Effects of hook and bait on bycatch and target catches in a southern Atlantic swordfish longline fishery

FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery i ACKNOWLEDGMENTS The present work was undertaken at the Portuguese Sea and Atmosphere Institute (IPMA, Portugal), in cooperation with Faculty of Sciences of the University of Porto. This study was carried out within the scope of the project “Redução das capturas acessórias na pescaria de palangre de superfície” - SELECT-PAL (Programa PROMAR Proj. 31-03-05-FEP-1). During the study period the present candidate hold a grant as research technician of the Data Collection Project co-funded by the European Commission under the Data Collection Framework (DCF) of the Common Fisheries Policy (CFP). This manuscript was a cooperative work where many different persons have contributed with diverse competences and the importance of each of them was extremely valuable. When you work in a project for so long there are far too many people to thank. There are some people that I honestly could not have completed this dissertation without, to whom I would like to express my sincere acknowledgements: - First, of course, is my supervisor, Dr. Miguel Neves dos Santos, for the opportunity, constant support, advice, and guidance to conduct this study. Also express my sincere acknowledgements for trusting me. - To Prof. Dr. Paulo Talhadas dos Santos for his coordination, advice, suggestions, guidance, and continual support throughout the course of this project. - To Pescarade, in particular to Mr. António Teixeira for her commitment in facilitating the access to the pelagic longliner vessel. Special gratitude to the crews of fishing vessel “Alma Lusa” (PN-1269-N) for their cooperation and commitment during the course of this study. Thanks a lot for the great times spent onboard. - Special recognition to all scientific fisheries observers, Israel Ribeiro, João Regala and Sérgio Góis for their assistance on data collection. I know, how good times we spent at sea, but how hard it can be. - To Rui Coelho for the training in R software environment for statistical computing and graphics. - To many friends and colleagues that work or worked in IPMA of Olhão and contributed in some way to keep looking forward, such as Marco Cerqueira, Alexandra Garcia, Israel Ribeiro, Sérgio Abreu, Pedro Lino, Joana Fernandez, Dr. Miguel Gaspar, Paulo Vasconcelos, João Cúrdia, Susana Carvalho, Fábio Pereira. - To José Luís Sofia, Tibério Simões, Fernando Martins, Luz Abreu, Maria de Lurdes and Lina Oliveira for the pleasant working environment and fabulous lunching times. - To my dear colleagues of work in IPMA of Porto, Dra. Fernanda Castilho, Emanuel Pombal, Paulo Castro, Marta Gonçalves, Mónica Felício, Diana Feijó, Alberto Rocha, FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery ii Georgina Correia, Filomena Pombal, Adelaide Resende, Raquel Milhazes e Maria do Céu Viegas for the pleasant working environment and for remember me, how good it sounds the northern accent. - To my parents and brother, without their constant kindness and support it would never had been made; - Finally, thanks to Filipa, for your support, encouragement, patience (a lot) and unconditional love. You are the main reason to keep carrying on. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery iii ABSTRACT Bycatch is a major problem in commercial fisheries. Despite pelagic longline being more selective than other fishing gears (e.g. trawl and driftnets) a wide variety of sea “megafauna” is caught, namely sea turtles and sharks. The present dissertation examined the effects of different hook styles and bait types combinations on the: i) sea turtle bycatch composition, bycatch rates and mortality and ii) catches of target, bycatch and discarded fishes from a Portuguese commercial longline fishery targeting swordfish in the South Atlantic Ocean. A total of 310 experimental longline sets were made between October 2008 and February 2012. Three different hook styles and two bait types were tested: the traditional J hook (9/0) 10º offset was compared to two 17/0 circle hooks (a non-offset and a 10º offset) and mackerel (Scomber spp.) bait was compared to squid (Illex spp.). Two species of sea turtles were caught, the leatherback (Dermochelys coriacea) and the loggerhead (Caretta caretta) with the latter comprising the majority of the catches. The highest mean bycatch per unit effort (BPUE) values for both species combined (1.693/1000 hooks) and for the individual species (1.505/1000 hooks for loggerheads) occurred with J-style hooks baited with squid. Hooking location was species-specific, with most loggerheads hooked by the mouth, while leatherbacks were mostly hooked externally by the flippers. Overall, 65% of all sea turtles were released alive (85% for leatherbacks compared to 63% for loggerheads). For the retained catch the effect of the different hook-bait combinations was species-specific, with bait being far more important than hook style. For the target species, swordfish (Xiphias gladius), the catch per unit effort (CPUE) was higher with J hooks baited with squid. However, for the elasmobranchs bycatch, particularly blue shark (Prionace glauca), the opposite effect was observed (higher catches with circle hooks baited with mackerel). For the discarded species, the at-haulback mortality was also species-specific, with the proportions of alive versus dead specimens of the protected bigeye tresher shark (Alopias superciliosus) varying significantly by hook style. Significant reduction of accidental sea turtle catches on the swordfish longline fisheries can be achieved by changing from J hooks to circle hooks, especially if baited with mackerel. However, such gear changes results in lower catch rates of swordfish and increases the catches of elasmobranchs. Therefore, from the fisheries management point of view, it is essential to assess the consequences of such gear modifications in a wider scale, prior to the implementation of the mandatory use of circle hooks on this fishery. Key-words: pelagic longline; circle hooks; bait types; bycatch; sea turtles; swordfish FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery iv RESUMO A captura acidental é um problema grave da pesca comercial. Apesar do palangre pelágico ser mais seletivo do que muitas outras artes de pesca (ex. redes de arrasto e redes de deriva) uma grande variedade de “megafauna” marinha é capturada, nomeadamente tartarugas marinhas e tubarões. A presente dissertação examinou os efeitos de diferentes combinações de anzóis e iscos na i) composição, taxas de captura acessórias e mortalidade de tartarugas marinhas e ii) capturas-alvo, capturas acessórias e rejeições de peixes de uma pescaria comercial Portuguesa de palangre pelágico dirigido ao espadarte no Atlântico Sul. Um total de 310 lances experimentais foram efetuados entre Outubro de 2008 e Fevereiro de 2012. Três tipos de anzóis e dois tipos de isco foram testados: o tradicional anzol J (9/0) 10º de inclinação foi comparado a dois anzóis 17/0 circulares (um sem inclinação e um com 10º de inclinação) e o isco cavala (Scomber spp.) foi comparada com pota (Illex spp.). Duas espécies de tartarugas marinhas foram capturadas, a tartaruga-decouro (Dermochelys coriacea) e a tartaruga-comum (Caretta caretta), com a última a compreender a maior parte das capturas. Os valores médios das capturas acessórias por unidade de esforço (BPUE) mais elevados para ambas as espécies combinadas (1.693/1000 anzóis) e para as espécies individuais (1.505/1000 anzóis para a tartarugacomum) ocorreram para os anzóis J iscados com pota. A localização do anzol foi espécieespecífico, com a maioria das tartarugas-comum a serem capturadas pela boca, enquanto as tartarugas-de-couro foram na sua maioria capturadas externamente pelas barbatanas. No total, 65% das tartarugas marinhas foram rejeitadas vivas (85% para as tartarugas-decouro em comparação com 63% para a tartaruga-comum). Para as capturas retidas a bordo o efeito das diferentes combinações anzol-isca foi espécie-específico, com a isca a ser bem mais importante do que o tipo de anzol. Para a espécie-alvo, espadarte (Xiphias gladius) a captura por unidade de esforço (CPUE) foi maior com os anzóis J iscados com pota. Contudo, para as capturas acessórias de elasmobrânquios, particularmente o tubarão-azul (Prionace glauca) foi observado o efeito oposto (capturas mais elevadas com anzóis circulares iscados com cavala). Para as espécies rejeitadas a mortalidade aquando da alagem também foi espécie-específico com as proporções de espécimes vivos versus mortos do protegido tubarão raposo-olhudo (Alopias superciliosus) a variar significativamente com o tipo de anzol. Reduções significativas nas capturas acidentais de tartarugas marinhas podem ser alcançados, mudando de anzóis J para anzóis circulares, especialmente se iscados com cavala. Contudo tal mudança de aparelho resulta em menores capturas de espadarte e um aumento das capturas de elasmobrânquios. Assim, do ponto de vista de gestão das pescas, é essencial analisar as consequências da alteração do FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery v aparelho de pesca numa maior escala, antes da implementação obrigatória dos anzóis circulares na pescaria de palangre pelágico dirigido ao espadarte no Atlântico sul. Palavras-chave: palangre pelágico; anzóis circulares; tipos de isca; capturas acessórias; tartarugas-marinhas; espadarte FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery xii Table 4.2 - Coefficients for the swordfish (A) and blue shark (B) CPUE Tweedie GLM with the respective odds-ratios, considering the covariates hook style and bait type. The parameter estimation of the model, the standard errors (SE), the Wald Statistic (Wald) and the respective p-values are presented. For the odds-ratios the point estimate with the lower and upper 95% confidence intervals (CI) are listed. Table 4.3 - Mean (standard deviation) size (LJFL and FL, in cm) for the species: swordfish (Xiphias gladius), bigeye (Thunnus obesus), yellowfin (Thunnus albacares), albacore (Thunnus alalunga), blue marlin (Makaira nigricans), blue shark (Prionace glauca) and shortfin mako (Isurus oxyrinchus). P-values refer to the Mann-Whitney tests comparing sizes with different baits and the Kruskal-Wallis tests comparing sizes with different hooks. J = 10º offset 9/0 hook; G = non-offset 17/0 circle hook; GT = 10° offset 17/0 circle hook. p<0.05 are in bold. Table 4.4 - Mean (standard deviation) catch per unit effort (CPUE, n/1000 hooks) of discarded species for the various hook-bait combinations. P-values refer to the MannWhitney tests comparing baits and the Kruskal-Wallis tests comparing hooks. J = 10º offset 9/0 hook; G = non-offset 17/0 circle hook; GT = 10° offset 17/0 circle hook. p<0.05 are in bold. NT = not tested. Table 4.5 – Percentage of alive versus dead specimens at-haulback for the most frequently discarded species: lancetfish (A. ferox), bigeye thresher shark (A. superciliosus), pelagic stingray (P. violacea), and crocodile shark (P. kamoharai). The statistical comparisons refer to Chi-square tests for contingency tables. J = 10º offset 9/0 hook; G = non-offset 17/0 circle hook; GT = 10° offset 17/0 circle hook. p<0.05 are in bold. NT= not tested FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 1 GENERAL INTRODUCTION AND OBJECTIVES Sea fisheries, a primary source of protein for billions of people globally (FAO, 2009) are the major anthropogenic influence on sea systems worldwide, affecting sea animal populations and ecosystem function (Pauly et al., 2005). The pelagic longline fishery is the world’s most widespread harvest activity, with approximately 5 million baited hooks set each day on 100.000 miles of line throughout the world’s oceans (Crowder and Myers, 2001). It is a passive fishing method based on fish attraction by means of bait and can be used from small-scale artisanal fishing boats to modern mechanized vessels. Pelagic longline fishing evolved in Japan during the 19th and early 20th centuries. Sailboats equipped with hemp longlines would venture as far as 30 nautical miles (nm) offshore from Japan in search of tuna and billfish. By 1912 there were over 100 registered sailboat tuna longliners in Japan and in 1920 the first diesel powered steel longline vessel appear (Beverly et al., 2003) expanding the fleet to the rest of the Pacific Ocean in the following decades. Global expansion of longline fisheries began in the 1950's and 1960's with Japanese tuna fleet spreading throughout the Atlantic Ocean and Mediterranean Sea (Watson and Kerstetter, 2006). The swordfish incidental catches of longlines targeting tuna were high, reason that attracted interest for it use and has enabled the development of a directed fishery for this species by the fleets of Canada, United States and Spain in the early 1960's. During the 1980's and 1990's, longliners increased swordfish catches to high levels, driven by the increase of fishing boats and fishing effort. Geographical expansion and improvements in fishing efficiency and targeting (e.g. introduction of monofilament gear, better freezing capacity and electronic aids to navigation and fish finding) contributed to the increase catch rates (Ward et al., 2000). Nowadays longlines are used to target many pelagic fish species throughout the world’s oceans and are responsible for most of the world’s swordfish (Xiphius gladius) catches and a large proportion of global tuna (Thunnus spp.) catches (Lewison et al., 2004). Human impacts on the world’s oceans are extensive and varied, warranting urgent and comprehensive management of sea resources in many places of the world (Halpern et al., 2008). Despite the widespread nature of longline fishing, only from the 1990's there has been global concern about the bycatch of sea turtles, sharks, birds and sea mammals in fishing operations (Lewison et al., 2004). The incidental mortality of these species has been widely held responsible for the declining populations and threatened conservation status of several species (Lewison et al., 2004). Although concerns about impacts of pelagic FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 2 longlining on both target and non-target species have already led to some changes in management and policy, there are lack of comprehensive and objective assessment of the problem. The knowledge of the status of biodiversity in the high seas is minimal compared with that on land and in coastal waters. Because pelagic longlines have the potential to catch a large number of both target and non-target organisms, some of which are already overfished and/or protected under international treaty’s, it is important to evaluate the impact of mitigation procedures on populations that are potentially vulnerable (Crowder and Myers, 2001). A variety of measures have been developed to reduce the bycatch mortality of sea turtles, sharks and seabirds in pelagic longline fisheries (Gilman et al., 2006). Techniques to address bycatch include: time-area closures; voluntary measures (e.g. moving after a by-catch); bycatch quotas; and more recently circle hooks and bait types intended to reduce both capture rate and post-release mortality (Lokkeborg, 2004; Gilman et al., 2006, Santos et al., 2012). Therefore in order to increase the area covered for circle hook studies in the Atlantic Ocean, the Portuguese Fisheries and Aquaculture Directorate and a private fishing company funded a project (SELECT-PAL: “Redução das capturas acessórias na pescaria de palangre de superfície”) to test the influence of different hook style and bait type combinations on the catch of target and non-target species caught by the Portuguese pelagic longline fishery operating in three major areas in the Atlantic Ocean: North-eastern Tropical, Equatorial and Southern Temperate. The aim of the present dissertation is to assess the effect of circle hook styles and bait types in reducing the mortality of sea turtles in the Portuguese pelagic longline fishery in Southern Atlantic Ocean. In addition, it was examined the effects of circle hooks and bait types on the catch of target species and on other non-target species taken as bycatch. To achieve the proposed objectives, the present thesis was structured in the following chapters:  Chapter I focus on the "state of the art" regarding the description of the fishery, including a brief description of the history, characteristics and catch species; evolution of the swordfish fisheries and global trends; characterization of the Portuguese pelagic longline fisheries and; mitigation measures of incidental catches;  On chapter II it was described the Material and Methods with the description of the study design, data collection and data analysis for sea turtles and fishes.  Chapter III focuses on the assessment of the effects of hook styles and bait types on sea turtles bycatch, particularly on catch composition by species, hooking location and mortality. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 3  Chapter IV concentrates on the influence of different hook styles and bait types combinations on the catches of target and non-target fish species, namely in terms of catch composition, catch rates, catch at size, and mortality at-haulback of discarded taxa. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 4 CHAPTER I – FISHERIES AND RESOURCES FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 5 1.1. THE PELAGIC LONGLINE FISHERY 1.1.1. Brief history of pelagic longline gear The most widespread form of pelagic longline gear appears to have been originally developed by the Japanese (Watson and Kerstetter, 2006 and reference therein). The introduction of the internal combustion engine in the early 1900´s, resulted in an expansion of fishing grounds, enabling the Japanese to target albacore in the central Pacific (Watson and Kerstetter, 2006). At the beginning of the 20th century, longline use was documented in the Mediterranean (Watson and Kerstetter, 2006 and reference therein) and in the mid1940´s an early form of pelagic longline in western North Atlantic was developed to target bluefin tuna (Thunnus thynnus) on Stellwagen Bank combining keg-line swordfish harpoon gear and halibut line-trawl (Wilson, 1960). In Norway in the late 1960´s a form of pelagic longline gear which used multifilament synthetic line floated just under the surface of the water was developed, to target porbeagle shark (Lamna nasus) (Gibson, 1998). During this period, small vessels using a very similar gear were also fishing swordfish at night off the coast of Cuba (Watson and Kerstetter, 2006 and reference therein). In the course of the 1950´s, the Japanese distance-water longliners operating in the north Pacific started targeting swordfish and albacore (Ward et al., 2000) and during late 1950´s and early 1960´s began the global expansion of longline fisheries. By then, many Japanese distance-water longliners started targeting tuna such as yellowfin and bigeye for sashimi markets throughout the Atlantic (North and South) Ocean and Mediterranean Sea. This expansion was initially driven by the Japanese tuna market and supported by international transportation, freezing technology, emerging markets for swordfish and shark fins that encouraged additional fleet expansion from others countries (Watson and Kerstetter, 2006). Multifilament nylon mainlines still dominate the international fishery, but the development of single-strand monofilament line in the 1970´s and the use of light dispositive (chemical light sticks) resulted in the expansion of the pelagic longline as the primary worldwide method of commercially harvesting of large pelagic fishes (Watson and Kerstetter, 2006). 1.1.2. Characteristics of the pelagic longline gear Pelagic longline gear is composed of a long length of mainline deployed across the ocean, to which numerous branch lines are attached, being suspended in the water column between regularly spaced floats. A branch line is a single line with a snap at one end and a hook at the other and connects the mainline to a single baited hook, with anywhere from 4 to 30 branch lines (baited hooks) between floats (Swenarton and Beverly, 2004). A typical longline FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 6 set from a medium-scale longliner would be about 30 to 40 nm long and have about 1200 to 2500 hooks (Beverly et al., 2003). Since the 1970´s longline fishing has evolved due to a better knowledge of vertical distribution of main target species, relationships of catches to temperature, dissolved oxygen, thermocline depth, and other environmental factors (Campbell et al., 1997; Hampton et al., 1998; Bertrand et al., 2002). These fisheries are opportunistic, switching gear style and making subtle changes to the fishing gear configuration to target the best available economic opportunity of each individual trip (Brothers et al., 1999). Depending on the target species, pelagic longlines can be set at a variety of depths from the surface layer, down to the thermocline, however even deep-set lines have a high percentage of their hooks (the ones nearest the floats) fishing in shallow water. Swordfish targeted longline gear, categorized “shallow-set” fishing, deploys surface gear by using usually four to five hooks between floats and no weight on the branch lines (Figure 1.1). Swordfish vessels make ample use of light emitting devices (lighsticks, battery powered light, etc.) placed near the hooks to attract fish (Swenarton and Beverly, 2004). To target “deep-dwelling” species such as bigeye tuna, more hooks are set between floats, small weight may be attached to each branch line, longer floatlines are used and the velocity of the vessel during the setting is slowed while the mainline is expelled from the boat at a high rate through a line shooter (Beverly et al., 2003). This gear configuration defines the “deep-set” tuna targeted gear, with hooks reaching depths of 300 meters (m) or more (Figure 1.1). Because several fish species exhibit dial rhythms in feeding activity, setting times affect catch rates (Løkkeborg and Pina, 1997). The “shallow-set” style gear is set at night to catch swordfish as they rise to surface waters to feed, with the gear being hauled in the day. “Deep-set” gear normally sets in the day and hauls at night in order to catch bigeye tuna (Beverly et al., 2003). Figure 1.1 - Set characteristics for “shallow-set” and “deep-set” pelagic longline gear (Source: Crowder and Myers, 2001) FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 7 Sharks comprise a large proportion of the total catch in some pelagic longline fisheries. For instance, sharks comprise more than 25% of the total catch in fisheries like the Australia longline tuna and Fiji longline tuna fishery. Location of fishing grounds, characteristics and methods of fishing gear are the primary factors determining a fleet’s shark catch rate (Gilman et al., 2007a). In recent years, has been observed a trend related to the gear design in order to maximize shark catches such as the use of wire leader and depth of baited hooks where “shallow-set” generally have higher shark catch rates than deeper setting fisheries (Gilman et al., 2007a). Some artisanal longline fisheries in the Pacific Ocean such as the Chilean and Peruvian artisanal longline, target dolphinfish (Coryphaena hippurus) during the austral summer and sharks are targeted from autumn to spring. Fishers target dolphinfish seasonally when this species is abundant in coastal waters due to shorter distance to fishing grounds and shorter trip length, with concomitant reduced costs in fuel and food. Wire leaders are not typically used during the dolphinfish season, whereas are used during the shark season to maximize shark retention and reduce gear loss (Gilman et al., 2007a). The same trend of the Pacific Ocean is observed in Atlantic Ocean (e.g. Portuguese coast and Azores islands) for the swordfish pelagic longline fishery where some vessels from spring to summer, making use of wire leaders on the branch lines direct their fishing effort to blue shark due to the lower availability of swordfish, reduced costs for fuel and rising market prices for shark meat and fins (pers. obs.) 1.1.3. The catch: Target and bycatch species The main target species of pelagic longline fisheries are tunas and billfishes, while other species including sharks are also an important component of the catch. The catch is normally divided into two distinct categories: target and bycatch. Tunas are by far the most important target species for pelagic longline and the main captured species are bluefin (Thunnus thynnus thynnus), bigeye (Thunnus obesus), yellowfin (Thunnus albacares), and albacore tuna (Thunnus alalunga). Some billfishes are also targeted, with broadbill swordfish (Xiphias gladius) being the most important, followed by striped marlin (Tetrapturus audax) (FAO, 2012). Bycatch are species that are caught incidentally (not targeted) that can be retained aboard for sale because of their commercial value, or discarded as they have no commercial value or are protected under management measures requiring they not to be landed. Some of those species include shortbill spearfish (Tetrapturus angustirostris), sailfish (Istiophorus albicans), dolphinfish (Coryphaena hippurus), wahoo (Acanthocybium solandri), escolar (Lepidocybium flavobrunneum), amongst others. A range of pelagic shark species such as blue shark (Prionace glauca), shortfin mako (Isurus oxyrinchus), oceanic whitetip FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 8 (Carcharhinus longimanus) and silky shark (Carcharhinus falciformis) are also taken as bycatch, although they are mainly prized for the value of their fins. The most common bycatch species that are discarded include lancetfish (Alepisaurus ferox) snake mackerel (Gempylus serpens), pelagic stingray (Pteroplatytrygon violacea), sea turtles and sea birds. Snake mackerel, lancetfish and pelagic rays can be taken at various depths on a longline, and are not associated with a particular type of pelagic longline style (Beverly et al., 2003). Sea turtles (e.g. loggerhead and leatherback turtles) and seabirds (such as albatrosses and petrels) are also caught, where sea turtles are often taken on the shallow hooks, generally near the floatline and the seabirds attack the baits on the gear as it is being set. 1.2. EVOLUTION OF THE SWORDFISH FISHERIES AND GLOBAL TRENDS 1.2.1. Global evolution of the swordfish fisheries The swordfish fisheries started around 1000 BC as near-shore subsistence activities of Mediterranean countries with the fishing methods mostly involved being the harpooning of large female specimens as they were basking at the sea surface (Ward et al., 2000). During the 1900´s, harpoon became more sophisticated in many areas, with the introduction of motorized boats, spotting planes and harpoons that give the fish a lethal electric shock (Ward et al., 2010). Nevertheless, most harpoon fisheries declined during 1980´s as a result of increased labour costs and introduction of more efficient fishing gears, such as the driftnets and longlines (Ward et al., 2000). Distance-water longliners targeting albacore tuna and swordfish started during the early 1950´s by Japanese fleet, operating in the north Pacific and in the late 1950´s many longliners vessels started targeting yellowfin and bigeye tuna for sashimi markets in Japan (Ward et al., 2000). In that period, more than half of the world´s swordfish catches were taken as incidental catch by the longliners targeting tuna (Ward and Elscot, 2000). The driftnet was introduced by Japanese fleet in early 1960´s to target pelagic species, particularly tunas. Encouraged by government incentives many small scale commercial fishers from Japan and Taiwan industrialized their fleets and in the mid-1980´s it was widely used to target pelagic species in international waters creating an outcry over wastage and incidental catch of sea wildlife that took in 1991 the United Nations (UN) banning the use of driftnets longer than 2.5 km long in international waters (Ward et al., 2000). FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 9 In the 1960´s smaller locally based longliners began to make shorter trips, storing their catches in ice to be sold in local markets and in the 1980´s with the improvement of air freight to distant markets the fresh-chill fleet quickly developed in many parts of the world (Ward et al., 2000), which contributed to the increased swordfish catches during that time. In the mid-1980´s the techniques to target swordfish became widespread (e.g. use of squid as bait, attached light, set at shallow depths at night, etc.) and catch rates became much higher than those for longliners targeting tuna. Improvements in gear such as the hauled speed, hydraulic powered reels and monofilament mainlines also contributed for an improvement on catch rates as well as in efficiency. Nowadays, swordfish is mainly caught with longline while fishing gears such as the harpoon represent a small portion of the total fishing effort (Ward et al., 2000). 1.2.2. Overview of the global trends of swordfish landings The swordfish landings started to increase during the mid-1950’s keeping this trend until 1970 when reached a peak of 39.000 Metric Tons (MT) (Figure 1.2), where the Pacific was responsible for 55%, the Atlantic 30%, the Mediterranean Sea 10% and the Indian Ocean less than 4% of the total swordfish landings (Figure 1.3). A 1971–78 reduction by the Food and Drug Administration (FDA) in the tolerance level for mercury in swordfish reduced imports and severely depressed consumption worldwide. However, in 1978 FDA revised its limits and from then swordfish landings return to exponentially increase, particularly in the Atlantic and Indian Oceans, where in 1980 the catches in the Atlantic waters overcome those from the Pacific Ocean (FAO, 2012). The redirection of the Taiwan fleet from the Pacific to Indian and Atlantic Oceans was the main factor contributing to the increase of swordfish catches in these oceans (Ward et al., 2000). In 1991 a slight decrease in the total swordfish landings occurred due to the United Nations (UN) ban on the use of driftnets longer than 2.5 km long in international waters. However, the vessels quickly shifted to the longline gear and total landings return to increase (Figure 1.2). In 2003 the historical maximum of the swordfish landing (119157 MT) was reached with the Indian, Pacific, Atlantic Oceans and Mediterranean Sea responsible for 33%, 31%, 21% and 13% of the total landings, respectively (Figure 1.2 and Figure 1.3). In the last decade a slightly decline on swordfish catches has been observed mainly due to the conservation and management measures implemented by tuna Regional Fisheries Management Organizations (tRFMO). FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 16 1.3.2. Description of the swordfish pelagic longline gear As mentioned before, the Portuguese pelagic longline fishery targeting swordfish began in the 1980´s and the fishing method has remained almost unchanged since then (Santos et al., 2012). Although a few changes have been incorporated in the last decade: i) fishermen shifting from the traditional gear, described by Rey and Alot (1984) to the so-called automatic or “American-style”, making use of mainlines and branch lines of monofilament and using flashlights and ii) in specific areas and seasons, pelagic sharks may be the target species, as a consequence the branch lines material are shifted to multifilament steel (wire leaders). There are no significant differences at the level of the material used in the preparation of the fishing gear, with the main difference being the automation of procedures carried out during the setting and hauling operations. As a result, the needs of skilled manpower are smaller compared to the traditional style longline. The automatic pelagic longline is based on a basic unit, consisting of four parts: the mainline, the branch lines, the hook and the bait. The mainline which varies in length, stretch for tens of kilometers rigged with a certain number of hooks off the branch line. Polyamide (nylon) monofilament is the most common material used for branch lines and mainline targeting swordfish, because the catch performance has been shown to be superior to those with multifilament (Brothers et al., 1999). The branch lines are connected to the mainline trough a snap. A swivel is placed immediately after the snap, which is attached to a nylon monofilament with 2.0 to 2.5 mm diameter. In the middle of the branch line a swivel with a lead of 60-80 g maybe adapted and a luminescent device is connected to it. Another swivel is than attached to the hook by a nylon monofilament of approximately 2 m with a diameter of 1.8 to 2.2 mm. The shape and size of hooks has been affected by both catch and bycatch considerations, however historically J-style hooks have been used. In Figure 1.7 a schematic drawing of the swordfish pelagic longline fisheries is shown. Depending on the target species and/or area, variations can be found in terms of type, length and dimension of branch line and mainline, type of hook and bait. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 17 Figure 1.7 - Schematic drawing of the swordfish pelagic longline fishing gear (Source: adapted from Santos et al., 2013). Monofilament gear is set from the stern of the boat with the aid of a line setter and can last for between 5 to 7 hours, depending on the number of hooks deployed and/or velocity of the vessel. As the line is sent out during setting, baited branch lines and floatlines with floats are attached at intervals usually controlled by an audible signal and the baited hooks remain at/or near the surface for a short period of time before they start sinking. Longlines targeting swordfish preferably set their gear at sunset due to adaptation to phototropism of this species. Once released, the gear drifts for a few hours (6 to 7 hours) until it is hauled, normally at sunrise. The hauling procedure is made at a velocity inferior to that of setting and normally takes between 7 to 10 hours, depending on the catches and constraints that might arise. The most used baits are squid (Illex spp.) and mackerel (Scomber spp.) that can be lured by the head or lower body. In Table 1.1 (Annex I) a summary of the main characteristics of the Portuguese pelagic longline fishery is shown. 1.4. MITIGATION OF INCIDENTAL CATCHES Drifting pelagic longlines catch a wide variety of bycatch, the unintended non-target organisms that are captured during fishing operations (Lewison et al., 2004). Despite the FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 18 differences in bycatch types and the magnitude of their effects from one fishery to another, bycatch can be a major driver from sea “megafauna” to lower trophic-level species, critical for the structure and functioning of the sea ecosystems and therefore the provision of ecosystems services (Gilman, 2011). Unsustainable bycatch fishing mortality of keystone species that play critical roles in regulating ecosystems processes can alter trophic interactions and change ecosystem structure and functioning, including reduced ecosystem resistance and resilience to environmental fluctuations, and possibly exceeding “tipping points”, where permanent regime shifts occur (Pauly et al., 1998). Under the “United Nations Convention on the Law of the Sea (UNCLOS)”, states are obligated to protect and preserve the sea environment and consider the effects of fishing on species associated with/or dependent upon commercially exploited species (United Nations, 1982). Additionally the Food and Agriculture Organization of the United Nations (“FAO Code of Conduct for Responsible Fisheries”) require nations to develop and apply safe and selective fishing gear to minimize waste, discards catch of non-target species and effects on associated or dependent species (FAO, 1995). Therefore, in recent years has been increased concern with regards to bycatch (Soykan et al., 2008) and several recent studies have addressed this issue in pelagic longline fisheries (Gilman et al., 2006). Most bycatch studies have focused on the more vulnerable, and charismatic sea “megafauna”, including sea turtles (e.g. Watson et al., 2005; Gilman et al., 2006; Santos et al., 2012), sea birds (e.g. Bugoni et al., 2008; Gilman et al., 2008; Jiménez et al., 2009) and sharks (e.g. Yokota et al., 2006; Ward et al., 2009; Coelho et al., 2012; Foster et al., 2012). Several measures to mitigate the incidental capture and bycatch have been proposed and/or implemented in different fisheries, including management measures such as i) regulatory controls on fishing effort, seasonal bycatch levels, fishing areas ii) programmes to avoid bycatch hotspots through voluntary fleet communications and iii) handling and release practices to increase post-survival (Gilman et al., 2006). Because of problematic turtle bycatch levels an area of western North Atlantic (including the productive Grand Banks) was partially closed to the US pelagic longline fleet in 2000, and completely closed in 2001 (US National Sea Fisheries Service 2000, 2001a, b) only reopening in 2004 after regulations were amended to require the use of recently tested turtle bycatch avoidance methods (US National Sea Fisheries Service 2004b). In the Hawaii longline swordfish fishery, similar restrictions were implemented, where it was closed for over 4 years and is now subject to strict management measures, including prescribed use of large circle hooks and fish bait, restricted annual effort, annual limits on turtle captures and 100% onboard observer coverage (US National Sea Fisheries Service 2004a). FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 19 In recent years scientists have been developing, testing and implementing fishing techniques and gear modifications to improve the selectivity and sustainability of pelagic longline fisheries and increase post-release survival (Soykan et al., 2008). This approach might be better accepted by fishers over other management strategies that reduce available fishing areas, such as time/area closures, which have predominated U.S. bycatch reduction measures. Longline gear operating characteristics including, fishing depth, gear soak time, bait type and hook style can have significant effects on the selectivity of the pelagic longline gear (Watson and Kerstetter, 2006). One such technique that has broadly been developed and tested to reduce bycatch is the circle hooks (see review by Graves et al., 2012). Circle hooks also seem to reduce the incidence of deep hooking trauma and post-release mortality in a variety of sea turtles (see review by Read, 2007) and freshwater/sea fishes (Cooke and Suski, 2004) and their overall benefit to commercial fisheries was recently reviewed by Graves et al., (2012). The tuna Regional Fisheries Management Organizations (tRFMOs) in the last decade have been encouraging their contracting parties and cooperating members (CPCs) to undertake research trials of circle hooks in their commercial pelagic longline fisheries. In January of 2010, Western and Central Pacific Fisheries Commission (WCPFC) was the first tRFMO to include the use of large circle hooks with an offset (that does not exceed 10º) as one available bycatch mitigation method required for implementation by all CPCs fishing for swordfish using shallow longline sets (WCPFC, 2008). However, conflicting results among studies (e.g. review paper by Read, 2007) conducted under several locations, seasons, and experimental protocols have hindered the development of regulations requiring the use of circle hooks in the others tRFMO´s. In some cases, the gear modifications can reduce the catches of the target species to such a degree that their use is impractical [e.g. the case study of the dolphinfish fishery in Ecuador reported by Largarcha et al. (2005)]. Given these complexities that involve possible modifications in fishing gears, different approaches have been recommended for different researchers. While Read (2007) recommends that bycatch mitigation measures should be tested (in rigorous experiments) before being mandatory in any fishery, Serafy et al. (2009) considered that this perspective is counter to the precautionary approach and rigorous field testing is preferable, considering that in cases of severely overfished, threatened or endangered populations, highly complex and time consuming field experimentation should not be a precondition for a given fishery change that could potentially benefit those populations. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 20 CHAPTER II – MATERIAL AND METHODS FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 21 2.1. STUDY DESIGN AND DATA COLLECTION For this study, a total of 310 longline sets were carried out during five trips along the Southern Atlantic region (Figure 2.1) that took place between October 2008 and February 2012. A commercial fishing vessel (“ALMA LUSA, PM-1269-N) from the Portuguese swordfish pelagic longline fleet participated in the study, with experimental fishing taking place between 11o to 34o S latitude and 044o W to 007o E longitude. The fishing gear consisted in a standard US style monofilament polyamide longline of 3.6 mm of diameter, approximately 62 nm (~ 110 km) long, with five polyamide branch lines between two buoy floats at intervals of approximately 80 meters (range from 70 - 90 m) and a depth of 20-50 m below the surface. Each branch line was 18 m in length and was composed by two sections, the first consisting of 2.5 mm monofilament (9 m length) connected by a swivel to a 2.2 mm monofilament gangion (9 m length) with a hook in the terminal tackle. A battery flashlight (green color) was attached to each gangion. On each set, gear deployment (1440 hooks) commenced around dusk (traditionally at 17:00 hours), with haulback starting around dawn of the next day (about 06:00 hours). Figure 2.1 – Study area and location of the 310 experimental longline sets in the Southern Atlantic region. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 22 Three different stainless steel hook styles (Figure 2.2), produced by WON YANG, Korea were used in each longline set, where the control corresponding to the traditional J hook on the fishery (EC-9/0-R), and the treatments corresponding to: GT hook, a 10º offset circle hook (H17/0-M-R); and G hook, a non-offset circle hook (H17/0-M-S). Figure 2.2 – Photograph of the three hook styles used during this study: (A) J – 10º offset 9/0 hook; (B) GT – 10º offset 17/0 circle hook and (C) G – non-offset 17/0 circle hook. Hook measurements and terminology in Table 2.1. The specifications of the different hooks are summarized in Table 2.1. In order to minimize the potential for confounding effects specific to a set (e.g. location, water temperature, chlorophyll, thermocline or other factors) hook style was alternated section by section of the longline. Each section was delimited by radio beacon buoys, and had 80 hooks of only one style. Furthermore, the hook style of the first section in the water changed every set, following a fixed scheme (i.e., J, G, GT, J, G, GT and so on). Two different bait types were used, mackerel (Scomber spp.) and squid (Illex spp.) but as recommended by Watson et al. (2005) only one type of bait was used in each set in order to avoid possible interaction effects. Standardized bait sizes were used in all longline sets. Mackerel and squid baits had 35.1±1.19 cm fork length and 27.8±0.97 cm mantle length, respectively (based on the measurement of 200 individuals of each genus). FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 23 Table 2.1 - Hook measurements and terminology of the different hook styles used in the study. The mean size of the hook parameters is based on the measurement of 50 hooks of each style. Standard deviation is indicated within parentheses. All characteristics of the fishing gear and practices (e.g. hook placement, flashlight color, bait size and hook manufacture) were standardized along the study. Length of mainline was allowed to slightly vary, as a result of sea conditions. The fishing effort in terms of deployed number of hooks of each style per set was kept constant throughout the study. For every set the first baskets of the gear deployment was monitored in order to ensure the correct execution of the experiment. All data was recorded by onboard scientific fisheries observers during hauling operations using standardize forms and procedures. For every experimental set, date, location (initial and final latitude and longitude) and number of hooks of each hook style was recorded. Sea Surface Temperature (SST) was also collected, being recorded at the beginning of haulback. Whenever a sea turtle was caught in the longline, the onboard observer identified the species, recorded the hook style and bait type used, the condition/status of the turtle athaulback (alive/dead), the type of interaction (i.e. location of the hook: flippers, mouth, esophagus or entangled) and the condition when released (alive/dead). When possible, turtles were boated with a large dip net. Further, and whenever possible, observers and crew attempted to remove fishing gear using long-handled de-hookers and line cutters. Observers attempted to remove all gear immediately. They were instructed to remove all external hooks and those in the mouth, as well as hooks in the esophagus when the insertion point of the barb could be seen. Whenever possible the sex of the specimen was determined and the curved carapace length (CCL) was measured to the nearest lower 1 cm. However, due to the size and weight of leatherback turtles, Dermochelys coriacea, only a limited number of specimens of this species were measured, with most specimens being immediately released by cutting-off the line without bringing the turtle onboard. Following Watson et al. (2005), power tests were carried out in order to estimate the experimental fishing effort required to detect a fishing method that has different degrees of effectiveness in reducing bycatch of sea Parameter Hook style J (EC-9/0-R) GT (H17/0-M-R) G (H17/0-M-S) Total length (mm) 87.2 (±1.11) 77.7 (±0.92) Front length (mm) 40.4 (±1.10) 43.9 (±0.45) Maximum width (mm) 43.3 (±0.64) 49.4 (±0.88) Gap (mm) 33.2 (±0.59) 27.0 (±0.51) Arm diameter (mm) 5.0 (±0.00) 5.0 (±0.00) Offset angle 10º 10º 0º FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 24 turtles in comparison with the control fishing method. The control fishing method was assumed to be the combination most commonly used in the fishery, specifically J-style hooks baited with squid, and the power calculations were based on the necessary number of hooks required to detect a 25% and 50% reduction in bycatch rate in the case of loggerheads and leatherbacks, respectively. For every fish captured the hook style and bait type were recorded and the species identified and measured to the nearest lower 1 cm. For swordfish and marlins length measurements were registered in lower jaw fork length (LJFL) and fork length (FL) for sharks, tunas and other teleosts. Information of catch disposition (retained/discarded), condition at-hauback (alive/dead), sex and its condition if discarded (alive/dead) was also collected. Large fishes (e.g. manta rays and sunfish) were released by cutting the leader as close to the animal as safely possible. Fish catches were placed into three categories depending on the species: target, bycatch or discard. For this fishery and using these specific fishing techniques, the main target species was swordfish. Bycatch included species that were not targeted, but were retained if caught such as tunas, billfishes, large pelagic sharks and other teleosts. Discards included species that were unintentionally caught, but not retained, mostly small sized elasmobranchs, teleosts with no commercial value and also larger elasmobranchs that are now forbidden to be retained aboard (e.g. thresher and hammerhead sharks) (Table 2.2, Annex II). 2.2. DATA ANALYSIS All statistical analyses were performed using the R project for Statistical Computing version 2.15.1 (R Development Core Team, 2012), primarily using functions available in the core R program. Exceptions were the Levene test for the homoscedasticity that is available in library “car” (Fox and Weisberg, 2011); non-parametric multiple test procedure for all-pairs comparisons available in library “nparcomp” (Konietschke, 2012); contingency table analysis that was performed with library “gmodels” (Warnes et al., 2011); Generalized Linear Model (GLM) fitting and maximum likelihood estimation of the index parameter using the Tweedie distributions available in library “tweedie” (Dunn, 2010) and the plots of means available in the “Rcmdr” (Fox et al., 2011). All maps in the thesis were created through ArcGis Desktop 10 software (ESRI, 2012). The shapefile (world borders) with the continental contours was obtained in the “Thematic mapping” website (Sandvik, 2009). A Pearson´s product-moment correlation coefficient between SST and latitude and longitude was performed in order to FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 25 evaluate the strength of linear dependence between these variables. For all tests performed, significant differences were declared at p<0.05. 2.2.1. Data analysis for sea turtles Catch rates were expressed as bycatch per unit effort (BPUE), calculated as the number of specimens caught per 1000 hooks. Given the lack of normality of the BPUE data verified with Kolmogorov-Smirnov tests with Lilliefors correction (Lilliefors, 1969) and heterogeneity in the variances (verified with Levene tests), Kruskal-Wallis tests were used to compare BPUE between different hook styles, and Mann-Whitney tests were used to compare BPUE between the two baits. A logistic-binomial generalized linear model (GLM) was used to determine the influence of hook style and bait type on turtle bycatch. Due to the small sample sizes, this model was only applied to the loggerheads. For this model, the response variable was the proportion of loggerhead catches in each longline set, calculated as the number of catches given the number of hooks used in each set. A binomial error distribution and a logit link function were used in the model. The explanatory variables tested were the hook style (J, G or GT) and the bait type (squid or mackerel), with their significance verified by the Wald statistic. The interaction between the two variables was tested with a likelihood ratio test and by comparing the Akaike information criterion (AIC) values of the models. This interaction was used in the final model because it was considered significant and relevant for interpreting the results. The odds-ratios of the parameters, with their respective 95% confidence intervals, were calculated considering the model parameters and the interaction. With regards to the size structure of the sea turtles caught, only the most abundant species (loggerheads) was analyzed, while the CCL of leatherbacks was not compared due to the small samples size. Loggerhead CCL was tested for normality and homogeneity of variances, and the skewness and kurtosis were calculated. Considering the results of these analyses, the application of parametric tests seemed reasonable, and therefore the mean CCL for the two different baits were compared with Student´s t-test, while the mean CCL for the three different hook styles were compared with Analysis of variances (ANOVAs). Additionally, the mean CCL for the different hooking locations were also compared using ANOVAs. When the ANOVA results were significant, Schefee post-hoc multiple comparisons were carried out. The relationship between hooking location and hook style was assessed using contingency tables and Chi-square tests of independence. Analyses were conducted for the two species FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 32 Figure 3.1 - Spatial distribution of BPUE by longline experimental set, for turtle species combined (top), loggerhead (TTL - C. caretta, middle) and leatherback (DKK - D. coriacea, bottom).The size of the circles is proportional to the BPUE and the dark crosses represent fishing sets with 0 catches. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 33 For the loggerhead sea turtle, and using the binomial modeling analysis, both the hook style and the bait type were significant for explaining the BPUE rates. Additionally, the interaction between hook style and the bait type was marginally significant (likelihood ratio test: diff. residual deviance=5.32, p=0.07), and produced a slightly lower AIC value (simple effects model AIC = 1404.3; model with hook-bait interaction AIC = 1402.9). When changing the bait type from squid to mackerel the odds-ratios of catching loggerhead sea turtles decreased significantly regardless of the hook style used, with these decreases ranging between 6782% (Table 3.3). However, and due to the interaction observed, changing from J-style to one of the circle hooks was only significantly different when using squid bait (with the odds-ratios decreasing between 54% and 63%), but not when using mackerel bait (with the 95% confidence intervals of the odds-ratios ranging between reductions of 65% to increases of 86%). Figure 3.2 - Plot of the mean BPUE (with the respective standard errors) observed with the different hook styles (J, G and GT) and bait types combinations, for the species combined, loggerheads (TTL - C. caretta) and leatherbacks (DKK - D. coriacea). On the bait type, M refers to mackerel and S refers to squid. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 34 Table 3.2 - Ratio between the mean BPUE obtained with the standard fishing gear (J hook baited with squid - control) and the different combinations of hook style (J – 10º offset 9/0 hook; G – non-offset 17/0 circle hook; GT – 10º offset 17/0 circle hook) and bait type (S – squid; and M – mackerel) tested, for species combined and for the two sea turtle species caught. Comparison Loggerhead Leatherback Combined species JS vs. GS 2.2 3.5 2.3 JS vs. GTS 2.7 3.5 2.7 JS vs. GM 5.6 - 6.3 JS vs. GTM 8.0 14.0 8.4 JS vs. JM 5.6 4.7 5.5 Table 3.3: Odds-ratios, with the respective 95% confidence intervals, for the effects of changing hook styles and bait types in the loggerhead (Caretta caretta) BPUE, accounting for the model interactions. Interaction Main factor Estimate Lower 95%CI Upper 95% CI Using squid bait Change from J to G 0.46 0.33 0.64 Change from J to GT 0.37 0.26 0.53 Using mackerel bait Change from J to G 1.00 0.54 1.86 Change from J to GT 0.70 0.35 1.39 Using J-style hook Change from Squid to Mackerel 0.18 0.11 0.29 Using G-style hook Change from Squid to Mackerel 0.38 0.23 0.64 Using GT-style hook Change from Squid to Mackerel 0.33 0.18 0.61 3.2.3. Bycatch at size and hooking location Loggerheads ranged in CCL from 41 to 78 cm and averaged 61.5 (±6.09) cm (N=260, n=234). Only 42% (N=26) of leatherback turtles (CCL from 48 to 140 cm and averaged 92.9 (±33.82) cm) were measured. For both species combined, the mouth was the most frequent hooking location (65.7%) regardless of the hook type used (Figure 3.3). However, when the species were analyzed separately it was possible to determine species-specific patterns of hooking locations. Leatherbacks were almost exclusively hooked by the flippers (73.1%) or entangled (19.2%) on the lines, whereas most loggerhead turtles bite the bait, with 71.5% hooked in the mouth and 17.7% hooked in the esophagus (Figure 3.3). The relative proportions of the different hooking locations were statistically different between hook styles (Figure 3.3), as confirmed by Chi-square tests between the two factors. This analysis was carried out for species combined (Chi-square=17.80, df=4, p=0.001) and for the loggerhead (Chi-square=20.87, df=4, p<0.001). On the contrary, the relative proportions of the different hooking locations were not statistically different between bait types for species FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 35 combined (Chi-square=1.72, df=2, p=0.424) and the loggerhead (Chi-square=1.74, df=2, p=0.418) (Figure 3.3). These analyses were not performed for the leatherback as most specimens were captured by the flippers, and the contingency tables had cells with zero values for most of the other combinations. Figure 3.3 - Hooking location per hook style (left) and bait type (right) for all species combined, loggerhead (C. caretta) and leatherback (D. coriacea). The bars refer to the percentage of each hooking location within each hook style or bait type. Numbers between brackets refer to the corresponding nominal catch of each hook style or bait type. For loggerheads the size distribution did not significantly vary depending on the bait type (tStudent: t=1.19, df=232, p=0.236) (Figure 3.4). However, significant differences in the size distributions were detected between hook styles (ANOVA: F=7.73, df=2, p<0.001), and hooking locations (ANOVA: F=8.71, df=3, p<0.001, Figure 3.4). Using Schefee post-hoc multiple comparison tests for hook styles, it was noted that significant differences only occurred between GT hooks and the other two hook styles (J and G), with GT hooks FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 36 capturing significantly larger specimens. With regards to the hooking location, significant differences were found between entangled and the remaining hooking locations, with the entangled specimens significantly smaller (Figure 3.4). Figure 3.4 - Boxplots with the size distribution (median, inter-quartil range, non-outliers range and outliers) for loggerhead (C. caretta), for each of the three factors considered (bait type, hook style and hooking location). 3.2.4. Mortality Overall, 65% of all sea turtles were alive at-haulback and were, therefore, released alive. The overall percentage of alive specimens at-haulback was higher for leatherbacks (85%) than for loggerheads (63%). The hooking location seems to have a great impact on mortality with most specimens caught by the flippers being alive at the time of haulback (88%), while the specimens entangled or hooked in the esophagus and in the mouth had lower percentages of alive specimens (22%, 48% and 66% alive at the time of haulback, respectively) (Figure 3.5). Because the two species tended to be hooked in different ways, hooking location reflected the species-specific mortality. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 37 Figure 3.5 - Percentage of fishing mortality at-haulback per hooking location (left), hook style (center) and bait type (right), for all species combined (top), loggerhead (C. caretta, middle) and leatherback (D. coriacea, bottom). The numbers between brackets refer to the corresponding nominal catch for each hooking location (n), hook style and bait type. For the factor hook style, and considering species combined, the GT-style hook had proportionally more turtles alive (83%) than dead (17%), with the percentage of alive specimens decreasing substantially for the J-style hooks (64%) and even more with the G hooks (53%), with those differences statistically significant (Chi-square Proportion test: Chisquare=13.27, df=2, p=0.001). When the loggerhead data were analyzed separately, the proportion of alive specimens was 83%, 62% and 50% for hook types GT, J and G, respectively (Figure 3.5), which was again statistically significant (Chi-square Proportion test: Chi-square=14.64, df=2, p<0.001). For leatherbacks, the proportions of alive specimens were very high for all hook styles; 100%, 82% and 80% for hook types G, J and GT, FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 38 respectively (Figure 3.5). For the factor bait type, the observed versus expected frequencies of dead and alive turtles were not significantly different for species combined (Proportion Chi-square with Yates correction: Chi-square=0.49, df=1, p=0.48), neither for loggerheads individually (Proportion Chi-square with Yates correction: Chi-square=0.21, df=1, p=0.65). 3.3. DISCUSSION The overall mean sea turtle BPUE observed in this study using the traditional gear configuration (1.694/1000 hooks) was similar to that reported by Sales et al. (2010) for another pelagic longline fishery targeting swordfish off southern Brazil (1.893/1000 hooks). However, our observed overall BPUE was higher than those reported by Pinedo and Polacheck (2004) off southern Brazil (1.48/1000 hooks), Pons et al. (2010) off Uruguay (average of 1.00 loggerheads/1000 hooks between 1998 and 2007), Petersen et al. (2009) off South Africa (0.04/1000 hooks) and Afonso et al. (2012) off northern Brazil (0.470.94/1000 hooks). Major interactions (22% of the sets) of loggerhead, and to a smaller extent leatherback sea turtles, seem to exist with the Southern Atlantic Portuguese pelagic swordfish longline fishery, particularly between 37ºW and 44ºW of longitude, as shown by the present study. A similar trend was found by Pinedo and Polacheck (2004) and Sales et al. (2010) in the South Atlantic, with loggerheads followed by leatherbacks also being the most captured species by the Brazilian and Uruguayan pelagic longline fleets. In the Equatorial Atlantic, the olive ridley (Lepidochelys olivacea) was the sea turtle species that interacted the most with the pelagic swordfish longline fishery, although other sea turtles were also present (Carranza et al., 2006; Sales et al., 2008; Santos et al., 2012). In comparison, for the North-west Atlantic region the loggerheads and leatherbacks seem to be the species most commonly caught in pelagic longlines (Watson et al., 2005; Foster et al., 2012). Hence, and as suggested by Gardner et al. (2008), the incidental capture of sea turtles seems to vary considerably by region, with the water temperature possibly playing a major role in this variability. The present study shows that sea turtle interactions can be significantly reduced by using mackerel in place of squid bait and to a lesser extent by employing circle hooks. A combination of circle hooks baited with mackerel can result in a reduction in sea turtle catches by 87.5% and 100% for loggerheads and leatherbacks, respectively. Still, the reductions observed in this study for leatherbacks should be interpreted with care, as the catches of that species in particular were very low. Similar findings were presented for the FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 39 South-east Atlantic by Anon. (2008), where it was suggested that bait type had the greatest influence on loggerhead turtle bycatch. Previous studies have also shown that changing the bait type from squid to mackerel (or other fish) and/or the traditional J to circle hooks, were effective measures to reduce sea turtle bycatch in different oceanic areas: in the North-west Atlantic (Watson et al., 2005; Foster et al., 2012); in North-west Pacific (Yokota et al., 2009); in the Equatorial Atlantic (Pacheco et al., 2011; Santos et al., 2012); and South Atlantic (Domingo et al., 2009; Sales et al., 2010). However, these comparisons should be carefully analyzed as the cited studies used slightly different hooks (in terms of sizes and shapes), covered different seasons and areas (with different ranges of temperature), and were based on substantially different numbers of sets. As loggerhead and leatherback turtles have different life histories, pelagic longlines impact both species differently, which can influence the size distribution of the captures. Leatherback sea turtles are pelagic/oceanic during all stages of their life (Bjorndal, 1997), thus a wide size range was observed in the captures, including adult specimens. It must be noted that the largest specimens captured were likely not measured due to difficulty in handling and boarding, thus no statistical inference should be made with regards to the sizes of the catches for this species. On the other hand, loggerheads typically frequent open waters feeding on pelagic invertebrates, where the juvenile development takes place, and after a decade or longer, sub-adults and adults move to neritic habitats near the continental coastline and start feeding upon benthic invertebrates (e.g. Mollusks) and fish (Bjorndal, 1997). As a result, based upon the information reported by Domingo et al., (2006) on the size at maturity for the South-western Atlantic Ocean, the captured loggerheads in our study were mostly juveniles. Similar catch-at-size of loggerheads were reported by other studies in the South Atlantic (Pinedo and Polacheck, 2004; Domingo et al., 2009; Sales et al., 2010). While bait type did not influenced loggerheads size distribution, significant differences were found in the size distribution between hook styles, with GT hooks capturing larger specimens. Stokes et al. (2012) alerts to the fact that when comparing hook type effects in size distribution a potential hook-size effect may be masked, due to the fact that most commonly used J hooks (7/0, 8/0, and 9/0) are slightly smaller than 16/0 and 18/0 circle hooks. Sales et al. (2010) also found significant differences in the sizes of captured loggerheads, with circle hooks capturing larger specimens, compared to those reported by Domingo et al. (2009) and Anon. (2008) who found no differences in the size distribution between hook types and bait in the same region. Likewise, no differences were found for the Olive ridley sea turtles caught in the Equatorial Atlantic by the Portuguese fishery (Santos et al., 2012). FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 40 Hooking location seemed to be mainly species-specific, which may be related to each species feeding behavior. While leatherbacks were almost exclusively hooked externally, mainly by the flippers (with all hook types), loggerheads were mostly hooked by the mouth in all treatments. Significant differences were found for this species in the relative proportions of the different hooking locations between hook styles, with J hooks showing a higher proportion of loggerheads retained by the esophagus, while the bait showed no significant differences. Likewise, in the North-west Atlantic, Watson et al. (2005) and Epperly et al. (2012) found no significant differences in hooking location for both loggerheads and leatherbacks upon switching between mackerel and squid bait. Anon. (2008) and Sales et al. (2010) also noted that deep-hooking involved more often J hooks than circle hooks, in the South Atlantic. In the North Atlantic, Stokes et al. (2012) also found significant differences in hooking location in loggerheads when comparing offset J hooks and non-offset and 10° offset circle hooks, with the latter hooking mostly loggerheads by the mouth while offset J hooks were swallowed more often. In contrast, Carruthers et al. (2009) found no significant differences in hooking location for loggerheads when comparing 16/0 circle hooks, non-offset J hooks, and offset (20°–30°) J hooks in the Canadian longline fishery for swordfish and tuna in the North Atlantic. The main factor that seemed to influence at-haulback mortality of sea turtles was the hooking location. Turtles hooked externally (by the flippers or entangled) showed a large proportion of specimens that were alive at-haulback, while specimens that were hooked in the mouth or deep hooked in the esophagus had a higher proportion of dead specimens at time of haulback. Hence, the type of circle hook appears to be an important factor in the mortality rate as well, as there were statistical differences between the three hook types tested. In loggerheads the GT hook showed the lowest at-haulback mortality, followed by J hook and G hook, whereas Sales et al. (2010) found no differences in loggerhead mortality among hook types. In the Equatorial Atlantic, Santos et al. (2012) also found differences in mortality between the same three hook types although, contrary to this study, for both sea turtle species (olive ridley and leatherback) the J hook had the highest mortality. Our reported mortality results represent the short term at-haulback mortality, and should be interpreted as minimum mortality estimates, as post-release mortality may occur. Estimating post-interaction survival is difficult given the variety of factors involved with each unique interaction. However, lightly hooked sea turtles (external and mouth hooked) have a higher chance of survival than sea turtles that swallow the hook, particularly when all gear is removed before release (Ryder et al., 2006; Swimmer et al., 2006). Results from satellite telemetry research also support the hypothesis that deeply hooked turtles have a higher FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 41 probability of mortality than lightly hooked turtles when all gear is removed (Chaloupka et al., 2004). Overall, the present study supports previous reported results on the reduction of sea turtle accidental catches in the swordfish longline fisheries, by changing the traditional configuration of J hook baited with squid to circle hooks baited with mackerel. It is important to note, however, that in this study the bait seemed to have more influence on the level of bycatch reduction than the hook style itself, and that in the case of mortality the effect of the hook type is not so evident (GT vs G instead of GT/G vs J). A high variability between results seems to exist in the literature, highlighting the influence of different aspects (e.g. region and consequently the species, season, fishery, etc.) in sea turtle accidental captures. For this reason extreme caution must be used when interpreting the results of this kind of studies. For example, Anon. (2008) mentions a remark of one observer alerting to the fact that circular hooks were much more difficult to remove than J-style hooks whether hooked in the mouth or internally. In addition, the observer stated that more traction was caused in the esophagus, producing tissue tears and hemorrhages when removing ingested circle hooks compared to J hooks. Parga (2012) illustrates another example of the uncertainty about the benefits of the use of circle hooks, stating that even though hooks in the mouth are generally considered low risk, sensitive structures are present in the mouth, such as the glottis or the jaw joint, that if damaged may cause death due to infection. The esophagus, on the other hand, has a strong muscular wall and is somewhat resistant to lesions, unless the hook lodges close to the heart or large blood vessels. The same author further noted that cutting short the branch line close to the mouth enables some deep hooked sea turtles to swallow and even expel the hooks without major harm. Therefore, gear removal seems to play a crucial role in turtle survival/mortality, and training vessel crews for onboard turtle management is essential for improving turtle survival at sea and maximizing possible positive effects of gear change. It is clear that circle hooks baited with mackerel significantly reduce sea turtle incidental catch on the Portuguese pelagic swordfish longline fishery in the southern Atlantic. However, from the fisheries management point of view, it is essential to assess the consequences of such gear modifications in a wider scale, prior to implementing them. The human dimensions of such modifications also have to be addressed (Campbell and Cornwell, 2008) and the economic impacts in the fishery have to be considered. For instance, on some cases, possible reductions in the target species catches may occur (e.g. Largarcha et al., 2005; Báez et al., 2010; Domingo et al., 2012), while in other cases the reductions in the target species catches are balanced by the gains in other marketable species (e.g. Coelho et al., FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 48 Table 4.1 - Mean catch per unit effort (CPUE, kg/1000 hooks) with respective standard deviation in parentheses for the various hook-bait combinations. J = 10º offset 9/0 hook; G = non-offset 17/0 circle hook; GT = 10° offset 17/0 circle hook. P-values are from Mann-Whitney tests comparing bait types and the Kruskal-Wallis tests comparing hook styles. p<0.05 are in bold. NT = not tested. FAO code Squid Mackerel Comparisions (p-values) Species name J G GT J G GT Bait Hook SWO Xiphias gladius - Target species 1135,1 (±1246,6) 834,3 (±987,2) 916,1 (±940,8) 910,6 (±1243,6) 565,4 (±614,7) 612,1 (±640,9) <0,01 0,01 ALB Thunnus alalunga 24,2 (±56,6) 53,9 (±150,4) 32,8 (±77,8) 13,3 (±76,3) 33,6 (±118,4) 33,5 (±133,6) <0,01 0,03 YFT Thunnus albacares 30,4 (±114,3) 45,4 (±165,8) 67,9 (±182,9) 22,0 (±98,5) 16,0 (±104,3) 10,1 (±63,8) <0,01 0,58 BET Thunnus obesus 33,6 (±107,7) 35.0 (±126,4) 19,5 (±91,5) 6,4 (±53,9) 16,6 (±84,2) 16,6 (±89,6) <0,01 0,46 BUM Makaira nigricans 121,7 (±407,1) 136,2 (±577,5) 80,1 (±285,3) 112,1 (±468,2) 75 (±284,8) 84,3 (±408,5) <0,01 0,32 WHM Kajikia albida 33,0 (±63,8) 24,7 (±66,7) 24,2 (±61,0) 19,2 (±53,1) 19,0 (±54,6) 18,7 (±60,5) <0,01 0,09 LEC Lepidocybium flavobrunneum 22,6 (±59,4) 18,1 (±65,3) 10,2 (±46,6) 33,0 (±97,7) 15,8 (±46,2) 14,1 (±38,4) 0,35 <0,01 DOL Coryphaena hippurus 8,9 (±25,5) 5,8 (±18,6) 4,5 (±14,3) 6,6 (±23,6) 4,5 (±14,4) 2,7 (±11,5) 0,01 0,05 TAS Taractes asper 1,4 (±10,5) 2,4 (±12,9) 1,4 (±10,6) 0,8 (±6,6) <0,01 0,6 AMB Seriola dumerili 0,2 (±3,0) 0,1 (±1,2) 0,3 (±3,4) <0,01 0,78 OIL Ruvettus pretiosus 1,0 (±7,4) 1,1 (±8,0) 0,9 (±6,1) 0,3 (±4,3) 0,9 (±6,1) 0,62 0,14 POA Brama brama 0,1 (±0,9) 0,1 (±0,5) 0,56 NT SAI Istiophorus platypterus 0,5 (±6,0) NT NT SPF Tetrapturus pfluegeri 0,3 (±3,7) 0,3 (±3,6) NT 0,61 WAH Acanthocybium solandri 1,7 (±9,7) 1,6 (±7,5) 1,6 (±8,7) 2,2 (±13,6) 1,1 (±6,5) 1,2 (±7,8) 0,32 0,72 Total teleosts bycatch 279,1 (±123,8) 324,3 (±174,2) 242,3 (±101,3) 216,4 (±136,5) 181,9 (±93,9) 183,3 (±121,8) <0,01 0,38 BSH Prionace glauca 701,9 (±598,3) 1016,8 (±976,8) 904,0 (±879,8) 1355,8 (±1138,5) 1778,3 (±1350,4) 1648,2 (±1303,4) <0,01 <0,01 CCA Carcharhinus altimus 0,3 (±2,9) 0,5 (±6,7) 0,2 (±2,6) 0,32 0,37 LMA Isurus paucus 2,0 (±24,3) 3,8 (±25,6) 0,2 (±2,2) 2,4 (±28,1) 0,03 0,41 POR Lamna nasus 1,4 (±17,3) NT NT SMA Isurus oxyrinchus 68,6 (±141,0) 101,5 (±221,2) 99,9 (±228,9) 100,1 (±183,9) 121,6 (±247,0) 136,8 (±300,2) 0,17 0,85 Total elasmobranch bycatch 772,2 (±340,5) 1118,9 (±598,8) 1005,9 (±538,2) 1459,6 (±741,3) 1900,3 (±930,9) 1787,5 (±880,9) <0,01 <0,01 Total retained catch 2186,5 (±1733,0) 2277,5 (±1499,3) 2164,2 (±1344,7) 2586,6 (±1999,7) 2647,6 (±1443,2) 2582,8 (±1457.3) <0,01 0,11 FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 49 Non-parametric statistical comparisons indicated that the significant factor on the catch rates of blue marlin (Makaira nigricans), white marlin (Kajikia albida) and dolphinfish (Coryphaena hippurus) was the effect of bait type (higher with squid) while for escolar (Lepidocybium flavobrunneum) only hook style was found to be a significant factor (Table 4.1; Figure 4.5). The effects tended to be opposite for elasmobranch bycatch, with most species having higher catch rates with circle hooks baited with mackerel. Particularly, for the blue shark significant differences were observed between bait (Mann-Whitney: W=146660, p<0.01) and hook types (Kruskal-Wallis: Chi-square=13.29, df=2, p<0.01; Table 4.1; Figure 4.6). Nonparametric multiple tests for all-pairs comparisons, enhanced that significant higher CPUE for the swordfish and escolar were obtained with J hooks, while the opposite was observed for albacore and blue shark (higher CPUE with circle hooks). Figure 4.6 – Plot design of the observed catch per unit effort (CPUE, Kg/1000 hooks) by the different hook styles (J, G and GT) and bait types (Squid and Mackerel) for the three main elasmobranch bycatch species in the fishery: blue shark (Prionace glauca), shortfin mako (Isurus oxyrinchus) and longfin mako (Isurus paucus). The vertical lines represent the range of values while the horizontal line represents the overall mean across all observations. Bait and hook effect were only significant (*) for blue shark. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 50 Figure 4.5 - Plot design of the observed catch per unit effort (CPUE, Kg/1000 hooks) by the different hook styles (J, G and GT) and bait types (Squid and Mackerel) for four teleost bycatch species in the fishery: blue marlin (Makaira nigricans), white marlin (Kajikia albida), dolphinfish (Coryphaena hippurus) and escolar (Lepidocybium flavobrunneum). The vertical lines represent the range of values while the horizontal line represents the overall mean across all observations. Bait effect was significant (*) for all species with exception for escolar were only hook effect was significant. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 51 For the swordfish and blue shark (the two species with the highest catch rates) plots and univariate non-parametric statistical tests (Figure 4.3 and 4.6; Table 4.1) demonstrated that both hook style and bait type were significant covariates. To model swordfish and blue shark catch rates along with these two explanatory variables it was applied a tweedie distribution, which confirmed that both hook style and bait type were significant factors (Tables 4.2 A and B). The catch rates (CPUE) of swordfish for hook styles G and GT (circle hooks) were lower than with J-style hooks, by factors of 0.68 (95% CI between 0.57 and 0.81) and 0.74 (95% CI between 0.62 and 0.89), respectively (Table 4.2). This represents an estimate reduction in the catch rates of 32% and 26% when changing from J-styles to G and GT hooks, respectively. For the blue shark, when changing from J-style to circle hooks, catch rates increased by factors of 1.37 (95% CI between 1.19 and 1.56) and 1.24 (95% CI between 1.08 and 1.43) for G and GT hooks, representing an estimated increase of 37% and 24%, respectively. Comparing the catch rates in terms of bait type, when changing from squid to mackerel the catches of swordfish were lower by a factor of 0.72 and for blue shark the catches were higher by a factor of 1.82 with 95% CI between 1.63 and 2.04 (Tables 4.2 A and B). Table 4.2 - Coefficients for the swordfish (A) and blue shark (B) CPUE Tweedie GLM with the respective odds-ratios, considering the covariates hook style and bait type. The parameter estimation of the model, the standard errors (SE), the Wald Statistic (Wald) and the respective p-values are presented. For the odds-ratios the point estimate with the lower and upper 95% confidence intervals (CI) are listed. A - Swordfish Coefficients Odds-ratios Parameter Estimate SE Wald p-value Estimate Lower CI Upper CI (intercept) 7,09 0,07 100,7 <0,01 Hooh style G -0,39 0,09 -3,2 <0,01 0,68 0,57 0,81 Hook style GT -0,30 0,09 -2,8 <0,01 0,74 0,62 0,89 Bait Type mackerel -0,33 0,07 -2,9 <0,01 0,72 0,62 0,83 B – Blue shark Coefficients Odds-ratios Parameter Estimate SE Wald p-value Estimate Lower CI Upper CI (intercept) 6,55 0,08 80,9 <0,01 Hooh style G 0,37 0,11 3,4 <0,01 1,37 1,19 1,56 Hook style GT 0,25 0,11 2,3 <0,01 1,24 1,08 1,43 Bait Type mackerel 0,66 0,10 6,3 <0,01 1,82 1,63 2,04 FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 52 4.2.3. Size distribution of retained catch Overall, mean sizes for all the species studied were relatively similar between different hook styles and bait types. Significant differences were detected in size distribution when comparing hook styles for bigeye and yellowfin tuna. Using non-parametric multiple tests for all-pairs comparisons for bigeye tuna, it was detected that significant differences occurred between all hook styles, with J hooks capturing significantly larger specimens. For the yellowfin tuna, significant differences only occurred for the pairwise comparison J-G hooks, with higher mean size for circle hooks (Table 4.3 in Annex I; Figure 4.7). For swordfish and shortfin mako, mean size with circle hooks were slightly lower than those recorded for J hook catches (Table 4.3 in Annex I). When comparing bait type significant differences were found for the swordfish, yellowfin, albacore, blue shark and shortfin mako (Table 4.3 in Annex I; Figure 4.8). Yellowfin tuna was the only species where significant differences were detected between the two factors (hook style and bait type). Whereas the opposite (no significant differences) was observed for blue marlin. The mean size of tuna species caught with squid were slightly lower than that found when mackerel was used as bait. In contrast higher mean size for swordfish, blue marlin, blue shark and shortfin mako were registered when squid was used (Figure 4.8). Figure 4.7 - Boxplots with the size distribution (median, inter-quartil range, non-outliers range and outliers) for the species bigeye tuna (Thunnus obesus) and yellowfin tuna (Thunnus albacares) for each hook style. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 53 Figure 4.8 - Boxplots with the size distribution (median, inter-quartil range, non-outliers range and outliers) for the species swordfish (Xiphias gladius), yellowfin (Thunnus albacares), albacore (Thunnus alalunga), blue shark (Prionace glauca) and shortfin mako (Isurus oxyrinchus) for each bait type. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 54 4.2.4. Catch rates and mortality of discarded species In teleost discarded species no significant differences were observed on the catch rates when comparing hook styles, with exception for the snake mackerel (Gempylus serpens), as only for the lancetfish (Alepisaurus ferox) significant differences were detected in terms of bait type (Table 4.4). As observed with other large pelagic elasmobranches, the catch rates of bigeye thresher shark (Alopias superciliosus), hammerhead shark (Sphyrna zygaena) and manta rays (Myliobatidae) were higher when mackerel bait was used, but significant differences were only detected for the last two. For the hook style comparisons, significant difference was only detected for the pelagic stingray (Pteroplatytrygon violacea), which had the highest catch rates with J hooks (Table 4.4). For the species snake mackerel and pelagic stingray, results from the non-parametric multiple tests for all-pairs comparisons indicated that significant differences occurred between J hooks (higher catch rates) and circle hooks. Statistical analysis to the at-haulback mortality rates of the discarded species, detected significant differences between hook styles for lancetfish and bigeye tresher shark (Chisquare proportion tests: p<0.05, Table 4.5 in Annex I) with higher mortality rates in GT and G hook styles, respectively (Figure 4.9; Table 4.5 in Annex I). Regardless of the hook style used, most of the bigeye thresher sharks and lancetfish were dead at-haulback, while most of crocodile sharks (Pseudocarcharias kamoharai) were alive (Figure 4.9; Table 4.5 in Annex I). The statistical (Chi-square proportion test) analysis for the pelagic stingrays was not performed due to cells with zero values in the contingency table (100% alive at-haulback for every hook style). For all discarded species the Chi-square proportions tests (applying the Yates' continuity correction given that the contingency tables are of the 2x2 type), did not detect significant differences in the proportions of dead and alive specimens at-haulback with the two bait types (Chi-square proportion tests: p>0.05 on all cases). FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 55 Table 4.4 - Mean (standard deviation) catch per unit effort (CPUE, n/1000 hooks) of discarded species for the various hook-bait combinations. P-values refer to the Mann-Whitney tests comparing baits and the Kruskal-Wallis tests comparing hooks. J = 10º offset 9/0 hook; G = non-offset 17/0 circle hook; GT = 10° offset 17/0 circle hook. p<0.05 are in bold. NT = not tested. FAO code Squid Mackerel Comparisions (p-value) Species name J G GT J G GT Bait Hook ALX Alepisaurus ferox 0,50 (±1.59) 0,16 (±0.77) 0,13 (±0.78) 0,73 (±2.83) 0,42 (±1.75) 0,56 (±1.80) <0,01 0,08 CUP Cubiceps sp. 0,03 (±0.33) NT NT GSE Gempylus serpens 0,09 (±0.55) 0,07 (±0.50) 0,34 (±1.39) 0,08 (±0.58) 0,05 (±0.47) 0,13 0,01 LAG Lampris guttatus 0,05 (±0.33) 0,09 (±0.64) 0,08 (±0.58) 0,03 (±0.24) 0,01 (±0.17) 0,79 0,07 MOX Mola mola 0,05 (±0.47) 0,01 (±0.17) 0,03 (±0.33) 0,32 0,37 Total teleosts discards 0,65 (±0.79) 0,31 (±0.5) 0,24 (±0.45) 1,14 (±1.46) 0,52 (±0.84) 0,66 (±0.87) 0,08 <0,01 ALV Alopias vulpinus 0,01 (±0.17) 0,01 (±0.17) 0,01 (±0.17) 0,56 0,37 BTH Alopias superciliosus 0,17 (±0.82) 0,26 (±0.99) 0,24 (±1.09) 0,43 (±2.22) 0,17 (±0.95) 0,46 (±1.37) 0,23 0,33 FAL Carcharhinus falciformis 0,03 (±0.33) 0,03 (±0.33) 0,05 (±0.47) 1,00 0,17 ISB Isistius brasiliensis 0,03 (±0.33) 0,08 (±1.00) 1,00 NT MAN Myliobatidae 0,04 (±0.37) 0,05 (±0.53) 0,04 (±0.37) 0,16 (±1.13) 0,03 0,52 OCS Carcharhinus longimanus 0,03 (±0.24) 0,01 (±0.17) 0,03 (±0.33) 0,01 (±0.17) 0,05 (±0.41) 0,01 (±0.17) 0,74 0,71 PLS Pteroplatytrygon violacea 0,55 (±1.5) 0,03 (±0.24) 0,12 (±0.68) 0,42 (±1.69) 0,2 (±1.11) 0,32(±1.39) 0,95 <0,01 PSK Pseudocarcharias Kamoharai 0,3 (±1.27) 0,67 (±2.53) 0,52 (±1.80) 0,27 (±1.13) 0,67 (±2.53) 0,51 (±1.9) 0,89 0,3 SPZ Sphyrna zygaena 0,01 (±0.17) 0,05 (±0.41) 0,05 (±0.41) 0,15 (±0.89) <0,01 0,46 Total elasmobranch discards 1,09 (±0.75) 1,03 (±0.95) 0,95 (±0.77) 1,24 (±1.04) 1,29 (±1.07) 1,68 (±1.06) 0,09 0,17 Total discards 1,70 (±2.90) 1,30 (±3.01) 1,20 (±2.70) 2,38 (±4.58) 1,81 (±4.17) 2,34 (±3.90) 0,02 0,02 FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 56 Figure 4.9 – Mosaicplot of the percentage of alive versus dead specimens at-haulback for the most frequently discarded species: lancetfish (A. ferox), bigeye thresher shark (A. superciliosus), pelagic stingray (P. violacea), and crocodile shark (P. kamoharai). Hook effect was significant (*) for lancetfish and bigeye thresher shark. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 57 4.3. DISCUSSION The catch composition observed in this study was similar to those reported by other author´s for the same area of study (Anon., 2008; Sales et al., 2010; Domingos et al., 2012), where swordfish and blue shark comprised most of the retained catch. Significant higher swordfish CPUEs were found in western longitudes near the seamount chain “Rio Grande Rise”. There is limited information on physical characteristics of seamounts (e.g. minimum summit depth) that affect the abundance of pelagic fish, however depending on their physical characteristics and location, seamounts are an obstacle to flow, creating local currents and increasing upwelling around the seamount (White et al., 2007). Upwelling around seamounts can bring nutrients from the deeper ocean to enhance primary productivity, supporting a variety of life (White et al., 2007). Grubbs et al. 2002 hypothesized that pelagic fish aggregate at seamounts due to this enhanced primary productivity. In addition to more prey availability, pelagic species may aggregate at seamounts for spawning and nursery habitat (Allain et al., 2006). The results of increased catches within these longitudes may be explained not only by seamounts but also by other oceanographic considerations (e.g. SST, Subtropical Convergence, etc). For blue shark a significant increase catch rate was detected in eastern longitudes but no significant differences were detected within the range of latitude. Bigelow et al. (2009) in a research study addressed to analyze environmental effects on swordfish and blue shark catch rates, found that higher catch rates for blue shark were obtained in lower SST. In addition, the increased CPUEs of blue shark in eastern longitudes may also be related to other oceanographic factors (eg. thermal fronts and chlorophyll). The effects of environmental factors in the catches were not explored in the present study. Although, environmental influences on the distribution of fishery resources are important factors that should be used in fisheries management models (Carruthers et al., 2011). The main results, demonstrated that both hook style and bait type effects on the catch rates are species-specific, and that, the bait appears to be more important than hook style. For the swordfish, the primary target species of the Portuguese pelagic longline fishery, the results presented a significant decrease in catch rates, when switching from the traditional combination (J hooks baited with squid) to other combinations. Comparing the baseline reference level covariate J hook baited with squid, to the other levels of covariates, the catchability decreased, which is similar to the results reported by other authors (e.g. Watson et al., 2005; Anon., 2008; Sales et al., 2010; Coelho et al., 2012; Foster et al., 2012) for the Atlantic waters. Our results may be explained by the morphology of the circle hooks, with the point on circle hooks turned in towards to the hook shank, so the gap width (distance FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 64 GENERAL CONCLUSIONS This study represents advancement in knowledge of the use of circle hook styles and bait types as a mitigation measure in the reduction of the incidental bycatch of sea turtles in the Atlantic Ocean, particularly in the South hemisphere. It also contributed to the increase knowledge of the effect of hook-bait combinations on the catch of target and non-target species caught by the Portuguese swordfish pelagic longline fishery. The results obtained, showed that by changing the traditional configuration J hook baited with squid to circle hooks baited with mackerel, resulted in a reduction of BPUE on both species of sea turtles (Loggerhead and Leatherback). However, bait seems to have more influence on the bycatch reduction then the hook style (significant differences only occurred changing from J-style to one of the circle hooks when using squid bait). Hooking location was species-specific, with most leatherbacks hooked by the flippers or entangled, while loggerheads were mostly hooked by the mouth and esophagus. Mortality was also species-specific and reflected the hooking location, with most of the specimens caught by the flippers being alive, while the specimens hooked in the mouth and esophagus having higher percentages of dead specimens. The urgent need to address the bycatch of sea turtles, make of the technical measures, like the use of circle hooks and mackerel bait a very attractive management strategy. However, the behavior of the species caught, can be different from region to region, even when dealing with the same species, thus caution must be taken when interpreting these results. Technological changes with the aim of reducing the catch of undesirable species also affect the capture of target species in ways that vary depending on the region. Longline management strategies must not only be effective in reducing bycatch, but also be commercially viable. Our results indicate that hook and bait modifications (changing from J to circle hooks and from squid to mackerel) reduce bycatch of vulnerable sea turtles, but simultaneously result in lower catch rates of swordfish, the main target species of this fishery. However an increase catch of other marketable species (blue shark and shortfin mako) was observed. In this study no economic analysis were addressed so it was not possible to infer if the highest catches of elasmobranchs balanced the losses of swordfish lower catches. It should be noted that shark populations resilience to fishery exploitation, due to low fecundity and late age at maturity, is much lower than high fecundity species such as swordfish and Thunnus spp. Thus, predicting the effects of a gear change (circle hooks FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 65 baited with mackerel) on shark populations is urgent because any irrational exploitation can lead to the eventual collapse of these populations. The differences in mean size between hook style and bait type for the seven most caught species were of a few cm FL. However when comparing hook styles significant differences were detected for bigeye (larger specimens with J hooks) and yellowfin tuna (larger specimens with circle hooks). In relation to bait type significant differences were found in swordfish, blue shark and shortfin mako, with larger specimens caught in squid bait and yellowfin and albacore tuna with larger specimens in mackerel bait. For the discarded species mortality at time of fishing gear retrieval was species-specific with the proportions of alive versus dead varying significantly by hook style for the lancetfish and bigeye thresher shark. In order to promote a sustainable management of the pelagic longline fisheries, complementary studies must be undertaken in the near future. Although the present study provided new knowledge on the effect of different hook styles and bait types in sea turtle’s bycatch and target and non-target species in the South Atlantic Ocean, unanswered questions persist, such as the influence of environmental parameters in bycatch rates, therefore requiring further studies. Different results on the use of circle hooks have been reported, which mean that the effectiveness of each mitigation measure may be fisheryspecific, thus successful management will require a combination of alternative measures. The cooperation of stakeholders is essential to the biological, economic, and social success of fisheries management and regulation, and such is the case with the use of circle hooks. The most effective means of transferring the technology to date is not through management measures that are difficult to enforce, but through direct outreach to the fishermen´s in their fishing operations. Therefore, the expansion of mechanisms to engage stakeholders in circle hook education as well as participation in decision making processes has great potential to enhance effective and efficient incorporation of the technology. A holistic approach to bycatch reducing technology that engages social scientists as research collaborators should develop. This will increase compliance with management measures and ultimately increase the conservation benefits of circle hook use in pelagic fisheries. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 66 REFERENCES Afonso, A.S., Hazin, F.H.V., Carvalho, F., Pacheco, J.C., Hazin, H., Kerstetter, D.W., Murie, D., Burgess, G.H., 2011. Fishing gear modifications to reduce elasmobranch mortality in pelagic and bottom longline fisheries off Northeast Brazil. Fisheries Research. 108: 336–343. Afonso, A., Santiago, R., Hazin, H. and Hazin, F.H.V. (2012). Shark bycatch and mortality and hook bite-offs in pelagic longlines: Interactions between hook types and leader materials. Fisheries Research. 131-133: 9-14. Allain, V., Kirby, D., Kerandel, J. (2006). Seamount Research Planning Workshop Final Report. Report of the Seamount Research Planning Workshop Held at the Secretariat of the Pacific Community, Noumea, New Caledonia, 20-21 March 2006. Secretariat of the Pacific Community, Noumea, New Caledonia. Anon. (2008). Field study to assess some mitigation measures to reduce bycatch of marine turtles in surface longline fisheries. Report Reference No. FISH/2005/28A - Service Contract SI2.439703 ‘Assessment of turtle bycatch’. 217pp. Anon. (2009). Report of the 2009 Atlantic swordfish stock assessment session. ICCAT, Madrid, September 7 to 11. 11pp. Anon. (2012). Report of the Standing Committee on research and statistics (SCRS). ICCAT, Madrid, October 1 to 5. 300pp. Azevedo, M. (1990). Alguns aspectos da dinâmica populacional de espadarte (Xiphias gladius, L.) nas águas continentais Portuguesas. Dissertação das provas públicas para Assistente de Investigação, Instituto Nacional de Investigação das Pescas, Lisboa, 130pp. Báez, J.C., Real, R., Mácias, D., Serna, J.M., Bellido, J.J. and Camiñas, J.A. (2010). Captures of swordfish Xiphias gladius Linnaeus 1758 and loggerhead sea turtles Caretta caretta (Linnaeus, 1758) associated with different bait combinations in the Western Mediterranean surface longline fishery. Journal of Applied Ichthyology. 26: 126–127. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 67 Bertrand, A., Josse, E., Bach, P., Gros, P., and Dagorn, L. (2002). Hydrological and trophic characteristics of tuna habitat: consequences on tuna distribution and longline catchability. Canadian Journal of Fisheries and Aquatic Sciences. 59: 1002–1013. Beverly, S., Chapman L. and Sokimi W. (2003). Horizontal longline fishing methods and techniques: a manual for fishermen. Secretariat of the Pacific Community. 130 pp. Bigelow, K.A., Boggs, C.H., He, X., (1999). Environmental effects on swordfish and blue shark catch rates in the US North Pacific longline fishery. Fisheries Oceanography. 8: 178198 Bigelow, K.A., Kerstetter, D.W., Dancho, M.G., Marchetti, J.A. (2012). Catch rates with variable strength circle hooks in the Hawaii-based tuna longline fishery. Bulletin of Marine Science. 88(3): 425–447. Bjorndal, K.A. (1997). Foraging ecology and nutrition of sea turtles. In The Biology of Sea Turtles, Lutz PL, Musick JA (eds). CRC Press: New York; 199–231. Brothers N., Cooper J., and Løkkeborg S. (1999). The incidental catch of seabirds by longline fisheries: worldwide review and technical guidelines for mitigation. Rome: FAO Fisheries Circular no. 937. Bugoni, L., Mancini, P.L., Monteiro, D.S., Nascimento, L., Neves, T.S. (2008). Seabird bycatch in the Brazilian pelagic longline fishery and a review of capture rates in the southwestern Atlantic Ocean. Endangered Species Research. 5: 137−147. Campana, S., Joyce, W. and Manning, M.J., (2009). Bycatch and discard mortality in commercially caught blue sharks Prionace glauca assessed using archival satellite pop-up tags. Marine Ecology Progress Series. 387: 241–253. Campbell, R., Whitelaw, W. and McPherson, G. (1997). Domestic Australian Longline Fishing Methods and the Catch of Tunas and Non-Target Species off North Eastern Queensland. Western Pacific Yellowfin Research Group 7. 28 pp. Campbell, L.M and Cornwell, M.L. (2008). Human dimensions of bycatch reduction technology: current assumptions and directions for future research. Endangered Species Research. 5: 325–334. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 68 Candy S.G. (2004). Modeling catch and effort data using generalized linear models, the tweedie distribution, random vessel effects and random stratum-by-year effects. CCAMLR Science. 11:59–80. Carranza, A., Domingo, A. and Estrades, A. (2006). Pelagic longlines: A threat to sea turtles in the Equatorial Eastern Atlantic. Biological Conservation 131: 52-57. Casale, P. (2011). Sea turtle by-catch in the Mediterranean. Fish and Fisheries 12: 299–316. Carruthers, E.H., Schneider, D.C. and Neilson, J.D. (2009). Estimating the odds of survival and identifying mitigation opportunities for common bycatch in pelagic longline fisheries. Biological Conservation. 142: 2620–2630. Carruthers, E.H., Neilson, J.D., Smith, S.C., (2011). Overlooked bycatch mitigation opportunities in pelagic longline fisheries: soak time and temperature effects on swordfish (Xiphias gladius) and blue shark (Prionace glauca). Fisheries Research. 108: 112–120 Chaloupka, M., Parker, D., Balazs, G. (2004). Modelling post-release mortality of loggerhead sea turtles exposed to the Hawaii-based pelagic longline fishery. Marine Ecology Progress Series. 280: 285–293. Chuenpagdee, R. Morgan, L.E., Maxwell, S.M., Norse, E.A., Pauly, D. (2003). Shifting gears: assessing collateral impacts of fishing methods in US waters. Frontiers in Ecology and the Environment. 1: 517–524. Coelho, R., Santos, M.N., Amorim, S. (2012). Effects of hook and bait on targeted and bycatch fishes in an equatorial Atlantic pelagic longline fishery. Bulletin of Marine Science, 88 (3): 449-467. Compagno, L.J.V. (2001). Sharks of the world. An annotated and illustrated catalogue of shark species known to date. Volume 2. Bullhead, Mackerel and Carpet Sharks (Heterodontiformes, Lamniformes and Orectolobiformes). FAO, Rome. Cooke, S.J. and Suski C.D. (2004). Are circle hooks an effective tool for conserving sea and freshwater recreational catch-and-release fisheries? Aquatic Conservation: Marine and Freshwater Ecosystems. 14: 299–326. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 69 Crowder, L.B., and Myers, R.A. (2001). A Comprehensive Study of the Ecological Impacts of the Worldwide Pelagic Longline Industry. First Annual Report to the Pew Charitable Trusts. Curran, D., Bigelow, K. (2011). Effects of circle hooks on pelagic catches in the Hawaiibased tuna longline fishery. Fisheries Research. 109: 265–275. De Metrio, G., Megalofonu, P. (1988). Mortality of marine turtles (Caretta caretta L., and Dermochelys coriacea L.) consequent to accidental capture in the Gulf of Taranto. Rapports et ProcessVerbaux des Reunions Conseil International pour l’Exploration de la Mer Mediterranee, Monaco. Diaz, G.A. (2008). The Effect of Circle Hooks and Straight (J) Hooks on the Catch Rates and Numbers of White Marlin and Blue Marlin Released Alive by the U.S. Pelagic Longline Fleet in the Gulf of Mexico. North American Journal of Fisheries Management. 28: 500–506. Domingo, A., Bugoni, L., Prosdocimi, L., Miller, P., Laporta, M., Monteiro, D.S., Estrades, A. And Albareda, D. (2006). The impact generated by fisheries on Sea Turtles in the Southwestern Atlantic. WWF Progama Marino para Latinoamérica y el Caribe, San José, Costa Rica. Domingo, A., Barceló, C., Swimmer, Y., Pons, M. and Miller, P. (2009). Anzuelos circulares vs. anzuelos “J” en la flota palangrera uruguaya. Collective Volumes of Scientific Papers. ICCAT, 64(7): 2427-2442. Domingo, A., Pons, M., Jiménez, S., Miller, P., Barceló, C., Swimmer, Y. (2012). Bulletin of Marine Science. 88(3): 499–511. Dunn, P.K. (2010). Tweedie: tweedie exponential family models. R package version 2.0.7. Accessed on 10 June 2012. Available in: http://www.r-project.org/. Epperly, S.P., Watson, J.W., Foster, D.G. and Shah, A.K. (2012). Anatomical hooking location and condition of animals captured with pelagic longlines: the grand banks experiments 2002-2003. Bulletin of Marine Science 88: 513-527. FAO (1995). Code of Conduct for Responsible Fisheries. Rome, FAO Fisheries Department. 41pp. Accessed on 22 December 2012. Available in: http://www.fao.org/ FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 70 FAO. (2009). Guidelines to reduce sea turtle mortality in fishing operations. FAO Fisheries Department, Rome. Accessed on 22 December 2012. Available in: http://www.fao.org/ FAO. (2012). FAO Fisheries Department, Fishery Information, Data and Statistics Unit. FISHSTAT Plus: Universal software for fishery statistical time series. Version 2.3. Accessed on 23 December 2012. Available in: http://www.fao.org/fishery/statistics/software/fishstat/en Foster D.G., Epperly S.P., Shah A.K. and Watson J.W. (2012). Evaluation of hook and bait type on the catch rates in the western north atlantic ocean pelagic longline fishery. Bulletin of Marine Science. 88: 529–545. Fox J, Andronic L., Ash M., Boye T., Calza S., Chang A., Grosjean P., Heiberger R., Jay Kerns G., Lancelot R., et al. (2011). Rcmdr: R Commander. R package version 2.3. Accessed on 06 June 2012. Available in: http://CRAN.R-project.org/ package=Rcmdr Fox, J. and Weisberg S. (2011). An {R} companion to applied regression. 2nd ed. Thousand Oaks CA: Sage. Accessed on 06 June 2012. Available in: http://cran.r-project.org/web/ packages/car/ Gardner, B., Sullivan, P.J., Epperly, S. and Morreale, S.J. (2008). Hierarchical modeling of bycatch rates of sea turtles in the western North Atlantic. Endangered Species Research, 5: 279–289. Gao, X., Alvo, M., Chen J. and Li, G. (2008). Nonparametric Multiple Comparison Procedures for Unbalanced One-Way Factorial Designs. Journal of Statistical Planning and Inference. Volume 138, 2574–2591. Gibson, C.D. (1998). The broadbill swordfishery of the Northwest Atlantic: an economic and natural history. Camden, Maine: Ensign Press. 139 pp. Gilman E., Zollett E., Beverly S., Nakano H., Davis K., Shiode D., Dalzell P. and Kinan I. (2006). Reducing sea turtle by-catch in pelagic longline fisheries. Fish and Fisheries 7(1): 223. Gilman, E., Clarke, S., Brothers, N., et al. (2007a). Shark Depredation and Unwanted Bycatch in Pelagic Longline Fisheries: Industry Practices and Attitudes, and Shark FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 71 Avoidance Strategies. Western Pacific Regional Fishery Management Council, Honolulu, USA. Gilman, E., Kobayashi, D., Swenarton, T., Brothers, N., Dalzell, P. and Kinan-Kelly, I. (2007b). Reducing sea turtle interactions in the Hawaii-based longline swordfish fishery. Biological Conservation 139: 19-28. Gilman, E., Kobayashi, D. and Chaloupka, M. (2008). Reducing seabird bycatch in the Hawaii longline tuna fishery. Endangered Species Research. 5:309–323. Gilman, E., Clarke, S., Brothers, N., et al., (2008). Shark interactions in pelagic longline fisheries. Mar. Policy 32, 1–18. Fisheries. Marine Policy 32: 1-18. Gilman, E. (2011). Bycatch governance and best practice mitigation technology in global tuna fisheries. Marine Policy. 35(5): 590-609. Godin, A.C., Carlson, J.K. and Burgener, V. (2012). The effect of circle hooks on shark catchability and at-vessel mortality rates in longlines fisheries. Bulletin of Marine Science. 88(3): 469-483. Graves, J.E., Horodysky, J.E., and Kerstetter, D.W. (2012). Incorporating circle hooks into Atlantic pelagic fisheries: case studies from the commercial tuna/swordfish longline and recreational billfish fisheries. Bulletin of Marine Science. 88(3): 411-422. Grubbs, R. D., Holland, K. N., Itano, D. (2002). Comparative trophic ecology of yellowfin and bigeye tuna associated with natural and man-made aggregation sites in Hawaiian waters. 15th Meeting of the Standing Committee on Tuna and Billfish, SCTB15, Honolulu, Hawai'i, 22-27 July 2002. YFT-6, 1-21. Guilherme R. G., Selene L., Roberto G., Alberto F. A. (2012). Stomach contents analysis of swordfish (Xiphias gladius) caught off Southern Brazil: A bayesian analysis. Collective Volume of Scientific Papers. ICCAT, 68(4): 1594-1600. Halpern, B.S., Walbridge S., Selkoe K.A. et al. (2008) A global map of human impact on sea ecosystems. Science 319, 948-952. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 72 Hampton, J., Bigelow, K., and Labelle, M. (1998). Effect of longline fishing depth, water temperature and dissolved oxygen on bigeye tuna (Thunnus obesus) abundance indices. Oceanic Fisheries Programme, Secretariat of the Pacific Community, New Caledonia. 18 pp. Hazin, H. (2006). Influência das variáveis oceanográficas na dinâmica populacional e pesca de espadarte, Xiphias gladius, Linnaeus 1758, capturado pela frota brasileira. Tese de doutoramento, ramo Ciências do Mar. Universidade do Algarve, Faro, Portugal. 216pp. Hillestad, H.O., Richardson J.I., McVea C. Jr. and Watson J.M. Jr. (1995). Worldwide incidental capture of sea turtles. In Biology and conservation of sea turtles. Revised edition. Bjorndal KA (ed). Smithsonian Institute Press Washington D.C.; 489–495. Horodysky, A.Z. and J.E. Graves. (2005). Application of pop-up satellite archival tag technology to estimate post release survival of white marlin (Tetrapturus albidus) caught on circle and straight-shank (“J”) hooks in the western North Atlantic recreational fishery. Fishery Bulletin. 103(1): 84-96. IUCN. 2012. IUCN Red List of Threatened Species. Version 1/2012. Accessed on 04 July of 2012. Available in: http://www.iucnredlist.org Jiménez, S., Domingo, A. and Brazeiro, A. (2009). Seabird bycatch in the southwest Atlantic: interaction with the Uruguayan pelagic longline fishery. Polar Biol. 32: 187–196. Julian, F. and Beeson, M. (1998). Estimates of marine mammal, turtle, and seabird mortality for two California gillnet fisheries: 1990–1995. Fisheries Bulletin. 96: 271–284. Kerstetter, D.W. and J.E. Graves. (2006). Survival of white marlin (Tetrapturus albidus) released from commercial pelagic longline gear in the western North Atlantic. Fishery Bulletin. 104(3): 434-444. Konietschke, F. (2012). nparcomp: Perform multiple comparisons and compute simultaneous confidence intervals for the nonparametric relative contrast effects. R package version 2.0. Accessed on 19 February 2013. Available in: http://cran.r-project.org/web/packages/ nparcomp Largarcha, E., Parrales, M., Rendón, L., Velasquez, V., Orozco, M. and Hall M. (2005). Working with the Ecuadorian fishing community to reduce the mortality of sea turtles in FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 73 longlines: the first year March 2004–March 2005. Western Pacific Regional Fishery Management Council. Lewison, R.L., Crowder, L.B., Read, A.J., and Freeman, S.A., (2004). Understanding impacts of fisheries bycatch on sea megafauna. Trends in Ecology and Evolution. 19: 598– 604. Lewison, R.L. and Crowder, L.B. (2007). Putting longline bycatch of sea turtles into perspective. Conservation Biology. 21: 79-86. Lilliefors, H.W. (1969). On the Kolmogorov-Smirnov test for the exponential distribution with mean unknown. Journal of the American Statistical Association. 64: 387–389. Lokkeborg, S. (2004). A review of existing and potential longline gear modifications to reduce sea turtle mortality. Pages 165 – 169 in: Papers presented at the Expert Consultation on Interactions between Sea Turtles and Fisheries within an Ecosystem Context. FAO Fisheries Report No. 738, Supplement. Food and Agriculture Organization, Rome. 238 pp. Løkkeborg, S. and Pina, T. (1997). Effects of setting time, setting direction and soak time on longline catch rates. Fisheries Research. 32: 213-222. Lutcavage, M.E., Plotkin, P., Witherington, B. and Lutz, P. (1996). Human impacts on sea turtle survival. In The biology of sea turtles, Lutz PL, Musick JA (eds). CRC Press: New York; 387–409. Magnuson, J.J., Bjorndal, K.A., Dupaul, W.A., Graham, G.L., Owens, F.W., Peterson, C.H., Pritchard, P.C.H., Richardson, J.I., Saul, G.E., West, C.W. (1990). Decline of the sea turtles: causes and prevention. 1st ed. National Academy Press: Washington D.C. Mejuto, J. and Hoey, J. (1991). An approach to a stock hypothesis for the swordfish (Xiphias gladius) of the Atlantic Ocean. Collective Volume Scientific Papers. ICCAT, 35(2): 482-501. Ménard F., Labrune C., Shin Y., Asine A., Bard F. (2006). Opportunistic predation in tuna: a sizebased approach. Marine Ecology Progress Series. 323: 223–231. FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 80 ANNEXES Annex I – Tables Table 1.1 – General parameters and measurements of the Portuguese pelagic longline fishing gear (American-style). Gear Characteristics Swordfish Pelagic Longline (“American-style”) Mainline material Nylon monofilament (3.2 – 3.6 mm Ø) Mainline lenght 35 - 60 nautical miles (~ 60 - 110 km) Mainline deployment Shooter Distance between buoys 400 - 700 m Distance from buoy to mainline (floatline length) 12 – 18 m Branch line length 8 - 20 m Branch line material Nylon monofilament (1.8 - 2,0 mm Ǿ) Wire leader (trace) used on branch lines? No when targeting swordfish and tuna Yes when targeting sharks Number of hooks per set 800 - 1500 Hooks per basket (between buoys) 4 – 8 Maximum depth of mainline when set 50 m Typical min depth of hooks 20 m Typical max depth of hooks 70 m Common mainline sink rate 40 m in 8 seconds Common timing of set, soak, and haul Set: 5 PM (for 7 hours) Soak: 7 hours Haul: 6 AM (for 10 hours) Lightstick use Battery Hook setting interval 8 - 16 seconds (70-100 m) Radio beacons buoys Yes Hook type “Stell” 10/0 - ref. 39960 J16/0 Ancora (no offset) - ref. 722 J17/0 Ancora (no offset) - ref. 722 J18/0 Ancora (no offset or 10º offset) - ref. 722 J Mustad 9 (10º offset) Weight size and location 60 - 80 g. at connection of mainline with float line. 65 - 75 g. swivel at top of leader Clip size and type 320 - 350S Bait type Squid (Ilex spp.) and mackerel (Scomber spp.) FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 81 Table 4.3 - Mean (standard deviation) size (LJFL and FL, in cm) for the species swordfish (Xiphias gladius), bigeye (Thunnus obesus), yellowfin (Thunnus albacares), albacore (Thunnus alalunga), blue marlin (Makaira nigricans), blue shark (Prionace glauca) and shortfin mako (Isurus oxyrinchus). P-values refer to the Mann-Whitney tests comparing sizes with different baits and the KruskalWallis tests comparing sizes with different hooks. J = 10º offset 9/0 hook; G = non-offset 17/0 circle hook; GT = 10° offset 17/0 circle hook. p<0.05 are in bold. Bait type Hook style Comparisions (p-value) Species name Squid Mackerel J G GT Bait Hook Xiphias gladius 166,5 (±35,0) 162,2 (±36,1) 165,0 (±36,7) 163,7 (±34,5) 164,9 (±35,1) <0,01 0,42 Thunnus obesus 124,2 (±30,6) 126,2 (±35,8) 144,5 (±26,6) 128,7 (±30,1) 110,5 (±30,6) 0,59 <0,01a Thunnus albacares 155,3 (±10,1) 158,6 (±12,3) 154,4 (±9,1) 159,2 (±9,1) 155,0 (±12,3) 0,01 0,01b Thunnus alalunga 105,7 (±9,4) 111,7 (±5,9) 106,6 (±9,6) 108,0 (±8,3) 109,0 (±8,7) 0,01 0,13 Makaira nigricans 229,1 (±36,8) 223,0 (±28,3) 222,8 (±29,3) 231,9 (±35,8) 224,1 (±34,4) 0,86 0,15 Prionace glauca 197,8 (±26,5) 194,8 (±30,6) 195,1 (±29,8) 196,9 (±29,3) 195,4 (±28,7) <0,01 0,08 Isurus oxyrinchus 176,8 (±35,6) 167,7 (±38,1) 175,2 (±33,6) 168,5 (±38,6) 171,4 (±38,3) <0,01 0,10 a Non-parametric multiple tests for all-pairs comparisons. Significant difference between all hooks. b Non-parametric multiple tests for all-pairs comparisons. Significant differences between G and the other two style of hooks (J and GT). FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 82 Table 4.5 – Percentage of alive versus dead specimens at-haulback for the most frequently discarded species: lancetfish (A. ferox), bigeye thresher shark (A. superciliosus), pelagic stingray (P. violacea), and crocodile shark (P. kamoharai). The statistical comparisons refer to Chi-square tests for contingency tables. J = 10º offset 9/0 hook, G = non-offset 17/0 circle hook, GT = 10° offset 17/0 circle hook. p<0.05 are in bold. NT= not tested J G GT Comparisions Species name n % dead % alive n % dead % alive n % dead % alive Chisquare pvalue Alepisaurus ferox 89 82,02 17,98 47 57,45 42,55 48 87,50 12,50 14,48 <0,01 Alopias superciliosus 43 37,21 62,79 34 67,65 32,35 52 53,85 46,15 7,17 0,03 Pteroplatytrygon violacea 72 0,00 100,00 17 0,00 100,00 33 0,00 100,00 NT NT Pseudocarcharias kamoharai 42 11,90 88,10 100,00 8,00 92,00 77 10,39 89,61 0,61 0,74 FCUP Effects of hook and bait on bycatch and target catches in a Southern Atlantic swordfish longline fishery 83 Annex II – List of fish species Table 2.2 – List of fish species, with FAO codes, common names, scientific names (with authority) and category. In the category, T-B refers to targeted bony fish species; BC-B to bony fish bycatch; BC-E to elasmobranch bycatch; D-B to bony fish discards and D-E to elasmobranch discards FAO Code Category Common name Scientific name (Authority) SWO T-B Swordfish Xiphias gladius (Linnaeus, 1758) BET BC-B Bigeye tuna Thunnus obesus (Lowe, 1839) YFT BC-B Yellowfin tuna Thunnus albacares (Bonnaterre, 1788) ALB BC-B Albacore Thunnus alalunga (Bonnaterre, 1788) AMB BC-B Greater amberjack Seriola dumerili (Risso, 1810) BUM BC-B Blue marlin Makaira nigricans (Lacepède, 1802) DOL BC-B Dolphinfish Coryphaena hippurus (Linnaeus, 1758) LEC BC-B Escolar Lepidocybium flavobrunneum (Smith, 1843) OIL BC-B Oilfish Ruvettus pretiosus (Cocco, 1833) POA BC-B Atlantic pomfret Brama brama (Bonnaterre, 1788) SAI BC-B Sailfish Istiophorus platypterus (Shaw 1792) SPF BC-B Longbill spearfish Tetrapturus pfluegeri (Robins & de Sylva, 1963) TAS BC-B Rough pomfret Taractes asper (Lowe, 1843) WAH BC-B Wahoo Acanthocybium solandri (Cuvier, 1832) WHM BC-B White marlin Kajikia albida (Poey, 1860) BSH BC-E Blue shark Prionace glauca (Linnaeus, 1758) CCA BC-E Bignose shark Carcharhinus altimus (Springer, 1950) LMA BC-E Longfin mako Isurus paucus (Guitart, 1966) POR BC-E Porbeagle Lamna nasus (Bonnaterre, 1788) SMA BC-E Shortfin mako Isurus oxyrinchus (Rafinesque, 1810) ALX D-B Lancetfish Alepisaurus ferox (Lowe, 1833) CUP D-B Cubiceps sp GSE D-B Snake mackerel Gempylus serpens (Cuvier, 1829) LAG D-B Opah Lampris guttatus (Brünnich, 1788) MOX D-B Sunfish Mola mola (Linnaeus, 1758) ALV D-E Common thresher Alopias vulpinus (Bonnaterre, 1788) BTH D-E Bigeye thresher Alopias superciliosus (Lowe, 1841) FAL D-E Silky shark Carcharhinus falciformis (Müller & Henle, 1839) ISB D-E Cookie cutter shark Isistius brasiliensis (Quoy & Gaimard, 1824) MAN D-E Devil and manta rays Myliobatidae (family) (Bonaparte, 1838) OCS D-E Oceanic whitetip Carcharhinus longimanus (Poey, 1861) PLS D-E Pelagic stingray Pteroplatytrygon violacea (Bonaparte, 1832 PSK D-E Crocodile shark Pseudocarcharias kamoharai (Matsubara, 1936) SPZ D-E Smooth hammerhead Sphyrna zygaena (Linnaeus, 1758)