey C. Mangel Incidental capture of sea turtles in the artisanal gillnet fishery in Sechura Bay, northern Peru Short Communications Javier Quiñones, Sixto Quispe & Oscar Galindo Illegal capture and black market trade of sea turtles in Pisco, Peru: the never-ending story

4 Lat. Am. J. Aquat. Res., 45(3): , 2017 Sea Turtle Research and Conservation in Latin America Preface 501 Jeffrey Mangel, Jeffrey Seminoff, Bryan Wallace & Ximena Vélez-Suazo (Guest Editors) DOI: /vol45-issue3-fulltext-1 Preface Filling the gaps in sea turtle research and conservation in the region where it began: Latin America Ximena Velez-Zuazo 1, Jeffrey C. Mangel 2,3, Jeffrey A. Seminoff 4 Bryan P. Wallace 5,6 & Joanna Alfaro-Shigueto 2,3,7 1 Center for Conservation and Sustainability, Smithsonian Conservation Biology Institute National Zoological Park, Washington, DC, USA 2 ProDelphinus, Lima, Perú 3 Centre for Ecology and Conservation, University of Exeter, Penryn, Cornwall, United Kingdom 4 NOAA-National Marine Fisheries Service, Southwest Fisheries Science Center La Jolla, California, USA 5 Conservation Science Partners, Inc. Fort Collins, CO, USA 6 Nicholas School of the Environment, Duke University Marine Lab, Beaufort, NC, USA 7 Universidad Científica del Sur, Facultad de Biología Marina, Lima, Perú ABSTRACT. The first documented long-term sea turtle research and conservation project in the world was officially launched in Latin America (Tortuguero, Costa Rica) in Despite the enormous advances in research and conservation in the nearly seven decades since, many questions still remain unanswered about fundamental aspects of ecology and population dynamics that hinder the conservation of sea turtles in the region. To catalyze further dissemination of information and improvement of sea turtle conservation, this Special Issue presents 10 papers solely focused on studies conducted in Latin America. This Special Issue resulted from an initiative launched to celebrate the 36 th Annual Symposium on Sea Turtle Biology and Conservation, held in Peru in March the first time this event was held in South America. The articles featured present novel results for four of the five species of sea turtles present in this region, with data collected as far back as 1971 and as recent as The studies cover diverse subjects including the nesting ecology for the most endangered populations of sea turtles in the world -the Eastern Pacific hawksbill turtle (Eretmochelys imbricata) and leatherback turtle (Dermochelys coriacea); the origins and connectivity of nesting and foraging populations of hawksbills and green turtles (Chelonia mydas); the detection of a new foraging ground for hawksbills in the Eastern Pacific; and the pervasive occurrence of incidental capture as well as illegal retention of sea turtles. The recovery of these imperiled marine reptiles relies on information to design and implement sound conservation actions; in this regard, the papers in this Special Issue are making a vital contribution, following the initial efforts launched nearly 70 years ago. Keywords: sea turtle research, nesting ecology, population genetics, bycatch, illegal take. In the mid 1950 s, Dr. Archie Carr, a young professor from the University of Florida, visited the remote black sand beaches of Costa Rica s Caribbean coast at a site that would soon host the first long-term sea turtle study in Latin America and perhaps the most famous sea turtle nesting beach in the world, Playa Tortuguero. Here, Dr. Carr started a pivotal long-term study to investigate basic aspects of the biology and ecology of the nesting population of green sea turtles (Chelonia mydas) (Carr & Ogren, 1960), thus setting the stage for future efforts in Latin America. Now, almost 70 years later, and with more than 50 organizations working in the region, the field of sea turtle biology and conservation has matured tremendously. Indeed, Dr. Carr s original goal of protecting and recovering the imperiled populations of sea turtles of Latin America has continued thanks to the efforts of a cadre of passionate sea turtle researchers and conservationists throughout the region. All five sea turtle species present in Latin America are listed, either at the species or population-level in a threatened category by the IUCN Red list (www. iucnredlist.org, Table 1). The populations of the leatherback turtle (Dermochelys coriacea), the hawksbill turtle (Eretmochelys imbricata), and the loggerhead turtle (Caretta caretta) are considered

5 502 Latin American Journal of Aquatic Research Table 1. Overall and population-level classification of five sea turtle species by IUCN Red List of Endangered Species ( CR: critically endangered, EN: endangered, VU: vulnerable, LC: least concern. Red text indicates a classification in a threatened category. Latin America region 1 Species/Populations Overall East Pacific Southwest Atlantic Northwest Atlantic Dermochelys coriacea CR CR CR LC Chelonia mydas EN EN EN EN Caretta caretta VU CR 1 LC LC Eretmochelys imbricata CR CR CR CR Lepidochelys olivacea VU VU VU VU 1 This classification applies only to the south Pacific populations. Critically Endangered, while the populations of the green turtle and the olive ridley turtle (Lepidochelys olivacea) are considered Endangered and Vulnerable, respectively. The Eastern Pacific (EP) populations of leatherbacks and hawksbills are considered among the most endangered sea turtle populations in the world. Long-term studies have estimated a decline of more than 97% for the nesting population of the EP leatherback turtle, largely due to incidental capture and direct use of eggs (Wallace et al., 2013). While contemporary nesting populations of hawksbill turtles were recently rediscovered, they are estimated to reflect a small fraction of the historic population (Gaos et al., 2010). In contrast, there is encouraging evidence of recovery for other species. Many nesting populations of leatherback turtles in the Caribbean have experienced significant increases after years of conservation efforts and are considered as Least Concern by the Red List (Tiwari et al., 2013), while EP green turtles and olive ridleys are also increasing in abundance following historic depletions (Plotkin et al., 2012; Delgado-Trejo & Alvarado-Figueroa, 2012). The sea turtles of Latin America tell contrasting stories, but also astounding ones. For example, regardless of the habitat they occupy -either Caribbean coral reefs or mangrove ecosystems in the EPhawksbill turtles have a particular taste for sponges and in the EP they even devour mangrove shoots and seeds (Van Dam & Diez 1997; Carrión-Cortez et al., 2013; A. Gaos, unpublish. data). Separate populations of loggerhead turtles traverse the Pacific Ocean during their life history, tracking oceanic gyres between nesting and foraging grounds in the North and South Pacific Oceans (Bowen et al., 1995; Boyle et al., 2009). But, despite the tremendous discoveries made since the first studies began many decades ago in Tortuguero and other Caribbean locations, many fundamental questions about sea turtle biology, ecology, life history, and population dynamics remain unanswered. One reason for this is the paucity of peer-reviewed publications from Latin American scientists, a situation noted before (Nature Index, 2015) and that represents one of the main challenges to achieving a better understanding of sea turtles in Latin America and to effective conservation actions and policy. This disparity between the copious data being collected daily throughout the region and the data being published was one of the main motivations for launching this Special Issue Sea turtle research and conservation in Latin America. The 36 th Annual Symposium on Sea Turtle Biology and Conservation, held in Peru in March 2016, inspired this Special Issue. This was the first time the venue for this international event was in South America, and it attracted a remarkable number of researchers from throughout Latin American that arrived eager to share their most recent results. We conceived this Special Issue as a space to showcase the research conducted in the region. We invited manuscripts from research colleagues and groups engaged in long-term, multispecies, or multidisciplinary studies as these have demonstrated their utility to provide integrated perspectives and to aid management and conservation actions (e.g., Wallace et al., 2010; Raymond et al., 2015; Chin et al., 2017). We wanted the Special Issue to focus particularly on in-water studies and threats, quantification of threats by species and life stage; population connectivity at local and regional levels; and studies from understudied nesting beaches. The ten articles featured in this Special Issue comprise studies conducted in South and Central America and the Caribbean and present novel results covering as much as 45 years of data collection efforts. We identified three main subject groups covered by the manuscripts presented here. The first group includes assessments of long-term nesting data for the two most threatened populations of sea turtles in the world. Two studies addressed important information gaps of nesting abundance from unreported and understudied beaches. Gaos et al. (2017) analyzed 33 years of nesting records for the hawksbill turtle from nine

6 Preface 503 sites along the eastern Pacific. They reported a larger number of deposited nests, contrasting with previous estimates that were based upon more limited information (Gaos et al., 2010). Santidrian-Tomillo et al. (2017), on the other hand, looked towards secondary beaches for the eastern Pacific leatherback and analyzed nesting records from as far back as Secondary beaches are not usually considered when determining population trends, but for this species, these secondary beaches may represent higher-thanexpected abundance relative to the entire population, as well as potential sources for the recolonization of index beaches that are near extirpation. Both studies valued every inch of beach in their efforts to identify opportunities for the recovery of depleted rookeries. Foraging grounds are as important as nesting beaches for the conservation of sea turtles, if not more so. These areas, which can include neritic (e.g., Velez- Zuazo et al., 2014) and oceanic habitats (e.g., Briscoe et al., 2016), are pivotal for sea turtle development, recruitment, and population dynamics. The second group of manuscripts focused on these important areas, but approached their research questions from different angles. Two of the studies addressed open questions about the connections between feeding habitats and nesting beaches for the EP hawksbill and the Caribbean green turtle, respectively. Applying tools drawn from the field of molecular ecology, these studies provide further evidence of the large geographic areas sea turtles traverse during their ontogenic migrations and the regional perspective local initiatives should take when designing and implementing conservation actions. Zúñiga-Marroquín & Espinosa de los Monteros (2017) found that Eastern Pacific hawksbills establish local and trans-pacific connections during their life history, while Patrício et al. (2017) filled a gap for understanding the complex spatial connections Caribbean green sea turtles establish as a result of their migratory behavior. Two studies revealed new foraging habitats. Llamas et al. (2017) confirmed the existence of a very important resident foraging aggregation of EP hawksbills in the protected waters of the Coiba Island National Park, Panama. Alvarez el al. (2017) determined that the southernmost foraging aggregation of green turtles -found in the waters off the coast of Chilewere from the Galapagos nesting rookery. And Jimenez et al. (2017) added to our knowledge of what green turtles eat in a tropical foraging habitat, but also reported a high occurrence of plastics in their diet, adding to the warning signs about the pervasive problem of plastics pollution in our oceans. The third group of studies focused on several of the long-standing anthropogenic threats facing sea turtle: incidental catch (also known as bycatch) and illegal take (Wallace et al., 2011). Liles et al. (2017) investigated the impact of the lobster gillnet fishery in Central America and estimated one of the highest rates of gillnet bycatch and mortality for hawksbill turtles anywhere in the world. Likewise, Pingo et al. (2017) found high rates of bycatch of green sea turtles in gillnet fishery operations in northern Peru, with entanglements occurring in almost every observed fishing set. Bycatch may also lead to the illegal retention of the sea turtles, and according to Quiñones et al. (2017) this is not a rare event in the port of Pisco, in central Peru. In a period of six years, Quiñones et al. (2017) reported more than 900 sea turtle carapaces from four species with evident signs of slaughtering, most of which were observed at public dumpsites. Furthermore, the study detected an ongoing black market for sea turtle meat, despite the national regulations that prohibit its sale. The studies in this group serve as reminders that long-standing threats still jeopardize the recovery of sea turtle populations despite notable advances in sea turtle conservation. Seventy years of sea turtle research and conservation in Latin America have yielded enormous amounts of information that have helped set and achieve conservation milestones throughout the region. These range from starting conservation programs in almost every country in the region to influencing government policies to protect sea turtles throughout their life histories. Thanks to these efforts, some sea turtle populations are showing undeniable signs of recovery. However, as the studies presented in this Special Issue demonstrate, conservation work is still required to sustain positive trends and to reverse ongoing declines, and more discoveries await. To these ends, this Special Issue conveys several important messages: 1) that every beach counts for female sea turtles and that it is important to monitor and protect these areas to help the recovery of depleted nesting populations; 2) that sea turtle foraging grounds await discovery and probably require immediate protection; 3) that concerted regional efforts are essential to secure the connectivity of populations; and 4) that we must continue to work toward reducing bycatch and any form of intentional harvest of sea turtles. This Special Issue helps fill information gaps and heightens global awareness for sea turtle conservation, especially in Latin America, the region where sea turtle conservation was born. ACKNOWLEDGEMENTS We thank the authors, reviewers, and editorial staff of the Latin American Journal of Aquatic Research for making this Special Issue possible. Special thanks to the International Sea Turtle Society, the Whitley Fund

7 504 Latin American Journal of Aquatic Research for Nature, the University of Exeter, Universidad Científica del Sur, the DEFRA Darwin Initiative, and United States Fish and Wildlife Service for funding the publication of this issue. REFERENCES Alvarez-Varas, R., J. Contardo, M. Heidemeyer, L. Forero-Roso, B. Brito, V. Cortés, M.J. Brain, S. Pereira & J. Abreu-Vianna Ecology, health and genetic characterization of the southernmost green turtle (Chelonia mydas) aggregation in the Eastern Pacific: implications for local conservation strategies. Lat. Am. J. Aquat. Res., 45(3): Boyle, M.C., N.N. FitzSimmons, C. Limpus, S. Kelez, X. Velez-Zuazo & M. Waycott Evidence for transoceanic migrations by loggerhead sea turtles in the southern Pacific Ocean. Proc. R. Soc. Lond. B, 276: Bowen, B.W., F.A. Abreu-Grobois, G.H. Balazs, N. Kamezaki, C.J. Limpus & R.J. Ferl Trans- Pacific migrations of the loggerhead turtle (Caretta caretta) demonstrated with mitochondrial DNA markers. Proc. Natl. Acad. Sci., 92: Briscoe, D.K., M.D. Parker, S. Bograd, E. Hazen, K. Scales, G.H. Balazs, M. Kurita, T. Saito, H. Okamoto, M. Rice, J.J. Polovina & L.B. Crowder Multiyear tracking reveals extensive pelagic phase of juvenile loggerhead sea turtles in the North Pacific. Mov. Ecol., 4: 23. Carr, A. & L. Ogren The ecology and migrations of sea turtles 4. Bull. Am. Mus. Nat. Hist., 121: Carrión-Cortez, J., C. Canales-Cerro, R. Arauz & R. Riosmena-Rodríguez Habitat use and diet of juvenile eastern Pacific hawksbill turtles (Eretmochelys imbricata) in the North Pacific coast of Costa Rica. Chelonian Conserv. Biol., 12: Chin, A., C.A. Simpfendorfer, W.T. White, G.J. Johnson, R.B. McAuley & M.R. Heaped Crossing lines: a multidisciplinary framework for assessing connectivity of hammerhead sharks across jurisdictional boundaries. Sci. Rep., 7: doi: /srep Delgado-Trejo, C. & J. Alvarado-Diaz Current conservation status of the black sea turtle in Michoacán, México. In: J.A. Seminoff & B.P. Wallace (eds.). Sea turtles of the Eastern Pacific: advances in research and conservation. University of Arizona Press, Tucson, pp Gaos, A.R., F.A. Abreu-Grobois, J. Alfaro-Shigueto, D. Amorocho, R. Arauz, A. Baquero, R. Briseño, D. Chacón, C. Dueñas, C. Hasbún & M. Liles Signs of hope in the eastern Pacific: international collaboration reveals encouraging status for a severely depleted population of hawksbill turtles Eretmochelys imbricata. Oryx, 44: Gaos, A., M.J. Liles, V. Gadea, A. Peña de Niz, F. Vallejo, C. Miranda, J. Jessica Darquea, A. Henriquez, E. Altamirano, A. Rivera, S. Chavarría, D. Melero, J. Urteaga, C.M. Pacheco, D. Chácon, C. LeMarie, J. Alfaro-Shigueto, J.C. Mangel, I. Yañez & J.A. Seminoff Living on the Edge: hawksbill turtle nesting and conservation along the eastern Pacific rim. Lat. Am. J. Aquat. Res., 45(3): Jiménez, A., S. Pingo, J. Alfaro-Shigueto & J.C. Mangel Feeding ecology of the green turtle Chelonia mydas in northern Peru. Lat. Am. J. Aquat. Res., 45(3): Liles, M.J., A.R. Gaos, A.D. Bolaños, W.A. López, R. Arauz, V. Gadea, J. Urteaga, I. Yañez, C.M. Pacheco, J.A. Seminoff & M.J. Peterson Survival on the rocks: high bycatch in lobster gillnet fisheries threaten hawksbill turtles on rocky reefs along the Pacific coast of Central America. Lat. Am. J. Aquat. Res., 45(3): Llamas, I., E.E. Flores, M. Abrego, J.A. Seminoff, C.E. Hart, R. Donadi, B. Peña, G. Alvarez, W. Poveda, D.F. Amorocho & A. Gaos Distribution, size range and growth rates of hawksbill turtles at a major foraging ground in the eastern Pacific ocean. Lat. Am. J. Aquat. Res., 45(3): Nature Index Developing partnerships. Nature, 527(7577): S60-S63. doi: /527s60a. Patrício A.R., X. Velez-Zuazo, R.P. van Dam & C.E. Diez Genetic composition and origin of juvenile green turtles foraging at Culebra, Puerto Rico, as revealed by mtdna. Lat. Am. J. Aquat. Res., 45(3): Pingo, S., A. Jimenez, J. Alfaro-Shigueto & J.C. Mangel Incidental capture of sea turtles in the artisanal fishery in Sechura Bay, northern Peru. Lat. Am. J. Aquat. Res., 45(3): Plotkin, P.T., R. Briseño-Dueñas & F.A. Abreu-Grobois Interpreting signs of olive ridley recovery in the Eastern Pacific. In: J.A. Seminoff & B.P. Wallace (eds.). Sea turtles of the Eastern Pacific: advances in research and conservation. University of Arizona Press, Tucson, pp Quiñones, J., S. Quispe & O. Galindo Illegal capture and black market trade of sea turtles in Pisco, Peru: The never-ending story. Lat. Am. J. Aquat. Res., 45(4): Raymond, B., M.A. Lea, T. Patterson, V. Andrews Goff, R. Sharples, J.B. Charrassin, M. Cottin, L. Emmerson, N. Gales, R. Gales & S.D. Goldsworthy Important marine habitat off east Antarctica revealed by two decades of multi species predator tracking. Ecography, 38:

8 Preface 505 Santidrian-Tomillo, P., N.J. Robinson, L.G. Fonseca, W. Quirós-Pereira, R. Arauz, M. Beange, R. Piedra, E. Velez, F.V. Paladino, J.R. Spotila & B.P. Wallace Secondary nesting beaches for leatherback turtles on the Pacific coast of Costa Rica. Lat. Am. J. Aquat. Res., 45(3): Tiwari, M., B.P. Wallace & M. Girondot Dermochelys coriacea (Northwest Atlantic Ocean subpopulation). The IUCN Red List of Threatened Species 2013: e.t a [ doi.org/ /iucn.uk rlts.t A en.]. Reviewed: 15 June Van Dam, R.P. & C.E. Diez Diving behavior of immature hawksbill turtles (Eretmochelys imbricata) in a Caribbean reef habitat. Coral Reefs, 16: Velez-Zuazo, X., J. Quiñones, A.S. Pacheco, L. Klinge, E. Paredes, S. Quispe & S. Kelez Fast growing, healthy and resident green turtles (Chelonia mydas) at two neritic sites in the central and northern coast of Peru: implications for conservation. PloS ONE, 9(11): e Wallace, B.P., M. Tiwari & M. Girondot Dermochelys coriacea (East Pacific Ocean subpopulation). The IUCN Red List of Threatened Species 2013: e.t a [ 2305/IUCN.UK RLTS.T A en]. Reviewed: 15 June Wallace, B.P., A.D. DiMatteo, B.J. Hurley, E.M. Finkbeiner, A.B. Bolten, M.Y. Chaloupka, B.J. Hutchinson, F.A. Abreu-Grobois, D. Amorocho, K.A. Bjorndal & J. Bourjea Regional management units for marine turtles: a novel framework for prioritizing conservation and research across multiple scales. PLoS ONE, 5(12): e Wallace, B.P., A.D. DiMatteo, A.B. Bolten, M.Y. Chaloupka, B.J. Hutchinson, F.A. Abreu-Grobois, J.A. Mortimer, J.A., Seminoff, D. Amorocho, K.A. Bjorndal, J. Bourjea, B.W. Bowen, R. Briseño- Dueñas, P. Casale, B.C. Choudhury, A. Costa, P.H. Dutton, A. Fallabrino, E.M. Finkbeiner, A. Girard, M. Girondot, M. Hamann, B.J. Hurley, M. López- Mendilaharsu, M.A. Marcovaldi, J.A. Musick, R. Nel, N.J. Pilcher, S. Troëng, W. Witherington & R. Mast Global conservation priorities for marine turtles. PLoS ONE, 6(9): e Zuñiga-Marroquin, T. & A. Espinosa de los Monteros Genetic characterization of the Critically Endangered hawksbill turtle (Eretmochelys imbricata) from the Mexican Pacific region. Lat. Am. J. Aquat. Res., 45(3):

9 Lat. Am. J. Aquat. Res., 45(3): , 2017 Sea turtles illegal captures & black market in Pisco, Peru 615 Sea Turtle Research and Conservation in Latin America Jeffrey Mangel, Jeffrey Seminoff, Bryan Wallace & Ximena Vélez-Suazo (Guest Editors) DOI: /vol45-issue3-fulltext-11 Short Communication Illegal capture and black market trade of sea turtles in Pisco, Peru: the never-ending story Javier Quiñones 1, Sixto Quispe 1 & Oscar Galindo 1 1 Laboratorio Costero de Pisco, Instituto del Mar del Perú, Callao, Perú Corresponding author: Javier Quiñones (javierantonioquinones@gmail.com) ABSTRACT. The Pisco-San Andrés area (13 44 S, W) in central Peru is known for a traditional historic sea turtle fishery. To determine if illegal captures and black market trade exist, we carried out bi-weekly sampling in dumpsites and coastal areas from 2009 to A total of 953 carapaces were encountered, which included mainly black turtles (Chelonia mydas, 92.2%) and to a lesser extent, olive ridley turtles (Lepidochelys olivacea, 4.3%), leatherback turtles (Dermochelys coriacea, 1.4%), and a single hawksbill turtle (Eretmochelys imbricate, 0.1%). The mean curved carapace length (CCL) was 59.1 for black turtles, 60.4 for olive ridleys and 113 cm for leatherbacks. For all species, most of turtles reported were juveniles and came largely from illegal captures (89%) and not from stranding reports (1.4%). Mean mortality was 8.1 carcasses km -1 year -1 at beaches and carcasses year -1 at dumpsites. Main consumed prey items in black turtles were silverside fish eggs (47.9%), Chondracanthus seaweed (31.4%) and Paranthus sp., anemone (16.2%). Despite the big sampling effort, mortality estimates could be underestimated since big percentages are butchered and discarded at sea. Still, numbers remains high with almost 1000 turtles in a five-year period and an illegal trade persists. Urgent measures are needed to recover this endangered species. Keywords: Chelonia mydas, illegal captures, black market, Pisco, Perú. The sea turtle consumption in the Pisco, San Andrés area (13 72 S, W), ~130 km south of Lima has a long history. Records go back to the pre-hispanic era ( AC), when around 30 turtles were reported from an archaeological site around Paracas (13 83 S, W). During the Middle Ages, the Spanish Jesuit chronicler Bernabé Cobo reported the capture of ~90 turtles with a beach seine in Pisco (Frazier & Bonavia, 2000). Later, during the 1960s and 1970s, a report (Frazier, 1979) on the current state of the sea turtle fishery in Peru and the East Pacific, stated that the greatest numbers of turtles were captured in the port of Pisco, roughly 142 ton were landed, representing more than half the landings for the entire country from period. But it is during late 1970s when a sophisticated turtle trafficking network thrived in the Pisco area. The fishery was operated by 7 to 10 boats exclusively dedicated to catch sea turtles, using nets specially designed to catch turtles, and with a wellestablished trafficking structure. The average catch was between 10 to 30 turtles per vessel per day and the captures were mainly composed by green turtles and, to a lesser extent, leatherbacks (Frazier, 1979; Hays- Brown & Brown, 1982). During the 1987 El Niño up to 110 boats were recorded landing turtles in Pisco (Zeballos, pers. obs.), standing as the largest sea turtle harvest ever recorded in Peru. In just 10 months, over 20,000 turtles were landed (Aranda & Chandler, 1989). Sea turtle extraction was a legal activity until 1995 when a total ban on use of all the sea turtle species was established by Peruvian legislation (Morales & Vargas, 1996). However, the hunting still continues in Pisco by means of incidental and directed illegal captures as was preliminary evidenced (De Paz et al., 2002, 2004, 2007; Quiñones et al., 2010). In order to collect evidence of sea turtle illegal captures in Pisco, the regional laboratory of the Peruvian Marine Research Institute (IMARPE) conducted an assessment. From November 2009 until March 2015, bi-weekly visits to local dumpsites and local beaches in the surrounding area of Pisco and San Andrés were conducted and all encountered turtle remains were registered and geo-referenced. The carcasses, carapaces and stranded individuals were identified to species level and the curved carapace length (CCL) was measured and registered. For black sea turtles (Chelonia mydas agassizii), life stages were characterized as follows: juvenile (<69 cm CCL), sub-adult (>69 cm and <85 cm CCL), and adult turtles (>85 cm CCL), according to the minimum breeding size (CCL = 69 cm) and mean breeding size (CCL = 85 cm) reported for 1037 nesting females in the Galapagos rookery during

10 616 Latin American Journal of Aquatic Research (Zárate et al., 2013). For olive ridleys (Lepidochelys olivacea), individuals >57 cm CCL were considered adults, using the minimum size reported for nesting females in Nancite, Costa Rica (Marquez, 1990). For leatherback sea turtles (Dermochelys coriacea), individuals >123 cm CCL were considered adults, using the minimum size reported for nesting females in Parque Nacional Marino las Baulas in Costa Rica (Reina et al., 2002). In the case of hawksbill turtles, individuals >66 cm CCL were considered adults, based on the minimum size reported for nesting females at Estero Padre Ramos in Nicaragua (Altamirano et al., 2014). All carapaces were painted with red spray to avoid recounting. To identify the primary anthropogenic-derived threats for green turtles, the type of record obtained was used to determine the potential cause of death. If encountered at a dumpsite the turtle was considered to have been illegally captured, whereas if a turtle was found stranded at the beach the cause of death was classified either as boat collision, incidental entanglement in fisheries gear, or undetermined. Boat collisions were identified based on clear boat strike wounds or propeller injuries, whereas entanglements were considered when clear net derived injuries were present. We calculated annual mortality rates by dividing the number of new carcasses found at dumpsites (mean n carcasses year -1 ) and at the beach (mean n carcasses km -1 year -1 ) by the time elapsed between surveys, using the methodology of Senko et al. (2014). In addition, for each sea turtle reported, a decomposition stage was determined using a scale according to USGS standards (Work, 2000) which is: 1) Stranded dying animal; 2) fresh animal recently deceased; 3) animal moderately decomposed; 4) animal highly decomposed; and 5) skeletal or mummified animal. Necropsies were performed on fresh individuals in Category 2 and esophagus and stomachs were retrieved and immediately transported to the IMARPE regional laboratory in Pisco to analyze contents to determine diet composition. From stomach contents in black turtles, prey items were identified to the lowest possible taxon. Plant matter, mollusks, fish, crustaceans, jellyfish, and actinarian anemones were identified according to Dawson et al. (1964), Alamo & Valdivieso (1987), Chirichigno (1974), Retamal (1981), and Sanamyan et al. (2004). Quantitative assessment of diet was based on the relative wet weight (ww) in each sample, and was calculated as follows: %ww = (wet weight of a diet item/total weight of all items) 100. Finally, in order to determine the dimension of the black market and extent of illegal trade, we made several visits to the Pisco central market and conducted 70 structured qualitative interviews with local fishermen; main topics were by-catch and illegal trade. Between November 2009 and March 2015, five inland dumping places were reported, mainly in the central Pisco-San Andrés area. The majority of sea turtles (85%) were registered at the main dumping site (13 74 S, W) and surrounding areas and the remainder (12.7%) of turtles were reported along the shorelines of Caucato, Pisco, San Andrés and Paracas (between S and S). A total of 953 sea turtles were registered: 92.2% (n = 898) were black turtles, 4.3% (n = 41) were olive ridleys, 1.4% (n = 13) were leatherbacks, and 0.1% (n = 1) were hawksbills. The mean CCLs were 59.2 ± 9.5 cm (range: cm) for black turtles, 60.4 ± 6.2 cm CCL (range: cm) for olive ridleys, 113 ± 18.2 cm CCL (range: cm) for leatherbacks; the single juvenile hawksbill turtle had a CCL of 50.2 cm. Based on sizes observed, the life stage was determined: for black turtles, 83% were juveniles, 15.4% sub-adults and only 1.6% adults (Fig. 1), 38.5% of olive ridleys were juveniles, and 70% of leatherbacks were juveniles. The main identified cause of death of black turtles was illegal capture (89%, n = 899). For this species, stranding events were represented by only 1.4% (n = 13), and within this 0.9% were not determined (n = 8), 0.3% were due to collisions (n = 3), and 0.2% due to entanglements (n = 2) (Fig. 1). For olive ridleys, 97.5% were illegal captures and only 2.5% were standings where the cause of death was not determined. For leatherbacks, 85% were illegal captures and 15% stranded. Of the stranded leatherbacks, half showed evidence of collision and the rest was not determined. Our mean black turtle mortality estimates were 8.1 carcasses km -1 year -1 at beaches and carcasses year -1 at dumpsites (Fig. 2). According to the decomposition scale used, no dying animal was reported, 24.7% (n = 177) were classified as freshly deceased, 50.8% (n = 364) were moderately decomposed, 22.8% (n = 163) highly decomposed, and 1.7% (n = 12) skeletal or mummified. Five esophageal and stomach content analyses were performed during october The most consumed item, expressed in percentage of wet weight (% ww), was silverside fish (Odontesthes regia regia) eggs (47.9% ww), followed by the seaweed yuyo (Chondracanthus chamissoi, 31.4% ww) and the actinarian anemone Paranthus sp. (16.2% ww). Crustaceans, polychaetes, fish, and green seaweed remains were registered in lesser amounts (Table 1). The existence of illegal trade was demonstrated. Recovered evidence at the Pisco Central fish-market observations during summer and autumn months indicates there were five ladies offering sea turtle meet

11 N of Green Turtles Sea turtles illegal captures & black market in Pisco, Peru Adult Min 69 Adult Mean Curved Carapace Length (CCL) Causes of death in Chelonia mydas (98.6%) (1.4%) ilegal capture strandings No determined collisions entaglements 8 (0.9%) 3 (0.3%) 2 (0.2%) Figure 1. a) Size structure and causes of death of black turtles (Chelonia mydas agassizii), based in information from dumpsites and strandings in the Pisco area for the period , b) pie chart of main causes of death in black turtles, numbers represent the number of cases with their percentages. Table 1. Prey wet weights (g) in both esophageal and stomach samples (n = 5) of black turtles (Chelonia mydas) and percentage of wet weight (%ww) in the Pisco area, during October Prey item Total %ww Animal matter Silverside eggs (fish) Paranthus sp. (anemone) Silverside (fish) Squilla stillirostris (crustacean) Diopatra (Polychaete) Crustacean remains Other Polychaetes Vegetal matter Chondracanthus chamissoi Enteromorpha oficinalis Ulva papenfusi Caulerpa filiformis Mineral matter Mud for sale. Each one had approximately two turtles for sale and the price per kg varied from US$ 7 to 10. Based on our observations, we infer that between 10 to 20 turtles, mainly black turtles, are sold per week during summer and autumn months, whereas during winter and early spring season the amounts of turtle for sale were less and some weeks were totally absent from the market. As a result of the interviews, the illegal turtle trade was divided into three main activities: i) Local consumption, subdivided in family consumption by fishermen and subsistence consumption on board fishing vessels by fishermen; ii) market and other types of commercialization, subdivided in turtle meat and fins sold in the Pisco market as well as door to door sales of turtle meat in the fishing town of San Andrés; and iii) demand from other towns, subdivided in demands from Lima, the capital city, and demands from Pisco nearby towns like Chincha, Cañete, and Ica. In addition, as a result of combining answers from interviews and night observations, two fishermen, pretending to catch silverside fish at Paracas, were observed landing turtles hidden inside large plastic bags. These fishermen had artisanal gillnets with a mesh size of 60 cm (the traditional measure for turtle nets). The majority of the prey items consumed by the turtles captured were animal matter, mainly represented by silverside fish eggs and anemones (Paranthus sp.) during spring, whereas during summer months the diet was mainly composed of jellyfish. There was a clear seasonality in prey consumption, with jellyfish Chrysaora

12 618 Latin American Journal of Aquatic Research Figure 2. Pictures of illegal captures of sea turtles in the Pisco area. Clockwise from the first picture top left: leatherback captured for human consumption at the beach; and the three remaining ones were black turtles in the main dumpsite of Pisco. plocamia consumption occurring mainly during the summer-autumn seasons and sea anemone Paranthus sp. consumption during the winter-spring seasons (Paredes, 2015). The human activity that has the largest impact on sea turtles is fisheries bycatch (Lewinson et al., 2004, 2013; Wallace et al., 2011), even worst, direct harvest remains a major threat to sea turtles population worldwide (Campbell, 2003; Lewison et al., 2013), as was previously stated a legal sea turtle fishery existed in Pisco until mid-90s, after that illegal poaching continues in the country, however sea turtles poaching does not occurred exclusively in Peru, in several countries in the Americas this illegal activity still occurs nowadays, like in the Mexican Pacific (Mancini & Koch, 2009; Mancini et al., 2011; Senko et al., 2014), the wide Caribbean area (Bell et al., 2006; Richardson et al., 2009; Lagueux et al., 2014) and Brazil (Geubert et al., 2013), in addition, this illegal activity occurs elsewhere worldwide, like in Africa (Riskas & Tiwari, 2013); Southeast Asia (Joseph et al., 2014); Oceania (Maison et al., 2010) among other places. Direct extraction, together with by-catch interaction with small scale gillnets fisheries (SSGF), is a strong factor of mortality in Peru mostly due to economic

13 Sea turtles illegal captures & black market in Pisco, Peru 619 needs, lack of law enforcement and strong traditional consumption, as was reported in BC, Mexico (Mancini et al., 2011). The SSGF usually operates in nations where there are few protective measures and limited enforcement capabilities (Chuenpagdee et al., 2006; Dutton & Squires, 2008). Peru is a typical example of this situation, were an estimated of ~6000 turtles are extracted annually by SSGF (Alfaro-Shigueto et al., 2011), particularly in the Pisco area, where still nowadays fishermen are manufacturing redes tortugueras a traditional gill net for direct turtle harvest. In addition, retaining bycatch is a common practice in the area. Some artisanal fishermen go at sea with a wide array of legal nets; however, redes tortugeras are also taken and placed in hidden places on board, in order to use it just in case the target capture is not enough. Despite the high sampling effort, the mortality estimates presented here are underestimated for at least two reasons. First, poachers are killing turtles at sea and throwing the carcasses into the water where they go unnoticed, similar to reports from BC, Mexico (Mancini & Koch, 2009). Second, the remains could be burned between the sampling days, thereby not allowing us to find the carcasses. Yet despite this, almost 1000 carapaces were encountered in the Pisco area. The high mortality rates of 160 turtles year -1 in dumpsites and 8.1 km -1 year -1 at beaches are the consequence of a strong sampling effort (two times per week), a good spatial coverage (almost 100% of dumpsites in Pisco), and sea turtle aggregations in Pisco restricted to a small geographic area of less than m 2 and ~200 km of coastline (Velez-Zuazo et al., 2014; Quiñones et al., 2017). Is suggested that mortality rates in Pisco are likely underestimated considering that this area overlaps and interacts with a strong SSGF. There are an average of 200 SSGF boats operating in San Andrés and 100 boats in Tambo de Mora. The majority of the turtles illegally taken in Peru were small individuals, recruited from oceanic epipelagic areas to neritic areas (Luschi et al., 2003) like in the shallow waters of Paracas Bay, indeed, our black turtle size structure reflected a predominance of juveniles (over 80%), however if we compared it with information from 1987, a size decrease of ~9 cm of CCL was noticed, in marine resources fisheries population assessments revealed that age and sexual maturation (ASM) could decrease as a consequence of fishing pressure (Olsen et al., 2004; Swain et al., 2007), in the case of sea turtles, a synergy of different issues like: fishing pressure, physical and biological characteristics of the environment and density dependent factors could influence the ASM and the growth rates of sea turtles (Avens & Snover, 2013). Regarding the black market trade, the situation seems to be declining since the total ban in 1996 (Morales & Vargas, 1996). Nevertheless, we found that there are still local/regional markets that supply sea turtle meat and derived products. Though a quite strong regional trade persist in the area and this activity represents an extra income to fishermen in Pisco, where a turtle of average weight (30 kg) can be sold up to US$60 demands of some turtle derived products persist, for instance, oil and turtle blood were thought to be effective remedies against flu and other diseases in the past, even nowadays the turtle oil is still demanded by local people, each bottle of 500 ml is sold at ~US$10, instead turtle blood ingesting is no longer practiced. The trafficking structure fluctuates seasonally with a well-structured network during summer and autumn seasons and a weak and unstructured network during winter and spring seasons, however the current demand is much lower than in the former years. All these illegal activities are prone to continue unless localized law enforcement and political measures are strongly implemented in the area. Even if our data is only a fraction of the whole illegal captures and trade in the country, it provides a valuable understanding of the current situation in the Pisco area. The presented results could be useful for further management and recommendation actions. REFERENCES Alamo, V. & V. Valdivieso Lista sistemática de moluscos marinos del Perú. Boletín Instituto del Mar del Perú & Deutsche Gesellschaft fuf Technische Zusammenarbeit, Volumen Extraordinario, Callao, Perú: Alfaro Shigueto, J., J.C. Mangel, F. Bernedo, P.H. Dutton, J.A. Seminoff & B.J. Godley Small scale fisheries of Peru: a major sink for marine turtles in the Pacific. J. Appl. Ecol., 48: Altamirano, E., V. Gadea, A. Gaos & M. Liles Protocolo de monitoreo, protección de nidos y manejo de vivero de tortuga carey (Eretmochelys imbricata) en la RN Estero Padre Ramos y Aserradores, Temporada Fauna y Flora Internacional e ICAPO, 34 pp. Aranda, C.A. & M.W. Chandler Las tortugas marinas del Perú y su situación actual. Bol. Lima, 11: Avens, L. & M.L. Snover Age and age estimation in sea turtles. In: J. Wyneken, J. Lohmann & J.A. Musick (eds.). The biology of sea turtles. CRC Press, Boca Ratón, pp

14 620 Latin American Journal of Aquatic Research Bell, C.D., J.M. Blumenthal, T.J. Austin, J.L. Solomon, G. Ebanks-Petrie, A. Broderick & B.J. Godley Traditional Caymanian fishery may impede local marine turtle population recovery. Endanger. Species. Res, 2: Campbell, L.M Contemporary culture, use, and conservation of sea turtles. In: P.L. Lutz, J.A. Musick & J. Wyneken (eds.). Biology of sea turtles. CRC, Boca Raton, pp Chirichigno, N Clave para identificar los peces marinos del Perú. Informe Instituto del Mar del Perú, Callao, 44: Chuenpagdee, R., L. Liguori, M.L.D. Palomares & D. Pauly Bottom up, global estimates of smallscale fisheries catches. Fisheries Centre Research Report, British Columbia University, 14: 105 pp. Dawson, E.Y., C. Acleto & N. Foldvik The seaweeds of Perú. Nova Hedwigia, 13: De Paz, N., J.C. Reyes & M. Echegaray Datos sobre captura, comercio y biología de tortugas marinas en el área de Pisco-Paracas. Memorias I Jornada Científica Reserva Nacional de Paracas, Universidad Nacional Agraria La Molina, Lima, pp De Paz, N., J.C. Reyes & M. Echegaray Capture and trade of marine turtles at San Andrés, Southern Peru. In M.S. Coyne & R.D. Clark (eds.). Proceedings of the Twenty-First Annual Symposium on Sea Turtle Biology and Conservation. NOAA Technical Memorandum NMFS SEFSC, 528: De Paz, N., J.C. Reyes, M. Echegaray, M. Ormeño & H. Anchante Identificación y manejo de hábitats críticos de tortugas marinas en Perú: Paracas, estudio de caso. In: C. Guerra-Correa, A. Fallabrino, P. Bolados-Díaz & C. Turner (eds.). Estado actual y perspectivas de la investigación y conservación de las tortugas marinas en las costas del Pacífico Sur- Oriental. Antofagasta, 28 pp. Dutton, P.H. & D. Squires Reconciling biodiversity with fishing: a holistic strategy for Pacific sea turtle recovery. Ocean Develop. Int. Law, 39: Frazier, J Marine turtles in Peru and the east Pacific. Office of Zoological Research, National Zoological Park, Smithsonian Institution (unpublished manuscript). Frazier, J. & D. Bonavia Prehispanic marine turtles in Peru: where were they? In: F.A. Abreu-Grobois, R. Briceño-Dueñas, R. Márquez & L. Sarti (eds.). Proceedings of the Eighteenth International Sea Turtle Symposium. U.S. Department Commerce, NOAA Tech. Memo NMFS-SEFSC-436: Geubert, F.M., M. Barletta & M.F. da Costa Threats to sea turtle populations in the Western Atlantic: poaching and mortality in small-scale fishery gears. J. Coast. Res., 65: Hays-Brown, C. & W.M. Brown Status of sea turtles in the southeastern Pacific: emphasis on Peru. In: K.A. Bjorndal (ed.). Biology and conservation of sea turtles. Smithsonian Institution Press, Washington DC, pp Joseph, J., C.Y. Kuen, P.M. Palaniappan & L.H. Chark Genetic investigation of green turtles (Chelonia mydas) harvested from a foraging ground at Mantanani, Sabah, Malaysia. Herpetol. Conserv. Biol., 3: Lagueux, C.J., C.L. Campbell & S. Strindberg Artisanal green turtle, Chelonia mydas, fishery of Caribbean Nicaragua: I. Catch rates and trends, PLoS ONE, 9: e Lewison, R.L., L.B. Crowder, A.J. Read & S.A. Freeman Understanding impacts of fisheries bycatch on marine megafauna. Trends Ecol. Evol., 19: Lewison, R.L., B. Wallace, J. Alfaro-Shigueto, J.C. Mangel, S.M. Maxwell & E. Hazen Fisheries bycatch of marine turtles, lessons learned from decades of research and conservation. In: J. Wyneken, K.J. Lohmann & J.A. Musick (eds.). The biology of sea turtles. Vol. 3. CRC, Boca Raton, pp Luschi, P., G.C. Hays & F. Papi A review of longdistance movements by marine turtles, and the possible role of ocean currents. Oikos, 103: Maison, K.A., I.K. Kelly & K.P. Frutchey Green turtle nesting sites and sea turtle legislation throughout Oceania. US Department of Commerce, National Oceanic and Atmospheric Administration, National Marine Fisheries Service. NOAA Technical Memorandum NMFS-F/SPO-110. Mancini, A. & V. Koch Sea turtle consumption and black market trade in Baja California Sur, Mexico. Endanger. Species Res., 7: Mancini, A., J. Senko, R. Borquez-Reyes, J.G. Póo, J.A. Seminoff & V. Koch To poach or not to poach an endangered species: elucidating the economic and social drivers behind illegal sea turtle hunting in Baja California Sur, Mexico. Hum. Ecol., 39: Márquez, R Sea turtles of the world. An annotated and illustrated catalogue of sea turtle species known to date. FAO Fish. Synop. No. 125, Vol. 11, Rome, 81 pp. Morales, V.R. & P. Vargas Legislation protecting marine turtles in Peru. Mar. Turtle Newslett., 75: Olsen, E.M., H. Knutsen, J. Gjøsæter, P.E. Jorde, J.A. Knutsen & N. Stenseth Life history variation among local populations of Atlantic cod from the Norwegian Skagerrak coast. J. Fish. Biol., 64:

15 Sea turtles illegal captures & black market in Pisco, Peru 621 Paredes, E Hábitos alimentarios de la tortuga verde del Pacífico este Chelonia mydas agassizii (Boucort, 1868) en la bahía de Paracas, Ica, Perú, durante el año Tesis Universidad Nacional Mayor de San Marcos, San Marcos, 61 pp. Quiñones, J., S. Quispe, E. Paredes Black turtle population dynamics in Paracas, Peru during a six year in water monitoring. Proceedings of the Thirty Six Annual Symposium on Sea Turtle Biology and Conservation. Lima, Perú, March (in press). Quiñones, J., V.G. Carman, J. Zeballos, S. Purca & H. Mianzan Effects of El Niño-driven environmental variability on black turtle migration to Peruvian foraging grounds. Hydrobiologia, 645: Reina, R.D., P.A. Mayor, J.R. Spotila, R. Piedra & F.V. Paladino Nesting ecology of the leatherback turtle, Dermochelys coriacea, at Parque Nacional Marino Las Baulas, Costa Rica: to Copeia, 2002: Retamal, M.A Catálogo ilustrado de los crustáceos decápodos de Chile. Universidad de Concepción, Gayana, 44: Richardson, P.B., M.W. Bruford, M.C. Calosso, L.M. Campbell, W. Clerveaux, A. Formia, B.J. Godley, A.C. Henderson, K. McClellan, S. Newman, K. Parsons, M. Pepper, S. Ranger, J. Silver, L. Slade & A. Broderick Marine turtles in the Turks and Caicos Islands: remnant rookeries, regionally significant foraging stocks, and a major turtle fishery. Chelonian Conserv. Biol., 8: Received: 30 August 2016; Accepted: 13 January 2017 Riskas, K.A. & M. Tiwari An overview of fisheries and sea turtle bycatch along the Atlantic coast of Africa. Munibe, 1: Sanamyan, N., K. Sanamyan & D. Schories Actiniaria.com. Available in [ Reviewed: 18 January Senko, J., A. Mancini, J.A. Seminoff & V. Koch By-catch and directed harvest drive high green turtle mortality at Baja California Sur, Mexico. Biol. Conserv., 169: Swain, D.P., S.F. Sinclair & J.M. Hanson Evolutionary response to size-selective mortality in an exploited fish population. Proc. R. Soc. Lond. B, 274: Velez-Zuazo, X., J. Quiñones, A.S. Pacheco, L. Klinge, E. Paredes, S. Quispe & S. Kelez Fast growing, healthy and resident green turtles (Chelonia mydas) at two neritic sites in the central and northern coast of Peru: implications for conservation. PLoS ONE, 9: e Wallace, B.P., A.D. DiMatteo, A.B. Bolten, M.Y. Chaloupka, B.J. Hutchinson, F.A. Abreu-Grobois, J.A. Mortimer et al Global conservation priorities for marine turtles. PLoS ONE, 6:e Work, T.M Sea turtle necropsy manual for biologists in remote refuges. National Wildlife Health Center, Hawaii Field Station, 25 pp. Zárate, P., K.A. Bjorndal, M. Parra, P.H. Dutton, J.A. Seminoff & A.B. Bolten Hatching and emergence success in green turtle Chelonia mydas nests in the Galápagos Islands. Aquat. Biol., 19:

16 Lat. Am. J. Aquat. Res., 45(3): , 2017 Incidental capture of sea turtles in Sechura Bay, Peru 6061 Sea Turtle Research and Conservation in Latin America Jeffrey Mangel, Jeffrey Seminoff, Bryan Wallace & Ximena Vélez-Suazo (Guest Editors) DOI: /vol45-issue3-fulltext-10 Research Article Incidental capture of sea turtles in the artisanal gillnet fishery in Sechura Bay, northern Peru Sergio Pingo 1, Astrid Jiménez 1, Joanna Alfaro-Shigueto 1,2,3 & Jeffrey C. Mangel 1,2 1 ProDelphinus, Lima, Perú 2 Centre for Ecology and Conservation, University of Exeter, Penryn, United Kington 3 Facultad de Biología Marina, Universidad Científica del Sur, Lima, Perú Corresponding autor: Sergio Pingo (sergio@prodelphinus.org) ABSTRACT. Gillnets are recognized globally as one of the fishing gears with the highest levels of bycatch and mortality of sea turtles. Through onboard observer monitoring from July 2013 to June 2014 we assessed the bycatch of sea turtles by an artisanal gillnet fishery operating from Sechura Bay, Peru. One hundred and four sea turtles were incidentally caught in 53 observed fishing sets. The observed species composition of bycatch was green turtle Chelonia mydas (n = 100), hawksbill Eretmochelys imbricata (n = 3) and olive ridley Lepidochelys olivacea (n = 1). Bycatch occurred in 62.3% of monitored sets, with an average of 1.96 turtles caught per set. For all sea turtles combined, 28.8% of individuals were dead and 71.2% were alive at the time of retrieval. The majority of individuals caught were classified as juveniles and sub-adults, with an average carapace length (CCL) of 57.3 ± 0.9 cm for green turtles and 40.2 ± 2.4 cm for hawksbills. The mean annual catch per unit effort (CPUE) of sea turtles was 1.11 ± 0.31 turtles km h -1 ), but varied by seasons. These results suggest that Sechura Bay is an important developmental habitat for juvenile and sub-adult green turtles and hawksbill turtles, but one subject to intense fishing interaction pressure. The development of monitoring programs, local awareness-raising activities, and enhanced management and protection of this critical foraging area and developmental habitat is recommended. Keywords: sea turtles, CPUE, gillnet, bycatch, Sechura Bay, Peru. INTRODUCTION Five species of sea turtles are known to occur in the Peruvian waters, the olive ridley (Lepidochelys olivacea), green (Chelonia mydas), hawksbill (Eretmochelys imbricata), leatherback (Dermochelys coriacea) and loggerhead (Caretta caretta) (Hays-Brown & Brown, 1982; Eckert & Sarti, 1997; Alfaro-Shigueto et al., 2004, 2010a; López-Mendilaharsu et al., 2006; Castro, et al., 2012). Research suggests that the Peruvian waters are primarily used as a foraging habitat (Hays- Brown & Brown, 1982; Alfaro-Shigueto et al., 2002; Santillan, 2008), although recent studies have confirmed the presence of green and olive ridley turtles nesting along Peru s highly developed northern coastline, making Peruvian coast the southernmost sea turtle nesting habitat in the eastern Pacific (Kelez et al., 2009; Velez-Zuazo et al., 2014, SWOT, 2015). In recent years, it has become apparent that vessels from small-scale fisheries (SSF) using trawls (Lewison et al., 2004), gillnets (Murray, 2009), seine nets, pound nets (Gilman et al., 2010), longlines (Casale, 2008, 2010; Alfaro-Shigueto et al., 2011), and many other gears types all incur in sea turtle bycatch (Moore et al., 2010). Fisheries bycatch has been identified as an important factor in many population declines, included of sea turtles. These populations can decline over short timescales, often without detection (Lewison et al., 2004). This situation poses a serious threat to many sea turtle populations and their conservation efforts (Lewison et al., 2004; Alfaro-Shigueto et al., 2008; Dutton & Squires, 2008; Koch et al., 2013). Within the Peruvian fisheries sector, SSF are particularly important because of their role in food security, but also as a source of employment (Mangel et al., 2010; FAO, 2010; Alfaro-Shigueto et al., 2011). Operating along the entire Peruvian coastline, the gillnet fishery comprise the largest component of Peru s small-scale fleet and it is conservatively estimated to set km of net per year (Alfaro-Shigueto et al., 2010b). Recent studies show that gillnet fisheries in Peru have high interaction rates with sea turtles and exert significant pressure on sea turtle populations throughout the Pacific (Wallace et al., 2010; Alfaro- Shigueto et al., 2011; Lewison et al., 2014). The frequency of interactions depends on spatiotemporal

17 2607 Latin American Journal of Aquatic Research overlap between critical habitat for a given species and fishing activities, encompassing a wide range of fishing methods and gear characteristics (Wallace et al., 2008, 2010). The purpose of the present research was to evaluate the incidental capture of sea turtles in the artisanal gillnet fishery in Sechura Bay, northern Peru, considering this bay is an important area for development of small scale fishery, but also an important foraging area of juvenile sea turtles. MATERIALS AND METHODS Study area and data collection Sechura Bay is located on the northern coast of Peru in Sechura Province, Piura Department ( S and W) (Fig. 1). Is the largest bay in Peru and an important and traditional zone of artisanal fishing and mariculture (GORE - Piura, 2012; Morón et al., 2013). The study was conducted in Sechura Bay from July 2013 to June Data was collected by trained onboard observers as part of a program to monitor the small-scale bottom set gillnet fleet operating from Constante port (5 35 S, W). Fishing boats ranged in length from 6 to 10 m and each trip consisted of setting of bottom set gillnets. Bottom set gillnets were made of multifilament twine and were composed of multiple net panes that measured 56.4 m long by 2.8 m high, with a stretched mesh of approximately 24 cm (Alfaro-Shigueto et al., 2010b; Ortiz et al., 2016). Typical to this fishery, nets were deployed in the late afternoon, soaked overnight and retrieved the following morning. The soak time ranged from 12 to 24 h (López-Barrera et al., 2012; Ortiz et al., 2016). The target species in this fishery are flounder Paralichthys spp., guitarfish Rhinobatos planiceps and other species of ray from the Batoidea superorder as common stingray Dasyatis spp. and round ray Urotrygon spp. (Tume et al., 2012; Ortiz et al., 2016). Onboard observers recorded specific data about the fishery operation, including information on gear characteristics (e.g., net size and number of panes, number of sets), environmental data for each set (e.g., location, time of set and haul, sea surface temperature, water depth, and water visibility), and information on each sea turtle bycatch event. Incidental capture, morphometric data and sea turtle handling Incidentally captured sea turtles were brought onboard the boat for handling. We proceeded to untangle each individual and assessed its basic condition (alive, inactive/drowned or dead). Those individuals recorded as inactive/drowned, were rehabilitated following the handling and resuscitation techniques described on the NOAA Southeast Fisheries Science Center website for onboard observers ( Information collected for each turtle included species identification, the geographical position (latitude and longitude) of capture, capture condition and final fate (released alive or discarded dead), and curved carapace length (CCL; measured from the nuchal notch to posterior-most tip) (Bolten, 2000). Measurements were made using a metric tape (±0.1 cm). Sea turtles determined to be in good condition were tagged with Inconel tags applied to the trailing edge of both front flippers and were released. Dead turtles were measured and then discarded at sea. For all sea turtle individuals, skin sample were taken for further studies. Individuals of C. mydas, with a CCL 69 cm. were considered as juveniles, individuals with 69 CCL <85 cm. were considered as sub-adults, and individuals with a CCL 85 cm. were categorized as adults (Zarate et al., 2013). Individuals of E. imbricata, with a CCL 74 cm were considered as juveniles, individuals with 74 CCL <81.6 cm. were considered as sub-adult, and individuals with a CCL 81.6 cm were categorized as adults (Liles et al., 2011). Finally, individuals of L. olivacea, with a CCL 59.2 cm were considered as juveniles, individuals with 59.2 CCL < 64.9 cm. were considered as sub-adult and individuals with a CCL 64.9 cm were considered as adults (Barrientos-Muñoz et al., 2014). Data analysis Sea turtle bycatch per unit effort (CPUE) was determined as: CPUE = number of turtles captured / (net length [km]) (soak time of net [12 h]) (Wang et al., 2013). Gillnet bycatch data for the study was grouped by month in order to derive monthly stratified CPUE estimates. These data were calculated in terms of catch set -1 (Mangel et al., 2010). However, to facilitate comparison with other studies, catch per km h -1 was also calculated. Descriptive statistics are presented as mean ± standard deviation (SD). The annual bycatch rate in Constante port was also calculated, according to Alfaro-Shigueto et al. (2011) applying their same estimates of fleet size and fishing effort (8 fishing vessels, 30 sets per month), the best available estimates this fishing fleet s size and effort for this port. Maps of fishing effort and turtle captures were prepared using MAPTOOL (Seaturtle.org, V. 2002, available at

18 Incidental capture of sea turtles in Sechura Bay, Peru 6083 Figure 1. Location of gillnet sets in Sechura Bay, Peru. Sets without bycatch of sea turtles ( ); sets with bycatch ( ) (Seaturtle.org Maptool, V. 2016). RESULTS Fifty-three fishing sets were monitored (Fig. 1), 15 on winter (April to June), 10 on spring (October to December), 14 on summer (January to March) and 14 on autumn (July to September). Nets averaged 1.12 ± 0.02 km in length (range = km) and ± 0.45 h of soak time (range = h). Sea turtle bycatch totaled 104 individuals. Onehundred individuals were C. mydas (96.2%), three individuals were E. imbricata (2.9%) and one individual was a L. olivacea (0.9%) (Table 1). Bycatch occurred in 62.3% of monitored sets (Fig. 1) with an overall bycatch rate of 1.96 ± 0.44 turtles set -1 (range = 0-16 turtles set -1 ). The number of turtles caught varied by season. The largest number of captures occurred during winter (n = 39), followed for autumn (n = 29), spring (n = 21), and summer (n = 15) (Fig. 2). The month with the highest number of caught turtles was July (n = 34) while the months with the lowest number of captures were December and June (n = 1, each month) (Table 2). Logistical constraints and poor weather conditions precluded the gathering of observer data of the fishing trips for the months of November 2013 and January 2014 at Constante port.

19 N Individuals Latin American Journal of Aquatic Research Table 1. Number of sea turtles incidentally caught with gillnets in Sechura Bay in 53 fishing sets, morphometric measures and animal fate, July 2013 to June CCL: curved carapace length. Species Bycatch CCL (cm) Fate Mean ± SD Range Alive Dead Chelonia mydas ± Eretmochelys imbricata ± Lepidochelys olivacea na Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Winter Spring Summer Autumn Season Figure 2. Numbers of incidentally captured sea turtles by season and by month in the Sechura Bay, *Notice that November and January do not have fishing sets. Of the 104 turtles caught, we obtained complete morphometric data from 99 individuals (Fig. 3). The remaining five animals for which data collection was not possible were all C. mydas. The observed CCL by species was 57.3 ± 0.9 cm (range: 40.5 cm to 79.6 cm) for C. mydas; 40.2 ± 2.4 cm (range: 36.5 to 44.6 cm) for E. imbricata, while the only olive ridley L. olivacea measured 64.1 cm CCL. Of all turtles captured, 28.8% (30 individuals) were recovered dead (28 C. mydas and 2 E. imbricata). The remaining 71.2% (74 individuals) were captured alive, tagged and released. These consisted of 72 C. mydas, one E. imbricata, and one L. olivacea. The overall CPUE observed was 1.11 ± 0.31 turtles km h -1 or 0.11 ± 0.03 turtles km h -1. Sets during the winter had the highest observed CPUE with a mean of 1.94 ± 1.58 turtles km h -1 or 0.19 ± 0.10 turtles km h -1. Similarly, the CPUE varied among months. July recorded the highest CPUE with 5.11 ± 2.65 turtles km h -1 or 0.43 ± 0.22 turtles km h -1, while December had the lowest CPUE with 0.16 ± 0.1 turtles km h -1 or 0.01 ± 0.01 turtles km h -1 (Table 2). The annual bycatch for the Constante Port was 183 sea turtles. DISCUSSION Sea turtle catch rates Recent declines of large marine vertebrates, such as sea turtles, seabirds and marine mammals, have focused attention on the ecological impacts of incidental take, or bycatch, in global fisheries (Oravetz, 2000; Wallace et al., 2010; Lewison et al., 2014). Sea turtles are incidentally captured in almost all fishing gear, including trawl nets, gillnets, pelagic and bottom longlines (Lewison et al., 2004; Rosales et al., 2010). Among these, gillnet fisheries may be the single largest threat to sea turtle populations (Gilman et al., 2010; Wallace et al., 2010). In Peru, gillnets were reported as the main source of turtle mortalities in artisanal fisheries from 1986 to 1999 (Estrella & Guevara- Carrasco, 1998a, 1998b; Estrella et al., 1999a, 1999b) and have been the focus of attention in recent years (Mangel et al., 2010; Wang et al., 2013; Ortiz et al.,

20 Frecuency (n) Incidental capture of sea turtles in Sechura Bay, Peru 6105 Table 2. Monthly and seasonal capture per unit effort (CPUE; turtles km h -1 and turtles km h -1 ), July 2013 to June Season Month Bycatch (n) Fishing set CPUE (turtles km h -1 ) CPUE (turtles km h -1 ) Month Season Month Season Winter July ± ± 0.22 August ± ± ± ± 0.10 September ± ± 0.02 Spring October ± ± ± 0.72 December ± ± ± 0.03 Summer February ± ± ± 0.10 March ± ± ± 0.01 April ± ± 0.01 Autumn May ± ± ± ± 0.06 June ± ± Chelonia mydas Eretmochelys imbricata Lepidochelys olivacea CCL (cm) Figure 3. Size-classes of sea turtles captured, by species, in the Sechura Bay, July 2013 to June ). Our results suggest that gillnets are an important source of bycatch and mortality of sea turtles in Sechura Bay, being a threat to the sea turtles populations in this important foraging and developmental habitat (de Paz & Alfaro-Shigueto, 2008; Santillán, 2008). Studies of the sea turtle bycatch suggested that bycatch rates reported for gillnets in Sechura Bay are among the highest in the world (Wallace et al., 2010; Alfaro-Shigueto et al., 2011). Alfaro-Shigueto et al. (2011) reported that is notable the high proportion of bycatch-positive sets and high CPUE for green turtles in the bottom set nets at Constante port (56%; 2.78 turtle per set). Caceres et al. (2013) observed that all monitored trips with sea turtle interactions were by bottom set net boats. These bycatch rates are similar to the present study, reporting that 62.3% of observed sets had bycatch. The mean CPUE was approaching two turtles per set and the mortality rate was 28.8%. These values are of concern and we anticipate that would be higher in the absence of onboard observers. Our bycatch results are in agreement with other studies investigating sea turtle bycatch by net fisheries in Peru. A study in Pisco-Paracas, De Paz et al. (2002) reported a total of 204 sea turtles caught in gillnets during 276 monitored days, with the bycatch consisting of C. mydas (67.8%), L. olivacea (27.7%) and D. coriacea (2.9%). Castro et al. (2012) monitored 265 fishing operations from Lambayeque, from which a total of 383 sea turtles were captured: being 80.4% olive ridleys, 19.3% green turtles and 0.2% hawksbill turtles. Cáceres et al. (2013), collected data from the Constante Port and recorded that 43 green turtles were

21 6611 Latin American Journal of Aquatic Research captured during 14 monitored trips. Rosales et al. (2010) in Tumbes registered 95 specimens belonging to four sea turtles species (Chelonia mydas, Lepidochelys olivacea, Dermochelys coriacea and Eretmochelys imbricata); the most registered species were C. mydas (64.2%) and L. olivacea (30.5%). In each of these studies, C. mydas was one the most frequently caught species of sea turtle. Our results reinforce these findings, we observed the capture of 104 sea turtles, of which the vast majority were C. mydas (96.2%), followed by E. imbricata (2.9%), and one individual L. olivacea (0.9%). The results reported in this research suggest that the bycatch in gillnets is one of the main cause of mortality of sea turtles in this area. We reported a mortality percentage of 28.8, being higher than the reported in industrial shrimp trawlers vessels of eastern Venezuela (17.5%) (Alio et al., 2010). The CPUE of 0.61 ± 0.22 turtles set -1 reported by Rosales et al. (2010) in Tumbes for a research of three years was lower than the reported in this study (1.96 ± 0.44 turtles set -1 ). However the annual bycatch rate was lower than the report in the same area by Alfaro-Shigueto et al. (2011), reporting a high proportion of bycatch-positive sets and obtained an annual bycatch rate by this fishery of 368 sea turtles. This study also reported a high bycatch per unit effort (BPUE) for green turtle (2.78 turtles per set), and a mortality rate of 41%, being it higher than the results obtained in this study. As part of experimental research carried out from 2011 to 2013, also in the Sechura Bay demersal gillnet fishery, Ortiz et al. (2016) obtained a CPUE of 1.40 ± 0.16 green turtles km h -1 in control nets, this bycatch rate is similar to our observed CPUE of 1.11 ± 0.31 turtles km h -1. Given these high rates of observed bycatch, in order to secure long-term population viability and to conform with international guidelines for responsible fisheries (FAO, 2009), sea turtle bycatch mitigation solutions for these fishery need to be identified to minimize the number of bycatch mortalities (Nguyen et al., 2013). While net modifications have, in some cases, resulted in megafauna bycatch mitigation in certain fisheries without substantial reductions in target catch, mitigating net bycatch has proven challenging because nets are inherently nonselective (Peckham et al., 2015 & Gilman et al., 2010). However, recent research, conducted in-part in the Constante demersal set-net fishery, suggests that sea turtle bycatch in gillnets could be reduced through illuminating nets (Wang et al., 2013 & Ortiz et al., 2016). Sea turtle size classes The sizes of C. mydas captured in Sechura Bay in this study corresponded to a population consisting of juveniles (89.5%) and sub-adults (10.5%). The mean CCL = 57.3 ± 0.9 cm (range = 40.5 to 79.6). In this same area, Santillán (2008), analyzed the fishery bycatch from Constante Port and found a mean CCL for green turtles of 63.6 ± 1.6 CCL (range: 47.5 to 88 cm; n = 45), indicating a concentration of juvenile and subadult turtles. Cáceres et al. (2013), obtained a mean CCL of 60.2 ± 6.8 cm (range: 52 to 92 cm), and thus considered all as juvenile turtles. Paredes et al. (2015) in Virrilá estuary reported a mean CCL of 59.2 ± 10.2 cm (range = cm) indicating a population represented by juveniles (62.6%) Our results suggest that this bay harbor an immature population. In the Eastern Pacific Ocean (EP), the hawksbill turtle has been reported as once common from Mexico to Ecuador (Alfaro-Shigueto et al., 2010a; Gaos et al., 2010) recruited to neritic habitats (Scales et al., 2011). Studies in Máncora, Constante and Parachique, from 2000 to 2005, found a mean CCL of 38.9 ± 5.9 cm (range cm, n = 11), indicating a population of mostly immature individuals (Alfaro et al., 2010a). Quiñones et al. (2011) in the San Andrés area, reported a mean CCL size of 45.2 ± 3.2 cm, and they concluded that juveniles and sub-adults used this area as a foraging ground. We reported the incidental capture of three E. imbricata, with a mean CCL of 40.2 ± 2.4 cm (range: cm). This size corresponded to juvenile individuals and is similar to other reports for the northern coast of Peru. Kelez et al. (2003) measured 16 carapaces of L. olivacea from Tumbes to Ancash (Peru) from 2001 to 2002 and reported a mean CCL of 66.6 cm. For the year 2008, in Tumbes, measurements of 47 L. olivacea carapaces yielded a mean CCL of 63.3 ± 4.5 (range: cm; n = 47) indicating the presence of juveniles, subadults and adults (Forsberg, 2012). Our observed bycatch of one L. olivacea with a CCL of 64.1 cm is consistent with a sub-adult sized individual. CONCLUSION AND RECOMMENDATIONS Gillnets are a significant source of sea turtle bycatch in Sechura Bay. Our results indicate that Sechura Bay is an important foraging area and developmental habitat for green turtles and also possibly for critically endangered hawksbill turtles. This research found that a majority of sets having bycatch, given these catch rates; we recommend the identification and implementation of mitigation measures to reduce sea turtle bycatch, like illuminating nets with LED lights, shark silhouettes and use float lines without buoys (Gilman et al., 2010; Wang et al., 2010, 2013; Ortiz et al. 2016). To help maximize their uptake and effectiveness, such efforts to identify solutions should involve small-scale

22 Incidental capture of sea turtles in Sechura Bay, Peru fishermen as well as scientists and other stakeholders and decision-makers. Enhanced management and protection of this bay that acknowledges its importance as a developmental habitat and foraging ground is recommended. To decrease sea turtle captures and commerce, efforts are needed to offer fishermen new economic alternatives. Additional efforts should include an education and research program targeting the Sechura Bay community. ACKNOWLEDGEMENTS We want to acknowledge the fishermen of Constante Port for their generous collaboration in the monitoring trips and Pro Delphinus staff biologists who provided data and field work. To biologist Armando Ugaz of National University of Piura for his unconditional support in this research. This study was conducted with support of the National Fish and Wildlife Foundation- NFWF, Whitley Fund for Nature-WFN, Darwin Initiative Sustainable Artisanal Fisheries Initiative in Peru- DEFRA, University of Exeter and Prodelphinus. Besides sample collection was made possible by General Direction Resolution Nº SERFOR- DGGSPFFS. REFERENCES Alfaro-Shigueto, J., P. Dutton, J. Mangel & D. Vega First confirmed occurrence of loggerhead turtles in Peru. Mar. Turtle Newslett., 103: Alfaro-Shigueto, J., M. Van Bressem, D. Montes & K. Onton Turtle mortality in fisheries off the coast of Peru. In: A. Mosier, A. Foley & B. Brost (eds.). Proceedings of the Twentieth Annual Symposium on sea turtle biology and conservation. NOAA Tech. Memo. NMFS-SEFSC-477, 369 pp. Alfaro-Shigueto, J., J. Mangel, F. Bernedo, P. Dutton, J. Seminoff & B. Godley Small-scale fisheries of Peru: a major sink for marine turtles in the Pacific. J. Appl. Ecol., 48: Alfaro-Shigueto, J., J. Mangel, M. Pajuelo, C. Cáceres, J. Seminoff & P. Dutton Bycatch in Peruvian artisanal fisheries: gillnets versus longlines. In: A. Rees, M. Frick, A. Panagopoulou & K. Williams (eds.). Proceedings of the Twenty-Seventh Annual Symposium on Sea Turtle Biology and Conservation. NOAA Tech Memo NMFS-SEFSC-569, 262 pp. Alfaro-Shigueto, J., J. Mangel, C. Cáceres, J. Seminoff, A. Gaos & I. Yañez. 2010a. Hawksbill turtles in Peruvian coastal fisheries. Mar. Turtle Newslett., 129: Alfaro-Shigueto, J., J. Mangel, M. Pajuelo, P. Dutton, J. Seminoff & B. Godley. 2010b. Where small can have a large impact: structure and characterization of smallscale fisheries in Peru. Fish. Res., 106: Alió, J., L. Marcano & D. Altuve Incidental capture and mortality of sea turtles in the industrial shrimp trawlingfishery of northeastern Venezuela. Cienc. Mar., 36(2): Barrientos-Muñoz, K., C. Ramírez-Gallego & V. Páez Nesting ecology of the olive ridley sea turtle (Lepidochelys olivacea) (Cheloniidae) at El Valle Beach, Northern Pacific, Colombia. Acta Biol. Colomb., 19(3): Bolten, A Técnicas para la medición de tortugas marinas. In: K. Eckert, K. Bjorndal, F. Abreu-Grobois & M. Donnelly (eds.). Técnicas de investigación y manejo para la conservación de las tortugas marinas. Traducida al español. UICN/CSE Grupo Especialista en Tortugas Marinas, pp Cáceres, C., J. Alfaro & J. Mangel Green turtle captured in net fisheries in the Port of Constante, Peru. In: J. Blumenthal, A. Panagopoulou & A. Rees (eds.). Proceedings of the Thirtieth Annual Symposium on Sea Turtle Biology and Conservation. NOAA Tech Memo NMFS-SEFSC-640, 177 pp. Casale, P Incidental catch of marine turtles in the Mediterranean Sea: captures, mortality, priorities. WWF Italy, Rome, 64 pp. Casale, P Sea turtle by-catch in the Mediterranean. Fish Fish., 12: Castro, J., J. de la Cruz, P. Ramírez & J. Quiñones Captura incidental de tortugas marinas durante El Niño , en el norte del Perú. Lat. Am. J. Aquat. Res., 40(4): De Paz, N. & J. Alfaro-Shigueto Foraging grounds for sea turtles in inshore Peruvian waters. In: H. Kalb, A. Rohde, K. Gayheart & K. Shanker (eds.). Proceedings of the Twenty-Fifth Annual Symposium on Sea Turtle Biology and Conservation. NOAA Technical Memorandum NMFS-SEFSC-582, 204 pp. De Paz, N., J. Reyes & M. Echegaray Datos sobre captura, comercio y biología de tortugas marinas en el área de Pisco - Paracas. In: J. Mendo & M. Wolff (eds.). Memorias I Jornada Científica "Bases ecológicas y socioeconómicas para el manejo de los recursos vivos de la Reserva Nacional de Paracas". Universidad Nacional Agraria La Molina, pp Dutton, P. & D. Squires Reconciling biodiversity with fishing: a holistic strategy for Pacific sea turtle recovery. Ocean Develop. Int. Law, 39: Eckert, S. & L. Sarti Distant fisheries implicated in the loss of the world s largest leatherback population. Mar. Turtle Newslett., 78: 2-7.

23 613 8 Latin American Journal of Aquatic Research Estrella, C. & R. Guevara-Carrasco. 1998a. Informe estadístico anual de los recursos hidrobiológicos de la pesca artesanal por especies, artes, caletas y meses durante Inf. Inst. Mar Perú, 131: Estrella, C. & R. Guevara-Carrasco. 1998b. Informe estadístico anual de los recursos hidrobiológicos de la pesca artesanal por especies, artes, caletas y meses durante Inf. Inst. Mar Perú, 132: Estrella, C., R. Guevara-Carrasco & J. Palacios. 1999a. Informe estadístico de los recursos hidrobiológicos de la pesca artesanal por especies, artes, caletas y meses durante el segundo semestre de Inf. Inst. Mar Perú, 143: Estrella, C., R. Guevara-Carrasco, J. Palacios, W. Ávila & A. Medina. 1999b. Informe estadístico de los recursos hidrobiológicos de la pesca artesanal por especies, artes, meses y caletas durante el primer semestre de Inf. Inst. Mar Perú, 148: Food and Agriculture Organization (FAO) Guidelines to reduce sea turtle mortality in fishing operations. Food and Agriculture Organization, Rome, 128 pp. Food and Agriculture Organization (FAO) Operaciones pesqueras. Mejores prácticas para reducir las capturas incidentales de aves marinas en la pesca de captura. Orientaciones técnicas de la FAO para la pesca responsable. Food and Agriculture Organization, Rome, 1(2): Forsberg, K Assessing sea turtle bycatch and mortality in North Peru: a community conservation initiative. In: L. Belskis, M. Frick, A. Panagopoulou, A. Rees & K. Williams (eds.). Proceedings of the Twenty-ninth Annual Symposium on Sea Turtle Biology and Conservation. NOAA Technical Memorandum NOAA NMFS-SEFSC-630, 192 pp. Gaos, A., A. Abreu-Grobois, J. Alfaro-Shigueto, D. Amorocho, R. Arauz, A. Baquero, et al Signs of hope in the eastern Pacific: international collaboration reveals encouraging status for a severely depleted population of hawksbill turtles Eretmochelys imbricata. Oryx, 44: Gilman, E., J. Gearhart, B. Price, S. Eckert, H. Milliken, J. Wang, Y. Swimmer, D. Shiode, O. Abe, S. Peckham, M. Chaloupka, M. Hall, J. Mangel, J. Alfaro-Shigueto, P. Dalzell & A. Ishizaki Mitigating sea turtle by-catch in coastal passive net fisheries. Fish. Fish., 11: Gobierno Regional Piura (GORE-Piura) Ámbito de gestión de la zona marino costera de Piura. Documento elaborado en el marco del Programa Regional de Manejo Integrado de la Zona Marino Costera de Piura. Cooperación Alemana al Desarrollo-Agencia GIZ en el Perú. [ accion=verelemento&idelementoinformacion=570& verpor=&idtipoelemento=35&idtipofuente=&idfue nteinformacion=48]. Reviewed: 10 October Hays-Brown, C. & W. Brown Status of sea turtles in the South eastern Pacific: emphasis on Peru. In: K. Bjorndal (ed.). Biology and conservation of sea turtles. Smithsonian Institution Press, Washington DC, pp Kelez, S., X. Vélez-Zuazo & C. Manrique New evidence on the loggerhead sea turtle Caretta caretta (Lineaus, 1758) in Peru. Ecol. Appl., 2(1): Kelez, S., X. Vélez-Zuazo, F. Angulo & C. Manrique Olive ridley Lepidochelys olivacea nesting in Peru: the southernmost records in the eastern Pacific. Mar. Turtle Newslett., 126: 5-9. Koch, V., H. Peckham, A. Mancini & T. Eguchi Estimating at-sea mortality of marine turtles from stranding frequencies and drifter experiments. PLoS ONE, 8(2): e Lewison, R., L. Crowder, A. Read & S. Freeman Understanding impacts of fisheries bycatch on marine megafauna. Trends Ecol. Evol., 19(11): Lewison, R., L. Crowder, B. Wallace, J. Moore, T. Coxe, R. Zydelis, S. McDonald, A. DiMatteo, D. Dunn, C. Kot, R. Bjorkland, S. Kelez, C. Soykan, K. Stewart, M. Sims, A. Boustany, A., Read, P. Halpin, W. Nichols & C. Safina Global patterns of marine mammal, seabird, and sea turtle bycatch reveal taxa-specific and cumulative megafauna hotspots. Proc. Nat. Acad. Sci. USA, 111(14): 1-6. Liles, M., M. Jandres, W. López, G. Mariona, C. Hasbún & J. Seminoff Hawksbill turtles Eretmochelys imbricata in El Salvador: nesting distribution and mortality at the largest remaining nesting aggregation in the eastern Pacific Ocean. Endang. Species Res., 14(1): López-Barrera, E., G. Longo & E. Monteiro-Filho Incidental capture of green turtle (Chelonia mydas) in gillnets of small-scale fisheries in the Paranaguá Bay, Southern Brazil. Ocean Coast. Manage., 60: López-Mendilaharsu, M., A. Estrades, M. Caraccio, V. Calvo, M. Hernández & V. Quirici Biología, ecología y etología de las tortugas marinas en la zona costera uruguaya. In: R. Menafra, L. Rodríguez- Gallego, F. Scarabino & D. Conde (eds.). Bases para la conservación y el manejo de la Costa Uruguaya. Montevideo, Uruguay: Vida Silvestre Uruguay, pp Mangel, J., J. Alfaro-Shigueto, K. Van Waerebeek, C. Cáceres, S. Bearhop, M. Witt & B. Godley Small cetacean captures in Peruvian artisanal fisheries: high despite protective legislation. Biol. Conserv., 143:

24 Incidental capture of sea turtles in Sechura Bay, Peru 6149 Moore, J., T. Cox, R. Lewison, A. Read, R. Bjorkland, S. McDonald, L. Crowder, E. Aruna, I. Ayissi, P. Espeut, C. Joynson-Hicks, N. Pilcher, C. Poonian, B. Solarin & J. Kiszka An interview-based approach to assess marine mammal and sea turtle captures in artisanal fisheries. Biol. Conserv., 143: Morón, O., F. Velazco & L. Beltrán Características hidrográficas y sedimentológicas de la Bahía de Sechura. Inf. Inst. Mar Perú, 40(3-4): Murray, K Characteristics and magnitude of sea turtle bycatch in US mid-atlantic gillnet gear. Endang. Species Res., 8: Nguyen, V., S. Larocque, L. Stoot, N. Cairns, G. Blouin- Demers & S. Cooke Perspectives of fishers on turtle bycatch and conservation strategies in a smallscale inland commercial fyke net fishery. Endang. Species Res., 22: Oravetz, C Reducción de la captura incidental en pesquerías. In: K. Eckert, K. Bjorndal, F. Abreu- Grobois & M. Donnelly (eds.). Técnicas de investigación y manejo para la conservación de las tortugas marinas. Traducida al español. UICN/CSE Grupo Especialista en Tortugas Marinas, pp Ortiz, N., J. Mangel, J. Wang, J. Alfaro-Shigueto, S. Pingo, A. Jiménez, T. Suarez, Y. Swimmer, F. Carvalho & B. Godley Reducing green turtle bycatch in small-scale fisheries using illuminated gillnets: the cost of saving a sea turtle. Mar. Ecol. Prog. Ser., 545: Paredes, E., J. Quiñones, S. Quispe & V. Bachmann Black and hawksbill turtle strandings in estuarine waters in the Peruvian Northern coast. In: Y. Kaska, B. Sonmez, O. Turkecan & C. Sezgin (comp.). Book of abstracts of 35 th Annual Symposium on Sea Turtle Biology and Conservation. Macart Press, Turkey, 250 pp. Peckham, S., J. Lucero-Romero, D. Maldonado-Diaz, A. Rodriguez-Sanchez, J. Senko, M. Wojakowski & A. Gaos Buoyless nets reduce sea turtle bycatch in coastal net fisheries. Conserv. Lett., 9(2): Quiñones, J., J. Zeballos, S. Quispe & L. Delgado Southernmost records of hawksbill turtles along the East Pacific coast of South America. Mar. Turtle Newslett., 130: Received: 14 July 2016; Accepted: 22 April 2017 Rosales, C., M. Vera & J. Llanos Varamientos y captura incidental de tortugas marinas en el litoral de Tumbes, Perú. Rev. Peru. Biol., 17(3): Santillán, L Análisis de la dieta de Chelonia mydas agassizii tortuga verde del Pacífico en la Bahía de Sechura, Piura-Perú. Tesis de Magister, Universidad Agraria La Molina, Lima, 70 pp. Scales, K., J. Lewis, R. Graham & B. Godley The hawksbill turtle, Eretmochelys Imbricata, at Lighthouse Reef Atoll, Belize. Testudo, 7(3): Seaturtle Maptool Program for Analysis and Graphics. [ Reviewed: 15 September The State of the World s Sea Turtles (SWOT) Special Feature South America. SWOT Report XI. Oceanic Society. SeaturtleStatus.org 11: 48 pp. Tume, J., A. Ibaceta, M. Cortez & J. Santisteban Recursos de la zona béntica de la bahía de Sechura. Cienc. Desarrollo, 15(1): 50 pp. Velez-Zuazo, X., J. Quiñones, A. Pacheco, L. Klinge, E. Paredes, S. Quispe & S. Kelez Fast growing, healthy and resident green turtles (Chelonia mydas) at two neritic sites in the central and northern coast of Peru: implications for conservation. PLoS ONE, 9(11): Wallace, B., S. Heppell, R. Lewison, S. Kelez & L. Crowder Impacts of fisheries bycatch on loggerhead turtles worldwide inferred from reproductive value analyses. J. App. Ecol., 45: Wallace, B., R. Lewison, S. McDonald, R. McDonald, C. Kot, S. Kelez, R. Bjorkland, E. Finkbeiner, S. Helmbrecht & L. Crowder Global patterns of marine turtle bycatch. Conserv. Lett., 3: Wang, J., S. Fisler & Y. Swimmer Developing visual deterrents to reduce sea turtle bycatch in gill net fisheries. Mar. Ecol. Prog. Ser., 408: Wang, J., J. Barkam, S. Fisler, C. Godinez-Reyes & Y. Swimmer Developing ultraviolet illumination of gillnets as a method to reduce sea turtle bycatch. Biol. Lett., 9: 1-4. Zárate, P., K. Bjorndal, M. Parra, P. Dutton, J.Seminoff & A. Bolten Hatching and emergence success in green turtle Chelonia mydas nests in the Galápagos Islands. Aquat. Biol., 19:

25 Lat. Am. J. Aquat. Res., 45(3): , Hawksbills 2017 at a major foraging ground in the eastern Pacific 597 Sea Turtle Research and Conservation in Latin America Jeffrey Mangel, Jeffrey Seminoff, Bryan Wallace & Ximena Vélez-Suazo (Guest Editors) DOI: /vol45-issue3-fulltext-9 Research Article Distribution, size range and growth rates of hawksbill turtles at a major foraging ground in the eastern Pacific Ocean Israel Llamas 1, Eric E. Flores 2,3, Marino E. Abrego 2,9,11,12, Jeffrey A. Seminoff 4, Catherine E. Hart 5 Rodrigo Donadi 6, Bernardo Peña 2, Gerardo Alvarez 7, Wilfredo Poveda 2 Diego F. Amorocho 8 & Alexander Gaos 9,10 1 Eco-Mayto A.C., Cabo Corrientes, Jalisco, México 2 Ministerio de Ambiente de Panamá 3 Sistema Nacional de Investigación de Panamá (SNI), Panamá 4 National Oceanic and Atmospheric Administration, Southwest Fisheries Science Center La Jolla, California, USA 5 Grupo Tortuguero de las Californias A.C., La Paz, Baja California Sur, México and Investigación, Capacitación y Soluciones Ambientales y Sociales A.C. Tepic, Nayarit, México 6 Ciudad de Panamá, Panamá 7 Tortuguias, Panamá 8 World Wilflife Fund for Nature, Secretariado Latino America y el Caribe, Cali, Colombia 9 Eastern Pacific Hawksbill Initiative, San Diego, California, USA 10 San Diego State University, San Diego, California, USA 11 Universidad Marítima Internacional de Panamá, UMIP 12 Fundación Agua y Tierra, FUNDAT, Panamá Corresponding author: Israel Llamas (israel_llamas@hotmail.com) ABSTRACT. Hawksbill sea turtles (Eretmochelys imbricata) inhabiting the eastern Pacific Ocean are one of the world s most threatened marine turtle management units. Despite the fact that knowledge about the status of sea turtles at foraging grounds is a key element for developing the effective conservation strategies, comprehensive studies of hawksbills at foraging habitats in the eastern Pacific remain lacking. For many years anecdotal information indicated Coiba Island National Park in Panama as a potentially important hawksbill foraging ground, which led to the initiation of monitoring surveys in September Ongoing mark-recapture surveys to assess population status, generate demographic data and identify key foraging sites have been conducted every six months in the park since that time. To date, a total of six monitoring campaigns consisting of four days each have been conducted, leading to the capture and tagging of 186 hawksbills, 51 of which were recaptured at least once. The size range of captured individuals was 30.0 to 75.5 cm and largely comprised of juveniles. Somatic growth rates of individual hawksbills were highly variable, ranging from to 7.1 cm year -1. To our knowledge, these are the first published growth rates for juvenile hawksbill turtles in the eastern Pacific Ocean. When these growth data are combined with information on hawksbill demography and distribution, our findings indicate Coiba Island National Park is one of the most important known foraging sites for hawksbill sea turtles in the eastern Pacific Ocean. Keywords: marine turtle, Eretmochelys imbricata, demography, management, conservation, Panama. INTRODUCTION Knowledge about the status of sea turtle populations at coastal foraging areas is a key element for developing effective conservation strategies. In contrast to nesting beach surveys that focus solely on adult females, foraging area assessments can provide demographic information on a broad range of age-classes of both sexes, thus giving insights about present and future population abundance trends (Diez & Van Dam, 2002; Blumenthal et al., 2009). Additionally, detailed inwater diagnostic evaluations of sea turtle populations are imperative to determine growth and survival rates, thus further assisting the development of best management options (Manly, 1990; NRC, 2010). The need for status assessments in foraging areas has been high-

26 598 Latin American Journal of Aquatic Research lighted (Chaloupka & Limpus, 2001; Rees et al., 2016), and such studies have become much more frequent in recent years (León & Diez, 1999; Eguchi et al., 2010; Burkholder et al., 2011). However, in-water surveys have tended to focus on easily accessible sites such as mainland coastal regions, whereas foraging grounds located in remote and hard-to-access settings, particularly insular habitats, remain understudied. The hawksbill turtle, Eretmochelys imbricata, is a highly endangered sea turtle with a circumtropical distribution (Mortimer & Donnelly, 2008). The species is best described by its elongated beak and imbricate scutes on the carapace and plastron, especially during juvenile and subadult life stages. Known as tortoiseshell or bekko, these plates have caused the hawksbill turtle to be the target of an exhaustive harvest for artisanal uses throughout the world (Groombridge & Luxmoore, 1989; Shattuck, 2011). This demand, coupled with the loss of nesting habitat and harvest of eggs and meat for human consumption has caused hawksbill turtle populations to plummet worldwide (Meylan & Donnelly, 1999). In the eastern Pacific Ocean, hawksbill turtles are listed as critically endangered and remain one of the world's most endangered regional management units (Wallace et al., 2011). However, in contrast to most of today's modern conservation scenarios, the conservation narrative for hawksbill turtles in the eastern Pacific Ocean is evolving into a positive one. This stems from the fact that only a decade ago the conservation community believed the species had been nearly extirpated from the region (Mortimer & Donnelly, 2008). This unfavorable conservation outlook began to change in 2007 with the discovery of several critical hawksbill nesting beaches and foraging areas (Gaos et al., 2010) and the subsequent establishment of conservation projects at many of these critical sites has provided hope for recovery. Although numerous hawksbill foraging grounds have been identified in the eastern Pacific, the majority of reports have resulted from opportunistic studies focused on other subjects (e.g., fisheries bycatch, nesting beach conservation, etc.) or have been shortterm in nature and of limited sample sizes (Table 1) (Alfaro-Shiqueto et al., 2010; Carrión-Cortez et al., 2010; Quiñones et al., 2011; Brittain et al., 2012; Chacón-Chaverri et al., 2014; Heidemeyer et al., 2014; Tobón-López & Amorocho, 2014). As a result, comprehensive information on hawksbill demographics from foraging grounds in the eastern Pacific remains extremely limited. Considering the vast resources that have been invested in conservation at nesting beaches over the last decades (Gaos et al., 2017), structured on-going surveys at foraging grounds Table 1. Published hawksbill in-water monitoring studies available for the eastern Pacific Ocean, including author, country and sample size (n). Author Country n In this study Panama 186 Chacón-Chaverri et al. (2014) Costa Rica 62 Heidemeyer et al. (2014) Costa Rica 11 Tobón & Amorocho (2014) Colombia 16 are needed to understand rates of recruitment and population trends in order to gauge the effectiveness of conservation efforts on the nesting beaches, as well as to better guide future conservation actions. Coiba National Park (CNP) is a marine reserve composed of 39 major islands located in the Gulf of Chiriqui, Republic of Panama (ANAM, 2009) (Fig. 1). Coiba Island, the primary island within CNP, served as a penal colony from 1919 to 2004, the unintentional result being the preservation of natural resources in the area. This includes the largest extent of coral aggregations in the eastern Pacific (Glynn, 1997). The pristine environmental conditions of the archipelago led to the declaration of the area as a National Park in 1992, enforced by law in 2004, and as a UNESCO World Heritage Site in Yet despite its intact nature, few studies on the status of sea turtles are available due to its remote location. Surveys carried out in 2011 to evaluate sea turtle presence around Coiba Island identified the area as an important nesting and foraging site for several species (Ruiz & Rodriguez, 2011), but species-specific quantifications remained unavailable. Based on these findings and the presence of the pristine coral reefs in the CNP, which represent a primary hawksbill foraging habitat (Meylan, 1988; Van Houtan et al., 2016), a three-day rapid in-water assessment was carried out in September 2014 to evaluate the presence of hawksbill turtles. The research team, composed of national and international researchers and organizations, observed 103 individual hawksbills during that time, immediately revealing CNP as one of the most important hawksbill foraging areas known in the eastern Pacific. These findings led to the subsequent establishment of a consistent in-water monitoring program at the site in the coming years. Here we present the results of hawksbill in-water monitoring at CNP between 2014 and 2016, which represent the most comprehensive set of in-water monitoring data for hawksbills at foraging grounds in the eastern Pacific, as well as the first for a hawksbill foraging ground in Panama. These results will serve as a baseline for long-term studies of hawksbill status at this insular marine protected area and also serve as a

27 Hawksbills at a major foraging ground in the eastern Pacific 599 Figure 1. Map of Coiba National Park, Panama, with insets a) and b) showing the location and details of several hawksbill survey sites. point of comparison for other hawksbill foraging areas throughout the eastern Pacific. METHODS AND MATERIALS Study area The CNP belongs to the Panamic Biogeographic Province, which extends from the Gulf of Guayaquil in Ecuador (3ºS) to the Gulf of Tehuantepec in Mexico (16ºN) (Cortes, 1997). The climate of the region is humid-tropical monsoonic, with a rainfall of up to 3500 mm year -1, an average temperature of 25.9ºC, and marked seasonality; it has dry (from mid-december to mid-april) and rainy seasons. The islands are covered by tropical rainforest, and they have several rivers with variable flows and hydrographic basin sizes. Coral reefs in the CNP are generally small in area, shallow (<15 m) and structurally simple, and possess low scleractinian coral species richness (Cortes, 1997). A total of 56 hard coral species, 20 species of scleractinian corals (dominated by Pavona spp.), and two species of hydrocorals are present. In addition to these corals a variety of frondose and turf macroalgal species are present; algal communities consist mainly of Gelidiopsis intricata, Hypnea pannosa, Dictyota spp. and Amphiroa beauvosii. A variety other invertebrates including sponges are also present (Guzman et al., 2004).

28 600 Latin American Journal of Aquatic Research Survey methodology, capture procedures and data collection We conducted systematic hawksbill monitoring campaigns at CNP every six months from September 2014 to September 2016 (six total field visits), each lasting four days. During each visit to CNP we used a 25-ft. skiff with outboard motor to visit previouslyidentified coral reefs in the northwestern, northeastern and southern portions of CNP (Fig. 1). Hawksbills were spotted during diurnal and nocturnal snorkel surveys and when possible, were captured by hand while free diving. Captured turtles were immediately brought aboard the skiff, where measurements of curved carapace length (CCL) and curved carapace width (CCW) were taken. Each turtle was tagged with Inconel tags (Style 681, National Band and Tag Company, Newport, Kentucky, USA) on the front left flipper. During the last three campaigns, we also tagged all hawksbills subcutaneously with a passive integrated transponder (PIT) (Avid, Norco, CA, USA) along the turtles front left flipper, and post proper functioning of the PIT tag was confirmed through the use of a scanning device (AVID Power Tracker IV). When possible, body weight (kg) was also measured using a spring balance. After tags were applied and measurements taken, we used sanitary techniques to collect epidermal skin tissue biopsies from the dorsal surface of the neck region (Dutton, 1996) for later genetic and stable isotope analyses, then released the hawksbills at their original site of capture. Capture per unit effort (CPUE) Capture per unit effort (CPUE) was calculated based on the total number of hawksbills encountered (i.e., captures and recaptures) at each site per monitoring campaign. This total was then divided by the total time (h) spent during diving and the number of divers participating in surveys, expressed as captures per person per hour. Data analysis Because CCL and CCW were significantly correlated (Spearman, S = ; rho = 0.79; P < 0.01), and in our dataset some CCW values were missing, CCL was used as the sole biometric variable for size. Mean growth rate was calculated from the difference in CCL recorded at first capture and last recapture of each individual, with a minimum 60 day interval between capture events (Hawkes et al., 2014). Growth rates were calculated from the mark-recapture profiles for each foraging site sampled. Negative or zero growth rates were also included, since these are part of the measurement error. All statistical analyses were performed in R v3.3.0 (R Core Team, 2016). When possible, data was transformed to meet parametric assumptions, otherwise non-parametric tests were conducted. All values reported in the results are means (±SE) unless otherwise indicated. RESULTS Captures and recaptures per site Hawksbills were captured at all 13 monitored foraging sites (Fig. 1, Table 2). In total, 186 individual hawksbill turtles were captured and tagged, including 24 in September 2014, 28 in March 2015, 26 in September 2015, 40 in January 2016, 36 in July 2016 and 32 in September 2016 (Table 2). The most hawksbills (n = 53) were captured at Playa Blanca (Fig. 1b, Table 2), followed by Canales de Afuera (n = 36), Bahía Rosario (n = 32) and Granito de Oro (n = 30) (Table 2). A total of 51 out of 186 turtles were recaptured, with turtles at large from 0.21 to 2.05 year (mean = 2.0 ± 1.2; median = 1.7). Granito de Oro was the site with the most recaptures (n = 66) (Fig. 2a, Table 2), followed by Canales de Afuera (n = 18), Playa Blanca (n = 9) and Bahía Rosario (n = 9) (Table 2). Sex was determined for 20 of the 186 captured turtles (females: 11; males: 9), the remaining 166 turtles were juveniles or subadults. Size class distribution and growth rates Mean size of captured individuals was CCL = ± 0.71 cm (range: cm), and a CCW = ± 0.85 cm (range: cm) (mean ± SE). Initial CCLs ranged from 30.0 to 75.5 cm (mean: 45.6 cm), with the most common size class (n = 44) being cm CCL (Fig. 2a). Overall there was a significant difference in CCL among the study sites (ANOVA: Site, F 12,172= 1.88, P = 0.038). There were 23 multiple recaptures (we recaptured eight turtles twice, five turtles three times, five turtles four times and five turtles five times), which resulted in 74 growth-rate measurements. Mean individual growth rates (n = 51) ranged from to 7.08 cm year -1 (Fig. 2b). Fastest growth rates were found in turtles measuring cm CCL and the slowest growth rates were recorded for hawksbills with CCL of cm (Fig. 3). Capture per unit effort An average of five people were part of each snorkel team during each monitoring campaign, performing a total of 59 snorkeling sessions (Table 3). Total survey time was 58.8 h. Mean capture per unit effort was 0.92

29 Hawksbills at a major foraging ground in the eastern Pacific 601 Table 2. Number of hawksbill turtles captured for the first time and number of recaptures during each survey at CNP. C: capture, RC: recapture. *All but one recapture occurred at the site of original capture. Site September March September January July September Total C C RC C RC C RC C RC C RC C RC Playa Blanca Canales de Afuera Bahía Rosario Granito de Oro Del Centro Bahía Gambute Ranchería El María La Isla Uva Central Isla Brincanco Cocos Total a captures/person/hour during the six campaigns, with a maximum per-researcher capture rate of 3.29 captures/ hour. On average the first campaign showed the lowest CPUE = 0.21 ± 0.32 captures/person/ hour, whereas the sixth campaign showed the highest CPUE = 1.43 ± 1.28 captures/person/ hour. However, capture effort among campaigns was not significantly different (Kruskal- Wallis test, X 5 2 = 10.35, P = 0.066). The Playa Blanca and Granito de Oro sites both showed the highest CPUE during the sixth campaign, with 3.67 and 3.00 captures/person/hour, respectively. DISCUSSION b Figure 2. a) Hawksbill turtle size distribution at first capture and b) growth rate by size class of hawksbills recaptured one or more times at CNP marine reserve. This study represents the most comprehensive mark and recapture effort to date for hawksbill turtles at foraging grounds in the eastern Pacific. While we documented more individuals compared to previous studies in the eastern Pacific (Chacón-Chaverri et al., 2014; Heidemeyer et al., 2014; Tobón-López & Amorocho, 2014), we did so during only six monitoring campaigns, highlighting the high presence of hawksbill turtles at CNP. We continued to document new individual hawksbills with each campaign, indicating the overall population using the area may be in the thousands. The majority of hawksbills were juveniles of the smallest size class, highlighting the importance of the area as a nursery ground. Although satellite tracking has gained considerable attention over the last decade to determine marine turtle residency and habitat use, flipper tagging remains a more financially feasible tool for monitoring multiple individuals of a population and can provide data una-

30 602 Latin American Journal of Aquatic Research Table 3. Curved carapace length (CCL) and width (CCW) of hawksbill turtles captured and flipper tagged at each study site in CNP. Values are mean at first capture ± SD. Figure 3. Distribution of hawksbill growth rates at CNP marine reserve. ttainable via satellite technologies (Hart et al., 2015). For eastern Pacific hawksbills, flipper tagging is particularly useful for population monitoring due to their small home ranges (Gaos et al., 2012a) and this is supported by our study, where all but one of the recaptured individuals were recaptured at their original capture site. The strong fidelity to foraging areas is punctuated by the fact that many of the sites are located within only a few km from one another, yet turtles generally were recaptured at the same site as their initial capture (Fig. 2b). In the case of the single individual that changed locations, it movements covered >20 km, demonstrating the site fidelity is not absolute. Small home ranges have previously been described for juveniles (Carrión-Cortez et al., 2013), but even adult eastern Pacific hawksbills have some of the shortest migration movements of any sea turtle (Gaos et al., 2012a). Turtles were largest at Playa Blanca (CCL: 51 cm) and smallest at Isla Brincanco (CCL: 35 cm). However, adult female hawksbills in the eastern Pacific do move greater distances during their inter-nesting period (Gaos et al., 2012a) which may account for us not recapturing the six turtles with CCL >70 cm (range: cm) that were possibly visiting the area between nesting events rather than being residents of Coiba. Alternatively, the low adult recapture rate may also be because adult hawksbills only visit CNP as a stopover during longer migrations to areas outside of the Park. Although CNP represents an important hotspot for hawksbill turtles in the eastern Pacific, we did not find any individuals that had been previously marked in other foraging/nesting areas, despite tagging programs being carried out at nesting beaches and foraging grounds in various neighboring countries (e.g., Altamirano, 2014; Chacón-Chaverri et al., 2014; Tobón-López & Amorocho, 2014; Heidemeyer et al., Site CCL CCW N Playa Blanca 49.9 ± ± Canales de Afuera 43.7 ± ± Granito de Oro 44.8 ± ± Bahía Rosario 44.0 ± ± Del Centro 50.4 ± ± Bahía Gambute 39.5 ± ± Isla Uva 42.2 ± ± Central 43.5 ± El María 44.1 ± ± La ± ± Cocos Isla Brincanco 35.0 ND 1 Ranchería 41.6 ± ± ; Liles et al., 2015). Anecdotal and confirmed reports of limited hawksbill nesting at various beaches along continental Panama, including Playa Malena and Mata Oscura (J. Rodriguez, pers. comm.), which are located along the south coast of Veraguas province and where nesting beach monitoring is carried out by local community groups and NGOs (D. Pinto, pers. comm.). Despite limited attempts to evaluate nesting during our in-water monitoring campaigns, we were unable to confirm hawksbill nesting at beaches within CNP. However, CNP is an archipelago composed of 39 small islands and multiple beaches that are conducive to hawksbill nesting. We therefore recommend a full survey of potential nesting beaches during the putative nesting season (June/July; Gaos et al., 2017). The diverse range of marine and coastal habitats present at Coiba Islands include coral reefs, seagrasses and several mangrove estuaries (ANAM, 2009). The coral reefs cover approximately 1700 ha and are in good condition due to the protected status of the park s Marine Protected Area (MPA) and previous history as a penal colony, where fishing vessels were prohibited from coming near the island. During our study, all monitoring activities were conducted in coral reef habitats. Coral reefs are known to be the primary habitat for hawksbill turtles worldwide (Meylan & Donnelly, 1999; Wood et al., 2013; Reising et al., 2015), and hawksbills are believed to play an important role in maintaining the health of these systems (Leon & Bjorndal, 2002). Hawksbills have also been identified foraging on coral reefs in other parts of the eastern Pacific (Carrion-Cortez et al., 2013; Heidemeyer et al., 2014; Chacón-Chaverri et al., 2014). However, in the eastern Pacific the hawksbill turtle is renowned for nesting and foraging within mangrove estuaries (Gaos et al., 2012b; Liles et al., 2015). Whether hawksbills

31 Hawksbills at a major foraging ground in the eastern Pacific 603 use mangrove estuaries in CNP for nesting or foraging remains unknown as we were unable to monitor these systems during the study timeframe. In-water monitoring in mangrove estuaries of CNP is complicated by the presence of crocodiles (Crocodylus acutus) within these habitats. Of note however, is that the largest mangrove stands on the main Coiba island can be found on the east coast, where the primary coral reefs are also located (Fig. 1), thus the possibility that hawksbill utilize both habitats is likely. It is largely assumed that hawksbill turtles pass their first years of life in oceanic habitats (Reich et al., 2007), during a stage commonly referred to as the lost years (Carr, 1987), before recruiting to neritic habitats at a size of 20 to 35 cm CCL (Witzell, 1983). However, recent research suggests hawksbills in the eastern Pacific may lack pelagic phase during early posthatchling development (Liles et al., 2017; Gaos et al., 2017). Our research suggests that the Coiba archipelago is a recruitment site for young juvenile hawksbills, whether they originate from pelagic or other neritic ocean systems, as the smallest turtle tagged during this study was 30.0 cm CCL Growth rate data for hawksbill turtles between 1980 and 2013 from the West Atlantic indicated a mean annual growth rate of 3.1 ± 2.3 cm year -1 (Bjorndal et al., 2016). This is greater than the mean growth rate we documented of 2.8 cm year -1 for hawksbills in CNP. However, individual growth rates in CNP ranged from to 7.08 cm year -1, which is within the range found in the west Atlantic population (-2.1 to 22.6 cm year -1 ; Bjorndal et al., 2016), and the smaller annual mean may be a result of the comparatively short timeframe of our study. However, growth rates can also differ between populations of conspecifics as a result of habitat and food availability within foraging areas. Hawksbills in the west Pacific have peak growth rates later in life (60 SCL, Bjorndal & Bolten, 2010) than their Atlantic counterparts (around 35 cm SCL) (Chaloupka & Limpus, 2001), which is closer to that of turtles in CNP, where the size distribution cm CCL had the highest growth rate (3.6 cm year -1 ). Until now, the vast majority of our monitoring efforts in CNP have focused on the east coast of the main island (i.e., Coiba Island) and the smaller outlying islands towards the mainland (Fig. 2b), primarily due to accessibility. The west coast of Coiba Island has yet to be monitored for hawksbills due to high wind exposure, which creates difficult conditions for boats and monitoring teams. Nonetheless, it is important that future monitoring activities include this portion of the archipelago, which likely also hosts hawksbill habitats, and doing so will be important to understanding the full significance of CNP for the species. CPUE increased during our study and may have resulted from the experience gained by team members in accessing habitats and in capturing turtles with each successive monitoring campaign. Turtles were most easily captured at Playa Blanca and Granito de Oro, making them priority sites for continued monitoring if financial resources become limited. Granito de Oro is a popular snorkeling site in CNP for tourists coming over for day trips from the mainland, which may have desensitized hawksbills to human presence. While this may facilitate monitoring of the species, considering the high value of tortoiseshell (Mortimer & Donnelly, 2008), this could also facilitate poaching of the species, thus management measures should be taken to ensure hawksbills turtles and their habitat remain secure. ACKNOWLEDGEMENTS We thank the Ministry of Environment of Panama and the Scientific Committee of CNP for the scientific permit SEA/A , and entrance permits to the park station in CNP. Special thanks to Conservation International in Panama and MarViva Foundation for providing supplemental financial support for monitoring expeditions. Finally, we are grateful to all the personal in CNP including rangers and the directive board. We especially acknowledge Mali Mali for his guidance and insights during this project. REFERENCES Alfaro-Shigueto, J., J.C. Mangel, C. Caceres, J.A. Seminoff, A.R. Gaos & I.L. Yañez Hawksbill Turtles in Peruvian Coastal Fisheries. Mar. Turtle Newslett., 129: Altamirano, E Informe del Proyecto de Conservación de tortuga Carey (Eretmochelys imbricata) en la RN Estero Padre Ramos, Nicaragua. Temporada Fauna y Flora International & EP Hawksbill Initiative. Managua Techn. Rep., 34 pp. Autoridad Nacional del Ambiente (ANAM) Plan de Manejo del Parque Nacional Coiba J.L. Maté, D. Tovar, E. Arcia & Y. Hidalgo (comp.). STRI, 168 pp. Bjorndal, K.A. & A.B. Bolten Hawksbill sea turtles in seagrass pastures: success in a peripheral habitat. Mar. Biol., 157: Bjorndal, K.A., M. Chaloupka, V.S. Saba, C.E. Diez, R.P. van Dam, B.H. Krueger, J.A. Horrocks, A.J. Santos, C. Bellina, M.A. Marcovaldi & M. Nava Somatic growth dynamics of west Atlantic hawksbill sea turtles: a spatio temporal perspective. Ecosphere, 7(5). Blumenthal, J.M., T.J. Austin, C.D.L. Bell, J.B. Bothwell, A.C. Broderick, G. Ebanks Petrie, J.A. Gibb, K.E.

32 604 Latin American Journal of Aquatic Research Luke, J.R. Olynik, M.F. Orr, J.L. Solomon & B.J. Godley Ecology of hawksbill turtles Eretmochelys imbricata on a western Caribbean foraging ground. Chelonian Conserv. Biol., 8: Brittain, R., S. Handy & S. Lucas Two reports of juvenile hawksbill sea turtles (Eretmochelys imbricata) on the southeast coast of Guatemala. Mar. Turtle Newslett., 133: Burkholder, D.A., M.R. Heithaus, J.A. Thomson & J.W. Fourqurean Diversity in trophic interactions of green sea turtles Chelonia mydas on a relatively pristine coastal foraging ground. Mar. Ecol. Prog. Ser., 439: Carrión-Cortez, J., C. Canales-Cerro, R. Arauz, R. Riosmena-Rodríguez Habitat use and diet of juvenile eastern pacific Hawksbill Turtles (Eretmochelys imbricata) in the North Pacific Coast of Costa Rica. Chelonian Conserv. Biol., 12(2): Carr, A. 1987, New perspectives on the pelagic stage of sea turtle development. Conserv. Biol., 1: Chacón-Chaverri, D., D.A. Martínez-Cascante, D. Rojas & L.G. Fonseca Golfo Dulce, Costa Rica, un área importante de alimentación para la tortuga carey del Pacífico Oriental (Eretmochelys imbricata). Rev. Biol. Trop., 63: Chaloupka, M. & C. Limpus Trends in the abundance of sea turtles resident in southern Great Barrier Reef waters. Biol. Conserv., 102(3): Cortés, J Biology and geology of coral reefs of the eastern Pacific. Coral Reefs, 16: S39-S46. Diez, C.E. & R.P. van Dam Habitat effect on hawksbill turtle growth rates on feeding grounds at Mona and Monito Islands, Puerto Rico. Mar. Ecol. Prog. Ser., 234: Dutton, P.H Methods for collection and preservation of samples for marine turtle genetic studies. In: B.W. Bown & W.N. Witzell (eds.). Proceedings of the International Symposium on marine turtle conservation genetics. NOAA Tech Memo NMFS-SEFSC- 396, National Technical Information Service, Springfield, pp Eguchi, T., J.A. Seminoff, R.L. LeRoux, P.H. Dutton & D.L. Dutton Abundance and survival rates of green turtles in an urban environment: coexistence of humans and an endangered species. Mar. Biol., 157: Gaos, A.R., R.L. Lewison, B.P. Wallace, I.L. Yanez, M.J. Liles, W.J. Nichols, A. Baquero, C.R. Hasbun, M. Vasquez, J. Urteaga & J.A. Seminoff. 2012a. Spatial ecology of critically endangered hawksbill turtles Eretmochelys imbricata: implications for management and conservation. Mar. Ecol. Prog. Ser., 450: Gaos, A.R., R.L. Lewison, I.L. Yañez, B.P. Wallace, M.J. Liles, W.J. Nichols, A. Baquero, C.R. Hasbún, M. Vasquez, J. Urteaga & J.A. Seminoff. 2012b. Shifting the life-history paradigm: discovery of novel habitat use by hawksbill turtles. Biol. Lett., 8: Gaos, A., M. Liles, V. Gadea, A. Peña de Niz, F. Vallejo, M. Cristina, J. Darquea, A. Henriquez, A. Rivera, S. Chavarria, D. Melero, J. Urteaga, C. Pacheco, D. Chacon, C. LeMarie, J. Alfaro-Sigueto, J. Mangle, I. Yañez & J. Seminoff Living on the Edge; Hawksbill Turtle Nesting and Conservation Along the Eastern Pacific Rim. Lat. Am. J. Aquat. Res., 45(3): Gaos, A.R., A. Abreu, J.A. Alfaro, D. Amorocho, R. Arauz, A. Baquero, R. Briseño, D. Chacón, C. Dueñas, C. Hasbún, M. Liles, G. Mariona, C. Muccio, J.P. Muñoz, W.J. Nichols, J.A. Seminoff, M. Vásquez, J. Urteaga, B. Wallace, I.L. Yañez & P. Zárate Signs of hope in the eastern Pacific: international collaboration reveals encouraging status for a severely depleted population of hawksbill turtles Eretmochelys imbricata. Oryx, 44: Glynn, P.W Assessment of the present health of coral reefs in the eastern Pacific. In: R.W. Grigg & C. Birkeland (eds.). Status of coral reefs in the Pacific Ocean. UNIHI Sea Grant, CP-98-01, Hawaii, pp Groombridge, B. & R. Luxmoore The green turtle and hawksbill (Reptilia: Cheloniidae): world status, exploitation and trade. United Nations Environment Programme, Secretariat of the Convention on International Trade in Endangered Species of Wild Fauna and Flora, Cambridge, 601 pp. Guzman, H.M., C.A. Guevara & O. Breedy Distribution, diversity, and conservation of coral reefs and coral communities in the largest marine protected area of Pacific Panama (Coiba Island). Environ. Conserv., 31(2): Hart, C.E., G.S. Blanco, M.S. Coyne, C. Delgado-Trejo, B.J. Godley, T.T. Jones, A. Resendiz, J.A. Seminoff, M.J. Witt & W.J. Nichols Multinational tagging efforts illustrate regional scale of distribution and threats for East Pacific green turtles (Chelonia mydas agassizii). PLoS ONE 10(2): e doi: / journal.pone Hawkes, L.A., A. McGowan, A.C. Broderick, S. Gore, D. Wheatley, J. White, M.J. Witt & B.J. Godley High rates of growth recorded for hawksbill sea turtles in Anegada, British Virgin Islands. Ecol. Evol., 4: Heidemeyer, M., R. Arauz-Vargas & E. Lopez-Aguero New foraging grounds for hawksbill (Eretmochelys imbricata) and green turtles (Chelonia mydas) along the northern Pacific coast of Costa Rica, Central America. Rev. Biol. Trop., 62:

33 Hawksbills at a major foraging ground in the eastern Pacific 605 Leon, Y.M. & K.A. Bjorndal Selective feeding in the hawksbill turtle, an important predator in coral reef ecosystems. Mar. Ecol. Prog. Ser., 245: León Y.M. & C.E. Diez Population structure of hawksbill turtles on a foraging ground in the Dominican Republic. Chelonian Conserv. Biol., 3: Liles, M.J., A.R. Gaos, A.D. Bolaños, W.A. Lopez, R. Arauz, V. Gadea, J. Urteaga, I.L. Yañez, C.M. Pacheco, J.A. Seminoff1 & M.J. Peterson Survival on the rocks: high bycatch in lobster gillnet fisheries threatens hawksbill turtles on rocky reefs along the Eastern Pacific coast of Central America. Lat. Am. J. Aquat. Res., 45(3): Liles, M.J., M.J. Peterson, J.A. Seminoff, E. Altamirano, A.V. Henríquez, A.R. Gaos, V. Gadea, J. Urteaga, P. Torres, B.P. Wallace & T.R. Peterson One size does not fit all: importance of adjusting conservation practices for endangered hawksbill turtles to address local nesting habitat needs in the EP Ocean. Biol. Conserv., 184: Manly, B.F.J Stage-structured populations, sampling, analysis and simulation. Chapman & Hall, New York, 187 pp. Meylan, A Spongivory in hawksbill turtles: a diet of glass. Science, 239: Meylan, A.B. & M. Donnelly Status justification for listing the hawksbill turtle (Eretmochelys imbricata) as critically endangered on the 1996 IUCN Red List of Threatened Animals. Chelonian Conserv. Biol., 3: Mortimer, J.A. & M. Donnelly Marine Turtle Specialist Group 2007 IUCN Red List Status Assessment Hawksbill Turtle (Eretmochelys imbricata), 121 pp. [http: Reviewed: 10 July National Research Council. (NRC) Assessment of Sea-Turtle Status and Trends: Integrating Demography and Abundance. The National Acadamies Press. Washington, D.C., 174 pp. Quiñones, J., J. Zeballos, S. Quispe, L. Delgado Southernmost records of hawksbills turtles along the East Pacific Coast of South America. Mar. Turtle Newslett., 130: R Core Team R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL [ Rees, A.F., J. Alfaro-Shigueto, P.C.R. Barata, K.A. Bjorndal, A.B. Bolten, J. Bourjea, A.C. Broderick, et al Are we working towards global research priorities for management and conservation of sea turtles? Endang. Species Res., 31: Reich, K.J., K.A. Bjorndal & A.B. Bolten The lost years of green turtles: using stable isotopes to study cryptic life stages. Biol. Lett., 3: Reising, M., M. Salmon & S. Stapleton Hawksbill nest site selection affects hatchling survival at a rookery in Antigua, West Indies. Endang. Species Res., 29: Ruiz, A. & J. Rodríguez Caracterización de playas de anidación de tortugas marinas en el Parque Nacional Coiba, provincia de Veraguas, Panamá. Conservación Internacional, 39 pp. Shattuck, E.F Geographic origins of illegally harvested hawksbill sea turtle products. MS Thesis. Michigan State University, Michigan, 63 pp. Tobón-López, A. & D.F. Amorocho Population study of the hawksbill turtle Eretmochelys imbricata (Cheloniidae) in the southern Pacific region of Colombia. Acta Biol. Colomb., 19(3): Van Houtan, K.S., D.L. Francke, S. Alessi, T.T. Jones, S.L. Martin, L. Kurpita, C.S. King & R.W. Baird The developmental biogeography of hawksbill sea turtles in the North Pacific. Ecol. Evol., 6: Wallace, B.P., A.D. DiMatteo, B.J. Hurley, E.M. Finkbeiner, A.B. Bolten, M.Y. Chaloupka, et al Regional Management Units for marine turtles: a novel framework for prioritizing conservation and research across multiple scales. PLoS ONE 5: e15465 doi: /journal.pone Witzell, W.N Synopsis of biological data on the hawksbill turtle Eretmochelys imbricata (Linnaeus, 1766). FAO Fish. Synop., 137: 78 pp. Wood, L.D., R. Hardy, P.A. Meylan & A.B. Meylan Characterization of a hawksbill turtle (Eretmochelys imbricata) foraging aggregation in a highlatitude reef community in southeastern Florida, USA. Herpetol. Conserv. Biol., 8: Received: 14 March 2017; Accepted: 16 Abril 2017

34 Lat. Am. J. Aquat. Res., 45(3): , 2017 Feeding ecology of Chelonia mydas in Northern Peru 585 Sea Turtle Research and Conservation in Latin America Jeffrey Mangel, Jeffrey Seminoff, Bryan Wallace & Ximena Vélez-Suazo (Guest Editors) DOI: /vol45-issue3-fulltext-8 Research Article Feeding ecology of the green turtle Chelonia mydas in northern Peru Astrid Jiménez 1, Sergio Pingo 1, Joanna Alfaro-Shigueto 1,2,3 Jeffrey C. Mangel 1 & Yuri Hooker 4 1 ProDelphinus, Miraflores, Lima, Perú 2 Centre for Ecology and Conservation, University of Exeter, Penryn, Cornwall, UK 3 Facultad de Biología Marina, Universidad Científica del Sur, Panamericana, Lima, Perú 4 Laboratorio de Biología Marina, Universidad Peruana Cayetano Heredia, Lima, Perú Corresponding autor: Astrid Jiménez (astrid@prodelphinus.org) ABSTRACT. Diet and food preferences of the green turtle Chelonia mydas were analyzed based on digestive tract contents of dead specimens caught incidentally by an artisanal gillnet fishery in Sechura Bay, northern Peru. We examined 27 digestive tracts and identified 35 prey items. The sampled turtles were all juveniles (CCL = 53.7 ± 1.2 cm, range cm). The prey items were grouped into six categories: algae, cnidarians, mollusks, arthropods, chordates and garbage/anthropogenic debris. The items with the highest Frequency of Occurrence values (% FO) were: Caulerpa filiformis (77.8%), Loligo gahi (eggs) (51.9%) and Rhodymenia corallina (44.4%). By weight (% W), the most important items, were L. gahi (eggs) (33.3%), Stomolophus sp. (7.3%) and Aphos porosus (6.5%). According to the Preponderance Index (%IP), the preponderant item was L. gahi (eggs) with 6.1% and 61.2% during winter-spring and summer-autumn, respectively. According to the Resultant Weight index (Rw) of wet items, the most important items were: C. filiformis (13.1%), L. gahi (eggs) (10.5%), R. corallina (7.4%), plastic (7.5%), Gigartina chamissoi (5.1%). Garbage/anthropogenic debris was common in the digestive tracts analyzed. Plastic items had a frequency of occurrence of 44.4%. A greater diversity of food items was observed during summer and autumn. This study shows that juvenile C. mydas forage on a variety of resources. We recommend that conservation plans, land use planning and future management plans in the Sechura Bay include green turtles as a sentinel species for monitoring biodiversity of marine resources and the degree of pollution in the Bay. Keywords: diet, Chelonia mydas, Sechura, Peru, small-scale fisheries. INTRODUCTION From their emergence as hatchlings through to their adult life, sea turtles experience ontogenetic changes in habitat use that includes nesting beaches and juvenile and adult feeding areas (FAO, 2011). Currently, multiple sea turtle populations are declining (Amorocho & Reina, 2007; Carrión-Cortez et al., 2010), due largely to anthropogenic activities including commercial overfishing, bycatch, destruction of critical habitat for feeding and nesting; and most recently, pollution (Gilman et al., 2006, 2010; Boyle & Limpus, 2008; Rodríguez, 2010). The green turtle (Chelonia mydas) is a circumglobal species that is susceptible to overexploitation as a food resource, incidental mortality in fisheries (Alfaro- Shigueto et al., 2002), and coastal foraging habitat degradation, all of which have contributed to its listing as Endangered on the IUCN Red List (Lemons et al., 2011). The species is found year-round in shallow waters of coasts, bays and lagoons (Bjorndal, 1980; Plotkin et al., 1993), which are important habitats for growth and development (Musick & Limpus, 1997). C. mydas is distributed along the west coasts of North and South America (Marquez, 1990; Aranda & Chandler, 1989). Marquez (1990) reported coastal primary foraging areas from San Diego Bay, USA to Mejillones, Chile and more recently, the species southward distribution was extended to Valparaíso, Chile (Troncoso-Fierro & Urbina-Burgos, 2007). Neritic habitats in Peru also form important feeding areas for the species (Alfaro-Shigueto et al., 2002, 2004; Velez- Zuazo et al., 2014). The feeding ecology of this species has been studied throughout most of its distribution (Quiñones et al., 2015a). At their juvenile developmental habitats, green turtles can take on a herbivorous diet, feeding primarily on seagrass and algae (Plotkin et al., 1993; Seminoff et al., 2002a, 2003; López-Mendilaharsu et al., 2005; Sampson & Giraldo, 2014). Other studies have noted

35 586 Latin American Journal of Aquatic Research that C. mydas juveniles can also be carnivorous consumers, foraging mainly on cnidarians (Carrión- Cortez et al., 2010), sponges (Seminoff et al., 2002b), and tunicates (Amorocho & Reina, 2007, 2008). Additional studies have also identified an omnivorous diet, based mainly on algae but with an additional wide variety of animals, such as crustaceans, fish eggs, mollusks, and, to a lesser extent, jellyfish (Hays-Brown & Brown, 1982; De Paz et al., 2004; Kelez et al., 2004; Santillán, 2008; Quiñones et al., 2015a, 2015b). The diet composition at different feeding grounds depends on the availability of resources (Bjorndal, 1980), and to some extent, by foraging preferences, especially for certain species of algae (López- Mendilaharsu et al., 2008). This feeding strategy, where juveniles are herbivorous/omnivorous, allows for accelerated growth of recruits (Quiñones et al., 2015a). In the Eastern Pacific Ocean (EPO), there is baseline diet information for some green turtle feeding grounds, including those located in the Baja California Peninsula (Mexico) (Seminoff et al., 2002a, 2003; López-Mendilaharsu et al., 2005), the Galapagos Islands (Seminoff et al., 2002b; Carrión-Cortez et al., 2010), and the Gorgona Island (Colombia) (Amorocho & Reina, 2007, 2008). In Peru, C. mydas feeding grounds have been identified in Tumbes (Rosales et al., 2010), El Ñuro (Velez-Zuazo et al., 2014), Sechura Bay and Virrilá estuary (Santillán, 2008; De Paz & Alfaro-Shigueto, 2008; Paredes et al., 2015), Isla Lobos de Tierra (Quiñones et al., 2015a), Chimbote (Alfaro-Shigueto et al., 2004), Pisco (Hays-Brown & Brown, 1982; Quiñones et al., 2010), and the Paracas Bay area (Hays- Brown and Brown, 1982; De Paz et al., 2007; Paredes, 2015; Quiñones et al., 2010; Quiñones et al., 2015a, 2015b). Virrilá estuary and Sechura Bay are known to be important aggregation areas for juvenile and subadult green turtles (Santillán, 2008; Alfaro-Shigueto et al., 2011; Paredes et al., 2015; Ortiz et al., 2016). In addition to marine algae and seagrasses (Hays-Brown & Brown, 1982; Alfaro-Shigueto et al., 2004; Seminoff et al., 2002b; Amorocho & Reina, 2008), at many of these sites, green turtles consume large quantities of invertebrates (like scyphozoan jellyfish and sessile anemones) (Quiñones et al., 2010; Paredes, 2015), fish and mollusks. Sechura Bay is also an important and traditional site of small-scale fishing (Alfaro-Shigueto et al., 2010; Morón et al., 2013) for demersal and benthic resources. More recently, the bay has seen the development of mariculture, mainly of Argopecten purpuratus (Mendo et al., 2008; Mendo, 2011). Small-scale fisheries (SSF) are often poorly managed and also have generated environmental problems such as pollution (IMARPE, 2011), and impacts on endangered species, including sea turtles (Alfaro-Shigueto et al., 2011; Ortiz et al., 2016). The purpose of this research was to study the diet composition of juvenile and subadult green turtles in this bay, an area of particular importance as a foraging ground for the species in northern Peru, but also an area important for small-scale fisheries. MATERIALS AND METHODS Study area Sechura Bay (05 12 to S and to W) is delimited in the north by Punta Gobernador and Punta Aguja to the south, has an approximate extension of 89 km 2 (IMARPE, 2007; Morón et al., 2013), and is within the Piura Region (Fig. 1). Turtle bycatch and sizes of individuals The study was conducted between July 2013 and June 2014 as part of a program to monitor the small-scale fleet operating from the port of Constante (05 35 S, W). Fishing vessels ranged in length from 6 to 10 m and each trip consisted of setting of bottom set gillnets. Gillnets were made of multifilament twine and were composed of multiple net panes that measured 56.4 m long by 2.8 m high, with a stretched mesh of approximately 24 cm. Nets were typically deployed in the late afternoon, soaked overnight and retrieved the following morning (Alfaro-Shigueto et al., 2010; Ortiz et al., 2016). We collected digestive tract contents from dead specimens of C. mydas incidentally caught in this fishery. Each turtle was measured for the curved carapace length (CCL cm) from the anterior notch to the tip of the posterior-most marginal scutes (Bolten, 2000; Wyneken, 2001). Individuals with CCL < 69 cm. were considered as juveniles, individuals with 69 < CCL < 85 cm. were considered as sub-adults, and individuals with CCL > 85 cm. as adults (Zárate et al., 2013). Diet study All digestive tract contents were collected and stored in 10% formaldehyde in seawater solution (Jacobson, 2000; Work, 2000; Wyneken, 2001). The identification of categories and items was made with the help of identification guides reaching the lowest possible taxonomic level. Prey species identification guides used included Chirichigno (1974), Mendez (1981), Alamo & Valdivieso (1997), Acleto & Zúñiga (1998), Chirichigno & Cornejo (2011), Moscoso (2012, 2013), Tume et al. (2012) and Guiry & Guiry (2015).

36 Feeding ecology of Chelonia mydas in Northern Peru 587 Figure 1. Sechura Bay, Piura, Peru, showing the geographical location of individuals of C. mydas caught incidentally during (Seaturtle.org Maptool, 2002). For all consumed items, we estimated the Frequency of Occurrence (%FO) and wet weight percentage (%W) (Carrión-Cortez, 2010), according to the following equation: Frequency of Occurrence (%FO) and wet weight percentage (%W) were combined to calculate the Preponderance Index (%IP) (Mohan & Sankaran, 1988). Additionally, diet data were grouped into two seasonal groups: summer-autumn (January to June), and winter-spring (July to December) (Paredes, 2015), toward establishing the order of numerical dominance of items within the diet, according to the following: where: Pi = total wet weight of diet item in all samples; Pt = total wet weight of all samples; Ni = number of samples containing diet item; Nt = total number of samples. where: %FO = frequency of occurrence percentage; %W = weight percentage.

37 588 Latin American Journal of Aquatic Research To establish the order of importance for the entire array of foods ingested (Mohan & Sankaran, 1988; Carrión-Cortez, 2010), was also calculated the weighted resultant index (Rw), according to the following: where: %FO = frequency of occurrence percentage; %W = weight percentage. The Resultant weight index (Rw) can be graphically represented as a function of the θº angle. This index allows one to interpret the importance of each item considering the values of %W and %FO. Items with a uniform representation of %W and %FO have angles close to 45º. The values of the Rw vary between 0 and 100. Items with values close to 100 represent the most important item in the diet (Mohan & Sankaran, 1988; Carrión-Cortez, 2010). RESULTS We examined 27 digestive tracts of C. mydas. The sampled specimens were 100% juveniles (CCL = 53.7 ± 1.2 cm, range cm) (Fig. 3). The largest number of digestive tracts collected were in the summer (January-March) and autumn (April-June), with a total of 12 and seven digestive tracts, respectively. Four digestive tracts were collected in both the winter (July- September) and spring (October-December). Thirty-five food items were identified and grouped into six categories: algae, cnidarians, mollusks, arthropods, chordates and garbage/anthropogenic debris (Table 1, Fig. 2). Of all items identified, five had the highest frequency of occurrence (% FO): Caulerpa filiformis thread-like algae (77.8%), Loligo gahi common squid (eggs) (51.9%), Rhodymenia corallina rose seaweed (44.4%), Gigartina chamissoi tongueweed (29.6%), and Ulva lactuca sea lettuce and Gelidium congestum jelly-weed (22.2% each). By weight (% W), green turtles consumed mainly five food items: L. gahi (eggs) (33.3%), Stomolophus sp. jellyfish (7.3%), Aphos porosus monkfish (6.5%), R. corallina (5.1%), and Sinum cymba (eggs) (4.6%). According the Preponderance Index (%IP), during winter-spring, the only preponderant item was L. gahi (eggs) (6.1%), while during summer-autumn the preponderant items were L. gahi (eggs) (61.2%), R. corallina (16.7%), G. chamissoi (12.8%), C. filiformis (11.5%) and Stomolophus sp. (10.5%) (Table 1). Prey items found varied seasonally (Table 2). The three most important items by season according to the weighted resulting index (Rw) were: L. gahi (eggs), Hexaplex sp. rock snail (egg capsules) and Hepatus chilensis crabs in winter; C. filiformis, L. gahi (eggs) and Cronius ruber blackpoint sculling crab in spring; C. filiformis, L. gahi (eggs) and R. corallina in summer; and L. gahi (eggs), R. corallina, G. chamissoi and Pseudosquillopsis lessoni mantis shrimp in autumn. Debris from anthropogenic activities, such as nylon monofilament, rope, and plastic bags, was common in the digestive tracts analyzed (56%). Within these garbage/anthropogenic debris items, plastic bags were the most common, with a FO of 44.4%. The degree of importance of preys consumed by C. mydas for the study period, according to the weighted resulting index (Rw) in decreasing order was C. filiformis, L. gahi (eggs), R. corallina, Plastic, G. chamissoi and U. lactuca (Fig. 4). DISCUSSION The green turtle uses the Peru coast mainly as a foraging ground (Hays-Brown & Brown, 1982; Santillán, 2008; Quiñones et al., 2010, 2015b; Paredes et al., 2015). Bays along the coast offer protection and food resources for juvenile green turtles (Hatase et al., 2006; Vander et al., 2013). Sechura Bay and Paracas Bay are two areas identified in past studies as green turtle foraging areas (De Paz & Alfaro-Shigueto, 2008; Santillán, 2008; Cáceres et al., 2013). These bays have similar characteristics, such as shallow water and abundant algae and seagrasses that provide habitat for invertebrates, fish and mollusks (Paredes, 2015). The size distribution of green turtles captured in Sechura Bay in this study corresponded to an immature population consisting of juveniles (100%). Our results suggest that this bay is likely to harbor juveniles almost exclusively. This may indicate that small green turtles spend more time in nearshore areas than larger sized individuals (Carrión-Cortez et al., 2010). Smaller turtles have higher relative energy demands than adults (Koch et al., 2007; Carrión-Cortez, 2010), and have been shown to prefer sheltered areas where net energy expenditure during foraging activities is less than in high-energy oceanic zones (Seminoff et al., 2003; Koch et al., 2007; Santillán, 2008). Research conducted in Baja California (Mexico) indicates that the green turtle has an herbivorous diet, feeding on red algae, green and seagrass (Seminoff et

38 Feeding ecology of Chelonia mydas in Northern Peru 589 Table 1. Frequency of occurrence (%FO), weight percentage (%W) and weighted resultant index (Rw) and preponderance index (%IP a = winter-spring; %IP b = summer-autumn) of prey groups recovered from digestive tracts of C. mydas caught from Sechura Bay, (n = 35). Category/ Phylum Item/Components N %FO %W Rw %IP a %IP b Codium peruvianum Rhodymenia corallina Ulva lactuca Plantae Caulerpa filiformis Eisenia cokeri Gelidium congestum Gigartina chamissoi Porphyra sp Cnidaria Stomolophus sp Mytilidae Aplysia sp Loligo gahi (eggs) Mollusca Loligo gahi (individuals) Hexaplex sp. (egg capsules) Tagelus peruvianus Octopus sp Sinum cymba (eggs) Portunidae Penaeidae Arthropoda Hepatus chilensis Acanthonyx petiverii Cronius ruber Pseudosquillopsis lessoni Actinopterygii Ascidiacea Odontesthes regia (eggs) Pyrosoma sp Chordata Engraulis ringens Ophichthus pacifici Aphos porosus Normanichthys crockeri Garbage / anthropogenic debris Urotrygon sp Plastic Feathers Rope al., 2002a, 2002b; Koch et al., 2007; López- Mendilaharsu et al., 2008). This trend toward algae and/or sea grasses has also been well documented in adult green turtles in the Caribbean (Bjorndal, 1980). On the Pacific coast of South America, direct observations in the Galapagos Islands of sub-adult and adult green turtles indicated that they predominantly fed on algae, including Ulva, Padina, Gellidium and Gracilaria spp. (Green, 1994). Sampson et al. (2013) found in esophageal lavages of C. mydas juveniles at Gorgona National Park that the most abundance items were Povillopora damicornis, rhodoliths, Cladophora sp. and algae mats. However, in coastal waters of Peru, Hays-Brown & Brown (1982) in Pisco, found a significant amount of animal prey items (mollusks, polychaetes, jellyfish, amphipods, sardines and anchovies) in the stomach contents of sub-adult and adult green turtles, in addition to algae. In San Andres (Pisco), Quiñones et al. (2010) found in the stomach and esophagus of C. mydas that jellyfish was the most consumed prey item, followed by mollusks and macroalgae. This feeding behavior is similar to the present study, where C. mydas shows an omnivorous diet, composed mainly of items of animal origin (68.6%).

39 590 Latin American Journal of Aquatic Research Figure 2. Proportion of items by taxonomic category (n = 35) present in 27 digestive tracts of C. mydas from Sechura Bay, Table 2. The five most important food items found in digestive tracts of C. mydas, Sechura Bay, , represented by the frequency of occurrence percentage (% FO), weight percentage (%W) and weighted resultant index (Rw) by seasons. N = number of digestive tracts by seasons. Season Items O %FO %W Rw Loligo gahi (eggs) Hexaplex sp. (eggs capsules) Winter Plastic N = 4 Hepatus chilensis Odontesthes regia (eggs) Spring N = 4 Summer N = 12 Autumn N = 7 Caulerpa filiformis Loligo gahi (eggs) Cronius ruber Gelidium congestum Plastic Caulerpa filiformis Loligo gahi (eggs) Plastic Rhodymenia corallina Ophichthus pacifici Loligo gahi (eggs) Rhodymenia corallina Gigartina chamissoi Pseudosquillopsis lessoni Caulerpa filiformis Early studies of the diet of C. mydas suggested macroalgae as their main prey item (Bjorndal, 1997), with the genus Codium and Rhodymenia extensively reported as diet components (López-Mendilaharsu et al., 2005; Rodríguez, 2010). In Paracas, De Paz et al., (2007) reported algae consumption of the genus Ulva in greater proportion. Our analysis reported algae consumption in all seasons, including Caulerpa filiformis, Rhodymenia corallina, Codium peruvianum and Ulva lactuca. The most important of these according to the Resultant Weighted Index was C. filiformis (Rw= 13.1%).

40 Feeding ecology of Chelonia mydas in Northern Peru 591 Figure 3. Size-classes of C. mydas captured in the Sechura Bay, The dotted lines represent the minimum and average size reported for green turtle adults in the Galapagos Islands, the largest nesting colony of C. mydas near Peru (Source: Galapagos Islands size data). Santillán (2008) analyzed 11 stomach contents from Sechura Bay and reported the highest values of frequency of occurrence for Gracilaria sp. (37.8%), C. filiformis (35.6%), Codium sp. (33.3%), and eggs of Loligo gahi (22.2%). Quiñones et al. (2010) found in 192 stomach and esophagus samples from San Andres (Pisco) that the most frequently consumed prey items were jellyfish (70.8%), mollusks (62%), crustaceans (47.4%) and macroalgae (37.5%). Our study also reports C. filiformis (74.2%) and L. gahi (eggs) (48.4%), but both at approximately twice the frequency reported by Santillán (2008). Alfaro-Shigueto et al. (2004) examined 11 stomach contents from Chimbote Bay, and in seven found fish eggs, squids, Engraulis ringens anchovy and brachiopods. Hays-Brown & Brown (1982) analyzed 39 stomach contents and reported the occurrence of fish in 23% (mainly sardines, anchovies and fish eggs). Santillán (2008) also reported the presence of engraulids, especially serranids and carangids. In the Virrilá estuary in Sechura Bay, Quiñones (Pers. comm.) found great percentages of fish like Mugil cephalus mullet and a smaller percentage of Anchoa nasus white anchovy in stomach contents of green turtles. Our results indicate a %FO of 59.3% for fishes. This includes both bony fishes like E. ringens, elasmobranchs of the genus Urotrygon round ray and fish eggs from Odontesthes regia silverside. This indicates that consumption of fish is important in the diet of C. mydas. Invertebrates comprised 43% of the total abundance of items in the stomachs analyzed here. Some invertebrates were found only once (Aplysia sp. sea slug and Tagelus peruvianus saltwater clams ) and did not warrant consideration as a major diet item (Rw = 0.85 and Rw = 0.62 respectively). Hays-Brown & Brown (1982) found crustaceans in five of 39 stomach contents examined, representing a %FO of 13%. Likewise, Alfaro-Shigueto et al. (2004) found crustaceans (Hyperia medusarum, Euphylax dovii) in four of 11 stomach contents analyzed. In our study, Crustacea was reported for 14 digestive tracts (51.9%), consisting mainly of Pseudosquillopsis lessoni and Portunidae. This represents FO values of 18.5% each, but only P. lessoni had a value as weight (3.5%). In another part of their investigation, Hays-Brown & Brown (1982) found mollusks (%FO = 64%), mainly Nassarius, Mytilus and Semele in 25 stomach contents. Alfaro-Shigueto et al. (2004), found Nudibranquia eggs, Aplysia sp., Sinum cymba, Chione sp., Natica sp., Nassarius grayi, Mactra sp., Semimytilus algosus and other mytilids in 10 of 11 stomach contents analyzed. In our research, mollusks were found in 18 digestive tracts (66.7%). High FO and weight values were found for eggs of L. gahi (51.9% and 33.3% respectively), followed by eggs of S. cymba (11.1% and 4.6% respectively). A study in Magdalena Bay (Mexico) reported changes in prey diversity in juvenile turtle diets which coincided with seasonal changes in vegetal biomass (López-Mendilaharsu et al., 2008). Similary, Santillán (2008) reported that the herbivorous diet would predominate during spring and summer, while the carnivorous diet would increase during autumn and winter. Our results show that green turtles consumed L. gahi (eggs) throughout the year but, mainly during summer-autumn (%IP = 61.2%). This could be due to their seasonally high abundance in Sechura Bay (IMARPE, 2007) which in turn could be explained by the L. gahi (eggs) spawning seasons which peak in spring or early of summer and autumn (Villegas, 2001). It has been suggested that the dominance of some food items over others could be related to increased algae abundance and that algae are the habitat of diverse organisms favoring their proliferation (Paredes, 2015; Gribben et al., 2009). Box (2008) stated that algae from the Caulerpa genus favors the growth of various organisms, mollusks in particular. Our work provides support for this assertion because C. filiformis and L. gahi (eggs) were mostly found in the same samples. Both items had high Resultant Weighted Index values (Rw) (C. filiformis = 13.1% and L. gahi (eggs) = 10.5%), and were the most important items in the C. mydas diet observed in our study. During winter-spring, according the Preponderance Index (%IP), the only preponderant item was L. gahi (eggs) (6.1%), while during summer-autumn the prepon-

41 592 Latin American Journal of Aquatic Research Figure 4. Weighted Resultant Index (Rw) plotted against the angle for food items in digestive tracts of C. mydas from Sechura Bay, Food items with uniform representation in both values, %W and %FO fall around the 45, whereas those with uneven representation spread on either side of the middle line (close to 0 : high %FO and less %W; close to 90 : high %W and less %FO). derant items were L. gahi (eggs) (61.2%), R. corallina (16.7%), G. chamissoi (12.8%), C. filiformis (11.5%) and Stomolophus sp. (10.5%). This could be due, because during winter-spring the green turtle individuals go somewhere close like Virrilá estuary to feed on great percentages of M. cephalus mullet, Ulva spp. and A. nasus "anchovy white" (Paredes, unpubl. data) whereas during summer-autumn go out of estuary to feed of mentioned items. In the present study, 55.6% of digestive tracts analyzed contained garbage and/or anthropogenic debris. The materials found include: plastic (bags/ packing) and rope the majority of which was found in the final portion of the intestine. Marine debris is a growing problem for wildlife. It has been documented to affect more than 267 species worldwide (Schuyler et al., 2012) and can have lethal and sub-lethal effects on sea turtles and other wildlife (Schuyler et al., 2013). The occurrence of debris from anthropogenic activities (especially plastic bags) in digestive tracts of C. mydas has been reported since the 1980 s (Schuyler et al., 2012). In Peru, a study from Chimbote reported plastic bags and traces of nylon in 91% of stomach contents analyzed (Alfaro-Shigueto et al., 2004). In Sechura, Santillán (2008) reported that green turtle stomach contents contained approximately 26.7% plastic debris. The high incidence of garbage and anthropogenic debris found in green turtles in our study may reflect the pollution of areas within the bay due to human presence and activities (e.g., coastal community proximity, mariculture area, fisheries). CONCLUSIONS AND RECOMMENDATIONS Our results highlight the importance of neritic habitats, especially bays, as key habitats for the development of juvenile and sub-adult green turtles in the Eastern Pacific Ocean. Juvenile of C. mydas in Sechura Bay had an omnivorous diet and foraged on a variety of resources, but mainly on animal prey items like mollusks, arthropods and chordates. They did not appear to focus on any particular prey species. Future research could include a detailed assessment of the composition of species in the benthic areas of Sechura Bay, including their spatial and temporal distributions. Future use and development of bay areas, including of Sechura Bay, should take into account the vital role these habitats play in the development of juvenile green turtles. We recommend conducting complementary studies to characterize and quantify marine debris and formulating management plans toward reducing plastics pollution in Sechura Bay. Species like the green turtle can also be used as sentinels for biodiversity and pollution within bays and coastal areas.

42 Feeding ecology of Chelonia mydas in Northern Peru 593 ACKNOWLEDGEMENTS We would like to thank the fishing communities who participated in this study. Our thanks also to biologists Armando Ugaz and Humberto Rivera (National University of Piura), David Montes (Ricardo Palma University) for their generous support during the project and for their help in identifying the samples collected. We also want to thank ProDelphinus staff for their help in data collection. This study was conducted with support from the National Fish and Wildlife Foundation (NFWF), Whitley Fund for Nature (WFN), Darwin Initiative - Sustainable Artisanal Fisheries Initiative in Peru (DEFRA), University of Exeter and ProDelphinus NGO. REFERENCES Acleto, C. & R. Zúñiga Introducción a las algas. Editorial Escuela Nueva, Lima, 383 pp. Álamo, V. & V. Valdivieso Lista sistemática de Moluscos Marinos del Perú. Instituto del Mar del Perú, Callao, 183 pp. Alfaro-Shigueto, J., M. Van Bressem, D. Montes & K. Onton Turtle mortality in fisheries off the coast of Peru. In: A. Mosier, A. Foley & B. Brost (eds.). Proceedings of the Twentieth Annual Symposium on Sea Turtle Biology and Conservation. NOAA Tech. Memo. NMFS-SEFSC-477, 369 pp. Alfaro-Shigueto, J., D. Montes, C. Acleto, R. Zuñiga & P. Huamán Diet analysis from green turtle Chelonia mydas agassizii from central Peruvian Coast. In: M. Coyne & R. Clark (eds.). Proceedings of the Twenty-first Annual Symposium on Sea Turtle Biology and Conservation. NOAA Technical Memorandum NMFS-SEFSC-528, 368 pp. Alfaro-Shigueto, J., J. Mangel, M. Pajuelo, P. Dutton, J. Seminoff & B. Godley Where small can have a large impact: Structure and characterization of smallscale fisheries in Peru. Fish. Res., 106: Alfaro-Shigueto, J., J. Mangel, F. Bernedo, P. Dutton, J. Seminoff & B. Godley Small-scale fisheries of Peru: a major sink for marine turtles in the Pacific. J. Appl. Ecol., 48: Amorocho, D. & R. Reina Feeding ecology of the East Pacific green sea turtle Chelonia mydas agassizii at Gorgona National Park, Colombia. Endang. Species Res., 3: Amorocho, D. & R. Reina Intake passage time, digesta composition and digestibility in East Pacific green turtles (Chelonia mydas agassizii) at Gorgona National Park, Colombian Pacific. J. Exp. Mar. Biol. Ecol., 360: Aranda, C. & M. Chandler Las tortugas marinas del Perú y su situación actual. Bol. Lima, 62: Bjorndal, K Nutrition and grazing behavior of the green turtle Chelonia mydas. Mar. Biol., 56: Bjorndal, K Foraging ecology and nutrition of sea turtles. In: P. Lutz & J. Musick (eds.). The biology of sea turtles. CRC Press, Boca Raton, pp Bolten, A Técnicas para la medición de tortugas marinas. In: K. Eckert, K. Bjorndal, F. Abreu-Grobois & M. Donnelly (eds.). Técnicas de investigación y manejo para la conservación de las tortugas marinas. Traducida al español. UICN/CSE Grupo Especialista en Tortugas Marinas, pp Boyle, M. & C. Limpus The stomach contents of post-hatchling green and loggerhead sea turtles in the southwest Pacific: an insight into habitat association. Mar Biol., 155: Box, A Ecología de caulerpales: fauna y biomarcadores. Tesis Doctoral, Instituto Mediterráneo de Estudios Avanzados, Universidad Islas Baleares, Palma, 335 pp. Cáceres, C., J. Alfaro-Shigueto & J. Mangel Green turtle captured in net fisheries in the Port of Constante, Peru. In: J. Blumenthal, A. Panagopoulou & A. Rees (eds.). Proceedings of the Thirtieth Annual Symposium on Sea Turtle Biology and Conservation. NOAA Tech Memo NMFS-SEFSC-640, 177 pp. Carrión-Cortez, J., P. Zárate & J. Seminoff Feeding ecology of the green sea turtle (Chelonia mydas) in the Galapagos Islands. Mar. Biol. Assoc. UK, 1: 1-9. Chirichigno, N Clave para identificar los peces marinos del Perú. Informe Instituto del Mar del Perú, Callao, Informe 44, 108 pp. Chirichigno, N. & R. Cornejo Catálogo comentado de los peces marinos del Perú. Instituto del Mar del Perú, Publicación especial, Callao, 314 pp. De Paz, N., J. Reyes & M. Echegaray Capture and trade of marine turtles at San Andres, Southern Peru. In: M. Coyne & R. Clark (eds.). Proceedings of the Twenty-First Annual Symposium on Sea Turtle Biology and Conservation. NOAA Technical Memorandum NMFS SEFSC-528: pp De Paz, N. & J. Alfaro-Shigueto Foraging grounds for sea turtles in inshore Peruvian waters. In: H. Kalb, A. Rohde, K. Gayheart & K. Shanker (eds.). Proceedings of the Twenty-Fifth Annual Symposium on Sea Turtle Biology and Conservation. NOAA Technical Memorandum NMFS-SEFSC-582, 204 pp. De Paz, N., J. Reyes, M. Echegaray, M. Ormeño & H. Anchante Identificación y manejo de hábitats críticos de tortugas marinas en Perú: Paracas, estudio de caso. In: C. Guerra-Correa, A. Fallabrino, P. Bolados-

43 594 Latin American Journal of Aquatic Research Díaz & C. Turner (eds.). VII Simposio sobre Medio Ambiente. Estado actual y perspectivas de la investigación y conservación de las tortugas marinas en las costas del Pacífico Sur-Oriental. [ f]. Reviewed: 10 September Food and Agriculture Organization (FAO) Directrices para reducir la mortalidad de las tortugas marinas en las operaciones de pesca. Food and Agriculture Organization, Roma, 132 pp. Gilman, E., E. Zolett, S. Beverly, H. Nakano, K. Davis, D. Shiode, P. Dalzell & I. Kinan Reducing sea turtle by-catch in pelagic longline fisheries. Fish. Fish., 7: Gilman, E., J. Gearhart, B. Price, S. Eckert, H. Milliken, J. Wang, Y. Swimmer, D. Shiode, O. Abe, S. Peckham, M. Chaloupka, M. Hall, J. Mangel, J. Alfaro-Shigueto, P. Dalzell & A. Ishizaki Mitigating sea turtle bycatch in coastal passive net fisheries. Fish. Fish, 11: Green, D Galapagos sea turtles: an overview. In: B. Schroeder & D. Witherington (eds.). Proceedings of the Thirteenth Annual Symposium on Sea Turtle Biology and Conservation. NOAA Technical Memorandum NMFSC-SEFSC-341, Jekyll Island, GA, pp Gribben, P., J. Byers, M. Clements, L. McKenzie, P. Steinberg & J. Wright Behavioural interactions between ecosystem engineers control community species richness. Ecol. Lett., 12(11): Guiry, M. & G. Guiry Algae Base. World-wide electronic publication, National University of Ireland, Galway. [ Reviewed: 15 September Hays-Brown, C. & W. Brown Status of sea turtles in the South Eastern Pacific: emphasis on Peru. In: K. Bjorndal (ed.). Biology and conservation of sea turtles. Smithsonian Institution Press, Washington DC, pp Hatase, H., K. Sato, M. Yamaguchi, K. Takahashi & K. Tsukamoto Individual variation in feeding habitat use by adult female green sea turtles (Chelonia mydas): are they obligately neritic herbivores? Oecologia, 149(1): Instituto del Mar del Perú (IMARPE) Estudio de línea base del ámbito marino de la Bahía de Sechura, Perú. Instituto del Mar del Perú, Callao, 116 pp. Instituto del Mar del Perú (IMARPE) Informe nacional sobre la conservación de las tortugas marinas en el Perú. Instituto del Mar del Perú, Callao, 72 pp. Jacobson, E Toma de muestras de tejidos y técnicas para la necropsia. In: K. Eckert, K. Bjorndal, F. Abreu- Grobois & M. Donnelly (eds.). Técnicas de investigación y manejo para la conservación de las tortugas marinas. Traducida al español. UICN/CSE Grupo Especialista en Tortugas Marinas, 4: Kelez, S., C. Manrique, X. Velez-Suazo & M. Williams Green turtle (Chelonia mydas agassizii) diet differences in two Peruvian coastal localities. In: M. Coyne & R. Clark (eds.). Proceedings of the Twenty- First Annual Symposium in Sea Turtle Biology and Conservation. NOAA Technical Memorandum NMFS SEFSC-528: 368 pp. Koch, V., L. Brooks & W. Nichols Population ecology of the green/black turtle (Chelonia mydas) in Bahia Magdalena, Mexico. Mar. Biol., 153: Lemons, G., R. Lewison, L. Komoroske, A. Gaos, C. Lai, P. Dutton, T. Eguchi, R. LeRoux & J. Seminoff Trophic ecology of green sea turtles in a highly urbanized bay: insights from stable isotopes and mixing models. J. Exp. Mar. Biol. Ecol., 405: López-Mendilaharsu, M., S. Gardner, J. Seminoff, & R. Riosmena-Rodríguez Identifying critical foraging habitats of the green turtle (Chelonia mydas) along the Pacific coast of the Baja California Peninsula, Mexico. Aquat. Conserv. Mar. Freshw. Ecosyst., 15: López-Mendilaharsu, M., S. Gardner, R. Riosmena- Rodríguez & J. Seminoff Diet selection by immature green turtles (Chelonia mydas) at Bahia Magdalena foraging ground in the Pacific coast of the Baja California Peninsula, México. J. Mar. Biol. Assoc. UK, 88(3): Márquez, R Sea Turtles of the World. An annotated and illustrated catalogue of the sea turtle species known to date. FAO Fisheries Synopsis. Rome, 125(11): 81 pp. Méndez, M Claves de identificación y distribución de los langostinos y camarones (Crustacea: Decapoda) del mar y ríos de la costa del Perú. Boletín Instituto del Mar del Perú, Callao, 172 pp. Mendo, J. (ed.) Libro de Resúmenes del Seminario Bases Científicas para el Manejo y Repoblamiento de la Concha de Abanico en la Bahía de Sechura. Universidad Nacional de Piura (UNP) el día 11 de noviembre del Proyecto FINCYT-UNALM, 65 pp. Mendo, J., M. Wolff, W. Carbajal, I. Gonzáles & M. Badjeck Manejo y explotación de los principales bancos naturales de concha de abanico (Argopecten purpuratus) en la costa peruana. In: A. Lovatelli, A. Farías & I. Uriarte (eds.). Estado actual del cultivo y manejo de moluscos bivalvos y su proyección futura: factores que afectan su sustentabilidad en América Latina. Taller Técnico Regional de la FAO. Chile. FAO Actas de Pesca y Acuicultura, Roma, pp

44 Feeding ecology of Chelonia mydas in Northern Peru 595 Mohan, M. & T. Sankaran Two new indices for stomach content analysis of fishes. J. Fish. Biol., 33: Morón, O., F. Velazco & L. Beltrán Características hidrográficas y sedimentológicas de la Bahía de Sechura. Inf. Inst. Mar Perú, 40(3-4): Moscoso, V Catálogo de Crustáceos Decápodos y Estomatópodos del Perú. Bol Inst. Mar Perú, Callao, 27(1-2): Moscoso, V Clave para identificación de crustáceos decápodos y estomatópodos del Perú. Boletín Instituto del Mar del Perú, Callao, Perú, 28(1-2): Musick, J. & C. Limpus Habitat utilization and migration in juvenile sea turtles. In: P. Lutz, J. Musick & J. Wyneken (eds). The biology of sea turtles 1. CRC Press, New York, pp Ortiz, N., J. Mangel, J. Wang, J. Alfaro-Shigueto, S. Pingo, A. Jiménez, T. Suarez, Y. Swimmer, F. Carvalho & B. Godley Reducing green turtle bycatch in small-scale fisheries using illuminated gillnets: the cost of saving a sea turtle. Mar Ecol. Prog. Ser., 545: Paredes, E Hábitos alimentarios de la tortuga verde del pacífico este Chelonia mydas agassizii (Boucort, 1868) en la Bahía de Paracas, Ica, Perú, durante el año Tesis Licenciatura, Universidad Nacional Mayor de San Marcos, Lima, 69 pp. Paredes, E., J. Quiñones, S. Quispe & V. Bachmann Black and hawksbill turtle strandings in estuarine waters in the peruvian Northern coast. In: Y. Kaska, B. Sonmez, O. Turkecan & C. Sezgin (eds.). Book of Abstracts of 35 th Annual Symposium on Sea Turtle Biology and Conservation. Macart Press, Turkey, 250 pp. Plotkin, P., M. Wicksten & A. Amos Feeding ecology of the loggerhead sea turtle Caretta caretta in the Northwestern Gulf of Mexico. Mar. Biol., 115: 1-5. Quiñones, J., V. Gonzales, J. Zeballos, S. Purca & H. Mianzan Effects of El Niño-driven environmental variability on blackturtle migration to Peruvian foraging grounds. Hydrobiologia, 645: Quiñones, J., I. García-Godos, M. Llapapasca, F. Van Oordt, & E. Paredes. 2015a. The Black Sea turtle (Chelonia mydas agassizii) at Lobos de Tierra Island, Northern Peru: high densities in small areas. South Am. J. Herpetol., 10(3): Quiñones, J., E. Paredes, S. Quispe & L. Delgado. 2015b. Tortugas marinas durante el 2010 en Pisco, Perú. Inf. Inst. Mar Perú, 42(4): Rodríguez, J Afinidad trófica a zonas de alimentación de la tortuga verde (Chelonia mydas) en la costa occidental de Baja California Sur, México. La Paz, Baja California Sur. Tesis Maestría, Instituto Politécnico Nacional, La Paz, 98 pp. Rosales, C., M. Vera & J. Llanos Varamientos y captura incidental de tortugas marinas en el litoral de Tumbes, Perú. Rev. Peru Biol., 17(3): Sampson, L. & A. Giraldo Annual abundance of salps and doliolids (Tunicata) around Gorgona Island (Colombian Pacific), and their importance as potential food for green sea turtles. Rev. Biol. Trop., 62(1): Sampson, L., A. Giraldo & D. Amorocho Diet, food availability and selectivity of Chelonia mydas juveniles at Gorgona National Park, Colombian Pacific. In: T. Tucker, L. Belskis, A. Panagopoulou, A. Rees, M. Frick, K. Williams, R. LeRoux & K. Stewart (eds.). Proceedings of the Thirty-third Annual Symposium on Sea Turtle Biology and Conservation. NOAA Technical Memorandum NOAA NMFS- SEFSC-645: 263 pp. Santillán, L Análisis de la dieta de Chelonia mydas agassizii tortuga verde del Pacífico en la Bahía de Sechura, Piura-Perú. Tesis de Magister, Universidad Agraria La Molina, Lima, 70 pp. Schuyler, Q., B. Hardesty, C. Wilcox & K. Townsend To eat or not to eat? Debris selectivity by marine turtles. PLoS ONE, 7: 7 pp. Schuyler, Q., B. Hardesty, C. Wilcox & K. Townsend Global analysis of anthropogenic debris ingestion by sea turtles. Conserv. Biol., 28: Seminoff, J., A. Reséndiz & W. Nichols. 2002a. Home range of the green turtle (Chelonia mydas) at a coastal foraging ground in the Gulf of California, México. Mar. Ecol. Progr. Ser., 242: Seminoff, J., A. Reséndiz & W. Nichols. 2002b. Diet of East Pacific green turtles (Chelonia mydas) in the Central Gulf of California, Mexico. J. Herpetol., 36(3): Seminoff, J., T. Jones, A. Reséndiz, W. Nichols & M. Chaloupka Monitoring green turtles (Chelonia mydas) at a coastal foraging area in Baja California, Mexico: multiple indices describe population status. Mar. Biol. Assoc. UK, 83: Troncoso-Fierro, F. & B. Urbina-Burgos Actualización del registro de tortugas marinas para el sur de Chile. In: C. Guerra-Correa, A. Fallabrino, P. Bolados- Díaz & C. Turner (eds.). Libro de Resúmenes VII Simposio Sobre Medio Ambiente. Estado Actual y Perspectivas de la Investigación y Conservación de las Tortugas Marinas en las Costas del Pacífico Sur Oriental, Antofagasta. Available at: [ seaturtle.org/documents/vii_simp_medio_ambiente. pdf]. Reviewed: 22 August 2016.

45 596 Latin American Journal of Aquatic Research Tume, J., A. Ibaceta, M. Cortez & J. Santisteban Recursos de la zona béntica de la bahía de Sechura. Ciencia y Desarrollo, 15(1): 50 pp. Vander, H., K. Arthur, A. Bolten, B. Popp, C. Lagueux, E. Harrison, C. Campbell & K. Bjorndal Trophic ecology of a green turtle breeding population. Mar Ecol. Prog. Ser., 476: Velez-Zuazo, X., J. Quiñones, A. Pacheco, L. Klinge, E. Paredes, S. Quispe & S. Kelez Fast growing, healthy and resident green turtles (Chelonia mydas) at two neritic sites in the Central and Northern Coast of Peru: implications for conservation. PLoS ONE, 9(11): Villegas, P Aspectos biológicos-pesqueros del calamar común, Loligo gahi d Orbigny, 1835 en el área del Callao durante 1996 y Tesis Biología, Universidad Nacional Mayor de San Marcos, Lima, 84 pp. Work, T Manual de necropsias de tortugas marinas para biólogos en refugios o areas remotas. U.S. Geological Survey, National Wildlife Health Center. Hawaii. [ Products/Turtle%20manual%20spanish.pdf]. Reviewed: 10 December Wyneken, J The Anatomy of sea turtles. U.S. Department of Commerce NOAA Technical Memorandum NMFS-SEFSC, 172 pp. Zárate, P., K. Bjorndal, M. Parra, P. Dutton, J. Seminoff & A. Bolten Hatching and emergence success in green turtle Chelonia mydas nests in the Galápagos Islands. Aquat. Biol., 19: [ Reviewed: 20 August Received: 25 June 2016; Accepted: 18 December 2016

46 Lat. Am. J. Aquat. Res., 45(3): , 2017 Hawksbill nesting along the Eastern Pacific Rim 5721 Sea Turtle Research and Conservation in Latin America Jeffrey Mangel, Jeffrey Seminoff, Bryan Wallace & Ximena Vélez-Suazo (Guest Editors) DOI: /vol45-issue3-fulltext-7 Research Article Living on the Edge: Hawksbill turtle nesting and conservation along the Eastern Pacific Rim Alexander R. Gaos 1,2,3,14, Michael J. Liles 4,14, Velkiss Gadea 5,14, Alejandro Peña de Niz 6,14 Felipe Vallejo 7,14, Cristina Miranda 7,14, Jodie Jessica Darquea 8,9,14, Ana Henriquez 14 Eduardo Altamirano 5,14, Alejandra Rivera 5,14, Sofía Chavarría 14, David Melero 5,14, José Urteaga 5,10,14 Carlos Mario Pacheco 14 Didiher Chácon 11,14, Carolina LeMarie 7,14, Joanna Alfaro-Shigueto 12,14 Jeffrey C. Mangel 12, Ingrid L. Yañez 14 & Jeffrey A. Seminoff 13,14 1 San Diego State University, San Diego, California, USA 2 University of California Davis, Davis, California, USA 3 Ocean Associates Inc, under contract to the Southwest Fisheries Science Center National Marine Fisheries Service, National Oceanic and Atmospheric Administration La Jolla, California, USA 4 University of Texas at El Paso, El Paso, USA 5 Fauna & Flora International, Managua, Nicaragua 6 Centro de Protección y Conservación de Tortugas Marinas Playa Teopa, Playa Teopa, Jalisco, Mexico 7 Fundación Equilibrio Azul, Quito, Ecuador 8 Ecuador Mundo Ecologico, Ecuador 9 Scripps Institute of Oceanography, San Diego, California, USA 10 Emmett Interdisciplinary Program in Environment and Resources, Stanford University, California, USA 11 Latin American Sea Turtles, Tibas, San Jose, Costa Rica 12 Universidad Cientifica del Sur, Facultad de Biologia Marina, Lima, Peru./ ProDelphinus Lima 18/ University of Exeter, School of Biosciences, UK 13 Southwest Fisheries Science Center, National Marine Fisheries Service, National Oceanic and Atmospheric Administration, La Jolla, California, USA 14 Eastern Pacific Hawksbill Initiative, San Diego, California, USA Corresponding author: Alexander R. Gaos (gaos@hawksbill.org) ABSTRACT. Prior to 2007, efforts to monitor and conserve hawksbill turtles (Eretmochelys imbricata) in the eastern Pacific Ocean were opportunistic and records were virtually non-existent. The first abundance estimates were published in 2010, but contained limited data on the species. Ongoing research since that time has led to the identification of several rookeries, including sites containing large proportions of the overall hawksbill nesting currently known to occur in the region. Monitoring projects were established at several sites and have since provided substantial nesting data on the species. Here we summarize data collected between 1983 and March 2016 from all sites (n = 9) confirmed to host >10 nests in any given season to provide an update on hawksbill nesting in the eastern Pacific. We documented a total of 3,508 hawksbill nests, 265,024 hatchlings and 528 individual nesting females in the region. The vast majority of these records (99.4%, 99.9% and 99.6%, respectively) were generated subsequent to 2007, coinciding with the discovery of eight of the nine rookeries included in this study and the organization of monitoring efforts at those sites, which led to the increased documentation conferred here. Our findings should not be misconstrued as increases in actual nesting or signs of recovery, which could diminish the ongoing need for conservation actions, but rather as optimism, that there is still an opportunity to restore the species in the eastern Pacific. The top three sites in terms of average annual number of nests were Estero Padre Ramos (Nicaragua; ± 47.6 nests), Bahia de Jiquilisco (El Salvador; ± 46.7 nests) and Aserradores (Nicaragua; ± 24.0 nests), and all three sites are located in mangrove estuaries in Central America, highlighting the importance of these rookeries/habitats for the survival and recovery of hawksbills in the region. The remaining six sites received between 6.9 ± 7.3 nests (Costa Careyes, Mexico) and 59.3 ± 17.7 nests (Los Cobanos, El Salvador) annually. By integrating data collected on nesting hawksbills with local conservation realities at the most important known hawksbill rookeries in the eastern

47 573 2 Latin American Journal of Aquatic Research Pacific, we provide a more holistic view of the conservation status and management needs of the species in this ocean region. Keywords: population status, reproductive biology, demography, critically endangered, marine conservation, mangrove estuary. INTRODUCTION The hawksbill turtle (Eretmochelys imbricata) has historically been the focus of intense human exploitation, and in some cases remains heavily exploited, leading to global population declines of >80% and its inclusion as Critically Endangered (CR) on the Red List of the International Union for the Conservation of Nature (IUCN) (Mortimer & Donnelly, 2008). Despite the blanket classification of this and other marine turtle species on the IUCN Red List, such an approach fails to capture the regional differences in population status that often arise from impacts that vary spatially (Seminoff & Shanker, 2008). The importance of evaluating the differential marine turtle statuses and levels of ongoing threats at sub-global spatial scales has been recognized by the IUCN (IUCN, 2003), and this has led to regional assessments of both leatherback (Dermochelys coriacea) and loggerhead (Caretta caretta) turtles, with area-specific evaluations of other species within the taxon slated to follow (Wallace et al., 2013; Casale & Tucker, 2015). Less than a decade ago the hawksbill turtle was considered extremely rare in the eastern Pacific Ocean and no consistent nesting was thought to occur at that time (Mortimer & Donnelly, 2008). Notwithstanding this perspective, the discovery of important remnant nesting and foraging strongholds beginning in 2007 (Vásquez & Liles, 2008; Gaos & Urteaga, 2010; Liles et al., 2011; Gaos & Yañez, 2012) led to a regional data sharing workshop in 2008, where data were compiled and subsequently published, providing the first comprehensive analysis of hawksbill nesting (as well as in-water sightings, strandings and bycatch records) observations in the region (Gaos et al., 2010). Catalyzed by the early findings, the sudden revelations surrounding hawksbills in the eastern Pacific prompted increased attention by the international conservation community and the formation of various, and in most cases, multi-organizational actions aimed at learning about and recovering the species. These escalated efforts led to ongoing discoveries of hawksbill nesting and foraging grounds throughout the region (Alfaro-Shiqueto et al., 2010; Gaos & Urteaga, 2010; Quiñones et al., 2011; Brittain et al., 2012; Carrión-Cortez et al., 2013; Chacón-Chaverri et al., 2014; Tobón-López & Amorocho, 2014; Heidemeyer et al., 2014; Kelez et al., 2016). The rapid surge in data has quickly improved researchers understanding of the species and their critical habitats, behooving updated regional assessments. This is particularly true for nesting as evaluating the size and dynamics of wildlife populations is a core underpinning of conservation ecology (Galliard et al., 1998) and monitoring the number of females that come ashore to nest is the most logistically and economically feasible method to evaluate sea turtle populations (Balazs & Chaloupka, 2004). Recent studies on nesting beaches have demonstrated differences in habitat selection (i.e., mangrove estuaries vs open-coast beaches) and local conservation realities (e.g., protection vs extraction) among sites, with implications for conservation and recovery of the species (Liles et al., 2015a, 2015b, 2016). Furthermore, whereas hawksbills in the eastern Pacific were previously referred to as a single population (Gaos et al., 2010, 2012a, 2012b; Liles et al., 2011, 2015a, 2015b; Wallace et al., 2011), recent genetic research has revealed the existence of multiple stocks in the region, as well as a unique reproductive ecotype represented by hawksbills nesting in mangrove estuaries (Gaos et al., 2016). These findings have heightened interest in understanding differences in nesting levels and conservation scenarios at the various rookeries. In this study we collected data from nine rookeries in the eastern Pacific, ranging from Mexico to Ecuador. We use our findings to 1) present an updated assessment of hawksbill nesting activity in the eastern Pacific region, 2) discuss nesting biology of the species, and 3) describe the conservation realities and challenges at the nesting sites. These data provide a baseline with which to compare future nesting abundance and site-specific population trends. MATERIALS AND METHODS Study sites This study encompasses all hawksbill rookeries (n = 9) confirmed to host >10 nests during any given season along the eastern Pacific Rim (Fig. 1). The nesting sites are distributed across five countries, including Mexico (MX; n = 1; classified as the North America region), El Salvador (ES; n = 3), Nicaragua (NI; n = 2) and Costa Rica (CR; n = 1) (collectively classified as the Central

48 Hawksbill nesting along the Eastern Pacific Rim Figure 1. Locations of the hawksbill nesting sites and beaches confirmed to host >10 nests year -1 along the Pacific coast of North (A), Central (B) and South (C) America. America region), and Ecuador (EC; n = 2; classified as the South America region). We classified individual nesting beaches according to whether coastlines were interrupted by natural barriers (e.g., rivers, rock outcrops) and/or were given distinct local nomenclature, and we grouped beaches together into individual sites as recognized by local inhabitants (Liles, 2015). All sites were classified as either mangrove estuaries (located within a mangrove, salt-water forest estuary) or open-coast (facing the open ocean).

49 575 4 Latin American Journal of Aquatic Research Data collection and analysis Beach surveys to encounter nesting hawksbills and their nests consisted of a combination of nocturnal and/or early morning beach patrols. When hawksbills were encountered (and when feasible), we measured curved carapace length (CCL: nuchal notch to posterior-most tip of marginal scutes), applied Inconel flipper tags (National Band & Tag, Newport, KY, USA) to both front or rear flippers, as well as an internal passive integrated transponder (PIT) tag to a single front flippers to allow for ongoing identification, and recorded the date of the emergence event. Tagged turtles observed across years were classified as remigrants, with remigration intervals representing the number of years between successive nesting events. Internesting intervals were calculated by the number of days between successive nests by an individual turtle within a season. We documented all successful nesting events (whether the turtle was observed or not) by encountering nests during or posterior (i.e., within 12 h) to oviposition, then calculated the fate of all nests as either protected (in situ, relocated along the beach, or relocated to a hatchery by personnel) or lost (poached, predated, washed away by tides). Nests initially left in situ or relocated along the beach were marked and tracked throughout the incubation period, and were reclassified as lost if poached, predated or washed away at a later date. Nest protection rates were calculated by dividing the number of nests protected by the total number of nests recorded for each season. Clutch frequency (i.e., the number of clutches laid per season) was calculated by recording the total number of nests for known (i.e., observed/tagged) individual turtles. When feasible, we conducted post-hatching nest excavations to quantify hatchlings and calculate a hatching success rate, which we determined by dividing the total number of live hatchlings (emerged and/or found in nest during excavation) by the total number of eggs deposited in a nest. Eggs relocated to hatcheries or along beaches were counted at the time of relocation, while the total number of eggs for nests left in situ was determined by only counting shells consisting of 50% or more. For sites implementing the use of hatcheries, live hatchlings emerging from nests or encountered during excavations were released to the sea under the cover of darkness (Eckert et al., 1999). Because nesting in South America begins toward the latter end of the calendar year and extends into the following year, for tables and figures we classified data for nesting seasons at these sites corresponding to the year during which the season started. All data were tabulated, analyzed and graphed using Microsoft Excel v Mean values, standard deviations (SD) and ranges are reported throughout the present paper. RESULTS There were differences in the years monitored, monitoring effort, quantity and spatial extent of individual nesting beaches/sites, habitat type, and other parameters at the study sites (Table 1). A combined total of 3,508 hawksbill nests were recorded in the eastern Pacific during the study timeframe (Table 1), with virtually all nests (99.4%) recorded since 2008 (Fig. 2a) and the majority occurring at Bahia de Jiquilisco (ES; 38.4%) and Estero Padre Ramos (NI; 36.5%) (Fig. 2b). An average annual total of ± 71.0 nests were deposited across the region, with Estero Padre Ramos (NI) receiving the greatest average annual number of nests (213.2 ± 47.6), followed by Bahia de Jiquilisco (168.5 ± 47.6), Aserradores (NI; ± 24.0), Los Cobanos (ES; 59.3 ± 17.7), Osa Peninsula (CR; 52.0, only one year of data available), El Pelado (EC; 46.0, only one year of data available), Punta Amapala (ES; 23.3 ± 10.8), Machalilla (EC; 20.4 ± 10.0) and Costa Careyes (MX; 6.9 ± 7.3) (Table 1). Of the 3,508 nests documented across all sites, 3,048 (86.9%) were protected and 460 (13.1%) were lost, with an average annual protection rate across all sites of 89.8% (±19.0) (Table 1). The average annual nest protection rate was above 90% at Estero Padre Ramos (NI; 94.0 ± 4.4%), Costa Careyes (MX; 98.7 ± 4.1%), Machalilla (EC; 99.7 ± 4.7%) and El Pelado (EC; 100%, only one year of data available), and below 70% at Los Cobanos (ES; 61.2 ± 16.7%) and Aserradores (NI; 66.9 ± 36.6%) (Table 1). A total of 528 individual nesting female hawksbills were observed and tagged at the study sites between 1983 and March 2016 (Table 1), with most females (99.4%) encountered since 2010 (Fig. 2c). Of this total, 215 (40.7%) were tagged in Estero Padre Ramos (NI), 150 (28.4%) in Bahia de Jiquilisco (ES), 70 (13.3%) in Aserradores (NI), 44 (8.3%) in Machalilla (EC), and all other sites representing <5% (Fig. 2d). Conservation efforts at these sites led to the production of an average total of 43,481.2 ± 6,473.4 hatchlings per year and a total of 265,024 hatchlings across all years (Table 1), with 99.9% of those records coming since 2008 (Fig. 2e). The majority of hatchlings were produced at Estero Padre Ramos (42.4%) and Bahía de Jiquilisco (39.1%), with remaining sites accounting for 5% of all hatchlings produced (Fig. 2f). The average annual number of hatchlings produced per site was above 12,000 at Estero Padre Ramos (NI: 18,451.3 ± 4,728.0) and Bahía de Jiquilisco (ES; 12,827.1 ± 6,907.2), and below 2,000 at Punta Amapala

50 Table 1. Information on the nine hawksbill rookeries included in this study, including site name, country, region (NA: North America, CA: Central America, SA: South America), habitat type (ME: mangrove estuary, OC: open coast), coastline (km covered), number of individual nesting beaches, year monitoring initiated, average annual number of hawksbill nests, average annual nest protection rate, average annual number of hatchlings produced, average annual hatching success rate, total number of nests recorded, total number of hatchlings produced, and total number of reproductively active females observed/tagged. Standard deviation (SD) of averages shown in parenthesis for reference. Dash indicates no data available. Hawksbill nesting along the Eastern Pacific Rim (ES: 1,790.8 ± 519.4) and Costa Careyes (MX: ± 663.3) (Table 1). The average hatching success rate across all sites was 56.7% (±20.0), with average annual hatching success rates lower than 55% experienced at both Costa Careyes (MX; 52.2% ± 23.7) and Bahía de Jiquilisco (ES; 52.5% ± 16.5) and rates higher than 70% experience at only Punta Amapala (ES; 72.3% ± 11.6) (Table 1). The majority of hawksbill nests (80.6%), hatchlings (75.1%) and females (82.4%) were recorded at sites located in mangrove estuaries compared to sites located along the open coast. Overall by region, Central America hosted the overwhelming majority of hawksbill nesting (91.9%), followed by South America (6.2%) and North America (1.9%). We found differences in the peak nesting months for hawksbills in the different regions, with a peak in June and July for Central America (slightly elevated values in the latter month), in August for North America, and in January and February for South America (Fig. 3). The average CCL (84.6 ± 7.3 cm), clutch size (159.1 ± 36.9 eggs), clutch frequency (2.0 ± 1.1 clutches), internesting interval (20.1 ± 9.3 days) and remigration interval (2.2 ± 0.7 years) for hawksbills varied among sites (Table 2). DISCUSSION Nesting distribution and abundance Prior to 2008, encountering nesting hawksbills in the eastern Pacific was rare and strictly opportunistic, with no programs designed specifically to monitor and conserve the species. This contributed to major data deficiencies in the region and few directed conservation interventions. Since that time, we have established ongoing efforts aimed at monitoring nesting hawksbills at eight of the nine rookeries included in this study. It is important to recognize that the apparent increase in the number of hawksbill nests and nesting females highlighted in this study (Figs. 2a, 2c) are the result of increased monitoring efforts targeting the species, and hence increased documentation. Our findings should not be interpreted as a recovery trend. Hawksbills in the eastern Pacific are recognized as one of the world s most imperiled and least resilient Regional Management Units (RMU) for all marine turtle species (Wallace et al., 2011; Fuentes et al., 2013). Our findings further confirm the status of this RMU, with only 528 reproductive females documented to date across the region. Despite these findings, considering that the species was only recently considered functionally extirpated in the region, there is now optimism regarding potential recovery.

51 6577 Latin American Journal of Aquatic Research Figure 2. a) Total number of hawksbill nests by year, b) percentage (with totals overs bars) of hawksbill nests by study site, c) total number of individual reproductive female hawksbills identified and tagged by year and d) percentage (with totals overs bars) of individual reproductive female hawksbills identified and tagged by study site, and e) total number of hawksbill hatchlings by year and f) percentage (with totals overs bars) of hawksbill hatchlings by study site, between 1983 and 2015 (data for Machalilla and El Pelado extend into March 2016). Dash indicates no data available. Ongoing monitoring at the known nesting sites will undoubtedly lead to the documentation of additional nesting females. However, with the exception of Los Cobanos (ES) and to a lesser extent Aserradores (NI), the currently monitored nesting sites in Central America (which combined host 85% of the known annual hawksbill nesting in the region), as well as Machalilla (EC) in South America, are quickly approaching tag saturation of non-neophyte females. Annual nesting at Costa Careyes (MX) in North America, Osa Peninsula (CR) in Central America and El Pelado (EC) in South America is relatively limited,

52 Hawksbill nesting along the Eastern Pacific Rim 5787 Figure 3. Percentage of hawksbill nests laid by month in North (n = 52), Central (n = 2804) and South America (n = 204). thus encountering significant numbers of additional females is not likely. We estimate a maximum of additional females have yet to be tagged across all of our study sites. Thus even with these additions, the total number of nesting females would remain <700 individuals across the region. It is possible that additional rookeries, and hence females, will be identified, but efforts have been carried out along much of the eastern Pacific coastline and reports of nesting remain minimal (Gaos et al., unpubl. data). Ongoing conservation actions continue to produce hatchling cohorts, a portion of which will presumably mature into adults. If coupled with decreases in the mortality of adults and juveniles (as a result of conservation actions), we would expect the number of nesting females to increase with time. However, given that hawksbills are estimated to require 17+ years to reach sexual maturity (Boulon, 1994; Chaloupka & Musick, 1997; Crouse, 1999; Snover et al., 2013) and that the bulk of hawksbill conservation in the eastern Pacific was only initiated in 2008, it may be decades before we see an appreciable increase in the number of nesting females. Hawksbills in the eastern Pacific nest on average every 2.2 ± 0.7 years, but in contrast to most other rookeries where nesting in consecutive years is rare or non-existent (e.g., Mortimer & Bresson, 1999; Richardson et al., 2006; Moncada et al., 2010; Prince & Chaloupka, 2012), annual nesting appears relatively common (10.7% of remigrants; n = 22) in both the Central and South American regions of the eastern Pacific (Fig. 4; data unavailable for North America). Oviposition is an energy-demanding process and female sea turtles often skip reproductive years if they do not have sufficient fat reserves, instead remaining in foraging areas to continue to feed and accumulate the energy required (see refs in Santos et al., 2010). Our remigration intervals were pooled from three of our study sites located in the Central (n = 2) and South (n = 1) American regions (Table 2), from which a total of 18 post-nesting hawksbills have been tracked via satellite telemetry, the majority (n = 15; 83.3%) to foraging grounds located in mangrove estuaries (Gaos et al., 2012a, 2012b; Gaos & Seminoff, unpubl. data). Mangrove estuaries are one of the most resource-rich habitats along the eastern Pacific coastline (Dewalt et al., 1996; Gaos et al., 2012b) and an abundance of highquality prey (e.g., invertebrates) in these habitats may provide sufficient energy resources to enable nesting during consecutive years at these study sties. Additionally, adult hawksbills in the eastern Pacific undertake relatively short post-nesting migrations between nesting and foraging grounds (maximum = km) or are non-migratory altogether (Gaos et al., 2012a, 2012b), a behavior that would further facilitate nesting during consecutive seasons. In general, hatching success for hawksbills in the eastern Pacific is on par with conspecifics in other regions (Bjorndal et al., 1985; Marcovaldi et al., 1999; Richardson et al., 1999; Hitchens et al., 2004; Ditmer & Stapleton, 2012). However, with the human consumption of eggs consistently approaching 100% prior to the start of conservation efforts (except for Ecuador) (Vásquez et al., 2008; Altamirano, 2014), and this level of exploitation being maintained for decades along Pacific Latin America (Cliffton et al., 1982; Cornelius, 1982), it is reasonable to assume that little hatchling recruitment occurred during that time. Excessive egg harvest might have set up a population time-bomb (Mortimer, 1999) for hawksbills in which population numbers could suddenly collapse due to lack of immature turtles that would serve to replace older adults as they die (Seminoff & Shanker, 2008). Our reports of nesting females demonstrate that we likely have not reached that point, at least at the three primary rookeries in Central America (i.e., Bahia de Jiquilisco, ES; Estero Padre Ramos, NI; Aserradores, NI). In contrast, it is possible that the extremely low nesting numbers in North and South America (and smaller rookeries in Central America) may be direct evidence that such a collapse is already occurring. However, the low nesting in North and South America could also be due to these rookeries being located at the extremes of hawksbill distribution in the region (Alfaro-Shiqueto et al., 2010; Quiñones et al., 2011). Regardless, conservation efforts since 2008 have led to the release of more than 265,000 hawksbill hatchlings across the region (Table 1). If even a small proportion of these hatchlings represent juveniles in the pipeline that will eventually reach maturity (Seminoff & Shanker, 2008), they should improve the conservation outlook for hawksbills in the eastern Pacific. Ensuring

53 Table 2. Average values and standard deviations for CCL, clutch size, clutch frequency, renesting intervals and remigration intervals for hawksbills at the individual study sites (where available). Standard deviation (SD) of averages shown following average value and range values (min-max) provided in parenthesis. Sample size (n) used in calculations provided in right column for each variable. Dash indicates no data available Latin American Journal of Aquatic Research the protection of hawksbills at their foraging grounds will be fundamental to recovery efforts as sea turtles spend the majority of their lives in these habitats and efforts focused solely on nesting beaches can be insufficient (e.g., Benson et al., 2015). Conservation challenges and opportunities A combined total of 80.6% of the hawksbill nesting documented to date in the eastern Pacific has occurred in the mangrove estuaries at Bahia de Jiquilisco (ES), Estero Padre Ramos (NI) and Aserradores (NI), highlighting the relative importance of these habitats for hawksbills in the region. Without question, major degradation to these sites, which are under increased anthropogenic pressures (Gaos et al., 2012b; Altamirano, 2014, Liles et al., 2015b), would represent a major hindrance to hawksbill recovery. Of primary concern is the conversion of mangrove habitat to shrimp aquaculture ponds. While such conversions remain relatively limited at our study sites, similar habitats in the region (e.g., southern Gulf of Fonseca) have been severely impacted by these practices (Dewalt et al., 1996, Blázquez & Navarrete, 1996). Unfortunately, despite protected area status at two of these three nesting sites, as well as mangrove protection legislation in both countries, many of these protected areas are only considered as such by law (i.e., paper reserves sensu Valdez, 2008), while in actuality receiving limited enforcement by government authorities and continuing to experience high rates of biodiversity and habitat loss (Liles et al., 2016). Considering the critical nature of the hawksbill rookeries in these areas, their urgent, permanent and authentic protection is vital. Climate change is also of marked concern to hatching success in these habitats as the nesting beaches throughout the estuaries are of low relief, with minimal elevation change, and eggs are laid <1 m from the water table (Liles et al., 2015b), making them particularly susceptible to inundation and mortality (Fuentes et al., 2013). For the first time at any of the three mangrove estuary sites, in August of 2015 spring tides led to the lethal flooding of more than 30 nests positioned on the highest sand berms in Bahia de Jiquilisco (ES). Anthropogenic-induced global warming is predicted to lead to more extreme climactic events and an overall increase in average global ocean levels (IPCC, 2013), which may increase nest inundation and mortality events at these sites. Aligning diverse stakeholder interests to hawksbill conservation needs is a challenging task (Fabbri, 1998; Olsen et al., 1997). Landowners and development are often viewed as threats to conservation (Jacobson et al., 2010), yet antagonistic views of these stakeholders can

54 Hawksbill nesting along the Eastern Pacific Rim 5809 Figure 4. Pooled remigration intervals of hawksbills at Bahia de Jiquilisco and Estero Padre Ramos in Central America and Machalilla in South America. be unwarranted. The primary hawksbill nesting habitat in Aserradores (NI) is on the private property of a single land-trust. While several years ago the trust built a small marina within the estuary, they have committed to protecting other parts of the estuary, including the primary hawksbill nesting habitats. Additionally, the trust actively participates in and contributes to hawksbill conservation, which is further facilitated by only having to deal with a single entity rather than numerous landowners who often have conflicting interests. Notwithstanding this perspective, conflicts between several local stakeholders and the marina over resource rights do exist and threaten to impact the effectiveness of conservation efforts. In a different development scenario, but one with similar results, the coastline of Costa Careyes (MX) consists primarily of exclusive resorts and high-end homes, limiting beach access to the general public. While Costa Careyes does not have any official protective status, the restricted access provides a certain degree of pseudo-protection and only three hawksbills nests have been recorded as lost, all due to the predation by ants. However, monitoring at this site is limited and thus it is likely that some nests are poached and/or go undocumented. Development obviously does present risks to habitat as well. Most of the Machalilla nesting complex (EC) is part of a national park and the primary nesting beach (La Playita) is well protected. However, the large coastal community of Puerto Lopez is excluded from the park and the waters directly in front of the town are important foraging areas for adult and juvenile hawksbills. In 2014 there was an initiative to build a solid cement pier on the southern end of the beach, with construction planned directly on top of parts of the primary foraging habitat. While the effort was initially thwarted thanks to the actions of a local conservation organization, plans to build the pier have recently been revived (Ministry of the Environment report MAE- SGMC , 22 June 2015). Hawksbills also occasionally nest on the beach of Puerto Lopez, where the government is currently building a cement boardwalk along most of the beach and the impacts the boardwalk may have on hawksbill nesting remain unknown. It is important to point out that recent genetic research has identified Machalilla as a distinct management unit (MU) from MUs located in Central America (n = 3; Gaos et al., 2016). Coupled with apparent differences in size (CCL) and nesting season (Table 2 and Fig. 3, respectively), increased efforts to ensure this already depleted rookery persists may be warranted. Prior to 2015 Machalilla was the only known nesting site along Pacific South America. However, the pilot monitoring program initiated in El Pelado (EC) in 2015 led to the documentation of a total of 46 nests and 10 females in a single season of monitoring. While this fledgling project was somewhat rudimentary in terms of beach coverage and data collection, the initial findings bode well for the conservation outlook of hawksbills along Pacific Ecuador. Efforts are being made to increase coverage of the nesting beach in an effort to observe additional nesting females, evaluate hatching success, collect genetic samples and gather additional information. Plans to develop a condominium complex directly in front of the primary hawksbill nesting beach of Playa Rosada in El Pelado threaten to negatively impact nesting females/hatchlings and should be closely scrutinized. The relatively low representation of Aserradores (NI) in terms of overall number of nests, hatchlings and females (Figs. 2b, 2d, 2f) across years masks its relative importance. Monitoring of hawksbills was only initiated in 2014 and during the subsequent two years (i.e., ), 200 nests and 70 nesting females recorded at the site during that time. Indeed, with an average of 100 ± 24.0 nests deposited annually, Aserradores is the third largest known rookery in the region (Table 1). Due to the relative infancy of the program, illegal collection (i.e., poaching) of nests remains a challenge (Table 1), as is the case with many of the programs during their initial years, and it is expected that conservation results will improve as interpersonal relationships are built with advancement of the program (e.g., Hawksbill Cup competition, Liles et al., 2016). Similar to Aserradores, the regional importance of Los Cobanos (ES) in the overall hawksbill nesting totals is also somewhat misleading, but for quite a different set of circumstances. After a successful inaugural 2008 season during which monitoring resulted in the recording of 79 nests, the program only functioned for one out of the subsequent five years. The

55 Latin American Journal of Aquatic Research program was re-initiated in 2014 and has continued to function since that time. Nonetheless, the presence of local gangs (referred to in El Salvador as Maras ) on the beaches that are often involved in the illegal collection and sale of turtle eggs, discourages program collaborators from patrolling certain beaches and/or during certain times (Liles et al., 2011). This reduces our ability to collect information on the species and protect a greater percentage of nests (poaching at Los Cobanos is approximately 40% of total nests annually). Despite being the fourth largest known hawksbill rookery in the eastern Pacific in terms of average annual number of nests (Table 1), there are no immediate solutions to this issue in sight. The beaches identified along the Osa Peninsula (CR) collectively represent the only site not currently being monitored for hawksbill nesting. This is due to the sparse and relatively low levels of nesting across what constitutes a relatively large geographic area (approximately 25 km, Table 1). Coupled with most of the beaches being extremely remote and difficult to access, these realities render ongoing monitoring efforts unfeasible. Indeed, identifying and protecting hawksbill nesting grounds throughout the eastern Pacific remains challenging due to the disparate, lowdensity, cryptic nesting demonstrated by the species (Vásquez & Liles, 2008; Gaos et al., 2010; Liles et al., 2011, 2015a) as these nesting characteristics necessitate logistical, temporal and/or financial investments that often are unattainable by researchers. Regardless, efforts to identify nesting habitats can be facilitated by focusing acute, rapid assessments during the peak nesting seasons (Fig. 3) (SWOT, 2011). Sites of primary importance for such assessments in North America (i.e., Mexico) include the Islas Tres Marias, an island chain off the coast of the Mexican state of Nayarit that has long been rumored to host important numbers of nesting hawksbills (Sweifel, 1960; Woodes, 1711 in Saenz-Arroyo et al., 2006), but which has yet to be properly investigated (Gaos & Yañez, 2012), as well as much of the rocky coastline along the Mexican state of Michoacan, which is interspersed with numerous small sandy coves. In Central America, the coastline along northwest Costa Rica (particularly in Santa Rosa National Park), as well as the beaches of Coiba Island, Las Perlas Archipelago and the Darien Gap in Panama, are sites that are difficult to access and remain poorly researched. In South America much of the coast of Colombia remains under investigated, particularly in the northern Choco region. Long-term, consistent data collection at the rookeries highlighted in this study is necessary to evaluate population trends over time. Additionally, rigorous data on survival and growth rates are needed for the development of accurate population models (Chaloupka & Limpus, 2005; Eguchi et al., 2010) and long-term in-water monitoring programs recently began at several key hawksbill foraging grounds in an effort to generate the necessary information. Further, these in-water monitoring programs seek to evaluate whether increased numbers of young life-stage hawksbills are recruiting to foraging grounds as a result of previous and ongoing nest/hatchling conservation efforts at rookeries. CONCLUSIONS In this study we provide the most current and robust nesting dataset for hawksbill turtles in the eastern Pacific. Only a few years ago records of nesting by the species in the region were virtually non-existent. The movement to discover, learn about and conserve hawksbills in this ocean region was precipitous and has advanced rapidly since commencement. This has led to increased documentation of hawksbills, demonstrating ongoing persistence of the taxa in the region and providing hope for recovery. Nonetheless, as we ve demonstrated here, the species remains in a highly precarious state (Table 1) and remains largely conservation dependent (Scott et al., 2005). Ongoing efforts to protect hawksbills at our study sites are essential to recovery of the population. ACKNOWLEDGEMENTS We thank the National Fish and Wildlife Foundation, U.S. National Fish and Wildlife Foundation, U.S. National Oceanic and Atmospheric Administration, U.S. Agency for International Development, International Seafood Sustainability Institute, SeeTurtles.org, Darwin Initiative, Seaworld Busch Gardens Conservation Fund, The William H. Donner Foundation, Inc., Marina del Sol, Asociación Mangle, Ayuda en Acción, CODEPA, FUNDARRECIFE, FUNZEL, The Rufford Foundation, Puerto Barillas, Vivazul, Plant-A-Fish,?!Careyes Foundation, Bodega Williams & Humbert, Conservation International-Ecuador, Machalilla National Park, The Ocean Foundation and the Hawksbill Committee of Estero Padre Ramos for logistical/ financial support. We are grateful to national environmental authorities in Mexico (CONANP), El Salvador (MARN), Nicaragua (MARENA), Costa Rica (SINAC-ACOSA) and Ecuador (MAE) for permits. We are indebted to the community members and careyeros of the study sites for their support. In particular we would like to thank the following individuals: Perla Torres, Luís Manzanares, Neftalí Sanchez, Melissa Valle, Wilfredo Lopez, Celina Dueñas, Luis Mera,

56 Hawksbill nesting along the Eastern Pacific Rim Ricardo Gonzales, Alexis Lopez, Alan Baez, Katty Rendn, Paulina Lopez, Michelle Pico and Earl Possardt. REFERENCES Altamirano, E Informe del Proyecto de Conservación de tortuga Carey (Eretmochelys imbricata) en la RN Estero Padre Ramos, Nicaragua. Temporada Fauna y Flora Internacional e ICAPO. Managua, 34 pp. Alfaro-Shigueto, J., J.C. Mangel, C. Caceres, J.A. Seminoff, A.R. Gaos & I. Yañez Hawksbill turtles in Peruvian coastal fisheries. Mar. Turtle Newslett., 129: Balazs, G.H. & M. Chaloupka Thirty-year recovery trend in the once depleted Hawaiian green sea turtle stock. Biol. Conserv., 117: Benson, S.R., R.F. Tapilatu, N. Pilcher, P.S. Tomillo & L.S. Martínez Leatherback turtle populations in the Pacific Ocean. In: J.R. Spotila & P.S. Tomillo (eds.). The Leatherback turtle: biology and conservation. Johns Hopkins Universtiy Press, Baltimore, Maryland, pp Bjorndal, K.A., A. Carr, A.B. Meylan & J.A. Mortimer Reproductive biology of the hawksbill Eretmochelys imbricate at Tortuguero, Costa Rica, with notes on the ecology of the species in the Caribbean. Biol. Conserv., 34(4): Blázquez, L.A. & G. Navarrete Identificación de un proyecto Araucaria XXI en el Golfo de Fonseca. Tech Mem Agencia Española de Cooperación Internacional para el Desarollo (AECID), Madrid, and Comisión Centro Americana de Ambiente y Desarrollo (CCAD), Antiguo Cuscatlán, El Salvador. Boulon, R.H Growth rates of wild juvenile hawksbill turtles, Eretmochelys imbricata, in St. Thomas, United States Virgin Islands. Copeia, 1994(3): Brittain, R., S. Handy & S. Lucas Two reports of juvenile Hawksbill sea turtles (Eretmochelys imbricata) on the southeast coast of Guatemala. Mar. Turtle Newslett., 133: Carrión-Cortez, J., C. Canales-Cerro, R. Arauz & R. Riosmena-Rodríguez Habitat use and diet of juvenile Eastern Pacific hawksbill turtles (Eretmochelys imbricata) in the North Pacific coast of Costa Rica. Chelonian Conserv. Biol., 12(2): Casale, P. & A.D. Tucker Caretta caretta. The IUCN Red List of Threatened Species 2015: e.t3897a [ UK RLTS.T3897A en]. Reviewed: 4 April Chacón-Chaverri, D., D.A. Martínez-Cascante, D. Rojas & L.G. Fonseca Golfo Dulce, Costa Rica, un área importante de alimentación para la tortuga carey del Pacífico Oriental (Eretmochelys imbricata). Rev. Biol. Trop., 63: Chaloupka, M. & C. Limpus Estimates of sex- and age-class-specific survival probabilities for a southern Great Barrier Reef green sea turtle population. Mar. Biol., 146: Chaloupka, M.Y. & J.A. Musick Age, growth, and population dynamics. In: P.L. Lutz & J.A. Musick (eds.). The biology of sea turtles. CRC Press, Boca Raton, pp Cliffton, K., D.O. Cornejo & R.S. Felger Sea turtles of the Pacific coast of Mexico. In: K.A. Bjorndal (ed.). Biology and conservation of sea turtles. Smithsonian Institution Press, Washington DC, pp Cornelius, S.E Status of sea turtles along the Pacific coast of Middle America. In: K.A. Bjorndal (ed.). Biology and conservation of sea turtles. Smithsonian Institution Press, Washington DC, pp Crouse, D.T Population modeling and implications for Caribbean hawksbill sea turtle management. Chelonian Conserv. Biol., 3: Dewalt, B.R., P. Vergne & M. Hardin Shrimp aquaculture development and the environment: people, mangroves and fisheries on the Gulf of Fonseca, Honduras. World Develop., 24(7): Ditmer, M.A. & S.P. Stapleton Factors affecting hatch success of hawksbill sea turtles on Long Island, West Indies. PLoS ONE, 7(7): e Eckert, K.L., K.A. Bjorndal, F.A. Abreu-Grobois & M. Donnelly (eds.) Research and management techniques for conservation of sea turtles. IUCN/SSC Marine Turtle Specialist Group Publication No. 4. Eguchi, T., J.A. Seminoff, R.A. LeRoux, P.H. Dutton & D.L. Dutton Abundance and survival rates of green turtles in an urban environment: coexistence of humans and an endangered species. Mar. Biol., 157: Fabbri, K.P A methodology for supporting decision making in integrated coastal zone management. Ocean Coast. Manage., 39(1-2): Fuentes, M., D.A. Pike, A.D. Dimatteo & B.P. Wallace Resilience of marine turtle regional management units to climate change. Global Change Biol., 19(5): Gaillard, J.M., M. Festa-Bianchet & N.G. Yoccoz Population dynamics of large herbivores: variable recruitment with constant adult survival. TREE, 13(2): Gaos, A.R., F.A. Abreu-Grobois, J. Alfaro-Shigueto, D. Amorocho, R. Arauz, A. Baquero, R. Briseno, D. Chacon, C. Duenas, C. Hasbun, M. Liles, G. Mariona, C. Muccio, J.P. Munoz, W.J. Nichols, M. Pena, J.A. Seminoff, M. Vasquez, J. Urteaga, B. Wallace, I.L.

57 Latin American Journal of Aquatic Research Yañez & P. Zarate Signs of hope in the eastern Pacific: international collaboration reveals encouraging status for a severely depleted population of hawksbill turtles Eretmochelys imbricata. Oryx, 44(4): Gaos, A.R. & J. Urteaga New conservation project for hawksbill turtles in Estero Padre Ramos Natural Reserve, Nicaragua. Oryx, 44: Gaos, A.R. & I. Yañez Saving the Eastern Pacific hawksbill from extinction: last chance or chance lost. In: J.A. Seminoff & B.P. Wallace (eds.). Marine turtles of the Eastern Pacific Ocean. University of Arizona Press, Arizona, pp Gaos, A.R., R.L. Lewison, I.L. Yañez, B.P. Wallace, M.J. Liles, W.J. Nichols, A. Baquero, C.R. Hasbún, M. Vásquez, J. Urteaga & J.A. Seminoff. 2012a. Shifting the life-history paradigm: discovery of novel habitat use by hawksbill turtles. Biol. Lett., 8: Gaos, A.R., R.L. Lewison, B.P. Wallace, I.L. Yañez, M.J. Liles, W.J. Nichols, A. Baquero, C.R. Hasbun, M. Vasquez, J. Urteaga & J.A. Seminoff. 2012b. Spatial ecology of critically endangered hawksbill turtles Eretmochelys imbricata: implications for management and conservation. Mar. Ecol. Prog. Ser., 450: Gaos, A.R., R.L. Lewison, M.J. Liles, V. Gadea, E. Altamirano, A.V. Henríquez, P. Torres et al Hawksbill turtle terra incognita: conservation genetics of eastern Pacific rookeries. Ecol. Evol., 6(4): Heidemeyer, M., R. Arauz-Vargas & E. López-Agüero New foraging grounds for hawksbill (Eretmochelys imbricata) and green turtles (Chelonia mydas) along the northern Pacific coast of Costa Rica, Central America. Rev. Biol. Trop., 62: Hitchins, P.M., O. Bourquin & S. Hitchins Nesting success of hawksbill turtles (Eretmochelys imbricata) on Cousine Island, Seychelles. J. Zool., 264(4): Intergovernmental Panel on Climate Change (IPCC) Climate Change 2013: The physical science basis. In: T.F. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex & P.M. Midgley (eds.). Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, pp International Union for the Conservation of Nature (IUCN) Guidelines for Application of IUCN Red List Criteria at Regional Levels: Version 3.0. IUCN Species Survival Commission. IUCN Gland Switzerland and Cambridge, UK, 26 pp. Jacobson, C.A., J.F. Organ, D.J. Decker, G.R. Batcheller & L. Carpenter A conservation institution for the 21st century: implications for state wildlife agencies. J. Wildl. Manage., 74(2): Kelez, S., X. Velez-Suazo & A.S. Pacheco First record of hybridization between green Chelonia mydas and hawksbill Eretmochelys imbricata sea turtles in the southeast Pacific. Peer. J., 4: e1712. doi: / peerj Liles, M.J., S. Chavarría, E. Altamirano, I. Yañez, J. Urteaga & A.R. Gaos Community construction through culturally rooted celebration: turtles all the way down. In: A.M. Feldpausch-Parker, H. Bergea, T.R. Peterson & K. Raitio (eds.). Environmental communication and community: constructive and deconstructive dynamics of social transformation. Routledge, London, pp Liles, M.J., M.J. Peterson, Y.S. Lincoln, J.A. Seminoff, A.R. Gaos & T.R. Peterson. 2015a. Connecting international priorities with human wellbeing in lowincome regions: lessons from hawksbill turtle conservation in El Salvador. Local Environ., 20(11): Liles, M.J., M.J. Peterson, J.A. Seminoff, E. Altamirano, A.V. Henríquez, A.R. Gaos, V. Gadea, J. Urteaga, P. Torres, B.P. Wallace & T.R. Peterson. 2015b. One size does not fit all: importance of adjusting conservation practices for endangered hawksbill turtles to address local nesting habitat needs in the eastern Pacific Ocean. Biol. Conserv., 184: Liles, M.J Nesting ecology and interactions between local and international priorities for hawksbill sea turtle (Eretmochelys imbricata) conservation on the Pacific coast of Central America. PhD. Dissertation, Texas A&M University, Texas, 148 pp. Liles, M.J., M.V. Jandres, W.A. Lopez, G.I. Mariona, C.R. Hasbun & J.A. Seminoff Hawksbill turtles Eretmochelys imbricata in El Salvador: nesting distribution and mortality at the largest remaining nesting aggregation in the eastern Pacific Ocean. Endang. Species Res., 14: Marcovaldi, M.A., C.F. Vietas & M.H. Godfrey Nesting and conservation management of hawksbill turtles (Eretmochelys imbricata) in Northern Bahia, Brazil. Chelonian Conserv. Biol., 3(2): Moncada, F.G., G. Nodarse, Y. Medina & E. Escobar Twelve years of monitoring hawksbill turtle (Eretmochelys imbricata) nesting at Doce Leguas Keys and Labyrinth, Jardines de la Reina Archipelago, Cuba. Mar. Turtle Newslett., 127: 6-8. Mortimer, J.A. & R. Bresson Temporal distribution and periodicity in hawksbill turtles (Eretmochelys imbricata) nesting at Cousin Island, Republic of Seychelles, Chelonian Conserv. Biol., 3: Mortimer, J.A. & M. Donnelly Eretmochelys imbricata. The IUCN Red List of Threatened Species, Version

58 Hawksbill nesting along the Eastern Pacific Rim Olsen, S., J. Tobey & M. Kerr A common framework for learning from ICM experience. Ocean Coast. Manage., 37(2): Prince, R.I.T. & M. Chaloupka Estimating demographic parameters for a critically endangered marine species with frequent reproductive omission: hawksbill turtles nesting at Varanus Island, Western Australia. Mar. Biol., 159: Quiñones, J., J. Zeballos, S. Quispe & L. Delgado Southernmost records of hawksbill turtles along the east Pacific coast of South America. Mar. Turtle Newslett. 130: Richardson, J.I., R. Bell & T.H. Richardson Population ecology and demographic implications drawn from an 11-year study of nesting hawksbill turtles, Eretmochelys imbricata, at Jumby Bay, Long Island, Antigua, West Indies. Chelonian Conserv. Biol., 3(2): Richardson, J.I., D. Hall, P. Mason, K. Andrews, R. Bjorkland, Y. Cai & R. Bell Eighteen years of saturation tagging data reveal a significant increase in nesting hawksbill sea turtles (Eretmochelys imbricata) on Long Island, Antigua. Anim. Conserv., 9: Saenz-Arroyo, A., C.M. Roberts, J. Torre, M. Cariño- Olvera & J.P. Hawkins The value of evidence about past abundance: marine fauna of the Gulf of California through the eyes of 16 th to 19 th century travelers. Fish. Fish., 7: Santos, A.J.B., E.M.X. Freire, C. Bellini & G. Corso Body mass and the energy budget of gravid hawksbill turtles (Eretmochelys imbricata) during the nesting season. J. Herpetol., 44(3): Scientific Advisory Board (SWOT) The state of the World s Sea Turtles (SWOT) minimum data standards for nesting beach monitoring, version 1.0. Handbook, 28 pp. Scott, J.M., D.D. Goble & J.A. Wiens Recovery of imperiled species under the Endangered Species Act: the need for a new approach. Front. Ecol. Environ., 3: Seminoff, J.A. & K. Shanker Marine turtles and IUCN Red Listing: a review of the process, the pitfalls, and novel assessment approaches. J. Exp. Mar. Biol. Ecol., 356: Snover, M.L., G.H. Balazs, S.K.K. Murakawa, S.K. Hargrove, M.R. Rice & W.A. Seitz Age and growth rates of Hawaiian hawksbill turtles (Eretmochelys imbricata) using skeletochronology. Mar. Biol., 160: Sweifel, R.G Herpetology of the Tres Marías Islands. Results of the Puritan-American Museum of Natural History Expedition to Western Mexico. Bull. Am. Mus. Nat. Hist., 119(2): Ch. 9. Tobón-López, A. & D.F. Amorocho Population study of the hawksbill turtle Eretmochelys imbricata (Cheloniidae) in the Southern Pacific region of Colombia. Acta Biol. Colomb., 19(3): Valdez, R Wildlife conservation in China: preserving the habitat of China's wild west, by Richard B. Harris. J. Wildl. Manage., 72: Vásquez, M. & M.J. Liles Estado actual de las tortugas marinas en El Salvador. Mesoamericana, 12(3): 53. Vásquez, M., M.J. Liles, W. López, G. Mariona & J. Segovia Sea turtle research and conservation, El Salvador. FUNZEL-ICMARES/UES & National Fish and Wildlife Foundation, San Salvador, Final Report, 13 pp. Wallace, B.P., M. Tiwari & M. Girondot Dermochelys coriacea. The IUCN Red List of Threatened Species 013:e.T6494A [ dx.doi.org/ /iucn.uk rlts.t6494a en]. Reviewed: 4 January Wallace, B.P., A.D. DiMatteo, A.B. Bolten, M.Y. Chaloupka, B.J. Hutchinson, F.A. Abreu-Grobois, J.A. Mortimer et al Global conservation priorities for marine turtles. PLoS ONE, 6(9): e Received: 5 July 2016; Accepted: 15 November 2016

59 Lat. Am. J. Aquat. Res., 45(3): , 2017 Leatherback turtle secondary nesting beaches in Costa Rica 563 Sea Turtle Research and Conservation in Latin America Jeffrey Mangel, Jeffrey Seminoff, Bryan Wallace & Ximena Vélez-Suazo (Guest Editors) DOI: /vol45-issue3-fulltext-6 Research Article Secondary nesting beaches for leatherback turtles on the Pacific coast of Costa Rica Pilar Santidrián-Tomillo 1,2, Nathan J. Robinson 2,3, Luis Gabriel Fonseca 4, Wagner Quirós-Pereira 4,5 Randall Arauz 6, Madeleine Beange 6, Rotney Piedra 7, Elizabeth Vélez 8, Frank V. Paladino 2,3 James R. Spotila 2,9 & Bryan P. Wallace 10,11 1 Population Ecology Group, Institut Mediterrani d Estudis Avançats IMEDEA (CSIC-UIB), Esporles, Spain 2 The Leatherback Trust, Playa Grande, Costa Rica 3 Department of Biology, Indiana University-Purdue University Fort Wayne, Indiana, USA 4 Biocenosis Marina (BIOMA), Moravia, San José, Costa Rica 5 Bren School of Environmental Science and Management, University of California Santa Barbara, Santa Barbara, California, USA 6 Programa de Restauración de Tortugas Marinas, PRETOMA, Costa Rica 7 Sistema Nacional de Áreas de Conservación (SINAC), Costa Rica 8 Asociación Kuemar, Heredia, Costa Rica 9 Department of Biodiversity, Earth and Environmental Science Drexel University Philadelphia, Pennsylvania, USA 10 Abt Associates, Boulder, USA 11 Nicholas School of the Environment, Duke University, Beaufort, USA Corresponding author: P. Santidrián Tomillo (bibi@leatherback.org) ABSTRACT. Leatherback turtles (Dermochelys coriacea) have declined dramatically in the eastern Pacific Ocean (EP) in recent decades. Traditionally, population assessments have relied on the numbers of turtles on the beaches with the highest abundance of turtles (index beaches) and often disregarded the importance of nesting beaches with smaller, but still regular, numbers of nesting turtles (secondary beaches). We characterize leatherback nesting on secondary beaches throughout Pacific Costa Rica. Nesting distribution was significantly reduced since the 1990s and it currently appears to be constricted to the Santa Elena and Nicoya peninsulas. Over the past five years, nesting abundance on secondary beaches was low, ranging between 0.4 ± 0.5 and 5.3 ± 1.5 females and 3.8 ± 5.2 and 22.8 ± 10.8 nests per beach and per year. There was some exchange of turtles between beaches. The exchange rate (percentage of females that nested at least once on a different beach) ranged between 7% and 28%. While Caletas still registers multiple clutches that are laid by 1-2 females in some years, it may no longer qualify as a secondary beach due to the infrequent nature of nesting events registered recently and the total absence of nests in some of those years. Although nesting abundance is relatively low at secondary beaches, they host at least ~25% of total leatherback nesting abundance in Costa Rica. As the EP leatherback turtle declines, not only do the numbers of nesting turtles decrease but local extirpations are occurring on, previously categorized, secondary beaches. The critically low number of turtles at present may prevent recolonization of sites where they have been extirpated. Keywords: metapopulation, sea turtle, index beach, secondary beach, endangered, declining population. INTRODUCTION Metapopulation theory focuses on the interactions and distribution of spatially-separated subpopulations, the persistence of which depends on the balance between extinctions and recolonizations (Hanski & Gilpin, 1991; Hanski, 1998). Genetic flow and recolonizations are driven by the dispersion of individuals from extant subpopulations via the rescue effect (Ebenhard, 1991). The rate at which this may occur is, in turn, dependent on the 1) distance between sites and 2) dispersion rates (Ebenhard, 1991). However, recolonization may not be possible in declining populations that have reached critically low abundances.

60 564 Latin American Journal of Aquatic Research Understanding metapopulation dynamics and conservation biogeography is one of the global research priorities for the conservation of sea turtles in the 21 st century (Hamann et al., 2010). In particular, the factors behind the dispersion and demography of subpopulations within metapopulations are considered as critical information for population assessments. Sea turtles nest on tropical and subtropical beaches, and population assessments of sea turtles are generally based on the number of females nesting on a particular beach. However, some exchange of individuals between nesting sites within a region occurs (Miller, 1997). Furthermore, sea turtle metapopulations could be complex and the characteristics of the different subpopulations, as well as the relationships among each other, may play an important role in their dynamics. Metapopulation assessments may therefore be particularly important when the number of nesting turtles reaches critically low levels that may prevent exchange between nesting sites and thus result in local extirpations. The nesting range for leatherback turtles (Dermochelys coriacea) along the coast of the eastern Pacific Ocean (EP) extends from Northern Mexico to Panama, with peak activity concentrating in Mexico and Costa Rica. Conservation efforts in Mexico have been focused on four index beaches (Mexiquillo, Tierra Colorada, Cahuitán and Barra de la Cruz; Sarti- Martínez et al., 2007), while conservation efforts in Costa Rica have been focused at Parque Nacional Marino Las Baulas (PNMB), a complex comprising three nesting beaches (Playa Grande, Ventanas and Langosta; Santidrián-Tomillo et al., 2007). In Costa Rica, the level of exchange between the beaches in PNMB is ~20% per season (Santidrián-Tomillo et al., 2007), but the number of nesting turtles and the level of exchange have yet to be quantified on other beaches of secondary importance outside of PNMB. Based on annual surveys at index beaches, the overall abundance of EP leatherbacks has declined exponentially (>90%) across their nesting range since the 1980s (Santidrian- Tomillo et al., 2007; Sarti-Martinez et al., 2007; Wallace et al., 2013) Due to the decreasing abundance of EP leatherbacks at index sites, we need to identify additional opportunities for targeted conservation efforts to protect remaining nesting females and their offspring at secondary nesting sites as well. The term index beach is used to refer to selected beaches or sites where nesting activity is most intense and/or sites that are monitored regularly and over the long-term to provide an indication of population trends (SWOT, 2011). Thus, secondary beaches can be considered nesting sites where turtles nest regularly, are used by the same subpopulation of turtles that nest on the index beach(es) and are of secondary importance due to lower intensity of nesting activity. While the benefit of allocating resources for research and conservation on index beaches is clear, the relative importance of secondary beaches to the metapopulation is less obvious. Due to funding and/or logistic limitations, secondary beaches are often disregarded. However, given the urgent need to protect nesting females and enhance hatchling production of endangered sea turtle populations such as EP leatherbacks, a thorough understanding of the complete nesting distributions, including secondary beaches, and how turtles utilize available nesting habitat within those distributions, would inform expanded conservation efforts that include secondary beaches. We hereby conduct a first assessment of secondary nesting beaches for leatherback turtles in Costa Rica. In particular, we report on nesting abundance and protection level of secondary beaches, and compare current nesting activities to estimates obtained in the 1990s from aerial and ground surveys. Finally, we discuss the relative importance of these secondary nesting sites for the Costa Rican subpopulation of EP leatherback turtles. As leatherback turtles are critically endangered in the EP, understanding the dynamics among nesting sites and the conservation challenges of each site is essential to dictate policy directed towards the prevention of local extirpations and allow recoveries. MATERIALS AND METHODS As a first step toward characterizing the complete nesting distribution in Costa Rica, we conducted a workshop in March 2013 to identify sites where leatherback turtles nested outside PNMB. At the workshop, we identified 8 sites that could qualify as secondary beaches due to the occurrence of regular nesting events. These were: Naranjo, Cabuyal, Nombre de Jesús, Marbella, Junquillal, Ostional, Camaronal and Caletas. Other sites were also identified but these beaches are not considered secondary sites as nesting was sporadic. For this study, we acquired information from 5 of identified secondary nesting beaches: Naranjo, Cabuyal, Nombre de Jesús, Ostional and Caletas. For each site, we determined: 1) nesting abundance, 2) protection level, and when possible 3) exchange rate between nesting sites. Data for Cabuyal and Nombre de Jesús were only available from 2010 and 2011 respectively, while data from Ostional and Caletas were available from 2004 and 2002 respectively. The longest data set was provided for Naranjo, which was intermittently covered from 1971 to 2015.

61 Leatherback turtle secondary nesting beaches in Costa Rica 565 Nesting abundance was estimated as the number of verified nests and/or nesting females. When number of nests was not available, we estimated it by multiplying the number of body pits by the mean nesting success (90%) determined by the multi-decade monitoring program in PNMB as estimated in Reina et al. (2002) and corrected with recent data. Females were identified with Passive Integrated Transponders (PITs) (Reina et al., 2002) and occasionally metal ID tags. We compared our results to those obtained from aerial surveys and ground surveys conducted in seasons and along the entire Pacific coast of Costa Rica (Mayor, 1998). Protection level was categorized by official protection status as: 1) National Park, 2) Wildlife Refuge or 3) not protected under an official management category in accordance with the management categories of the National System of Conservation Areas (SINAC) of Costa Rica. RESULTS Our comparison of historical (1990s) and present day leatherback nesting activities in Pacific Costa Rica revealed a contraction of the population s overall nesting distribution from more than 10 sites to 5 sites that host regular nesting, including PNMB. In the 1990s, the 10 most important identified leatherback secondary nesting beaches outside PNMB were (North to South): Potrero Grande, Naranjo, Cabuyal, Matapalo, Camaronal, Caletas, Hermosa, Carate, Río Oro and Pejeperro/Piro (Fig. 1) (Mayor, 1998). Of those, Hermosa, Carate, Río Oro, Pejeperro/Piro and possibly Caletas seem to currently only be receiving sporadic nesting. Based on preliminary information gathered at the 2013 workshop, Camaronal and Junquillal could qualify as secondary nesting beaches, although the number of nests was low at Camaronal, but we could not obtain detailed updated information on these sites for this study to verify their status. Levels of nesting at Potrero Grande and Matapalo are unknown and therefore, could either be secondary or sporadic. In contrast to the historic distribution of leatherback nesting along the Pacific coast of Costa Rica, secondary nesting beaches are currently restricted to the Nicoya and Santa Elena peninsulas (Fig. 2). We identified four important secondary beaches (North to South): Naranjo, Cabuyal, Nombre de Jesús and Ostional, where leatherback turtles nested regularly with at least one nesting event registered per year throughout the period of study (Fig. 2). Caletas may no longer be considered as a secondary nesting beach, despite having registered nesting events regularly throughout the 2000s, as no leatherbacks nested there over three consecutive years ( to ) and the mean annual number of females over the last five years was lower than one turtle. As expected, the number of turtles and nests registered on secondary beaches were low (Tables 1-2). On average, 62.8 ± 7.0 nests (25%) were registered per year in the last five years on secondary beaches and ± 44.1 nests (75%) per year at PNMB over the same time period. The site with the highest number of turtles and nests was Ostional, whilst the site with lowest abundance was Caletas. Mean (± SD) annual number of turtles over the last five years ranged between 3.4 ± 1.9 turtles at Cabuyal and 5.3 ± 1.5 turtles at Ostional and was less than one turtle per year for Caletas (0.4 ± 0.5 turtles) (Table 1). Mean (± SD) annual number of nests over five years ranged between 8.6 ± 7.7 nests at Nombre de Jesús and 22.8 ± 10.8 nests at Ostional, and 3.8 ± 5.2 at Caletas (Table 2). At Naranjo and Caletas, the two sites with the longest running programs, the number of females and nests declined (Fig. 3). In general, most females identified on secondary beaches remained at the same site, although we did detect a low level of exchange between beaches. At Ostional, 12 out of the 43 identified females were observed at least once at other sites (28%). At Naranjo, 4 out of 21 females were also observed at other beaches (19%). At Caletas, 1 out of 15 turtles were observed elsewhere (7%) and at Cabuyal, 2 out of 15 turtles were observed at other beaches (13%). Protection level varied among sites. Some secondary beaches were protected as National Parks (Naranjo in Santa Rosa National Park), which is the highest level of management protection, Wildlife Refuges (Ostional and Caletas) or were not protected under an official management category (Cabuyal and Nombre de Jesús) (Fig. 2). DISCUSSION EP leatherbacks have precipitously declined since monitoring projects started in Mexico and Costa Rica in the 1980s, which led to their classification as Critically Endangered on the IUCN Red List of Threatened Species TM (Wallace et al., 2013). That assessment was based on the decreasing number of nesting turtles and nests quantified on the index beaches of both countries (Santidrián-Tomillo et al., 2007; Sarti-Martínez et al., 2007). Our results indicate that as the population declined in numbers of nesting females and nests, some beaches that formerly hosted regular nesting now host only sporadic nesting (e.g., Caletas and beaches of the Osa Peninsula). As a result, the geographic distribution of leatherback nesting in Costa Rica has been reduced to the Santa Elena and

62 566 Latin American Journal of Aquatic Research Figure 1. Map of the 10 most important secondary nesting beaches for leatherback turtles along the Pacific coast of Costa Rica based on aerial and ground surveys conducted in and (Mayor, 1998). Figure 2. Map of secondary nesting beaches that currently host regular leatherback nesting every year. Secondary nesting beaches are protected as National Parks (NP), wildlife refuges (Ref) or under no official management category (NC).

63 Leatherback turtle secondary nesting beaches in Costa Rica 567 Table 1. Number of female leatherback turtles identified per season on each secondary nesting beaches along the Pacific coast of Costa Rica. The nesting season starts in October of the year shown and extends until the following March. Five year mean refers to the mean (± SD) number of females in the last five years. Season Naranjo Cabuyal Nombre de Jesús Ostional Caletas year mean ± SD 3.7 ± ± ± ± ± 0.5 Table 2. Number of leatherback turtle nests registered per season on secondary nesting beaches along the Pacific coast of Costa Rica. The nesting season starts in October of the year shown and extends until the following March. Five year mean refers to the mean (± SD) number of nests in the last five years. *Number of nests estimated based on body pit counts and 90% nesting success. Season Naranjo *Cabuyal *Nombre de Jesús Ostional Caletas yr mean ± SD 13.3 ± ± ± ± ± 5.2 Nicoya peninsulas and is currently constricted to around an area about ~50 km north and south of PNMB. We defined a secondary beach as a location where turtles nest regularly but in lower numbers than at the index sites. Consequently, beaches that do not receive nests regularly (i.e., every year), may not qualify as secondary sites. Since the mean annual number of turtles registered at Caletas over five nesting seasons was lower than one turtle, we think that nesting at this site may have become sporadic. Nesting levels likely depend on the overall nesting numbers at the population level, which are currently very low. Thus, we cannot be

64 568 Latin American Journal of Aquatic Research certain that Caletas has become a site with sporadic nesting, considering its importance as a secondary beach in the past. However, the lack of turtles over three consecutive years and the low mean annual number of females per year, serves as an indication that the beach may be in the process of becoming a site with sporadic nesting. Additionally, as the exchange rate registered at Caletas was the lowest among secondary sites, protection of nests on this beach may be essential, as future recolonization could be difficult. According to metapopulation theory, sites where subpopulations are extirpated and become extinct could be recolonized with immigrants from areas with extant subpopulations (Ebenhard, 1991). Movement of individuals between patches normally occurs from those areas that are of better quality and higher abundance (source areas) toward areas of lower quality that are maintained by immigration (sink areas) (Pulliam, 1998). Changes in dispersion patterns between subpopulations caused by shifts in source-sink roles (i.e., sink areas become source areas and vice versa) occur in some species, such as the Audouin s gull (Larus audouinii) (Oro, 2003). Thus, it is possible that historical secondary beaches such as those identified by Mayor (1998) in the 1990s that currently receive sporadic nesting, will be recolonized with individuals from index or secondary sites. The quality of beaches as suitable nesting habitat for leatherback turtles could also change overtime resulting in changes in source-sink roles. However, the number of nesting turtles has declined to critically low levels, with only 26 turtles identified nesting at PNMB in both and (historically lowest levels) and <5 turtles per beach per season on secondary beaches in the last five years. This historically low abundance likely jeopardizes the ability for recolonization of former secondary sites. The probability of extinction is very high for all locations, especially for those beaches with extremely low nesting abundance and with low exchange rate, making recolonization very unlikely after extinction. If current dispersion rates and patterns are maintained, we would expect that the sites where leatherback turtles exhibit the lowest exchange rate would be extirpated first. This might have been the case for Caletas, where we identified the lowest exchange rate (7%) historically and the mean number of females nesting per year dropped to less than one turtle per year in the last five years. In contrast to Caletas, but in support of the apparent importance of high exchange of females among beaches, the beach that registered the greatest exchange rate (Ostional) was also the beach with the highest number of females and nests among secondary sites. Figure 3. Trend in number of leatherback turtle nests registered at a) Parque Nacional Marino las Baulas (PNMB), b) Naranjo and c) Caletas. Naranjo and Caletas are the two longest running projects on secondary nesting beaches. As expected, a similar declining trend to the one documented at PNMB was observed at Naranjo and Caletas, the two longest running projects on secondary beaches (Fig. 3). This suggests that the declining subpopulations reflect an overall trend in the metapopulation, rather than a change in nesting distribution, caused by the same drivers of population decline. These drivers have been identified as high rates of mortality related to interactions with fishing gear, particularly in small-scale fisheries in South America (Alfaro-Shigueto et al., 2011), and unsustainable egg harvest on nesting beaches throughout EP leatherbacks nesting range in Mexico and Central America (Sarti- Martinez et al., 2007; Santidrian-Tomillo et al., 2008).

65 Leatherback turtle secondary nesting beaches in Costa Rica 569 This would also apply to leatherback distribution and trends at the regional scale. The nesting populations of Mexico and Costa Rica are genetically indistinguishable, which could be due to imprecise natal homing (Dutton et al., 1999). However, nesting exchange between the two countries is rare. In Mexico, interbeach exchange is common but the decline in nesting turtle abundance nationwide has mirrored that of Costa Rica (Sarti-Martínez et al., 2007; Wallace et al., 2013). We found that there was some exchange of females between secondary beaches (7% to 28%). Beach exchange was previously reported to occur within index beaches of PNMB (Santidrián-Tomillo et al., 2007) and is frequent in populations of Atlantic leatherback turtles, where they can move between beaches separated by several hundred km located in different countries, such as in between French Guiana and Surinam (Girondot et al., 2007) and within Caribbean countries (Troëng et al., 2004; Ordoñez et al., 2007). In addition to beach exchange by nesting females, there could be country wide and regional wide mixing between life stages. For example, a vast majority of females nesting on secondary beaches remained nesting on the same beach, but these females could have been hatchlings produced elsewhere. Female hatchlings produced at one beach could return to nest on a different beach as adults, or male turtles could mate with females that nest on different beaches, facilitating gene flow between sites. Conservation implications Comprehensive, long-term monitoring of several nesting sites that all serve the same population is critical for informing population-scale conservation strategies. It is possible to develop monitoring protocols to rapidly assess nesting abundance and trends on multiple sites (Delcroix et al., 2013). Aerial surveys, boat surveys, and drones can also enable scientists to quickly assess nesting activity over large regions (Witt et al., 2009; Bevan et al., 2015; Metcalfe et al., 2015). Optimizing efforts to increase coverage, while reducing time spent on each beach is especially desirable when funding is limited. However, from a conservation perspective, it is critical to maintain a permanent presence on beaches where poaching pressure is high. Monitoring programs on index beaches have been essential elements of conservation efforts for protecting the most important sea turtle nesting sites, and also provide a platform for generating estimates of population trends. However, given the dire status of the EP leatherback population, protection of index beaches is insufficient by itself as a beach-based conservation strategy without expanding efforts to secondary beaches. Marine Protection Areas (MPAs) designed to protect index beaches might not always work at recovering populations of sea turtles because pressures can change over time and space (Nel et al., 2013). The effect of localized conservation efforts on some sites for the overall population is not clear. The level of protection varied among nearby leatherback nesting sites. Secondary beaches were either protected as National Parks, Wildlife Refuges or lacked official protection. PNMB was created in 1991 and ratified in 1995 to protect the site with the highest abundance of leatherback turtle nests in Costa Rica and thus, funding was accordingly allocated to its preservation over the years. However, secondary beaches could also contribute to overall population status (e.g., changes in source-sink roles emphasize the value of small sites, Oro, 2003). Thus, the lack of management on some secondary sites could have detrimental implications for the leatherback turtles in PNMB, and the metapopulation as a whole. Egg poaching is a common threat to sea turtles in Costa Rica and was one of the main drivers of the population collapse at PNMB. Approximately, 90% of leatherback clutches were poached for ~20 years before the Park was established (Santidrian-Tomillo et al., 2008). This anthropogenic source of mortality also contributed to the population decline of leatherback turtles in Mexico (Sarti-Martínez et al., 2007). Levels of poaching in Mexico have been consistently reduced over the years although, poaching pressure is still high and effective conservation depends on human presence on the nesting beaches (López & Sarti, 2016). Likewise, egg poaching was gradually eradicated at PNMB following establishment of the National Park, but continued on most secondary beaches and still occurs on some of them. Although protection of nests has been effective in controlling poaching within PNMB, failure to reduce or eliminate this threat at other beaches where leatherbacks nest also hinders the recovery of the nesting population at PNMB and the population overall. Although, healthy sea turtle populations should be able to sustain some level of poaching (i.e., natural predation of eggs), the combination of very high levels of human egg poaching and high rates of fisheries induced mortality is unsustainable and has had an overwhelming effect that has brought the EP leatherback turtle to the brink of extinction. Other threats such as climate change and urban development of nesting beaches could additionally impact EP populations of leatherback turtles as these become depleted preventing recoveries (Saba et al., 2012; Roe et al., 2013). Finally, the total disappearance of nesting beaches and not only declining numbers, also have an effect on sea turtle populations, especially since low nesting

66 570 Latin American Journal of Aquatic Research numbers make recolonization difficult. This type of disappearance occurred in the Caribbean where some historically large and small nesting populations were extirpated (McClenachan et al., 2006). The green turtle (Chelonia mydas) population of the Cayman Islands, once one of the largest populations in the world, was extirpated in the 1800s. Despite the implementation of protection 200 years ago, the number of nesting turtles has not yet recovered and nesting is currently sporadic (Aiken et al., 2001). Likewise, the large population of leatherback turtles that nested at Terengganu, Malaysia in the 1950s is currently considered functionally extinct and it only receives sporadic nesting (Chan & Liew, 1996; Benson et al., 2015). Unfortunately, we do not yet understand metapopulation ecology of leatherback turtles and the number of turtles is too low in the EP to identify patterns properly. The study of population dynamics at the metapopulation level of Atlantic leatherback turtles could help filling these gaps in knowledge. For example, differences among subpopulations could exist in beach exchange, immigration rates, fluctuations in nesting numbers and changes in source-sink roles. This information would help in the development of more effective conservation plans at the metapopulation level. As specific information on the EP metapopulation is currently lacking, protection should include as many beaches as possible and conservation strategies should pay attention to the complete loss of nesting sites where nesting abundance is currently low because local extinction may be irreversible. Beach protection, although not fully achieved, has been insufficient to prevent population declines of EP leatherback turtles. As interactions with fisheries have likely largely contributed to the population collapse (Spotila et al., 2000), reducing fishery bycath in the ocean is essential for beach protection to be effective. Thus, an integral strategy to protect EP leatherback turtles in the ocean and on the beaches is needed to stop declining trends and allow recoveries. ACKNOWLEDGEMENTS We thank the numerous research assistants and volunteers that have collaborated in beach protection and participated in the data collection at all sites. Research permits were obtained from the Guanacaste Conservation Area (ACG) and the Tempisque Conservation Area (ACT), of Costa Rica. We thank Roger Blanco, Nelson Marín and Roberto Zúñiga for facilitating the processes. Funding was provided by The Leatherback Trust, Rufford Small Grants Programme, International Students Volunteers (ISV), The National Fish and Wildlife Foundation (NFWF) and the U.S. Fish and Wildlife Service (USFWS). The U.S. Fish and Wildlife Service s Marine Turtle Conservation Fund supported the initial workshop in We additionally thank Fundecodes, Ostional National Wildlife Refuge and the communities of Ostional and Cabuyal. REFERENCES Aiken, J.J., B.J. Godley, A.C. Broderick, T. Austin, G. Ebanks-Petrie & G.C. Hays Two hundred years after a commercial marine turtle fishery: the current status of marine turtles nesting in the Cayman Islands. Oryx, 35: Alfaro-Shigueto, J., J.C. Mangel, F. Bernedo, P.H. Dutton, J.A. Seminoff & B.J. Godley Small-scale fisheries of Peru: a major sink for marine turtles in the Pacific. J. Appl. Ecol., 48: Benson, S.R., R. Tapilatu, N. Pilcher, P. Santidrián- Tomillo & L. Sarti-Martínez Leatherback turtle populations in the Pacific Ocean. In: J.R. Spotila & P. Santidrián-Tomillo (eds.). Biology and conservation of leatherback turtles. John Hopkins University Press, Baltimore, pp Bevan, E., T. Wibbels, B.M.Z. Najera, M.A.C. Martinez, L.A.S. Martinez, F.I. Martinez, J.M. Cuevas, T. Anderson, A. Bonka, M.H. Hernandez, L.J. Pena & P.M. Burchfield Unmanned Aerial Vehicles (UAVs) for monitoring sea turtles in near-shore waters. Mar. Turtle Newsl., 145: Chan, E.H. & H.C. Liew Decline of the leatherback population in Terengganu, Malaysia. Chelonian Conserv. Biol., 2: Delcroix, E., S. Bédel, G. Santelli & M. Girondot Monitoring design for quantification of marine turtle nesting with limited effort: a test case in the Guadeloupe archipelago. Oryx, 48: Dutton P.H., B.W. Bowen, D.W. Owens, A. Barragan & S.K. Davis Global phylogeography of the leatherback turtle (Dermochelys coriacea) J. Zool., 248: Ebenhard, T Colonization in metapopulations: a review of theory and observation. Biol. J. Linn. Soc., 42: Girondot, M., M.H. Godfrey, L. Ponge & P. Rivalan Modeling approaches to quantify leatherback nesting trends in French Guiana and Surinam. Chelonian Conserv. Biol., 6: Hamann, M., M.H. Godfrey, J.A. Seminoff, K. Arthur, P.C.R. Barata, K.A. Bjorndal, A.B. Bolten et al Global research priorities for sea turtles: informing management and conservation in the 21 st century. Endang. Species Res., 11: Hanski, I Metapopulation dynamics. Nature, 396:

67 Leatherback turtle secondary nesting beaches in Costa Rica 571 Hanski, I. & M. Gilpin Metapopulation dynamics: brief history and conceptual domain. Biol. J. Linn. Soc., 42: López, K. & L. Sarti (comps.) Proyecto Laúd: conservación de la tortuga laúd Dermochelys coriacea en el Pacífico mexicano. Temporada de anidación Dirección de Especies Prioritarias para la Conservación. CONANP-SEMARNAT. Kutzari, Asociación para el Estudio y Conservación de las Tortugas Marinas, 22 pp. Mayor, P Nesting of the leatherback turtle (Dermochelys coriacea) at the las Baulas and Santa Rosa National Park and population estimations for the entire Pacific coast of Costa Rica. Diploma Thesis, Swiss Federal Institute of Technology (ETH), Zurich Department of Environmental Sciences, 59 pp. McClenachan, L., J.B.C. Jackson & M.J.H. Newman Conservation implications of historic sea turtle nesting beach loss. Front. Ecol. Environ., 4: Metcalfe, K., P.D. Agamboué, E. Augowet, F. Boussamba, F. Cardiec, J.M. Fay, A. Formia et al Going the extra mile: ground-based monitoring of olive ridley turtles reveals Gabon hosts the largest rookery in the Atlantic. Biol. Conserv., 190: Miller, J.D Reproduction in sea turtles. In: P.L. Lutz & J.A. Musick (eds.). The biology of sea turtles. CRC Press, Boca Raton, pp Nel, R., A.E. Punt & G.R. Hughes Are coastal protected areas always effective in achieving population recovery for nesting sea turtles? PLoS ONE, 8(5): e Ordoñez, C., S. Troëng, A. Meylan, P. Meylan & A. Ruiz Chiriqui beach, the most important leatherback nesting beach in Central America. Chelonian Conserv. Biol., 6: Oro, D Managing seabird metapopulations in the Mediterranean: constraints and challenges. Sci. Mar., 67: Pulliam, H.R Sources, sinks, and population regulation. Am. Nat., 132: Reina, R.D., P.A. Mayor, J.R. Spotila, R. Piedra & F.V. Paladino Nesting ecology of the leatherback turtle, Dermochelys coriacea, at Parque Nacional Marino Las Baulas, Costa Rica: to Copeia, 2002: Roe, J.H., P.R. Clune & F.V. Paladino Characteristics of la leatherback nesting beach and implications for coastal development. Chelonian Conserv. Biol., 12: Saba, V.S., C.A. Stock, J.R. Spotila, F.V. Paladino & P. Santidrián-Tomillo Projected response of an endangered marine turtle population to climate change. Nature Climate Change, 2: Santidrián-Tomillo, P., E. Vélez, R.D. Reina, R. Piedra, F.V. Paladino & J.R. Spotila Reassesment of the leatherback turtles (Dermochelys coriacea) nesting population at Parque Nacional Marino Las Baulas, Costa Rica: effects of conservation efforts. Chelonian Conserv. Biol., 6: Santidrian-Tomillo, P., V.S. Saba, R. Piedra, F.V. Paladino & J.R. Spotila Effects of illegal harvest of eggs on the population decline of leatherback turtles in Las Baulas Marine National Park, Costa Rica. Conserv. Biol., 22: Sarti-Martínez, L., A.R. Barragán, D. García Muñoz, N. García, P. Huerta & F. Vargas Conservation and biology or the leatherback turtle in the Mexican Pacific. Chelonian Conserv. Biol., 6: Spotila, J.R., R.D. Reina, A.C. Steyermark, P.T. Plotkin & F.V. Paladino Pacific leatherback turtles face extinction. Nature, 405: Scientific Advisory Board (SWOT) The State of the World s Sea Turtles (SWOT). Minimum data standards for nesting beach monitoring, version 1.0. Handbook, 28 pp. Troëng, S., D. Chacón & B. Dick Possible decline in leatherback turtle Dermochelys coriacea nesting along the coast of Caribbean Central America. Oryx, 38: Wallace, B.P., M. Tiwari & M. Girondot Dermochelys coriacea. The IUCN Red List of Threatened Species 2013: e.t6494a [ dx.doi.org/ /iucn.uk rlts.t6494a en]. Reviewed: 20 July Witt, M.J., B. Baert, A.C. Broderick, A. Formia, J. Fretey, A. Gibudi, G. Avery M. Mounguengui, C. Moussounda, S. Ngouessono, R.J. Parnell, D. Roumet, G.P. Sounguet, B. Verhage, A. Zogo & B.J. Godley Aerial surveying of the world s largest leatherback turtle rookery: a more effective methodology for largescale monitoring. Biol. Conserv., 142: Received: 28 July 2016; Accepted: 9 November 2016

68 Lat. Am. J. Aquat. Res., 45(3): , 2017 Genetic characterization of Mexican Pacific hawksbill turtle 555 Sea Turtle Research and Conservation in Latin America Jeffrey Mangel, Jeffrey Seminoff, Bryan Wallace & Ximena Vélez-Suazo (Guest Editors) DOI: /vol45-issue3-fulltext-5 Research Article Genetic characterization of the Critically Endangered hawksbill turtle (Eretmochelys imbricata) from the Mexican Pacific region Tania Zuñiga-Marroquin 1,2 & Alejandro Espinosa de los Monteros 1 1 Departamento de Biología Evolutiva, Instituto de Ecología, A.C., Xalapa, Veracruz, México 2 Departamento de Ciencias Agropecuarias, Universidad del Papaloapan, Oaxaca, México Corresponding author: Alejandro Espinosa de los Monteros (alejandro.espinosa@inecol.mx) ABSTRACT. The hawksbill turtle (Eretmochelys imbricata) is a Critically Endangered species and has been a species of interest for decades. Only in recent years attention has been focused on the populations of the Eastern Pacific Ocean. We present a genetic characterization of this species in the Mexican Pacific, based on mitochondrial DNA sequences. Six localities were sampled along the Pacific Coast, from the Gulf of California to Chiapas, between 2002 and Seventeen individuals found in marine habitats at six localities and six nests laid at three nesting sites were sampled along the Mexican Pacific. Our results show five haplotypes of 766 bp, three previously identified and two that to date were not reported. Genetic diversity indices indicate moderate to low variation for this region. Even with the small sample size reported here, our results show important relationships between the Mexican Pacific hawksbills and nesting populations of Central America and foraging areas along the Eastern and Indo-Pacific. These results, along with updated information on ecology and behavior, are essential for the future approach to conservation and management programs resulting in the recovery of this species in the Eastern Pacific. Keywords: genetic diversity, mitochondrial DNA, hawksbill turtle, endangered species, Eastern Pacific. INTRODUCTION Hawksbill turtles (Eretmochelys imbricata) are medium sized sea turtles with a circumglobal distribution throughout tropical and, to a lesser extent, subtropical waters of the Atlantic, Indian and Pacific oceans. Like other sea turtles, adult hawksbills migrate between foraging and nesting habitats; but are unique in their spongivory and algivory foraging behavior (Meylan, 1988; Leon & Bjorndal, 2002; Bell, 2013). They usually nest alone on sandy beaches, primarily under vegetation and in some cases, in mangrove estuary habitats (Gaos et al., 2012). Post-hatchlings, small juveniles (<20 cm carapace length), and migrating animals are found in pelagic areas, while larger juveniles and adults forage in benthic habitats that include coral reefs, other hard bottom habitats, sea grasses, algal beds, mangrove bays, creeks, and mud flats (Leon & Bjorndal, 2002; Gaos et al., 2012; Bell, 2013). Bowen et al. (2007) suggested that some immature hawksbills could settle in foraging areas close to their natal beaches where they will come back to nest. This is due to the phylopatric characteristic of sea turtle species (Mortimer & Donnelly, 2008). Recent studies in Eastern Pacific populations have shown a novel habitat use by adult hawksbills, in which they settle within confined inshore estuarine bays and even nest in mangrove estuaries (Gaos et al., 2012, 2016; Liles et al., 2015). This species is categorized on the IUCN Red List as Critically Endangered due to a drastic population decline throughout its range (Mortimer & Donnelly, 2008). The larger populations are concentrated in the Caribbean region in the Atlantic basin and in Indonesia and Australia for the Pacific Ocean. In the Eastern Pacific, the hawksbill turtle is particularly threatened, with several populations approaching extirpation (Mortimer & Donnelly, 2008). Until relatively recent times, the species was considered common along the Pacific coast of the Americas (Cliffton et al., 1982). The decline of the hawksbill population in the Eastern Pacific is tied closely to the tortoiseshell trade, because the coastal indigenous populations in the Gulf of California and along the coast of México traded tortoiseshell with the Spaniards during the colonial era (Sáenz-Arroyo et al., 2006). The Caribbean region and Indo-Pacific rookeries are the most studied populations and the best known to date. Based on those studies we know about the general

69 556 Latin American Journal of Aquatic Research nesting (Horrocks & Scott, 1991; Ditmer & Stapleton, 2012; Nishizawa et al., 2012) and feeding (Bowen et al., 1996; Blumenthal et al., 2009; Velez-Zuazo et al., 2008; Nishizawa et al., 2010) behaviors of the species, and the genetic characteristics of hawksbills in both the Indo-Pacific (Nishizawa et al., 2010, 2012, 2016; Tabib et al., 2011, 2014; Vargas et al., 2016) and Atlantic (Reece et al., 2005; Lara-Ruiz et al., 2006; Monzón- Argüello et al., 2010; LeRoux et al., 2012; Trujillo- Arias et al., 2014) basins. These genetic studies have demonstrated strong population structure within the Caribbean basin ( st = 0.64, P < 0.01) and nearly fixed differences between nesting areas in northeast and northwest Australia. The evidence suggests that natal homing predominates, but breeding populations may encompass several proximal nesting sites (Bass et al., 1996; Bowen et al., 2007; Gaos et al., 2016). Recent studies in remote sites like the Persian Gulf, islands in South East Asia and Japan, have revealed the presence of endemic haplotypes in hawksbill nesting and feeding colonies (Nishizawa et al., 2012, 2016; Tabib et al., 2014). Unfortunately, the Eastern Pacific has remained under-studied until the last decade. Seminoff et al. (2003) suggest that the Tres Marías Islands may have been an important breeding area for the hawksbill turtle. Nonetheless, when compared to historical accounts of marine life from the 16 th to the 19 th centuries in the Gulf of California, current data reveal a significant decline of the populations in this region, even though apparently the Eastern Pacific hawksbill was never as common as in others regions (Gaos et al., 2010). There are reports of small numbers of nesting hawksbills along the Eastern Pacific from the Gulf of California to the northern coast of Ecuador; but these nesting events have diminished in the last 20 years (Cornelius, 1982; Seminoff et al., 2003; Gaos et al., 2006; Mortimer & Donelly, 2008). Recent exhaustive monitoring efforts along the Eastern Pacific, led by Iniciativa Carey del Pacífico Oriental (ICAPO) through its network collected historical data ( ) of the presence of hawksbill turtles in this region, identifying six countries with nesting sites, with El Salvador, Mexico and Costa Rica identified as the most important, and eight with in-water sightings, with major abundances in Mexico, Ecuador and Colombia. From 2007 to 2009, through its monitoring efforts, the ICAPO network reported 540 verified nests, of which 430 (79.6%) occurred in El Salvador and with lower percentages in Costa Rica, Ecuador, Nicaragua and Mexico. However, most reports (63%) of at sea observations of hawksbills occurred in Mexico (Gaos et al., 2010). These findings help set the baseline for implementing monitoring projects both in nesting and foraging sites along the Eastern Pacific, and highlighted the importance of strengthening the participation of key people such as fishermen, local communities, ecotourism operators and of promoting efforts to identify, study and protect nesting sites and in-water aggregations and identify new ones. Since 2010, several studies have focused on the Eastern Pacific hawksbill. Liles et al. (2011) recorded more than 300 nesting events on the beaches of El Salvador in the 2008 season. Trujillo-Arias et al. (2014) identified two haplotypes from seven juveniles found in Colombian Pacific waters. Gaos et al. (2016) report the first genetic characterization of nine hawksbill rookeries in the Eastern Pacific region, reporting a low genetic and haplotypic diversity compared with Indo- Pacific and Caribbean populations, but found strong stock structure among the main rookeries in Central America. They included eight samples from Costa Careyes in Mexico, all with the same haplotype (EiIP33); this haplotype was found widespread throughout Central America and the Indo-Pacific nesting sites (Vargas et al., 2016) and is present in Southeast Asia foraging grounds (Nishisawa et al., 2016). ICAPO and the Marine Turtle Specialist Group (MTSG) emphasized the need to develop studies in the Eastern Pacific region to help better understand the nature of this population and its current decline, as this region is likely one of the most endangered sea turtle populations (Mortimer & Donnelly, 2008). The goal of this paper was to characterize the genetic identity of the Mexican Pacific hawksbill at nesting and foraging sites, and to investigate their relationship to rookeries recently characterized in the Eastern Pacific and others regions. MATERIALS AND METHODS Collection sites Between 2002 and 2007, hawksbill turtle localities were surveyed along the Pacific coast of Mexico, chosen based on local knowledge of the occurrence of the species. For nesting sites (i.e., Oaxaca, Guerrero and Jalisco) samples consisted of collecting a small amount of blood from live hatchlings or a small amount of muscle from dead hatchlings recovered from nests, sampling one hatchling per nest. In marine localities (i.e., Chiapas, Oaxaca, Guerrero, Jalisco, Sinaloa and Gulf of California) hawksbill turtles were located and captured by hand during daytime snorkeling or scuba diving surveys (Guerrero and Jalisco), stranded individuals (Gulf of California) or in captivity (Chiapas, Oaxaca, Sinaloa). Blood (about 1 ml) was taken from the dorsal cervical sinus of individuals, following the protocol

70 Genetic characterization of Mexican Pacific hawksbill turtle 557 outlined by Owens & Ruiz (1980) and stored in a lysis buffer (ph 8) (Longmire et al., 1997); tissues were stored in 90% ethanol for transport to the laboratory. After obtaining samples, juveniles and adults individuals were tagged using standard metal tags. Molecular data Approximately 2 mm of collected tissue was ground in Chelex 5% (w/v solution) for total genomic DNA extraction following the method suggested in Singer- Sam et al. (1989). A fragment of the mtdna control region (837 bp) was amplified with polymerase chain reaction (PCR) using primers LTEi9 (5 -gggaataatcaa aagagaagg-3 ) and H950 (5 -gtctcggatttaggggttt-3 ) and internal primers (DLint1H and DLint2L) developed by Abreu-Grobois et al. (2006). DNA amplification was conducted in Peltier-effect thermocyclers (ABI GeneAmp PCR system 2400) using the following parameters: one initial cycle at 95 C for 120 s, followed by 35 cycles of the following temperature regime: 30 s at 95 C, 30 s at 45 C, and 150 s at 72 C. In addition, we included a final extension step of 2 min at 72 C. All PCR reactions were conducted along with positive and negative controls to detect potential false positives because of contamination. Successful amplifications were purified using the GeneJET PCR Purification Kit (Thermo Scientific). Purified PCR products were subjected to cycle sequencing using the ABI Prism BigDye Terminator v3.1 Cycle Sequencing Kit, following the protocol suggested in the kit instructions. The excess of Taq dideoxy terminators was removed with Centri-Sep spin columns (Princeton Separations Inc.) in a variable speed microcentrifuge at 2500 rpm for 2 min. Final purifications were dried down in a vacuum centrifuge and suspended in 25 ml of loading solution. Sequencing products were subjected to capillary electrophoresis in the ABI Prism 310 DNA Sequencer (PerkinElmer, Waltham, MA, USA). Fragments were sequenced and evaluated for both forward and reverse reactions to ensure accurate data collection. Sequence files were analyzed and aligned using Sequencher v5.0 (Gene Codes Corp., Ann Arbor, MI, USA). Haplotypes were identified and assigned by collapsing the sequences using the computer software TCS v1.21 (Clement et al., 2000) and then searching in the database on GenBank ( for sequences within our reading frame. To assure broad comparisons of our results with previous studies, our alignment of the Control Region sequences had to be trimmed to a length of 766 bp (LeRoux et al., 2012). Each haplotype was compared to previously assigned haplotypes and named using the standardized nomenclature conventions recently established for the hawksbill turtles (LeRoux et al., 2012; Gaos et al., 2016; Nishizawa et al., 2016; Vargas et al., 2016). Molecular analyses The dataset was divided into two groups: the Nesting group (i.e., samples obtained from hatchlings at nesting sites), and the Marine group (i.e., samples obtained from individuals captured at marine habitats). This division was established because of the migratory behavior of the species. Therefore, the origin of the individuals caught in marine sites is unknown. The genetic variation for each group was evaluated using empirical descriptive values such as haplotype diversity (h), nucleotide diversity ( ) (Nei, 1987) and segregated sites (S) (Nei & Kumar, 2000). These values were obtained using the program DNAsp v5.0 (Rozas et al., 2003). We used the program TCS ver1.21 (Clement et al., 2000) to infer minimum-spanning (parsimonybased) networks for the haplotypes. RESULTS In spite of five years of collection effort, only 23 samples were obtained (Table 1). For the 23 individuals, we sequenced an 837 bp fragment that spanned 40 bp of the trna-pro and 797 bp of the 5 end of the mitochondrial control region. This region included the 380 bp and 766 bp segments previously sequenced in population genetic analyses of hawksbill turtles from other geographic zones, incorporating two new polymorphic sites for this extended segment. This survey identified six segregated sites among the sequences, four were transitions (G-A), and two were transversions (T-A). This variation recovers eight 837 bp undocumented haplotypes from the Mexican Pacific. The overall base composition for this fragment was A = 32.7%, C = 20.9%, G = 13.3%, and T = 33.1%. The eight 837 bp haplotypes identified were trimmed to 766 bp to be able to compare with previous studies, resulting in five haplotypes, three of which have been previously identified at rookeries and two of which that have not (Table 2). The three haplotypes previously identified, EiIP33, EiIP23 and EiIP106 (GenBank Accession Numbers KT934080, KT and KR respectively), are also found in nesting rookeries in Central America (Gaos et al., 2016); EiIP33 and EiIP23 have been identified in Indo-Pacific rookeries as well (Vargas et al., 2016), with EiIP33 being widespread in these two regions. Haplotypes EiIP33 and EiIP106 were identified in foraging grounds, EiIP33 at three Southeast Asian localities (Nishiwaza et al., 2016) and EiIP106 along the Colombian Pacific (Trujillo-Arias et al., 2014). The two new haplotypes identified, EiIP132 and EiIP133

71 558 Latin American Journal of Aquatic Research Table 1. Number and maturity stage of the hawksbill turtles sampled at nesting and marine sites at localities along the Mexican Pacific. The letter beside the sample size indicates maturity stage, H: hatchling, DE: dead embryos, I: immature, A: adults. CCL indicates curved carapace lengths average. Sites Marine sites Nesting sites Chiapas 2 (I) 0 Oaxaca 4 (I) 2 (H) Guerrero 3 (I) 2 (DE) Jalisco 2 (I) 2 (H) Sinaloa 4 (A) 0 Gulf of California 2 (I) 0 CCL 430 mm 87.5 mm (GenBank Accession numbers KJ and KJ ) were found in samples from Oaxaca and the Gulf of California, respectively. Genetic diversity at nesting sites The data gathered from six samples collected at nesting sites revealed the presence of two 766 bp haplotypes, EiIP23 and EiIP33 (Table 2). The genetic indices for these sites show moderate nucleotide diversity ( = ), and an average pairwise difference between sequences of k = 0.60 (Table 3). The parsimony network (Fig. 1a) shows the two haplotypes separated by one mutation. Both haplotypes have been identified in nesting sites in Central America (Gaos et al., 2016) and the Indo-Pacific region (Vargas et al., 2016). Gaos et al. (2016) identified the haplotype EiIP33 for one Mexican nesting rookery. Genetic diversity in marine survey The analysis of the 17 individuals sampled from marine surveys showed five haplotypes (Table 2). The genetic diversity indices ( = and k = 0.941; Table 3) were similar to those estimated for the haplotypes found at the nesting site. Three of the five haplotypes (EiIP23, EiIP33 and EiIP106) have been identified for nesting sites in Central America (Gaos et al., 2016), with EiIP23 and EiIP33 also found in the Indo-Pacific region. Haplotypes EiIP33 and EiIP106 were identified in foraging grounds in Southeast Asia and Colombian Pacific respectively (Trujillo-Arias et al., 2014; Nishizawa et al., 2016). Two new haplotypes EiIP132 and EiIP133 were identified and registered in GenBank under the accession numbers: KJ and KJ603534, respectively. Haplotype EiIP23 was the most common and widespread, observed in nine samples (53% of all marine samples) along the Mexican Pacific, from the Gulf of California to Oaxaca, followed by EiIP33 (23.5%) and EiIP106 (11.8%). Haplotypes EiIP132 and EiIP133 were found in only one juvenile hawksbill turtle at each of two sites; Oaxaca and the Gulf of California. The parsimony network (Fig. 1b) shows a pattern that supports a recent population expansion. The more common and widespread haplotype EiIP23 is associated with several closely related haplotypes. Based on the root probability criterion (Castello & Templeton, 1994) haplotype EiIP23 is identified as ancestral for the Mexican Pacific hawksbill. DISCUSSION Efforts at studying Eastern Pacific hawksbill turtles have suffered from low sample representation, with most samples coming from juveniles and stranding captures. Seminoff et al. (2003) reported 27 captures in four years along the Gulf of California, Gaos et al. (2006) reported eleven observations in nine years in Costa Rica. Trujillo-Arias et al. (2014) reported only seven juveniles from Pacific Colombia. In the Mexican Pacific, specifically, reports have been scarce. Gaos et al. (2016) in the study with the highest sample size to date in the Eastern Pacific (n = 269), only gathered eight samples from one locality in Mexico, identifying one haplotype. One explanation for this low sample size in Mexico is that in the Eastern Pacific this species occupies inshore estuaries, and even nests in mangrove estuaries. This is a recently described behavior (Gaos et al., 2012, 2016), which make them difficult to discover, and to sample nesting individuals. Additionally, the extreme difficulty in gathering hawksbill tissue samples (especially from nesting individuals) is a direct result of the low density of the species in the Eastern Pacific. Regardless, this study represents the first assessment of the genetic identity of this species at Mexican Pacific foraging sites and expands the knowledge about the genetic identity at nesting rookeries. The observed number of haplotypes at nesting sites within this geographic zone is important and similar to the number of haplotypes identified in others regions with low sample size. For instance, Gaos et al. (2016) identified three haplotypes in ten samples from the Osa Peninsula in Costa Rica and one haplotype from one sample from Azuero Peninsula in Panama; both sites have an estimated number of nesting females similar to that estimated for Mexico in the same study. Regardless, through their attempts to improve the sample size and find specific haplotypes or genetic structure from these nesting rookeries, the results of these studies show the need for and importance of developing local conservation programs for the recovery of these small populations.

72 Genetic characterization of Mexican Pacific hawksbill turtle 559 Table 2. Distribution and frequency of the haplotypes recovered for the hawksbill turtle along the Pacific coast of Mexico. N, M indicate Nesting and Marine sites respectively. *Haplotypes not previously identified in a nesting or foraging site. Chis: Chiapas, Oax: Oaxaca, Gue: Guerrero, Jal: Jalisco, Sin: Sinaloa, GC: Gulf of California. Position of the segregated site Frequency in each population Chis Oax Oax Gue Gue Jal Jal Sin GC GeneBank Haplotype M N M N M N M M M Accesion number EiIP23 A G A KT EiIP33 G KT EiIP106 A G KR EiIP132* G G KJ EiIP133* A KJ Table 3. Summary of population genetic statistics for the hawksbill in the Mexican Pacific. Parameter Nesting sites Marine sites Total Sample size Number of sites Number of haplotype Segregating sites Nucleotide diversity Average pairwise differences It is possible that the confluence of multiple migratory stocks in the Mexican Pacific may explain the haplotype diversity observed in hawksbill turtles sampled in marine sites. The information gathered here and in other recent studies let us provide putative place of birth for the specimens carrying haplotypes identified in the marine specimens and shared with nesting sites. Individuals widespread in Mexican Pacific waters carrying haplotype EiIP23 and EiIP33 could have their origin in Mexican nesting sites, but some of them could have migrated from the close Central American nesting sites (Gaos et al., 2016). Haplotype EiIP23 recovered widely in marine and nesting localities in Mexico is also found in rookeries from Solomon Islands (Vargas et al., 2016) and Central American (Gaos et al., 2016) nesting colonies. Haplotype EiIP33, found in nesting and marine surveys in Mexico, was found widespread in Indo-Pacific rookeries (Vargas et al., 2016) and the Persian Gulf (Tabib et al., 2011). Also, this haplotype was the most common haplotype in Central American rookeries (Gaos et al., 2016) and was identified in foraging grounds in Southeast Asia (Nishizawa et al., 2016). Haplotype EiIP106 has been identified in nesting individuals from El Salvador and Nicaragua in Central America, so the juveniles sampled in Oaxaca and Chiapas carrying this haplotype, could have their origin in those rookeries (Gaos et al., 2016). This haplotype was also identified at foraging grounds along the Colombian Pacific (Trujillo-Arias et al., 2014), suggesting that even when hawksbills nest at Central American rookeries, their foraging range extends from Mexico to Colombia. Finally, the two marine haplotypes, EiIP132 and EiIP133 have not been identified at any nesting sites; however, it is reasonable to assume that those individuals could come from some nesting colony in Mexico or Central America. This is supported by their close relationship to the haplotypes recovered from these nesting areas. This genetic relationship between Eastern Pacific hawksbills and those from other Pacific regions supports the hypothesis that some of the hawksbill hatchlings or juveniles from the Eastern Pacific could be moving to the Indo-Pacific, Persian Gulf and Southeast Asia for foraging, as this happen in others species. The recent discovery of an important nesting colony of hawksbills in El Salvador (Liles et al., 2011; Gaos et al., 2016) and others nesting rookeries suggests that the demographic dynamics of the hawksbill turtle in the Eastern Pacific is complex, and that further sampling efforts are necessary in order to understand the migration and dispersal patterns of this species. Even with the increased interest shown in the Eastern Pacific hawksbill in the last few years (Seminoff

73 560 Latin American Journal of Aquatic Research a b This project was made possible by the outstanding contributions of Instituto de Ecología, AC and CONACyT México. We are very thankful to the following individuals and institutions for supplying tissue samples: Centro Mexicano de la Tortuga, Universidad del Mar, Samantha Karam, Carolina Cidell, Felipe Becerril, Quiyari Santiago, Luis Mendoza, Susan Gardner, Alejandro Peña, Ana María Rivera and Alberto Abreu-Grobois. Especially we thank to the many fishermen who help in collecting samples and other fieldwork. TZM received a Doctoral Scholarship from the Mexican government (CONACyT N ). TZM submits this manuscript as partial fulfillment of the requirements for the Doctor of Science degree, at the Instituto de Ecología, AC. We recognize and appreciate the improvements to this manuscript thanks to suggestions and comments by Jeffrey Mangel and two anonymous reviewers. REFERENCES Figure 1. Minimum-length networks for the haplotypes of Eretmochelys imbricata in the Mexican Pacific. a) Nesting sites, and b) marine sites. Pie size is proportional to the frequency of the haplotype. Lines represent mutational steps between haplotypes. Position of the segregate site is indicated above the line. et al., 2003; Gaos et al., 2010, 2012, 2016; Liles et al., 2011, 2015), more extensive population surveys remain necessary. The results presented here provide an idea of the genetic identity of the nesting and foraging Mexican Pacific hawksbill turtles, but also shows the important role that Mexican Pacific waters play as habitat for juveniles and adults hawksbill turtles in the Eastern Pacific. Mexico has important estuarine systems along its Pacific coast that have not been explored as potential nesting sites and that, given the nesting behavior observed in neighboring populations, is required (Gaos et al., 2012). ACKNOWLEDGMENTS Abreu-Grobois, F.A., J.A. Horrocks, A. Formia, P. Dutton, R. LeRoux, X. Vélez-Zuazo, L. Soares & P. Meylan New mtdna dloop primers which work for a variety of marine turtle species may increase the resolution capacity of mixed stock analyses. In: M. Frick, A. Panagopoulou, A.F. Rees & K. Williams (eds.). Proceedings of the twenty-sixth annual symposium on marine turtle biology and conservation. International Marine Turtle Society, Crete, Greece, pp Bass, A.L., D. Good, K.A. Bjorndal, J.I. Richardson, Z.M. Hillis-Starr, J.A. Horrocks & B.W. Bowen Testing models of female reproductive migratory behaviour and population structure in the Caribbean Hawksbill turtle, Eretmochelys imbricata, with mtdna sequences. Mol. Ecol., 5(3): Bell, I Algivory in hawksbill turtles: Eretmochelys imbricata food selection within a foraging area on the Northern Great Barrier Reef. Mar. Ecol., 34(1): Blumenthal, J.M., T. Austin, C.D. Bell, J.B. Bothwell, A.C. Broderick, G. Ebanks-Petrie, J. Gibb, K. Luke, J. Olynik, M. Orr, J.L. Solomon & B.J. Godley Ecology of hawksbill turtles, Eretmochelys imbricata, on a western Caribbean foraging ground. Chelonian Conserv. Biol., 8(1): Bowen, B.W., W.S. Grant, Z.M. Hillis-Starr, D. Shaver, K.A. Bjorndal, A.B. Bolten & A.L. Bass Mixed-stock analysis reveals the migrations of juvenile Hawksbill turtles (Eretmochelys imbricata) in the Caribbean Sea. Mol. Ecol., 16(1): Bowen, B.W., A.L. Bass, A. Garcia-Rodriguez, C.E. Diez, R. Van-Dam, A.B. Bolten, K.A. Bjorndal, M. Miyamoto & R. Ferl Origin of hawksbill turtles in a Caribbean feeding area as indicated by genetic markers. Ecol. Appl., 6(2): Castelloe, J. & A.R. Templeton Root probabilities for intraspecific gene trees under neutral coalescent theory. Mol. Phylogenet. Evol., 3(2):

74 Genetic characterization of Mexican Pacific hawksbill turtle 561 Clement, M., D. Posada & K.A. Crandall TCS: a computer program to estimate gene genealogies. Mol. Ecol., 9(10): Cliffton, K., D.O. Cornejo & R.S. Felger Sea turtle of the Pacific coast of Mexico. In: K.A. Bjorndal (ed.). Biology and conservation of sea turtles. Smithsonian Institution Press, Washington DC, pp Cornelius, S.E Status of sea turtles along the Pacific coast of Middle America. In: K.A. Bjorndal (ed.). Biology and conservation of sea turtles. Smithsonian Institution Press, Washington DC, pp Ditmer, M.A. & S.P. Stapleton Factors affecting hatch success of hawksbill sea turtles on Long Island, Antigua, West Indies. PLoS ONE, 7: e Gaos, A.R., R. Arauz & I. Yañez Hawksbill turtles on the Pacific coast of Costa Rica. Mar. Turtle Newslett., 112: 14. Gaos, A.R., F.A. Abreu-Grobois, J. Alfaro-Shigueto, D. Amorocho, R. Arauz, A. Baquero, R. Briceño-Dueñas, et al Signs of hope in the Eastern Pacific: International collaboration reveals encouraging status for a severely depleted population of hawksbill turtles Eretmochelys imbricata. Oryx, 44(4): Gaos, A.R., R.L. Lewison, M.J. Liles, V. Gadea, E. Altamirano, A.V. Henríquez, P. Torres, et al Hawksbill turtle terra incognita: conservation genetics of Eastern Pacific rookeries. Ecol. Evol., 6(4): Gaos, A.R., R.L. Lewison, B.P. Wallace, I.L. Yañez, M.J. Liles, W.J. Nichols, A. Baquero, C.R. Hasbún, M. Vasquez, J. Urteaga & J.A. Seminoff Spatial ecology of critically endangered hawksbill turtles Eretmochelys imbricata: implications for management and conservation. Mar. Ecol. Prog. Ser., 450: Horrocks, J.A. & N. Scott Nest site location and nest success in the Hawksbill turtle Eretmochelys imbricata in Barbados, West Indies. Mar. Ecol. Prog. Ser., 69(1): 1-8. Lara-Ruiz, P., G.G. Lopez, F. Santos & L.S. Soares Extensive hybridization in hawksbill turtles (Eretmochelys imbricata) nesting in Brazil revealed by mtdna analyses. Conserv. Genet., 7(5): Leon, Y.M. & K.A. Bjorndal Selective feeding in the hawksbill turtle, an important predator in coral reef ecosystems. Mar. Ecol. Prog. Ser., 245: LeRoux, R.A., P.H. Dutton, F.A. Abreu-Grobois, C.J. Lagueux, C.L. Campbell, E. Delcroix, J. Chevalier, J.A. Horrocks, Z.-M. Hillis-Starr, S. Troeng, E. Harrison & S. Stapleton Re-examination of population structure and phylogeography of hawksbill turtles in the wider Caribbean using longer mtdna sequences. J. Hered., 103(6): Liles, M.J., M. Jandres, W. López, G. Mariona, C.R. Hasbún & J.A. Seminoff Hawksbill turtles Eretmochelys imbricata in El Salvador: nesting distribution and mortality at the largest remaining nesting aggregation in the Eastern Pacific Ocean. Endang. Species Res., 14(1): Liles, M.J., M.J. Peterson, J.A. Seminoff, E. Altamirano, A.V. Henríquez, A.R. Gaos, V. Gadea, J. Urteaga, P. Torres, B.P. Wallace & T.R. Peterson One size does not fit all: importance of adjusting conservation practices for endangered hawksbill turtles to address local nesting habitat needs in the eastern Pacific Ocean. Biol. Conserv., 184: Longmire, J.L., M. Maltbie & R.J. Baker Use of lysis buffer in DNA isolation and its implication for museum collections. Occas. Pap. Mus. Texas Tech. Univ., 163: 1-3. Meylan, A.B Spongivory in hawksbill turtles: a diet of glass. Science, 239(4838): Monzón-Argüello, C., C. Rico, A. Marco, P. López & L.F. López-Jurado Genetic characteri-zation of Eastern Atlantic hawksbill turtles at a foraging group indicates major undiscovered nesting populations in the region. J. Exp. Mar. Biol. Ecol., 387(1-2): Mortimer, J.A. & M. Donnelly Marine turtle specialist group 2007 IUCN Red List status assessment hawksbill turtle (Eretmochelys imbricata). IUCN/SSC Marine Turtle Specialist Group, 112 pp. Nei, M Molecular evolutionary genetics. Columbia University Press, New York, 512 pp. Nei, M. & S. Kumar Molecular evolution and phylogenetics. Oxford University Press, New York, 333 pp. Nishizawa, H., J. Joseph & Y.K. Chong Spatiotemporal patterns of mitochondrial DNA variation in hawksbill turtles (Eretmochelys imbricata) in Southeast Asia. J. Exp. Mar. Biol. Ecol., 474: Nishizawa, H., J. Okuyama, M. Kobayashi, O. Abe & N. Arai Comparative phylogeny and historical perspectives on population genetics of the Pacific Hawksbill (Eretmochelys imbricata) and green turtles (Chelonia mydas), inferred from feeding populations in the Yaeyama Islands, Japan. Zool. Sci., 27(1): Nishizawa, H., J. Okuyama, O. Abe, M. Kobayashi & N. Arai Mitochondrial DNA variation in hawksbill turtles (Eretmochelys imbricata) nesting on Ishigaki Island, Japan. Mar. Turtle Newsl., 132: 1-2. Owens, D. & G. Ruiz New methods of obtaining blood and cerebrospinal fluid from marine turtles. Herpetologica, 36(1): Reece, J.S., T.A. Castoe & C.L. Parkinson Historical perspectives on population genetics and conservation of three marine turtle species. Conserv. Genet., 6(2):

75 562 Latin American Journal of Aquatic Research Rozas, J., J.C. Sánchez-DelBarrio, X. Messeguer & R. Rozas DNASP, DNA polimorphism analyses by the coalecent and other methods. Bioinformatics, 19: Saénz-Arroyo, A., C.M. Roberts, J. Torre, M. Cariño- Olvera & J.P. Hawkins The value of evidence about past abundance: marine fauna of the Gulf of California through the eyes of 16 th to 19 th century travellers. Fish Fish., 7: Seminoff, J.A., W. Nichols, A. Resendiz & L. Brooks Occurrence of hawksbill turtles, Eretmochelys imbricata (Reptilia: Cheloniidae), near the Baja California peninsula, Mexico. Pac. Sci., 57(1): Singer-Sam, J., R.L. Tanguay & A.D. Riggs Use of chelex to improve PCR signals from a small number of cells. Amplifications: a forum for PCR users, 3: 11. Tabib, M., H. Zolgharnein, M. Mohammadi, M.A. Salari- Aliabadi, A. Qasemi, S. Roshani, H. Rajabi-Maham & F. Frootan MtDNA variation of the critically endangered hawksbill turtle (Eretmochelys imbricata) nesting on Iranian islands of the Persian Gulf. Genet. Mol. Res., 10(3): Tabib, M., F. Frootan & M. Askari-Hesni Genetic diversity and phylogeography of hawksbill turtle in the Persian Gulf. J. Biodivers. Environ. Sci., 4(4): Trujillo-Arias, N., D.F. Amorocho, D. López-Álvarez & L.M. Mejía-Ladino Relaciones filogeográficas de algunas colonias de alimentación y anidación de la tortuga carey (Eretmochelys imbricata) en el Pacífico y Caribe Colombianos. Bol. Invest. Mar. Cost., 43: Vargas, S.M., M.P. Jensen, S.Y.W. Ho, A. Mobaraki, D. Broderick, J.A. Mortimer, S.D Whiting, J. Miller, R.I.T. Prince, I.P. Bell, X. Hoenner, C.J. Limpus, F.R. Santos & N.N. Fitz-Simmons Phylogeography, genetic diversity, and management units of hawksbill turtles in the Indo-Pacific. J. Hered., 107(3): Velez-Zuazo, X., W. Ramos, R. Van-Dam, C.E. Diez, F.A. Abreu-Grobois & W.O. McMillan Dispersal, recruitment and migratory behaviour in a hawksbill sea turtle aggregation. Mol. Ecol., 17(3): Received: 1 May 2016; Accepted: 2 November 2016

76 Lat. Am. J. Aquat. Res., 45(3): , 2017 Characterization of the southermost green turtle 5401 Sea Turtle Research and Conservation in Latin America Jeffrey Mangel, Jeffrey Seminoff, Bryan Wallace & Ximena Vélez-Suazo (Guest Editors) DOI: /vol45-issue3-fulltext-4 Research Article Ecology, health and genetic characterization of the southernmost green turtle (Chelonia mydas) aggregation in the Eastern Pacific: implications for local conservation strategies Rocío Álvarez-Varas 1, Juan Contardo 1, Maike Heidemeyer 2,3, Lina Forero-Rozo 1, Beatriz Brito 1,4 Valentina Cortés 1, María José Brain 1, Sofía Pereira 1 & Juliana A. Vianna 5 1 Qarapara Tortugas Marinas Chile NGO, Santiago, Chile 2 Asociación Restauración de Tortugas Marinas PRETOMA, Tibás, Costa Rica 3 Centro de Investigación en Biología Celular y Molecular, Universidad de Costa Rica San Pedro, San José, Costa Rica 4 Facultad de Ciencias Veterinarias y Pecuarias, Universidad de Chile, Santiago, Chile 5 Departamento de Ecosistemas y Medio Ambiente, Facultad de Agronomía e Ingeniería Forestal Pontificia Universidad Católica de Chile, Santiago, Chile Corresponding author: Rocío Álvarez-Varas (ralvarez03@gmail.com) ABSTRACT. Bahía Salado, located in northern Chile (27 41 S, W), is the southernmost foraging ground for the endangered green turtle (Chelonia mydas) in the Eastern Pacific Ocean (EPO). To date, almost no information exists on its current status, nor on its connectivity with nesting rookeries in the EPO. This study aims to inform on the genetic characterization, health and ecology of Bahía Salado s green turtle aggregation in order to provide baseline information for local conservation strategies. We describe population structure and residency times using mark-recapture method. We also examine health parameters (body condition index, blood profile and blood copper-cu and lead-pb concentrations) and regional connectivity through genetic analyses. Our results indicate that this aggregation is composed exclusively of juveniles, with residency times varying between five to sixteen months. Turtles exhibited a very good body condition; however they showed the highest blood concentrations of Cu and Pb described for C. mydas and for almost all sea turtle species. Some biochemistry parameters (albumin, calcium, phosphorus, AST, triglycerides and creatinine) are also the highest ever reported for this species in the region. Analysis of the 770 bp (base pairs) control region of the mitochondrial DNA revealed four haplotypes, suggesting a strong genetic connectivity to the Galapagos rookery. Our study indicates that Bahía Salado s aggregation represents a developmental foraging ground, where juvenile green turtles thrive. Although Bahía Salado s ecosystem seems to be a very suitable habitat for the species, the high levels of Cu and Pb, together with elevated AST, demand further research on the negative impacts of heavy metals on this aggregation. Our results highlight the importance to protect this bay from anthropological activities, evaluate pollution sources and other local threats to this particular coastal ecosystem. We recommend year-round monitoring of the green turtle aggregation and other components of this ecosystem, incorporating participation of local seaweed collectors and the fishing community. Keywords: green turtle, juvenile aggregation, foraging ground, body condition index, heavy metals, blood chemistry, mitochondrial DNA, natal origin, Chile. INTRODUCTION Green turtle (Chelonia mydas) is listed as globally endangered in the IUCN Red List (Seminoff, 2004). In the Eastern Pacific Ocean (EPO), the green turtle is distributed along the west coast of North and South America (Quiñones et al., 2010) and is commonly known as black turtle, due to morphological and color variations (Chassin-Noria et al., 2004). Throughout this paper, C. mydas will be referred to as Eastern Pacific green turtle. Factors driving green turtle distribution in the EPO are numerous and may vary from site to site (Jensen et al., 2012). In their initial life stage, hatchlings are dispersed by ocean currents and remain in the pelagic realm for several years until recruiting to coastal areas (Reich et al., 2007). These areas can be shared with adults or frequented only by juveniles (Amorocho et

77 541 2 Latin American Journal of Aquatic Research al., 2012), constituting a discrete benthic developmental habitat, as stated by Meylan et al. (2011). There is growing evidence for a by-size spatial segregation pattern between green turtle foraging grounds in the EPO (Seminoff et al., 2003; López- Mendilaharsu et al., 2005; Koch et al., 2007; Velez- Zuazo et al., 2014). Particularly in the southern EPO (Peru), Velez-Zuazo et al. (2014) observed latitudinal habitat segregation, where the northern location was composed mainly of sub-adults and adults, and the southern one almost exclusively of juveniles, corresponding to a benthic developmental foraging ground. Although the causes of this segregation remain poorly understood, it seems to be related with resource partitioning, and proximity to rookeries, among other factors (Bjorndal et al., 2000; López-Mendilaharsu et al., 2005; Koch et al., 2007; Meylan et al., 2011; Velez- Zuazo et al., 2014). Green turtle feeding areas are generally composed of individuals recruited from multiple nesting sites (Meylan et al., 2011; Amorocho et al., 2012). Genetic studies carried out in the EPO show that the major contributing rookeries to foraging areas in South America appear to be the Galapagos Islands (Ecuador), Michoacán (Mexico) and Costa Rica, and to a lesser extent, Revillagigedo (Mexico) and Hawaii (United States; Velez-Zuazo & Kelez, 2010; Alfaro-Shigueto et al., 2011; Amorocho et al., 2012; Veliz et al., 2014). Nevertheless, flipper tag returns, as well as genetic and telemetry studies conducted in Peru and Chile, suggest that the Galapagos Archipelago seems to be the principal rookery source for individuals of these South American green turtle foraging grounds (Velez-Zuazo & Kelez, 2010; Alfaro-Shigueto et al., 2011; Veliz et al., 2014; Donoso et al., 2016; Dutton et al., 2016). In Chile, six neritic aggregation areas have been described for green turtles, all of them located in the north of the country: Playa Chinchorro (18 28 S, W), Bahía Chipana (21º19 S, 70º03 W), Bahía Mejillones del Sur (23 05 S, W), Caleta Constitución (23 24 S, W), Poza Histórica de Antofagasta (23 35 S, W), and Bahía Salado (27 41'S, 70 59'W; Grupo Nacional de Trabajo de Tortugas Marinas, unpubl. data). All these habitats present soft-bottom areas with dominant macroalgae in relatively sheltered locations (Veliz et al., 2014; Sarmiento-Devia et al., 2015). Bahía Salado stands out as the southernmost aggregation and the bottom is dominated by macroalgae and an endemic seagrass, Zostera chilensis (Zavala et al., 2009). Historical evidence shows that green turtles are present in Chilean coastal waters year-round (Frazier & Salas, 1984; Bolados-Díaz et al., 2007; Brito et al., 2007). Although data on residency times of this species in local neritic habitats are almost unknown, there are recapture records between six months and three years from Playa Chipana and Bahía Mejillones del Sur, respectively (Bolados-Díaz et al., 2007; Veliz et al., 2014). Studies from other neritic areas remain lacking; nevertheless, Bahía Salado s green turtle aggregation seems to be permanent, at least since the 1980s (Brito et al., 2007). Worldwide, coastal environments have been affected by organic and inorganic pollution stemmed from a wide range of industrial, agricultural and urban sources (Komoroske et al., 2011). This is a particular issue in northern Chile, where trace-metals naturally occur and are also released into the environment by anthropic activities related mainly to mining (Ramirez et al., 2005; Castillo & Valdés, 2011). Sea turtles inhabiting neritic areas may be particularly sensitive to marine pollution due to their delayed maturation and longevity (Komoroske et al., 2011; Camacho et al., 2013). Recent evidence demonstrates that heavy metals decrease the immune response, leading to an increase in disease vulnerability, particularly in sea turtles (Day et al., 2007; Camacho et al., 2013; Carneiro da Silva et al., 2016). Alterations of red blood cell count and biochemistry parameters have also been reported in these species (Day et al., 2007; Camacho et al., 2013). Trace-metals analyses of green turtles blood and carcasses from the Antofagasta Region in northern Chile, revealed above-average concentrations of arsenic (As), copper (Cu), lead (Pb) and mercury (Hg), reflecting industrial mining activity and subsequent heavy metal pollution in the region (Plaza-Araya et al., 2010; Canales-Cerro & Álvarez-Varas, 2015). Unfortunately no studies on pollutants or blood parameters in Bahía Salado s green turtles exist, although this region has been historically impacted by the mining industry (Ramirez et al., 2005; Castillo & Valdés, 2011). Additionally, to a lesser extent, local pollution sources related to vessel movement for macroalgae extraction exist in the bay (SUBPESCA, 2010). Marine pollution does not seem to be the only threat for Bahía Salado s green turtle aggregation. Preliminary data based on fishermen interviews from seven fishing coves adjacent to Bahía Salado, suggested a moderate bycatch rate, mainly associated to gillnets (49% of the 53 interviews; Cortés, unpubl. data). On the other hand, although algae extraction by local communities constitutes the major economic activity in this bay, its high degree of isolation and oceanographic characteristics have attracted big companies to develop non-renewable energy projects (i.e., thermoelectric plants based on coal and natural gas). Despite the above, the current status of this green turtle aggregation remains unknown, thus making it difficult to formulate

78 Characterization of the southermost green turtle 5423 appropriate local conservation plans in order to avoid or mitigate potential population impacts. Based on the latitudinal habitat segregation observed by Velez-Zuazo et al. (2014), we predict that the Bahía Salado s green turtle aggregation will represent a benthic developmental habitat dominated by juvenile individuals. Likewise, we expect a permanent turtle residence reflected in high recapture rates and wide recapture intervals, as reported for other Chilean aggregations of this species. Also, we believe that turtles will exhibit elevated levels of heavy metals in blood, concordant with the local historical pollution, and accordingly, alterations in blood and biochemistry parameters. Finally, taking into account the regional pattern of genetic contribution to foraging grounds in the EPO, it is probable that Bahía Salado s turtles carry haplotypes dominant or endemic to the Galapagos Archipelago, being the most proximate nesting rookery. Here, we provide new information in terms of ecology and health for the southernmost neritic aggregation of Chelonia mydas in the EPO. The overall objectives of the present study were to gather data on the population status of the green turtle in Bahía Salado, northern Chile, to provide baseline information for the development of fact-based local conservation strategies. Specifically, we describe: 1) population structure and residency times, 2) body condition index (BCI), blood parameters and heavy metals in blood as indicators of population health (Labrada-Martagón et al., 2010; Camacho et al., 2013; Suarez-Yana et al., 2015), and ultimately 3) connectivity of Bahía Salado s green turtle aggregation with nesting rookeries of the region using molecular markers. MATERIALS AND METHODS Study area Bahía Salado (27 41'S, 70 59'W) is a bay located in the Atacama Region in northern Chile (Fig. 1). This region is characterized by a semiarid climate with a dense coastal cloud cover. Rain is rare and usually influenced by El Niño Southern Oscillation (ENSO), which manifests itself as a superficial dispersion to the south of equatorial waters with high salinity and high temperatures (Squeo et al., 2008). Bahía Salado is formed by shallow waters, reaching maximum depths up to 10 m. High algae coverage with seagrass (Zostera chilensis) patches near the high tide line can be found on the inner side of the bay where depths reach 3 m maximum (Zavala et al., 2009). Annual sea temperatures oscillate between 13º to 21 C in this area and the highest green turtle density is found at Playa La Hedionda (Álvarez-Varas, unpubl. data, Fig. 1). Sea turtle capture A total of four field trips of 10 days each, were carried out in spring 2013 (October), summer 2014 (March), spring 2014 (November) and summer 2015 (February) completing a period of sixteen months. Green turtles were captured in shallow areas (<2 m depth) using two entanglement nets ( m, mesh size of 35 cm stretched). Nets were set during 8 h per day, and checked constantly from the coastline and by two apnea divers every 30 min. Upon capture, sea turtles were taken to shore to apply a standard monitoring protocol, including morphometry, weighing, tissue and blood sampling and identification. Turtles were tagged on each front flipper using Inconel tags (Style 681, National Band and Tag Company, Newport, Kentucky; Zárate et al., 2013) prior to release (holding time never exceeded 40 min per animal). All captures were authorized by the Chilean Sub- Secretariat of Fishing (SUBPESCA, by its Spanish abbreviation), through a Research Capture Permit granted in April 2013 (Exempt Resolution N 917) and renewed in July Catch per unit effort An estimation of catch per unit effort (CPUE) was calculated for each field trip by dividing the total number of sea turtles caught on each sampling occasion by the number of effort units. One unit effort was equivalent to one 8 h in-water set for two 50 m long nets (Koch et al., 2007). Morphological data The following measurements were taken for each turtle (Eckert et al., 1999): minimum curved carapace length (CCL min), curved carapace length notch to tip (CCL), curved carapace width (CCW), straight carapace length (SCL), straight carapace width (SCW), plastron length (PL), plastron width (PW), head length (HL), head width (HW), tail total length (TTL) and post-cloacal tail length (PTL). Curved and straight measurements were obtained using a metric tape and a calibrated forester s caliper (±0.1 cm, straight measurements), respectively. Body mass was obtained using a spring balance (±100 g). Life stage determination Green turtles were classified in different life stages based on the mean CCL size of nesting females in Galapagos (Zárate et al., 2013). Turtles with CCL <85 cm were classified as juveniles and CCL 85 cm as adults.

79 543 4 Latin American Journal of Aquatic Research Figure 1. Location of the Bahía Salado study area in Chile. Body condition index A body condition index (BCI = body mass 10,000 / SCL 3 ) was calculated to evaluate the relative fatness of captured turtles (Bjorndal et al., 2000). This index was used as an indirect predictor of the nutritional status and/or health condition of the animal (Bjorndal et al., 2000; Velez-Zuazo et al., 2014). Hematology, blood biochemistry analysis and hemoparasite detection Blood samples (5 ml) were collected from each turtle from the dorsal cervical sinus (Mader, 2006). Three ml of blood samples (3 ml) were stored in heparinized tubes (BD Vacutainer, NJ USA, 68 USP) and refrigerated until analysis. Type of processing, storage and detailed analyses of hematological and biochemical variables (including hemoparasites) are shown in Appendix 1 and 2. All laboratory blood analyses were carried out at the Laboratorio de Hematología y Bioquímica Clínica, Facultad de Ciencias Veterinarias, Universidad de Chile, Santiago, Chile. Heavy metals analysis Two ml (2 ml) of the original 5 ml blood sample were also stored in heparinized tubes, placed in cryotubes and maintained in liquid nitrogen (-196ºC). Copper (Cu) and lead (Pb) contents in blood analyses were carried out at the Laboratorio Veterinario Especializado (Vetlab ), Santiago, Chile. Samples were processed using Atomic Absorption Spectrophotometry (AAS) and metal concentrations were expressed in µg g -1. Samples and standards were analyzed using a duplicate, in different series, and were read against target reagents in a spectral range λ nm AAS Shimatzu AA integrated to AWizard Software. For Cu determination, a calibration curve was designed using 10 dilutions with values between µg g -1 which were obtained from a standard pattern of Cu certified concentration (SeronomTM Trace Elements Serum). Samples and standards were read at λ nm, according to the standardized Cu protocol. For Pb, the calibration curve was designed with 10 concentration points in a range of µg g -1, which were obtained from a standard pattern of Pb certified concentration (SeronomTM Trace Elements Whole Blood). Measurements were done at λ nm. Genetic analysis Skin samples (5 mm) were collected from the neck area of each turtle using a sterile scalpel. Samples were stored in ethanol at environment temperature. In order to determine the possible natal origin of Bahía Salado s green turtles, we amplified haplotypes of the mitochondrial DNA (mtdna) control region. Therefore, DNA was isolated using the Aljanabi & Martínez (1997) protocol. The control region (D-loop; approx. 773bp) was amplified using primers LCM15382 (5'- GCTTAACCCTAAAGCATTGGO3') and H950g (5'- GTCTCGGATTTAGGGGTTTGO3') designed by Abreu-Grobois et al. (2006). Reactions were carried out in a total volume of 25 µl with 2 µl DNA, 1 buffer reaction, 200 µm dntp, 0.5 µm of each primer, 0.8 Platinum Taq DNA polymerase (Invitrogen) units and 1.4 mm of MgCl 2. The Polymerase Chain Reaction (PCR) protocol was as follows: 10 min at 95 C, 95 C for 15 s, a touchdown at C for 30 s, 72 C for 45 s, with 2 cycles at each annealing temperature, and 35 amplification cycles of 95 C for 15 s, 50 C for 30 s,

80 Characterization of the southermost green turtle C for 45 s, followed by a final extension period of 30 min at 72 C. The PCR product was visualized using electrophoresis in 1% agarose with red gel. The previously described procedures were carried out at the Laboratorio de Biodiversidad Molecular, Departamento de Ecosis-temas y Medio Ambiente, Pontificia Universidad Católica de Chile, Santiago, Chile. Final products were purified and sequenced bilaterally at Macrogen Inc., Seoul, Korea. Raw sequences were edited and corrected manually using the GENEIOUS version program ( Kearse et al., 2012) and truncated to the standard length of 773 bp. Sequences were aligned using the ClustalX algorithm implemented in GENIOUS, and haplotypes were identified after running a BLAST search implemented in the GenBank database (National Center for Biotechnology Information, USA: NCBI Home page nih.gov). The DnaSP program (Librado & Rozas, 2009) was used to calculate haplotype and nucleotide diversity for the neritic aggregation. RESULTS CPUE, morphological data, life stage and body condition A total of 320 net-set hours (equivalent to 40 unit effort) yielded 14 captured turtles consisting of seven different individuals, four of which were recaptured in a period of five and sixteen months (Tables 1, 2). Overall CPUE for first-time captures was 0.18 turtles per capture unit, thus equaling one turtle every 45 h of netting (Table 1). According to Zárate et al. (2013), all turtles were juveniles and ranged between cm CCL (mean size of 66.5 ± 9.8 cm), and weighed from 19.5 to 76.0 kg (mean of 39.6 ± 20.0 kg; Table 2). The body condition index (BCI) ranged between 1.19 and 2.02 (mean of 1.66 ± 0.28). Morphological and BCI data for each turtle are shown in Table 2. Hematology, blood biochemistry analysis and hemoparasite detection In all cases, blood smear examinations were negative for hemoparasites. Biochemistry analyses showed elevated levels of albumin, calcium, phosphorus, AST, triglycerides and creatinine. Results from hematology and blood biochemistry analyses for Bahía Salado s green turtles and other studies on this species from the EPO are shown in Table 3. Heavy metals analysis Cu showed a mean blood concentration of 2.26 ± 0.10 µg g -1 and Pb of 1.11 ± 0.06 µg g -1, indicating high levels of these metals (Table 4). Genetic analysis The sequences of the seven turtles exhibited three polymorphic sites, all of which were transitions. Four previously described haplotypes were identified in the sequences, where haplotype CmP-4.6 (GenBank accession number: KC ) was found in three individuals, CmP-4.7 (GenBank accession number: KC ) in two individuals and haplotypes CmP (GenBank accession number: KC ) and CmP-4.1 (GenBank accession number: KC ) each in one individual respectively. Most of the analyzed haplotypes were dominant or endemic to the Galapagos Archipelago (Dutton et al., 2014). Haplotype diversity (h) was high with 0.81 ± 0.13 and nucleotide diversity (π) was low with ± DISCUSSION Bahía Salado as a benthic development habitat for Eastern Pacific green turtles Several studies show a small-scale, size-based foraging habitat segregation in green turtles from the northern EPO (Gulf of California, Mexico), where smaller juveniles inhabit relatively protected and shallow areas, and large juveniles and adults generally in deeper habitats (Seminoff et al., 2003; López-Mendilaharsu et al., 2005; Koch et al., 2007). Habitat segregation has also been observed in Peru, with sub-adults and adults occurring at the most northern locations and smaller juveniles dominating foraging areas at the southern coasts (Velez-Zuazo et al., 2014). Historical data indicate that C. mydas in Chilean waters correspond exclusively to juvenile and sub-adult individuals (Sarmiento-Devia et al., 2015), which is congruent with southern Peruvian foraging areas. Bahía Salado, located in northern Chile, harbors the southernmost neritic aggregation of green turtles in the EPO. Our results showed that this aggregation was composed only by juveniles according to Zárate et al. (2013), where sizes varied between 54.0 and 83.1 cm CCL (mean 66.5 ± 9.8 cm) and weights between kg (mean 39.6 ± 20.0 kg). Thus, supporting the benthic developmental hypothesis stated by Meylan et al. (2011), and our prediction on exclusive presence of juveniles in this Chilean green turtle aggregation. Velez-Zuazo et al. (2014) reported individuals with sizes between 44.9 and 84.5 cm CCL (mean 57.7 ± 8.7 cm) at Paracas (~14ºS), southern Peru. On the other hand, in the northern coast of Chile, Veliz et al. (2014) recorded green turtles between 47.0 and 75.7 cm CCL at Playa Chinchorro (~18ºS), and Donoso et al. (2016), from cm CCL at Bahía Mejillones del Sur (~23ºS). With this information, a pattern of latitudinal segregation becomes evident. Southern Peru and northern

81 6545 Latin American Journal of Aquatic Research Table 1. Netting effort and number of turtles caught and recaptured on each sampling date. Catch per unit of effort (CPUE) calculated for first-time captures and total of captures. Field trip date Number of turtles CPUE Capture effort First captures Recaptures Total First captures Total October March November February Total Table 2. Morphological data, body condition index and life stage of green turtles from Bahía Salado between 2013 and For recaptured turtles, data correspond to the last capture. SD: standard deviation, CCL min: minimum curved carapace length, CCL n-t: curved carapace length notch to tip, CCW: curved carapace width, SCL: straight carapace length, SCW: straight carapace width, PL: plastron length, PW: plastron width, HL: head length, HW: head width, TTL: tail total length, PTL: post-cloacal tail length, BCI: body condition index. All morphological measurements are shown in centimeters (cm) and weight in kilograms (kg). Individual 1* 2** 3* 4*** Mean ± SD Tags CCL min ± 9.6 CCL n-t ± 9.8 CCW ± 9.2 SCL ± 8.6 SCW ± 6.4 PL ± 5.8 PW ± 4.8 HL ± 1.9 HW ± 1.2 TTL ± 3.6 PTL ± 1.2 Weight ± 20 BCI ± 0.28 Life stage Juvenile Juvenile Juvenile Juvenile Juvenile Juvenile Juvenile - *one recapture, **two recaptures, ***three recaptures Chile serve as green turtle developmental areas dominated by juveniles, whereas in northern Peru and likely in foraging areas closer to the equatorial realm, adults would be much more common (Velez-Zuazo et al., 2014; Veliz et al., 2014; Donoso et al., 2016). As ectotherms, green turtles are constrained by climatic, as well as, physical factors affecting the surrounding environment (Spotila et al., 1997). A previous study suggested that waters 25 C may represent the thermal threshold below which migrating adult females actively avoid surface waters in the EPO (Seminoff et al., 2008). However, thermal thresholds for juvenile turtles in the region remain poorly understood. In the northern Gulf of California (~28 N) green turtles are known to hibernate during colder months (Felger et al., 1976), while in Bahía Magdalena (~24 N) this does not seem to happen. Although sea temperatures decrease to 18ºC during winter, in this place juvenile turtles only are less active and probably forage less (Koch et al., 2007). In Bahía Salado, sea temperatures drop below 13ºC during the cold season; nevertheless, there are active turtles year-round (Brito et al., 2007). This, together with the presence of exclusively juveniles in this location, suggests that bysize latitudinal habitat segregation could also be related to ocean temperatures and thermal restraints. Feeding preferences have been linked with habitat segregation in the EPO as well (López-Mendilaharsu et al., 2005; Koch et al., 2007; Velez-Zuazo et al., 2014). Therefore, future studies taking into account these factors will allow further comprehension of the relevance of temperature in foraging ground segregation. On the other hand, the presence of so few green turtle individuals in Bahía Salado during our study, may be

82 Characterization of the southermost green turtle 5467 Table 3. Comparative values of hematology and blood biochemistry of juvenile green turtles from Bahía Salado and other locations of the EPO. SD: standard deviation, N: sample size, PCV: packed cell volume, AST: aspartate aminotransferase, LDH: lactate dehydrogenase, CK: creatine kinase. Table 4. Content of Cu and Pb in blood. Mean values (± SD) for sea turtles from the present study and other studies. due to that this location corresponds to the southern distribution limit of the EPO`s neritic aggregation of C. mydas. It is opposite to the high abundance found in the center of the distribution range for any species (Pianka, 1982). Coastal ecosystems off Peru and Chile are dominated by cold and nutrient-rich waters associated to the Humboldt Current System (Quiñones et al., 2010); however, in the region there is a high oceanographic and climatic heterogeneity due to seasonal and inter-decadal variations, such as El Niño Southern Oscillation events (ENSO; Quiñones et al., 2010). During ENSO, warmer waters (22º-28ºC) reach the coast of South America probably facilitating sea turtle approach to Peruvian and Chilean coasts (Quiñones et al., 2010; Sarmiento-Devia et al., 2015). Likewise, local-superficial circulation and the Gunter Sub-surface Current that transport warm waters southward from tropical latitudes, could be modulating the proximity of sea turtles to the coast or the time that they spend in Chilean waters (Sarmiento-Devia et al., 2015). Our results showed a high recapture rate in Bahía Salado (~27ºS), with four of the seven turtles recaptured in a period of five and sixteen months (Table 1, 2). In Peru, Velez-Zuazo et al. (2014) reported a mean recapture rate of 26% and 12.5% and a maximum recapture interval of 1015 days and 680 days (during three years of study) for El Ñuro (~4ºS) and Paracas (~14ºS), respectively. In northern Chile, a recapture rate of 3.9% with a maximum time of residence of three years was reported by Bolados-Diaz et al. (2007) in Bahía Mejillones del Sur (~23ºS) between 2003 and Moreover, Veliz et al. (2014) mentioned only one recapture with an interval of six months from a total of 18 turtles captured in Playa Chipana (~18ºS) throughout almost two years of studies. As initially

83 547 8 Latin American Journal of Aquatic Research predicted our results showed a high recapture rate and wide recapture intervals, as reported for other Peruvian and Chilean neritic aggregations. It is plausible, considering that Bahía Salado s aggregation is small, and it suggests a high residence of juveniles in this bay. On the other hand, preliminary data based on stable isotope analysis, indicated that Bahía Salado s green turtles are feeding on the endemic seagrass Zostera chilensis (Álvarez-Varas, unpubl. data). Therefore, all these results highlight the relevance of this location as a developmental foraging ground for C. mydas in the southern EPO. Green turtle as bioindicator of heavy metal pollution in Bahía Salado s ecosystem Blood has been recognized as an indicator of recent exposition to pollutants in sea turtles, unlike tissues such as skin, carapace or some internal organs, which constitute a proxy of chronical exposition (Day et al., 2007; Ikonomopoulou et al., 2011; Komoroske et al., 2011). Vast evidence indicates that the route of entry of pollutants in these species mainly occurs through food intake (Torrent et al., 2004). Thus, blood samples could provide an approach on contamination of the site where turtles feed. Concentrations of Cu and Pb found in the blood of Bahía Salado`s green turtles (2.26 ± 0.10 µg g -1 and 1.11 ± 0.06 µg g -1, respectively) corresponded to one of the highest values described for C. mydas and for almost all sea turtle species (Table 4). Green turtles at Poza Histórica de Antofagasta (~23ºS) presented similar values (2.80 ± 0.40 µg g -1 and 0.70 ± 0.40 µg g -1, respectively; Canales-Cerro & Álvarez-Varas, 2015; Table 4). Likewise, Plaza-Araya et al. (2010) reported elevated levels of As, Cu, Pb and Hg in liver and kidney of Chelonia mydas and Lepidochelys olivacea, also from Antofagasta. Elevated concentrations of heavy metals in turtles from Chilean neritic foraging grounds could be related to the intense and historic mining activity in the north of the country (Ramirez et al., 2005; Castillo & Valdés, 2011). In the same way, industries, ports, productive and touristic activities, may also contribute to this situation (Castillo & Valdés, 2011; Valdés & Castillo, 2014). Bahía Salado lacks large sources of local pollution; however, it is where the major forest of Macrocystis spp. of the Atacama Region is located, which mainly supplies the abalone industry of the region (SUBPESCA, 2010). Algae extraction is associated to a high amount of small vessels (~15 vessels day -1 during 9 h each day, Álvarez-Varas, pers. obs.) that transit daily throughout the bay. Previous studies carried out a few kilometers northward of this bay, showed elevated concentrations of Cu and Pb (among other trace-metals) in marine sediments and benthic organisms (Castillo & Valdés, 2011; Valdés & Castillo, 2014). Such results were attributed to an active atmospheric transport of heavy metals, local aquaculture activities and mineral characteristics of the area (Castillo & Valdés, 2011; Valdés & Castillo, 2014). In the particular case of Pb, Valdés & Castillo, (2014) suggested that the elevated levels found in marine sediments could be due to residues of fuel and paint used in aquaculture activities. Therefore, all these factors, together with the effect of coastal currents (Ramirez et al., 2005) may contribute to the high levels of these pollutants in our study s turtles. However, further research should incorporate other biological and environmental matrixes (e.g., heavy metals in sediments, water column, main preys of turtles, etc.) to better evaluate the extension of heavy metal inputs in the local environment. Moreover, in order to be able to attribute with certainty that the levels of Cu and Pb observed here are due to local pollution, it is necessary to understand movement patterns and residency times of turtles in the bay. Health parameters of green turtles from Bahía Salado Bahía Salado`s green turtles exhibited the highest values of BCI reported for foraging populations in the EPO (1.66 ± 0.28). Seminoff et al. (2003) reported BCI of 1.42 ± 0.02 for green turtles in Baja California Peninsula (~28ºN), Mexico; Koch et al. (2007) values of 1.35 ± 0.13 in Bahía Magdalena (~24ºN), Mexico; and Velez-Zuazo et al. (2014) of 1.50 for turtles at El Ñuro (~4ºS) and Paracas (~14ºS), Peru. Our results suggest that this location constitutes a very favorable habitat for this species probably due to that in temperate feeding grounds, the combination of younger stages, low temperatures and high prey availability may speed green turtle`s metabolism, thus they grow and gain weight relatively fast (Velez-Zuazo et al., 2014). Our results in hematology and blood biochemistry showed that Bahía Salado s green turtles exhibited several variables exceeding those reported for C. mydas from other locations of the EPO (Labrada-Martagón et al., 2010; Suarez-Yana et al., 2015; Table 3). Values of albumin, calcium, phosphorus, AST, triglycerides and creatinine reported in our study, were several magnitudes higher than those documented for juveniles by Labrada-Martagón et al. (2010) from Punta Abreojos (~26ºN), and Bahía Magdalena (~24ºN), Mexico (Table 3). Likewise, AST, total protein and creatinine from Bahía Salado s turtles were higher in comparison with values published by Suarez-Yana et al. (2015) from Sechura Bay (~5ºS), Peru (Table 3). In Baja California Sur, Labrada-Martagón et al. (2010) observed that during summer, juvenile green

84 Characterization of the southermost green turtle turtles had significantly higher concentrations of triglycerides, glucose, uric acid and total protein compared with those captured in winter; and during cold season s triglycerides and albumin decreased markedly. In addition, they suggested that elevated values of triglycerides, total protein, albumin and globulins, together with a good body condition of the turtles, may reflect a food rich environment (Labrada- Martagón et al., 2010). According to this, the elevated concentrations of triglycerides, albumins, and total protein reported here, could be due to high food availability (Labrada-Martagón et al., 2010), or to reflect the high productivity season when turtles were sampled. Our field trips were carried out during spring and summer months when sea temperatures ranged between 15-21ºC. As it is probable that these parameters decrease, or at least change during the cold months, it is fundamental to extend blood monitoring throughout the year at Bahía Salado. Sea temperatures reported by Labrada-Martagón et al. (2010) during winter in Baja California Sur were around 19ºC. This suggests that although water temperatures are low in spring and summer in Bahía Salado, green turtles were thriving and probably this environment has high food availability. On the other hand, concentrations of triglycerides, calcium and total protein may also increase significantly in postprandial turtles (Anderson et al., 2011; Phillips et al., 2015) and individuals with diets high in seagrass can exhibit elevated values for the latter (Whiting et al., 2007). Thus, elevated concentrations of all these parameters in Bahía Salado`s green turtle aggregation could be attributed to the moment turtles were captured since they are presumably feeding in the area, and/or diets based on seagrass. The good condition of Bahía Salado s green turtles is also supported by high levels of calcium and phosphorus. Labrada-Martagón et al. (2010) found that injured turtles from Punta Abreojos (~26ºN) had lower calcium, potassium and phosphorus levels compared to healthy turtles. Nevertheless, the high concentration of AST reported in the present study is striking. High values of this enzyme have been related to muscle damage and capture stress (Aguirre et al., 1995). Likewise, Labrada-Martagón et al. (2010) indicated AST is related to hepatocellular damage, and that its increase may be a physiological response to some contaminants in C. mydas. As our heavy metal analysis showed very high Cu and Pb concentrations in blood, AST values found in Bahía Salado s green turtles may be due to pollutant exposure or associated to turtle capture and handling (Aguirre et al., 1995; Suarez- Yana et al., 2015). In reptiles, unlike mammals, blood creatinine concentration is generally considered to be a poor indicator of renal function (McArthur et al., 2004; Mader, 2006). In this study, causes that could be associated to an increase of this metabolite remain unknown. However, there have been sick chelonian cases reported where creatinine values were below (and not above) the reference range for the species (McArthur et al., 2004). This could suggest that the values reported here not necessarily are associated to an abnormal or pathological condition. Connectivity of green turtles in the Eastern Pacific Ocean Foraging grounds of C. mydas commonly consist of genetically mixed stocks made up of turtles originating from different distant rookeries (Bolker et al., 2007; Bowen et al., 2007; Amorocho et al., 2012). Due to small sample size, we could not perform a mixed stock analysis (MSA) to determine statistically significant contributions from regional nesting rookeries. However, our overall picture of haplotypic characterization reflects what has been observed in other eastern and southeastern Pacific C. mydas foraging grounds and which are dominated by individuals deriving from the geographically closest nesting rookery in the Galapagos Archipelago. The very first MSA conducted in the EPO showed that juveniles foraging at Isla Gorgona (~2ºN), Colombia, were composed by more than 80% of the Galapagos stock (Amorocho et al., 2012). Similarly, a recent characterization of the C. mydas aggregation at Playa Chinchorro (~18ºS, northern Chile) recovered ca. 540 base pair long Cm-P4 (H1) as the most frequent haplotype, followed by Cm-P5 (H2), haplotype Cm- P17 and Cm-P93 (H3 and H4 respectively; Veliz et al., 2014). Considering the frequency of these shorter sequences in regional nesting stocks (Chassin-Noria et al., 2004; Dutton et al., 2008), the Galapagos rookery seems to be the principal source rookery for this South American feeding ground. Although larger fragments of 770 bp provide better resolution for natal origin estimations (Jensen et al., 2012), the observed pattern does not seem to differ. Using longer mtdna sequences, MSA on a northern Chilean foraging ground at Bahía Mejillones del Sur (~23ºS) and from incidentally caught green turtles in the Peruvian and Chilean longline fishery, both estimated a Galapagos contribution of more than 95% (Donoso et al., 2016, Dutton et al., 2016). Given that most of the individuals sampled at Bahía Salado also carried haplotypes dominant or endemic to the Galapagos Archipelago as was predicted (Dutton et al., 2014), this seems to be the major source population for South American green turtle foraging grounds. This supports previous observations on a regional structure in the distribution of northern and southeastern green turtle populations, where the Galapagos rookery contributes to foraging grounds from the South American to

85 Latin American Journal of Aquatic Research Central American coastline, and the Mexican rookery at Colola to the northeastern Pacific region (Seminoff et al., 2015). Therein, a recent MSA on the adult foraging ground at Poza del Nance in Guatemala (Chavarria, unpubl. data) marks the Central American boundary for individuals originating in the Galapagos Archipelago, with an estimated contribution of less than 20%, while at Isla del Coco, in the Pacific of Costa Rica, this rookery still contributes by more than 90% (Heidemeyer, unpubl. data). Thus, Bahía Salado represents the southernmost foraging ground known to date for green turtles nesting in the Galapagos Archipelago. CONCLUSIONS AND OUTLOOK Our study suggests that Bahía Salado constitutes a developmental habitat for C. mydas in northern Chile, supporting the hypothesis of a latitudinal pattern of bysize habitat segregation in the south of the EPO. Furthermore, this bay represents the southernmost aggregation area for the Galapagos rookery in the EPO described to date, therefore demanding further knowledge and protection, especially on migratory routes between both habitats. The high site fidelity reported here implies that Bahía Salado s green turtles could be staying for many years in this location before migrating to reproductive areas. Also, the good body condition shown in turtles from this bay indicates that the local ecosystem may be of great importance for their preparation for reproduction, highlighting the importance to conserve all of its associated biotic and abiotic components. The elevated heavy metal levels and alteration of some blood parameters in Bahía Salado s green turtles demand further research on the contamination extension and main pollution sources in the bay. Moreover, studies about sub-lethal effects (i.e., immunosuppression, higher susceptibility to diseases, lower growth rates, among others) eventually induced or caused by heavy metals in this aggregation are necessary. The above could directly impact the major rookery of green turtles in the EPO (Galapagos), thus affecting the population at a regional level. Likewise, it is fundamental to evaluate and quantify other threats that may generate negative impacts on the turtles and this ecosystem such as bycatch, seagrass meadow degradation, boat collisions, among others. Although our study showed that Bahía Salado s aggregation is small, it is probable that there is a continual arrival of individuals from adjacent foraging areas, as was observed during recent field trips, where new turtles were captured. Thus, the absence of local conservation strategies may impact the green turtle population in the region at a higher scale. Ultimately, our results point to increase monitoring and research of this ecosystem s quality in terms of pollution and food availability, as well as trophic ecology and movement patterns of turtles in this area. In the same way, it is necessary to involve local seaweed gatherers and the fishing community in local conservation strategies in order to achieve long-term protection of this extreme green turtle feeding ground and this valuable ecosystem. ACKNOWLEDGEMENTS We thank Qarapara Tortugas Marinas Chile NGO's team for their support in the field. Thanks to the Totoral community for their willingness, especially Karina Jorquera and John Michada for their help in activities conducted in Bahía Salado. We are also especially thankful to Betsy Pincheira, Carlos Olivares, Daly Noll, Ana María Ramírez, Universidad Mayor students of the Conservation, Biodiversity and Environment Internship, as well as the National Oceanic and Atmospheric Administration (NOAA), Rufford Small Grant, Idea Wild and the Chilean Government Environmental Protection Fund (FPA). Their valuable collaboration made this study possible. Finally, we thank Patricia Zárate, Aldo Pacheco and Ximena Velez-Zuazo for their comments and suggestions that helped us improve the manuscript. REFERENCES Abreu-Grobois, F.A., J.A. Horrocks, A. Formia, P.H. Dutton, R.A. LeRoux, X. Velez-Zuazo, L.S. Soares & A.B. Meylan New mtdna Dloop primers which work for a variety of marine turtle species may increase the resolution of mixed stock analysis. pp In: M. Frick, A. Panagopoulous, A.F. Rees & K. Williams (eds.). Proceedings of the 26 th Annual Symposium on Sea Turtle Biology, Island of Crete, Greece. [ turtlesymposium2006_abstracts.pdf]. Reviewed: 7 November Aguirre, A., G. Balazs, T. Spraker & T. Gross Adrenal and hematological responses to stress in juvenile green turtles (Chelonia mydas) with and without fibropapillomas. Physiol. Zool., 68(5): Alfaro Shigueto, J., J.C. Mangel, F. Bernedo, P.H. Dutton, J.A. Seminoff & B.J. Godley Small scale fisheries of Peru: a major sink for marine turtles in the Pacific. J. Appl. Ecol., 48(6):

86 Characterization of the southermost green turtle Aljanabi, S.M. & I. Martinez Universal and rapid salt-extraction of high quality genomic DNA for PCRbased techniques. Nucleic Acids Res., 25: Amorocho, D.F., F.A. Abreu-Grobois, P.H. Dutton & R.D. Reina Multiple distant origins for green sea turtles aggregating off Gorgona Island in the Colombian Eastern Pacific. PLoS ONE, 7(2): e doi: /journal.pone Anderson E.T., L.J. Minter, E.O. Clarke III, R.M. Mroch III, J.F. Beasley & C.A. Harms The effects of feeding on hematological and plasma biochemical profiles in green (Chelonia mydas) and kemp's ridley (Lepidochelys kempii) sea turtles. Vet. Med. Int. doi: /2011/ Bjorndal, K.A., Bolten, A.B. & M.Y. Chaloupka, Green turtle somatic growth model: evidence for density dependence. Ecol. Appl., 10(1): Bolados-Díaz, P., C. Guerra-Correa, C. Guerra & A. Silva Estudio poblacional de la congregación de tortuga verde (Chelonia mydas), presente en Bahía Mejillones del Sur, Antofagasta-Chile. In: C. Guerra- Correa, C.A. Fallabrino, P. Bolados-Díaz & C. Turner (eds.) VII Simposio sobre medio ambiente: estado actual y perspectivas de la investigación y conservación de las tortugas marinas en las costas del Pacífico Sur Oriental. Universidad de Antofagasta Chile. [ _Medio_Ambiente.pdf]. Reviewed: 14 November Bolker, BM, T. Okuyama, K.A. Bjorndal & A.B. Bolten Incorporating multiple mixed stocks in mixed stock analysis: many-to-many analyses. Mol. Ecol., 16(4): Bowen, B.W., W.S. Grant, Z. Hillis-Starr, D.J. Shaver & K.A. Bjorndal Mixed-stock analysis reveals the migrations of juvenile hawksbill turtles (Eretmochelys imbricata) in the Caribbean Sea. Mol. Ecol., 16: Brito, J.L., G. Domínguez, M. Marambio & P. Gysel La necesidad de proteger a las tortugas marinas de Chascos, Bahía Salado, Región de Atacama, Chile. In: C. Guerra-Correa, A. Fallabrino, P. Bolados-Díaz & C. Turner (eds.). VII Simposio sobre medio ambiente: estado actual y perspectivas de la investigación y conservación de las tortugas marinas en las costas del Pacífico Sur Oriental. Universidad de Antofagasta Chile. [ Reviewed: 14 November Bucchia, M., M. Camacho, M.R. Santos, L.D. Boada, P. Roncada, R. Mateo & L.A. Henríquez-Hernández Plasma levels of pollutants are much higher in loggerhead turtle populations from the Adriatic Sea than in those from open waters (Eastern Atlantic Ocean). Sci. Total. Environ., 523: doi: /j.scitotenv Camacho, M., L.D Boada, J. Orós, P. López, M. Zumbado, M. Almeida-González & O.P. Luzardo. 2014a. Monitoring organic and inorganic pollutants in juvenile live sea turtles: results from a study of Chelonia mydas and Eretmochelys imbricata in Cape Verde. Sci. Total Environ., 481: Camacho, M., J. Orós, L.A. Henríquez-Hernández, P.F. Valerón, L.D. Boada, A. Zaccaroni & O.P. Luzardo. 2014b. Influence of the rehabilitation of injured loggerhead turtles (Caretta caretta) on their blood levels of environmental organic pollutants and elements. Sci. Total Environ., 487: Camacho, M., J. Orós, L.D. Boada, A. Zaccaroni, M Silvi, M., C. Formigaro & O.P. Luzardo Potential adverse effects of inorganic pollutants on clinical parameters of loggerhead sea turtles (Caretta caretta): results from a nesting colony from Cape Verde, West Africa. Mar. Environ. Res., 92: Canales-Cerro, C. & R. Álvarez-Varas History, science and conservation of sea turtles in Chile. In: M.M. Lara-Uc, J.M. Rguez-Baron & R. Riosmena- Rodríguez (eds.). Successful conservation strategies for sea turtles: achievements and challenges. Nova Science Publishers, New York, pp Carneiro da Silva, C., R. Klein, I. Barcarolli & A. Bianchin Metal contamination as a possible etiology of fibropapillomatosis in juvenile female green sea turtles Chelonia mydas from the southern Atlantic Ocean. Aquat. Toxicol., 170: doi: /j.aquatox Castillo, A. & J. Valdés Metal contents in Cancer polyodon (Crustacea: Decapoda) in a bay system of northern Chile (27 S). Lat. Am. J. Aquat. Res., 39(3): doi: /vol39-issue3-fulltext-7. Chassin-Noria, O., A. Abreu-Grobois, P.H. Dutton & K. Oyama Conservation genetics of the east Pacific green turtle (Chelonia mydas) in Michoacán, Mexico. Genetica, 121: Cortés-Gómez, A.A., G. Fuentes-Mascorro & D. Romero Metals and metalloids in whole blood and tissues of olive ridley turtles (Lepidochelys olivacea) from La Escobilla Beach (Oaxaca, Mexico). Mar. Pollut. Bull., 89(1): Day, R.D., A.L. Segars, M.D. Arendt, A.M. Lee & M.M. Peden-Adams Relationship of blood mercury levels to health parameters in the loggerhead sea turtle (Caretta caretta). Environ. Health Persp., 115(10): doi: /ehp Donoso, M., P.H. Dutton & E. LaCasella Nesting population origin of a Green turtle foraging aggre-

87 Latin American Journal of Aquatic Research gation in northern Chile determined from mtdna analysis: drawing new boundaries to management units in the southeastern Pacific. 36 th Annual Symposium on Sea Turtle Biology and Conservation, 29 February-4th March 2016, Lima, Peru. Dutton P.H., E. LaCasella, Alfaro-Shiguetto, J. Paz- Campos, M. Donoso & J.C. Mangel Stock origin of leatherback, loggerhead and green turtles foraging in the southeastern Pacific: insights into their trans-oceanic connectivity. 36th Annual Symposium on Sea Turtle Biology and Conservation, 29 February- 4th March 2016, Lima, Peru. Dutton, P.H., G.H. Balazs, R.A. LeRoux, S.K. Murakawa, P. Zárate & L.S. Martinez Composition of Hawaiian green turtle foraging aggregations: mtdna evidence for a distinct regional population. Endang. Species Res., 5: doi: /esr Dutton, P.H., M.P. Jensen, A. Frey, E. LaCasella, G.H. Balazs, P. Zárate, O. Chassin-Noria, A.L. Sarti- Martinez & E. Vélez Population structure and phylogeography reveal pathways of colonization by a migratory marine reptile (Chelonia mydas) in the Central and Eastern Pacific. Ecol. Evol., 4: doi: /ece Eckert, K., K. Bjorndal, F. Abreu-Grobois & M. Donnelly (eds.) Research and management techniques for the conservation of sea turtles. IUCN/SSC Marine Turtle Specialist Group Publication 4. Felger, R.S., K. Clifton & P. Regal Winter dormancy in sea turtles: independent discovery and exploitation in the Gulf of California by two local cultures. Science, 191: Frazier, J. & S. Salas Tortugas marinas del Pacífico Oriental: el recurso que nunca acabará? Simposio de Conservación y Manejo de Fauna Silvestre Neotropical. IX Congreso Latinoamericano de Zoología, Arequipa, Perú, pp Guirlet, E., K. Das & M. Girondot Maternal transfer of trace elements in leatherback turtles (Dermochelys coriacea) of French Guiana. Aquat. Toxicol., 88(4): doi: /j.aquatox Ikonomopoulou, M.P., H. Olszowy, C. Limpus, R. Francis & J. Whittier Trace element concentrations in nesting flatback turtles (Natator depressus) from Curtis Island, Queensland, Australia. Mar. Environ. Res., 71(1): doi: /j.marenvres Jain, N. (ed.) Schalm's veterinary hematology. Lea & Febiger, Philadelphia, 1221 pp. Jensen, M.P., N.N. FitzSimmons & P.H. Dutton Molecular genetics of sea turtles. In: J. Wyneken, K.J. Lohmann & J.A. Musick (eds.). The biology of sea turtles 3. CRC Press, Boca Raton, pp Kearse M., R. Moir, A. Wilson, S. Stones-Havas, M. Cheung, S. Sturrock, S. Buxton, A. Cooper, S. Markowitz, C. Duran, T. Thierer, B. Ashton, P. Meintjes & A. Drummond Geneious basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics, 28(12): doi: /bio informatics/bts199. Kenyon, L.O., A.M. Landry Jr. & G.A. Gill Trace metal concentrations in blood of the kemp's ridley sea turtle (Lepidochelys kempii). Chelonian Conserv. Biol., 4(1): Koch, V., L.B. Brooks & W.J. Nichols Population ecology of the green/black turtle (Chelonia mydas) in Bahía Magdalena, Mexico. Mar. Biol., 153(1): doi: /s Komoroske, L.M., R.L. Lewison, J.A. Seminoff, D.D. Deheyn & P.H. Dutton Pollutants and the health of green sea turtles resident to an urbanized estuary in San Diego, CA. Chemosphere, 84(5): doi: /j.chemosphere Labrada-Martagón, V., L. Méndez-Rodríguez, S. Gardner, M. López-Castro & T. Zenteno-Savín Health indices of the green turtle (Chelonia mydas) along the Pacific coast of Baja California Sur, Mexico. I. Blood biochemistry values. Chelonian Conserv. Biol., 9(2): doi: Ley-Quiñónez, C., A.A. Zavala-Norzagaray, T.L. Espinosa-Carreon, H. Peckham, C. Marquez-Herrera, L. Campos-Villegas & A.A. Aguirre Baseline heavy metals and metalloid values in blood of loggerhead turtles (Caretta caretta) from Baja California Sur, Mexico. Mar. Pollut. Bull., 62(9): doi: /j.marpolbul Librado, P. & J. Rozas DnaSP v.5: A software for comprehensive analysis of DNA polymorphism data. Bioinformatics, 25(11): doi: / bioinformatics/btp187. López-Mendilaharsu, M., S.C. Gardner, J.A. Seminoff & R. Riosmena-Rodríguez Identifying critical foraging habitats of the green turtle (Chelonia mydas) along the Pacific coast of the Baja California Peninsula, Mexico. Aquat. Conserv. Mar. Freshw. Ecosys., 15: doi: /aqc.676. Mader, D.R Reptile medicine and surgery. Elsevier Saunders, Canada, 1242 pp. McArthur, S., R. Wilkinson & J. Meyer Medicine and surgery of tortoise and turtles. Blackwell Publishing, Oxford, 579 pp. Meylan, P.A., A.B. Meylan & J.A. Gray The ecology and migrations of sea turtles 8. Tests of the developmental habitat hypothesis. Bull. Am. Mus. Nat. Hist., 357: 1-70.

88 Characterization of the southermost green turtle Páez-Osuna, F., M.F. Calderón-Campuzano, M.F. Soto- Jiménez & J.R. Ruelas-Inzunza Lead in blood and eggs of the sea turtle, Lepidochelys olivacea, from the Eastern Pacific: concentration, isotopic composition and maternal transfer. Mar. Pollut. Bull., 60(3): doi: /j.marpolbul Phillips, B.E., S.A. Cannizzo, M.H. Godfrey, B.A. Stacy & C.A. Harms Exertional myopathy in a juvenile green sea turtle (Chelonia mydas) entangled in a large mesh gillnet. Case Rep. Vet. Med., 2015: 1-6. doi: /2015/ Pianka, E.R Ecología evolutiva. Ediciones Omega, Barcelona, 536 pp. Plaza-Araya, C., C. Guerra-Correa & B. Helena-Soto Bioacumulación de metales trazas en dos especies de tortugas marinas: Chelonia mydas y Lepidochelys olivacea que habitan las costas de Chile. In: M.C. Díaz-Granados, B. Wallace & L.H. Rodríguez (eds.). IV Simposio Regional de las Tortugas Marinas del Pacifico Sudoriental y I Simposio Colombiano de Tortugas Marinas. III Congreso Colombiano de Zoología, Asociación Colombiana de Zoología, 21 al 26 Noviembre de 2010, Medellín, 327 pp. Quiñones, J., V. González-Carman, J. Zeballos, S. Purca & H. Mianzan Effects of El Niño-driven environmental variability on black turtle migration to Peruvian foraging grounds. Hydrobiologia, 645(1): doi: /s Ramirez, M., S. Massolo, R. Frache & J.A. Correa Metal speciation and environmental impact on sandy beaches due to El Salvador copper mine, Chile. Mar. Pollut. Bull., 50(1): doi: /j.marpolbul Reich, K.J., K.A. Bjorndal & A.B. Bolten The lost years of green turtles: using stable isotopes to study cryptic lifestages. Biol. Lett., 3(6): doi: /rsbl Sarmiento-Devia, R.A., C. Harrod & A.S. Pacheco Ecology and conservation of sea turtles in Chile. Chelonian Conserv. Biol., 14(1): org/ /ccab Seminoff, J.A. (Southwest Fisheries Science Center, U.S.) Chelonia mydas. The IUCN Red List of Threatened Species 2004: e.t4615a [ dx.doi.org/ /iucn.uk.2004.rlts.t4615a en]. Reviewed: 19 November Seminoff, J.A., T.T. Jones, A. Resendiz, J.W. Nichols & M. Chaloupka Monitoring green turtles (Chelonia mydas) at a coastal foraging area in Baja California, Mexico: multiple indices describe population status. J. Mar. Biol. Assoc. UK, 83: Seminoff, J.A., P. Zárate, M. Coyne, D.G. Foley, D. Parker, B.N. Lyon & P.H. Dutton Post-nesting migrations of Galapagos green turtles Chelonia mydas in relation to oceanographic conditions: integrating satellite telemetry with remotely sensed ocean data. Endanger. Species Res., 4: doi: /esr Seminoff, J.A., C.D. Allen, G.H. Balazs, P.H. Dutton, T. Eguchi, H.L. Haas, S.A. Hargrove, M. Jensen, D.L. Klemm, A.M. Lauritsen, S.L. MacPherson, P. Pay, E.E. Possardt, S. Pultz, E. Seney, K.S. Van Houtan & R.S. Waples Status review of the green turtle (Chelonia mydas) under the U.S. Endangered Species Act. NOAA-TM-NMFS-SWFSC-539. [ nmfs.noaa.gov/pr/species/status%20reviews/green_t urtle_sr_2015.pdf]. Reviewed: 8 May Spotila J.R., M.P. O Connor & F.V. Paladino Thermal biology. In: P.L. Lutz & J.A. Musick (eds.). The biology of sea turtles. CRC Press, Boca Raton, pp Squeo, F.A., G. Arancio & J.R. Gutiérrez Libro Rojo de la flora nativa y de los sitios prioritarios para su conservación: Región de Atacama. Ediciones Universidad de La Serena, La Serena, pp Suarez-Yana, T., D. Montes, I.R. Zuñiga, J.C. Mangel & J. Alfaro-Shigueto Hematologic, morphometric, and biochemical analytes of clinically healthy green sea turtles (Chelonia mydas) in Peru. Chelonian Conserv. Biol., 15(1): doi: /CCB Subsecretaria de Pesca, Ministerio de Economía, Fomento y Turismo (SUBPESCA) Informe Técnico (R. Pesq.) N 36/10. Plan de Manejo Bahía Chasco, III Región de Atacama. pp 29. [ cl/institucional/602/articles-83762_documento.pdf]. Reviewed: 14 April Torrent, A., O. González, P. Monagas & J. Orós Trace distribution of metals in loggerhead turtles (Caretta caretta) stranded in the Canary Islands, Spain. Mar. Pollut. Bull., 49: Valdés, J. & A. Castillo Evaluación de la calidad ambiental de los sedimentos marinos en el sistema de bahías de Caldera (27 S), Chile. Lat. Am. J. Aquat. Res., 42(3): doi: /vol42-issue3-fulltext- 10. Van de Merwe, J.P., M. Hodge, H.A. Olszowy, J.M. Whittier & S.Y. Lee Using blood samples to estimate persistent organic pollutants and metals in green sea turtles (Chelonia mydas). Mar. Pollut. Bull., 60(4): doi: /j.marpolbul Veliz, D., P. Salinas, W. Sielfeld, D. Contreras, C. Azócar, M. Tobar & J. Gallardo Estudio poblacional y genético de la tortuga Chelonia mydas agassizii (Sauria: Cheloniidae) en La Playa Chinchorro, Arica,

89 Latin American Journal of Aquatic Research Chile. Rev. Biol. Mar. Oceanogr., 49(3): doi: /S Velez-Zuazo, X. & S. Kelez Multiyear analysis of sea turtle bycatch by Peruvian longline fisheries: a genetic perspective. In: J.A. Blumenthal, A. Panagopoulou & A.F. Rees (eds.). Proceedings of the 30 th Annual Symposiumon Sea Turtle Biology and Conservation, April 2010, Goa, India. Velez-Zuazo, X., J. Quiñones, A.S. Pacheco, L. Klinge, E. Paredes, S. Quispe & S. Kelez Fast growing, healthy and resident green turtles (Chelonia mydas) at two neritic sites in the central and northern coast of Peru: implications for conservation. PLoS ONE, 9(11): e pone Whiting, S., M. Guinea, C. Limpus & K. Fomiatti Blood chemistry reference values for two ecologically distinct populations of foraging green turtles, eastern Indian Ocean. Comp. Clin. Path., 16: doi: /s y. Zárate, P., K.A. Bjorndal, M. Parra, P.H. Dutton, J.A. Seminoff & A.B. Bolten Hatching and emergence success in green turtle Chelonia mydas nests in the Galapagos Islands. Aquat. Biol., 19(3): doi: /ab Zavala, P., H. Díaz & P. Araneda Determinación de la biomasa de Macrocystis integrifolia (huiro canutillo), Lessonia trabeculata (huiro palo) y Heterozostera chilensis (pasto marino), mediante técnicas de teledetección aeroespacial en Bahía Chascos, región de Atacama. Geosensing Ltda., Viña del Mar, pp. 43. [ biblioteca/589/articles-64468_documento.pdf]. Reviewed: 14 November Received: 31 May 2016; Accepted: 30 October 2016 Appendix 1. Hematological and blood biochemistry processing including location, method, storage sample and analysis. Process Location Method Storage Analysis Blood smears Field Blood films were done by wedge smear technique. These were fixated with methanol for five minutes, then air dried. Blood smear staining Packed cell volume (PCV) Obtaining plasma for blood biochemistry Blood biochemistry Laboratory Each blood smear was stained with May- Grünwald-Giemsa stain, following Schalm s Veterinary Hematology protocol (Jain 1986). Field Non-heparinized micro-hematocrit capillary was filled with heparinized blood, reaching 3/4 of the tube, in order to centrifuge it for 5 minutes at 12,000 rpm in a Hawksley micro-centrifuge (Hawksley y Sons Ltd., England). Field Heparinized blood was processed to obtain plasma, in a period of time that did not exceed one hour after blood extraction. Blood samples were centrifuged at 3,500 rpm for 15 minutes, using a field centrifuge. Plasma was transferred to cryotubes with a sterile Pasteur pipette. Laboratory Blood plasma was warmed up until it reached environmental temperature and then was immediately analyzed with a Microlab 100 (MERK ) equipment, following the methodologies described in Appendix 2. Slide transport box/ Environmental temperature Slide box/ Environmental temperature Immediately analyzed Liquid nitrogen tank / -196 C Deep freeze / -80 C Further processing and analysis was performed at the laboratory. Under an optic microscope (Carl Zeiss, Standard 20 model), with 100x immersion lens, 20 fields of each smear were observed in search of hemoparasite presence. PCV lecture was performed using a micro-hematocrit capillary scale, where the capillary was placed to obtain the PCV expressed in percentages. Further processing and analysis was performed at the laboratory. Total proteins, albumin, globulins, glucose, calcium, phosphorus, aspartate aminotransferase (AST/GOT), lactate dehydrogenase (LDH), creatine kinase (CK) triglycerides and creatinine were obtained.

90 Characterization of the southermost green turtle Appendix 2. Methods and products used in blood biochemistry analysis. AST/GOT: aspartate aminotransferase, LDH: lactate dehydrogenase, CK: creatine kinase. Parameter Method Commercial name Manufacturer Total protein Biuret endpoint Total protein BioMed Egy Chem, Egypt Albumin Endpoint colorimetric method Albumin Endpoint colorimetric method Globulins Globulins = Total protein - Albumin - - Glucose GOD-Trinder method Glucose liquiform Labtest Diagnostica S.A., Brasil Calcium Arsenazo III Colorimetric method Calcium-Arsenazo III BioMed, Egy Chem, Egypt Phosphorus UV Endpoint Phosphorus UV K068 Bioclin, Quibasa Química Básica Ltda., Brasil AST/GOT Kinetic method GOT BioMed, Egy Chem, Egypt LDH Kinetic method LDH liquiform Labtest Diagnostica S.A., Brasil CK Kinetic method CK-NAC liquiform Labtest Diagnóstica S.A., Brasil Triglycerides Colorimetric method Triglycerides L.S BioMed, Egy Chem, Egypt Creatinine Fixed rate method Creatinine BioMed, Egy Chem, Egypt

91 Lat. Am. J. Aquat. Res., 45(3): , Hawksbill 2017 turtles bycatch in Eastern Pacific lobster gillnets 521 Sea Turtle Research and Conservation in Latin America Jeffrey Mangel, Jeffrey Seminoff, Bryan Wallace & Ximena Vélez-Suazo (Guest Editors) DOI: /vol45-issue3-fulltext-3 Research Article Survival on the rocks: high bycatch in lobster gillnet fisheries threatens hawksbill turtles on rocky reefs along the Eastern Pacific coast of Central America Michael J. Liles 1,2, Alexander R. Gaos 3, Allan D. Bolaños 4, Wilfredo A. Lopez 1,5 Randall Arauz 1,4, Velkiss Gadea 1,6, José Urteaga 1,7, Ingrid L. Yañez 1, Carlos M. Pacheco 1 Jeffrey A. Seminoff 1,8 & Markus J. Peterson 2 1 Eastern Pacific Hawksbill Initiative (ICAPO), San Salvador, El Salvador 2 University of Texas at El Paso, El Paso, USA 3 San Diego State University, San Diego, USA, University of California, Davis, USA 4 Programa Restauración de Tortugas Marinas (PRETOMA), San José, Costa Rica 5 Fundación para la Conservación y Protección del Arrecife de Los Cóbanos (FUNDARRECIFE) 6 Fauna and Flora International, Managua, Nicaragua, Eastern Pacific 7 Stanford University, Stanford, CA, USA 8 National Oceanic and Atmospheric Administration, Southwest Fisheries Science Center La Jolla, CA, USA Corresponding author: Michael J. Liles (mliles@hawksbill.org) ABSTRACT. Small-scale coastal fisheries can cause detrimental impacts to non-target megafauna through bycatch. This can be particularly true when high-use areas for such species overlap with fishing grounds, as is the case with hawksbill turtle (Eretmochelys imbricata) aggregations at lobster gillnet fishing sites in El Salvador and Nicaragua. We quantified hawksbill bycatch by partnering with local fishers to record data for 690 gillnet sets on rocky reefs at Los Cóbanos Reef Marine Protected Area ( ) and Punta Amapala ( ) in El Salvador, and La Salvia ( ) in Nicaragua. Based on 31 observed hawksbill captures, the mean bycatch-per-unit-effort (0.0022; individuals per set = ) and mortality (0.74) are among the highest reported for the species across fishing gear types and oceanic regions worldwide, and we conservatively estimate that at least 227 juvenile hawksbill captures occurred in lobster gillnet fishing fleets at our sites during the study. Estimated mortality for the 227 hawksbills -which could approach the 74% observed mortality of total capturesfrom interactions with lobster gillnet fisheries at these sites during the study period may constitute the greatest single source of human-induced in-water mortality for juvenile, sub-adult, and adult hawksbills in the eastern Pacific, and is of grave concern to the population. Based on our findings, we discuss neritic habitat use by hawksbills during their lost years and offer recommendations for bycatch reduction strategies, including community-based efforts to enhance sustainable self-governance via the establishment of locally crafted conservationist norms and marine protected areas at important developmental habitat. Keywords: bycatch rate, foraging hotspot, developmental biogeography, marine conservation, communitybased governance, habitat protection. INTRODUCTION Unintentional capture of non-target species, or bycatch (Hall et al., 2000), constitutes a significant proportion of global catches in marine fisheries (Kelleher, 2005; Pauly & Zeller, 2016) and is considered a primary driver in population declines of many long-lived marine megafauna, including marine mammals, seabirds, elasmobranchs, and sea turtles (Zydelis et al., 2009; Mangel et al., 2010; Wallace et al., 2010, 2013; Dapp et al., 2013; Lewison et al., 2014; Peckham et al., 2016). For decades, research on bycatch taxa and mitigation efforts have focused on large-scale, industrial fisheries, given their high fishing effort and centralized infrastructure (Lewison et al., 2004; Soykan et al., 2008; Zydelis et al., 2009). Indeed, large-scale fisheries are recognized as a serious threat to some marine vertebrate populations that require urgent intervention. For example, pelagic longlining vessels in the North Pacific appear to be the largest known source of mortality for the black-footed albatross, killing at least 5,000-14,000 individuals annually, which represents

92 522 Latin American Journal of Aquatic Research the removal of % of the population per year (Lewison & Crowder, 2003). Similarly, pelagic longlines used by industrial fleets are a primary source of mortality for older, reproductively valuable leatherback turtles (Dermochelys coriacea) in the Pacific Ocean (Lewison et al., 2004; Donoso & Dutton, 2010), where fewer than 3,000 adult females nest in the entire oceanic basin (Benson et al., 2015). However, smallscale, artisanal fisheries also are coming under increasing scrutiny as important sources of bycatch across taxa commensurate with or surpassing levels found in large-scale fisheries (Zydelis et al., 2009; Alfaro-Shigueto et al., 2010b, 2011; Gilman et al., 2010; Peckham et al., 2016). Mangel et al. (2010), for instance, reported cetacean bycatch-per-unit-effort (BPUE) in Peruvian artisanal drift gillnet fisheries to be comparable to BPUE for the large-scale Moroccan driftnet fleet in the southwest Mediterranean. Additionally, Peckham et al. (2007) reported annual bycatch levels of ~1,000 loggerhead turtles (Caretta caretta) for two small-scale fleets in Baja California, Mexico, which is on par with bycatch levels incurred by the entire North Pacific large-scale fisheries. Small-scale, nearshore fisheries provide important livelihood support and poverty alleviation for coastal residents in low-income regions worldwide (Béné, 2006). At the same time, fishing pressure is among the most substantial of human impacts on coastal ecosystems (Stewart et al., 2010) and can result in a high frequency of interactions with marine vertebrates whose primary habitat commonly overlaps with smallscale fishing activities (Wallace et al., 2008; Lewison et al., 2014). Despite growing evidence of significant threats posed by small-scale fisheries, inherent logistical difficulties associated with monitoring and managing small-scale fleets (e.g., diffuse effort, decentralized infrastructure, political and economic marginalization, little regulation) result in few available data from which to base interventions (Chuenpagdee et al., 2006). There is an urgent need to fill these data gaps, as even relatively small fisheries with few fatal interactions could constitute a major threat to severely depleted populations of marine vertebrates. Similar to other long-lived marine megafauna, sea turtles exhibit life history characteristics -including slow maturation, low fecundity, and high adult survivorship- such that relatively low levels of bycatch mortality can hamper population viability (Heppell, 1998; Soykan et al., 2008; Wallace et al., 2008). This is particularly true for small populations of highly threatened species that are inherently susceptible to decline from mortality, such as hawksbill turtles (Eretmochelys imbricata) in the eastern Pacific Ocean. Hawksbills are critically endangered globally (Mortimer & Donnelly, 2008) and in the eastern Pacific Ocean are considered among the most endangered sea turtle populations in the world (Wallace et al., 2011). Fewer than 700 adult female hawksbills are estimated to exist in the eastern Pacific region (Mexico-Peru), with >90% of all known nesting activity occurring in El Salvador and Nicaragua (Gaos et al., 2010, 2017; Liles et al., 2015b). Contrary to their conspecifics in other oceanic regions that utilize long-distance (>2,000 km), offshore migrations (e.g., Miller et al., 1998; Van Dam et al., 2008), eastern Pacific hawksbills employ short (<300 km), nearshore (<4.2 km) migrations between nesting and foraging areas (Gaos et al., 2012a). The unique and highly neritic life history of hawksbills in the eastern Pacific may make them particularly vulnerable to nearshore in-water threats. Indeed, fisheries bycatch is considered a serious threat to eastern Pacific hawksbills (Gaos et al., 2010), including artisanal lobster gillnets on rocky reefs at Los Cóbanos Reef Marine Protected Area (Los Cóbanos) in El Salvador (Liles et al., 2011). Little is known, however, about the spatiotemporal distribution and bycatch rates at Los Cóbanos and similar rocky reef systems at Punta Amapala in El Salvador and La Salvia in Nicaragua, where previous research signaled hawksbill bycatch occurs in lobster gillnet fisheries (Liles & Thomas, 2010; Gaos et al., 2014b). Further, eastern Pacific hawksbills are considered a high risk Regional Management Unit (RMU; Wallace et al., 2011) with high vulnerability to bycatch mortality and, therefore, a top priority for bycatch monitoring and mitigation, particularly in small-scale fisheries operating in or near critical hawksbill habitats (Wallace et al., 2010, 2013), such as at rocky reefs where eastern Pacific hawksbills congregate (Gaos et al., 2012a; Carrión-Cortez et al., 2013; Tóbon-López & Amorocho-Llanos, 2014). For these reasons, we investigated the distribution and magnitude of hawksbill bycatch in artisanal lobster gillnet fisheries on rocky reefs along the Pacific coast of Central America to provide a better understanding of threats confronting the species in the region. Specifically, we quantified hawksbill bycatch and mortality, and examined spatiotemporal, environmental, and gillnet fishing predictor variables to help explain the incidental capture of hawksbills in lobster gillnet fisheries at Los Cóbanos and Punta Amapala in El Salvador, and La Salvia in Nicaragua. Based on our findings, we discuss the potential population-level impacts of lobster gillnet bycatch on eastern Pacific hawksbills and offer insight into the developmental biogeography of this species in the region. We conclude by considering potential community-based conservation strategies to mitigate hawksbill bycatch in lobster gillnet fisheries along the Pacific coast of Central America.

93 Hawksbill turtles bycatch in Eastern Pacific lobster gillnets 523 MATERIALS AND METHODS Study site description Previous studies identified lobster gillnet fishing as potentially interacting with hawksbills at Los Cóbanos and Punta Amapala in El Salvador, Isla Amapala in Honduras, and La Salvia in Nicaragua (details for how and where these interview data were collected are available at Liles & Thomas, 2010 and Gaos et al., 2014b). We included all of these sites in the present study, except for Isla Amapala, which we excluded because of logistical constraints. Los Cóbanos (13 31 N, W), Punta Amapala (13 08 N, W), and La Salvia (13 00 N, W) (Fig. 1) are comprised primarily of submerged volcanic reef formations at depths ranging from 0 to 30 m and host diverse marine communities, including corals, sponges, and fishes (Orellana-Amador, 1985; Domínguez-Miranda, 2010). These rocky reef systems are under increasing pressure from local fishers, sedimentation loads transported by surrounding rivers, and wastewater discharge from nearby communities (Reyes-Bonilla & Barraza, 2003). Los Cóbanos and Punta Amapala also constitute some of the most important open-coast nesting beaches for hawksbills in the eastern Pacific (Liles et al., 2011; Gaos et al., 2017), and all three sites are situated in the migration corridor of post-nesting hawksbills (Gaos et al., 2012a), increasing the potential for fisheries interactions. Punta Amapala and La Salvia flank the western and eastern sides of the Gulf of Fonseca, respectively (Cortés, 2007), which also is critical hawksbill foraging habitat (Gaos et al., 2012b). Fisheries observations Onboard observations of lobster gillnet fisheries occurred from 8 November 2008 to 29 September 2009 at Los Cóbanos, 21 November 2012 to 29 April 2014 at Punta Amapala, and 11 September 2012 to 12 May 2014 at La Salvia. We selected lobster gillnet fishers (1, 4, and 4 crews at Los Cóbanos, Punta Amapala, and La Salvia, respectively) who were willing to participate in the study and trained them as observers to collect data on each gillnet they set during fishing trips. These data included net characteristics (i.e., length, height, mesh size, and line test), georeferenced location using a handheld GPS unit, substrate type, depth, weight of lobster catch, and number of hawksbill bycatch events. While we have bycatch data for other species (e.g., olive ridley turtles (Lepidochelys olivacea)), this study was focused on describing bycatch of hawksbills at our study sites. For each hawksbill capture, observers measured curved carapace length (CCL) from nuchal notch to posterior-most tip of marginal scutes, and spatially referenced the location of capture using a GPS unit. Observers tagged live turtles with inconel tags (Model 681, National Band & Tag, Newport, KY, USA) on the second proximal scale at the edge of both front flippers, photographed them for confirmation, and subsequently released the individuals. Dead turtles were measured, photographed, and buried. We used previously established diagnostic approaches to determine whether a turtle was bycaught in fisheries (e.g., Caillouet Jr. et al., 1996; Koch et al., 2006). Specifically, we inferred the source of mortality for hawksbills found adrift or stranded to be lobster gillnets when the turtle exhibited at least two of the following characteristics: 1) carcass presented evidence of gillnet interactions, such as marks and/or lesions on extremities, 2) carcass encountered adjacent to (i.e., 2.0 km considering prevailing northward surface current) known lobster gillnet fishing sites, 3) size class of carcass corresponded to those that have been directly observed as bycatch in lobster gillnets at the study sites and 4) month carcass was found coincided with lobster gillnet fishing season (September-April) at the study sites. As compensation for collaboration, we covered fuel costs for fishing activities of the observers at Los Cóbanos and provided fishing gear (e.g., life jackets, rain gear, handheld GPS units) to observers at Punta Amapala and La Salvia. Local fishers typically worked in 2-3 person crews from 6-8 m skiffs powered with 40 horsepower outboard motors and used bottom-set gillnets to target primarily Pacific green spiny lobster (Panulirus gracilis), which have high commercial value on local markets, but also retained incidentally captured species (families Carangidae, Haemulidae, Kyphosidae, Portunidae, Scombridae, and Serranidae) of lesser commercial value for sale or subsistence. Gillnets measured 600-4,000 m in length by 1 m in height with cm mesh monofilament mesh and kg line test, and were soaked 1-31 h at depths of 1-43 m (Table 1). All gillnets were anchored to the bottom using rocks, cement-filled plastic bottles, or anchors at both ends, with lead weights attached along the foot-rope. At Los Cóbanos and Punta Amapala, local fishers used gillnets with small floats along the head-rope that produced an inverse pendulum effect, where the head-rope suspended <1 m above the foot-rope anchored to the substrate and swung back and forth with the current. At La Salvia, gillnets were fished without floats along the head-rope, where the net laid across the bottom and entangled in marine invertebrates (e.g., sponges and corals) and substrate. At Los Cóbanos and La Salvia, local fishers typically set nets in the afternoon and retrieved them the following morning, whereas at Punta Amapala, nets were set in the morning and retrieved the following morning.

94 524 Latin American Journal of Aquatic Research Figure 1. Locations of observed lobster gillnet sets (white circles), hawksbill captures (triangles), and hawksbill strandings (crosses and black lines) at a) Los Cóbanos Reef Marine Protected Area, El Salvador, (n = 133 sets, 3 captures, and 8 strandings), b) Punta Amapala, El Salvador, (n = 300 sets, 14 captures, and 7 strandings), and c) La Salvia, Nicaragua, (n = 257 sets and 14 captures). Black circles represent lobster fishing landing sites and black box delimits potential community-based marine protected area at Poza de la Gata, Punta Amapala. Given the dynamic nature of many small-scale fishing fleets (Béné, 2006) and the complications associated with managing fishers that use informal landing sites in small communities (Salas et al., 2007), accurate country-wide estimates for the entire artisanal lobster gillnet fishing fleets in El Salvador and Nicaragua are unavailable. However, the lobster gillnet fishing fleet numbered approximately 14 boats (28 to 42 fishers) at Los Cóbanos in , with the communities of Los Cóbanos, Barra Ciega, and Barra Salada contributing five, two, and seven boats, respectively (W. Lopez, pers. obs.). Based on surveys conducted with lobster fishers at Punta Amapala and La Salvia in 2015 (C. Pacheco, pers. obs.), the fleet at Punta Amapala consisted of an estimated 53 boats (103 fishers) in , 50 of which departed from the community of Playas Negras and three from El Maculís, and at La Salvia in , nine boats (20 fishers) operated out of the community of La Salvia. Fishing effort and frequency of fishing trips at the three sites depended on lobster catches. During the primary lobster fishing season (September-April), boats set gillnets 4-7 days per week during periods of high lobster catches and 2-3 days per week during periods of low catches (CENDEPESCA, 2012). During the offseason, boats generally fished 0-2 days per week, but would increase frequency if lobster catches were high. Data analyses We calculated hawksbill BPUE for each net as the number of hawksbill captures per gillnet fishing effort ([net length 100 m -1 ] [net soak time 12 h -1 ]) (Wang et al., 2013) and hawksbill mortality as the proportion of hawksbills killed. Additionally, to facilitate comparisons with other published data, we reported hawksbill bycatch per gillnet set (Wallace et al., 2010). However, we used the BPUE calculation that incorporated net length and soaktime for higher resolution in fishing effort for all estimates. To estimate the minimum annual hawksbill bycatch across the lobster gillnet fishing fleet at each site, we made two calculations and then reported the lower value: 1) multiplied the mean hawksbill

95 Hawksbill turtles bycatch in Eastern Pacific lobster gillnets 525 BPUE for each net per boat per day by the estimated fishing effort (minimum observed gillnet fishing effort among boats that fished per day) for the estimated minimum number of boats fishing during a given year, and 2) multiplied the average number of observed hawksbill bycatch events per boat by the estimated minimum number of boats fishing during a given year (see Fisheries observations subsection for the estimated number of boats per study site; Table 3). We formulated competing nominal logistic regression models using each of 13 spatio-temporal, environmental, and gillnet fishing variables (Table 1) singly, in pairs, collectively, and with interactions to explain hawksbill bycatch in observed lobster gillnet sets. Models were then compared within an informationtheoretic approach to model selection using Akaike s information criterion corrected for small sample size (AIC c), and calculating the associated Akaike weights (w i) for each model (Burnham & Anderson, 2002). We evaluated model performance by calculating the areaunder-the-curve (AUC) of the receiver operating characteristics (ROC) plot (Field & Bell, 1997), and considered AUC values of , , , and >0.9 as poor, acceptable, excellent, and outstanding agreement between predictions and observations, respectively (Swets, 1988; Hosmer & Lemeshow, 2000). We computed all statistical analyses and spatial analyses using JMP Pro 12.2 (SAS Institute, Cary, NC, USA) and ArcMap 10.4 (Esri, Redlands, CA, USA), respectively. For descriptive statistics, values are presented as mean ± standard deviation (SD). RESULTS Observers monitored 690 gillnet sets (total observed fishing effort = 20,018.5 effort units (see Data analyses subsection for formula)) in lobster gillnet fisheries at Los Cóbanos (n = 133 sets), Punta Amapala (n = 300 sets) and La Salvia (n = 257 sets) (Fig. 1). Most observed fishing effort occurred from September to April (86.5% of total, n = 17,312.5 effort units), which coincides with the lobster fishing season at the three sites, and nearly 50% (n = 319) of total observed gillnet sets occurred in 2013, which was the only full year monitored. The mean net length and soaktime per observed set across sites was 1,926.0 ± 1,014.5 m and 17.5 ± 5.3 h, respectively, with the longest nets deployed at La Salvia (mean = 2,330.1 ± m) and longest soaktimes at Punta Amapala (21.3 ± 2.1 h, Table 2). Gillnet sets occurred primarily on rocky substrate (91.4%, n = 631 sets) and the mean depth of observed gillnets across sites was 10.6 ± 6.0 m (n = 686 sets), with nets set shallowest (2.7 ± 0.9 m, range = , n = 133 sets) and deepest (14.8 ± 5.2 m, range = , n = 299 sets) at Los Cóbanos and Punta Amapala, respectively. The mean distance of nets from shore was likewise closest (0.52 ± 0.3 km, range = , n = 133 sets) at Los Cóbanos and furthest (1.68 ± 0.5 km, range = , n = 300 sets) at Punta Amapala, with an average distance of 1.33 ± 0.8 km (n = 690 sets) at the three sites. Hawksbill capture in observed gillnet sets was a relatively rare event -662 sets (96.0% of total) captured no hawksbills, 25 sets (3.6%) captured one hawksbill, and three sets captured two hawksbills (0.4%). Combined, the 690 observed gillnet sets captured 31 hawksbills (BPUE = ), of which 90.3% (n = 28) occurred at Punta Amapala (n = 14 hawksbills, BPUE = ) and La Salvia (n = 14 hawksbills, BPUE = ; Table 2). Despite fewer hawksbill captures (n = 3), observed gillnet sets at Los Cóbanos registered the highest BPUE (0.0032) of the three sites. Hawksbill capture per gillnet set was across sites, with values of , , and registered at Los Cóbanos, Punta Amapala, and La Salvia, respectively. Overall, hawksbill bycatch events were highest in December-January (n = 18) and September-October (n = 9), representing 69.2% of total hawksbill captures, and lowest in May and July-August with zero hawksbill captures, which reflected fishing effort and lobster catches (Fig. 2). The mean depth and distance from shore of observed hawksbill captures across sites was 9.7 ± 5.3 m and 0.84 ± 0.6 km, respectively, with bycatch events shallowest (2.2 ± 0.8 m, range = , n = 3) and closest to shore (0.39 ± 0.1 km, range = ) at Los Cóbanos, and deepest (12.4 ± 5.7 m, range = , n = 14) and furthest from shore (0.98 ± 0.4 km, range = ) at Punta Amapala. Hawksbill captures at Punta Amapala and La Salvia followed a pattern similar to the frequency of gillnet sets and lobster catches with respect to depth, except at shallow depths of 1-5 m, where the proportion of hawksbill captures was considerably higher than both net sets and lobster catches (Fig. 3). Of the 31 hawksbill captures, 74.2% (n = 23) were landed dead, with 100% mortality (n = 14) at Punta Amapala (Table 2). An additional 16 hawksbill carcasses were opportunistically found adrift or stranded on adjacent beaches during the study period, with eight, seven, and one dead hawksbills encountered at Los Cóbanos, Punta Amapala, and La Salvia, respectively. Pooling observed hawksbill bycatch and stranding events across sites, 44.7% (n = 21) of total hawksbill gillnet interactions and 67.7% (n = 21) of total observed mortality occurred at Punta Amapala. Measured hawksbills retrieved from gillnets and found stranded averaged 35.1 ± 12.1 cm CCL (range = 14-54, n = 34) at the three sites, with smaller (31.6 ± 14.7 cm

96 526 Latin American Journal of Aquatic Research Table 1. Information on 13 variables included in predictive models for hawksbill capture in observed lobster gillnet sets at Los Cóbanos Reef Marine Protected Area (LC, ) and Punta Amapala (PA, ), El Salvador, and La Salvia (LS, ), Nicaragua (n = 678 sets). Variable type Variable name Description Mean + SD or percentage of sets Range Spatio-temporal Site Geographic area where net was set 20% (LC), 43% (PA), 37% (LS) - Year Year in which net was set 4% (2008), 16% (2009), 23% (2012), 46% (2013), 11% (2014) , Day Day number on which net was set Distance Distance (m) of net from the shore 1.3 ± Environmental Depth Depth (m) at which net was set 10.6 ± Substrate Type of substrate where net was set 92% (rock), 5% (mud), 3% (sand) - Tide Water height (m) when net was set relative to 0.5 ± historic annual mean sea level Gillnet fishery Length Horizontal length (m) of net 1924 ± SoakTime Number of hours the net was soaked in water 17.5 ± LineTest Strength (kg) of monofilament fishing line 1% (1.4), 30% (1.8), 45% (2.3), 3% - used for net (2.7), 20% (4.5), 1% (13.6) MeshSize Size (cm) of mesh opening of net 30% (7.6), 1% (8.9), 67% (10.2), 2% (12.7) - NetPanel No. of parallel net panel layers used for 93% (single), 7% (double) - gillnet LobstersCaptured Weight (kg) of lobsters captured per set 7.8 ± CCL, n = 20) and larger (47.0 ± 5.6 cm CCL, n = 3) turtles at Punta Amapala and Los Cóbanos, respectively (Table 2). We estimate that a minimum of 227 hawksbills were captured in the lobster gillnet fisheries at Los Cóbanos ( ), Punta Amapala ( ), and La Salvia ( ). The majority of these estimated bycatch events (69.7%, n = 158 turtles) occurred at Punta Amapala, with >100 hawksbills captures occurring in 2013 (Table 3). The best-approximating model for predicting hawksbill capture in observed lobster gillnet sets included four of the 13 predictor variables (w i = 0.126), which had <2 times the empirical support as the secondthrough fourth-ranked models (Table 4). The top 23 models constituted the 90% confidence set and were plausible, with a cumulative ΔAIC c <7 (Burnham et al., 2011). Diverse combinations of the same variables in the models demonstrated their similarity in how they affect the system. Depth and year were the most strongly supported predictor variables and appeared in 23 and 21 of the 23 best-approximating models, respectively, followed by soaktime (10 models) and distance from shore (9 models). AUC values for 22 of the top 23 models ranged from 0.70 to 0.76, indicating acceptable model performance, with only one model (0.66) considered poor. Low hawksbill abundance and relative rarity of bycatch events at our study sites often resulted in the disparity in hawksbill bycatch rates in similar spatiotemporal and biophysical conditions. Nevertheless, many of the top models had sufficient empirical support and acceptable AUC values, indicating that the variables included in the models are indeed important. DISCUSSION There is growing concern that small-scale fisheries are driving declines in some sea turtle populations around the world (Soykan et al., 2008; Alfaro-Shigueto et al., 2010b; Moore et al., 2010). However, the paucity of bycatch data from small-scale fisheries globally continues to impede conservation interventions, particularly for hawksbills in the eastern Pacific (Wallace et al., 2010). Our study provides the first assessment of the spatiotemporal distribution patterns and magnitude of hawksbill bycatch and mortality in artisanal lobster gillnet fisheries in El Salvador and Nicaragua. Based on 31 observed hawksbill captures, of which 23 were lethal, we conservatively estimate that lobster gillnet fishing fleets captured at least 227 hawksbills on rocky reefs at Los Cóbanos ( )

97 Table 2. Annual observed fishing effort and hawksbill bycatch in lobster gillnet fisheries at Los Cóbanos Reef Marine Protected Area (LC, ) and Punta Amapala (PA, ), El Salvador, and La Salvia (LS, ), Nicaragua. BPUE, mean bycatch-per-unit-effort calculated for each net as the number of hawksbills captured per gillnet fishing effort ([net length 100 m -1 ] [net soak time 12 h -1 ]); Mort, mortality calculated for each year as the number of hawksbills killed per total number of hawksbills captured; Strand, dead hawksbill found adrift or on beach likely killed by lobster gillnets; CCL, curved carapace length. *Previously published data (Liles et al., 2011) on juvenile hawksbills measuring cm CCL incidentally captured in lobster gillnets not included in size values. Hawksbill turtles bycatch in Eastern Pacific lobster gillnets 527 and Punta Amapala ( ) in El Salvador, and La Salvia ( ) in Nicaragua. Estimated mortality for the 227 hawksbills -which could approach the 74% observed mortality of total captures- from interactions with lobster gillnet fisheries at these sites during the study period may constitute the greatest single source of human-induced in-water mortality for juvenile, subadult, and adult hawksbills in the eastern Pacific (Wallace et al., 2010, 2013). Although our observer data only encompassed a relatively small proportion of overall estimated lobster fishing fleets (7% at Los Cóbanos, 8% at Punta Amapala, and 44% at La Salvia), we expect our calculations underestimate actual hawksbill bycatch at the three sites because 1) we used minimum values for all factors except hawksbill BPUE, for which we used observed BPUE, 2) we used minimum values for calculations from fisheries observations covering only a portion of each year, except for 2013, 3) the number of boats we used to calculate annual hawksbill bycatch was a conservative estimate, as there were likely additional boats operating at lobster fishing sites (e.g., 32 boats at Acajutla in Los Cóbanos, CENDEPESCA, 2012) that we were unable to confirm and thus were omitted from our calculation, and 4) boats with observers may deviate from conventional fishing practices to avoid turtle captures or may conceal interactions with hawksbills (Gilman et al., 2005, 2007). Assigning cause of mortality for marine animals, including sea turtles, found adrift or stranded on a beach is challenging given the diverse array of human and natural threats operating in dynamic aquatic environments (Hart et al., 2006). To minimize these inherent uncertainties, we incorporated multiple indicators to infer the cause of mortality for hawksbills at our study sites (see Fisheries observations subsection) and our observations suggest that hawksbill interaction with lobster gillnets is the most likely cause for hawksbills found dead-stranded in this study. Whereas all 16 turtles were found adjacent to ( 1.78 km to the north) lobster gillnet fishing grounds (Fig. 1) during the months of peak lobster fishing effort (September-January; Fig. 2) and pertained to smaller size classes that corresponded with those directly observed as bycatch in the lobster gillnet fishery (Table 2), only a small subset of these individuals had clear external lesions or net scars. The general absence of lesions on bycaught and stranded turtles in our study is consistent with other studies with more robust datasets (Peckham et al., 2008; Alfaro-Shigueto et al., 2011; Snape et al., 2013; Senko et al., 2014). Because hawksbills generally partition vertical habitat by size, where juvenile turtles tend to forage at shallower water habitats than adults (Blumenthal et al., 2009), it is not

98 528 Latin American Journal of Aquatic Research Table 3. Minimum annual fishing effort and hawksbill bycatch in lobster gillnet fisheries at Los Cóbanos Reef Marine Protected Area (LC, ) and Punta Amapala (PA, ), El Salvador, and La Salvia (LS, ), Nicaragua. Min. effort: minimum gillnet fishing effort, Bycatch: number of hawksbills captured in lobster gillnet fisheries. Country Year Observed Estimated total Site Boats Min. effort Bycatch Boats Min. effort Bycatch El Salvador LC Total , PA , , Total , Nicaragua LS Total , Overall 9 17, , Figure 2. Observed lobster gillnet fishing effort (n = 20,018.5 effort units), hawksbill captures (n = 31), and lobster catches (n = 5,322.5 kg) by month at Los Cóbanos Reef Marine Protected Area ( ) and Punta Amapala ( ), El Salvador, and La Salvia ( ), Nicaragua. Hawksbill captures and lobster catches are expressed as proportion of total. surprising that lobster gillnet fisheries operating in shallow waters at our study sites captured turtles of smaller size classes when fishing activities and hawksbill foraging habitat overlapped. Further, eastern Pacific hawksbills are highly neritic and their movements typically occur <2 km from shore (Gaos et al., 2012a), making interactions with artisanal longlines and gillnets targeting pelagic fish species highly unlikely (e.g., De Paz & Sui, 2008). This is corroborated by informal interviews conducted with local fishers regarding hawksbill capture in these fisheries (A. Gaos, pers. obs.). Similarly, bycatch of juvenile hawksbills in industrial shrimp trawlers is improbable because the extension of rocky reefs at our study sites prevents access to nearshore waters where hawksbills congregate (Arauz, 1996). Characterizing bycatch patterns for sea turtles across large geographic regions is necessary to understand potential cumulative impacts to species and is essential to guide conservation actions that may ameliorate threats (Lewison et al., 2014). Following the grouping of bycatch data into three broad fishing gear categories (i.e., nets, longlines, and trawls) used by Wallace et al. (2010), which allows for comparisons of bycatch in uncommon or underrepresented fisheries within and among regions, the mean BPUE (individuals per set; ) and mortality (0.74) we report for hawksbills in El Salvador and Nicaragua are among the highest reported for the species across fishing gears and oceanic regions worldwide, including net (e.g., western Atlantic, median BPUE = 0.008, median mortality = 0.50; eastern Pacific, 0.005, 0.39), longline (e.g., eastern Pacific, 0.118, 0.00; southwestern Indian, 0.034, 0.16), and trawl (northwestern Indian, 0.024, not determined; eastern Pacific, 0.011, 0.63) fisheries (Wallace et al., 2013). More recent studies on hawksbill bycatch support this assertion. Huang (2015) reported that two Taiwanese pelagic longline fishing fleets targeting tuna set 24,346,000 hooks and incidentally captured 156 hawksbills ( turtles per 1,000 hooks, 33.3% mortality) in the Pacific from 2008 to Based on commercial logbook records, Riskas et al. (2016) reported six hawksbill captures in pelagic gillnet fisheries near key foraging and nesting grounds off the coast of Australia in the western Pacific between 2000 and 2013, which comprised 3.1% of all reported

99 Hawksbill turtles bycatch in Eastern Pacific lobster gillnets 529 Figure 3. Observed lobster gillnet sets, hawksbill captures, and lobster catches by depth at a) Los Cóbanos Reef Marine Protected Area ( ), El Salvador (n = 133 sets, 3 hawksbills, and kg of lobster), b) Punta Amapala ( ), El Salvador (n = 299 sets, 14 hawksbills, and 2,244.2 kg of lobster), c) La Salvia ( ), Nicaragua (n = 256 sets, 14 hawksbills, and 2,822.7 kg of lobster). Hawksbill captures and lobster catches are expressed as proportion of total. sea turtle interactions. In the southeastern Pacific, Ortiz et al. (2016) found a relatively high bycatch rate of hawksbills per set (n = 6 turtles) in 228 bottomset gillnets deployed from 11 boats using typical fishing practices as part of regular fishing trips during net illumination trials on the coast of Peru in Given existing gaps in hawksbill bycatch data in some fisheries and oceanic regions, there are likely unreported or unstudied high-bycatch fisheries that may represent a serious threat to this species in other geographic areas. Regardless, known bycatch rates and mortality for hawksbills in the eastern Pacific - including our data on lobster gillnet fisheries- currently account for the highest or second highest rates across all three fishing gear categories in 13 oceanic regions (Wallace et al., 2013), underscoring the intense cumulative threat confronting the severely depleted population of hawksbills in the region. To mitigate sea turtle captures in fisheries effectively, conservationists need to understand spatiotemporal factors contributing to bycatch rates and mortality for species in diverse geographic areas. We identified several spatial patterns in the distribution of hawksbill captures and observed gillnet sets across the three sites. Overall, bycatch events tended to occur in gillnets deployed an average of 1 m shallower (9.7 m, range = ) and nearly twice as close to shore (0.84 km, range = ) than total gillnet sets (depth, 10.6 m, range = ; distance, 1.33 km, range = ; Fig. 1, 3-4). Model-selection supported this finding, with the predictor variables depth and distance from shore represented in 23 and 9 of the 23 bestapproximating models, respectively (Table 4). Shallow depths of hawksbill captures on rocky reefs at our study sites dovetail with published data on the foraging ecology and dive behavior of the species. Hawksbills that forage at coral reefs primarily consume sponges and macroalgae, which are typically found at highest densities in shallow water (Meylan, 1988; Leon & Bjorndal, 2002; Bell, 2013), and rarely utilize habitats deeper than 20 m (Houghton et al., 2003; Witt et al., 2010). Similar to their conspecifics in other ocean regions, juvenile (Carrión-Cortez et al., 2013) and adult (Gaos et al., 2012c) hawksbills in the eastern Pacific predominately use shallow water habitat <20 m deep, where 92% (n = 18,389.4 effort units and 654 sets) of total fishing effort and 97% (n = 30) of total hawksbill captures occurred in our study, further highlighting the threat of lobster gillnet fishing activities operating in nearshore waters of El Salvador and Nicaragua. Temporal patterns in hawksbill bycatch at our study sites varied across months and years. Hawksbill capture typically reflected the level of fishing effort, with higher bycatch occurring during the lobster-fishing season (September-April) and fewer bycatch events during the off-season (May-August) (Fig. 2). This is consistent with seasonal patterns of sea turtle bycatch corresponding to fishing effort reported elsewhere (Peckham et al., 2007; Gardner et al., 2008; Kot et al., 2009; Alfaro-Shigueto et al., 2011). Interannual variability in hawksbill bycatch often differed within and across sites (Table 2). For example, 43% (n = 6) of total hawksbill bycatch events occurred at La Salvia during a three-week period (21 December January 2014), when zero hawksbill captures occurred at the same site during the same period in High

100 530 Latin American Journal of Aquatic Research Table 4. Comparison of the 23 best-supported predictive models a for hawksbill capture in observed lobster gillnet sets at Los Cóbanos Reef Marine Protected Area (LC, ) and Punta Amapala (PA, ) in El Salvador, and La Salvia (LS, ) in Nicaragua (n = 678 sets). Rank a Predictor variables b K c AIC c d Δ i e w i f w 1/w i g AUC h 1 Depth + Year + LobsCapt + SoakTime Depth + Year Depth + Year + LobsCapt + Length Depth + Year + SoakTime Depth + Year + Day Depth + Year + LobsCapt + SoakTime + Dist Depth + Year + LobsCapt + SoakTime + Length Depth + Year + Length Depth + Year + Dist Depth + Year + LobsCapt + Length + Dist Depth + Year + SoakTime + Dist Depth + Year + LobsCapt + Site Depth + Year + SoakTime + Length Depth + Year + Day + Dist Depth + Year + LobsCapt + SoakTime + Length + Dist Depth + Year + Length + Dist Depth + Year + Site Depth + Year + LobsCapt + Dist + Site Depth + Year + SoakTime + Length + Dist Depth + Site Depth + Year + Day + Site Depth + Year + SoakTime + Site Depth + SoakTime + Length a Model ranked 1 is the best-approximating model according to AIC c among those considered. Models ranked 1-23 constitute the 90% confidence set based on summed w i. b Intercept is included in all models. Predictor variable definitions: Day: day number on which net was set, Depth: depth at which net was set, Dist: distance of the net from the shore, Length: horizontal length of net; LobsCapt: weight of lobster catch per net, Site: geographic area where net was set, LC and PA in El Salvador, and LS in Nicaragua, SoakTime: number of hours the net was soaked in water, Year: year in which net was set ( ; ). c Number of parameters in model. d Akaike s information criterion corrected for small sample size. e Delta AIC c, difference in AIC c value from best-approximating model. f Akaike weight, probability that current model is best-approximating model among those considered. g Evidence ratio, relative likelihood of each model in relation to best-approximating model. h AUC, area-under-the-curve of the receiver operating characteristic plot for response variable. interannual variability among sites was exemplified in 2014, where 7% (n = 1) and 43% (n = 6) of total hawksbill captures, constituting 30.3% (n = 3,281.3 effort units) and 10.8% (n = effort units) of total fishing effort, occurred at Punta Amapala and La Salvia, respectively. The observed variability in bycatch patterns may be explained by underlying climatic and oceanographic conditions (e.g., El Niño Southern Oscillation) that structure productivity, prey, and predator distributions, and associated fishing effort (Kot et al., 2009). Given the potential variability in hawksbill bycatch among years, it is not surprising that all plausible models included year as an important predictor of hawksbill capture in lobster gillnet fisheries at our study sites (Table 4). Effective assessments of population-level effects of fisheries bycatch, including bycatch rates, mortality, reproductive values of affected individuals (i.e., relative importance of individual to current and future reproduction), and population characteristics (e.g., population abundance and trends, geographic distribution, life history traits of species), are imperative for identifying conservation priorities and recovery strategies for depleted sea turtle populations (Crouse et al., 1987; Wallace et al., 2008; Lewison et al., 2013). Even without accounting for potential sublethal effects of hawksbill captures, such as fitness and reproduction impairment (Wilson et al., 2014), or post-release mortality (Snoddy & Southwood-Williard, 2010), which could further exacerbate impacts to the population,

101 Hawksbill turtles bycatch in Eastern Pacific lobster gillnets 531 Figure 4. Mean distances from shore and mean sizes of hawksbills (n = 27) captured in observed lobster gillnet sets at Los Cóbanos Reef Marine Protected Area ( ) and Punta Amapala ( ), El Salvador, and La Salvia ( ), Nicaragua. Error bars show standard deviation of the mean and dashed line represents mean curved carapace length of nesting hawksbills at Los Cóbanos and Punta Amapala (Gaos et al., 2017). Curved carapace lengths listed in 1 cm increments. the high observed bycatch rates and morality we report indicate that by themselves, lobster gillnet fisheries are a major threat to the persistence of hawksbills in the eastern Pacific. Our results underscore the threats that lobster gillnets can pose to non-target species, which is particularly troubling given their prevalent use in smallscale fisheries throughout Latin America. In the western Atlantic, artisanal fishers include nonselective gillnets in the wide array of main gear used to target lobsters in coastal waters, including traps, aggregation devices, and diving (Ehrhardt, 2005; FAO, 2006; Salas et al., 2011; Giraldes et al., 2015). Indeed, Aucoin & León (2007) reported the capture of three juvenile hawksbills in a 640 m lobster gillnet used in experimental bycatch trials during 24-h periods over four days in the Dominican Republic, and encountered a lobster gillnet with an additional seven dead juvenile hawksbills. In the eastern Pacific, small-scale fishers primarily use gillnets to capture lobsters (FAO, 2007; Pérez-González, 2011; Salas et al., 2011; Carrión- Cortez et al., 2013), although in Mexico, Colombia, and Ecuador lobster fishers also employ traps and diving (Beltrán, 2005; Salas et al., 2011). We expect hawksbill bycatch to be greatest in lobster gillnet fisheries along the coast of Central America, where intense use of lobster gillnets occurs in areas relatively adjacent to major nesting grounds for this species in the region (Gaos et al., 2017), which employs short migrations between nesting and foraging areas (Gaos et al., 2012a). Size distributions of individuals captured at our study sites, corresponding to juveniles (Table 2, Fig. 4), demonstrate that rocky reefs in El Salvador and Nicaragua provide important developmental habitat for hawksbills in the region. Because sea turtle populations are sensitive to relatively small changes in survival rates of large juveniles (Crouse et al., 1987; Heppell et al., 2005; Wallace et al., 2008), the high frequency of interactions and mortality of juvenile hawksbills in lobster gillnet fisheries is particularly alarming given the extremely small size of the eastern Pacific population, and could undermine conservation gains from the protection of nesting sites (Gaos et al., 2017). We recognize that bycatch mortality of reproductively mature individuals constitutes a greater threat to sea turtle population viability than the loss of turtles from smaller size classes (e.g., Lewison & Crowder, 2007). However, the capture of at least 227 juvenile hawksbills during our study from a severely depleted population consisting of fewer than 700 mature females is a cause for concern. While the absolute bycatch at our study sites appears to be relatively high for hawksbills (Wallace et al., 2013), it is comparatively lower than some other sea turtle species captured in net fisheries (Peckham et al., 2007). This disparity in absolute bycatch may reflect the relative abundance of species in a given ocean basin, where there can be a difference of orders of magnitude among populations and species (Donoso & Dutton, 2010). For example, absolute hawksbill bycatch from our study is overshadowed by numbers of hawksbills harvested for the tortoiseshell trade in the Indo-Pacific (Van Dijk & Shepherd, 2004) and by the number of green turtles (Chelonia mydas) consumed by humans at Baja California Sur, Mexico in the North Pacific (Senko et al., 2014), where hawksbills and green turtles are more abundant than eastern Pacific hawksbills (Seminoff, 2004; Mortimer & Donnelly, 2008). When combining our bycatch data of juveniles in lobster gillnet fisheries with those of other fisheries in the region, particularly juveniles killed in bottom-set gillnets in Mexico (Seminoff et al., 2003; Koch et al., 2006), Ecuador (Alava et al., 2005), and Peru (Alfaro- Shigueto et al., 2010a, 2010b; Ortiz et al., 2016), and juvenile and adult mortality in bottom-set longlines (M. Liles, pers. obs.) and blast fishing (Liles et al., 2011) in mangrove estuaries in El Salvador, hawksbill interactions with small-scale fisheries reportedly incur the largest cumulative impact to the species in the eastern Pacific and are of grave concern. Studies aimed at describing bycatch also can help elucidate cryptic life stages of species that occur in unknown or relatively inaccessible locations, such as the oceanic pelagic stage of post-hatchling and small juvenile sea turtles (Reich et al., 2007; Putman & Naro- Maciel, 2013; Van Houtan et al., 2016). Despite few available data, the existence of an oceanic pelagic stage during at least the first 1-2 years of a sea turtle s life, commonly referred to as the lost years (Carr, 1986), is a life history hypothesis widely accepted for all sea turtle species, except the flatback (Natator depressus)

102 532 Latin American Journal of Aquatic Research in Australia (Carr, 1987; Musick & Limpus, 1997; Bolten, 2003; Reich et al., 2007; Witherington et al., 2012; Mansfield et al., 2014). However, because direct observations of post-hatchling and small juvenile sea turtles in developmental habitat are so rare, particularly for hawksbills, researchers often must infer a species presence or absence from indirect findings (Putman et al., 2014). Van Houtan et al. (2016), for example, inferred from diverse indirect sources -including strandings, stomach contents, satellite drifters, and longline bycatch records- that hawksbills 0-4 years of age measuring 8-34 cm straight carapace length occurred predominantly in coastal pelagic waters of Hawaii and that absence of hawksbill bycatch in longlines may indicate a lack of prolonged presence in oceanic habitats in the North Pacific. Biases inherent in utilizing indirect approaches to determine presence of specific size classes of sea turtle species, such as demographic selectivity of fishing gear, targeted fishing areas, and population size, can increase uncertainty of findings but often are required when fisheries-independent approaches or direct observations are unavailable (Alfaro Shigueto et al., 2008). Direct observations of hawksbill captures at our study sites revealed that 11% (n = 3) and 19% (n = 5) of measured hawksbills (n = 27) had a mean carapace length of 15 cm and <20 cm, were captured at depths of 6.5 ± 4.0 m and 7.0 ± 3.7 m, and were located 0.51 ± 0.3 km and 0.69 ± 0.4 km from shore, respectively (Fig. 4). This is the first direct evidence that the lost years, or pelagic juvenile phase, for eastern Pacific hawksbills may be shorter than previously thought, or nonexistent, and questions whether hawksbills in this region conform to life-history patterns exhibited by other species and populations that undergo an ontogenetic shift from oceanic habitat to neritic habitat or whether an early oceanic-stage is entirely absent and recruitment to nearshore foraging areas occurs immediately after hatching. However, we recognize that our small sample size (n = 8 hawksbills) inhibits us from drawing firm conclusions and it is possible that these smaller turtles represent the smallest size class for new recruits, as some sea turtle species can partition size based on depth with smaller turtles occurring in shallower water (e.g., Koch et al., 2007). Regardless, our findings underscore the urgent need to mitigate inwater threats at rocky reefs utilized by all life stages of eastern Pacific hawksbills. Bycatch of long-lived marine megafauna in smallscale fisheries presents a formidable challenge and underscores the need for cross-sector collaboration to safeguard the distinct life stages of species across diverse habitats. Because many small-scale fisheries occur in low-income regions, depressed socioeconomic conditions can dictate the utilization of certain fishing practices (e.g., use of destructive fishing gear at coral reef ecosystems; Cinner, 2010). Bycatch reduction strategies that incorporate local fishers into decisionmaking processes could minimize threats to fishers livelihoods and increase the long-term effectiveness of strategy implementation. CONSERVATION IMPLICATIONS In low-income regions where natural resource management and environmental law enforcement are typically weaker, primary resource users often selfgovern resource use in their local environment (Dietz et al., 2003), such as lobster gillnet fishing in El Salvador and Nicaragua. This local reality highlights the power and control local fishers wield in determining the success or failure of hawksbill conservation initiatives (Liles et al., 2015a) and underscores the importance of understanding the factors that motivate or discourage their participation in conservation (Liles et al., 2016; Peterson & Liles, in press), especially as it relates to bycatch mitigation activities. Further, direct participation of local fishers that draws from their large repository of knowledge will be essential for identifying effective long-term solutions to bycatch. Such solutions must simultaneously maintain or increase target catches, if they are to be accepted among fishing communities (Gilman et al., 2005; Lewison et al., 2011; Peckham et al., 2016). Close collaboration with lobster gillnet fishers at our study sites offers important opportunities to assess potential lowerbycatch alternatives to lobster gillnets that have been successful in other regions, such as lobster traps (Shester & Micheli, 2011) and net illumination (Wang et al., 2013; Ortiz et al., 2016). Although these alternatives were not specifically tested in areas with operating lobster gillnet fisheries, they provide instructional opportunities for gear experimentation in areas where lobster gillnet fishing is prevalent, such as El Salvador and Nicaragua. Experimental lobster trap trials at Punta Amapala and La Salvia in successfully eliminated hawksbill captures, but were ineffective at maintaining lobster yields (n = 0 kg total), regardless of bait types and spatiotemporal variation of sets (Gaos et al., 2014a), and preliminary net illumination trials at Punta Amapala in have been inconclusive due to the absence of hawksbill captures in paired net trails (C. Pacheco, pers. obs.), which might be confounded by environmental conditions at our study sites, such as high turbidity of water and strong currents. Despite discouraging results from initial bycatch reduction trials with traps and net illumination, spatiotemporal patterns in observed hawksbill bycatch

103 Hawksbill turtles bycatch in Eastern Pacific lobster gillnets 533 and fishing effort during our study provide insight into potential community-based solutions to bycatch that could be tailored to local realities. At Punta Amapala and La Salvia, 21% (n = 6) of bycatch events occurred at depths of 1-5 m, while only 6% (n = kg) and 7% (n = 37) of total lobster catches and gillnet sets occurred in the same range of depths, respectively (Fig. 3). Engaging local fishers to develop and adhere to shared conservationist norms that prohibit gillnet fishing in waters 1-5 m deep could drastically reduce hawksbill interactions across sites without substantially affecting lobster catches. Similarly, reducing soaktimes of nets (overall mean of 17.5 h) might reduce hawksbill bycatch and mortality from drowning, particularly at Punta Amapala where fishers soaked nets an average of 20.3 h with 100% hawksbill mortality, compared to nets set at Los Cóbanos and La Salvia for 13.4 h with 0% mortality and 16.5 h with 64% mortality, respectively (Table 2). Model-selection supports this assertion, as soaktime was in 10 of the 23 bestsupported models, including the top model, highlighting its importance in predicting hawksbill capture in lobster gillnets and the likely effectiveness of soaktime reduction (Table 4). However, local fishers often view maximizing fishing effort (i.e., long net lengths and soaktimes) as a way to reduce economic expenditures -longer soaktimes reduce trips to fishing sites resulting in lower costs attributed to transport- and therefore might be reluctant to adopt a practice perceived as an economic burden. However, conducting educational outreach via workshops with fishers that highlight the precarious conservation status of sea turtles, particularly hawksbills, and the low sustainability of bycatch, coupled with information on the proper handling of incidentally captured sea turtles (e.g., Sea Turtle Handling Guidebook for Fishermen; Gerosa & Aureggi, 2001), could help persuade fishers to modify their spatiotemporal fishing activities, reduce soak times, and enable them to reduce post-release mortality. In addition to potential spatiotemporal adjustments in fishing effort to mitigate bycatch, local fishers at Punta Amapala expressed interest in establishing a community-based protected area at Poza de la Gata (C. Pacheco, pers. obs.), which is a portion of rocky reef where relatively high numbers of hawksbills forage and rest in shallow subtidal caves (Domínguez-Miranda, 2010), and where nearly 40% (n = 5) of hawksbill capture and mortality occurred with only marginal fishing effort (Fig. 1). MPAs and area closures that eliminate gillnet fishing can produce positive population responses of both non-target and target species within demarcated borders (Regular et al., 2013) and across a spillover-induced density gradient outside the borders of the protected area (Gell & Roberts, 2003; Goñi et al., 2006; Forcada et al., 2009). This is particularly true for coral and rocky reef ecosystems in low-income regions, where ubiquitous use of destructive fishing gears, such as gillnets, can degrade habitat and capture high proportions of immature individuals of target species ultimately leading to reduced yields (Cinner, 2010), which is currently observed in lobster gillnet fisheries at Los Cóbanos and Punta Amapala (CENDEPESCA, 2012). Additionally, because high-maintained fishing effort can suppress biomass and alter the structure of target species populations (Campbell & Pardede, 2006), establishing communally-accepted no-take areas may relieve fishing pressure and create sustained nursery areas for lobster development, which could stimulate increases in biomass and restore the structural integrity of lobster populations across rocky reefs where overfishing occurs. However, local realities within which lobster fishers operate could challenge their ability to collectively manage resources effectively, including potential deficiencies in social cohesion and organization (Gutiérrez et al., 2011), disproportionate distributions of power and political influence in communities (Agrawal & Gibson, 1999), and lack of government support in regulating and sanctioning offenders (Liles et al., 2015a). Additionally, because gang- and drug-related crime is pervasive throughout El Salvador, which has situated the country among the most violent nations in the world (Carcach, 2015) and has impacted conservation at nesting beaches along the Salvadoran coast (Liles et al., 2011; Gaos et al., 2017), local fishers could be hesitant to engage in peerpressure of nonconformists and rule enforcement due to fear of reprisals. Nonetheless, fisher interest in establishing a community-based protected area at Poza de la Gata to safeguard developmental areas of hawksbills and lobsters might surmount these impediments, which would represent the first protected area in El Salvador that was designed (e.g., delimit borders and devise rules) and managed (e.g., monitor congruence between established rules and fisher behavior) by local resource users, and could serve as a model for locally-driven conservation schemes at Los Cóbanos and La Salvia, as well as other areas where effective management and enforcement of resource use are limited. For future natural science studies, we recommend that researchers continue partnering with local fishers to assess spatiotemporal patterns and demography of hawksbill bycatch in lobster gillnet fisheries along the Pacific coast of Central America, particularly at Los Cóbanos and Isla Amapala in Honduras, over multiple years to better guide evidence-based bycatch reduction efforts and provide more accurate estimates of cumulative impacts to the population. Additionally,

104 534 Latin American Journal of Aquatic Research future research should quantify the amount and composition of non-hawksbill bycatch in lobster gillnet fisheries to evaluate potential impacts to marine invertebrates and benthic habitats, which would facilitate the development of ecosystem-based management strategies at rocky reefs. Finally, we suggest future studies focus on elucidating habitat use by posthatchling and early juvenile hawksbills in the eastern Pacific, particularly at inshore mangrove estuaries in El Salvador and Nicaragua where >80% of nesting activity occurs (Gaos et al., 2017), to inform protection measures at critical developmental habitats in the region and to determine the existence of a pelagic juvenile phase for this population. Future social science research should assess the socioeconomic, political, and cultural characteristics of lobster fishing communities to better understand their adaptive capacity in collectively responding to resource fluctuations and ability to diversify livelihood strategies. We also suggest that research focus on how lobster fishers view their fishing practice and associated bycatch, what motivates their participation in lobster gillnet fishing, and what interpersonal connections facilitate or impede their activities in El Salvador in Nicaragua, to help identify locally relevant alternatives to damaging fishing practices that aim to reduce hawksbill mortality and negative impacts on rocky reef ecosystems. CONCLUSIONS Unlike other areas of the eastern Pacific where opportunistic direct take of in-water hawksbills is prevalent (e.g., Baja California Peninsula, Mexico; Seminoff et al., 2003; Koch et al., 2006; Peckham et al., 2008; Peru; Alfaro-Shigueto et al., 2011), this practice is relatively rare along the Pacific coast of El Salvador and Nicaragua. Indirect take of hawksbills through fisheries bycatch, however, is common and threatens the species in nearshore, developmental habitat. Our findings demonstrate that lobster gillnet fisheries constitute a major source of mortality that likely has seriously negative impacts on the severely depleted eastern Pacific hawksbill population and support the growing body of literature that highlights small-scale fisheries as a key threat to the recovery of some sea turtle populations. Spatiotemporal patterns in fishing effort and bycatch observed in our study - including greater likelihood of hawksbill capture in nets set at shallow depths with longer soaktimes during the peak lobster fishing season- offer guidance for potential community-based mitigation strategies in lobster gillnet fisheries in El Salvador and Nicaragua. Limited capacity of many Central American governments to enforce environmental laws that protect endangered species elevates the role local fishers play in selfgoverning resource use and underscores the importance of partnering with local fishers to effectively conserve and recover hawksbills in the eastern Pacific. ACKNOWLEDGEMENTS We thank local fishers at Los Cóbanos, Punta Amapala, and La Salvia for their trust and collaboration, this study would have been impossible without their knowledge and participation. We acknowledge A. Gutiérrez, C. Dueñas, E. López, G. Mariona and M. Vásquez for assistance. Constructive comments on earlier drafts from J. Mangel, J. Senko, and one anonymous reviewer were greatly appreciated. We are grateful to national environmental authorities in El Salvador (MARN) and Nicaragua (MARENA) for permits. We thank the Disney Worldwide Conservation Fund, Sungevity, U.S. Agency for International Development, U.S. National Fish and Wildlife Foundation, U.S. National Oceanic and Atmospheric Administration, and The Ocean Foundation for funding or logistical support. REFERENCES Agrawal, A. & C.C. Gibson Enchantment and disenchantment: the role of community in natural resource conservation. World Develop., 27: Alava, J.J., P. Jiménez, M. Peñafiel, W. Aguirre & P. Amador Sea turtle strandings and mortality in Ecuador: Mar. Turtle Newslett., 108: 4-7. Alfaro-Shigueto, J., J.C. Mangel, J.A. Seminoff & P.H. Dutton Demography of loggerhead turtles Caretta caretta in the southeastern Pacific Ocean: fisheries-based observation and implications for management. Endang. Species Res., 5: Alfaro-Shigueto, J., J.C. Mangel, F. Bernedo, P.H. Dutton, J.A. Seminoff & B.J. Godley Small-scale fisheries of Peru: a major sink for marine turtles in the Pacific. J. Appl. Ecol., 48: Alfaro-Shigueto, J., J.C. Mangel, C. Caceres, J.A. Seminoff, A. Gaos & I. Yañez. 2010a. Hawksbill turtles in Peruvian coastal fisheries. Mar. Turtle Newslett., 129: Alfaro-Shigueto, J., J.C. Mangel, M. Pajuelo, P.H. Dutton, J.A. Seminoff & B.J. Godley. 2010b. Where small can have a large impact: structure and characterization of small-scale fisheries in Peru. Fish. Res., 106: Arauz, R A description of the Central American shrimp fisheries with estimates of incidental capture and mortality of sea turtles. In: Proceedings of the 15 th

105 Hawksbill turtles bycatch in Eastern Pacific lobster gillnets 535 Annual Symposium on Sea Turtle Biology and Conservation, Hilton Head, South Carolina, pp Aucoin, S. & Y. León Preliminary data on hawksbill turtle (Eretmochelys imbricata) bycatch in an artisanal gillnet used near Jaragua National Park, Dominican Republic. Proceedings of the 60th Gulf and Caribbean Fisheries Institute, Punta Cana, Dominican Republic, pp Bell, I Algivory in hawksbill turtles: Eretmochelys imbricata food selection within a foraging area on the Northern Great Barrier Reef. Mar. Ecol., 34: Beltrán, C Evaluación de la pesca de pequeña escala y aspectos de ordenación en cinco países seleccionados de América Latina: El Salvador, Costa Rica, Panamá, Colombia y Ecuador. Períodos, FAO Circular de Pesca, N 957/2, Rome, 71 pp. Béné, C Small-scale fisheries: assessing their contribution to rural livelihoods in developing countries. FAO Fisheries Circular No. 1008, Rome, 48 pp. Benson, S.R., R.F. Tapilatu, N. Pilcher, P.S. Tomillo & L.S. Martínez Leatherback turtle populations in the Pacific Ocean. In: J.R. Spotila & P.S. Tomillo (eds.). The leatherback turtle: biology and conservation. JHU Press, Baltimore, pp Blumenthal, J.M., T.J. Austin, J.B. Bothwell, A.C. Broderick, G. Ebanks-Petrie, J.R. Olynik, M.F. Orr, J.L. Solomon, M.J. Witt & B.J. Godley Diving behavior and movements of juvenile hawksbill turtles Eretmochelys imbricata on a Caribbean coral reef. Coral Reefs, 28: Bolten, A.B Variation in sea turtle life history patterns: neritic versus oceanic developmental stages. In: P.L. Lutz, J.A. Musick & J. Wyneken (eds.). The biology of sea turtles. CRC Press, Boca Raton, pp Burnham, K.P. & D.R. Anderson Model selection and multi-model inference: a practical informationtheoretic approach. Springer-Verlag, New York, 488 pp. Burnham, K.P., D.R. Anderson & K.P. Huyvaert AIC model selection and multimodel inference in behavioral ecology: some background, observations, and comparisons. Behav. Ecol. Sociobiol., 65: Caillouet Jr., C.W., D.J. Shaver & W.G. Teas Relationship between sea turtle stranding rates and shrimp fishing intensities in the northwestern Gulf of Mexico: versus Fish. Bull., 94: Campbell, S.J. & S.T. Pardede Reef fish structure and cascading effects in response to artisanal fishing pressure. Fish. Res., 79: Carcach, C A spatio-temporal model of homicide in El Salvador. Crime Sci., 4. doi: /s Carr, A Rips, FADS, and little loggerheads. BioScience, 36: Carr, A New perspectives on the pelagic stage of sea turtle development. Conserv. Biol., 1: Carrión-Cortez, J., C. Canales-Cerro, R. Arauz & R. Riosmena-Rodríguez Habitat use and diet of juvenile eastern Pacific hawksbill turtles (Eretmochelys imbricata) in the North Pacific coast of Costa Rica. Chelonian Conserv. Biol., 12: Centro de Desarrollo de la Pesca y Acuicultura (CENDEPESCA) Caracterización de la pesquería de la langosta (Panulirus gracilis) en Playas Negras, La Unión, y Puerto Acajutla, Sonsonate, El Salvador. Centro de Desarrollo de la Pesca y Acuicultura, San Salvador, 23 pp. Chuenpagdee, R., L. Liguori, M.L.D. Palomares & D. Pauly Bottom-up, global estimates of smallscale marine fisheries catches. Fisheries Centre Research Reports 14, Vancouver, 105 pp. Cinner, J.E Poverty and the use of destructive fishing gear near east African marine protected areas. Environ. Conserv., 36: Cortés, J Coastal morphology and coral reefs. In: J. Bundschuh & G.E. Alvarado (eds.). Central America: geology, resources and hazards. Taylor & Francis, London, pp Crouse, D.T., L.B. Crowder & H. Caswell A stagebased population model from loggerhead sea turtles and implications for conservation. Ecology, 68: Dapp, D., R. Arauz, J.R. Spotila & M.P. O'Connor Impact of Costa Rican longline fishery on its bycatch of sharks, stingrays, bony fish and olive ridley turtles (Lepidochelys olivacea). J. Exp. Mar. Biol. Ecol., 448: De Paz, C. & S. Sui Observadores abordo. San Salvador, El Salvador, 58 pp. Dietz, T., E. Ostrom & P.C. Stern The struggle to govern the commons. Science, 302: Domínguez-Miranda, J.P Caracterización biofísca del área marina frente a Playa Las Tunas, Playas Negras, Playas Blancas, Playa Maculís, y las Mueludas, Municipio de Conchagua, Departamento de La Unión, El Salvador. Technical Report, San Salvador, 83 pp. Donoso, M. & P.H. Dutton Sea turtle bycatch in the Chilean pelagic longline fishery in the southeastern Pacific: opportunities for conservation. Biol. Conserv., 143:

106 536 Latin American Journal of Aquatic Research Ehrhardt, N.M Population dynamic characteristics and sustainability mechanisms in key western central Atlantic spiny lobster, Panulirus argus, fisheries. Bull. Mar. Sci., 76: Food and Agriculture Organization (FAO) Resumen informativo sobre la pesca por países. La República de Nicaragua. Food and Agriculture Organization, Rome, 23 pp. Food and Agriculture Organization (FAO) Resumen informativo sobre la pesca por países. La República de Panamá. Food and Agriculture Organization, Rome, 25 pp. Field, A.H. & J.F. Bell A review of methods for the assessment of prediction errors in conservation presence-absence models. Environ. Conserv., 24: Forcada, A., C. Valle, P. Bonhomme, G. Criquet, G. Cadiou, P. Lenfant & J.L. Sánchez-Lizaso Effects of habitat on spillover from marine protected areas to artisanal fisheries. Mar. Ecol. Prog. Ser., 379: Gaos, A.R., M.J. Liles & J. Urteaga. 2014b. Eastern Pacific hawksbill initiative: phase III. Final Programmatic Report. San Diego, 10 pp. [ hawksbill.org/wp-content/uploads/2010/2009/gaos-et -al.-2014_final-programmatic-report_icapo_ pdf]. Gaos, A.R., A. Bolaños, M.J. Liles & R. Arauz. 2014a. Hawksbill bycatch in lobster gillnet fisheries operating along the Pacific coast of Central America. Annual Symposium on Sea Turtle Biology and Conservation, New Orleans, 99 pp. Gaos, A.R., R.L. Lewison, B.P. Wallace, I.L. Yañez, M.J. Liles, W.J. Nichols, A. Baquero, C.R. Hasbun, M. Vasquez, J. Urteaga & J.A. Seminoff. 2012a. Spatial ecology of critically endangered hawksbill turtles Eretmochelys imbricata: implications for management and conservation. Mar. Ecol. Prog. Ser., 450: Gaos, A.R., R.L. Lewison, I.L. Yañez, B.P. Wallace, M.J. Liles, W.J. Nichols, A. Baquero, C.R. Hasbun, M. Vasquez, J. Urteaga & J.A. Seminoff. 2012b. Shifting the life-history paradigm: discovery of novel habitat use by hawksbill turtles. Biol. Lett., 8: Gaos, A.R., R.R. Lewison, B.P. Wallace, I.L. Yañez, M.J. Liles, A. Baquero & J.A. Seminoff. 2012c. Dive behaviour of adult hawksbills (Eretmochelys imbricata Linnaeus, 1766) in the eastern Pacific Ocean highlights shallow depth use by the species. J. Exp. Mar. Biol. Ecol., 432: Gaos, A.R., M.J. Liles, V. Gadea, A. Peña de Niz, F. Vallejo, C. Miranda, J.J. Darquea, A. Henríquez, E. Altamirano, A. Rivera, S. Chavarría, D. Melero, J. Urteaga, C. M. Pacheco, D. Chácon, C. LeMarie, J. Alfaro-Shigueto, J.C. Mangel, I.L. Yañez & J. A. Seminoff Living on the edge: hawksbill nesting and conservation along the eastern Pacific rim. Lat. Am. J. Aquat. Res., 45(3): Gaos, A.R., F.A. Abreu-Grobois, J. Alfaro-Shigueto, D. Amorocho, R. Arauz, A. Baquero, R. Briseño et al Signs of hope in the eastern Pacific: international collaboration reveals encou-raging status for a severely depleted population of hawksbill turtles Eretmochelys imbricata. Oryx, 44: Gardner, B., P.J. Sullivan, S.J. Morreale & S.P. Epperly Spatial and temporal statistical analysis of bycatch data: patterns of sea turtle bycatch in the North Atlantic. Can. J. Fish. Aquat., 65: Gell, F.R. & C.M. Roberts Benefits beyond boundaries: the fishery effects of marine reserves. Trends Ecol. Evol., 18: Gerosa, G. & M. Aureggi Sea turtle handling guidebook for fishermen. UNEP/MAP RAC/SPA, Tunis, 31 pp. Gilman, E., N. Brothers & D. Kobayashi Principles and approaches to abate sea bird bycatch in longline fisheries. Fish Fish., 6: Gilman, E., D. Kobayashi, T. Swenarton, N. Brothers, P. Dalzell & I. Kinan-Kelly Reducing sea turtle interactions in the Hawaii-based longline swordfish fishery. Biol. Conserv., 139: Gilman, E., J. Gearhart, B. Price, S. Eckert, H. Milliken, J. Wang, Y. Swimmer, D. Shiode, O. Abe, S.H. Peckham, M. Chaloupka, M. Hall, J. Mangel, J. Alfaro-Shigueto, P. Dalzell & A. Ishizaki Mitigating sea turtle by-catch in coastal passive net fisheries. Fish Fish., 11: Giraldes, B.W., A.Z. Silva, F.M. Correa & D.M. Smyth Artisanal fishing of spiny lobsters with gillnets - a significant anthropogenic impact on tropical reef ecosystem. Global Ecol. Conserv., 4: Goñi, R., A. Quetglas & O. Reñones Spillover of spiny lobsters Palinurus elephas from a marine reserve to an adjoining fishery. Mar. Ecol. Prog. Ser., 308: Gutiérrez, N.L., R. Hilborn & O. Defeo Leadership, social capital and incentives promote successful fisheries. Nature, 470: Hall, M.A., D.L. Alverson & K.I. Metuzals Bycatch: problems and solutions. Mar. Pollut. Bull., 41: Hart, K.M., P. Mooreside & L.B. Crowder Interpreting the spatio-temporal patterns of sea turtle strandings: going with the flow. Biol. Conserv., 129:

107 Hawksbill turtles bycatch in Eastern Pacific lobster gillnets 537 Heppell, S.S Application of life-history theory and population model analysis to turtle conservation. Copeia, 1998: Heppell, S.S., S.A. Heppell, A.J. Read & L.B. Crowder Effects of fishing on long-lived marine organisms. In: E.A. Norse & L.B. Crowder (eds.). Marine conservation biology: the science of maintaining the sea's biodiversity. Island Press, Washington, pp Hosmer, D.W. & S. Lemeshow Applied logistic regression. John Wiley & Sons, Hoboken, 375 pp. Houghton, J.D.R., M.J. Callow & G.C. Hays Habitat utilization by juvenile hawksbill turtles (Eretmochelys imbricata Linnaeus, 1766) around a shallow water coral reef. J. Nat. Hist., 37: Huang, H Incidental catch of seabirds and sea turtles by Taiwanese longline fleets in the Pacific Ocean. Fish. Res., 170: Kelleher, K Discards in the world's marine fisheries: an update. FAO Fisheries Technical Paper N 470, Rome, 131 pp. Koch, V., L.B. Brooks & W.J. Nichols Population ecology of the green/black turtle (Chelonia mydas) in Bahía Magalena, Mexico. Mar. Biol., 153: Koch, V., W.J. Nichols, H. Peckham & V. de la Toba Estimates of sea turtle mortality from poaching and bycatch in Bahia Magdalena, Baja California Sur, Mexico. Biol. Conserv., 128: Kot, C.Y., A.M. Boustany & P.N. Halpin Temporal patterns of target catch and sea turtle bycatch in the US Atlantic pelagic longline fishing fleet. Can. J. Fish. Aquat., 67: Leon, Y.M. & K.A. Bjorndal Selective feeding in the hawksbill turtle, an important predator in coral reef ecosystems. Mar. Ecol. Prog. Ser., 245: Lewison, R.L. & L.B. Crowder Estimating fishery bycatch and effects on a vulnerable seabird population. Ecol. Appl., 13: Lewison, R.L. & L.B. Crowder Putting longline bycatch of sea turtles into perspective. Conserv. Biol., 21: Lewison, R.L., L.B. Crowder, A.J. Read & S.A. Freeman Understanding impacts of fisheries bycatch on marine megafauna. Trends Ecol. Evol., 19: Lewison, R.L., B. Wallace, J. Alfaro-Shigueto, J.C. Mangel, S.M. Maxwell & E.L. Hazen Fisheries bycatch of marine turtles: lessons learned from decades of research and conservation. In: J. Wyneken, K. Lohmann & J. Musick (eds.). The biology of sea turtles. CRC Press, Boca Raton, pp Lewison, R.L., C.U. Soykan, T. Cox, H. Peckham, N. Pilcher, N. LeBoeuf, S. McDonald, J. Moore, C. Safina & L.B. Crowder Ingredients for addressing the challenges of fisheries bycatch. Bull. Mar. Sci., 87: /bms Lewison, R.L., L.B. Crowder, B.P. Wallace, J.E. Moore, T. Cox, R. Zydelis, S. McDonald, A. Dimatteo, D.C. Dunn, C.Y. Kot, R. Bjorkland, S. Kelez, C. Soykan, K.R. Stewart, M. Sims, A. Boustany, A.J. Read, P. Halpin, W.J. Nichols & C. Safina Global patterns of marine mammal, seabird, and sea turtle bycatch reveal taxa-specific and cumulative megafauna hotspots. PNAS, 111: Liles, M.J. & C. Thomas Sea turtle priority conservation areas in the coastal waters of El Salvador. San Salvador, 45 pp. [ Reviewed: 10 June Liles, M.J., S. Chavarría, E. Altamirano, I. Yañez, J. Urteaga & A.R. Gaos Community construction through culturally rooted celebration: turtles all the way down. In: A.M. Feldpausch-Parker, H. Bergea, T.R. Peterson & K. Raitio (eds.). Environmental communication and community: constructive and deconstructive dynamics of social transformation. Routledge, London, pp Liles, M.J., M.V. Jandres, W.A. Lopez, G.I. Mariona, C.R. Hasbun & J.A. Seminoff Hawksbill turtles Eretmochelys imbricata in El Salvador: nesting distribution and mortality at the largest remaining nesting aggregation in the eastern Pacific Ocean. Endang. Species Res., 14: Liles, M.J., M.J. Peterson, Y.S. Lincoln, J.A. Seminoff, A.R. Gaos & T.R. Peterson. 2015a. Connecting international conservation priorities with human wellbeing in low-income nations: lessons from hawksbill turtle conservation in El Salvador. Local Environ., 20: Liles, M.J., M.J. Peterson, J.A. Seminoff, E. Altamirano, A.V. Henríquez, A.R. Gaos, V. Gadea, J. Urteaga, P. Torres, B.P. Wallace & T.R. Peterson. 2015b. One size does not fit all: importance of adjusting conservation practices for endangered hawksbill turtles to address local nesting habitat needs in the eastern Pacific Ocean. Biol. Conserv., 184: Mangel, J.C., J. Alfaro Shigueto, K.V. Waerebeek, C. Cáceres, S. Bearhop, M.J. Witt & B.J. Godley Small cetacean captures in Peruvian artisanal fisheries: high despite protective legislation. Biol. Conserv., 143: Mansfield, K.L., J. Wyneken, W.P. Porter & J. Luo First satellite tracks of neonate sea turtles redefine the 'lost years' oceanic niche. Proc. R. Soc. B, 281: doi: /rspb Meylan, A Spongivory in hawksbill turtles: a diet of glass. Science, 239:

108 538 Latin American Journal of Aquatic Research Miller, J.D., K.A. Dobbs, C.J. Limpus, N. Mattocks & A.M. Landry Long-distance migrations by the hawksbill turtle, Eretmochelys imbricata, from northeastern Australia. Wild. Res., 25: Moore, J.E., T.M. Cox, R.L. Lewison, A.J. Read, R. Bjorkland, S.L. McDonald, L.B. Crowder, E. Aruna, I. Ayissi, P. Espeut, C. Joynson-Hicks, N. Pilcher, C.N.S. Poonian, B. Solarin & J. Kiszka An interview-based approach to assess marine mammal and sea turtle captures in artisanal fisheries. Biol. Conserv., 143: Mortimer, J.A. & M. Donnelly Eretmochelys imbricata. IUCN Red List of Threatened Species: e.t8005a UK.2008.RLTS.T8005A en, Available at: [ Reviewed: 20 September Musick, J.A. & C.J. Limpus Habitat utilization and migration in juvenile sea turtles. In: P.L. Lutz & J.A. Musick (eds.). The biology of sea turtles. CRC Press, Washington, pp Orellana-Amador, J.J Peces marinos de Los Cóbanos, El Salvador. Fundación Sigma, Buffalo, 126 pp. Ortiz, N., J.C. Mangel, J. Wang, J. Alfaro-Shigueto, S. Pingo, A. Jiménez, T. Suárez, Y. Swimmer, F. Carvalho & B.J. Godley Reducing green turtle bycatch in small-scale fisheries using illuminated gillnets: the cost of saving a sea turtle. Mar. Ecol. Prog. Ser., 545: Pauly, D. & D. Zeller Catch reconstructions reveal that global marine fisheries catches are higher than reported and declining. Nature: /ncomms Peckham, S.H., D. Maldonado-Díaz, A. Walli, G. Ruiz, L.B. Crowder & W.J. Nichols Small-scale fisheries bycatch jeopardizes endangered Pacific loggerhead turtles. PLoS ONE, 2: e1041. Peckham, S.H., J. Lucero-Romer, D. Maldonado-Díaz, A. Rodríguez, J. Senko, M. Wojakowski & A.R. Gaos Buoyless nets reduce sea turtle bycatch in coastal net fisheries. Conserv. Lett., 9: Peckham, S.H., D. Maldonado-Díaz, V. Koch, A. Mancini, A. Gaos, M.T. Tinker & W.J. Nichols High mortality of loggerhead turtles due to bycatch, human consumption and strandings at Baja California Sur, Mexico, 2003 to Endang. Species Res., 5: Pérez-González, R Catch composition of the spiny lobster Panulirus gracilis (Decapoda: Palinuridae) off the western coast of Mexico. Lat. Am. J. Aquat. Res., 39: Peterson, T.R. & M.J. Liles. Hawksbill turtle conservation in El Salvador: weighing biodiversity and livelihood risks. In: T. Avtgis, D. Rodríguez & C. Liberman (eds.). Casing crisis and risk communication. Kendall Hunt Publishing, Dubuque (in press). Putman, N.F. & E. Naro-Maciel Finding the 'lost years' in green turtles: insights from ocean circulation models and genetic analysis. Proc. R. Soc. B, 280: doi: /rspb Putman, N.F., F.A. Abreu-Grobois, A.C. Broderick, C. Ciofi, A. Formia, B.J. Godley, S. Stroud, T. Pelembe, P. Verley & N. Williams Numerical dispersal simulations and genetics help explain the origin of hawksbill sea turtles in Ascension Island. J. Exp. Mar. Biol. Ecol., 450: Regular, P., W. Montevecchi, A. Hedd, G. Robertson & S. Wilhelm Canadian fishery closures provide a large-scale test of the impact of gillnet bycatch on seabird populations. Biol. Lett., 9: doi: /rspb Reich, K.J., K.A. Bjorndal, A.B. Bolten The 'lost years' of green turtles: using stable isotopes to study cryptic life stages. Biol. Lett., 3: Reyes-Bonilla, H. & J.E. Barraza Corals and associated marine communities from El Salvador. In: J. Cortéz (ed.). Latin American coral reefs. Elsevier, Amsterdam, pp Riskas, K.A., M.M.P.B. Fuentes & M. Hamann Justifying the need for collaborative management of fisheries bycatch: a lesson from marine turtles in Australia. Biol. Conserv., 196: Salas, S., R. Chuenpagdee, A. Charles & J.C. Seijo (eds.) Coastal fisheries of Latin America and the Caribbean. Food and Agricultural Organization, Technical paper N 544, Rome, 430 pp. Salas, S., R. Chuenpagdee, J.C. Seijo & A. Charles Challenges in the assessment and management of small-scale fisheries in Latin America and the Caribbean. Fish. Res., 87: Seminoff, J.A Chelonia mydas. IUCN Red List of Threatened Species 2004: e.t4615a dx.doi.org/ /iucn.uk.2004.rlts.t4615a en, Available at: [ org/]. Reviewed: 20 September Seminoff, J.A., W.J. Nichols, A. Resendiz & L.B. Brooks Occurrence of hawksbill turtles, Eretmochelys imbricata (Reptilia: Cheloniidae), near the Baja California Peninsula, Mexico. Pac. Sci., 57: Senko, J., A. Mancini, J.A. Seminoff & V. Koch Bycatch and directed harvest drive high green turtle mortality at Baja California Sur, Mexico. Biol. Conserv., 169: Shester, G.G. & F. Micheli Conservation challenges for small-scale fisheries: bycatch and habitat impacts of traps and gillnets. Biol. Conserv., 144:

109 Hawksbill turtles bycatch in Eastern Pacific lobster gillnets 539 Snape, R.T.E., D. Beton, A.C. Broderick, B.A. Cicek, W.J. Fuller, O. Ozden & B.J. Godley Stranding monitoring and anthropological surveys provide insight into marine turtle bycatch in small-scale fisheries of the eastern Mediterranean. Chelonian Conserv. Biol., 12: Snoddy, J.E. & A. Southwood-Williard Movements and post-release mortality of juvenile sea turtles released from gillnets in the lower Cape Fear River, North Carolina, USA. Endang. Species Res., 12: Soykan, C.U., J.E. Moore, R. Zydelis, L.B. Crowder, C. Safina & R.L. Lewison Why study bycatch? An introduction to the Theme Section on fisheries bycatch. Endang. Species Res., 5: Stewart, K.R., R.L. Lewison, D.C. Dunn, R.H. Bjorkland, S. Kelez, P.N. Halpin & L.B. Crowder Characterizing fishing effort and spatial extent of coastal fisheries. PLoS ONE, 5: e Swets, J.A Measuring the accuracy of diagnostic systems. Science, 240: Tóbon-López, A. & D.F. Amorocho-Llanos Population study of the hawksbill turtle Eretmochelys imbricata (Cheloniidae) in the southern Pacific region of Colombia. Acta Biol. Colomb., 19: Van Dam, R.P., C.E. Diez, G.H. Balazs, L.A. Colón Colón, W.O. McMillan & B. Schroeder Sexspecific migration patterns of hawksbill turtles breeding at Mona Island, Puerto Rico. Endang. Species Res., 4: Van Dijk, P.P. & C.R. Shepherd Shelled out? A snapshot of bekko trade in selected locations in Southeast Asia. A Traffic South-east Asia Report, 29 pp. Van Houtan, K.S., D.L. Francke, S. Alessi, T.T. Jones, S.L. Martin, L. Kurpita, C.S. King & R.W. Baird The developmental biogeography of hawksbill sea turtles in the North Pacific. Ecol. Evol., 6: Received: 11 June 2016; Accepted: 26 September 2016 Wallace, B.P., S.S. Heppell, R.L. Lewison, S. Kelez & L.B. Crowder Impacts of fisheries bycatch on loggerhead turtles worldwide inferred from reproductive value analyses. J. Appl. Ecol., 45: Wallace, B.P., C.Y. Kot, A.D. DiMatteo, T. Lee, L.B. Crowder & R.L. Lewison Impacts of fisheries bycatch on marine turtle populations worldwide: toward conservation and research priorities. Ecosphere, 4: Wallace, B.P., R.L. Lewison, S.L. McDonald, R.K. McDonald, C.Y. Kot, S. Kelez, R.K. Bjorkland, E.M. Finkbeiner & L.B. Crowder Global patterns of marine turtle bycatch. Conserv. Lett., 3: Wallace, B.P., A.D. DiMatteo, A.B. Bolten, M.Y. Chaloupka, B.J. Hutchinson, F.A. Abreu-Grobois, J.A. Mortimer et al Global conservation priorities for marine turtles. PLoS ONE, 6: e Wang, J., J. Barkan, S. Fisler, C. Godinez-Reyes & Y. Swimmer Developing illumination of gillnets as a method to reduce sea turtle bycatch. Biol. Lett., 9: doi: /rspb Wilson, S.M., G.D. Raby, N.J. Burnett, S.G. Hinch & S.J. Cooke Looking beyond the mortality of bycatch: sublethal effects of incidental capture on marine animals. Biol. Conserv., 171: Witherington, B., S. Hirama & R. Hardy Young sea turtles of the pelagic Sargassum-dominated drift community: habitat use, population density, and threats. Mar. Ecol. Prog. Ser., 463: Witt, M.J., A. McGowan, J.M. Blumenthal, A.C. Broderick, S. Gore, D. Wheatley, J. White & B.J. Godley Inferring vertical and horizontal movements of juvenile marine turtles from time-depth recorders. Aquat. Biol., 8: Zydelis, R., J. Bellebaum, H. Osterblom, M. Vetemaa, B. Schirmeister, A. Stipniece, M. Dagys, M. van Eerden & S. Garthe Bycatch in gillnet fisheries - an overlooked threat to waterbird populations. Biol. Conserv., 142:

110 Lat. Am. J. Aquat. Res., 45(3): , 2017 Origin of Puerto Rico green turtle aggregations 5061 Sea Turtle Research and Conservation in Latin America Jeffrey Mangel, Jeffrey Seminoff, Bryan Wallace & Ximena Vélez-Suazo (Guest Editors) DOI: /vol45-issue3-fulltext-2 Research Article Genetic composition and origin of juvenile green turtles foraging at Culebra, Puerto Rico, as revealed by mtdna Ana R. Patrício 1,2, Ximena Vélez-Zuazo 3, Robert P. van Dam 4 & Carlos E. Diez 5 1 Centre for Ecology & Conservation, College of Life and Environmental Sciences University of Exeter, Cornwall Campus, Penryn, United Kingdom 2 MARE-Marine and Environmental Sciences Centre, ISPA-Instituto Universitário 3 Center for Conservation and Sustainability, Smithsonian Conservation Biology Institute National Zoological Park, Washington DC, USA 4 Chelonia Inc., Puerto Rico 5 Programa de Especies Protegidas, DRNA-PR, San Juan, Puerto Rico Correspondings authors: Rita Patrício (r.patrício@exeter.ac.uk) ABSTRACT. Marine migratory species encounter a range of threats as they move through coastal and oceanic zones. Understanding the connectivity and dispersal patterns of such species is critical to their effective conservation. Here we analyzed the temporal genetic composition and the most likely origin of juvenile green turtles foraging at Puerto Manglar and Tortuga Bay, Culebra, Puerto Rico, using mitochondrial DNA control region sequences. We identified 17 haplotypes, of which CM-A3 (51.5%), CM-A5 (19.4%) and CM-A1 (13.6%) were the most common. Haplotype (h) and nucleotide (π) diversities were and 0.008, respectively. There was no evidence of significant variation in the genetic composition of these aggregations throughout seven years ( ), suggesting that relative contributions from source populations did not significantly change during this period. Mixed Stock Analysis (MSA), incorporating 14 Atlantic nesting populations as possible sources, indicated four main contributing stocks to the Culebra foraging grounds: Costa Rica (34.9%), Mexico (29.2%), East Central Florida (13.2%), and Suriname (12.0%). The regional pattern of connectivity among Wider Caribbean rookeries and Culebra was further evidenced by a second MSA using Atlantic Regional Management Units (RMUs) as sources, with 94.1% of the mixed stock attributed to this area. This study addresses the information gap on the connectivity of the green turtle in the North Atlantic, and establishes an important baseline that can be used to determine future changes in stock composition. Keywords: Chelonia mydas, connectivity, mixed stock analysis, mtdna, foraging ground. INTRODUCTION Anthropogenic activities in the world s oceans are leading to a rapid decline of species and marine ecosystems health (Halpern et al., 2008). Marine migratory animals, such as whales (Rasmussen et al., 2007), sharks (Bonfil et al., 2005), seabirds (Catry et al., 2011), and sea turtles (Hays & Scott, 2013), are among the most vulnerable due to the range of threats they encounter during their extensive movements (Lascelles et al., 2014). Understanding the temporal and spatial distribution of these species and the connectivity between geographic areas is therefore essential for an integrated management and the conservation of marine ecosystems. Sea turtles carry out some of the greatest migrations across ocean basins (Hays & Scott, 2013), going through habitat changes during their lifecycle (Heppell et al., 2002; Bowen & Karl, 2007). The green turtle Chelonia mydas immediately after hatching at the beach, reaches the ocean and begins an oceanic period coupled with pelagic habitat and epipelagic feeding (Heppel et al., 2002), which may last 3-5 years in the Greater Caribbean (Reich et al., 2007). During this phase, known as the lost years, the distribution and movements of the turtles are poorly known, but they seem to be shaped by a balance between association with oceanic currents (Lahanas et al., 1998; Putman & Naro-Maciel et al., 2013) and directed swimming (Putman & Mansfield, 2015). At cm straightcarapace-length (SCL), juveniles recruit to shallow neritic areas and shift to benthic feeding (Heppell et al., 2002; Bolten, 2003). Neritic zones are used as developmental habitats, where turtles spend several

111 507 2 Latin American Journal of Aquatic Research years foraging until reaching a size or maturity stage that triggers them to migrate (Bjorndal et al., 2005a). Sexually mature individuals move periodically from foraging grounds to nesting beaches and mating areas, often separated by hundreds to thousands of kilometres (Bowen et al., 1992; Bowen & Karl, 2007). The composition of sea turtles at both the nesting beaches and foraging grounds has been assessed with genetic markers. The maternally inherited mitochondrial DNA (mtdna) has been most widely used (Bowen & Karl, 2007; Lee, 2008; Jensen et al., 2013), revealing that near-shore aggregations of immature green turtles are mixed stocks composed by individuals from multiple nesting colonies, whereas nesting beaches form largely isolated populations (Bowen & Karl, 2007). This structure among rookeries results from the natal philopatry exhibited by marine turtles, in which the reproductive females return to the beaches where they hatched to nest (Meylan et al., 1990), and it enables estimating the sources of turtles sampled at foraging grounds, through the use of Bayesian mixed stock analysis (MSA; Pella & Masuda, 2001). MSA iteratively compares the distribution of haplotype frequencies between a foraging ground and each putative rookery of origin, and may incorporate ecological information such as rookery size, improving model estimates. In the Greater Caribbean region, unsustainable harvesting of marine turtles during and prior to the 20 th century led to the decline of several rookeries. Some of these nesting populations have been recovering over the past decades, following protection from human hazards (e.g., Tortuguero in Costa Rica, Archie Carr Refuge in Florida, Aves Island in Venezuela, Chaloupka et al., 2008, García-Cruz et al., 2015), which consequently should be reflected in the recruitment to juvenile aggregations. MSAs have looked into the origin of foraging grounds in Florida (East Central Florida, Hutchinson Island, St. Joseph Bay and Dry Tortugas and Everglade), Texas, the Bahamas, Barbados, and Nicaragua (Bass & Witzell, 2000; Foley et al., 2007; Naro-Maciel et al., 2012; Proietti et al., 2012; Prosdocimi et al., 2012; Anderson et al., 2013; Naro- Maciel et al., 2016). Developmental foraging habitats are further known from several other areas (e.g., Belize, Bonaire, British and American Virgin Islands, Puerto Rico, St Kitts and Nevis), but they remain genetically uncharacterized. Of additional importance is the understanding of the temporal variation on genetic composition of mixed stocks. In the Bahamas, variability in the frequency of mtdna haplotypes of a green turtle juvenile aggregation was detected over a 12-year period and attributed to increased recruitment (Bjorndal & Bolten, 2008). Temporal variability in source contributions has been attributed to very low hatching success at a major source elsewhere (Jensen et al., 2016). Other studies with green turtles in Brazil (Naro-Maciel et al., 2007) and Florida (Naro-Maciel et al., 2016), and with hawksbill turtles in Puerto Rico (Velez-Zuazo et al., 2008), however, found no temporal variation on the genetic composition of juvenile aggregations. In Puerto Rico, Puerto Manglar and Tortuga Bay at Culebra, are recognized as important developmental habitats for juvenile green turtles (Diez et al., 2010; Patrício et al., 2011, 2014). Turtles as small as 23 cm SCL are known to recruit into these coastal bays, where they spend over a decade, departing before the onset of sexual maturity (Patrício et al., 2011, 2014). Here we investigate the genetic composition of these foraging aggregations during a period of seven years and estimate the most likely origins of these stocks using a MSA, including 14 Atlantic nesting populations as potential sources. This study addresses the information gap on juvenile foraging ground composition in the Caribbean and sets a baseline for the Puerto Rico aggregations, allowing comparisons with future monitoring. MATERIALS AND METHODS Study site and sampling Puerto Manglar (18.30 N, W) and Tortuga Bay (18.32 N, W) are two foraging grounds for immature green turtles, located at Culebra and Culebrita Islands, respectively, within the boundaries of a critical habitat for the green turtle, designated by the Endangered Species Act (NMFS-NOAA, 1998) in Puerto Rico (see Fig. 1 in Patrício et al., 2011). The Department of Natural and Environmental Resources of Puerto Rico (DNER-PR) has conducted a capturemark-recapture program at these sites, since From 2000 to 2006 we collected samples from 103 green turtles foraging in these bays [2000 (18), 2001 (16), 2002 (2), 2003 (17), 2004 (13), 2005 (25), 2006 (12)]. Turtles were captured with an entanglement net (200 m long, 5 m deep, nylon twine, 25 cm stretch mesh size) deployed for ~1 h sets at <5 m depth, with the help of a motor boat. Swimmers snorkelled continually along the net to locate and disentangle trapped turtles. Turtles were kept in the shade and covered with wet towels while captive and until processing. Handling time averaged 15 min per individual, after which turtles were released close to their capture location. Tissue samples were collected from the shoulder area using a disposable biopsy punch (4-6 mm diameter, Acuderm ). Samples were preserved in 95% ethanol or saltsaturated 20% DMSO-20% EDTA and stored at room temperature. SCL of sampled individuals was measured

112 Origin of Puerto Rico green turtle aggregations with Haglof tree calipers to the nearest 0.1 cm. All turtles were applied a unique ID tag in both front flippers to avoid misidentification and sample duplication. Sequencing and haplotype assignment DNA was extracted using the DNeasy Blood & Tissue kit (Qiagen) following manufacturer s instructions, and eluted in a final volume of 50 µl per sample. DNA concentrations were quantified with a spectrophotometer (NanoDrop ND-3300) and a 735 bp fragment of the mtdna control region was amplified by Polymerase Chain Reaction (PCR) with primers LTEi9 and H950 (Abreu-Grobois et al., 2006). Amplifications were performed in a total volume of 10 μl, with 1 μl genomic DNA at a concentration of ~10 ng μl -1, 4.0 µl of Qiagen Taq Master Mix, 0.5 µm of each primer at 10 µm and 2.0 µl MilliQ water. PCR started with an initial denaturing step of 5 min at 94ºC, followed by 30 cycles of 30 s at 94ºC, 30 s at 52ºC, and 1 min at 72ºC, with a final hold at 72ºC for 5 min. All PCR reactions included positive and negative controls. PCR products were purified with ExoSAP-IT (Affymetrix) and sequenced in both forward and reverse directions using a BigDye Terminator v.3.1 (Bioanalytical Instruments) and the automated sequencer station ABI 3130xl (Applied Biosystems) at the Sequencing and Genotyping Facility of the University of Puerto Rico, Río Piedras. Sequences were assembled and aligned by eye using Sequencher 4.5 (Gene Codes). To identify unique haplotypes and estimate absolute haplotype frequencies we used DNAspv4.10 (Rozas et al., 2003). Haplotypes were identified using the Basic Local Alignment Search Tool (BLAST) from the National Center for Biotechnology Information ( and named following the standardized nomenclature of the Archie Carr Center for Sea Turtle Research. Diversity estimates Haplotype (h) and nucleotide diversities (π), pairwise genetic distances among groups (F ST), and exact tests of differentiations (Raymond & Rousset, 1995) were estimated using Arlequinv 3.1 (Excoffier & Lischer, 2010) for two sets of groups: 1) sample years at our study sites (n = 6), and 2) Atlantic green turtle foraging grounds (n = 18, Fig. 1). A false discovery rate (FDR) correction, following Narum (2006), was applied to calculate the most fitting threshold for the P-value significance, considering the number of comparisons involved in the analysis, under an expected original threshold of P < The sample size for 2002 was too small (i.e., n = 2) for robust statistic comparisons among years, so it was excluded from the temporal analysis. We truncated the DNA fragments to 491 bp length, the fragment historically explored and for which most genetic information is currently available, to compare diversity estimates with other foraging aggregations. Geographic variability and genetic diversity To investigate how mithocondrial control region diversity is partitioned among foraging aggregations, we conducted a spatial analysis of molecular variance (SAMOVA, Dupanloup et al., 2002), incorporating geographic positions obtained through Google Earth, and using 100 simulated annealing processes. This analysis defines geographic groups that are maximally differentiated (rather than defining a priori groupings). The F CT statistic from AMOVA (calculated a posteriori) was then compared among different values of groups (K), ranging from 2 to 18 foraging grounds, to assess the most likely number of K, corresponding to the highest F CT (Dupanloup et al., 2002). Additionally, genetic distances between foraging sites were included in a principal coordinate analysis (PCoA) using the package Genalex (Peakall & Smouse, 2012), to plot variability in a two-dimensional space. Mixed stock analysis (MSA) The most likely origin of the studied aggregations was estimated through a one-to-many MSA using BAYES (Pella & Masuda, 2001). We compiled the available genetic information from green turtle Atlantic nesting populations and used it as baseline information for the MSA (See Fig. 1 for sites included in this study, site abbreviations, and literature sources, and Table 1 for genetic composition). Rookery size, defined as the number of nesting females per rookery (Seminoff et al., 2015), was used to establish weighted priors. Previous studies have shown that there is significant structure among most of the genetically characterized Atlantic green turtle rookeries (Bolker et al., 2007, Shamblin et al., 2012, 2014), supporting the applicability of a MSA. There is however a lack of genetic differentiation at the mtdna control region between some individual rookeries (e.g., Suriname and Aves Island, Naro- Maciel et al., 2016), so we also ran a MSA pooling the individual rookeries into Regional Management Units (RMUs, Wallace et al., 2010), which group multiple nesting populations based on their genetic similarities, for conservation management. Following Naro-Maciel et al. (2016), the RMUs were defined as: 1) Northwest Atlantic - EcFL, SFL, MEX, CUB, CR; 2) Central Atlantic - BUC, AV, SUR; and 3) South and East Atlantic - RC/FN, ASC, TRI, GB, BIO, STP. Four independent chains with different starting points were run for 30,000 iterations, with a burn-in of 15,000 steps. We used the Gelman-Rubin diagnostic to assess conver-

113 509 4 Latin American Journal of Aquatic Research Figure 1. Atlantic green turtle (Chelonia mydas) foraging grounds (n = 18, dark triangles and black star for study site) and nesting populations (n =14, gray circles) included in this study, with respect to major ocean currents: GfC: Gulf Current, NEC: North Equatorial Current, SEC: South Equatorial Current, BrC: Brazil Current, GC: Guinea Current, BgC: Benguela Current. Nesting populations: EcFL and SFL: Florida, USA (Shamblin et al., 2014); CUB: southwest Cuba (Ruiz-Urquiola et al., 2010); MEX: Quintana Roo, Mexico (Encalada et al., 1996); CR: Tortuguero, Costa Rica (Bjorndal et al., 2005b; Encalada et al., 1996); SUR: Matapica and Galibi, Suriname (Encalada et al., 1996; Shamblin et al., 2012); AV: Aves Island (Lahanas et al., 1998, 1994; Shamblin et al., 2012), Venezuela; BUC: Buck Island (Shamblin et al., 2012); RC/FN: Rocas Atoll and Fernando Noronha (Bjorndal et al., 2006; Encalada et al., 1996), Brazil; ASC: Ascension Island (Encalada et al., 1996; Formia et al., 2007); TRI: Trindade Island, Brazil (Bjorndal et al., 2006); GB: Poilão, Guinea-Bissau (Patrício et al., 2017); BIO: Bioko Island, Equatorial Guinea (Formia et al., 2006); STP: Sao Tome and Principe (Formia et al., 2006). Foraging grounds: NC: North Carolina (Bass et al., 2006), HI: Hutchinson Island, Florida (Bass & Witzell, 2000), DT+EP: Dry Tortugas + Everglades Park, Florida (Naro-Maciel et al., 2016), SJ: St. Joseph Bay, Florida (Foley et al., 2007), TEX: Texas (Anderson et al., 2013), USA; BHM: Bahamas (Lahanas et al., 1998), CUL: Culebra, Puerto Rico (this study), BRB: Barbados (Luke et al., 2004), ALF: Almofala, Brazil (Naro-Maciel et al., 2007), RC: Rocas Atoll, Brazil (Naro-Maciel et al., 2012), FN: Fernando Noronha, Brazil (Naro-Maciel et al., 2012), BA: Bahia, Brazil (Naro-Maciel et al., 2012), ES: Espirito Santo, Brazil (Naro-Maciel et al., 2012), UB: Ubatuba, Brazil (Naro-Maciel et al., 2007), AI: Arvoredo Island, Brazil (Proietti et al., 2012), CB: Cassino Beach, Brazil (Proietti et al., 2012), BuA, Buenos Aires, Argentina (Prosdocimi et al., 2012), CV: Cape Verde (Monzón-Argüello et al., 2010). gence of the chains to the posterior distribution, assuming that there was no evidence of non-convergence at values <1.2 (Pella & Masuda, 2001). RESULTS At Puerto Manglar (n = 60) mean SCL was 47.4 ± 8.8 cm (mean ± SD, range: cm, Fig. 2), and at Tortuga Bay (n = 43) it was 44.7 ± 11.0 cm (mean ± SD, range: cm, Fig. 2). There was no significant difference in SCL distribution between the two groups (t 101= , P = ). We detected 17 polymorphic sites at the 735 bp mtdna fragment, one transversion, 16 transitions and one insertion (position 617), defining 17 haplotypes, 13 of them previously described (Supplemental Table 1). After truncating the sequences the total number of haplotypes dropped to 10 (Table 1). In both aggregations the haplotype CM-A3 was dominant (PM: 43%; TB: 63%), followed by haplotypes CM-A5 (PM: 22%; TB: 16%), CM-A1 (PM: 15%; TB: 12%), and CM-A8 (PM: 7%; TB: 5%). We also identified rare haplotypes with frequencies of 1-3%: CM-A2, CM-A16, CM-A17, CM-A18, CM-A27 and an orphan haplotype, CM-A26,

114 Origin of Puerto Rico green turtle aggregations 5105 Figure 2. Straight-carapace-length (SCL, cm) distributions for immature green turtles (Chelonia mydas) captured between 2000 and 2006 at Puerto Manglar (n = 60, dark gray) and Tortuga Bay (n = 43, light gray) foraging grounds, Puerto Rico. not yet reported in a nesting population, emphasizing that some stocks still lack genetic studies or have not yet been adequately sampled. A randomized Chi-square (χ 2 = 6.05, P = 0.89) and an exact test of differentiation (P = 0.88) indicated no significant genetic structure between the two aggregations, so these were pooled for further analyses, referred to henceforth as the Culebra foraging ground. We found no significant temporal variation in the haplotype composition of the Culebra foraging ground among sampling years over seven-year period (Table 2). There seems to be an increase in haplotype CM-A5 with time, however (Supplemental Fig. 1). The haplotype (h) and nucleotide (π) diversities at Culebra foraging grounds were comparable to those of Atlantic green turtle aggregations (Table 3). Culebra was significantly different from all other foraging sites except the Bahamas (Table 4). The SAMOVA suggested that the 18 foraging aggregations were partitioned into two or three main groups, with F CT = for K = 2, and F CT = for K = 3. The estimates of F CT decreased faster as K increased, after K = 3 (Supplemental Fig. 2). Because the percentage variation between populations within groups increased from 1.5% for K = 3 to 2.5% for K = 2 (Supplemental Fig. 2) by including Barbados with the south Atlantic foraging grounds, we consider that K = 3 is a better grouping. This was consistent with the PCoA. The SAMOVA (K = 3) and the PCoA separated foraging areas geographically, highlighting three groups: 1) all South American foraging grounds and Cape Verde, 2) Northwest Atlantic foraging grounds, and 3) Barbados (Fig. 3). Using this a priori grouping in the AMOVA, a highly significant structure was observed among the groups (F ST = , P < 0.001). The MSA using RMUs as potential sources estimated that 77.9% of the green turtles foraging at Culebra recruit from the Northwest Atlantic RMU (95% CI: %), 16.2% from the Central Atlantic RMU (95% CI: %) and 5.9% from the South and East Atlantic RMU (95% CI: %) (Fig. 4a). The MSA using individual nesting populations estimated that 34.9% of the Culebra turtles originated from Tortuguero, Costa Rica (95% CI: %); 29.2% from Mexico (95% CI: %); 13% from East Central Florida (95% CI: %); 12% from Suriname (95% CI: %), 3% from South Florida (95% CI: %), 3% from Cuba (95% CI : %) and 3.5 % from Guinea-Bissau (95% CI : 0-9.9%) (Fig. 4b, and Supplemental Table 2). DISCUSSION Understanding the links between developmental habitats and the source populations of migratory species is critical to assess threats at their different life stages, and develop effective conservation policies. Here we analyzed the genetic composition of two important developmental aggregations for green turtles in the Caribbean (Culebra, Puerto Rico), over a period of seven years, and predicted the most likely connectivity of these aggregations to Atlantic nesting populations, using mtdna control region sequences and a MSA, improving our understanding on the movements of green turtles in the North Atlantic. Genetic structure among foraging aggregations The similarity in the genetic composition of Tortuga Bay and Puerto Manglar suggests that there is no differential recruitment between the two foraging grounds, which was expected given that these are only 2 km apart. There was also no significant genetic differentiation between Culebra and the Bahamas. This foraging ground also has major contributions from Northwest Atlantic rookeries, but not from Central Atlantic rookeries (Putman & Naro-Maciel, 2013), contrary to what we estimated for Culebra. At greater distances however, there is structure among foraging grounds, and we found two major groups, represented by the northwest Atlantic and the south and east Atlantic. The Barbados mixed stock was distinct from both groups, as it receives equal contributions from both north and south Atlantic nesting populations (Luke et al., 2004), potentially due to its position relative to the coalescence of the North Equatorial and South Equatorial currents (Luke et al., 2004).

115 6511 Latin American Journal of Aquatic Research Table 1. mtdna haplotype frequencies at the study site and at 14 Atlantic green turtle nesting populations, with total number of samples and haplotypes per area, and total number of nesting females at rookeries. See Fig. 1 for site abbreviations. Table 2. Sample size (n), total number of haplotypes (hap), and haplotype (h) and nucleotide diversities (π) per year, at Culebra foraging ground (Puerto Rico), for immature green turtles, throughout a seven year period, and pairwise comparisons among sampling years: exact test P-values (P > 0.05) in the above diagonal and F ST values in the below diagonal. Year Year n Hap h π ± ± ± ± ± ± ± ± ± ± ± ±

116 Origin of Puerto Rico green turtle aggregations 5127 Table 3. Sample size (n), haplotype number (hap) and haplotype (h) and nucleotide (π) diversity estimates ± SD of Atlantic green turtle (Chelonia mydas) foraging grounds (n = 18), using a fragment of 491 bp of the control region of the mitochondrial DNA as a marker. The study population is represented in bold. Juvenile foraging grounds n hap h (π) Culebra, Puerto Rico a ± ± North Carolina, USA b ± ± Hutchinson island, FL, USA c ± ± St. Joseph, FL, USA d ± ± Dry Tortugas and Everglades, FL, USA e ± ± Texas, USA f ± ± Bahamas g ± ± Barbados h ± ± Ubatuba, Brazil i ± ± Almofala, Brazil i ± ± Rocas, Brazil j ± ± Fernando Noronha, Brazil j ± ± Bahia, Brazil j ± ± Espirito Santo, Brazil j ± ± Arvoredo Island, Brazil k ± ± Cassino Beach, Brazil k ± ± Buenos Aires, Argentina l ± ± Cape Verde m ± ± a This study, b Bass et al. (2006), c Bass & Witzell (2000), d Foley et al. (2007), e Naro-Maciel et al. (2016), f Anderson et al. (2013), g Lahanas et al. (1998), h Luke et al. (2004), i Naro- Maciel et al. (2007), j Naro-Maciel et al. (2012), k Proietti et al. (2012), l Prosdocimi et al. (2012), m Monzón-Argüello et al. (2010). Regional connectivity among Culebra and Wider Caribbean populations The MSAs indicated that the Culebra aggregations originate from multiple rookeries within the Wider Caribbean region. This strong regional connectivity agrees with the closest to home hypothesis, where immature turtles tend to move to and settle in foraging grounds closest to their natal beach after recruiting to neritic habitats (Bowen et al., 2004; Bolker et al., 2007). Similar patterns of regionalized recruitment have already been observed in Atlantic green turtles (Bass et al., 2006; Bolker et al., 2007; Naro-Maciel et al., 2012) and in other marine turtle species (Bowen & Karl, 2007). However, this pattern may be influenced by the geographic position of foraging areas and nesting beaches relative to major oceanic currents (Luke et al., 2004). The connectivity within the Wider Caribbean estimated in the MSA is supported by several tag returns from foraging and nesting adult turtles (Fig. 5). Most of these tags were recovered at Nicaragua (n = 8), at foraging grounds long known to be used by the nesting population of Tortuguero (i.e., Miskito Cays, Carr & Ogren, 1960; Bjorndal, 1980), but also at Venezuela (n = 1), Colombia (n = 1), and Florida (n = 1). In the latter, a turtle first tagged as a juvenile at Tortuga Bay in 1997, was found nesting in 2014 (Bagley, pers. comm.), further confirming this connectivity. Interestingly, there was also a tag return in 2006 from the north of Brazil (State of Ceará, >3500 km, Lima et al., 2008), so more distant links can exist. Temporal variability Throughout the seven years of this study we could not detect a significant variation on the frequency of the mtdna haplotypes at the Culebra aggregation, which could suggest that there were no changes in the overall contributions from the major source populations (i.e., Costa Rica, Mexico, East Central Florida and Suriname). These results are not conclusive however, because our annual sample size may have been too small to detect significant change. We did observed a slight increase in the frequency of haplotype CM-A5, which could potentially be associated with the positive trend in population growth at rookeries where this is the dominant haplotype, i.e., Suriname and Aves Island (García-Cruz et al., 2015; Turny, pers. comm.). At Puerto Manglar, a positive trend on abundance with a mean annual increase of 10.9% was observed over the course of 15 years ( , Patrício et al., 2014), more accentuated from 2006, owing to increased recruitment. This reflects the positive trend in the source populations (Chaloupka et al., 2008), which may

117 Table 4. Pairwise exact test P-values (above diagonal) and pairwise FST values (below diagonal) among 18 Atlantic green turtle foraging aggregations, based on ~490 bp sequences of the control region of the mtdna. The study site is in bold, and abbreviations follow those in Figure 1. Asterisks indicate statistically significant comparisons (*P < 0.05, **P < 0.01, ***P < 0.001). i) prior to corrections, in the low diagonal, ii) after FDR correction, in the above diagonal. Non-significant values, after FDR correction, are marked in bold (for a P < 0.05, FDR = 0.009; P < 0.01, FDR = 0.002, for a P < 0.015, FDR = 0.000; Narum, 2006) Latin American Journal of Aquatic Research

118 Origin of Puerto Rico green turtle aggregations 5149 Figure 3. PCoA of 18 Atlantic green turtle (Chelonia mydas) foraging grounds using F ST genetic distances inferred from control region mitochondrial DNA haplotypes. The percentage of the variability explained by each coordinate is shown in brackets. Foraging grounds: NC: North Carolina, USA; EcFL: East Central Florida, USA; BHM: Bahamas; CUL: Culebra, Puerto Rico; BRB: Barbados; ALF: Almofala, RC: Rocas Atoll, FN: Fernando Noronha, BA: Bahia, ES: Espirito Santo, UB: Ubatuba, Brazil, AI: Arvoredo Island, and CB: Cassino Beach, Brazil; BuA, Buenos Aires, Argentina; CV: Cape Verde. lead to changes in the relative contributions from Atlantic rookeries to the Culebra aggregation, particularly if they are not all recover at the same pace. Impact for nesting and breeding recruitment Both Tortuga Bay and Puerto Manglar foraging grounds are recruitment sites for post-pelagic individuals, where minimum sizes found are 22.8 and 29.8 cm SCL, respectively (Diez et al., 2010). A longterm capture-mark-recapture (CMR) program has revealed that immature turtles remain in these bays for several years (ca. 10 to 17 years, Patrício et al., 2014), and that larger immature turtles (>65 cm SCL) permanently emigrate, potentially to subadult foraging sites closer to their breeding grounds (Patrício et al., 2011). As turtles spend such a long period of their early life at these developmental sites, mortality there can impact the multiple rookeries to which they are linked. Juvenile green turtles at Culebra s aggregations have high survival probability (0.83; CI 95% = , Patrício et al., 2011), comparable to estimates found for juvenile mixed stocks in areas virtually free of human impacts (Bjorndal et al., 2003; Chaloupka & Limpus, 2005). Occasional stranding s of immature green turtles with evidence of boat collisions or of fibropapilloma tumors have occurred; otherwise no direct hazards for green turtles are known at the study sites. Habitat degradation, however, may have a negative impact, as both coastal urban development and recreational boats continue to increase in the area. Fibropapillomatosis (FP) is endemic to Culebra s aggregations and in 2003 disease prevalence reached 75% at the most affected Figure 4. Mean proportion and 95% confidence intervals (error bars) of green turtles (Chelonia mydas) foraging at Culebra, Puerto Rico, attributed to a) three Atlantic Regional Management Units (RMUs): Northwest Atlantic (CR, MEX, EcFL, CUB, SFL), Central Atlantic (SUR, AV, BUC) and South and East Atlantic (GB, ASC, TRI, RC/FN, BIO, STP), and b) each of 14 Atlantic nesting populations, estimated by a mixed-stock-analysis. Nesting populations: CR: Tortuguero, Costa Rica; MEX: Quintana Roo, Mexico; EcFL: East Central Florida, USA; SUR: Matapica and Galibi, Suriname; CUB: southwest Cuba; SFL: Florida, USA; GB: Poilão, Guinea-Bissau; ASC: Ascension Island; AV: Aves Island, Venezuela; BUC: Buck Island; TRI: Trindade Island, Brazil; RC/FN: Rocas Atoll and Fernando Noronha, Brazil; BIO: Bioko Island, Equatorial Guinea; STP: Sao Tome and Principe.

119 Latin American Journal of Aquatic Research Figure 5. Map showing green turtle (Chelonia mydas) rookeries in the wider Caribbean region that contribute to the Culebra (Puerto Rico) foraging aggregations (dashed arrows, contributions 3%), and locations of tag returns from turtles resident at Culebra (solid arrows). Mean percentage contributions by the different nesting populations, as estimated through Bayesian mixed-stock-analysis (MSA) are indicated in bold, as well as number of tag returns (in parenthesis). Note: the pathways shown are not indicative of migratory corridors. EcFL: East Central Florida, SFL: South Florida, USA; CUB: southwest Cuba; MEX: Quintana Roo, Mexico; CR: Tortuguero, Costa Rica; AV: Aves Island, Venezuela; SUR: Matapica and Galibi, Suriname; RC/FN: Rocas Atol and Fernando Noronha, Brazil; and CUL: Culebra foraging aggregation (Map created using foraging site (i.e., Puerto Manglar, Diez et al., 2010). It was shown, however, that FP did not affect survival rates (Patrício et al., 2011), and that individual recovery was likely (Patrício et al., 2016). CONCLUSIONS Green turtles, once abundant in the Caribbean, faced major population decline of possibly 99%, since the arrival of European (Jackson, 1997). Thanks to conservation efforts of the past decades, major green turtle populations worldwide are now rapidly recovering (Chaloupka et al., 2008). This has been particularly noticeable in the wider Caribbean region, where long-term data allows for robust abundance trend estimates of major populations, e.g., Costa Rica, Florida, and Mexico (Seminoff et al., 2015). A positive abundance trend was also detected at Puerto Manglar, as mentioned earlier (Patrício et al., 2014). Turtles are however still harvested in some regions in the wider Caribbean (Humber et al., 2014). Most notably at Nicaragua there is a large legal artisanal fishery of green turtles aimed for local consumption (Humber et al., 2014; Lagueux et al., 2014), but additional commercialization of turtle meat continues to occur due to lack of law enforcement, and this fishery was estimated to take ca turtles per year, and considered to be unsustainable (Lagueux et al., 2014). The majority of tag returns from the Culebra aggregation came from Nicaragua, which poses a conservation paradox if efforts are conducted to protect these juvenile aggregations but unsustainable harvesting at later stages of their life occurs elsewhere. Our study emphasizes, therefore, the widely recognized need for a comprehensive regional conservation strategy (Wallace et al., 2011). ACKNOWLEDGMENTS Samples analysed in this study were obtained with the help of numerous field assistants and volunteers. We